Full text of "Fire Assaying"

See other formats

OU160177

W
>

OSMANIA UNIVERSITY LIBRARY

Call No. ^ /f
Author

This book should be returned on or before tlw date l&st marked below,

METALLURGICAL TEXTS

FIRE ASSAYING

METALLURGICAL TEXTS

A SERIES OF TEXTBOOKS OUTLINED BY THE

FOLLOWING COMMITTEE

BRADLEY STOUGHTON, Chairman and Consulting
Editor

Dean Emeritus of Engineering, Professor of
Metallurgy, Lehigh University.

ERNEST A. HERSAM,

Professor of Metallurgy, University of Cali-
fornia.

ZAY JEFFRIES,

Consulting Metallurgist, Cleveland.

DORSEY A. LYON

351 California Street, San Francisco, Calif.

GEORGE B. WATERIIOUSE,

Professor of Metallurgy, Massachusetts In-
stitute of Technology.

ALBERT E. WHITE,

Director of Research, University of Michigan.

FIRE ASSAYING

ORSON CUTLER SHEPARD

Associate Professor of Mining and Metallurgy, Stanford University;

Associate Professor of Metallurgy (Visiting}, Massachusetts

Institute of Technology

AXD

WALDEMAR F. DIETRICH

Teacher of Mining and Metallurgy, Sacramento Junior College

FIRST EDITION
SECOND IMPRESSION

McGRAW-HILL BOOK COMPANY, INC.

NEW YORK AND LONDON
1940

PRINTED IN THE UNITED STATES OF AMERICA

parts thereof, may not be reproduced

in any form without permission of

the publishers.

THE MAPLE PRESS COMPANY, YORK, PA.

PREFACE

The objectives of this book are to present the subject of fire
assaying in such a manner as to serve the following needs: (1) as
a self-study text of the practical art of assaying for individuals
who may not have had a college course in chemistry; (2) as a
text for technical institute students of assaying who do not
intend to follow a complete academic curriculum in metallurgy;
(3) as a reference for practicing assayers ; and (4) as a college text
for the student of metallurgy.

For the first and second of the above objectives, the book is
arranged so that a working knowledge of the practice of fire
assaying may be obtained without the need for a comprehensive
understanding of the physicochemical principles involved; yet
a sound foundation is laid for the intelligent application of
scientific principles in such matters as crucible charge calcula-
tions, cupellation and scorification procedures, and other essential
phases of the subject. For the practical assayer, Chap. VI,
dealing with the more theoretical aspects of assay fusions, may
be omitted entirely.

As a reference for practicing assayers, sufficient material is
presented so that workable procedures can be evolved for the
determination of gold and silver in almost any type of natural
or artificial material. Complete schemes are given for the
determination of the individual members of the platinum group
in order to bring together many scattered references that are
unlikely to be available except in the larger technical libraries.
A brief resume* of the fire assay of base metals is also given.
At the suggestion of assayers in the field, a short chapter on the
preparation of bullion from amalgam and cyanide precipitates
is presented, as well as appendixes listing assay equipment, assay
supplies, and the location of assay supply houses.

A foundation for the correlation of fire assaying with the
principles and processes of extractive pyrometallurgy is given
in Chap. VI. This should prove of interest to college instructors
who wish to utilize a course in fire assaying as a laboratory
introduction to pyrometallurgy.

vi PREFACE

Every effort has been made to serve the interests of simplicity,
clarity, and convenience. To this end the authors have intro-
duced step procedures wherever possible and have prepared
various tabular summaries to facilitate the selection of definite
procedures and the readjustment of faulty techniques to avoid
abnormalities.

The authors and their respective students have verified and
demonstrated the workability of all the important procedures
recommended in the book and furthermore have done a con-
siderable amount of research to eliminate ambiguous and indefi-
nite statements that have appeared in earlier works on fire
assaying.

In addition to specific acknowledgments throughout the book,
the authors wish to express their appreciation for the generous
cooperation of many practicing assayers, who have been con-
sulted from time to time by one or the other of the authors, and
for the assistance of Miss Mabel E. Clark in the preparation of the
manuscript.

ORSON CUTLER SHEPARD,
WALDEMAR F. DIETRICH.

STANFORD UNIVERSITY, CALIF.;
SACRAMENTO, CALIF.;
September, 1940.

CONTENTS

PAGE

PREFACE v

CHAPTER I
THE SCOPE OF FIRE ASSAYING 1

CHAPTER II
SAMPLING 14

CHAPTER III
WEIGHING 39

CHAPTER IV
CUPELLATION 46

CHAPTER V
PARTING 74

CHAPTER VI

THEORETICAL DISCUSSION OF ASSAY FUSIONS AND RELATED
SMELTING PROCESSES 85

CHAPTER VII
THE CRUCIBLE ASSAY ^ 121

^ CHAPTER VIII
THE SCORIFICATION ASSAY 165

CHAPTER IX
THE ASSAY OF BULLION FOR GOLD AND SILVER 172

CHAPTER X

THE ASSAY OF MATERIALS REQUIRING PRELIMINARY ACID
TREATMENT 184

CHAPTER XI
ASSAY OF SOLUTIONS FOR GOLD AND SILVER 194

CHAPTER XII

THE FIRE ASSAY FOR THE PLATINUM METALS 204

vii

viii CONTENTS

PAGE

CHAPTER XIII
FIRE ASSAY METHODS FOR BASE METALS 220

CHAPTER XIV
THE ACCURACY OF THE FIRE ASSAY FOR GOLD AND SILVER. 227

CHAPTER XV
THE PREPARATION OF GOLD AND SILVER BULLION FROM

AMALGAM AND CYANIDE PRECIPITATE 242

CHAPTER XVI

ASSAY EQUIPMENT AND SUPPLIES 248

APPENDIX A
MINIMUM EQUIPMENT FOR A SMALL ASSAY OFFICE 264

APPENDIX B
LIST OF ASSAY SUPPLIES AND REAGENTS 266

APPENDIX C
ASSAY SUPPLY HOUSES 268

INDEX 271

FIRE ASSAYING

CHAPTER I
THE SCOPE OF FIRE ASSAYING

The term "fire assaying" is applied to quantitative determina-
tions in which a metal or metals are separated from impurities by
fusion processes and weighed in order to determine the amount
present in the original sample. The metals recovered are those
which emerge from the fusions in the metallic state. For this
reason, fire assaying is especially adapted to the determination
of the noble metals : gold, silver, and the platinum group. Except
for silver-rich alloys, fire assaying is the only method in common
use for the determination of the noble metals. The method can
also be used for the determination of metals that are compara-
tively easy to reduce, such as copper, tin, lead, bismuth, and
antimony; but wet analytical methods for these metals are more
accurate and are generally preferred.

The materials that are subjected to fire assay originate in a
prospect, mine, or a metallurgical process, or in the recovery of
scrap metal. These materials may be received in gross form in
large lots, and the first consideration is to obtain a small sample
that will be representative of the entire lot. For brittle mate-
rials, this is accomplished by a series of progressive particle-size
reductions accompanied by the rejection of a part of the original
volume of material at each stage, according to the principles
discussed in Chap. II. The final assay sample, or pulp, consists
of J4 Ib. or more which is ground sufficiently fine so that small
portions can be weighed out for assay without introducing errors.
The sampling of solid metals and alloys is done by a systematic
method of drilling or sawing, conducted in such a manner as to
obtain a representative sample.

Objectives of Assaying. The most important outcomes of
assay results are as follows:

2 FIRE ASSAYING

1. Valuation of mining property.

2. Basis for buying or selling ores and metallurgical products.

3. Guide to prospecting and development of ore bodies.

4. Delimitation of boundaries of marginal ore.

5. Check against waste dilution in mining operations.

6. Control of average grade of ore mined and milled.

7. Investigation and control of metallurgical processes.

8. Accounting for all metals in process.

The degree of accuracy required in assaying depends upon
which of the above purposes is involved; and the assay er should
make a study of the permissible limits of error for each class of
assay and devise methods that are as rapid as possible for each
degree of accuracy required.

The highest degree of accuracy on individual samples is neces-
sary in buying and selling ores and metallurgical products.
On such materials most control assayers, acting for the buyer
or seller, take the average of two or preferably three deter-
minations. If the control assayers for buyer and seller do not
agree within predetermined limits, a duplicate pulp is submitted
to an umpire who usually reports the average of four or more
separate determinations (see Chap. XIV).

Samples for mine valuation require rigid precautions to avoid
gross or constant positive errors attributable to intentional or
accidental salting; but the individual assays do not require a
high degree of precision, since the final value of an ore body is
determined by the average of a large number of assays, and the
error involved in taking an individual sample is greater than the
normal errors of assaying. Hence, single assays of valuation
samples usually suffice and, as a precaution against the introduc-
tion of serious errors in individual assays, composite assays
should be made of suitable groups of samples (see Chap. XIV).

In order to guard against salting, 1 the assay er should frequently
check the precious-metal content of fluxes and reagents and
should insert blanks or previously assayed samples at irregular
positions in the sequence of valuation samples. All valuation
samples should be kept under constant surveillance by a responsi-
ble party during their progress through the assay office.

Some forms of salting can be detected by a study of the sam-
ples. Valuation samples should be regularly tested by: (1)

1 The extraneous introduction of precious metals.

THE SCOPE OF FIRE ASSAYING 3

panning and examination of the concentrate for unusual sub-
stances such as gold filings or amalgam, and (2) leaching with
water followed by an assay of the water to find soluble precious
metal salts.

The other purposes of assaying as outlined above do not
generally require a high degree of precision. Single determina-
tions, with ordinary precautions against errors, are usually
adequate particularly if periodical composites of important
assays are made or if a metallurgical check is available, as by
checking the metal content of the products against the feed to a
metallurgical process. Mill-tailings samples or smelter slags
require special care in assaying, since a small error will seriously
affect the metallurgical balance sheet on account of the tonnage
involved.

General Methods of Fire Assaying. Figure 1 is an idealized
flow sheet of the fire assay for gold and silver, with references to
the chapters in this text where detailed descriptions of each part
of the process are given.

The fire assay of ores and metallurgical products for noble
metals generally starts with a crucible fusion or a scorification to
slag the impurities and to collect the precious metals into a lead
button. In the crucible fusion, fluxes are added to form with
impurities in the ore a slag whose essential constituents are
soda (Na2O) or litharge (PbO), or both, and silica (Si02). Boric
oxide (BaOs), in the form of borax glass (NaAO?), is usually
added to replace or supplement the silica. In scorification the
amount of added fluxes is very small, the major flux being
litharge derived from the oxidation of lead which is mixed with
the ore at the outset. The size of the lead button is controlled
in the crucible process by control of the oxidation-reduction
reactions, so that just the right amount of lead is reduced from
the added litharge in the charge. In scorification the button
size is determined by the relationship between the size of scorify-
ing dish (scorifier) used and the amount of granulated lead added.
The crucible fusion allows the use of a larger sample of ore and is
applicable to a wider range of materials than scorification and
hence is the most generally used.

Prior to a fusion process, some materials such as those
containing metallic zinc or copper require a preliminary acid
treatment for removing undesirable impurities. Materials

FIRE ASSAYING

high in oxidizable impurities arc sometimes given a preliminary
roast.

Brittle materials in form and tonnage as received

HjO sample soon offer original
weighing of lot ( Chapter H)

Weighed, dried, and weighed
again for r^O determmatron

Rejects from sampling refurned
to t reatment process

.Primary assay sample cut
^(Chapter J)

^.Successive parficle site
and bulk reduction

Final assay sample, '/ 4 Ib
or more, dried, ground to
100- mesh or finer,
thoroughly mixed

Solid metals or
alloys.orsolidifi'ed
ladle sample of
molten metoils

Aqueous solutions,
as from the
cyanide p

;d by drill
wing (Chap '

le process

Sampled by drilling
or sowing ( Chapter J)

Representative
sample in form of
borings or filings

Assay portion weighed out
t

Cumulative sample
from continuous
process or dip
sample from batch

Representative sample
of solution

Assay portion by
volume measure

Optional (rarely used)
preliminary roast of
ores high in S.As.Sb
(Chapter M)

Materials with excessive
amounts of interfering
elements,especiolly Zn.Ni.Cu

Bullions rich in
gold or silver
(Chapter DC)

mica I ti

Preliminary acid treatment
of certain difficult
materials (Chapter X)

Chemical treatment
to collect Au+Ag
(ChapterXI)

Prepared for cupel lation (sometimes for
scarification ) by addition of granulated
lead and wrapped in lead foil

Assay fusion by alternative methods

a. Crucible fusion

(Chapters VI an
Fluxes added, plus PbO for
lead button, and oxidizers or
reducers for control of
buttbVi size

b. Scarification fusion
( Chapter W
Oxidizing fusion with
granulated lead and small
quantities of flux

Poured into molds or on flat plate, cooled until solid

djrsi

Slag rejected, oV saved for
corrected assay (Chapter XDf)

Lead button ,containing Auand Ag,
separated from slag and shaped
for convenient handling

l discarded,

Inauartation

Silver added if necessary for
parting ralio, 3 to 5 Ag to 1 Au
at any of points indicated
(Chapter V)

r Jd, or saved for

"corrected assay ( Chapter XET)

Washings poured into solution
containing chlorides, to
\r silv<

recover silver

Cupellation: Removal of
lead by oxidizing fusion

(Chapter flf)

^*- 1

Bead, cleaned, flattened,
weighed for Au 4 Ag
( Chapter HI)

.Parted in HNcJ, washed with H 2
( Chapter V)

Gold dried .ignited at red
heat and weighed on
bead balance

Silver calculateS by deducting
weight of Au and weight of
Au+Ag in reagents

FIG. 1. Idealized flow sheet of fire assay for gold and silver.

After the lead button is obtained, the lead is removed by an
oxidizing fusion during which the lead oxide is partly volatilized

THE SCOPE OF FIRE ASSAYING 5

but is mainly absorbed by the vessel in which the process is
conducted. This is known as "cupellation," and the vessel is
known as a " cupel."

The end product of cupellation is a bead containing the noble
metals. It is weighed on a bead balance, in order to obtain the
gross weight of the noble metals present, and is then treated with
acids which separate the metals from each other. This opera-
tion is known as "parting." In the absence of members of the
platinum group, parting involves a solution of the silver in nitric
acid, leaving the gold unattacked. The gold is then washed,
dried, and weighed; and the weight of silver is calculated by
subtracting the weight of gold from the combined weight.

A minimum ratio of three to five times as much silver as gold
is necessary for parting. If this proportion of silver is not
present in the original sample it must be added at one of the
alternative points in the process indicated on Fig. 1. This
procedure is known as "inquartation." The separation of
members of the platinum group involves special chemical
methods that arc described in Chap. XII.

Bullions rich in noble metals but containing minor amounts
of base-metal impurities are subjected directly to the cupellation
process. Silver in silver-rich alloys is frequently determined by
volumetric chemical methods, which are more accurate for this
type of material than arc fire methods.

Assay Equipment and Supplies. The choice of equipment for
an assay office is dependent upon the scale, scope, and perma-
nence of operations. At some small mines the assay er may have
numerous other duties and may devote only a few hours each
week to the actual assaying, yet a well-equipped assay office may
be justified for handling emergency assays and peak loads.

All the equipment and supplies for assaying can be purchased
from assay supply houses, and the first step in establishing an
assay laboratory is to send to one or more of the nearest supply
houses for illustrated catalogues describing and quoting the
equipment. A list of the leading distributors of assay equipment
and supplies throughout the world is given in Appendix C.

The equipment for an assay laboratory is divided into the
following categories:

1. Crushing, grinding, screening, and sampling equipment
(Chap. II).

6 FIRE ASSAYING

2. Assay furnace and accessory equipment (Chap. XVI).

3. Weighing equipment (Chap. III).

4. Parting equipment (Chap. V).

5. Chemical equipment, for special methods.

As an aid to the selection of equipment and to the local
fabrication of certain items, brief descriptions of typical equip-
ment are given in Chap. XVI and throughout the text. A
representative list of complete assay-office equipment, with
cost estimates, is given in Appendix A.

The supplies used in assaying, in relation to their applications,
are discussed throughout the text; a check list of the essential
supplies is given in Appendix B.

The layout of an assay office shoiild be planned with a view to
convenience, efficiency, and safety. The building should be
located as far as possible from heavy machinery and railroad
lines, to avoid the effect of vibrations on delicate weighing opera-
tions. Assay bead balances should be mounted on rigid founda-
tions separated from the building structure, and a separate
balance room is desirable. The sampling room, with its crushers
and other equipment, should be, if possible, in a room separate
from the rest of the laboratory, thus avoiding the possibility
of dust contamination of assay reagents and dust damage to
delicate balances. In large assay offices it is customary to
provide a separate room for parting and other chemical work,
an office, and separate storerooms for samples and for supplies.
If the assayer is responsible for the retorting of amalgam and
bullion melting, a separate room should be provided for that
purpose. The handling of rich materials in the assay office
requires great care to avoid salting.

Duties of Assayer. The primary duty of an assayer is to
assume responsibility for the accurate and prompt determination
of valuable metals in samples submitted to him. At plants
producing bullion he also is frequently charged with the duty
of preparing the bullion from amalgam or cyanide precipitates
and of ascertaining the weight and fineness of the bullion that
is shipped. In addition to these fundamental responsibilities,
the assayer at small mines may devote a part of his time to
other engineering activities such as surveying, mapping, mechani-
cal drafting, sampling, metallurgical calculations, metallurgical

THE SCOPE OF FIRE ASSAYING 7

testing, etc. In large organizations the assayer's duties are
more narrowly confined to assaying only.

It is important that the limits of the assayer's responsibility be
clearly defined with respect to the responsibilities of other persons
in the organization who have any connection with the source of
samples submitted to the assayer, and with the interpretation
and use of the assayer's reports.

Part of the organization chart of a highly specialized mine
and mill staff is given in Fig. 2, upon which the origin of assay
samples is superimposed. In this example the chief sampler
at the mine and the metallurgist at the mill are charged with the
primary responsibility of organizing the scope and methods of
all necessary routine sampling for the purpose of supplying data

General Manager

f~ 1 * - T 1

Geologist Mine Superintendent Assayer Mill Superintendent

Foreman-*! Chief ->-* Metallurgist Shift

Sampjer

Bosses
Millmen

T j t

Shift ' Sampling

Bosses *] Crew

| _Y ^

Solid lines show hierarchy of responsibility
Dotted lines show origin of assay samples
FIG. 2. Partial organization chart of a typical mining and milling company.

for the effective operation of the mine and mill. Their work is
subject to direct supervision by the mine and mill superintend-
ents, respectively, who may desire more or fewer samples or
special samples from time to time. At some properties the chief
sampler and the metallurgist particularly the latter are staff
officers responsible directly to the general manager. This
arrangement is satisfactory if the company is operating a number
of different mines and is conducting a campaign to acquire new
properties, but in any event the mine superintendent of each
mine and the mill superintendent of each separate milling unit
should have the authority to demand such sampling as they may
require.

The geologist, unless also serving as engineer in charge of
development, generally has no responsibility for routine samples
but is usually given blanket approval to submit to the assayer

8 FIRE ASSAYING

within specified limits any samples that he needs for his geological
studies. He may also suggest to the chief sampler, through the
mine superintendent, that the regular sampling crew take certain
samples. When new properties are under investigation, the
geologist or the examining engineer (who may be the chief
engineer, the chief sampler, or a subordinate) decides upon the
samples to be assayed.

The foreman and shift bosses in the mine are usually permitted
to take supplementary grab samples from newly blasted develop-
ment headings or from chutes that deliver ore to the main haulage
levels. These samples are usually transferred to the chief
sampler, who includes them with his regular samples. At mines
not employing a separate sampling crew, such samples by fore-
men and bosses may be the only ones used to guide mine develop-
ment and operation.

The routine mill sampling is usually carried out by automatic
devices or by the millmen, under the immediate supervision of the
shift boss acting upon the metallurgist's advice, as approved by
the mill superintendent. Millmen and mill shift bosses are
seldom permitted to take and submit assays without the express
approval of the superintendent or metallurgist.

At custom mills and smelters, in buying ores and concentrates,
the sampling procedure of shipments purchased is of vital
importance and is usually placed under the direction of a separate
official who is responsible directly to the manager.

The general manager is always in a position to order a revision
of the sampling and assaying procedure or to obtain special
samples personally or through subordinates. The mine and mill
superintendents are in a similar relationship to their respective
divisions.

It is customary for the assayer to submit a complete copy of
all assays to the general manager, mine superintendent, and mill
superintendent. Another copy may be divided so that the chief
sampler obtains a record of mine samples and the metallurgist
a copy of the mill data; or the offices may be arranged so that
these officials refer to the superintendent's copies. The geologist
and others are generally given a separate record for such assays
as they have personally submitted. Confidential reports to
the general manager only are sometimes requested. In some
organizations the routine reports for the manager are transmitted

THE SCOPE OF FIRE ASSAYING 9

to the chief clerk, who prepares such tabulations and sum-
maries as the general manager may require.

In the organization under discussion the assayer assumes no
responsibility whatsoever for the validity of the samples that
he receives. Most of this responsibility is divided between the
chief sampler and the metallurgist, each of whom has independent
means of testing the general reliability of the assayer's work.
On the other hand, serious sampling errors, in either the mine or
mill, will be disclosed by reliable assay results if the actual mine
or mill production does not check the calculated estimates. The
principal disadvantage of divided authority over sampling and
assaying is that the assayer has little opportunity to understudy
for more important positions and tends to become technically
narrowed to his specific duties.

The status of the assayer at small mines and properties in the
development stage merits some discussion. At some properties
he combines the functions of assayer with those of one or more
of the following: engineer, sampler, geologist, and metallurgist.
If he is in charge of all the sampling and engineering work, he
is placed in a position where his integrity is of paramount impor-
tance, since the management has no satisfactory independent
check of his reliability. In such a capacity, the assayer must
use exceptional precautions to ensure the honesty and accuracy
of his work; and he is justified in keeping a personal file of all
relevant data so as to safeguard himself against becoming
involved in fraudulent promotions.

One of the commonest tricks of fraudulent promotion involves
the appointment of a dishonest mine superintendent, who
personally takes all samples from development headings and is
in a position to falsify ore-reserve estimates and other critical
data. Even in legitimate promotions the superintendent or
other responsible official may prolong his tenure by falsely
encouraging reports of new development work. If the assayer
suspects chicanery, he owes it to his future professional status to
ascertain the facts.

Organization of Assaying Routine. It is essential that the
assayer organize every step in his daily routine to the end that
maximum efficiency and acceptable accuracy are ensured. The
general principle to be applied in the development of an efficient
routine is that every effort be made to keep the assays in active

10 FIRE ASSAYING

progress throughout the entire procedure, from the time the
samples are received until the final results are reported. The
principal details to observe are that the furnaces be hot when
needed, that equipment and supplies be ready for use, that
accidental delays or interruptions be avoided, and that the
laboratory be neat and orderly. Since some of the individual
steps in the assay process are cyclic and involve waiting periods,
certain operations may be dovetailed together, and much of the
incidental work of preparing flux mixtures and of keeping the
laboratory in x order may be done during these waiting periods.
For example, a set of fusions remains in the furnace for from 15 to
20 min. If another set is to follow, the preparation of the second
set can be made while the first set is in the furnace. If only a
single furnace is available, it should be loaded with cupels
immediately after the last set of fusions is poured, in order that
the cupels will be hot when the buttons are ready for them.
Cupellation requires from 25 to 30 min., and the experienced
assayer learns how to control the furnace so that little attention
is required except at the beginning and end of the process; he is
therefore free to perform other duties during the greater part
of the cupellation period. An interval, varying from 5 to 15 min.,
is also available during parting.

Assay Sequence. It is of utmost importance that the assayer
devise, and strictly adhere to, a good system of arranging arid
handling the daily assays in a predetermined order throughout
the entire process, from the time the samples are received until
the final report is made. Errors caused by misplacement of a
sample, crucible, cupel, or parting cup are undetectable by
ordinary means, with the result that high-grade samples might
be reported in place of low-grade samples, and vice versa. To
avoid such errors, every assayer should follow a system .by which
the order of samples is automatically maintained. If one man
carries through all operations himself, his system may be devised
to suit his own peculiarities; but, if helpers are employed, the
system must be such that neither the assayer nor the helpers can
confuse the sequence of assays without detection.

THE SCOPE OF FIRE ASSAYING 11

be no guaranty in itself that the buttons would be placed in the
proper cupel. Parting cups may be marked permanently, but
time is lost in arranging them in the proper sequence. Thus a
satisfactory routine should be developed that is independent of
numbered marks on the assay receptacles.

When the day's set of pulverized and mixed samples has been
assembled, the assayer arranges them in suitable order and
records the sample numbers or description on his report form or
notebook. From then on, until the final report is prepared, the
assays are handled in the same sequence and may be assigned
serial numbers from 1 upward.

Ruled lines may be provided in the notebook to separate each
furnace load of crucibles, each set of cupellations, and each tray
of parting cups. The number of assays in each of these units is
usually different, but if special parting-cup trays are made they
should hold the same number of parting cups and be in the same
arrangement as a set of cupels in the furnace.

All carrying trays for crucibles, buttons, cupels, and parting
cups should be designed so that the two ends are distinctive,
thus avoiding the danger of reversing the entire set. The
arrangement of assays in rows in the? carrying tray should read
from left to right, as printed matter is read, not zigzagged as in
section numbering in a township. If assistants of uncertain
reliability arc employed, the assayer may insert a blank assay of
known silver and gold content in a key position in each furnace
set. If this assay appears in its proper place when the dor6
beads are weighed it is reasonable assurance that no misplace-
ments have been made in the previous operations. For con-
trolling the fusion and cupellation sequence a small quantity of
some element may be added to one or more of the assays, that
will give a stain on the cupel that is not present in any of the
other assays. Copper or nickel is suggested for this purpose
and may be added to any of the regular assays if the amount is
just sufficient to give a distinctive cupel stain but not enough to
cause cupellation loss or freezing of the button.

Records. The principal technical records needed in an assay
office are a suitable laboratory notebook, or work sheet, and a
report form. The' laboratory work sheet is used for keeping a
record of the sequence of assays, the weight of ore tken, the
amount of silver added for inquartation (if any), and the respec-

FIRE ASSAYING

tive weights of the dore* and gold beads, together with the
calculated gold and silver content of the samples. For beginners
a ruled form on which to enter the amount of fluxes used for
each sample is helpful. Such a form is illustrated in Fig. 3.
Experienced assay ers do not need to keep details of flux charges,
but a column for remarks on the work sheet should be provided
for recording abnormalities.

The report form gives only the original description or number
of the sample and the gold and silver content in weight units,
or in monetary value, or both. In the United States, gold and

ASSAY LABORATORY

ASSAY REPORT

GOLD AT $35.00 PER 02.TROY LITHARGE f MG. AG PER 2 A.T. PflO
SILVER AT $ PER 02. TROY CORRECTION! MG Au

SAMPLE
MO.

FLUXING DATA

fee

WT. BEAD IN MG.

ASSAY.OZ PERT.

VALUE IN | PER TON

DESCRIPTION

ORE

Peo

SODA

BORAX

BG
A.T

FLOUR
6.

NITER
G.

K;to

DORE

TOTAL

% '*

REMARKS: DATE

ASSAYER

FIG. 3. Form for detailed record of assays.

silver are reported in troy ounces per avoirdupois short ton
(2,000 Ib.) and, for convenience, a special system of weights is
used in which the assay ton (A.T.), weighing 29.166 g., bears the
same relation to a milligram as the avoirdupois short ton bears
to the troy ounce. Hence, when a 1-A.T. sample is weighed
out, the weight of gold or silver in milligrams is the content of
gold or silver in troy ounces per avoirdupois ton. In some
British countries, gold and silver are reported in pennyweight
(dwt.) per long ton (2,240 Ib.) ; and in countries using the metric
system, gold and silver are reported in grams per metric ton.

At assay offices connected with a mine, mill, or smelter a
printed report form is usually prepared on which the daily results
of all routine assays, as well as additional special assays, may be

THE SCOPE OF FIRE ASSAYING 13

recorded. Also, small report forms are needed for reporting the
results of individual assays or special groups of assays to staff
members, shippers of ore for treatment, and others. It is usually
necessary to provide several copies of assay reports, and one copy
should always be -retained in the assay office. Any specially
printed forms, therefore, should be assembled in convenient form
for duplicating, and the use of distinctively colored paper for
each copy is desirable.

CHAPTER II
SAMPLING

When a metal-content determination of a large amount of
material is required, a sample is taken and sent to the assay
office, where, by further sampling operations, a small portion is
obtained for the assay process.

The proportion of the required metal that is found in the assay
portion is attributed to the material from which the sample was
taken. For example, if an assay portion of 1 assay ton (29.166 g.)
is used in the assay of an ore, and if 0.23 mg. of gold are recovered,
the ore is said to assay 0.23 oz. of gold per ton of ore. In order
that the assay may be accurate the/assay portion must contain the
same relative amount of gold as the average of all the material
from which the sample was taken. The difficulty of obtaining
an assay portion small enough to be economically treated in the
assay process and yet representative of the average of several
tons of a heterogeneous material varying in grade from barren
rock to particles of native metal is often greater than the diffi-
culties encountered in the actual assay process.

Types of Samples. There are three types of samples : the spot
sample^ _the. random,, sample, and the stratified sample. A spot
or grab sample is taken from one place in the material to be
sampled and tends to be less reliable than a random sample.
The random sample consists of portions taken at random from
various places in the material to be sampled. The stratified
sample consists of portions taken at regular intervals throughout
the material to be sampled. Stratified samples tend to be more
reliable than random samples, and, for that reason, they are
frequently used.

THE ACCURACY OF SAMPLES

Random and stratified samples are made up of a number of
cuts or portions taken from different parts of the material to be
sampled. Each cut consists of a number of individual fragments

SAMPLING 15

of material. Either or both of these sampling units, cut and
fragments, may vary individually in grade from the average of
the material being sampled. Accuracy of sampling can be
investigated from the standpoint of the sample cuts or from the
standpoint of the individual fragments. With either point of
view the factors influencing the reliability- of a sample are: (1)
the dispersion, or the variation in grade of the units in the mate-
rial to be sampled, and (2) the size of the sample, or the number
of units taken into the sample. In general the reliability of a
sample is improved by increasing its weight, in order to increase
the number of units, or by crushing and mixing the material
before sampling, to decrease the dispersion per unit of weight.
Increases in either the number of units taken into the sample or
the amount of crushing and mixing increases the cost of sampling,
and, furthermore, any increase in the size of the final assay
portion increases the cost of assaying. The balance between
cost and accuracy that has been found suitable at the average
mine has resulted in customary sampling procedures. These
should not be followed blindly; all important sampling operations
should be planned to suit the material to be sampled and to
produce the accuracy desired.

Limit of Error. In any sample taken from a heterogeneous
material there is the possibility of some chance variation; conse-
quently a precise limit of error cannot be established for a particu-
lar sampling process. If a considerable number of samples were
taken in the same way from the same material, and the assay
results arranged according to the frequency of occurrence of
results in various ranges, it would be found that the intermediate
result would occur most frequently and there would be less and
less of both higher and lower results.

Another sample taken in exactly the same way from the same
material might deviate from the true result by a greater amount
than any of the previous samples; but the probability of its doing
so is small.

There is increasingly less chance of obtaining a sample with an
increasingly greater error, but there is no limit to the possible
error except the richest or poorest part of the material to be
sampled. In lieu of a definite limit of error an artificial or
practical limit may be taken at some point where the probability
of greater error is considered so small that it may be disregarded.

16 FIRE ASSAYING

The probability of error is conveniently expressed by the
standard deviation <r, which is the square root of the mean of the
squares of the individual deviations from the average and
expresses in one figure the scattering of individual results.

With random sampling the distribution of results of duplicate
samples should approximate the "normal curve." Two-thirds
of the results will be within one standard deviation from the
average, 19 out of 20 will be within two standard deviations
from the average, 369 out of 370 will be within three standard
deviations from the average, and 16,666 out of 16,667 will be
within four standard deviations from the average. In most
statistical work, three times the standard deviation is taken as
the practical limit of error, or the greatest extent to which a
result is liable to be in error due to chance variation. 1 Greater
error is possible, but the chance of its occurrence (1 in 370) is so
small that it may usually be disregarded.

The sampling error considered here is that due to chance
variation in an honest random sample. If the sampling
method is unfair the error cannot be decreased by increasing
the size of sample or by averaging a large number of incorrect
samples.

Standard Deviation. The standard deviation of a sample
may be determined by

1. Calculation from the individual deviations of a large number
of duplicate samples.

2. Assaying the individual units or portions that go to make
up the sample and, from the standard deviation of the individual
units, calculating the standard deviation of the average of all the
units.

3. Estimating the assay of the sample and the metal content
of the pieces of rich ore, from which can be calculated the stand-
ard deviation of a sample composed of a very large number
of fragments and containing only a comparatively few rich
fragments.

The standard deviation is defined as the square root of the
. mean of the squares of the deviation of individual samples from
the average of a number of duplicate samples. The first method
of calculation is expressed by the relation

1 EZEKIEL, M., "Methods of Correlation Analysis/' p. 23, John Wiley &
Sons, Inc., New York, 1930.

SAMPLING 17

where <r x = the standard deviation of the sample (milligrams per

assay ton).
x = the deviation of each individual sample from the

average (milligrams per assay ton).
n = the number of duplicate samples investigated.*
The second method is derived from the first by considering the
sample to consist of the average of a number of units, where each
unit has the standard deviation of cr x . The reliability of this
sample is expressed by the standard deviation of the mean (of the
units in the sample) a m . The relation between the standard
deviation of the mean <r m , the standard deviation of the individual
unit <r x , and the number of units included in the average n', is

* m = A/rT^t (2 ^

The third method of calculating the standard deviation of a
sample depends upon the standard deviation of the number of
rich particles (o-p) where there is the chance of success (p) and the
chance of failure (q) in obtaining rich particles that are present
in the proportion p in a sample of n" particles,

The chance of success plus the chance of failure is equal to one,
so that (1 p) may be substituted for q and the equation
becomes

<T P = \ / ^ri ff p n"p 2 (4)

In almost any ore sample the number of rich particles is very
small in proportion to the total number of particles ; consequently
p is very small and n"p 2 may be neglected. The average number
of rich particles in the sample M is equal to the total number of
particles in the sample n" multiplied by the proportion of rich

* The chance of error is somewhat greater than normal, when calculated
from a standard deviation based on less than 30 samples.

1 EZEKIEL, op. cit., p. 25.

2 YULE, G. U., "An Introduction to the Theory of Statistics," 7th ed.,
p. 257, Charles Griffin & Company, Ltd., London, 1924.

18 FIRE ASSAYING

particles p. Substituting M for n"p and neglecting n"p z give
the Poisson relationship

* P = VM (5)

The standard deviation in the number of rich particles in a
sample of n" particles <r p , multiplied by the average metal
content of the rich particles in milligrams (7, is the deviation in
milligrams, which also equals the standard deviation in milligrams
per assay ton <r x , multiplied by the weight of sample in assay tons
W.

<r p C = <r x W
Therefore

W .

The average number of rich particles in the sample M is equal to
the assay of the ore (in milligrams per assay ton) A, multiplied
by the weight of the sample in assay tons W, and divided by the
average metal content of the rich particles in milligrams C.

(7)

Substituting (7) and (6) in Eq. (5) and simplifying give

[AC

Examples of the Determination and Use of the Standard
Deviation. As an illustration of the methods for determining the
standard deviation of a sample and its use in estimating the
limit of error of a sample, or the minimum weight of sample to
achieve any required reliability, consider the selection of the
assay portion from a gold-ore pulp that has been pulverized to
pass a 100-mesh sieve. Usually no detectable sampling error is
found in this operation, but with some gold ores an alarming
variation is found between duplicate assay portions. At the
Alaska Juneau mine, for example, 10 separate assay portions
from the same samples gave the results shown in Table I. 1 It

1 BRADLEY, P. R., Mining Methods in the Alaska Juneau Mine, Trans.
A.I.M.E., vol. 72, p. 106, 1925.

SAMPLING

is obvious that with this material any single assay portion of 1
assay ton is liable to deviate from the true result by more than
can be allowed. By calculating the standard deviation of these
samples, the practical limit of error in an individual assay ton can
be determined ; if it is more than permissible the size of the sample
necessary to give the desired accuracy can be calculated. The
computation of the standard deviation of a 1 -assay-ton portion
taken from sample 1 is given in Table II.

TABLE I. RESULTS OF 10 ASSAYS ON SAMPLES FROM ALASKA JUNEAU MILL

Assay results, ounces of gold per ton of ore

Assay

number

Sample 1

Sample 2

Sample 3

Sample 4

0.02

0.04

0.03

0.09

0.02

0.01

0.07

0.06

0.05

0.08

0.11

0.07

0.26

0.06

0.00

0.04

0.26

0.14

0.02

0.04

0.06

0.00

0.10

0.03

0.01

0.04

0.24

0.08

0.02

0.03

0.19

0.04

0.01

0.02

0.19

0.08

Average

0.029

0.034

0.148

0.075

The standard deviation of a single assay-ton portion as taken
from sample 1 is shown to be 0.0236. On the average of two
times out of three, any single assay-ton portion may be expected
to give an assay between 0.006 and 0.053 oz. per ton. Only
once in 370 times will the error exceed three standard deviations
or 0.07 oz. per ton. Therefore, 0.07 may be taken as the practical
limit of error in a single sample of 1 assay ton. If greater
accuracy is desired, it can be obtained only by increasing the size
of the sample, as the gold particles cannot be crushed to reduce
the dispersion. The size of sample n f necessary to give a result
within E ounces per ton of the true value is calculated by substi-
tuting E/3 for <r m in Eq. (2), and solving for n'. Thus, to find
the required number of assay tons of sample 1 to give an assay
portion accurate to within 0.02 oz. per ton, substitute 0.02/3 for
<T m and 0.0236 for <r x .

FIRE ASSAYING

TABLE II. STANDARD DEVIATION OF A I-ASSAY-TON PORTION
FROM SAMPLE 1

Assay result,

Assay number

ounces of gold

Deviation X

per ton

0.02

0.009

0.000081

0.02

0.009

0.000081

0.06

0.031

0.000961

0.07

0.041

0.001681

0.00

0.029

0.000841

0.02

0.009

0.000081

0.06

0.031

0.000961

0.01

019

0.000361

0.02

0.009

0.000081

0.01

0.019

0.000361

Average

0.029

2 = 0.005490

From Eq. (2),
Substituting,

(O 2

O.QOQQ43 + 0.000549
0.000043

(9)

= 14 assay tons

According to the calculation, an assay portion of at least 14
assay tons is required from this material, in order that the sample
be accurate to 0.02 oz. per ton of ore. In practice this is usually
accomplished by making seven separate fusions of 2 assay tons
each.

Frequently it is desirable to know the standard deviation of a
sample without going to the trouble and expense of taking a
series of duplicate samples for assay. When the approximate
assay of the ore is known and the average metal content of the
rich particles can be estimated, the standard deviation can be
roughly approximated by substituting in Eq. (8).

For example, the samples of Table I might have been known to
assay about 0.07 oz. of gold per ton. A knowledge of the metal

SAMPLING 21

content of the rich particles could have been obtained by panning
a portion of the crushed ore and examining the gold. If a con-
siderable number of the gold particles were of such a size as to
weigh about 0.02 mg., the standard deviation of a 1 assay-ton
sample would be:

(0.07) (0.02)

This value for the standard deviation compares with 0.024 for
sample 1, 0.019 for sample 2, 0.085 for sample 3, and 0.0311 for
sample 4. Gold particles weighing 0.02 mg. are about the
heaviest found in minus 100-mesh material, so that the examples
have been of an unusually difficult ore. With many ores, even
though an occasional large piece of gold is found, the majority of
the gold particles are so small that duplicate assay portions of 1
assay ton will generally check to 0.01 oz. of gold per ton of ore.

If the sample is to be taken from material that has not been
finely crushed, Eq. (8) can be used to estimate the minimum
weight of sample to give any desired reliability. To illustrate,
a calculation of the minimum weight of sample that may be
taken from a certain silver ore is given below:

x Calculation of Minimum Sample Weight
Conditions

1. The ore is expected to assay about 15 oz. of silver per ton
of ore.

2. The ore has been crushed to 1-in. pieces, and an accuracy of
0.5 oz. of silver is desired.

3. Rich pieces of ore weighing about 50 g. and known to assay
about 100 oz. of silver per ton of ore are observed in the material.

Solution

The silver in the rich pieces of ore is calculated from the assay-
ton proportion and the weight of the pieces of ore to be about 171
mg. 1 The standard deviation allowed is 0.5/3 or 0.17. The
grade of the ore is 15 oz.

1 1 A.T. = 29.166 g., in which quantity of ore 1 mg. of silver is equivalent
to 1 oz. per ton. Hence, 29.166:50 = 100: (milligrams of silver), and milli-
grams of silver (in the rich pieces) 171 mg.

22 FIRE ASSAYING

Substituting in Eq. (8)

0.17 '" X

Solving for TF,

W = 89,000 A.T.
or

W = 5,740 Ib. minimum weight.

The calculation just made indicates that a sample of at least
5,700 Ib. must be taken from the silver ore described, in order not
to exceed a practical limit of error of 0.5 ounce per ton.

Mine samples are frequently crushed to J^ in. and divided on a
Jones riffle to obtain a final assay sample of about 1 Ib. The
practical ' limit of error for an assay sample taken in this way
from the above-mentioned silver ore is calculated below.

Calculation of Limit of Error
Conditions

1. The ore is expected to assay 15 oz. of silver per ton.

2. A final assay sample of about 1 Ib. or 16 A.T. is to be split
from a sufficiently representative portion of the ore that has been
crushed to y in.

3. Crushing finer than 1 in. has not released richer pieces of
ore, so that the richest pieces still assay about 100 oz. per ton.

Solution

The weight of silver in the rich V-in. pieces is about ^4 of
that in the 1-in. pieces, or 17 M4 == 27 mg.
Substituting in Eq. (8),

^ 16
5 oz. of silver per ton

The practical limit of error = 3<r x 15 oz. of silver per ton.

The calculation shows that as often as one time in three, the
assay sample taken by standard practice from this silver ore will
be in error by more than 5 oz. of silver per ton. The practical
limit of error is as great as the assay itself. This is an unusually
large error, and would occur on an average of only one time in

SAMPLING 23

370. In most cases a considerable amount of the rich silver
particles are much smaller than Y m - but even if the average
rich particle contained only 5 mg. of silver, the standard devia-
tion would be 2 oz. of silver per ton of ore.

Probably most assayers do not realize the unreliability of a
final assay sample, consisting of about 1 Ib. of material when it is
split from an ore, which has been crushed to about ^ in. and
consists of particles differing considerably in grade. After
experimenting with some 136 triplicate assay samples split with
the Jones riffle, from a gold ore crushed to % in. and containing
no visible free gold, Boericke 1 concluded that samples split from
a gold ore of variable content show variations of as much as 25 or
30 per cent between assays, unless the original sample is crushed
much finer than minus ^ in. before splitting.

When such large errors are commonly present in single assay
samples, one may wonder how assays can be used to evaluate ore
bodies reliably or how assays can be used to obtain monthly
metal balances in milling operations. Both of these operations
are based upon the average of a considerable number of separate
samples. Each separate sample may be regarded as a unit of a
larger sample that is composed of the average of all the units.
The accuracy of the average may then be expressed in terms of
the standard deviation of the mean, from Eq. (2).

where a m the standard deviation of the mean.

<r x = the standard deviation of the individual units.
n' = the number of units in the average.

For example, samples from a gold mine might individually
have a standard deviation <r x of about 0.03 oz. of gold per ton and
a practical limit of error of 0.09 oz. of gold per ton, yet the average
of 80 samples would have a limit of error of only 0.01 oz. of gold
per ton. Similarly the feed sample of ore going to a mill for one
day is likely to be considerably in error, but the monthly average
of the daily assays is usually sufficiently reliable.

Effect of Particle Size on the Minimum Weight of Sample.
The minimum weight of sample may be decreased as the indi-

1 BOERICKE, W. F., The Jones Riffle in Cutting Down Samples, Eng.
Min. Jour., vol. 140, No. 6, p. 55, 1939.

24 FIRE ASSAYING

vidual particles in the material to be sampled are crushed smaller.
The relation between particle size and minimum weight can be
derived from Eq. (8)

* = A \^ (8)

Consider sampling an ore of assay A, where a constant relia-
bility of sample v x is desired. The weight of sample W must be
proportional to C, the precious-metal content of the large rich
pieces of ore. Disregarding the change in grade of the rich
pieces of ore due to the liberation of pure valuable mineral on
crushing, the metal content of the rich pieces of ore is proportional
to their volume or to the cube of their diameters. Therefore

W = kd* (10)

In Eq. (10), W is the minimum weight of sample when the ore
contains large pieces of diameter d. The magnitude of the
proportionality constant k depends upon the reliability of the
sample and the units used for sample weight and diameter of
particles.

Rich valuable minerals are usually broken free at a compara-
tively small size and increasingly larger pieces of rich ore contain
a proportionately greater amount of gangue minerals, gradually
approaching the average assay of the ore.

For this reason, the metal content of the rich pieces of ore is
seldom proportional to the cube of their diameters but is approxi-
mately proportional to some power less than the cube of the
diameter. Richards 1 found that the weights of samples used in
practice varied approximately as the square of the diameter of
the largest pieces. This gives the convenient and practical
relation

W = kd* (11)

The value of the proportionality constant can be found from
the weight of sample that is satisfactory at some particular
particle size. A 1-assay-ton sample is usually considered satis-
factory for the final assay portion of ores that have been crushed
to pass a 100-mesh sieve. Substituting 1 for W, and 0.0058 in.

RICHARDS, R. H., "Ore Dressing," Vol. 2, p. 850, McGraw-Hill Book
Company, Inc., New York, 1906.

SAMPLING

(the aperture of the 100-mesh sieve) for d, gives 30,000 for the
proportionality constant.

In Table III, minimum sample weights are given, which have
been calculated from Eq. (11) with 30,000 substituted for k.
This table should be used only for rough work on ordinary ores.
The minimum weight of sample should be calculated by the
method described on page 21 for important sampling operations.

TABLE III. MINIMUM SAMPLE WEIGHTS*

Size of particles

Weight of samples

Diameter, inches

Mesh

Assay tons

Pounds

240,000

15,432

120,000

7,716

1.5

67,500

4,340

1.0

30,000

1,929

0.5

7,500

482

0.25

1,875

120

0.065

127

0.0116

0.26

0.0058

100

0.06

* Minimum sample weights determined from the approximate relation W = 30,000d 2 .

SAMPLING PRACTICE

The usual types of material to be sampled for assaying are
listed in the following tabulation:

1. Brittle materials such as most ores and rocks:
o. That can be crushed as fine as desired.

6. Containing "metallics" or malleable scales that cannot be crushed.

2. Metals:
a. Solid.
6. Liquid.

3. Aqueous solutions such as cyanide solutions:

a. Homogeneous.

b. Mixed solids and solutions.

The Sampling of Brittle Materials

Practice in sampling brittle materials for assay is illustrated
by the procedures used in sampling broken rock. The material
to be sampled may comprise a large shipment of ore to a

26 FIRE ASSAYING

smelter or a small sample from a mine or prospect. In either
case the material is crushed and then divided into a sample
portion and a rejected portion. The sample portion is again
crushed and divided into repeated stages until the final assay
sample is obtained. If malleable particles are encountered dur-
ing the crushing operation the procedure for sampling material
containing metallics (page 32) must be used.

Mine samples are usually comparatively small, and assay
samples are prepared from them in a sampling room that is a part
of the assay office. The sampling of large amounts of ore, such
as the feed to a mill or a shipment to a smelter, is generally done
outside the assay office.

Preparation of Mine Samples for Assay. The sampling room
at a mine assay office should be equipped with an adjustable
crusher, either a disk pulverizer or a coffee-mill type of sample
grinder, a Jones riffle, a sample drier, and pans and buckets for
containing the samples during their preparation. An outline of
the usual steps in the preparation of mine samples for assay is
given in Fig. 4.

Mine samples are received in canvas bags, usually containing
from 10 to 50 Ib. of rock. One sample is opened at a time and
dumped into a large pan or a bucket. The identifying number
(on a piece of paper or a metal tag) is picked out from the mate-
rial, and placed in the drying pan to be used for this sample. The
sample is then fed with a hand scoop to a laboratory jaw crusher.
If the sample is larger than about 30 Ib. it is usually passed
through a Jones riffle, and half of it is rejected and half taken into
the sample.

The jaw crusher is then set to about J4 m -> an d the sample
portion, or all the material if it has not been split on the riffle, is
crushed to Y in. or finer. Then, by repeated passes through the
Jones riffle with half rejected at each pass, the sample is split
until only about 1 Ib. remains for the assay sample. This is
placed in the drying pan containing the sample number, and the
pan is placed in a drier heated to 105 to 110C. It is desirable to
have mine samples sent to the assay office in the afternoon, so
that they can be brought to this stage and left to dry during J
night.

After one sample has been crushed and split, the pans, riffle,
and crusher should be cleaned before starting with the next

SAMPLING 27

sample. When compressed air is available a pressure hose and
valve may be arranged so that a blast of air can be used in clean-
ing. When air is not available the cleaning must be done with
brushes. If the crushing room is not well ventilated the sampler
should wear a respirator to avoid continual breathing of fine rock
dust.

After the assay samples have dried they are passed through a
pulverizer set so that the product will pass a lOQ-mesh screen.

Mine Sample
Crusher Set to } /
Jones Riffle 1

Reject ha If Sample

Crusher Set to 1
Jones Riffle

Reject half Sample

Repeated Riffling

Rejected portion Sample I Ib.

Drier

Sample Grinder

Paper Envelope

Assay Sample to Fluxing Bench

, . ,

Standard Practice either Rapid Method

alternative I

Mix on Sampling Cloth Mix in Cocktail Shaker 2

Weigh Spatula Dips Weigh Single Scoop

Assay Portion Pulp to Storage Pulp to Storage Assay Portion

to Crucible to Crucible

orScorifier or Scarifier

' Bypass 1st riffle with small samples

2 - Use of cocktail shaker for mixing pulps suggested by H.R.Bramel

FIG. 4. Typical preparation of mine samples for assay.

Then the dried and pulverized samples, called " assay pulps,"
are placed in heavy paper envelopes, each marked with the
sample number.

Taking the Assay Portion. The assay pulps are transported
to the fluxing bench where the final assay portion is weighed.
Two methods of taking the assay portion are available : the usual
method with a sampling cloth and a rapid method with a cocktail
shaker.

28 FIRE ASSAYING

In the usual method of taking the assay portion the assay pulp
is poured onto a square sheet of rubberized cloth and mixed by
rolling the material back and forth from corner to corner of the
cloth. At least 60 rolls are generally given; the material is then
smoothed into a flat cake with a horizontal spatula stroke.
From the mixed and flattened assay pulp, small dips are taken
with a spatula at more or less regularly spaced intervals and
placed on the pan of the pulp balance. When just a little more
than the desired weight has been placed on the balance pan, a
small portion is removed with the spatula and held immediately
above the pan on the balance. Then the spatula is tapped with
the finger to shake material, little by little, onto the pan until
the desired weight is obtained. The remainder of the^ssay pulp
is transferred back to its bag, and the sampling cloth is brushed
clean before starting with the next sample.

When the rapid method of taking the assay portion is used, the
assay pulp is poured into a cocktail shaker and given about 20
shakes. Then, with a single scoop just a little more than the
desired assay portion is taken and placed on the balance pan. A
small portion of the material is then removed from the pan and
tapped back, little by little, onto the pan until the desired weight
is obtained. The remainder of the assay pulp is then poured
into its bag, and the shaker is brushed clean before starting with
the next sample.

The rapid method is easier and takes about half the time of the
usual method. Quicker and more thorough mixing of finely
ground materials can be obtained in a cocktail shaker than on a
rolling cloth, but taking the assay portion with a single scoop is
theoretically not quite so reliable as taking it with 10 or 15 small
spatula dips. In practice, no significant difference is found
between assay portions taken by either method from ordinary
samples.

Sampling of Large Amounts of Broken Rock. The methods
used for the sampling of large amounts of broken rock depend
upon the material to be sampled and the expense justified by the
importance of the sample. Highest accuracy is desired at custom
smelters where rich ore and concentrates are purchased. Pay-
ment for a shipment is based upon the valuable metal contained in
the shipment, and this is determined by (1) weighing, (2) taking a

SAMPLING 29

moisture sample to find the moisture content, and (3) taking an
assay sample to find the valuable-metal content.

A satisfactory moisture sample is difficult to obtain. Gener-
ally a grab sample of about 10 Ib. is taken for the moisture sample,
just after the shipment has been weighed and while it is being
unloaded. The moisture sample is weighed soon after it is
taken, after which it is dried at 110C., and the moisture is
considered to be the loss in weight. When the material con-
tains large chunks of rock as well as fine material a sample as
small as 10 Ib. is not likely to represent accurately the moisture
content.

This difficulty may be avoided by splitting the moisture sample
from the crushed material during the preparation of the assay
sample. When the moisture sample is taken from the crushed
material a correction is made for the drying that takes place
during the crushing and sampling operations. The average
percentage loss of moisture during sampling is determined by
experiment for different seasons of the year, and these corrections
are arbitrarily applied to subsequent moisture determinations.

The assay sample is taken by hand sampling or machine sam-
pling. Finely divided materials such as flotation concentrates are
usually sampled by hand sampling methods, while ores containing
coarse rock are passed through a sampling mill consisting of
alternate crushers and mechanical samplers.

A diagram typical of the sampling operations at a Western
lead smelter is given in Fig. 5. At mills or smelters where the
sampling operation does not justify the expense of mechanical
sampling equipment, hand sampling methods are used entirely.

Hand Sampling Methods. The common methods of hand sam-
pling are shovel sampling, coning and quartering, and pipe
sampling.

Shovel sampling or fractional shoveling consists in taking for
the sample a shovelful at regular intervals while the material to
be sampled is being shoveled from one place to another. This is
the best hand sampling method for lump ore and gives satis-
factory results, if the material has been crushed to at least 2 in.
and portions for the sample are taken at sufficiently close inter-
vals, and provided that the operator does not deliberately select
either rich or low-grade material for the sample.

FIRE ASSAYING

Coning and quartering has been widely used for sampling small
lots of crushed ore. It consists in shoveling the crushed ore into
a conical pile, which is then flattened and divided into quarters.

Shipment of Ore

Assign Lot Number

Weigh

Hand Sampling

Shjpmenf of Concentrate
Assign Lot Number
Weigh

Moisture Sample Moisture sample Assay
IOIb. llOlb. 11

Crusher Set 2"
1st. Mechanical Sampler

sample Reject

Reject

Sample 20% of Lot

Crusher Set 1"
2nd. Mechanical Sampler

Reject

Sample 4% of Lot

Crusher Set Vfc
3rd Mechanical Sampler

Reject

Sample 0.8% of Lot

Crusher Set '4"
4th. Mechanical Sampler

\r
Reject

Sample '16V. of Lot 1
1700lb.
L_

Original

Crush and Screen 20- mesh

Mechanical Mixer

Jones Riffle

Mix by shoveling
Jones Riffle

Duplicate

Hold for resampling in case
of d isagreement

Reject

Sample
Repeated Grinding .Screening, Mixing. Riffling

Drier

4 Ib. through 100-mesh
Jones Riffle

Jones Riffle

Reserve Sample Umpire Sample Sample for Seller Sample for Purchaser

' With small lots, bypass first samplers

FIG. 5. Typical sampling at a custom smelter.

Opposite quarters are rejected, and the remaining quarters are
taken for the sample.

When forming the cone it is essential that the apex of the cone
be kept in a vertical line. A segregation of fine ore takes place

SAMPLING 31

at the center of the cone, and, by drawing the cone to one side,
it is possible for the operator deliberately to vary the grade of the
sample. Coning and quartering is being replaced by shovel
sampling, as shovel sampling requires less work and gives samples
of greater reliability.

Pipe sampling is carried out by driving a pipe vertically down
through the material to be sampled and withdrawing a core of
material that clings to the inside of the pipe. The pipe is then
tapped with a hammer to remove the sample from the pipe.
This method can be used only for sampling finely crushed mate-
rial, such as flotation concentrates, through which the pipe can be
driven. The sample should be taken all the way from top to
bottom of the material being sampled, and the sample of a large
lot should be made up of material taken with the pipe driven at
regularly spaced intervals through the material. Pipe sampling
is a rapid and easy method of obtaining reliable samples of fine
material.

Mechanical Sampling. Mechanical sampling machines are
eitheFsTationary ui moving. Moving machine samplers, such
as the Vezin or Brunton type, cut across a falling stream of ore
at intervals, taking^alTthe stream a portion of the time. The
machines should be designed so that all parts of the ore stream
are represented in their correct proportions. To avoid clogging,
the width of the sample opening should be at least three times the
size of the largest pieces of ore. Moving machine samplers give
good results with materials that flow freely. They have a low
cost of operation and are widely used for sampling large amounts
of material.

Moving machine samplers are not suitable for sticky materials
that will clog the opening. -Consequently, concentrates are
usually sampled by hand.

Stationary machine samplers have a fixed cutter and take a
portion of the ore stream all the time. A falling ore stream is
likely to have the coarse and fine material segregated so that
one part of the stream may be richer or poorer than the rest.

The Jones riffle is the most widely used stationary mechanical
sampler. It avoids the difficulty caused by segregation, by
taking continuously small fractions from many parts of the ore
stream. It is generally used for dividing samples that have been
reduced to the point where they are handled by hand. The

32 FIRE ASSAYING

openings in the Jones riffle should be large enough so that they
do not become clogged and yet narrow enough so that many cuts
enter the sample. Usually alternate cuts enter the sample and
reject. There should be an even number of divisions so that the
division at one end enters the sample and that at the other end
enters the reject.

Sampling Material Containing Metallics. Ores that contain
coarse malleable particles should not be sampled and assayed in
the ordinary manner. The malleable particles cannot be readily
broken into small pieces, and the chance presence or absence of
large fragments of valuable minerals in the assay portion causes
serious variations in the assay result. Malleable particles likely
to be encountered in ores are native gold, native silver, native
copper, and cerargyrite. When large particles of any of these
minerals are present, a "metallics assay" should be made.

Metallics Assay. The metallics assay consists of the following
steps :

1. Crush the ore sample to about the size of the malleable
particles.

2. Take a metallics-assay portion so large that chance variation
in distribution of valuable minerals will not disturb the reliability
of the sample more than can be allowed.

3. Separate the coarse metallics from the bulk of the sample
and weigh them.

4. Recover the gold and silver from the coarse metallics.

5. Assay the remaining bulk of the mctallics-assay portion by
ordinary methods.

6. Calculate the total gold and silver in the original metallics-
assay portion^

7. From the total gold and silver in the metallics-assay portion
calculate, by proportion, the amount per assay ton.

Crushing a Metallics Sample. Materials containing metallics
can be crushed by ordinary methods to about the size of the
metallic particles. Further crushing by grinding between two
rubbing surfaces, as in a disk pulverizer, causes the malleable
particles to roll into cylindrical and spherical shapes, and part of
the malleable minerals is rubbed into the grinding surfaces.
Crushing by rolling action is much to be preferred, as there is less
danger of salting the crushing equipment and because the malle-
able minerals are rolled into flat flakes, which can easily be

SAMPLING 33

separated by screening. Roll crushers, or a grinding pan with a
heavy iron roller, such as those used for working clays, are desir-
able for fine crushing samples containing metallics.

Taking the Metallics-assay Portion. The minimum allowable
weight of the metallics-assay portion depends upon the relia-
bility desired, the metal content of the large metallics, and the
grade of the ore. The relation of these factors to the weight of
sample is expressed by Eq. (8). An example of its use is given
below :

Conditions

\ . One kilogram of the crushed ore is concentrated by panning.
About 50 g. of concentrate is obtained, which consists of pyrite
with some free gold.

2. Six comparatively large pieces of free gold are visible. The
largest weighs close to 0.5 mg. and is estimated to be about 800
fine.

3. An assay accurate to 0.02 oz. of gold per ton of ore is desired.

Solution

Substitute values in the equation

[AC

<r x = one-third of the allowable error = 0.02/3 = 0.007
A = the assay of the ore due to coarse metallics
Gold in 1 kg. = (6 particles) (0.5 mg. weight) (0.8 fineness

correction) = 2.4 mg.

29 166
Gold per assay ton = A = (2.4) ' = 0.07 mg.

C = gold in one large metallic particle = (0.5) (0.8) = 0.4 mg.
Substituting and solving for TF,

0.007 =

W
W = 571 A.T. or about 37 Ib.

After the desired weight of the metallics-assay portion has been
determined, a sample of about that weight is cut from the crushed

34 FIRE ASSAYING

ore by standard sampling methods. This sample is then weighed
#nd its weight recorded.

Separation of the Coarse Metallics. Three processes are avail-
able for the separation of coarse metallic particles from the
metallics-assay portion:

1. Screening.

2. Concentration.

3. Amalgamation.

Screening is ordinarily used to separate the malleable particles
from the metallics-assay portion. It is carried out in stages,
starting as soon as the malleable particles begin to interfere with
crushing, and is continued during crushing, using increasingly
finer screens until finally a pulp free from coarse met allies is
obtained. The final screen size should be at least 100-mesh, and
for accurate work should be 150- or 200-mesh. Some brittle
minerals may be allowed to remain on the screen with the metal-
lies, but the group of met allies must be of small enough bulk so
that it can be put through the assay process. The weight of the
metallics should be recorded.

Concentration is commonly used to separate metallics in the
assay of placer gravels. It can also be used in the metallics
assay of ores, but, when a large amount of sulfide minerals is
present in the ore, the complete separation of the metallics by
this method is difficult. The concentration separation is made
with a gold pan or other gravity concentration device. With
placer gravels the separation is frequently so complete that only
the concentrate need be carried on through the assay process,
and the tailings may be discarded. A considerable amount of
black sand, however, often collects in the concentrate and makes
it too bulky to be economically put through the assay process. In
this case the concentrate itself may be given a metallics assay, using
either screening or amalgamation to separate the coarse gold.

Amalgamation is particularly suited to the recovery of metallics
from gold ores. It is carried out by grinding the metallics-assay
portion in a grinding pan or an amalgamation barrel with water
and mercury. The mercury, with the metallics that amalga-
mated, is then collected by panning or some other gravity
concentration method and is assayed to determine the gold and
silver that amalgamated. The tailings from amalgamation are
dried and assayed by ordinary methods.

SAMPLING 35

Recovering the Gold and Silver from the Coarse Metallics. All
the coarse-metallics fraction obtained from the metallics-assay
portion must be treated to recover its gold and silver. That from
screening or concentration is usually treated by the crucible assay
(Chap. VII). Scorification (Chap. VIII) could be used, but this
method is not suitable for handling more than about 2 g. of basic
impurities.

The mercury with amalgamated metallics from the amalgama-
tion separation can be assayed by the nitric acid method or by
distillation in a crucible assay charge. If gold is the only metal
sought, the nitric acid method is suitable. The mercury and
amalgam are heated in a beaker with nitric acid until the mercury
is dissolved. The residual gold is washed, dried, and weighed.
Inquartation and parting (Chap. V) are not required unless large
pieces of gold that did not amalgamate all the way to the center
are present.

When silver as well as gold is to be determined in mercury and
amalgam, distillation in a crucible assay charge is used for the
method of assay. The mercury and amalgam are placed in the
bottom of a 30-g. crucible and covered with a 1-assay-ton crucible
charge corresponding to a crucible charge for a quartz ore, using
assay silica and fluxes in accordance with the principles discussed
in Chap. VII.

Liquid mercury dissolves about 1 mg. of gold per gram of
mercury, so that even though the mercury used in the amalgama-
tion assay has been squeezed through a chamois it cannot be
assumed to be gold free. An assay should be made on another
portion of the mercury used in the process, and the gold and silver
found should be deducted from the amalgamation assay.

Calculating the Metallics Assay. The weight of the metallics
fraction is subtracted from the weight of the metallics-assay
portion to obtain the weight of the pulp after separation of the
metallics. The gold and silver content of this pulp is calculated
from its weight and assay. To this amount of gold and silver is
added that obtained from the metallics. The sum represents
the total gold and silver content of the metallics-assay portion.
The gold assay of the material is then found by dividing the gold
in the metallics-assay portion (in milligrams) by the weight of the
metallics-assay portion (in assay tons). An example is given
below:

36 FIRE ASSAYING

Cojiditions

1. Metallics-assay portion taken with a Jones riffle.
Weight = 18.365 kg.

2. Weight of metallics = 11 g.

3. Gold in metallics = 56.12 mg.

4. Gold assay of pulp = 0.34 oz. per ton

Solution

Weight of pulp = 18,365 - 11 = 18,354 g.
Gold in pulp: 18,354 = 0.34:29.166
Gold in pulp = 213.96 mg.
Total gold = 213.96 + 56.12 = 270.08 mg.

1 Q Q* P\

Assay-ton weight of metallics-assay portion = 7pr^ = 629.67

^y.ioo

270 08

Assay of the ore = A0n ' - = 0.43 oz. of gold per ton of ore
o^&y.o7

Usually the weight of the metallics fraction is negligible in
comparison with the weight of the metallics-assay portion. Then
the assay of the pulp plus a proportional part of the gold or silver
in the metallics equals the assay of the ore. The proportional
part of the gold or silver in the metallics is the gold or silver in
the metallics divided by the weight of the metallics-assay portion
in assay tons. In the example just given, the proportional part
of the gold in the metallics is 56.12/629.67 or 0.09 mg. of gold per
assay ton. The gold assay of the pulp, 0.34, plus 0.09 equals
0.43 oz. of gold per assay ton the gold assay of the ore.

Sampling Metals

Liquid metals are usually capable of dissolving other metals to
form homogeneous liquid solutions. The precious metals are
dissolved and collected in this way at a lead smelter by the blast-
furnace lead, and at a copper smelter by blister copper. These
bullions usually contain more lead or copper than the lowest
melting alloy in their series of alloys with the precious metals.
Consequently, as the bullions cool, the first solid to form is poorer
and the last is richer than the original solution. When the liquid
is cooled rapidly the constituents of different grade form in small
grains and do not have time to segregate. With slow cooling,

SAMPLING 37

as when the bullion is cast into a large bar or cake, segregation
takes place and the precious metals are not uniformly distributed.

Sampling Solid Metals. Solid metals are sampled by drilling,
sawing, or punching. The samples should be taken in a manner
that will secure representation of all parts of the metal, because
solid metals are likely to be segregated.

With saw sampling, every fifth or tenth bar poured from one
lot of molten metal is sawed across the middle. The sawdust is
mixed and sampled similarly to a crushed ore.

When metals arc sampled by drilling or punching, a template is
used to distribute holes over the surface of the bars. The holes
are drilled all the way through the bars. When a large number of
bars are poured from a melt, one hole is drilled in each bar,
following the order of the template. Thus the combined drillings
represent all parts of the bars.

Sampling Liquid Metals. Liquid metals are sampled by taking
a dip from the molten bath with a ladle or by batting shot from
a stream of metal as it is being poured. If a ladle is used it
should be preheated so that the sample will not start to solidify
until it is poured in water to granulate and cool it quickly. In
the method of batting shot from a stream of molten metal a
wooden paddle is used, and the shot is batted into a pan of water.

Shot produced by either method is fairly homogeneous because
it has been poured from a melt and quickly cooled. Low-grade
bullions require comparatively large assay portions, and the
assay portion is weighed directly from the dried shot. The assay
portion required in the assay of gold and silver bullion is com-
paratively small. Generally a few shot are rolled into thin strips
with bullion rolls, and the assay portion is snipped from the strips
with shears.

Sampling Aqueous Solutions

Sampling and assaying of aqueous solutions are required for
the control of hydrometallurgical processes. Solutions are
homogeneous with regard to dissolved constituents, and segrega-
tion does not take place in the liquid to be sampled except when
it contains suspended solids. Two types of fluids are sampled at
hydrometallurgical plants: (1) homogeneous solutions free from
solid particles that will settle, and (2) pulps consisting of solid
particles mixed with solutions.

38 FIRE ASSAYING

Sampling Homogeneous Solutions. Any portion taken from
a homogeneous solution represents instantaneously the solution
from which it was taken. Solutions in process may change in
grade with time; consequently, fractions of sample should be
taken at short time intervals to obtain a sample representative
of a solution during a considerable interval of plant operation
such as a shift or a day. Samples should be collected in con-
tainers that expose as little surface of liquid as possible to the
atmosphere, in order to minimize enrichment due to evaporation.

Instantaneous samples, dipped out once a shift, are usually
satisfactory for sampling large tanks of solution where the grade
does not change rapidly. *

An almost continuous sample can be taken from a falling
stream of solution by a wire mounted so that it slants downward
across the stream. Drops of solution run out along the wire and
are caused to drop into a narrow-mouthed sample bottle by means
of a string tied around the wire and hanging down into the bottle.
When both rate of solution flow and grade of solution vary from
time to time, the drip sample is not accurate, as it does not
accumulate at a rate proportional to the flow being sampled.

Several types of moving machine samplers, which take samples
proportional to the flow of a falling stream of solution or pulp,
are on the market. These samplers move a sample cutter across
the stream at regular intervals of time and remove a fraction of
sample at each cut which is proportional to the stream.

Sampling Mixed Solids and Solutions. A flowing stream of
solution containing suspended solids, such as the tailings from a
slime cyanide plant, is best sampled by means of a moving
machine sampler. After the complete sample has been taken it
is advisable to separate the solids and solution and to assay them
separately.

Solids and solution are separated by settling and decantation
or by filtration. A portion of the clean solution should be taken
for the solution sample before dilution with wash water. The
remaining solids should be washed before being dried for assay.

CHAPTER III
WEIGHING

In the fire assay of a material an assay portion is weighed from
which the gold and silver are recovered and weighed. The
proportion of weight of gold to that of ore and weight of silver to
that of ore expresses the assay of the ore. Different units are
used in different countries for the assay proportion.

Units Used in Assaying. In the United States, Canada, and
South Africa the assay proportion is expressed in troy ounces 1
of gold and silver per short or net ton (2,000 avoirdupois Ib.) of
ore.

A system of assay-ton weights is used in weighing the assay
portion. The assay ton contains the same number of mg.
(29,166%) as there are troy ounces in a short ton. Therefore,
1 assay ton : 1 mg. = 1 short ton : 1 troy oz. The milligrams of
precious metals per assay ton indicates the assay in troy ounces
per short ton.

In England and Australia the long or gross ton of 2,240 Ib. is
used, which has an equivalent assay long ton of 32.667 g. The
gold assay in British countries is commonly reported in penny-
weight (dwt.), a troy weight containing 1/20 troy oz.

In Mexico, South America, and other countries using the metric
system the gold assay is reported in grams per metric ton, and
the silver assay is reported in kilograms and grams per metric
ton. The metric ton contains 1,000,000 g. ; consequently, 1 mg.
of gold from a 10-g. assay portion indicates 100 g. of gold per
metric ton of ore. With larger assay portions the indicated
assay per milligram is divided by the coefficient used to increase
the assay portion, in order to find the assay per milligram of gold.
For example, 1 mg. from a 30-g. assay portion indicates 10 % or
33^ g. per ton of ore.
' A comparison of assay proportion units is given in Table IV.

1 The troy ounce contains 31.1035 g., and should not be confused with the
avoirdupois ounce of 28.3495 g.

FIRE ASSAYING

Assays of lead and copper bullion are reported in the same units
used for ores. Gold and silver bullion assays are universally
reported in fineness or parts per thousand. For example,
sterling silver contains 92.5 per cent silver and, consequently,
is 925 fine. The fineness of gold jewelry is reported in carats or
twenty-fourth parts. Twenty-four carat gold is pure gold,
and eighteen-carat gold is *%4 or 75 per cent gold. The term
" carat" is also used as a unit of weight for precious stones. An
international weight carat contains 200 mg.

TABLE IV. COMPARISON OF ASSAY PROPORTION UNITS

Ounces per

Pennyweight per

Grams per

short ton

long ton

metric ton

1.00

1.12

22.40

34.23

0.89

1.00

20.00

30.61

0.04

0.05

1.00

1.53

0.026

0.033

0.65

1 .00

Size of the Assay Portion. The size of the portion of ore
weighed for the assay process ranges from }fa assay ton to 3
assay tons. As the size of the assay portion is increased, the cost
of the assay increases because of a reduction in furnace capacity
and an increase in the fluxes required. On the other hand the
larger the assay portion the more reliably it represents the
material from which it was taken and the smaller the percentage of
error in weighing the metal obtained from the process. The gold
and silver recovered from the assay portion are weighed on an
assay balance, the best of which are sensitive only to about 0.002
mg. Therefore, if an accuracy of 1 per cent is required, the assay
portion should be large enough to produce at least 0.2 mg. of
metal. Most silver ores assay more than 5 oz. of silver per ton;
consequently, assay portions as small as Ho assay ton are satis-
factory from a weighing standpoint. Low-grade gold ores require
large assay portions, and frequently more than 3 assay tons should
be used. This can be accomplished by scorifying together the
lead buttons from two or more fusions of separate assay portions.

Weighing the Assay Portion. The balance used to weigh the
assay portion is called a " pulp balance." An unenclosed balance,
sensitive to 1 mg. and having a pan that will hold 60 g. of pulp, is
satisfactory. High accuracy is seldom required. The gold

WEIGHING 41

obtained from an assay is weighed to the nearest 0.005 mg. and
a variation of less than 0.002 mg. can scarcely be detected. If,
for example, an ore contains less than 3 oz. of gold per ton, a
deviation of 20 mg. in weighing a 1-assay-ton portion does not
affect the assay result by a weighable amount. This is shown by
the following proportion:

Assay portion : gold recovered = variation in assay portion :

minimum weighable gold
29,167 mg.:3 mg. = 20 mg.: 0.002 mg.

Aside from the minimum- weighable-bead consideration a varia-
tion of 20 mg. in a 1-assay-ton portion amounts to less than
0.07 per cent, an amount well below the other errors in an
ordinary assay. Rich materials, from which accurate assay
portions of less than J^ assay ton may be taken, should be
weighed on a balance of greater sensitivity than the ordinary
pulp balance. Large assay offices usually have an " analytical"
balance available for this purpose. Assay portions of gold or
silver bullion are commonly weighed on the balance used for
weighing the final gold and silver.

The methods of taking the assay portion from crushed ore
pulps are described on page 27. Taking an exact assay portion
from samples of metals causes difficulty, particularly when the
individual pieces of metal in the sample are comparatively large.
Occasionally time can be saved by taking a portion approximat-
ing the desired assay portion, weighing this portion exactly, and
then calculating each assay on the basis of the particular assay
portion used.

Weighing the Gold and Silver. The gold and silver recovered
in the fire assay are weighed on a sensitive balance known as the
"assay" or "button" balance. When both gold and silver are to
be determined, the dore bead from cupellation is weighed. It is
then parted, and the gold is weighed. Silver is determined by
subtracting the weight of gold from the weight of the dore" bead.
The dore* bead weight is used only for the determination of silver,
and the weighing need be no more accurate than that required
for the silver assay. For most work a variation in the silver
assay of as much as Ko z - P er ton is allowable; consequently,
with ^-assay-ton portions of ore the dor6 bead need be weighed
only to the nearest 0.05 mg. The gold is always weighed at least

42 FIRE ASSAYING

to the nearest 0.005 mg., which is about the limit of accuracy
of the assay balance.

Assay Balance. The assay balance consists of a light trussed
beam supported by a fulcrum knife-edge in the center and
carrying a knife-edge bearing on each end, from which hang
small pans by light short stirrups. The horizontal top of the
beam is graduated with 100 divisions on both the left-hand and
right-hand side for riders. Near the center, small star wheels
are provided on both sides for equilibrium adjustment, and,
extending downward from the center of the beam, a short
threaded shaft carries a threaded weight, which can be screwed
up or down to change the center of gravity of the balance.
The pointer usually extends vertically upward from the center
of the beam and terminates in front of a fixed scale for observing
the oscillations of the beam. A mechanism is provided for rais-
ing the beam to arrest its motion. This mechanism also sepa-
rates the knife-edge bearings and supports the pans. The beam
arrest is released by turning a knurled knob at the front of the
balance, which lowers the beam and pans onto the knife-edges,
allowing the beam to swing.

The balance is enclosed in a glass-paneled case equipped with
level bubbles and supported by leveling screws. Usually the
front panel or door of the balance must be raised to place the
bead on the pan or to change weights, beyond the 1-mg. variation
possible with the rider. Multiple- weight attachments, having
a keyboard arrangement for adding or subtracting weights
without opening the balance door, are available. One make of
balance can be purchased with a mechanical pan extractor for
carrying the weighing pan out through a crystal shutter in the
balance door. The mechanical pan extractor, together with
the multiple-rider attachment, facilitates weighing but nearly
doubles the cost of the balance.

Assay balances are built with maximum load capacities of 1,
2, and 5 g. The largest loads commonly used are 500-mg. assay
portions of gold or silver bullion, so that a capacity of 1 g. is
sufficient for all ordinary work.

The sensitivity of a balance is expressed in terms of the smallest
weight that causes a measurable deflection in the point of rest, or
in terms of the angle of deflection of the point of rest caused by a
unit of unbalanced load. Assay balances are rated by the first

WEIGHING 43

method, and balances can be obtained sensitive to
}^oo m g-> r Koo m g- The time of oscillation increases with
the sensitivity; a good Mocr m g- balance will swing to one side
and back in about 4 or 5 sec., while a Moo- m g- balance requires
7 or 8 sec. The sensitivity of a balance can be decreased below
its rated value, and at the same time faster weighing is obtained
by lowering the center of gravity. Raising the center of gravity
to increase the sensitivity of the balance beyond the manu-
facturer's rating causes the balance to become unstable. It is
desirable to have two assay balances: (1) a balance of compara-
tively low sensitivity for rapidly weighing dore beads, and (2)
a balance of high sensitivity for weighing gold. Assay balances
are constructed with the three knife-edges in one plane so that
the center of gravity is not changed with increase of weight on
the pans. Increased load decreases the sensitivity slightly,
owing to the increased mass of the loaded beam assembly.

Assay balances are very delicate and should be installed in a
room that is as free as possible from dust, vibration, and rapid
temperature changes. The balance should be mounted on a
rigid bench having a top about 26 in. from the floor. This
height of bench places the balance in a convenient position for an
operator sitting on a stool of ordinary chair height. If 2-in.
planks are used for the bench top, side support is not required
and the operator can place his knees under the bench. It is
advisable to support the balance bench on concrete columns
extending down through the floor to the ground. The effect of
vibration on the balance can be reduced by placing rubber
cushions under the balance rests that support the leveling
screws of the balance. Sunlight causes rapid temperature
changes in the balance room and should be excluded. In moist
climates it is desirable to keep a small vessel containing desiccant
in the balance case. Some assay ers ground the rider carrier to
discharge any static electricity that might accumulate on the
beam.

Operating the Assay Balance. Each time it is used the
assay balance should be tested for adjustment by the following
operations :

1. Brush the balance free from dust.

2. Pick up the balance pans with ivory- tipped forceps and
brush them with a small brush to make sure they are empty.

44 FIRE ASSAYING

3. See that the balance is level and that the balance beam,
hanger, stirrups, and pans are in their proper positions.

4. Move the right or weighing rider to zero on the right-hand
side of the balance beam.

5. Release the balance beam and adjust to "equal swings"
by moving the left or equilibrium rider along the left-hand side
of the balance beam. The balance door should always be closed
to exclude air currents when testing for equilibrium.

The beam arrest should be released only for testing equilib-
rium; it should be up to support the beam and stirrups whenever
changing weights or doing anything that will jar the balance.
Care should be taken to raise the beam arrest slowly, particularly
when the beam is tipped to one side.

Beads to be weighed are placed in the left or weighing pan
of the balance. Large dor6 beads may be picked up with forceps
and dropped on the pan while it is in place on the balance.
Gold and small dore beads are best added from a parting or
annealing cup. This is done by moving the weighing pan to the
front of the balance case. 1 Then tip the parting cup, tapping it
gently to cause the bead to slide from the cup to the pan. Par-
ticles of gold sometimes stick in the parting cup and must be
dislodged by gently poking them with the forceps or by brushing
with a feather. The pan containing the bead to be weighed is
then placed in position on the balance.

An estimate of the weight of the bead is made, and the esti-
mated weight is added to the right-hand side of the balance.
The beam is then released to test for equilibrium. A practiced
assayer can estimate, from the rate at which the beam moves
when released, the amount of weight to add or subtract to bring
about equilibrium. After the weight correction has been made,
the beam is again released. If the beam does not swing, or if
it swings only a short distance, the weight is close enough for
dore* and silver beads. This is known as weighing by "no
deflection."

When weighing gold the balance equilibrium should always be
checked by the method of "equal swings." To determine
equilibrium by equal swings, start the beam swinging by raising
the balance door and gently fanning one pan with the hand or,

1 Ivory-tipped forceps should be used to handle the balance pans as well
as for the manipulation of weights.

WEIGHING 45

with t^ie door closed, raise one rider from the beam for an instant-
Then, with the balance door closed, observe the movement
of the pointer as it oscillates from side to side. The amplitude
of swing slowly decreases, but at equilibrium the pointer swings
nearly the same amount on either side. When it swings con-
sistently farther to one side than the other, shift the weighing
rider in the required direction and again test for equal swings.
The process is continued until equilibrium is reached. Most
beginners have difficulty in estimating the amount of unbalanced
load by the way in which the balance swings and, consequently,
many trials are required to reach final equilibrium.

Facility in the use of the assay balance can be quickly acquired
by using the " deflection method" of weighing. For this method
of weighing, the beam release mechanism of the balance must be
adjusted so that it does not start the beam swinging when released
in balance. Then when released with a small unbalanced load
the beam will swing to the side of the unbalanced load, and the
amount of deflection is proportional to the unbalanced load.
The balance can be calibrated, and a chart of unbalanced load
for various amounts of deflection prepared for use in weighing.
To weigh by the deflection method the operator proceeds by
estimation and cut and try until within the maximum deflection
of the balance. The first deflection after release is measured,
and from the chart the corresponding unbalanced load is read.
The correction (positive if the swing is to the left and negative if
the swing is to the right) may be made arithmetically and the
result considered the true weight, but it is safer actually to make
the weight change and test for equilibrium by the method of
equal swings.

CHAPTER IV
CUPELLATION

Cupellation is an oxidizing fusion of an alloy of lead, gold, and
silver in an absorbent vessel known as a "cupel." During
cupellation the lead is oxidized to litharge (PbO), most of which
is absorbed by the cupel, leaving at the finish a bead of gold,
silver, and most of the platinum and associated metals that may
be present. The bead is subsequently treated to determine its
precious-metal content.

Cupellation is a fundamental part of all fire assay procedures,
whether the lead-precious-metal alloy is the button from a
crucible or scorification fusion of an ore sample, a sample of lead
bullion, or a synthetic mixture produced by adding lead to gold
or silver bullion or to chemical precipitates high in gold and
silver.

The process of cupellation is responsible for the greatest normal
errors in the fire-assaying process, and in order to obtain con-
sistently reliable results all factors in the cupellation process
must be controlled in such a manner as to favor low losses of
gold and silver.

The chief source of error in cupellation is the loss of silver and,
to a lesser extent, of gold by absorption into the cupel. This
loss is increased by increased cupellation temperature, especially
toward the end of the process, or by the presence of impurities,
some of them in minute amounts that either lower the surface
tension of precious metals or necessitate increased cupellation
temperature. The properties of the cupel also exert an influence
on cupel absorption. Unless cupellation temperatures are
abnormally high the loss of precious metals by volatilization is
negligibly small.

Cupels. A cupel is a porous cylinder or inverted-cone frustum
of refractory material with a cupped depression in the upper end
for retaining the lead button. In modern practice, cupels are
made of bone ash, cement, bone-ashcement mixtures, or mag-

CUPELLATION 47

nesia. Magnesia cupels are purchased as a finished product,
but the others are usually made at the assay office.

Manufactured cupels and cupel-machine dies are available in
a number of sizes, the commonest of which are \}/ in. and 1J^ in.
in diameter and hold, respectively, a maximum of approximately
32 and 45 g. of lead. The height of the cupel must be such that
the total volume of the cupel material is sufficient to absorb
most of the litharge produced in cupellation, as nearly 98.5 per
cent of the total litharge that is formed is absorbed by the cupel.
Bone-ash cupels absorb a weight of litharge about equal to their
own, cement cupels absorb slightly less than their weight, and
magnesia cupels absorb three-fourths of their w r eight. Magnesia
cupels are denser than bone-ash or cement cupels, hence a
magnesia cupel of a given volume absorbs as much litharge as
a bone-ash cupel of the same volume. A height of 1^ in. will
provide approximately sufficient volume to absorb the litharge
from buttons of the maximum size that can be handled in cupels
\Y or iy% in. in diameter. Cupellation losses vary slightly with
cupel height, particularly if there is much height variation in
a given row in the cupellation furnace. Hence it is preferable to
standardize the cupel height and to keep within a maximum
tolerance of He m - from the standard.

The shape of the cupel cup has been thoroughly investigated
by King, 1 who recommends either of the compound shapes
shown in Fig. 6, in preference to the continuously concave,
approximately spherical shapes in common use. King states
the advantages of the improved cups as follows :

a. The small deep finishing cup protects the alloy near the bead stage
from air draught, giving a slow oxidation rate and a consequent lower
temperature. This protection gives lower losses where the loss rate is
highest (at the finish), and reduces the hazards of sprouting.

6. The rapid cupellation at the start is due to the general convex
shape or elevation of the main upper areas. The higher temperature
thus produced materially assists in the elimination of impurities if
present, at a time when lead is sufficient to protect silver from heavy
loss.

c. The effective air draught is relatively automatically controlled;
this advantage held at different temperatures and with various amounts
of lead.

1 KING, J. T., The Influence of Cupels on Silver Loss, Univ. Toronto
Bull. 147, pp. 50-66, 1934.

FIRE ASSAYING

I ,1"

v - /* *

FIG. 6. Cupel cup shapes recommended by J. T. King. (Reproduced by per-
mission from Univ. Toronto Bull. 147.)

CUPELLATION 49

d. Since the loss is reduced, the bead more nearly represents the
original quantity of silver present.

With a standard 1^-in. bone-ash cupel having a simple spheri-
cal cup 0.201 in. deep, the cupellation loss with the improved
cups was 1.65 per cent on 100 mg. of silver, compared to 2.13 per
cent with the standard cupel. Similar results were obtained
with IJ^-in. cupels. King's work on cupel-cup shapes was done
with 100 per cent bone-ash cupels, and it is not conclusive that
his improved shape will show equal advantages with cupels of
other materials, particularly of magnesia, which has very different
thermal properties from bone ash or cement.

Bone-ash Cupels. Bone ash is made by the calcination of
animal bones and varies in composition according to the type
of bones and the care taken in the manufacturing process.
Clean, properly burned sheep bones contain approximately
90 per cent tricalcium phosphate, 5 per cent calcium oxide, 1
per cent magnesium oxide, and 3 per cent calcium fluoride.
Commercial bone ash may contain some silica, from incomplete
cleaning of the bones, and some carbon from incomplete calcina-
tion. Alkali sulfates and carbonates may also be present, owing
to insufficient removal of organic matter before burning the
bones. Calcium nitrate may be formed by the combination of
free CaO in the bone ash with nitrous fumes in the laboratory.
Calcium carbonate may also be formed in the same manner by
prolonged exposure to the CO 2 of the atmosphere. The presence
of organic matter and of carbonates, nitrates, and other salts
that decompose at cupellation temperature (850 to 900C.) or
lower is undesirable, as the evolution of gases during cupellation
causes loss by "spitting" of the lead.

Commercial bone ash is supplied in various grades, based upon
the particle fineness. The grades marketed by the Denver Fire
Clay Company are designated X, XX, XXX, and XXXX, in
the order of increasing fineness. The average screen analyses
of these four grades as reported by King 1 are given in Table V.
King was unable to detect any appreciable variations in silver
loss with cupels made from any of these grades, but he did find
that a gradation of particle size and the absence of particles
larger than 35-mesh were essential to low cupellation loss, 2

1 Ibid., p. 12.

2 Ibid., pp. 31-38.

FIRE ASSAYING

and that smooth cups are preferable to roughened cups. Coarse
particles should be avoided, as they cause high erratic losses and
adherence of the bead to particles of the cupel, making it difficult
to clean the bead.

TABLE V. AVERAGE SCREEN ANALYSES OF FOUR BONE-ASH GRADES*

Grade of bone ash

XXX

xxxx

Mesh

Per

cent

Per

cent

Per

cent

Per

cent

cumu-

cent

cumu-

cent

cumu-

cent

cumu-

lative

0.17

11.5

11.7

1.9

0.1-

0.1

0.02

13.7

25.4

10.8

12.7

0.8

0.9

0.33

0.39

100

18.3

43.6

21.0

33.6

18.1

19.0

14.6

15.0

150

13.7

57.4

15.3

48.9

18 7

37.7

22.3

37.3

200

5.6

63.1

5.2

54.1

6.9

44.5

44 3

270

7.1

70.0

7.2

61.3

9.8

54.3

7.9

52.2

-270

29.9

38.6

45.6

47 9

Total...

100.0

* KING, J. T., The Influence of Cupels on Silver Loss, Univ. Toronto Bull. 147, p. 12,
1934. Reproduced by permission.

The cost of bone ash in 1940 ranged from $9.50 to $10.75 per
100 lb., f.o.b. San Francisco. Cupels \Y in. in diameter and 1 in.
high require 30 g. of bone ash, which is equivalent to 15 cupels per
pound.

Cement or Composite Bone -ash-cement Cupels. Portland
cement alone is sometimes used for assay cupels; it is more com-
monly used with various percentages of bone ash. Cement
contains approximately 62 per cent CaO, 22 per cent Si() 2 ,
10 per cent AUOs + Fe 2 3 and minor amounts of MgO, S0 3 ,
and other impurities. When moistened with water and allowed
to dry, various aluminates and silicates are formed, which cause
the cement to harden or "set." The internal bonding of cement
gives cement cupels greater strength than pure bone ash. When
cement is used as all or part of the cupel composition the amount
of mixing water should be approximately 8 to 12 per cent. With

CUPELLATION 51

less than 5 or more than 15 per cent of water, cement cupels
may crack on drying and will usually crack or check on heating.

Pure cement cupels give higher cupellation loss than do bone-
ash cupels, but mixtures of cement and bone ash, up to 75 per
cent cement, give as low percentage of losses as those made of
pure bone ash. 1 A cupel containing 70 per cent cement and
30 per cent bone ash costs about one-third as much for materials
as a pure bone-ash cupel; also, it has much greater strength and
hence is less liable to breakage in handling. However, the
recovery of materials from damaged unused cupels is impractica-
ble if the cement content is high. This is due to the fact that
the setting reactions are not reversible, whereas damaged bone-
ash cupels can be pulverized easily and the bone ash used to
make new cupels.

The cost of cement, in less-than-carload lots, f.o.b. shipping
points in the United States, seldom exceeds $1 per 94-lb. bag or
$3 per 376-lb. barrel.

Magnesia Cupels. A number of manufactured cupels with a
magnesia base are on the market under various trade names,
such as Calmix, DFC Basic, Mabor, Morganite, and others.
Most of these are formed under high pressure and use a special
bond to increase their strength. Magnesia cupels have a higher
heat capacity and thermal conductivity than bone-ash or
cement cupels, and hence the heat of oxidation of the cupelling
lead is abstracted more rapidly. The alloy is therefore main-
tained at a lower temperature than with bone-ash or cement
cupels, but higher muffle temperatures must be maintained
throughout the cupellation cycle. Largely on account of lower
alloy temperature near the finish of cupellation the loss of silver
by cupel absorption is greatly decreased and is usually less than
half of the loss obtained with bone-ash cupels, under analogous
cupellation conditions. The gold loss with magnesia cupels is
the same as with bone-ash cupels.

In 1940 the price of one of the well-known brands of magnesia
cupels, f.o.b. San Francisco, Calif., was $2 per hundred for the
lj^-in. size and $2.40 per hundred for the 1^-in. size. The
cost of bone ash for one hundred l^-in. cupels is approximately
$0.70; for 30 bone-ash-70 cement cupels, the cost per hundred
is $0.25. Even if extra labor were employed to make them, the

1 Ibid., pp. 66-68.

52 FIRE ASSAYING

total cost of officemade bone-ash cupels should not exceed $1 per
hundred. On account of superior uniformity and lower and more
constant silver losses, magnesia cupels are preferred by many
assayers despite the higher cost.

Making Cupels. At present, suitable raw materials for making
magnesia cupels are not readily available to assayers; conse-
quently, few attempt to make this type of cupel at the assay
office. On the other hand, practically all bone-ash, or bone-ash-
cement or cement, cupels in use are made by the assay er or his
assistant.

Various types of cupel-making machines are on the market.
The simplest and leas.t expensive is a ring and plunger. The
ring is set upon an anvil and filled with the cupel mixture, then
the plunger or die is inserted, and the head of the plunger is struck
sharply several times with a wooden mallet or a hammer, slightly
turning the plunger within the ring after each blow. Considera-
ble experience is needed to make cupels efficiently with this
device. Bugbee 1 states that 100 cupels per hour can be made
by this method. Fulton and Sharwood 2 state that "200 cupels
an hour is a fair rate." Aside from limited production capacity
the chief objection to the hand mold is that it is difficult to apply
uniform pressure, with the result that the cupels are of uneven
density, which causes variations in silver absorption.

For greater capacity and uniformity a lever-type cupel
machine is used. A foot-lever machine is preferred to a hand-
lever machine, as it has greater potential capacity and requires
less physical exertion. An average operator should make at
least 200 cupels per hour with the hand-lever type, and 400 per
hour with the foot-lever type. Exceptionally efficient workmen
can double these rates. In addition to molding time, allowance
must be made for the preparation of the batch.

The effect of variations in molding pressure on silver losses was
studied by King, 3 who recommended that, for bone-ash cupels,
pressures of from 800 to 1,600 Ib. per square inch be used, with
mixing water from 10 to 12 per cent, and that, whatever pres-

1 BUGBEB, E. E., " Textbook of Fire Assaying/' 2d ed., p. 92, John Wiley
& Sons, Inc., New York, 1933.

2 FULTON, C. H., and SHARWOOD, W. J., "A Manual of Fire Assaying,"
3d ed., p. 101, McGraw-Hill Book Company, Inc., New York, 1929.

3 Op. cit., pp. 43-50.

CUPELLATION 53

sure-moisture combination is used, it should be kept constant.
The order of magnitude of the silver-loss variations, determined
by King for various pressure-moisture combinations, is not large
it ranges from a low of 1.65 per cent loss on 100 mg. of silver
with 4,000 Ib. per square inch pressure and 15 per cent moisture,
to a high of 1.98 per cent loss with 400 Ib. pressure and the same
moisture content. If desired, uniformity of pressure can be
attained in making cupels in foot- or hand-lever machines, by
providing an extension of the operating lever that may be
weighted with a known weight, such as a bar of lead or a bucket
of shot.

The amount of water used in mixing cupel materials should be
just sufficient so that the mixture will cohere when firmly pressed
in the hands. The optimum amount of water for any of the
bone-ash or cement cupel mixtures, or of either alone, is from
10 to 12 per cent, but a slightly larger, amount may be preferred
for pure bone ash, and cement will set better if the water content
is approximately 8 per cent.

If bone-ash-cement cupels are to be made, the dry ingredients
are mixed thoroughly on a mixing cloth, or in a mixing box or
barrel, and then sifted through a 10-mesh or finer screen.

To avoid puddling, water is added to the dry cupel batch in
small increments, and the batch is kneaded slightly after each
water addition. After all the water has been added, the entire
batch is thoroughly kneaded with the hands and passed through
a 10-mesh or other suitable screen to break up lumps and obtain
more uniform diffusion of the water. The batch should then
be covered with a damp cloth to prevent loss of moisture before
being used, and mixtures containing cement should be used as
soon as possible after adding water, in order to prevent premature
setting of the cement.

Finished cupels are placed in an air drying rack and are ready
for use whenever the free moisture has evaporated. Accelerated
drying, as with hot air, should be avoided, as this may cause
checking or cracking. Seasoning of cupels for longer periods is
entirely unnecessary, although the literature of assaying since
the days of Agricola (1556) 1 is replete with statements that
several weeks' to several months' aging of cupels is essential to

1 AGRICOLA, GEORGIUS, "De Re Metallica" (1556). Translated by H. C.
and Lou Henry Hoover, The Mining Magazine, London, 1912.

54 FIRE ASSAYING

good assay practice. Whatever free moisture may be present
in an air-dried cupel is expelled when the cupels are preheated
in the muffle just before using, a practice which is necessary in
all cases in order to remove combined water, CO 2, and other
volatile matter that would cause spitting and to avoid delayed
opening of the buttons if placed in cold cupels.

CUPELLATION PROCESS

The furnace operations involved in cupellation are divided
into the following stages:

1. Preheating of cupels.

2. Charging of buttons.

3. Opening of buttons.

4. Driving of lead.

5. Finish of cupellation.

6. Removal from muffle.

For the proper conduct of each of these operations the control
of temperature is of primary importance. Accurate temperature
measurements in cupellation are difficult to make, and much
of the literature of cupellation is confusing because of variations
in the technique of temperature measurement among different
investigators. The actual temperature of the button throughout
the entire process, culminating in the bead, cannot possibly be
measured with accuracy. An estimate of the most probable
value of button or bead temperature at any given time can be
made only by interpolation from the known melting-point data
of the metals and alloys involved, as checked against optical
pyrometers or by a correlation with temperatures measured at
some point in the cupel or muffle near the button. Temperature
measurements with a thermocouple pyrometer embedded in a
blank cupel are subject to error because of serious lag behind
temperature changes in the muffle and also because the heat of
oxidation of lead is not available to contribute heat to a blank
cupel, whereas it is an important factor in an active cupel. The
air temperature within a muffle, for example, at a point 2 or 3 in.
above the active cupels is subject to variations that depend upon
draft velocity and furnace characteristics and is greatly different
from the button temperature at all stages of the cycle. Probably
the most satisfactory method of temperature measurement for
the control of cupellation is by means of a thermocouple placed at

CUPELLATION 55

a point % in. above and just behind the central cupel in the row
of cupels under control. This is the method used by King 1
in an extensive series of experiments on the cupellation of silver
and is referred to by him as the "muffle temperature." In
King's experiments, another thermocouple was placed 2 in.
above the top of the central cupel in the row, and the readings
of this pyrometer were called the "air temperature/' which was
from 30 to 40C. less than the muffle temperature.

When muffle temperatures are measured in the manner used
by King, the actual temperature of the button will be practically
the same as the observed muffle temperature when no heat is
being evolved by the button, but during the active period of
oxidation of the lead ("driving") the temperature of the button
rises and reaches a temperature above that of its surroundings.
In this chapter, unless otherwise stated, "muffle temperature"
refers to the temperature as measured by King that is, by a
thermocouple placed J^ in. above and just behind the cupel
under observation.

Few assay offices afford the luxury of a pyrometer for measur-
ing the muffle temperature during cupellation. Consequently,
experienced assayers become skilled in estimating and controlling
muffle temperature and furnace draft at various stages of the
cupellation process. This is accomplished by observing the
temperature color of the muffle, the appearance or absence of
crystals of litharge "feathers" surrounding the cupeling lead,
the manner in which the litharge fumes are carried by the draft
current, and by various other observable phenomena. As a
guide to the beginner a color-temperature scale 2 is given in Table
VI, but the application of this scale is subject to considerable
uncertainty, depending upon the color sensitivity of the indi-
vidual, the conditions of illumination, and the emissivity of
the heated object.

Cupellation is usually performed in a muffle furnace, in which
the products of combustion of the flame do not come in contact
with the cupeling lead. A separate muffle draft, operated in
conjunction with the muffle door, is used to control the amount of
air passing over the cupels. The desirable temperature of

1 Op. rit., pp. 3-4 and Fig. 3, p. 5.

2 HOWE, H. M., Eng. Min. Jour., vol. 69, p. 75, 1900; WHITE, M., and
TAYLOR, F. W.: Trans. Amer. Soc. Mech. Eng., vol. 21, p. 628, 1900.

56 FIRE ASSAYING

TABLE VI. COLOR-TEMPERATURE SCALE

Degrees
Centigrade

Lowest red visible in the dark 470

Dark blood red or black red 532

Dark red, blood red, low red 566

Dark cherry red 635

Cherry red, full red 746

Light cherry, light red 843

Orange 900

Light orange 941

Yellow 1000

Light yellow 1080

White 1205

cupellation is not constant throughout the cycle; hence the
furnace must be designed so that a change of temperature may be
made quickly. This consideration, combined with the desire
for economy of design, has led to the development of furnaces
fired with a liquid or gaseous fuel in which uniformity of tem-
perature throughout the muffle is sacrificed to rapid heating* or
cooling. Liquid or gaseous fuels are better adapted to this
requirement than is coal, coke, or charcoal. On account of
the considerable temperature gradient from front to rear of the
muffle it is not possible to utilize the entire floor area of the muffle
for cupellation, and usually only the central third or the second
fourth of the muffle from the front can be held at a satisfactory
cupellation temperature. Thus, from two to four, or at the
most five, rows of cupels can be cupeled at one time without
overheating the rear rows. A baffle brick occupying the unused
rear of the furnace is an aid to temperature uniformity. The
unused front part of the muffle may be filled with blank cupels,
which are useful to cover large beads before removing them
from the furnace, or for the next batch of cupellations. The
judicious use of hot and cold cupels, or pieces of brick moved to
desired positions in the muffle, will aid in heating or cooling front
or rear rows of cupels in emergencies. If the practice of covering
finished cupels with hot cupels is not followed, a refractory
baffle may be used in front of the cupels.

Preheating of Cupels. Before cupellation the set of cupels
should be charged into the furnace and heated at 850 to 900C.
for 10 min., with the draft slightly open to provide an oxidizing
atmosphere. This will drive off free and combined water,

CUPELLATION 57

organic matter, carbon dioxide, and other volatile constituents
that would otherwise rise through the lead during cupellation
and cause the ejection of particles of lead. This phenomenon is
known as " spitting" and is a source of loss to the assay in ques-
tion, as well as the cause of salting other samples into which the
globules of lead may fall.

The common form of cupel tongs is made to handle one cupel
at a time. For loading blank cupels into large muffles, much
time is saved by the use of the loading device illustrated in Fig.
21, which consists of a tray and a retaining bar. The tray is of
suitable dimensions to hold the desired number of cupels. After
setting the cupels on the tray the retaining bar is placed in
position at the handle end of the tray, and the two are inserted
into the muffle. When the cupels are in the desired position the
retaining bar is held stationary while the tray is slid out from
under the cupels.

Charging of Buttons. After preheating the cupels the buttons
are charged. Although some assay ers make a practice of
charging the largest buttons first, in order that all may finish at
nearly the same time, this practice is unnecessary, and it is more
efficient to charge them in the same systematic order in which
the assays have been carried up to that point. Multiple-button-
charging devices have been designed for charging an entire set
of buttons at a single operation, but they have not been widely
adopted. With practice, more than 30 buttons can be charged
in 1 min. with a single pair of cupel tongs.

Opening of Buttons. After the buttons are charged, the
muffle draft and door should be closed. The lead melts and is
covered with a dark crust composed partly of litharge. The
draft is shut at this stage, to prevent oxidation, which would form
a heavier crust of litharge. As the temperature rises, the crust
melts, and the litharge is absorbed by the cupel. This should
occur within 1 or 2 min. after charging and is the phenomenon
known as " opening." The melting point of litharge is 888C.,
and the actual temperature of the cupel must reach this point
before cupel absorption of litharge begins. The muffle tempera-
ture required for reasonably rapid opening of the buttons is
approximately 900C.

If the opening of buttons is delayed beyond 2 or 3 min. it is
evident that the muffle temperature is too low, that the muffle

58 FIRE ASSAYING

atmosphere is strongly oxidizing, that the cupels were not
sufficiently preheated, or that the buttons contain an excessive
amount of copper, nickel, or other impurity of high melting
point. Buttons containing too much impurity for normal
opening should be rejected, as they can be cupeled only at abnor-
mally high temperature that would increase the loss of silver and
gold. If pure buttons fail to open promptly, even though the
furnace temperature is correct, a small splinter of wood may be
placed in front of the recalcitrant buttons and just enough air
admitted to burn the wood briskly, causing the flame to pass
over the unopened cupels. This reduces the litharge film and
permits the buttons to open.

Driving of Lead. When all the buttons have opened, as
observed through the muffle peephole or through a slight opening
of the door, the door is opened just sufficiently so that the buttons
may be observed, and the draft is adjusted so that there is a
positive current of air passing through the muffle. This is the
beginning of the " driving" stage of cupellation, during which the
lead is oxidizing rapidly.

The heat of oxidation of the lead causes the temperature of
the button to rise considerably above that of the cupel and the
muffle, in the case of bone-ash or cement cupels, but only slightly
above the cupel temperature with magnesia cupels. Therefore
with bone-ash or cement cupels the muffle temperature should
be lowered during the driving period in order to keep the lead
temperature from rising more than the minimum necessary for
the reaction to proceed. The muffle temperature may drop to
840 to 850C. without danger of " freezing," which is caused
by the solidification of a litharge film over the lead surface and
stops further cupellation. With large buttons and small beads
in bone-ash cupels the minimum muffle temperature may drop to
790 to 800C. for a brief period during the peak of the driving
stage. The actual lead temperature probably must be in excess
of 900C., in order that cupellation may continue. The cupel
temperature will be between that of the lead and the muffle.
Under such conditions the lead will have a much higher tem-
perature-color than the surroundings.

With magnesia cupels the temperature gradient from lead to
cupel to furnace is not so great as with bone-ash or cement cupels,
because of the greater heat capacity of magnesia compared with

CUPELLATION 59

the other materials. The button will appear slightly hotter
than the cupel, but the observable difference is by no means so
great as with the bone-ash or cement cupels. Since the minimum
lead temperature must be the same in all cases the muffle tem-
perature with magnesia cupels must be higher during the driving
period than with other types. Muffle temperatures of 870 to
880C., as measured % in. above the muffle floor behind the
cupel, are recommended for the driving stage of cupellation in
magnesia cupels, and temperatures as low as 830C. may often
be successfully used during the period of most active driving.
If temperature and draft conditions are correct during driving,
a fringe of litharge crystals, known as " feathers/' will appear
around the upper edge of a bone-ash cupel. This is due to the
condensation of that small part of the litharge which is volatilized
at the surface of the lead. If the temperature is too high the
volatilized litharge is carried away in the furnace gases or is
deposited on cooler projecting surfaces in the path of the air
stream as it passes through the furnace. If the air draft is too
strong, feathers may appear only on the side of the cupel toward
the draft, or may not form. As cupellation proceeds and the
lead button becomes smaller, concentric rings of feathers are
deposited within the original ring, but at ail times the tempera-
ture should be high enough so that there is a clear area between
the button and the feathers, otherwise there is danger that a
pool of litharge will form immediately around the button, which
will instantly prevent further absorption of litharge by the cupel,
and the litharge will soon completely cover the lead and solidify,
which will cause the button to freeze. Even though the assayer
fails to note the encroachment of feathers toward the button,
or the decrease in color temperature of the lead, the onset of
freezing is plainly evident in the oily appearance of the ring of
molten litharge at the outer periphery of the button, and the
temperature should be raised immediately to avoid freezing.
Individual cupels that show signs of incipient freezing may
be saved by placing a hot cupel or brick near or over them.
This is removed when the average furnace temperature has been
increased sufficiently; or they may be moved to a hotter part
of the furnace, although this latter operation must be conducted
with care to avoid spilling, and to avoid chilling of the cupel
still more by contact with cold tongs.

60 FIRE ASSAYING

Buttons that freeze after driving has once been started should
be rejected. To reopen them by raising furnace temperature,
closing the draft temporarily, and perhaps by using a wood flame
disturbs the normal cupellation of other buttons in the muffle,
and results on the reopened buttons themselves are usually low.
If wood is used for opening one or two frozen buttons in a muffle
rilled with others that are cupeling normally, freezing of the entire
lot may ensue. The freezing would be due to retardation of the
rate of oxidation in the reducing atmosphere, which then fails to
supply sufficient heat for cupellation to continue. Only by
raising the muffle temperature well above the melting point of
litharge before applying a reducing atmosphere can the frozen
buttons be opened without freezing others, and this practice
must be avoided as it will increase silver losses on the entire set.

It is seldom possible to obtain feathers on all cupels in more
than the first two rows in the furnace, unless exceptional precau-
tions are taken and considerable personal attention is given to
each button. To increase furnace capacity, two practices may
be employed:

1. Use bone-ash or bone-ash-cement cupels in the first two or
three rows, then magnesia cupels in an additional two or three
rows. A little experimentation will indicate the proper spacing
of the two types of cupels that will obtain satisfactory cupellation
temperature for all.

2. At the outset of a day's assaying the samples may be
arranged so that all low-silver samples, or samples in which
silver is not to be determined, will fall in such positions that they
will be cupeled in the rear row of the muffle. The losses of gold
are not of so great magnitude as the silver losses; if muffle tem-
peratures do not greatly exceed 900C. the gold loss will not be
excessive.

Among the criteria- for proper cupellation temperature and
draft control are the appearance of the lead and of the fume.
The lead, when driving, shows active motion, from which the
term " driving" was derived. This effect is caused by the rapid
formation of molten litharge at the lead surface, which creates
motion by sliding off the high meniscus of the lead to be absorbed
in the cupel. When the temperature is too high, the motion,
although still present, is less perceptible because the eye is less
sensitive to the difference in temperature color between the lead

CUPELLATION 61

and the molten litharge upon it. The litharge film is thinner,
owing to its greater fluidity. If the temperature is too low the
motion becomes sluggish, and at the same time the button tem-
perature color becomes progressively duller and less in contrast
to the cupel, which indicates that oxidation is not sufficiently
rapid to maintain the temperature of the lead. The fume
effects are less easily identified by the inexperienced assayer.
Under proper cupeling conditions as to temperature and draft
the fume rises sluggishly from the button, curls lazily above the
lead, and is wafted gently toward the rear of the muffle. If the
draft is too weak the fumes are not carried away positively;
if too strong they stream back from the cupels in a straight
line. 1

Finish of Cupellation. Toward the end of cupellation the pro-
portion of lead in the button decreases rapidly, and the melting
point of the alloy rises toward that of the finished bead. At the
same time the rate of oxidation of the lead decreases so that
considerably less heat is being supplied to the button, and its
temperature will drop unless additional heat is supplied from the
furnace. The melting point of pure silver is 961C.; of gold,
1063C.; of platinum, 1773C. The melting point of an alloy
of these metals is approximately the mean of the melting points
of the components. The melting points of lead-silver and of lead-
gold alloys are given in Fig. 7. Down to approximately 20
per cent lead, the melting point of lead-silver-gold buttons is
below the melting point of litharge (888C.); but with less than
20 per cent lead the melting point of the alloy increases toward
that of the gold-silver bead. If the platinum group of metals is
present in excess over gold and silver the melting point of the
alloy with lead rises above that of litharge at a higher percentage
of lead than with gold and silver.

It is not necessary that the muffle temperature be increased
at the finish to the actual melting point of the bead, as the
phenomenon of surfusion keeps the alloy in a molten state at a
temperature somewhat below the melting point. A maximum
of 76C. surfusion has been observed, 2 corresponding to a bead
temperature at the finish of 885C. To finish cupellation without

G, J. T., The Effect of Air on Cupellation Losses, Univ. Toronto Bull
6, 1926.

2 FULTON and SHARWOOD, op. tit., pp. 117-121.

FIRE ASSAYING

danger of freezing or of retention of lead in the bead the muffle
temperature should be 860C. for silver beads up to 100-mg.
weight and in bone-ash or bone-ash-cement cupels. For larger
beads a higher muffle temperature is needed on account of the
greater mass of metal to be kept in a state of surfusion for a
greater length of time. The finishing temperature for silver
beads of the order of 500 mg. or larger should be 900 to 910C.
In magnesia cupels the bead temperature and muffle temperature
are closer together, and a minimum muffle temperature of 890 to
900C. is necessary for completion of the cupellation. Finishing
temperatures should be higher than the foregoing minima when-
ever the bead is exceptionally high in gold or platinum.

1000

o
SJ800

-1
600

|400

9Dfi

s^
^^s

4<7

NT"

-^fc.

$SL<*//

*-ZPo

""-^.

\^t

'-.

P^L

s ^

J27 /

10 20

80 90

40 50 60 70
Weight percent of lead
FIG. 7. Melting points of lead-silver and lead-gold alloys.

100

The visual phenomena that appear near, and at the finish of,
cupellation are of considerable aid to the assayer. Since both
silver and gold have a higher surface tension than lead the button
becomes more rounded as the percentage of lead decreases; this
effect becomes pronounced as the proportion of precious metals
increases beyond 50 per cent. When the cupellation of large
beads approaches completion, oily-appearing drops of litharge
can be seen to collect on the surface of the bead. These particles
appear to move, and the movement becomes increasingly more
rapid until just before the finish, when the molten litharge forms
a thin film of variable thickness and creates interference colors.

CUPELLATION 63

The rainbow color bands move swiftly over the surface of the
button and give the illusion that the button is revolving about a
shifting axis. This is known as the "play of colors" and is
strikingly developed with large beads. When the last trace of
lead has been removed from the bead the play of colors disap-
pears and the bead becomes dull for a brief period, after which
it acquires a normal metallic luster. This change of luster is
known as " brightening" and is not always observable.

If the bead was in a state of surfusion at the finish and con-
sisted of nearly pure gold and silver it would solidify upon further
cooling with the emission of a flash of light (known as the "blick"
or "flash"). This flash is due to the sudden release of the latent
heat of fusion of the alloy at the moment of solidification, which
momentarily raises the temperature considerably. The blick
is seldom perceptible in beads larger than 700 mg. Small
amounts of lead or copper in the bead diminish the intensity
of the blick. All the platinum group of metals, with the excep-
tion of platinum and palladium, suppress the blick entirely.
If the blick is observable, it is a useful guide to the completion
of cupellation, but the assayer need not waste time trying to
observe the blick, as the end of cupellation is easily ascertained
by other observations. The beads may be left in the furnace for
a sufficiently long time thereafter to avoid the consequences
of removing them prematurely.

Liquid silver has the peculiarity of dissolving a large volume of
oxygen, which is expelled upon solidification. The solidifica-
tion of large beads starts at the surface. When the center of
silver beads solidifies, the oxygen is sometimes expelled violently,
spewing forth some of the interior silver and forming a cauli-
flo/ferlike growth on the bead. This mechanism is known as
"sprouting-" Sprouted beads should be rejected if there is any
reason to believe that any particles of metal have been lost. To
avoid sprouting, all large beads should be cooled slowly, either
by leaving them in the furnace after -the blick, and until certain
that they have cooled sufficiently, or by moving them to a slightly
cooler zone in the furnace and then covering them with a very
hot cupel before they have completely solidified. The hot cupel
melts the outer crust and allows the bead to solidify slowly, so
that the oxygen can escape from the interior without violence.
When the bead contains more than one-third of its weight in

64 FIRE ASSAYING

gold, it docs not sprout ; hence beads known to be of this composi-
tion may be removed from the furnace as rapidly as desired.
Sprouting of silver beads is considered a sign of purity, but this
indication is of no practical value to the assayer, as sprouting
should always be avoided.

Removal from Muffle. After the brightening, cupels may be
removed from the muffle if precautions are taken to avoid
sprouting. In the past, many assayers have thought that beads
should be removed from the muffle within 1 or 2 min. after the
blick in order to avoid loss. King 1 found that the increased loss,
due to leaving cupels with 100-mg. silver beads in the muffle for
1 hr. at the finishing temperature of 870C., was only 0.01 per
cent compared with a total cupellation loss of 2.07 per cent under
identical conditions when buttons were removed immediately
after the blick. Since other investigators have already agreed
that no loss of gold occurs by leaving beads in the muffle after
finishing, it is evident that all beads may be left in the furnace for
any reasonable time after the finish of cupellation, provided that
the muffle temperature is not increased greatly above the proper
finishing point. Presumably, if the temperature were raised
above the melting point of the bead, increased cupel absorption
and a slight volatilization loss would occur.

The lack of necessity for prompt removal of the cupels is a
great convenience to the assayer, because all cupels may be left
in their original position in the muffle until the last one has
finished, and their subsequent cooling and removal can be done
systematically and efficiently without danger of misplacing the
sequence of the assays or of interfering with the final stages of the
cupellation of unfinished beads. If desired, a scraper may be
used to pull the cupels forward, a row at a time, where theyVill
cool slowly without danger of sprouting thus dispensing with
the nuisance of covering them with a hot cupel. With this
procedure a refractory baffle brick may be used in place of blank
cupels in front of the muffle. This is removed jn a single motion
before drawing the finished cupel forward and is replaced before
the next batch of cupellations is started. While waiting for
the finished cupels to cool in the front of the muffle, fresh cupels
may be loaded in position for the next set.

1 KING, J. T., The Influence of Cupels on Silver Loss, Univ. Toronto
Bull. 147, pp. 40-41, 1934.

CUPELLATION 65

When the finished beads have cooled sufficiently, the cupels
are transferred in the established order to the cupel trays.

After the bead is cool enough to permit handling, it is gripped
with a curved-nose pliers and pulled away from the cupel, to
which it should adhere firmly. The under surface should appear
frosted but free from roots that extend into cracks in the cupel.
Particles of cupel material adhere to the frosted surface and are
removed first, by squeezing the bead strongly with the pliers in
a direction at right angles to the bottom surface or by flattening
large buttons slightly with a hammer and then brushing the bead
vigorously with a stiff bristle brush. The bead is then flattened
on a small clean anvil with a mineralogist's hammer and is
transferred to a clean parting cup. The flattening of large beads,
as in the bullion assay, is usually done in hardened steel rolls, with
frequent annealing to prevent cracking caused by work hardening.

During the operation of cleaning and flattening the bead the
assay er must exercise extreme vigilance to avoid losing the bead.
In order to minimize the risk of loss the work should be done on a
clean smooth bench top or on a sheet of glazed paper in which the
anvil is centered. A protecting shield or flange around the work-
ing space is desirable. The cupel should be placed close to the
anvil to avoid carrying the bead in the pliers any farther than
necessary. When the bead is brushed, the convex side of the
pliers should be steadied against the anvil, and only a downward
brushing movement should be employed. To flatten the bead
the blows of the hammer must fall squarely upon the bead, which
will eliminate the danger of projecting it away from the anvil.
When the bead is flattened, the parting cup is placed alongside
the anvil, the anvil is lifted and tilted over the cup, and the bead
is gently slid into the cup with the aid of the pliers.

Pure silver-gold beads, if properly cupeled, will have the
characteristic color of silver-gold alloys in which the gold color
is not easily apparent until at least 50 per cent of gold is present.
Since both gold and silver have a high surface tension the beads
will be well rounded, and the small beads will be nearly spherical
in shape. The bottom surface of the bead will be rough, and the
bead will adhere firmly to the cupel. Any deviation from these
conditions, particularly a dull luster, flattening, smooth bottom
surface, or lack of adherence to the cupel, may indicate the pres-
ence of lead, showing that the cupellation temperature at the

FIRE ASSAYING

finish was too low. The same abnormalities may, however, be
caused by the presence of other impurities. Therefore, at the
time the buttons are removed from the cupel and prepared for
weighing and parting, the bead and the cupel should be examined
tor evidence of impurities or for detection of the possible presence
of metals of the platinum group.

Summary of Cupellation -temperature Cycle. Characteristic
temperature cycles in cupellation are shown on Fig. 8, based on
measurements taken with a pyrometer J^ in. above the muffle
floor just behind the front row of cupels. These curves apply
particularly to buttons from 20 to 25 g. in weight and with beads

900

o -TJ860

SJB
*780

'700

-^^

860 $ ."

48 12 16 ' '12 8 40

Minutes after charging Minutes before finishing

FIG. 8. Characteristic cupellation-temperature cycles.

weighing 50 mg. or more, cupeled with feathers. The indicated
temperatures are subject to corrections based on the furnace
draft, heating and cooling lag, and the pyrometer position. The
minimum temperature allowable during the driving depends also
upon the type of cupel used and the size of the button and bead.
Large buttons that are cupeled rapidly in bone-ash cupels with
the ordinary shallow cup frequently permit minimum muffle
temperatures as low as 790C. and will finish at 820C. if the
beads are small. Magnesia cupels under similar conditions
require minima of 830C. in the driving trough, and 840 to
850C. to finish. The formation of feathers of litharge can be
observed readily with bone-ash or bone-ash-cement cupels and
serves to indicate proper driving temperature, but when copious

CUPELLATION 67

feathers form on magnesia cupels the temperature is dangerously
near the freezing point.

Finishing temperatures greatly in excess of 900C., as measured
in the manner indicated above, should be avoided in all cases,
as higher temperatures cause greatly increased losses of gold
and silver with all types of cupels and all variations in button
and bead weight.

Evidence of Impurities. The presence of impurities in lead
buttons usually influences adversely the accuracy of assaying.
In general the errors caused by impurities in cupellation may be
either positive, due to retention of impurities in the bead, or
negative, due to increased cupel absorption and to a slight extent
to increased vaporization loss. Increased absorption and
volatilization losses are usually due to the necessity for higher
cupellation temperature to finish the process, but increased
absorption of gold and silver may be caused by lowered surface
tension of the button (as with tellurium and selenium), and
increased volatilization loss may be caused by lowered vapor
pressure of the precious metals or by mechanical losses if the
impurity burns violently.

Elements below lead in the electromotive series, such as bis-
muth, copper, and tellurium, tend to become concentrated in
the lead during cupellation and to be retained to a certain extent
by the bead. Elements such as copper, nickel, tin, and zinc,
whose oxides are infusible at cupellation temperature and are
not greatly soluble in litharge, tend to form a crust over the
cupeling bead and inhibit the driving. Even if the buttons can
be finished without freezing, the higher cupellation temperature
required to finish such buttons increases the loss of silver and gold.

The observable cupellation phenomena that indicate impurities
in lead buttons are abnormal fumes and cupel stains, scoriae,
and bead peculiarities. Exceptionally dense fumes may be
observed from buttons containing antimony, arsenic, or zinc,
each of which burns during cupellation and forms an oxide that
condenses in the fume.

Any buttons that are hard or brittle may contain impurities
that are detrimental to cupellation, but not all such impurities
cause perceptible hardness or brittleness. Some of the embrit-
tling impurities, such as PbO or PbS, do not hinder successful
cupellation.

68 FIRE ASSAYING

The identification of the nature of impurities in lead buttons
can sometimes be made by observation of the slag obtained from
the assay fusion or may be deduced from the recognition of
minerals in the ore that are known to be a potential source of
button impurity. When the assay er fails to identify the inter-
fering element in the ore, slag, or button, an ex post facto identifi-
cation can sometimes be made by indications that appear during
cupellation. In a normal cupellation of pure lead the cupel will
be free from unabsorbed infusible residues (scoriae), and will
be stained a brownish-yellow color that is nearly uniform through-
out the entire stained portion of the cupel. The litharge feathers,
if present, will have a typical yellow color that the assayer soon
learns to recognize.

Abnormal cupel stains that persist throughout the entire
stained portion of the cupel are caused by such impurities as
copper, iron, and nickel, whose oxides, although infusible at
cupellation temperatures, are partly soluble in litharge. Abnor-
mal cupel or " feather" stains that appear mainly on the surface
of the cupel usually near the bead are caused by bismuth
or tellurium, which concentrate in the button toward the end of
cupellation, and hence their oxides do not have an opportunity to
penetrate far below the surface of the cupel. Scoriae are formed
whenever more of an infusible oxide is produced than can be dis-
solved by the litharge, leaving partly fused crusts on the cupel
surface. Practically all the interfering metals, with the excep-
tion of bismuth, will form scoriae when in excess.

The impurities likely to be found in lead buttons are antimony,
arsenic, bismuth, copper, nickel, selenium, and tellurium.
Cobalt, iron, manganese, and zinc are not likely to be found in
beads resulting from an assay fusion, as they are easily removed
in the slag. Metallic aluminum or zinc may be present in
cyanide precipitates that are cupeled directly, but if present in
appreciable amounts they should be removed by a prior fusion
or acid treatment.

Antimony is indicated by dense fumes of antimony trioxide
in the early driving stage, and by yellow scoria of lead antimo-
nate, which appear if more than 2 per cent antimony is present.
The lead antimonate expands upon solidification and causes
cupel cracking, which may result in leakage of lead from the
button. The effect of antimony on accuracy is to increase the
absorption loss of gold and silver. The increased loss is not

CUPELLATION 69

serious if the cupellation is conducted at normal temperature
and if scoriae and cupel cracking are absent.

Arsenic exhibits the same cupellation phenomena as antimony,
but the fumes are less perceptible and the scoriae are light yellow.
The effect on accuracy is nearly the same as with antimony.
Arsenic is seldom present in interfering proportions in lead but-
tons, as most of it is easily eliminated during the fusion process.

Bismuth acts like lead in cupellation, and cupels at a slightly
lower temperature. It is indicated by orange-yellow or blackish-
green feathers and by a brown stain on the cupel. If both lead
and bismuth are present the bismuth concentrates in the button
during cupellation, and the brown stain from bismuth will be
more intense near the bead. Gold and silver losses are not
excessive when cupellation is conducted in the presence of
bismuth, but some bismuth is retained in the bead, requiring a
correction to the silver result if high accuracy is desired.

Copper is one of the commonest impurities in assay buttons.
Copper oxide is infusible at cupellation temperatures, but it has
considerable solubility in molten litharge, and some copper is
therefore removed with the lead. Small amounts of copper give
a dirty greenish cupel stain, larger amounts give a dull-black
stain, and still larger amounts leave scoriae and cause freezing.
Cupel cracking is commonly observed if enough copper is present
to give a black stain. If the ratio of lead to copper is greater than
500 to 1, cupellation losses of gold and silver are nominal,
although some copper may be retained by the bead. Smaller
ratios of lead to copper cause noticeably increased absorption
loss, accompanied by increased retention of copper in the bead.
Most assayers accept cupellations from copper-bearing buttons
if the buttons finish at normal temperature without scoriae and
the cupel has a greenish, rather than a black, stain.

Nickel is similar in its cupellation action to copper, but its
effects are apparent at a much lower percentage. The stains
are greenish, and the scoriae are dark green. Comparatively
small amounts cause freezing at safe cupellation temperatures.
Nickel increases absorption losses in cupellation, and some
nickel is retained in the bead.

Tellurium and selenium are extremely detrimental to the
accurate recovery of gold and silver in cupellation, when present
in amounts greater than 1 per cent in lead buttons. Both ele-
ments lower the surface tension of the lead button and allow

70 FIRE ASSAYING

some of the alloy to be absorbed by the cupel; the rest of the
alloy may divide into separate particles, giving a number of
individual beads. This effect is minimized by the use of fine-
textured hard-finished cupels. Comparatively small amounts
of tellurium leave a pink stain on the hot cupel, which disap-
pears on cooling. If more than 15 per cent of tellurium is
present in the bead, it is dull and frosted. Gold losses in the
cupellation of buttons containing tellurium are much higher
than with other impurities mentioned herein, which affect silver
far more than gold. Selenium probably increases the volatiliza-
tion loss of gold and silver, 1 but tellurium does not cause abnormal
volatilization loss when the cupellation is carefully conducted.
The total loss of gold in the cupellation of buttons containing
selenium and tellurium is decreased by the presence of silver;
it is therefore advisable to add silver, if necessary, to ensure at
least three times as much silver as gold being present, which
ratio is necessary for parting in any event.

Evidence of Platinum -group Metals. The metals of the
platinum group of metals are ruthenium, rhodium, palladium,
osmium, iridium, and platinum. Each of these metals produces
distinctive effects on the structure and appearance of the cupella-
tion bead, and a few cause characteristic phenomena during the
latter stages of cupellation. Since all the metals of the group are
rare and have a considerable market value the assayer should
seek evidence of their presence in order to aid the development
of hitherto unknown sources of supply.

One of the significant effects of the presence of very small
quantities of ruthenium, rhodium, osmium, and iridium but
not of palladium or platinum is that flashing is prevented.

J. L. Byers 2 has amplified earlier observations by Lodge, 3
Bannister, 4 Bannister and Patchin, 5 and Bugbee, 6 by systema-

1 FULTON and SHARWOOD, op. cit., pp. 125-126.

2 BYERS, J. L., Surface Effects on Assay Beads Caused by Metals of the
Platinum Group, A.I.M.E. Preprint, 1932.

3 LODGE, R. W., " Notes on Assaying," 3d ed., pp. 239f., John Wiley
& Sons, Inc., New York, 1915.

4 BANNISTER, C. O., Trans. A.I.M.E., vol. 23, pp. 163-173, 1894.

5 BANNISTER, C. O., and PATCHIN, G., Jour. Chem., Met. Min. Soc. South
Africa, vol. 14, p. 478, 1913.

6 BUGBEE, E. E., "Textbook of Fire Assaying," 2d ed., pp. 112-114, John
Wiley & Sons, Inc., New York, 1933.

CUPELLATION 71

tizing the evidence of platinum metals in gold and silver beads
and establishing a rapid approximate quantitative method of
determination, dependent upon bead surface effects. For each
of the six metals of the group alloyed with silver and with gold,
Byers presents photomicrographs at 15 diameters of beads con-
taining four different amounts of rare metals.

Ruthenium. Ruthenium has a more marked effect on the
appearance of gold or silver beads than any other member of the
group. A blue-black encrustation of crystals of ruthenium
dioxide appears at the lower edge of the bead, but not on the
bottom, with as little as 0.004 per cent Ru; it spreads from that
region with increased amounts of ruthenium until the entire
bead is encrusted. Byers states that the encrusted area approxi-
mates 1 per cent of the surface of the bead not in contact with
the cupel, in the 0.1 per cent Ru-Au or Ru-Ag bead, 100 per cent
in the 2 per cent Ru-Au bead, and 100 per cent in the 2% per
cent Ru-Ag bead. The area covered is a direct function of the
ruthenium content. The unen crusted surface of Ru-Au beads
is scaly or platy, and the surfaces of the plated crystals are
striated, with a faint dendritic pattern. The gold grains are
moderately pitted. The unencrusted surface of Ru-Ag beads
has a silver luster, marked only by dendritic striations.

Rhodium. Bugbee 1 states that as little as 0.004 per cent of
rhodium in silver beads causes a distinct crystallization, which
is more apparent at 0.01 per cent Rh, and that 0.03 per cent Rh
causes unavoidable sprouting of silver beads. Byers does not
report sprouting but emphasizes iridescent color films on Rh-Au
beads and microscopic pits on Rh-Ag beads. The color film
on Rh-Au beads changes from gold, with a reddish tinge at
0.1 per cent Rh to an iridescent purple at 2 per cent Rh. With
Rh in excess of 0.1 per cent the surface structure of Rh-Au beads
becomes distorted, and a polyhedral structure develops, which
is strongly accentuated at 0.5 per cent Rh. At 1 per cent Rh the
spherical surface entirely disappears and the bead surface is
composed of conchoidal facets, up to 1 mm. in largest dimension.
Beyond 1 per cent Rh the number of facets increases and the bead
once more approaches the shape of a pure gold bead. In Rh-Ag
beads, iridescent films are not distinctive, but beads in the range
from 0.02 to 5 per cent Rh are covered with microscopic pits

1 Ibid.

72 FIRE ASSAYING

usually filled with a dark crystalline material, which entirely
covers the surface of beads containing more than 2 per cent Rh.
Rhodium-silver beads containing less than 2 per cent Rh also
show a slight dimpling, and those containing more than 2 per cent
Rh appear to be collapsed.

Palladium. The principal effect of palladium on gold or
silver beads is increasing lack of definition of grain boundaries,
accompanied by a somewhat pebbled appearance with increasing
amounts of palladium in the range from 0.1 per cent to 3 per cent
Pd. The color of Pd-Au beads varies from a brilliant gold at
0.02 per cent to a silvery bronze at 5 per cent. Palladium-gold
beads in the range from 3 to 5 per cent Pd lose their spherical
smoothness and develop ridgelike grain boundaries, giving a
polygonal appearance with slightly conchoidal facets. Pal-
ladium-silver beads retain their spherical surface, marked by
slight dimpling on the sides and top of the bead. The presence
of Pd in Au-Ag beads may be verified by an orange-colored
solution, due to colloidal Pd, in the nitric or sulfuric acid parting
of the bead.

Osmium and Iridium. It is difficult to distinguish between
osmium and iridium in assay beads without chemical tests.
Both metals, which usually occur together in nature, decrease
the surface tension of Au and Ag, which is first apparent in the
formation of a creased dimple in the central portion of the bead
surface. Dimpling appears in Os-Au beads containing as little
as 0.02 per cent Os, and in Ir-Au beads at 0.1 per cent Ir. With
5 per cent Os or Ir in gold beads, and with 5 per cent Ir in silver
beads, the shape of the bead becomes considerably flattened and
irregular because of extreme dimpling. At 10 per cent Ir in
Ir-Au beads the surface tension is so low that the deposit, after
cupellation, resembles scattered pieces of foil. In Os-Au, Ir-Au,
and Ir-Ag beads a lacy fretwork appears in and near the surface
of the dimples, whereas the rest of the bead surface is platy.
Osmium-silver beads do not exhibit marked differences from
pure silver beads, except for a slight decrease in grain size, faint
dendritic striations, gentle undulations of the bead surface, and
a silken luster. Beads containing 5 per cent or more of Ir in gold,
or 2 per cent or more in silver, contain a small amount of dark
crystalline material scattered over the bead surface, particularly
near the base. In beads containing 5 per cent Ir, or an alloy of

CUPELLATION 73

Os and Ir, vugs containing a pinkish-brown resinous material
may be found near the bottom of the bead.

Platinum. The general effect of Pt on gold and silver beads is
to develop a frosted appearance and a dulling of the luster due to
a decrease in the size of the surface grain, to a pitting of the
exposed grain surface, and to well-developed striations on the
grain face. These effects are not so well marked with Pt-Au
as with Pt-Ag beads. As little as 0.1 per cent Pt can usually
be detected under low-power magnification, even with com-
paratively small beads. At 5 per cent Pt in a Pt-Au bead,
the surface of the bead is broken by intertwining ridges and
valleys; and from 3 per cent Pt and upward in Pt-Ag beads,
the crystalline surface is gradually replaced by a surface with a
multitude of microscopic pits. Buttons containing a large
amount of Pt flatten toward the end of cupellation and may
freeze, leaving a gray, rough bead that sticks to the cupel.

CHAPTER V
PARTING

In the determination of gold and silver the beads from cupella-
tion are weighed and recorded as the weight of gold-silver alloy,
commonly called "dore*." The silver is then dissolved in acid,
an operation called "parting," and the gold is washed, dried,
annealed, and weighed. The weight of silver is determined by
subtracting the weight of the gold from the weight of the dorr*.

When an assay for silver is requested and gold is not to be
determined, it is customary to report the total weight of the
cupeled bead as silver and to omit parting. Parting cannot be
omitted in the assay of a gold ore because all gold ores contain
some silver, which would represent too great an error if it were
reported as gold.

The objectives of parting are to dissolve rapidly and remove
all the silver and any base metals that may be present in the
bead, and to leave the gold in a form that can be manipulated
without mechanical loss, preferably in a single piece. In order to
attain these objectives, the following factors should be considered :

1. Preparation of the bead for parting.

a. Ratio of silver to gold.

b. Inquartation.

c. Flattening the bead.

2. Choice of parting receptacles.
a. Parting cups.

6. Parting flasks and other glass receptacles.

c. Platinum or f used-silica porous or slotted thimbles.

3. Conditions of solution,
a. Kind of acid.

6. Purity of acid and wash water.

c. Acid temperature.

d. Acid quantity.

e. Acid concentration.
/. Time of contact.

g. Effect of base-metal impurities in the bead.
h. Indications of metals of the platinum group.

PARTING 75

4. Removal of silver-bearing acid from the gold.

a. Decantation and washing.

b. Floured gold from parting.

5. Preparation of gold for weighing,
a. Drying.

6. Annealing.

Ratio of Silver to Gold. A high proportion of silver in the alloy
to be parted increases the rate at which silver is dissolved, but
too high a proportion of silver tends to cause the gold to break up.
An alloy containing less than two times as much silver as gold
will not part, even in strong acid, and a still higher proportion of
silver is required for rapid parting in dilute acid. When parting
beads that contain more than six times as much silver as gold,
the gold may break up, particularly if the bead weighs more than
2 ing. The gold in small beads breaks up less readily and is
usually obtained in a single piece when beads smaller than 2 mg.
are parted, even though ten times as much silver as gold is
present.

Inquartation. In order to ensure complete parting without
breaking up of the gold, assay beads containing too little silver
must be inquarted with additional silver. Usually a total of
four or five times as much silver as gold is required, except when
masses of 200 mg. or more of gold are to be parted, as in the
gold-bullion assay, when the ratio of silver to gold is 2:1 or 3:1.
Early assayers believed that the best alloy for parting contained
Qiie-fourth gold and three-fourths silver. In fact the literal
translation of the Latin word for parting, quartatio, is "fourth-
ing." 1

If the original sample is known in advance to contain too low a
silver-gold ratio for parting, inquartation may be done in the
assay fusion or in cupellation. If silver is not to be determined,
the amount of silver required is estimated approximately and is
usually added as silver foil to the assay fusion or to the lead
button before cupellation. If the determination of silver is
required, the amount of silver added for inquartation must be
accurately known. In Western United States, Herman inquarts
are commonly used for this purpose. These consist of small
squares of sheet lead containing a known amount of silver.
Another common method is to add a standard silver nitrate

1 CRAMER, JOHN ANDREW, "Elements of the Art of Assaying Metals,"
2d ed., p. 194, 1764.

76 FIRE ASSAYING

solution to the crucible charge. It is convenient to prepare
solutions of two different strengths, one for low-grade samples
and one for concentrates and other high-grade samples, so that
in either case not more than a few milliliters of solution are
required. Some assay ers make a separate fusion for silver and
gold, adding sufficient silver at the outset for parting the gold
assay.

In some cases, the need for inquartation may not be suspected
until after the silver-gold bead is obtained, when the gold color
of the bead or the failure to part in strong acid may indicate that
the silver-gold ratio is too low. After the dore* weighing, inquar-
tation can be accomplished by cutting off a piece of silver foil
approximately equal to three times the weight of the bead,
wrapping the bead and silver together in a piece of lead foil, and
cupeling. The resulting bead is parted directly without reweigh-
ing. Inquartation by fusion of the bead and added silver on
charcoal with a blowpipe may be accomplished, but this method
is slow, and it is difficult to obtain completely homogeneous
alloys that will part without breaking up of the gold.

Although c.p. silver foil is generally free from gold, each lot
should be tested when used for inquartation to be certain that
there is not enough gold present to affect the gold assay. Five
or ten grams of silver should be used for the gold determination.
The silver is dissolved in nitric acid, diluted slightly, and the gold
is then carried down by the precipitation of a small part of the
silver with hydrochloric acid or sodium chloride. The precipitate
is settled, filtered, dried, scorified with lead, cupeled, and parted.

Flattening of the Bead. The thinner the silver-gold alloy the
more rapid and complete is the parting and the more coherent
the gold. Ordinary assay beads, up to perhaps 500 mg. weight,
are flattened with a small hammer on a steel anvil. Larger beads,
as in the bullion assay, are rolled thin in successive stages between
hardened and polished steel rolls, then coiled into a spiral called
a "cornet." Strain hardening of large beads, which might cause
them to crack in rolling, is removed by annealing at a red heat at
intervals between successive rollings.

In the gold-bullion assay the operation of flattening the bead
should be standardized to ensure uniformity of parting. The
United States mint method of preparing cornets is to place the
cleaned bead on an anvil used only for that purpose and flatten

PARTING 77

it by a middle blow and two end blows, then reduce to the
approximate thickness of a visiting card, nearly 0.01 in., in two
passes through the bullion roll. The cornet is then rolled so that
a uniform space is left between all turns of the spiral.

Parting Cups. Ordinary assay beads are usually parted in
parting cups, although some assay ers prefer glass flasks or test
tubes, and platinum containers are used at United States mints
for parting gold bullion.

Parting cups are glazed porcelain crucibles of about 10 or 15
ml. capacity. They are conveniently handled in groups by
means of parting-cup trays, which keep the cups in order and
hold them upright.

When using parting cups the dor6 beads are taken out of the
cupels, cleaned, flattened, and placed in order in a tray of parting
cups. The tray is then taken to the assay balance where the
dore beads are weighed. Here the parting cups have an advan-
tage over bead trays because the individual parting cups can be
picked up and the bead dumped onto the balance pan. This
operation is easier than picking up the beads with forceps,
particularly when the beads are small. After each bead is
weighed, it is dumped back into its parting cup and remains
therein during the subsequent operations of parting, washing,
drying, and annealing, until the final weighing of gold is made.

Parting Flasks or Test Tubes. The parting operation is
essentially the same in either glass flasks or test tubes, except
that flasks are usually placed directly on a hot plate during
parting, and test tubes are held in a rack placed in a vessel of
boiling water during parting. The capacity of the flasks or test
tubes generally used varies from 30 to 60 ml. They are especially
desirable for parting large beads, particularly when the silver is
near the minimum parting ratio, and prolonged heating at or near
the boiling temperature is required to secure complete parting.
The neck of the flask or test tube acts as a condenser and prevents
rapid evaporation to dryness. Furthermore the danger of
mechanical loss of gold by boiling over is minimized. The chief
disadvantage of flasks and test tubes is that the gold must be
transferred to another receptacle for annealing. Unglazed fire-
clay annealing cups are used for this operation. Such cups are
porous and absorb the small amount of water remaining with tjie
gold after decantation so that they can be heated rapidly to

78 FIRE ASSAYING

the annealing temperature without danger of loss of gold by
spattering.

To prevent bumping in glass or fused-silica parting receptacles
a small fragment of charcoal, electric-arc carbon, graphite
crucible, or other similar material is added with the cornet.
Bumping is absent when platinum vessels are used.

To transfer the parted and washed cornet from a flask into an
annealing cup, fill the flask with water, place the inverted anneal-
ing cup over the neck of the flask, and quickly invert the assem-
bly. The cornet will settle into the annealing cup and the flask
is then removed by quickly raising and inverting it. The water
in the cup is decanted off, and the cornet is ready for annealing.

Platinum or Fused-silica Thimbles. Perforated or slotted
platinum thimbles for parting gold bullion are used in United
States mints and at some refineries having a large number of daily
bullion assays. A number of thimbles are carried in a suitable
rack or frame of platinum, and the entire set is immersed in a
parting-acid bath contained in a platinum dish. Perforations or
slots in the bottoms of the thimbles permit the acid to enter and
attack the beads. When parted, the entire set is lifted out of the
acid, drained, and immersed in a second acid bath, followed by
successive washings. The principle of this method of parting
obviously could not be used if there were danger of any of the
gold breaking up, and in any event few assay offices could afford
the outlay for the large quantity of platinum ware required. It is
possible, however, to obtain fused silicaware of a similar character,
at considerably less cost.

Kind of Acid. Sulfuric acid is used in parting commercial
quantities of bullion, but nitric acid is almost universally used
in assaying because of the difficulty of handling hot sulfuric acid
and because of the danger of spattering during dilution of hot
sulfuric acid with water.

Purity of Acid and Wash Water. Particular precautions must
be taken to ensure that the acid, dilution water, and wash water
used in parting are free from chlorine or chlorides. Chlorine and
chlorides in the parting acid and wash water precipitate silver
chloride, which may coat the bead and prevent parting or may
precipitate in the pores of the parted gold and add to its weight.
If there is not sufficient silver present to precipitate all the chlo-
rine, some gold may be dissolved and lost.

PARTING 79

To ensure freedom from chlorine, all new lots of parting acid
should be tested, and the distilled water should be checked at
frequent intervals. Chlorine is present if a white turbidity or
precipitate of silver chloride is formed when a small amount of
silver nitrate is added. In conducting the test, which is very
delicate, the assay er- should be sure that the containing vessels
are thoroughly clean and that chlorine is not accidentally intro-
duced by perspiration from the hands or other parts of the body.

To test for chlorates in nitric acid, take 200 ml. or other con-
venient quantity of the acid, add a few milliliters of 1 per cent
silver nitrate and 4 g. of a metal such as zinc, aluminum, silver,
or copper, heat nearly to boiling to dissolve the metal, and reduce
the chlorate. Tho formation of a precipitate of silver chloride
indicates the presence of chloric acid.

If it is necessary to use acid or water containing chlorides, the
chlorine may be removed by precipitation with silver nitrate.
For this purpose, add silver nitrate solution, little by little, just
to the point where no more precipitate forms. Let the precipitate
settle and remove the chloride-free solution by decantation or
siphoning.

If nitric acid is used in parting, it should be free from sulfuric
acid, arid vice versa, as the two acids together dissolve some gold.
The presence of sulf ates in nitric acid is determined by the forma-
tion of a precipitate of barium sulf ate when barium chloride is
added.

Acid Concentration. The stronger the acid the more rapid
and complete is the parting, but the greater the tendency for the
gold to disintegrate. Hence, strong acid is necessary for parting
alloys having the minimum allowable ratio of silver to gold, and
weaker acid is desirable to avoid breaking up of the gold in silver-
rich alloys. For example, 200 to 1,000 mg. samples of bullion
containing 3 parts of silver to 1 part of gold may be started in
1:2 nitric acid (1 part concentrated acid to 2 parts of water by
volume, sp. gr. 1.17*) and finished in 1 : 1 (sp. gr. 1.25) or stronger
acid. Beads containing 10 mg. of gold and more than 6 parts of
silver to 1 part of gold, part almost completely without breaking
up of the gold in hot 1 :6 acid (sp. gr. 1.07). For more complete
removal of silver this should be followed by a 30-min. heating in
1 : 1 acid. Long-continued boiling in strong nitric acid will
dissolve an appreciable amount of gold and should be avoided.

80 FIRE ASSAYING

At most assay offices handling ordinary mine and mill products
other than bullions the parting-acid strength is arbitrarily chosen
to give good results on the majority of routine samples. The
most frequent strengths in use are from 1:4 to 1:8. The con-
venience of using a stock acid solution offsets the occasional
annoyance of broken gold partings, and if the acid proves insuffi-
ciently strong for certain beads, additional strong acid may be
added, drop by drop, until parting starts.

The necessity for a double acid treatment is avoided by many
assayers for all except large beads by ensuring that all samples
contain enough silver for complete parting by a single treatment
of 15 to 20 min. duration in the stock acid.

Temperature of Acid. Best results are obtained as to speed,
completeness of parting, and coherence of gold if the parting acid
is heated before use and is kept just below boiling during contact
with the alloy to be parted. Boiling is undesirable, as the bump-
ing causes mechanical disintegration of the gold, and boiling over
would cause actual loss. To save manipulative time by the use
of convenient acid-dispensing devices, which are not available
for handling hot acids, many assayers handling a large volume of
routine work start with cold acid of somewhat greater strength
than would be used with hot acid. Although parting with cold
acid at the start requires at least twice as long as when hot acid
is applied to the beads, this time is not lost when the assayer is
occupied with other duties during the interim. The tendency of
the gold to disintegrate may be accentuated when parting is
started in cold acid, but there is no proof that this is necessarily
true if the acid strength and silver-gold ratios are properly
adjusted.

Acid Quantity. The theoretical amount of silver that can be
dissolved by 10 ml. of 1 :6 nitric acid is 2.4 g., but in practice, on
account of loss of volume by volatilization, a 10-ml. parting cup
will not hold enough 1 : 6 acid to dissolve much more than 500
mg. of silver without replenishment of the acid. Hence for
large beads a larger parting cup is needed, unless a double acid
treatment is employed. Parting flasks are desirable if more than
1,000 mg. of silver is to be dissolved, as 30 to 40 ml. of acid may
be used.

Time of Contact. With high silver-gold ratios, parting is
practically complete when the evolution of gas bubbles ceases,

PARTING 81

provided the acid has not been consumed by then. The great
majority of routine assays containing up to 20 mg. of gold and
suitably high ratios of silver to gold will part with a sufficient
degree of completeness in 15 to 20 min. in 10 ml. of hot 1:4 to
1:8 acid. If more gold is present, or if the silver-gold ratio
approaches the critical minimum, it is necessary to decant off the
first acid after visible action ceases, then digest with stronger
acid for an additional 20 to 30 min. to dissolve the last removable
silver.

Effect of Base -metal Impurities in the Bead. The presence
of lead in assay beads increases the tendency of the gold to break
up in parting. In the gold-bullion assay, copper is added prior
to cupellation, if not already present, to ensure that all lead will
be removed during cupellation. It is possible that other base
metals have an effect similar to that of lead, but specific data are
lacking.

Indications of Metals of the Platinum Group. Certain of the
metals of the platinum group may be detected in parting, supple-
menting or verifying the evidence obtained during cupellation.

Nitric acid dissolves part of the platinum and palladium from
precious- metal alloys having a high silver content. Of these,
palladium is the most readily identified because the parting acid
becomes orange colored even with as little as 0.05 mg. of palla-
dium in a bead. Platinum imparts a brown or blackish color to
the parting acid, and the undissolved platinum discolors the
gold, making it steel gray. Platinum is also said 1 to have a
disintegrating effect on the gold, when not more than 5 per cent
of gold is present in the bead.

Iridium is not dissolved by nitric acid and appears in the part-
ings as detached black specks that remain black when annealed
and hence are difficult to distinguish from extraneous particles.

Decantation and Washing. Most assayers remove the spent
acid and washings after parting by decantation into a small
receptacle from which may be recovered any gold particles that
are accidentally decanted. Some assayers remove the acid and
wash water by suction through a glass nozzle attached by a
rubber tube to an aspirator bottle to which hydrochloric acid is
added, which collects the washings and precipitates the silver for
future recovery. This method is quicker and more complete

1 RAWLINS, Trans. Inst. Min. Met., vol. 23, p. 177, 1894.

82 FIRE ASSAYING

than decantation, and the accidental loss of gold is less than
by decantation. If desired, a small bottle could be inserted
in the suction line ahead of the solution bottle, to act as a
gold trap.

Whatever method of washing is used, extreme care must be
employed to avoid loss of gold. When the gold is excessively
powdered, some of the particles may be so small as to be invisible
to the naked eye, and it is virtually impossible to wash such
partings without loss. The recovery of extremely fine gold from
washings may be accomplished by adding a few drops of hydro-
chloric acid to precipitate a small amount of silver chloride,
which carries down the gold in settling. The precipitate is
filtered, scorified with lead, cupeled, and parted.

After removal of the waste acid the parting vessel should be
partly filled with warm distilled water, rinsing the sides of the
vessel with a fine jet at the same time. The wash water is
decanted, and washing repeated until at least three washes have
been given. Incomplete washing is indicated if black stains of
silver appear in the parting or annealing cup after annealing.
Before using the cups again, such stains should be removed by
boiling the cups in a saturated solution of potassium dichromatc
in sulfuric acid.

Drying. After washing, parting cups are dried by placing
them on a warm hot plate. Care must be taken to avoid too
rapid heating during this operation, as the violent evolution of
steam may project the gold from the cup. Fire-clay annealing
cups may be dried more rapidly, as the surplus moisture is
absorbed by the cup and is given off gradually when heated.

Annealing. Unannealed gold from parting is black in color
and cannot be distinguished readily from particles of dust, bone
ash, or other contaminants that may be present. The natural
gold color is developed by annealing the gold at a red heat. If
the annealed gold has a silvery tint the parting has been incom-
plete, and the gold should be inquarted and reparted. As already
noted, platinum causes discoloration of the gold and must be
removed by the methods described in Chap. XII.

Annealing does not change the weight of gold but merely
changes its allotropic form. Annealing would not be necessary
if it were not desirable to verify the completeness of parting,
to examine the gold for evidence of the platinum metals, and

PARTING 83

to avoid weighing extraneous matter as gold. Furthermore,
annealing destroys most of the porosity of the freshly parted
gold and hence avoids the possible effect of absorbed gases on
the weight.

One of the most convenient methods of annealing is to place
an entire tray load of parting cups in a muffle furnace. Some
assay ers object to keeping the main assay furnace under fire after
cupellation is finished and have installed small electric or gas-
fired annealing furnaces. For small-scale work, annealing may
be done individually over a single bunsen burner ; or a battery of
a number of burners closely spaced in a row, with a suitable rack
for holding the cups, may be provided. Eight to twelve burners
are sufficient to keep the annealing manipulations continuous, as
by the time the last parting cup is in place, the first one will be
annealed and may be removed from the flame. A similar anneal-
ing apparatus heated with electricity can be made from nichrome
resistance wire.

Recovery of Silver from Waste Liquor. The waste parting
acid and washings should be recovered in a bottle to which
hydrochloric acid or common salt has been added, to precipitate
the silver as silver chloride from which silver can ultimately be
recovered.

When it is desired to recover the silver from accumulated
precipitates, let them settle until the supernatant liquid is clear,
decant and wash two or three times by decantation with hot
water, and then transfer the precipitate to a filter paper and wash
until free from acid. The precipitate may then be transferred
to a graphite crucible and heated to decompose the chloride,
finally melting down the silver and pouring into a suitable mold.
A light cover with a borax-soda flux is desirable to aid in the
removal of impurities.

Summarized Procedure for Assay Bead Parting. The follow-
ing routinized procedure for parting in porcelain crucibles is
satisfactory for most assay work, except for special products,
such as precious-metal bullion and other rich materials producing
large beads. For such cases, see Chap. IX.

1. Fill the parting cup with warm parting acid of the stock
strength in use (usually 1:4 to 1:8 acid). Cold acid of greater
strength may be used if the assayer's routine permits the addi-
tional parting time required.

84 FIRE ASSAYING

2. Place on a hot plate. If any beads fail to react with the
parting acid after the boiling point has been reached, add 1:1
acid drop by drop until action starts. Beads that do not start
readily in hot 1 : 2 or stronger acid need inquartation and should
be removed for that purpose.

3. Keep just below boiling for 15 to 20 min., by which time all
visible action should have ceased, and the gold will appear as a
spongy mass, considerably shrunken from the original size of the
bead. If an exceptionally large (500 mg. or more) gold bead is
now disclosed, remove the first acid and repeat the digestion
with 1 :2 or stronger acid.

4. Remove from the hot plate, remove the acid by decantation
or aspiration, and wash three times with hot distilled water,
saving the washings for ultimate recovery of silver.

5. Dry carefully on a warm hot plate.

6. Anneal at a red heat.

CHAPTER VI

THEORETICAL DISCUSSION OF ASSAY FUSIONS
AND RELATED SMELTING PROCESSES

An assay er should have a general knowledge of pyrometallurgy
so that he can understand the processes that he is using and be
able to adapt them to the assay of unusual substances. The
materials that he may be required to assay range from any of the
rocks and minerals occurring in nature to all kinds of metallur-
gical products such as metals, artificial sulfides or arsenides, slag,
flue dust, cyanide solution, and precipitates from the cyanide
process. In fact, nearly every known element may be encoun-
tered either by itself or in chemical combination with others.

The elements may be divided into two main groups: the metals
and the nonmetals. In some instances the distinction is not
sharp but in general the oxides of members of each group have
distinctive characteristics both in water solutions and in non-
aqueous melts. In water solutions the metal oxides unite with
water to form hydroxides. These furnish an excess of hydroxyl
ions (OH) in the solution and make it basic. The oxides of the
nonmetallic elements unite with water to form acids and give
an excess of hydrogen ions in solution. There is also a strong
tendency for acids and bases to combine, forming salts. In non-
aqueous melts the metal oxides combine with the nonmetal
oxides and, because of the similarity to the action in aqueous
solutions, the metal oxides with a few exceptions are considered
basic, and the nonmetal oxides are considered acidic. Assay
and smelter slags are formed by the union of the acid nonmetallic
oxide silica with bases that, in smelting, are mainly the metal
oxides CaO and FeO. In assaying, Na2O and PbO are also
important bases ; and borax glass is used as an acid to supplement
silica.

In fire assaying as well as in smelting processes, increased
temperature is used primarily to melt the charge. Melting the
material has the advantage of releasing for possible recovery the

85 *

86 FIRE ASSAYING

precious metals that occasionally are so intimately disseminated
in the rock that they are only partially exposed by any practical
degree of crushing and grinding. It is this complete exposure of
valuable particles in a rock to the recovery process that gives
fire assaying its outstanding advantage over wet chemical deter-
minative methods for gold and silver. If refractories were
available that would stand the temperature required to melt
directly the materials to be assayed, and if a sufficiently high
temperature could be economically produced, assay fusions might
be carried out at high temperatures. Actually, practical con-
siderations of refractories and fuel have limited the usual assay
temperature to a maximum of about 1200C. Substances having
higher melting points are brought to a molten condition by the
formation of low melting compounds and mixtures.

Phases. In the melts used in assaying, smelting, and fire
refining there is the possibility of producing a number of separate
liquids that are not miscible with each other and, therefore,
segregate into separate layers called "phases." Any or all of the
phases metal, speiss, matte, slag, and molten alkaline salts
may be formed. The metal phase has the highest density and
forms the bottom layer; the other phases separate above the
metal phase according to their densities, which are usually in the
order given above. The rules governing the formation and com-
position of these phases furnish the guiding principles of assaying
and many pyrometallurgical operations.

The metal phase in fire assaying is a lead button that collects
the gold and silver. The lead button is the heaviest phase and
forms the bottom layer of an assay fusion. Its size and purity
are the chief concern of the assayer.

The speiss and matte phases are avoided in fire assaying by
proper conduct of the assay. Speiss consists of the arsenides
and antimonides of iron, cobalt, nickel, or copper; matte is a
mixture of fused sulfides, usually of iron and copper. Speiss is
heavier than matte but lighter than lead; hence it forms a layer
between the lead and the matte, if all three arc present.

The slag phase, consisting of metal oxides and silica or borax
glass, lies above the matte layer or, if matte and speiss are absent
as should be the case in an assay fusion the slag rests directly
upon the lead. Metallurgical slags are classified according to
the silicate degree, which is defined as the ratio of oxygen in the

THEORETICAL DISCUSSION OF ASSAY FUSIONS

acids to oxygen in the bases. This classification should not be
confused with that of silicate minerals formerly used by mineral-
ogists. They considered silicate minerals to be salts of various
hypothetical silicic acids and used such terms as "orthosilicate,"
"metasilicate," " sesquisilicate" and "bisilicate." The metal-
lurgical classification, which is followed in this text, is given in
Table VII. The silicate degree is expressed by name or by the
oxygen ratio. For example, a sesquisilicate is said to have a
silicate degree of 1.5.

TABLE VII. METALLURGICAL CLASSIFICATION OF SLAGS

Formula with

Silicate degree

Oxygen ratio
acid to base

SiO 2 and MO
base*

Na 2 O.2B,O 3 and
MO base*

Subsilicate

0.5

4MO.SiO 2

llMO.Na 2 O.2B 2 O 3

Monosilicate

1.0

2MO.SiO 2

5MO.Na 2 O.2B 2 O 8

Sesquisilicate

1.5

4MO.3SiO 2

3MO.Na 2 O.2B 2 O 3

Bisilicate

2.0

MO.SiO 2

2MO.Na 2 O.2B 2 O 8

Trisilicate

3.0

2MO.3SiO,

MO.Na 2 O.2B 2 O 3

* In this book, slag formulas are written with the constituents separated. Writing the
true formula (for example FeSiOa instead of FeO.SiOs) makes the silicate degree classifi-
cation difficult to use.

Alkaline salts, such as the chloride, sulfate or cyanide of the
alkali metals, are occasionally either formed or used in assaying.
These salts do not mix with molten silicate slags but appear as an
extremely fluid top layer in a fusion. The sodium sulfate pro-
duced in the fusion of sulfide ores is the alkaline salt most fre-
quently encountered in assaying. Sodium chloride has found
some use as a cover for crucible fusions but is seldom used today.
Cyanide salts are strong reducing agents and are used in the tin
assay.

METAL PHASE

The entire object of smelting, refining, or an assay fusion is to
produce a suitable metal phase that separates from the rest of the
molten charge because of its high density and forms the bottom
layer.

Whether a metal appears in the metal layer or in the slag
depends upon whether or not it is combined with oxygen. The
metal oxides are quite soluble in the slag and, in general, are

FIRE ASSAYING

insoluble in the metal. Consequently, when two different metals
are present, if one can be oxidized to metal oxide it will go into
the slag layer, while if at the same time the other metal can be
prevented from oxidizing it will appear in the metal layer, pro-
vided no speiss or matte is formed. This principle forms the
basis of the separations made in fire assaying, smelting, and fire
refining. Fortunately, there is a difference in the ease of oxida-
tion of the various metals, so that separations can be made.

The relative ease of oxidation of the metals can be measured by
their electrode potentials when immersed in solutions containing
1 g. equivalent weight of their ions per liter of solution. By these
measurements the elements have been listed in an order called
the " electromotive series of elements" (Table VIII).

TABLE VIII. PARTIAL ELECTROMOTIVE SERIES OF THE ELEMENTS

Element

Oxidation
product

Electrode
potential*

K+

2.92

Na+

2.71

Mg++

2.34

Mn+ +

1.05

Zn++

0.76

Fe ++

0.44

Cd ++

0.40

Co ++

0.28

Ni ++

0.25

Sn + +

0.14

Pb ++

0.13

H s

H+

0.00

Sb+ ++

-0.10

Bi+++

-0.23

As +++

-0.30

Cu+

-0.47

Te++++

-0.56

Ag+

-0.80

Hg + +

-0.85

Pt ++++

-0.86

Au+

- 1 . 50

* Values largely from W. M. Latimer, "Oxidation Potentials," Prentice-Hall, Inc., New
York, 1938.

The electromotive series furnishes the key to the possible
separations by oxidation or reduction. The metals low in the

THEORETICAL DISCUSSION OF ASSAY FUSIONS

series are called the "noble metals" and are difficult to oxidize.
Those above silver combine with oxygen when heated in the air,
and the action becomes more energetic as the list is ascended.
Considering the reduction of the metal oxides, it is found that
metal oxides up to and including silver lose their oxygen when
heated in the air; while if heated in a current of hydrogen, the
oxygen is easily removed from all the oxides up to and including
those of iron, leaving in each case the metal.

Actually the oxidation of a metal or the reduction of a metal
oxide is controlled by the pressure of oxygen, produced by
the tendency of the metal oxide to dissociate, compared with the
external oxygen pressure. When the dissociation pressure of the
metal oxide is lower than the external oxygen pressure, the metal
tends to oxidize; but if the dissociation pressure is greater than
the external oxygen pressure, the oxide is reduced to metal. The
dissociation pressure of metal oxides increases with the tempera-
ture, but only the noble metals can be reduced by heat alone.
For example, at room temperature the oxygen pressure due to
dissociation of silver oxide is less than the pressure of the oxygen
in a normal atmosphere, and silver oxide can exist. At a temper-
ature of only 227C. (Table IX) the oxygen pressure is much
TABLE IX. OXYGEN PRESSURE DUE TO DISSOCIATION OF METAL OXIDES

Temperature

Oxygen pressure, atmospheres, at equilibrium*

Degrees
absolute

Degrees
centigrade

Ag 2

Cu 2

PbO

NiO

FoO

ZnO

300
500
1000
2000

27
227
727
1727

8.4 X 10-5
24 9

0.56X10- 3 o
1 5 X 10-"
4 4 X 10" 1

3 1 X 10-
1 1 X 10-15
3 7 X 10-*

1 8 X 10-
8 4 X 10-20
3 3 X 10~

2.0 X 10~2*
1 6 X 10-

1.3X10-"
7.1 XIO" 31
9.5 X 10~ 12

* Values calculated by W. Stahl, M etdlurgie, vol. 4, 682, 1907. Taken from table in R. S. Dean,
"Theoretical Metallurgy," p. 159, John Wiley & Sons, Inc., New York, 1924, with the permission of the
author and publisher.

greater than atmospheric pressure, and silver oxide spontaneously
reduces to metal. To reduce base metals it is necessary to supply
a reducing agent something such as carbon, carbon monoxide,
hydrogen, or a metal higher in the series. These reducing agents
combine with oxygen and lower the external oxygen pressure
below the dissociation pressure of the metal oxide to be reduced.

90 FIRE ASSAYING

When the oxidation of a metal or the prevention of its reduc-
tion ivS desired, sufficient oxygen pressure must be supplied to
exceed the dissociation pressure of the metal oxide. The oxygen
can be supplied from the air as in cupellation and roasting, or
from niter as in the crucible assay of sulfide ores, or from a metal
oxide lower in the electromotive series as in the use of litharge as
an oxidizing agent in the softening of crude lead.

In an assay fusion the metal layer is the sole useful product.
It is possible to produce in one step a metal layer consisting of the
precious metals alone. This is not practical in either smelting
or assaying, because the unavoidable mechanical loss of small
amounts of the metal layer represents a value loss that is greater
than can be tolerated. To reduce the value loss in both smelting
and assaying a relatively large amount of a less valuable material
that will dissolve or alloy with the precious metals is provided,
so that the same loss in quantity of material from the layer
represents only a small value. Lead alloys readily with the
precious metals and is the less valuable material used to collect
gold and silver in both assaying and lead smelting. Lead is a
particularly suitable collector for the precious metals in assaying,
because it can readily be separated from the precious metals by
the simple process of cupellation. The lead alloy from the metal
layer of an assay fusion is known as the "lead button." It must
be produced under conditions that give:

1. Limited amounts of base-metal impurities.

2. Good collection of the precious metals.

3. A button close to the desired size.

IMPURITIES OF THE LEAD BUTTON

Base-metal impurities in the lead button vary in their effect.
Most base metals, with the exception of bismuth, will cause
trouble in cupellation if they are present in the lead button in
sufficient amount (Chap. IV). Fortunately, only a few metals
and metalloids are low enough in the electromotive series that
they can be reduced in an assay fusion and have the opportunity
of alloying with the lead. A glance at the electromotive series
shows that, besides the precious metals, there are only antimony,
bismuth, arsenic, copper, tellurium, and mercury below lead.
It is to be expected that when any of these elements are present
in an assay fusion, they will be at least partly reduced and may

THEORETICAL DISCUSSION OF ASSAY FUSIONS 91

contaminate the lead button. Mercury is reduced the easiest
of all these impurities, but it does not contaminate the lead
button because it boils at 356. 9C., and, consequently, in the
assay fusion it vaporizes and leaves the assay charge as soon as it
is reduced.

Bismuth. Bismuth is difficult to separate from lead, as it
behaves so nearly like lead. In the assaying or lead smelting of
ores containing bismuth, most of the bismuth is always reduced
and alloys with the metallic lead. The separation of bismuth
from lead is a difficult refining operation 1 even at lead smelters,
and the processes used cannot be adapted to fire assaying.
Bismuth can be cupeled similarly to lead, and a bismuth-lead
alloy cupels without difficulty, so that the presence of bismuth in
an ore does not interfere with the actual mechanism of the assay
process. The dor6 bead produced from high bismuth ores, how-
ever, contains some bismuth, because it is more difficult to oxidize
than lead. This residual bismuth adds to the silver assay but
does not affect the gold result. At the present time, there is no
known variation of the fire assay process that will eliminate
bismuth from the dore bead. Accurate silver results can be had
only by correcting for its bismuth content.

Antimony, Arsenic, Copper, and Tellurium. The other trouble-
some impurities, that is, antimony, arsenic, copper, and tellu-
rium, arc easier to oxidize than are bismuth and mercury. Their
concentration in the lead button can be lowered by the use of an
excess of litharge in the assay slag. In any event a certain frac-
tion of the impurity present in the charge will be reduced and
will contaminate the lead button. Hence, in assaying impure
ores, large samples cause much more difficulty than do small
samples. Occasionally a sample smaller than would ordinarily
be used must be taken, in order to reduce the concentration of
some impurity in the lead button.

Use of Excess Litharge. The effect of excess litharge in help-
ing to keep the impurities out of the lead button depends upon
its ability to keep them oxidized while some lead remains reduced ;
the action on copper is typical. Lead has a greater tendency to
combine with oxygen than has copper, and it is only by the mass

1 BETTERTON, JESSE O., and LEBEDEFF, YURII E., Debismuthizing Lead
with Alkaline Earth Metals, Including Magnesium, and with Antimony.
Trans. A.I.M.E., vol. 121, p. 205, 1936.

92 FIRE ASSAYING

action of excess litharge that a lead button can be reduced with-
out reducing nearly all the copper in the charge. In an assay
fusion of a copper-bearing ore the equilibrium constant of the
following reaction governs the amount of copper reduced.

PbO + 2Cu ^ Cu 2 + Pb (1)

According to the laws of mass action, we may write

A[Pb]A[Cu 2 0] = R (2)

where K is the equilibrium constant and A represents the activity.
The activity of the metallic lead and copper in the reaction is not
known, but with a constant amount of lead produced and with a
definite amount of copper allowed in the lead their activities
must be constant.

^[ pb ] _ v

Substituting (3) in (2),

A[Cu 2 0]

Then a new constant, K 2 . can be introduced for -^-;

A [PbO] " 2

The activities of the cuprous oxide and the lead oxide in the
charge are proportional to the concentrations present, and
therefore

Concentration Cu _ ^ ,^

Concentrat on PbO 2 U

Equation (6) shows that the amount of copper retained in
oxidized form, and therefore remaining in the slag, is in propor-
tion to the concentration of free PbO in the slag. The activity
of PbO, when combined with silica, is very low and is therefore
not effective in oxidizing copper. Accordingly, with a constant
amount of litharge in the charge an increase in silica decreases

THEORETICAL DISCUSSION OF ASSAY FUSIONS 93

the free PbO and consequently more copper enters the lead
bullion.

The problem of softening lead bullion in the refining of crude
lead by the removal of the hardening constituents arsenic,
antimony, copper, and tin is similar to that of keeping these
impurities out of the lead button in assaying. Of these metals,
tin does not contaminate the assay button nearly to the extent
that it enters the lead bullion in a blast furnace, because of the
difference in amount of reduction in the two processes. In all
but the uncontrolled reduction assay fusions, reduction and
oxidation are so adjusted that at the end of the operation both
lead and lead oxide exist together. In the lead blast furnace and
in uncontrolled reduction methods of assaying, much more reduc-
tion is carried out in an effort to reduce all lead oxide to lead.
For this reason these processes carry an excess of reducer,
which reduces a considerable amount of the tin in the charge,
even though tin is above lead in the electromotive series. In
the controlled reduction assay charge, a deficient amount of
reducer is used so that tin and other metals above lead in the
series are not reduced and, consequently, cannot contaminate the
lead button.

The other metals arsenic, antimony, and copper contami-
nate assay lead buttons as well as the smelter's lead bullion.
Two processes are available for the softening of lead bullion: (1)
softening in a reverberatory furnace with litharge and (2) soften-
ing in a kettle with soda salts (Harris process).

Softening in a reverberatory furnace with litharge depends upon
the oxidation effect of excess litharge, and the principle is exactly
the same as in the case where excess litharge is used in a crucible
assay charge. The refiner can carry the separation further than
can the assayer, as he continually removes the oxidized impurities
by skimming, and so disturbs the equilibrium and allows the
type reaction (1) to proceed in the direction of further oxidation
of impurities. The assayer does not have available this method
of changing equilibrium and can force the reaction in the desired
direction only by adding more litharge.

Use of Soda Salts. Softening in a kettle with soda salt?
depends upon the ease of oxidizing arsenic and antimony to form
arsenates and antimonates in a molten alkaline salt environment.
This process does not separate copper, and, consequently, when

94 FIRE ASSAYING

the Harris process is used in lead refining, copper must be removed
by an additional process.

The authors have tried crucible assays with very alkaline slags
in the hope of obtaining some of the refining effect of the Harris
process. On the whole the results were discouraging; normal
crucible assay charges caused less trouble from boiling over and
gave better slags. It should be remembered, however, that, in
an ordinary crucible assay of sulfide ores, arsenic oxidizes to
As20 5 and requires extra base to form arsenate. An additional
amount of sodium carbonate is usually added for this purpose.

Separation of Impurities by Scorification. Scorification con-
sists of heating the material to be scorified, whether it is an
impure lead button or an ore sample, with granulated lead in a
shallow fire-clay dish which is exposed to an oxidizing atmos-
phere. In the process, oxidation proceeds rapidly and a large
proportion of the lead present is oxidized to litharge. The
Scorification process has no greater ability to eliminate impurities,
except volatile ones, than has the crucible assay. The amount
of impurities remaining oxidized at the end of the fusion depends
mainly upon the proportion of free PbO existing in the slag at
the end of the operation. Whether the final state is reached by
partial oxidation of lead (scorification) or by partial reduction
of litharge (crucible assay) makes little difference. In the
scorification operation, silica is taken up from the scorifying dish
so that the slags do not contain so large an amount of free PbO
as one would expect. For this reason it is usually better to treat
impure lead buttons by melting them in a crucible with litharge
and then swirling for a few minutes than it is to scorify them.

COLLECTION OF THE PRECIOUS METALS

Complete collection of the precious metals by lead during an
assay fusion is desired. Actually, recovery is never quite com-
plete, and the ratio of loss to recovery is dependent upon one or
more of the following factors: (1) mechanical release of the
precious-metal particles by fusion of the surrounding material;
(2) contact of the released precious-metal particles with molten
lead; and (3) solubility of the precious metals in the various
phases present at the end of the fusion.

Release by Fusion. The material surrounding the precious-
metal particles must be fused to a mobile liquid, in order to

THEORETICAL DISCUSSION OF ASSAY FUSIONS 95

release the precious-metal particles for collection. Loss due to
incomplete fusion should not be tolerated, since lumpy and
viscous melts can be avoided by the use of properly proportioned
fluxes and a suitable temperature.

Contact with Lead. In order to be collected by lead, each
particle of precious metal must actually contact molten lead.
The action is quite similar to the amalgamation of gold by liquid
mercury at ordinary temperatures. Although granulated lead
is used in the scorification assay and small lead droplets are
reduced from litharge well distributed through the crucible assay
charge, it seems impossible to have sufficient contact between
the precious-metal particles and the lead particles by their
position in the assay charge. More likely, only a few precious-
metal particles find themselves actually in contact with lead
particles immediately upon release. The larger of the precious-
metal particles readily settle through the molten slag to the
metal phase below; the remainder probably encounter particles
of lead, or even the lead button itself, during their circulation
in the slag. Moreover, after the reactions have become com-
plete, circulation, probably due to convection currents, can be
observed in the molten slag. The ease with which good col-
lection of the precious metals is obtained in assay fusions is
remarkable.

Even in a fusion without any lead to serve as a collector there
is almost complete elimination of gold and silver from the slag.
After pouring such a melt the greater part of the precious metals
will be found in the crucible. They are distributed in many
small specks adhering to the crucible wall, so that the simple
addition of lead to the molten fusion does not suffice to collect
all of them. A study of collection subsequent to fusions, made
in the absence of metallic lead, 1 has shown that the reduction of
litharge to lead in the charge, and the stirring action of the
charge produced by the addition of litharge and iron filings, are
necessary for good collection. When lead is present at the time
that the fusion is made, good collection is easier to obtain. The
assayer may be certain that he is not seriously in error because of
lack of contact between the precious metals and molten lead when

1 SHEPARD, O. C., and DIETRICH, W. F., Oxidation-collection Method of
Assaying Sulphide Ores for Gold and Silver, in A.I.M.E. Tech. Pub. 997;
Metals Technology, vol. 6, No. 2 (1939).

96 FIRE ASSAYING

a fluid slag is produced, and when more than 15 g. of lead are
reduced. When less than 10-g. buttons are reduced in crucible
fusions, erratically low results are frequently obtained; obviously
the loss is due to lack of contact with lead for collection. In
scorification, good collection of gold and silver is obtained with
smaller buttons than in the crucible assay, partly because of
the smaller ore charge and partly because collection is nearly
complete while the button is still relatively large, and the ensuing
shrinkage of the button during scorification virtually does some
of the work of cupellation. Hence, buttons weighing 10 to 15 g.
are usually satisfactory, unless they are to be cupeled adjacent
to larger buttons.

In the crucible assay, increasing the amount of lead reduced
above 20 g. does not continue to reduce greatly the loss encoun-
tered in fusions. The fact that one fusion in which a 40-g. button
is reduced does not give so low a slag loss as a 20-g. button fusion,
followed by a reassay of the slag to produce an additional 20-g.
button, points to solubility of precious metals in the slag as the
major source of loss.

Whether the usual small loss in an assay fusion is due to un col-
lected values or to slag solubility is difficult to determine.
Ravitz and Fisher 1 have studied the distribution of gold and
silver between lead, speiss, matte, and slag, by fusing these
products with gold chloride and silver nitrate in a crucible.
With equal weights of lead, speiss, matte, and slag, definite ratios
were found between the gold and silver dissolved in the lead
and that dissolved in the speiss and matte. In general the
amount of silver in the speiss was close to 1.2 times that in the
lead, while the silver in the matte was about 0.54 times that in
the lead. Gold showed greater solubility in the speiss and less
solubility in the matte; the amount of gold in the speiss was
close to 1.6 times that in the lead, and the amount of gold in the
matte about 0.07 times that in the lead. The distribution of
values between the slag arid lead was erratic, and both the gold
and silver content of the slag varied from 0.001 to 0.01 times
that of the lead. This is roughly the order of magnitude of the
slag loss in assaying; but in assay fusions with different com-
positions of phases and with different proportions of slag to lead,

1 RAVITZ, S. F., and FISHER, K. E., Equilibrium in Lead Smelting, Trans.
A.I.M.E., vol. 121, p. 118, 1936.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 97

considerable difference in the distribution of gold and silver is
to be expected.

These results do show, however, the comparatively high
solubility of gold and silver in lead, matte, and speiss; and the
comparatively low solubility in slag. The control of these
phases to secure high recovery of the precious metals in the lead
button is discussed at greater length in the paragraphs on each
phase.

SIZE OF THE LEAD BUTTON

The assayer, by controlling the amount of metallic lead exist-
ing at the end of the assay fusion, controls the size of the lead
button. A size is desired that is a compromise between the
inconvenience and added cost of cupeling large buttons and the
increased loss of values occasioned by small buttons. Although
the loss of values increases markedly when lead buttons weighing
much less than 15 g. are produced in crucible assays, the recovery
does not increase greatly with increase in lead-button size
above 25 grams. In most cases a lead button weighing about
25 g. is desired. This is small enough to be economical and
large enough so that further increase in size, within practical
limits, gives scarcely appreciable added recovery. If extremely
high recovery is necessary it is better to re-treat the slag as in a
corrected assay rather than to increase the button size. An
important reason for desiring uniformity in the size of lead buttons
is the increased convenience and accuracy of cupeling buttons
of uniform size.

During an assay fusion the lead button is produced by either
(1) starting with lead oxide in the charge and reducing just a
sufficient amount of lead for the lead button, as in the crucible
assay ; or (2) starting with an excess of granulated lead in the charge
and oxidizing all but the amount required for the lead button,
as in the scorification assay. The production of a lead button
and the control of its size in either case depend upon the reac-
tions taking place during fusion.

Crucible Assay. Some crucible assay methods use excess
reduction and limit the amount of lead compounds in the charge,
in order to control the size of the lead button. If harmful impuri-
ties are present, their reduction to metal in such uncontrolled
reduction methods may cause button contamination or other

98 FIRE ASSAYING

difficulties. In controlled reduction assay methods, more litharge
is added to the charge than is required for the lead button. Lead
oxide is always available for the slag. The size of the lead buttons
in the controlled reduction methods depends upon production in
the assay fusion of just sufficient reduction to reduce the required
amount of lead. Reduction is affected by (1) minerals, some of
which have an oxidizing or reducing effect, and by (2) oxidizing
or reducing agents, one of which is usually required in the charge
to regulate the amount of reduction. Ores or materials to be
assayed may contain no important amounts of oxidizing or
reducing minerals to cause either oxidation or reduction of lead.
A knowledge of the amount of lead reduced by the reducing
minerals, and the amount of lead that is prevented from being
reduced by oxidizing minerals present in the ore, is essential in
order to determine the amount of oxidizing or reducing agent
that must be added to the charge to produce a lead button of
suitable size. For convenience in calculations, all figures for
oxidation or reduction are referred to the corresponding weight of
lead, even though the oxidizing or reducing agent may not react
directly with lead or litharge.

The reducing power (R.P.) of any substance is defined as the
weight of lead, in grams, reduced by 1 g. of that substance.
Power as used here refers only to the quantity of material. A
high reducing power does not mean ability to reduce metals
high in the electromotive series. -The reducing effect (R.E.)_ol
any substance is the weight of lead reduced by the total amount
of that substance in the charge. The oxidizing power (O.P.)
of any substance is defined as the weight of lead, in grams, that
is prevented from being reduced by 1 g. of that substance.
The oxidizing effect (O.E.) of any substance is the amount of
lead prevented from being reduced by the total amount of that
substance in the charge.

Reducing Agents. Reducing agents are substances added
to the charge to take up oxygen and reduce litharge to lead.
They are used only when the ore has insufficient reducing effect
to produce the size of lead button desired. Commonly used
reducing agents are: flour, charcoal, argol, sulfur, and metallic
iron; many other substances could be used,
f Ordinary flour contains about 15 per cent moisture, 70 per cent
starch, and smaller amounts of proteins, sugars, and other

THEORETICAL DISCUSSION OF ASSAY FUSIONS 99

organic compounds. Only a small error in reducing power is
made by lumping all the organic compounds in flour together,
and, considering flour to be 85 per cent starch, as represented
by the formula C 6 Hi O 5;

12PbO + C 6 HioO 5 = 12Pb + 6CO 2
162 2,486

In the above equation, 12 atoms of lead or 2,486 parts by
weight are reduced by 1 molecule of starch or 162 parts. The
theoretical reducing power of pure starch is 2,486/162 or 15.3.
If we consider flour to be 85 per cent starch, its theoretical
reducing power would be 13.0. The actual reducing power
varies somewhat with different types of flour. The reducing
power of a typical wheat flour was found to vary from 12.2 to
11.0, depending upon the acidity of the slag. Reducing agents
from which incompletely oxidized gaseous products may be
formed give lower reducing power in charges producing acid
slags than in basic charges. This occurs with all organic reduc-
ers, because carbon can escape from the charge as CO instead of
C02, and occurs with sulfur, which may be lost as elemental
sulfur or as S() 2 instead of forming sulfates. The actual reducing
power in basic charges is frequently close to the theoretical value.
In basic charges there is plenty of available litharge, so that
even a gaseous reducer cannot escape from the charge without
nearly complete reaction. In acid charges the litharge particles
are nearly all in contact with acids, and incipient reaction to form
borates and silicates reduces the amount of free PbO available
for lead formation before the reducing agent has completed its
work. When there is a deficiency of free PbO present the
efficiency of reducers that can escape from the charge is lowered.
On the other hand a reducer such as iron that cannot escape
from the charge will produce its full quota of lead, even from
acid charges and from lead silicate slags. Flour will not reduce
lead from an acid slag containing lead silicate.

The reducing power of charcoal depends mainly upon its
carbon content and varies with the purity.

2PbO + C = 2Pb + C0 2
12 414

100 FIRE ASSAYING

Theoretically, 1 g. of carbon will reduce 34.5 g. of lead. Most
charcoal in use contains a considerable amount of ash, so that
lower reducing powers are obtained in practice. If charcoal is
added separately to each crucible, it is desirable to dilute it with
an equal weight of silica, so that greater variation may be
allowed in measuring out the amount required.
/Argol consists of the crude potassium acid tartrate that is
deposited in the process of making wine. The reducing reaction
with pure potassium acid tartrate is given below :

lOPbO + 2KH(C 4 H 4 O 6 ) = lOPb + 8CO 2 + 5H a O + K 2 O
376 2,070

According to the reaction the reducing power of potassium acid
tartrate is 2,070/376 or 5.5. Impurities usually present in argol
increase the reducing power considerably above that of the pure
tartrate. The argol reducing reaction produces K 2 O, which
acts as a basic flux; but the amount produced is small and can
have little effect.

The reducing power of sulfur not only varies with the acidity
of the charge, but also with the sodium carbonate content of the
charge. In a very basic charge with sufficient Pb() but with no
sodium carbonate present, sulfur is oxidized to SO 2 .

2PbO + S = 2Pb + S0 2
32 414

In this reaction, sulfur has a reducing power of 12.9. With the
addition of soda to the same charge the reducing power of sulfur
increases to a maximum of 19.4 corresponding to the nearly
complete oxidation of sulfur to SO 3 . Sodium carbonate goes
through a series of reactions in an assay fusion with sulfur and
litharge, to produce finally sodium sulfate.

2PbO + S = 2Pb + SO 2 (1)

S0 2 + Na a C0 8 = Na 2 SOs + CO 2 (2)

Sodium sulfite is unstable at that temperature and forms
sodium sulfide and sodium sulfate.

4Na 2 S0 3 = 3Na 2 S0 4 + Na 2 S (3)

Sodium sulfate is an end product, but sodium sulfide reduces
more lead.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 101
3PbO + Na 2 S = 3Pb + Na 2 SO 3 (4)

The sodium sulfite formed in reaction (4) follows the last two
steps over and over until all the sulfur is converted to sulfate.
The series of reactions may be represented by a single equation :

3PbO + S + Na 2 C0 8 = 3Pb + Na 2 SO 4 + CO 2
32 621

from which the reducing power of sulfur is 19.4.

It is desirable to dilute strong reducers, such as sulfur, by
mixing them with a material having no oxidizing or reducing
effect to permit easier measuring of the required amount. If
sulfur is diluted with 3J parts of sodium carbonate the soda
required for the formation of sodium sulfate is automatically
provided. One gram-molecule of Na 2 C() 3 (106 g.) is required
for each atom of sulfur (32 g.) 106 -*- 32 = 3.34.
^Iron, as a reducing agent, differs from carbon, hydrocarbon,
or sulfur in that (1) it cannot escape from the charge as a gas
at any stage of the process, (2) it cannot migrate through an
unfused charge to meet and reduce lead compounds, and (3)
it has sufficient density to sink through a molten slag where it
contacts lead compounds and reduces them more effectively
than do reagents of low density. These characteristics make iron
a suitable reducing agent for the oxidation-collection method
and for the soda-iron method. Carbon, hydrocarbon, or sulfur
fails to give satisfactory results in either of the above processes.
In the oxidation-collection method, flour fails to give good col-
lection of the precious metals, even though a suitable-sized lead
button is reduced. The poor result with flour may be due to a
too rapid reaction, or to reduction from the surface instead of
down in the liquid melt. Tests comparing the use of iron with
that of other common reducing agents in the ordinary crucible
assay should be made. Iron will reduce lead at a later stage
in the process, which may or may not be an advantage. It
has the disadvantage of contributing its oxide to the slag.
The reducing power of iron in normal assay charges is about
3.7, as shown by the reaction:

PbO.SiO 2 + Fe = Pb + FeO.SiO 2
56 207

102 FIRE ASSAYING

Reducing Minerals. Reducing minerals are those minerals
which, if present in a crucible assay charge, will reduce litharge to
lead. Some ores contain graphite or sulfur, but the usual reduc-
ing minerals encountered in assaying are the metal sulfides. The
reducing power (R.P.) of metal sulfide minerals may be con-
sidered to be made up of two parts: (1) reduction caused by
oxidation of the metal and (2) reduction caused by oxidation of
the sulfur.

Pyrite has a reducing power equal to the amount of lead
reduced by the 0.466 g. of iron and 0.534 g. of sulfur contained in
1 g. of pyrite. The reducing power can be calculated from either
a knowledge of the reducing power of iron and sulfur or the
combining weights.

Calculations of the Reducing Power of Pyrite

1. Without sodium carbonate and with excess litharge, sulfur
oxidized only to SO 2 .

a. By the reducing power of the constituents.

0.466 g. of Fe X R.P. 3.7 = 1.7 g. of lead reduced
0.534 g. of S X R.P. 12.9 = 6^9 g. of lead reduced
1.000 g. of FeS 2 = 8 . 6 g. of lead reduced

(R.P. of pyrite)

b. By the combining weights.

One gram-molecule of pyrite, 120 g., requires 5 oxygen
atoms to form one FeO arid two S0 2 . Five gram-atoms
of lead are reduced, 207 X 5 = 1,035 g.
The lead reduced by 1 g. of pyrite or the R.P. of pyrite is
found from the proportion:

FeS 2 5Pb

120 : 1,035 : : 1 : R.P. of pyrite
R.P. of pyrite = 8.6

2. With excess sodium carbonate and litharge, sulfur oxidized
to S0 3 .

a. By the R.P. of the constituents.

0.466 g. of Fe X R.P. 3.7 =. 1.7 g. of lead reduced
0.534 g. of S X R.P. 19.4 = 10.4 g. of lead reduced
1.000 g. of FeS 2 = 12.1 g. of lead reduced

(R.P. of pyrite)

THEORETICAL DISCUSSION OF ASSAY FUSIONS 103

b. By the combining weights.

One gram-molecule of pyrite, 120 g., requires 7 oxygen
atoms to form one FeO and two S0 3 . Seven gram-
atoms of lead are reduced, 207 X 7 = 1,449 g.

FeS 2 7Pb

120 : 1,449 : : 1 : R.P. of pyrite
R.P. of pyrite = 12.1

Sometimes chemical equations are written to express the reduc-
tion of litharge to lead by pyrite in an assay fusion.
Without sodium carbonate:

FeS 2 + 5PbO + Si0 2 = 5Pb + FeO.SiO 2 + 2SO 2 |
With sodium carbonate in excess:

FeS 2 + 7PbO + 2Na 2 C0 3 + SiO 2 - 7Pb + FeO.Si0 2

+ 2Na 2 SO 4 + 2CO 2

These reactions should be regarded as the net result of the com-
plicated intermediate reactions that actually take place.

The general study of the reducing power of all sulfide minerals
is simplified by considering separately the reduction caused by
metal and sulfur.

Metal Effect. Many metals form more than one oxide, so
that in order to calculate the metal part of the reducing power of
sulfide minerals, the final stage of oxidation of the metal should
be known. This can be determined by experiment. Results
indicating the state of oxidation of metals and metalloids from the
reducing power of common sulfide minerals are given in Table X.
The measured reducing power given in the table is the highest
value found in very basic charges with excess sodium carbonate
and litharge. When sulfur alone is used in such charges, the full
reducing power of sulfur to SOs is obtained; therefore it is
assumed that the observed maximum reducing power of the sul-
fide minus the theoretical maximum reducing power of the sulfur
in the sulfide gives a measure of the reducing power of the metal
in the sulfide. Basic charges favor the formation of the higher
metal oxide, and yet the calculated reduction by most metals is
less than the theoretical value for the lowest metal oxide. Where
lower results are obtained, the difference is due to the assumption
of complete sulfur oxidation in the calculations. Both arseno-
pyrite and pyrite, however, give higher values than the theo-

104

FIRE ASSAYING

retical for the lowest oxide. This can be caused only by (at
least a part of) the metal going to the higher oxide. This con-
clusion has been verified in the case of arsenic by the fact that
arsenious oxide gives a reducing power of at least 1.7 in basic
charges, which is close to the theoretical value of 2.09 for conver-
sion of all of the trioxide to the pentoxide. Antimonious oxide
has no reducing power, and the experimental value for the metal
reducing power, as given in Table X, also verifies the conclusion
that antimony does not oxidize beyond the trivalent form in assay
charges.

TABLE X. INDICATED OXIDATION STATE OF METALS

Weight of lead reduced by the

oxidation of "M" in 1 g.

of mineral

Metal "M"

Mineral
investigated

Indicated
oxide

Theoretical

Meas-

rl *

M 2 O

M 2 S

M 2 5

Copper

Chalcocite

1 3

2 6

CiioO

Iron ....

Pyrite

1 7

2 6

2 2

FeaOsf

Lead.

Galena

0.9

PbO

Zinc

Sphalerite

2.1

2 1

ZnO

Antimony

Stibnite

1.8

3.0

1.7

Sb 2 3

Arsenic . .

Arsenopyrite

1 9

3.2

2.9

A& 2 O 5

* The values in this column were obtained by deducting the calculated reduction of lead,
due to oxidation of sulfur to SOs, from the maximum measured R.P. of minerals in veiy basic
charges. For arsenic the calculated reduction due to oxidation of iron to FeO was also
deducted from the measured R.P. of arsenopyrite.

t Oxidation of iron to FezOs takes place only in the very basic charges that were used in
this experiment in order to obtain nearly complete oxidation of sulfur to SO 3.

Sulfur Effect. A large part of the reducing power of most
sulfide minerals is due to sulfur, and the reducing power of sulfur
is in general more variable than that of metals. Consequently
the factors influencing the reducing power of sulfur are the
critical variables affecting the reducing power of sulfide minerals.
Within the range of practicable assay slags the most important
of these factors are: (1) the molecular ratio of litharge to soda
in the slag, (2) the silicate degree of the slag, (3) the extent to
which silica is replaced by an equivalent of borax glass, and (4)
the molecular ratio of total base fluxes to metal sulfides.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 105

1. The optimum molecular ratio of litharge to soda is 1.0, which
nearly corresponds to 2 parts of litharge and 1 part of sodium
carbonate by weight. If this ratio is increased without increas-
ing the total flux ratio the reducing power generally decreases
and becomes sensitive to variations in the silicate degree of the
slag. Decrease below 1 in the litharge-soda ratio, if not carried
to an extreme, has little effect on reducing power, but in prac-
tice, too much sodium carbonate is avoided on account of its
bulk and the evolution of C0 2 .

2. If the slags are more acid than the monosilicate the reducing
power of sulfides decreases with increased silicate degree and
becomes sensitive to a number of factors that are difficult to con-
trol, thus giving erratic results.

3. All or part of the silica may be replaced by the silicate
degree equivalent of borax glass in slags less acid than the mono-
silicate without significant effect on the reducing power of
sulfides, but, in more acid slags, borax glass is a more potent
depressant of the reducing power than silica.

4. The full theoretical reducing power of sulfides is not ap-
proached unless the ratio of base fluxes to sulfide is somewhat
higher than is necessary or convenient for satisfactory slag
formation and button purity. However, if the volume of fluxes
is sufficient for satisfactory fluxing, and if the three foregoing
conditions are satisfied, the reducing power will be sufficiently
constant, although somewhat less than the theoretical value.
This consideration usually implies the presence of soda for slag
formation in amount at least equal to the weight of ore taken for
assay.

Since any variations in the state of oxidation of the metal por-
tion of a sulfide mineral has been shown to have a minor effect on
reducing power, the chief effect of any variable tending to
decrease reducing power is incomplete oxidation of the sulfur.
Although the lack of complete oxidation is due mainly to escape
of incompletely oxidized sulfur from the charge, it is usually
necessary to proportion the constituents of an assay charge to
ensure maximum oxidation of sulfur and its conversion to sodium
sulfate, which forms an alkaline salt layer above the slag. If
this is not done the buttons may be brittle from entrapped sul-
fide or from the presence of reduced metals other than lead; or a
separate matte or speiss layer may be formed. Hence, condi-

106

FIRE ASSAYING

tions that favor maximum reducing power also favor maximum
button purity.

To illustrate the order of magnitude of the important factors
affecting the reducing power of sulfides, the reducing power of
stibnite (Sb 2 S 8 ) under various conditions is shown graphically
on Fig. 9. 1 It should be noted that for all the slag types illus-
trated the reducing power does not vary greatly in the sub- to
monosilicate range; and where the litharge-soda ratio is 1.0 or
less, the reducing power of the subsilicates approaches the

i.J

7.0
6.5
6.0
fe5.5
1.5.0

"S4.0

3.5
3.0
2.5

2 'o

sass;

=-=

-=5

iTS-

3r^

^C;

Xrt

-*x

\
v

**^^

^*N^

^"**

*^,,

**v

\ x

Curve

Slag Ratios
Nq 2 0:PbO (Na 2 0+PbO):Sb 2 3 B 2 3
1 12 C
1 6 C
1 12 '/
3 12 C
V 3 12
1

>
:SiO :
)

\ S

[

I1- J

4 0.5 0.6 0.7 0.8 0.9 1.0 1.

1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.C

Silicate degree of slag
FIG. 9. The reducing power of stibnite.

theoretical value of 7.3 for oxidation of the antimony to Sb2Oa,
and the sulfur to SO 3 . With slags more acid than a monosilicate,
however, the reducing power decreases notably with all slag
types; and this effect is intensified by lower flux ratios, higher
litharge-soda ratios, and by the substitution of borax glass for
silica. In the series of tests illustrated by Fig. 9, all buttons
produced under conditions that gave a reducing power much
below 6.0 were brittle. Embrittlement or hardening of the but-
tons occurred in similar series of tests with galena, pyrite,
arsenopyrite, and sphalerite whenever the reducing power was

1 Unpublished investigation by the authors.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 107

considerably below the theoretical value, which emphasized the
necessity for nearly complete oxidation of sulfur in crucible
fusions. With the copper minerals chalcocite and chalcopyrite
a considerable excess of litharge is needed to keep copper from
the button; therefore the litharge-soda ratio is much greater than
is recommended for other sulfides, but this is compensated for by
the considerably greater total flux ratio, and these sulfides, in
very basic charges, give nearly their true theoretical reducing
power.

In the crucible assay of sulfide ores, empirical rules are applied
to flux proportioning in order to ensure satisfactory slags and
buttons. These rules are developed more fully in a later section.
The conditions to be met in order to satisfy the need for nearly
complete oxidation of sulfides may be summarized as follows:
(1) provide sodium carbonate for slag at least equal to the weight
of the ore sample, (2) provide twice as much litharge for the slag
as sodium carbonate, and (3) keep the silicate degree of the slag
below 1.0. Additional sodium carbonate is needed for the forma-
tion of sodium sulfate, in the ratio of 3.34 sodium carbonate to
1 sulfur; and additional litharge must, of course, be provided for
the button. Whether or not additional fluxes, or altered flux
ratios, are required depends upon the nature of the metal portion
of the sulfide and the difficulty of carrying the metal oxide in the
slag.

Oxidizing Agents. Tn the controlled reduction methods of
assaying, oxidation is used to lower the reducing effect of sulfide
ores that would otherwise produce too large a lead button.
Oxidation can be produced by heating the ore when it is spread
out in a thin layer, exposed to the air (roasting). The roasting
process is slow and serves only to destroy the reducing power of
the ore so that the size of the lead button can be controlled in a
subsequent crucible fusion with a known reducing agent. This
method is seldom used in modern assaying. Much time and
effort can be saved by mixing an oxidizing agent with the ore
in a crucible charge so that both oxidation and fusion can be car-
ried out in the crucible. Either a single-stage or two-stage
process may be used. In the single-stage, or ordinary niter,
process just enough oxidizing agent is added to oxidize the excess
reducing effect (R.E.) of the ore and to produce the desired size of
lead button.

108 FIRE ASSAYING

The two-stage process, known as the " oxidation-collection
method/' has the advantage that the amount of oxidizing agent
does not need to be carefully proportioned to suit each individual
ore. This is avoided by mixing an excess of oxidizing agent with
the crucible charge.

Complete oxidation of the reducing minerals is accomplished
in the early stages of the fusion, and the excess oxidizing agent is
decomposed by heat. Near the end of the fusion, collection of
the precious metals is brought about by the addition of a briquette
containing litharge and sufficient iron filings to reduce a lead
button of the desired size.

The oxidizing agent universally used in either the single-stage
or two-stage process is niter (potassium nitrate). Niter decom-
poses when heated to over 400C. The first step consists of the
evolution of 1 atom of oxygen leaving potassium nitrite.

2KN0 3 = 2KN0 2 + O 2

With continued heating, KNO 2 decomposes, evolving more oxy-
gen and nitrous oxide and leaving potassium oxide. 1

2KN0 2 = K 2 + N 2 + 2

According to these reactions, 1 gram-molecule of niter (102 g.)
will furnish 1 gram-molecule of oxygen, which will oxidize
1 gram-molecule of carbon or equivalent reducer, and as a result
will prevent the reduction of 2 gram-molecules of lead (414 g.).
This gives a value of 41 ^io2> or 4.1 for the oxidizing power (O.P.)
of niter in terms of lead. In the usual assay charge, an oxidizing
power of about 4 is the value actually obtained. Fulton and
Sharwood 2 obtained an oxidizing power value of 5.1 for niter
in special charges containing only litharge, niter, and charcoal.
This corresponds to a theoretical oxidizing power of 5.12, and was
obtained by considering all the oxygen, except that required for
K 2 O, available for oxidation and that no oxides of nitrogen are
evolved. Whether 4.1 or 5.1 is the theoretical oxidizing power of
niter is of little importance, because the actual oxidizing power is

I PEESCOTT, A. B., and JOHNSON, J. C., "Qualitative Analysis," 7th ed.,
p. 288, D. Van Nostrand Company, Inc., New York, 1916.

2 FTTLTON, C. H., and SHARWOOD, W. J., "A Manual of Fire Assaying,"
3d ed., p. 78, McGraw-Hill Book Company, Inc., New York, 1929.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 109

close to 4, and this value is satisfactory in the range of slag
compositions required for sulphide ores.

Oxidizing Minerals. Some ores contain minerals such as hema-
tite, magnetite, and pyrolusite, which give up a part of their
oxygen in the fire assay. The oxygen given up oxidizes reducing
agents so that when a large amount of any of these oxidizing
minerals is present in the charge, the amount of reducing agent
must be increased.

The theoretical oxidizing power of hematite (FezOz) is derived
from the equations

2Fe 2 8 + C -* 4FeO + CO 2
2PbO + C -r 2Pb + C0 2

One gram-molecule of hematite combines with an amount of a
reducing agent that is equivalent to 1 gram-molecule of lead;
therefore the oxidizing power of hematite, in terms of lead, is
20 / x L60 r l-31. Analogous calculations show that the theoretical
oxidizing power of pyrolusite (MnO 2 ), when reduced to MnO, is
2.4; and the oxidizing power of magnetite (FeaO^, when reduced
to FeO, is 0.89. In practice, the full oxidizing power of oxidizing
minerals is seldom attained, either on account of incomplete
reduction or by loss of oxygen gas from the charge.

Summary. Size of Lead Button in the Crucible Assay. In
general the assayer desires to produce from a crucible assay
fusion, a lead button that weighs about 25 g. If the ore contains
no oxidizing or reducing minerals a reducing agent such as flour
or charcoal is mixed with the charge. The weight, in grams of
reducing agent required, is figured by dividing the weight of the
lead button desired by the reducing power of the particular
reducing agent used. When an ore contains either oxidizing or
reducing minerals, or both, the net oxidizing or reducing effect
of the ore must be determined if the ordinary single-stage assay
is to be used. The oxidizing or reducing effect can be calculated
from an estimate of the minerals present, and a knowledge of
their oxidizing or reducing powers; or it can be determined by a
preliminary assay. If the effect of the ore is oxidizing, a reducer
is added to give the desired size of lead button with sufficient
extra reducer to balance the oxidizing effect of the ore. If the
effect of the ore is reducing but not sufficient to produce a large
enough lead button, reducer is added to make up the difference.

110 FIRE ASSAYING

If the reducing effect of the ore is just right, neither an oxidizing
nor a reducing agent need be added. When the reducing effect
of the ore is so great as to produce oversize buttons the weight
desired is subtracted from the reducing effect of the ore, and
the excess reducing effect is balanced by adding niter. The
necessity for either an estimation or a preliminary fusion may
be eliminated by the new oxidation-collection method (Chap.
VII).

Scorification Assay. In the scorification assay a small ore
sample usually 0.10 assay ton is mixed with granulated lead
in a shallow fire-clay dish (scorifier) and heated in an oxidizing
atmosphere. A small amount of silica or borax glass is added
if the ore is free from silica, and the slag obtains additional
silica by attack upon the scorifier. Roasting reactions remove
sulfur and other reducing agents. A part of the lead is oxidized
progressively throughout the process, and the litharge produced
forms the greater part of the slag. At first, a ring of slag forms
at the outer periphery of the lead, which, on account of its high
surface tension, assumes an ellipsoidal shape, whereas the slag
surface is nearly flat. The process stops when sufficient slag has
formed to cover completely the diminishing pool of lead. Hence,
scorification is, in effect, a controlled oxidation method in which
the final volume of the slag and the size of the lead button bear a
geometrical relationship to each other; and the button size is
dependent chiefly upon the following factors; (1) the amount of
granulated lead added at the start, (2) the size and shape of the
scorifier, and (3) the amount of ore and flux added. The charac-
ter of the ore and minor variations in the furnace temperature do
not seriously influence the size of the lead button.

THE SPEISS PHASE

When arsenic or antimony is present in the lead smelter or in
assay fusions it can appear in one or more of the three phases:
slag, speiss, or metal. If arsenic and antimony are oxidized,
they can neither form a speiss nor enter the metal phase. On
the other hand, when they are reduced, they cannot enter the
slag and consequently must appear in either the speiss or metal
phase. Where they do appear, when reduced, depends upon the
reduced metals present. If lead is the only metal present,
reduced arsenic or antimony alloys with it and the lead is notice-

THEORETICAL DISCUSSION OF ASSAY FUSIONS 111

ably hardened. Speiss consists of compounds such as Fe 3 As,
NiAs, CuaAs, and CusSb, formed by arsenic or antimony with
one or more of the following metals: iron, nickel, cobalt, or
copper. In lead smelting the speiss-forming metals are usually
present, so that both speiss and hard lead are formed.

Speiss in Lead Smelting. Speiss dissolves precious metals,
and it is therefore undesirable in either an assay or a lead smelter.
The assay er can and must avoid speiss formation, but the lead
smelters are still troubled with this undesirable product. The
amount of reduction necessary in the two processes accounts for
its formation in one case and not in the other. In controlled
reduction assaying, only part of the lead in the charge is reduced ;
while in the blast furnace of a lead smelter, strongly reducing
conditions must be maintained in an effort to recover all the lead.
The relatively strong reduction in a lead blast furnace not only
prevents the oxidation of arsenic and antimony but also reduces
part of the speiss-forming metals iron, cobalt, nickel, and
copper some of which are always present in the blast-furnace
charge. Under these conditions, speiss forms.

Speiss in Assaying. In a controlled reduction assay fusion of
materials high in antimony and arsenic there is little danger of
speiss formation in the absence of copper. The other speiss-
forming metals are too high in the electromotive series to be
reduced in a normal controlled reduction assay. As a conse-
quence, when copper is absent, arsenic and antimony even
though reduced can only cause trouble by contaminating the
lead button; no speiss phase will be formed. In the presence of
copper, speiss will form in the assay fusion unless the charge is
proportioned with sufficient excess of litharge and soda to keep
the copper as well as the arsenic and antimony oxidized. Uncon-
trolled reduction methods should never be used for assaying
materials containing large amounts of arsenic or antimony. The
excess reduction used in uncontrolled reduction methods will
reduce arsenic and antimony, as well as some of the speiss-forming
metals, so that speiss is almost certain to form. When a speiss
phase forms, it is found as a brittle substance just above the lead
button. It can be distinguished from matte by its bright metallic
luster. The danger of loss of precious metals in the presence of
speiss occurs primarily through loss of small fragments of speiss
when separating the lead button from the slag; but even though

112 FIRE ASSAYING

all the speiss could be recovered, it cannot be cupeled
satisfactorily.

THE MATTE PHASE

Artificial sulfides of the heavy metals are called "matte." In
general, mattes are almost completely insoluble in slag and have
only slight solubility in metal or speiss phases, but they are
soluble in molten alkaline salts such as sodium carbonate and,
particularly, sodium sulfide.

Conditions of Matte Formation and Decomposition. Matte
can be formed by the fusion of metals with elemental sulfur, the
reduction of metal sulfates, or simply by melting sulfide minerals.

In assaying, as well as in metallurgical work, natural sulfide
minerals are usually present in the ore, and sulfur must be
removed in order to produce metal. Oxidation is the process
almost universally used for the separation of sulfur from metals.
The oxidation reactions are similar, whether the process is an
oxidizing roast (heating unmelted material in an excess of air), a
niter crucible assay (in which oxygen is supplied by niter instead
of air), or a copper converter operation (blowing molten matte
with a limited amount of air).

In an oxidizing roast, sulfide minerals are heated to above their
ignition temperatures, in the presence of air. The sulfur is
burned to SO 2 , and some is further oxidized to SO 3 . Metals
above silver in the electromotive series are oxidized, while the
more noble metals are left in the reduced condition. If the
temperature is held low, sulfates are formed by the combination
of metal oxides with sulfur trioxide. Sulfates dissociate at
higher temperatures, and, by heating to over 800C., nearly all
sulfate sulfur can be eliminated, provided that the SO 3 gas is
carried away. A few typical reactions are given below:

Argentite:

Ag 2 S + 2 = 2Ag + S0 2
Pyrite:

4FeS 2 + H0 2 = 2Fe 2 3 + 8SO 2

If pyrite is heated to about 500C., in the absence of air, 1 atom
of sulfur separates as elemental sulfur, leaving FeS. In the

THEORETICAL DISCUSSION OF ASSAY FUSIONS 113

presence of air, ignition takes place before the separation of the
sulfur atom.

Galena:

2PbS + 30 2 = 2PbO + 2S0 2
Sulfate Formation:

2S0 2 + 2 ^ 2S0 3

Equilibrium depends upon the temperature and the partial
pressure of the constituent gases.

PbO + S0 3 ^ PbS0 4

Equilibrium also depends upon the temperature and upon the
pressure of the sulfur trioxide gas.

Except for a difference in the ignition temperature and tem-
perature of decomposition of the sulfates, the other metal sulfides
oxidize in a manner similar to the reactions given.

Sulfur has a tendency to combine with oxygen that, at smelting
temperatures, is about the same as the tendency for lead to
oxidize. Consequently, metals above lead in the electromotive
series cannot be reduced by sulfur, but sulfur will take oxygen
away from the oxides of metals below lead in the series. Roast-
reaction and copper converting processes depend upon the use
of sulfur as a reducing agent.

Roast-reaction Process. The roast-reaction process can be
used to recover, from sulfide ores, metals low in the electromotive
series. The process consists of roasting the ore to produce metal
oxides and sulfates, and the roasted material is then mixed with
fresh sulfide ore and melted. The following reactions for the
smelting of lead in the reverberatory furnace are typical:

Roast:

Reaction:

2PbS + 30 2 = 2PbO + 2S0 2
PbS + 20 2 = PbS0 4

PbS + 2PbO 3Pb + S0 2
PbS + PbS0 4 = 2Pb + 2S0 2

114 FIRE ASSAYING

The roast-reaction process takes place in many smelting and
assaying operations. Processes depending entirely upon its
operation for metal production are now seldom used.

Converting. Converting finds its important use in the produc-
tion of copper and has no counterpart in assaying. The tend-
ency of lead to combine with oxygen is too close to that of sulfur
to allow selective oxidation of sulfur from PbS in a converter.
In copper smelting, a matte consisting of cuprous sulfide and
ferrous sulfide (Cu2S, FeS) is produced. While still liquid, the
matte is placed in a converter and blown with air. During the
first stage of the blow, iron and sulfur are selectively oxidized
and white metal, Cu2S, is left. In the second stage, blowing is
continued, selectively oxidizing sulfur and leaving blister copper.
Metals such as lead, nickel, iron, and others above copper in the
electromotive series are oxidized with the sulfur, and either go
out in the converter gas or are found in the slag. The precious
metals are collected by the copper matte and alloy with the
metallic copper that was produced in the converter.

Matte in Assaying. All processes of assaying sulfide ores,
except for the soda-iron method, require the almost complete
elimination of sulfide sulfur. If sulfide sulfur is not eliminated,
matte will form with at least one of the metals present in the
metallic state. The probable order of matte formation is given
by Dean 1 as MnS (greatest tendency to form), Cu2S, PbS, NiS,
FeS, Ag 2 S, and ZnS.

Matte in Uncontrolled Reduction. In the uncontrolled reduc-
tion (soda-iron) assay method, an excess of iron is used in an
attempt to replace all other metal sulfides with iron. Ferrous
sulfide dissolves in the alkaline salt slag. The amount of the
precious metals dissolved by ferrous sulfide and carried into the
slag is low, but most other sulfides would cause excessive loss of
precious metals.

The presence of copper-bearing minerals is particularly danger-
ous, as copper is reduced by the iron and will take sulfur from the
iron to form Cu2S a good solvent for the precious metals. Other
impurities such as nickel, cobalt, arsenic, antimony, bismuth, or
tellurium, when present in an ore, are reduced by the excess iron

1 DEAN, R. S., "Theoretical Metallurgy," p. 210, John Wiley & Sons, Inc.,
New York, 1924.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 115

and cause trouble either by forming matte or speiss or by con-
taminating the lead button.

Matte in Controlled Reduction. Lead is the only metal that
should be present in the reduced condition and available for
matte formation in controlled reduction methods of assaying.
The precious metals do not form sulfides, because lead has
a greater tendency to combine with sulfur, and there is always a
large excess of lead present to take any sulfur available. When
lead sulfide forms in an assay fusion it may not appear as a sepa-
rate matte phase, as lead and lead sulfide are mutually soluble
when liquid. A lead-lead sulfide alloy is formed in which the
components are soluble when liquid and insoluble when solid. 1
Only when the lead button is cooled very slowly do the con-
stituents have the opportunity to segregate according to their
density. Usually the presence of lead sulfide in the button is
evidenced only by the brittleness of the lead-lead sulfide alloy.
Brittle buttons are to be avoided since it is almost impossible to
separate them from the slag without loss.

Conditions Favoring Sulfide Elimination. The elimination of
sulfide sulfur is brought about by oxidation. In assaying, three
sources of oxygen are used: (1) air, as in scorification ; (2) niter,
as in the niter crucible assay; and (3) litharge, as in all assay
fusions to the extent that lead oxide is reduced to lead.

In the scorification assay, sulfur is eliminated in a manner
similar to oxidizing roasting. In addition to the oxidi zing-roast
reactions, liquid lead oxide attacks sulfides by the roast-reaction
process.

The amount of oxidation of sulfur and other substances that
oxidize as easily as lead can be increased by increasing the diam-
oter of the scorifying dish. Sufficient oxidation for a ^o- assav ~
ton sample of pure sulfide mineral is usually obtained in a 3-in.
scorifier, which will allow the oxidation of about 40 g. of lead.

In the crucible assay, complete elimination of sulfide sulfur and
the simultaneous reduction of a lead button are difficult, except in
the presence of excess sodium carbonate and litharge. Niter can
be used in the single-stage process only to oxidize sulfur above
that required for the lead button. Complete elimination of
sulfide sulfur depends upon the reaction between metal sulfides
and lead oxide. -

1 Ibid., p. 203.

116 FIRE ASSAYING

PbS + 2PbO ^ 3Pb + S0 2

This reaction will reach equilibrium without completely eliminat-
ing sulfide sulfur unless plenty of PbO is readily available and
unless the sulfur dioxide concentration of the environment is
kept low. 1 Sodium carbonate is added to the crucible assay
charge for the purpose of keeping the concentration of sulfur-
dioxide low. It does this by diluting the gas in the interstices
of the charge with carbon dioxide and also by the removal of
sulfur to the alkaline salt layer as sodium sulfate.

If either too small an amount of litharge or too little sodium
carbonate or too large an amount of acid fluxes is added to a
single-stage niter assay, the RP. of the sulfide minerals decreases.
Sulfide sulfur is not completely eliminated, and brittle lead but-
tons are produced. The effect of acid fluxes is due to the slag-
forming reactions, which reduce the amount of free soda and
litharge available for desulfurization.

THE SLAG PHASE

Slags, when liquid, consist of igneous solutions of basic and
acidic oxides and the chemical compounds that form between
them. Substances other than oxides are generally not miscible
with slags and enter some other phase. The primary require-
.ment of a slag is that it must take up the nonvolatile impurities
from the ore and hold them in a liquid condition, which will
allow the metal phase to separate completely at the furnace
temperature available. Other generally desirable slag features
are: (1) the slag should attack as little as possible the refractory
with which it is in contact and (2) the cost of reagents (fluxes)
required to impart the desired slag properties should be low.
In certain particular cases it is necessary that the slag carry
special reagents for their chemical effect on other phases. For
example, in the crucible assay of a copper ore or in the softening
of lead bullion an excess of litharge is carried in the slag for its
oxidizing effect.

Smelter Slags. Copper and lead smelter slags consist essen-
tially of the basic oxides CaO and FeO and the acidic oxide SiO 2 .
The total of these three constituents usually amounts to over

1 MAIER, C. G., Thermodynamic Data on Some Metallurgically Important
Compounds of Lead and the Antimony-group Metals and Their Applications;
U.S. Bur. Mines R.L 3262, p. 39, 1934.

THEORETICAL DISCUSSION OF ASSAY FUSIONS 117

80 per cent of the slag. The remainder of the slag consists of
particular oxides taken up from the material smelted. Analyses
of slags typical of lead blast-furnace smelting and copper rever-
beratory-furnace smelting are given below:

LEAD BLAST-FURNACE SLAG, MIDVALE 1935 AVERAGE*

Au,
oz./ton

Ag,

oz./
ton

Cu,

Pb,

Si0 2 ,

A1 2 8 ,

FeO&
MnO,

CaO,

MnO,

BaO,

ZnO,

0.00079

0.058

0.16

0.95

25.51

3.73

37.70

17.43

1.37

1.13

10.93

1.68

* CLBVENGEH, G. H., Blast-furnace Practice at Midvale, Utah, Trans. A.I.M.E., vol. 121,
p. 57, 1936.

COPPER REVERBERATORY-FURNACE SLAG, GARFIELD*

Au,
oz./ton

Ag,
oz./ton

Cu,

SiO 2 ,

A1 2 3 ,

Fe &

Mn, %

CaO,

(MnO, BaO, ZnO
not reported)

S, %

0.0016

0.25

0.39

40 3

6.7

31 5

8 7

* HAYWABD, C. R., "An Outline of Metallurgical Practice," p. 63, D. Van Nostrand Com-
pany, Inc., New York, 1929.

Assay Slags. Basic fluxes commonly used in assaying are
sodium oxide (from sodium carbonate), and litharge. The
principal acid flux is silica, supplemented by boric oxide, which is
obtained by adding either borax or borax glass. Except* for
the particular use of litharge for oxidation, and of sodium car-
bonate for desulfurization, fluxes are added for the purpose of
lowering the melting point and imparting a homogeneous fluidity
to the melted oxide impurities. The oxide impurities most
frequently encountered are silica (SiO 2 ) from quartz, lime (CaO)
from calcite, and the various metal oxides such as FeO, Fe 2 03,
Cu2O, PbO, ZnO, As 2 5 , Sb20 3 , and A1 2 O 8 , which may be present
as such in the ore or are formed during the process. Whether
the constituents occur separately in the ore or are combined as a
silicate mineral is immaterial in slag formation. In fact, artificial
silicate minerals of microscopic size frequently separate from
slowly cooled slags. The different artificial minerals or com-
pounds formed in slags are generally mutually soluble when liquid
and insoluble when solid. The melting point of mixtures of
substances like these (soluble when liquid and insoluble when

118

FIRE ASSAYING

solid) decreases according to Raoult's law, with the addition of
one substance to the other, and there is one composition having
the lowest melting point of the series. This is called the
"eutectic."

1700

1600

800

700

40 50 60 70 80 90
Composition, weight percent Si Oz

FIG. 10. Constitutional diagram, system Na 2 O.8iO2-SiO 2 . (From Inter-
national Critical Tables, Vol. IV, p. 87, McGraw-Hill Book Company, Inc.,
1928.)

For an example of fluxing, consider an ore containing free gold
and consisting almost entirely of quartz. Silica will not melt
at a temperature that can be obtained in the assay furnace,
so that a flux must be added to lower the melting point. Either
sodium carbonate or litharge could be used. Sodium carbonate,
when melted with silica, gives up carbon dioxide gas, and sodium
oxide unites with silica. The constitutional diagram of the

THEORETICAL DISCUSSION OF ASSAY FUSIONS 119

sodium oxide-silica series (Fig. 10) shows two compounds:
Na 2 O.SiO 2 and Na 2 O.2SiO 2 . The eutectic between Na 2 O.2SiO 2
and silica (containing 73.1 per cent Si0 2 ) melts at 793C.

This slag composition might be considered desirable for th,e
assay of a quartz ore, but it is unsatisfactory because, even though
it melts at a low temperature, the melt is too viscous at assay-
furnace temperatures to allow satisfactory separation of the

73V

900
d 850

1*800
o

|750

JTOO

650
600

0*5*

743

.754

725-

\ /

N/l

-732

^
*

V/5 -

7/0

Ij
1!

4!
^i

) 5 10 15 20 25 30 35

Composition, weight percent Si
FIG. 11. Constitutional diagram, system PbO-SiC>2. (Melting points from
Geller, Creamer, and Bunting, Bur. Standards, Jour. Research, vol. 13, p. 243,
1934.)

metals. Litharge also produces low melting compounds and
eutectics with silica (Fig. 11). Two compounds, PbO.SiO 2 and
2PbO.SiO 2 , are formed in this series. Either of these com-
pounds, or the eutectic between them, melt at low temperatures,
and the melts are sufficiently fluid for good separation of the
metals.

Soda may be used to replace a part of the litharge and obtain
equally satisfactory slags. The compounds and eutectic that
form when three components are used become quite complex.
In general, bisilicate or sesquisilicate slags are desired because
they give good fusion of most basic impurities, and also because
they scarcely attack the crucible. More basic slags are used only

120 FIRE ASSAYING

to provide excess litharge for its oxidizing effect on sulfur, copper,
arsenic, antimony, bismuth, or tellurium, and to provide excess
sodium carbonate for its ability to replace a part of the more
expensive litharge, and also for its desulfurizing effect. Many
bisilicates, for example, CaO.Si0 2 , are not sufficiently fluid in
an assay charge without the presence of other constituents such
as soda and litharge. These fluxes are used in nearly every
crucible assay, even when the ore itself is essentially basic. In
Chap. VII, practical rules are given for proportioning crucible
assay charges.

THE ALKALINE SALT PHASE

Slags are mainly composed of oxides, and, as a general rule, only
oxides will dissolve in them. In fusions containing compounds
such as chlorides, sulfates, or other similar metal salts, a separate
layer or phase is formed. This layer occurs mainly as salts of
the alkali metals, and, consequently, it is termed the "alkaline
salt phase." The sodium sulfate layer that forms at the top of a
crucible fusion of a sulfide ore is a typical example of this phase.

In the past, assayers added a cover of sodium chloride to
crucible assay charges, to form, deliberately, an alkaline salt
phase. Sodium chloride melts at a low temperature and is very
liquid when heated slightly above its melting point. The use of
a salt cover was thought to prevent dust loss, to prevent lead
globules or unfused ore particles from sticking to the upper part
of the crucible wall, to aid in preventing gas evolution from caus-
ing the charge to boil over, and to prevent either reducing or
oxidizing furnace gases from acting on the charge so as to affect
the size of the lead button. Ordinarily, no trouble is encountered
from any of these causes, even without the use of a salt cover,
and its use has been abandoned by most assayers because they
wish to avoid any chance of losing precious metals through the
formation of volatile chlorides. A few assayers continue the
cover idea by adding some borax glass or sodium carbonate, or a
mixture of the two on the top of the crucible charge. These
substances enter the body of the charge and become a part of
the slag as soon as boiling starts.

In smelting operations, alkaline salts are used in the Harris
process of softening lead bullion, but otherwise they are seldom
encountered.

CHAPTER VII
THE CRUCIBLE ASSAY

In the crucible assay a suitable portion of the sample is melted
with fluxes in a fire-clay crucible, arid oxidation-reduction reac-
tions are controlled so as to produce a lead button by reduction
of litharge. Gold and silver are collected by the lead button,
which is subsequently cupeled and further treated to ascertain
the precious-metal content of the sample.

The chief factors that must be controlled by the assayer in
order to ensure reliable results are:

1. The weight taken for assay.

2. The fluxing of the ore constituents.

3. The control of lead reduction.

All types of ores and metallurgical products may be assayed for
precious metals by the crucible assay, and for most materials
this process is preferred, but certain substances, particularly
alloys, may require preliminary acid treatment. The silver
content of silver-rich alloys is usually determined by chemical
methods. Some other materials may be assayed more con-
veniently by scorification (Chap. VIII).

With respect to fluxing requirements, ores may be broadly
classified into acidic ores, high in silica; and basic ores, containing
metallic compounds. With respect to oxidation-reduction reac-
tions, ores may be classified as neutral, reducing, or oxidizing.
In practice, any possible combination may be found : some of the
minerals in an ore may be acidic, others basic, some may have a
strong reducing power, others may be neutral or oxidizing.
The assayer is concerned with the composite fluxing requirements
of the sample and with the net oxidizing or reducing power.

PRELIMINARY INVESTIGATION OF SAMPLE

The general nature of the sample submitted for assay must be
known before the assayer can make an intelligent choice of
reagents for lead-button control and for fluxing.

121

122 FIRE ASSAYING

Visual observation of lump samples is the quickest and most
reliable method of estimating the approximate mineralogical
composition of a sample. If lump samples are not available, a
simple concentration of the pulverized sample in a pan or vanning
plaque may be applied, or the sample may be examined under
water with a low-power microscope.

Vanning Test. A 3- or 5-in. watch glass is quite suitable in
place of a pan or vanning plaque. One or two grams of the
sample is placed on the watch glass, covered with water, and then
allowed to stand until wetting is complete. Extreme fines
("slimes") that do not settle readily are removed by successive
shaking and decantation; the granular residue is -then stratified
by a gentle shaking motion, and the layers of different minerals
are spread out for observation by causing an accelerated flow of
water against one edge of the stratified sample. The heaviest
minerals will be on the bottom, the lightest on top and, after
fanning out, the heaviest minerals segregate at the apex of the
fan toward the side from which the water current was directed.

To make percentage estimates from visual observations, it is
necessary to take into consideration the relative specific gravity
of the different mineral groups. For this purpose the light-
colored gangue minerals, except barite and a few other minerals
of lesser importance, may be taken as unity. Galena has approxi-
mately three times the specific gravity of quartz or limestone,
other sulfides and the oxides of heavy metals have approximately
twice the specific gravity of quartz.

Qualitative Tests. Simple qualitative chemical tests on pul-
verized samples may be used by the assayer to aid in the classifica-
tion of unknown samples. The following tests are useful in the
identification of certain substances that are not readily identifi-
able by visual observation.

Effervescence and evolution of an odorless gas with cold dilute
(1:1) hydrochloric acid indicate the presence of calcite (CaC0 3 );
a similar effect with warm dilute hydrochloric acid indicates
dolomite [(Ca, Mg)CO 8 ]. This will serve to distinguish these
important basic carbonates from silicates, all of which are com-
paratively inert with acids.

Sphalerite (ZnS) is attacked by warm hydrochloric acid with
the evolution of hydrogen sulfide, which has a characteristic odor
of spoiled eggs. The light-colored varieties of sphalerite are

THE CRUCIBLE ASSAY 123

difficult to determine by visual observation of the pulverized
material.

All sulfides are attacked by warm concentrated (or dilute)
nitric acid, especially with the aid of a few crystals of potassium
chlorate. The rapidity of attack and the evolution of copious
brown fumes are characteristic.

Copper and iron are easily detected by cooling gfod diluting a
nitric acid solution, then neutralizing with ammonium hydroxide.
Iron is precipitated as a flocculent brown precipitate of ferric
hydroxide. Copper imparts to the solution an intense blue color.

The presence of nickel in ores can be confirmed by neutralizing
an acid solution of the sample with ammonium hydroxide until
the odor of ammonia faintly persists. Then filter and add an
alcoholic solution of dimethylglyoxime. A brilliant red precipi-
tate identifies nickel and forms in proportion to the amount of
nickel present.

Slag Colors. A preliminary assay fusion is often made to
determine the reducing or oxidizing power of an ore. The color
of the slag obtained from assay fusions frequently furnishes
qualitative information as to the presence of some elements.

To most vitreous slags, copper imparts a green color that is
particularly observable on the sides of the crucible or scoriiicr.
In basic slags, high in litharge, copper gives a brick-red color.

Manganese gives a purple color to assay slags. In high propor-
tions the color may be so deep as to appear black, but the purple
color becomes apparent in thin sections or in pulverized samples
of the slag. A black film may appear on the surface of basic
slags containing manganese.

Cobalt is a strong coloring agent and imparts a blue color to
slags of all types.

Antimony gives a greenish-yellow color to vitreous slags, but
this color is easily masked by small amounts of more highly
colored compounds.

Iron-bearing slags are yellow-brown, brown, or black, depend-
ing upon the amount present and the slag composition. Acid
slags containing small amounts of iron are pale green and trans-
parent. On account of the common occurrence of iron in ores,
the iron color may mask all other colors; and only copper, man-
ganese, and cobalt yield sufficiently strong colors to overcome the
effect of an approximately equal amount of iron. When iron

124 FIRE ASSAYING

and other strong coloring agents are absent, CaO, MgO, or ZnO
may give a white color to the slag.

WEIGHT TAKEN FOR ASSAY

The weight of sample to be taken for assay must be such that
the cumulative errors of weighing and manipulation are within the
desired limits of accuracy. The underlying principles involved
in the determination of sample weight are considered in Chap. II
on sampling.

For most mine samples, it is customary to use a J^-assay-ton
sample. The errors in taking the original sample are such that
extreme refinement of the assaying procedure is superfluous.
For materials that are known to be low grade, such as tailings,
at least 1 assay ton should be taken for assay; and 2 assay tons
is the usual sample weight for low-grade tailings. For con-
centrates and high-grade samples, 34-assay-ton samples are
commonly employed, and even smaller portions may be used for
exceptionally rich products.

If the sample is of material so complex that the required portion
cannot be fluxed in a crucible of convenient size, two separate
portions of one-half of the required total may be assayed and the
results combined to give practically the same degree of accuracy
as would a single assay of the same gross sample weight. This
procedure should not be confused with check assays that may be
required in addition.

FLUXING FOR THE CRUCIBLE ASSAY

When sulfide ores are assayed by the crucible process, the metal
originally combined with sulfur is converted to an oxide that,
with a few exceptions, acts as a base in assay slags to combine
with silica. Since carbonates are also converted to oxides in
the assay crucible, and since the metallic oxides in complex silicates
are already acting as base oxides, it is evident that the problems of
fluxing are virtually independent of the problems of controlling
the lead reduction.

As has been shown in Chap. VI, poly basic silicates or boro-
silicates containing sufficient soda or litharge or both satisfy most
of the requirements of a good assay slag. Such slags may be
successfully employed for practically all ores and metallurgical

THE CRUCIBLE ASSAY

125

3 *

S ^

- s

S"?

. TJ

I I

H- S

I 5

.2 4-

O
^.. +

<NO<N<N rHl

M^i

a q,

S15

5"E *

111

HI
US

s g g
S a *

S "O v fl

^55-2
J5 < H -g

126 FIRE ASSAYING

TABLE XII. SLAG FACTORS, EQUIVALENT WEIGHTS PER UNIT OF SILICA

Dominant
element

Compound

Original form

Combining
form

Wt. orig. form per
unit wt. SiOa

Silicate degree

Sub

Mono

Sesqui

Acid Fluxes

Boron
Boron

Borax glass
Borax

Na 2 O.2B*0 3
Na 2 O.2B 2 3 .10H 2 O

Na 2 O.2B 2 O 3 *
Na 2 O.2B 2 O 3 *

1.2
2.2

1 3
2.4

1 5
2.8

1 7
3 1

Basic Fluxes

Lead
Sodium
Potas-
sium

Litharge
Sodium carbonate
Potassium carbo-
nate

PbO

NaaCOs
K 2 C0 3

PbO
Na 2 O
K 2 O

14 9
7.1
9.2

7.4
3.5
4.6

4 9
2 4
3 1

3.7
1 8
2 3

Basic Ore Constituents

Antimony

Antimony triox-

Sb 2 O 3

Sb 2 3

G.5

3.2

2 2

1 6

ide

Stibnite

Sb 2 S 3

Sb 2 O 3

7.5

3 8

2 5

1 9

Calcium

Calcium oxide

CaO

3 7

1.9

1 2

0.9

(lime)

Calcite or lime-

CaC0 3

CaO

6 7

3.3

2 2

1 7

stone

Copper

Cuprous oxide

Cu 2 O

Cu 2

9 5

4 8

3 2

2 4

Cupric oxide

CuO

Cu 2 O

10.6

5 3

3 5

2.7

Chalcocite

Cu 2 S

Cu 2 O

5 3

3 5

2 7

Chalcopyrite

CuFeSa

Cu 2 O.2FeO

6 1

3 1

1 5

Iron

Ferrous oxide

FeO

4 8

2 4

1 6

1.2

Hematite

Fe 2 O 3

FeO

5 3

2.7

1 8

1.3

Magnetite

Fe 3 O4

FeO

4 9

2.5

1.6

1 2

Pyrite (rnarca-

FeSa

FeO

4.0

2.7

2.0

site)

Lead

Lead oxide (lith-

PbO

14.9

7.4

4.9

3.7

arge)

Galena

PbS

PbO

15.9

5.3

4.0

Magne-

Magnesium oxide

MgO

2.7

1.3

0.9

0.7

sium

(magnesia)

Magnesite

MgCOs

MgO

5.6

2 8

1.9

1.4

Manga-

Manganous oxide

MnO

4.7

2.4

1.6

1.2

nese

Pyrolusite

MnOz

MnO

5.8

2.9

1.9

1.4

Zinc

Zinc oxide

ZnO

5.4

2.7

1.8

1.3

Sphalerite

ZnS

ZnO

6.5

3 2

2.2

1.6

Note : The fluxing of aluminum and arsenic is explained in Table XI.
* The slag factors for borax and borax glass compensate the silicate degree for the base
NasO in the borax or borax glass.

THE CRUCIBLE ASSAY 127

products; but for certain materials, notably those high in copper,
oxide slags rich in litharge are preferable.

Although it is theoretically possible to calculate an ideal assay
slag for any given ore, the exact analysis of an ore is seldom known
to the assayer, and even if a complete analysis were available, the
time required to calculate the ideal charge is not warranted unless
a large number of assays are to be made on nearly similar material.
Hence, for simplicity and convenience, assayers use an excess of
fluxes and proportion these so that only minor variations are
needed to flux any of the samples currently received for assay.

The fluxing requirements of various types of ores are sum-
marized in Table XL

Table XII gives slag factors for use in calculating charge com-
positions. For convenience in calculating the equivalents of
compounds not listed in the table the first compound under each
dominant element of the basic ore components is the oxide that
combines with silica or borax in assay slags. The equivalents
of compounds not listed can be found by first calculating the
equivalent weight of the corresponding oxide and then dividing
this value into the factor from the table.

The use of the slag-factor table is illustrated in the following
examples :

Example 1 : Find the amount of sodium carbonate and litharge
in equal molecular proportions to form a sesquisilicate with 15 g.
of silica.

Since half of the SiO 2 will be fluxed by each base, 7.5 g. of
SiO 2 are to be combined with Na 2 CO 8 and 7.5 g. with PbO.

From the slag-factor table the sesquisilicate factor for Na2C03
is 2.4 and the factor for PbO is 4.9.

Therefore

(2.4) (7.5) = 18.0 g. of Na 2 CO 3 required
and

(4.9) (7.5) = 36.8 g. of PbO required

Example 2: Find the bisilicate silica equivalent of 15 g. of
limestone and the amount of borax glass required to replace
one-third of the required SiO 2 .

From the slag-factor table, 1.7 g. of CaCO 3 combines with 1 g.
of Si0 2 to form a bisilicate.

128 FIRE ASSAYING

Hence

15 g. CaC0 3

1.7

= 8.8 g. of SiO2 required.

The 8162 to be replaced by borax glass is 8.8/3 = 2.9 g., leav-
ing 8.8 - 2.9 = 5.9 g. of SiO 2 to be added as such.

From the table, 1.7 g. of borax glass is required to replace 1 g.
of SiO 2 in a bisilicate.

Hence (1.7) (2.9) = 4.9 g. of borax glass to be added.

Example 3: Find the monosilicate silica equivalent of the
bases in 15 g. of an ore containing 20 per cent chalcopyrite and
10 per cent bornite (FeS.2Cu 2 S.CuS).

From the table the monosilicate factor for chalcopyrite is
found to be 3.1.

Therefore, the Si0 2 required to produce a monosilicate with
the chalcopyrite in the sample is [(0.2)(15)]/3.1 = 1.0 g.

Bornite does not appear in the table, but the monosilicate
factor for bornite may be calculated by either of the following
methods :

a. Direct Calculation: For a monosilicate, 4(FeS.2Cu 2 S.CuS)
requires 14 oxygen atoms in the silica, or 7 moles of Si0 2 for
4 moles of bornite.

Therefore

Molecular weight 4(FeS.2Cu 2 S.CuS) = 4(501.93)
Molecular weight 7SiO 2 7(60.06)

= 4.77 g. bornite per g. Si0 2

and the Si0 2 required to produce a monosilicate with the bornite
in the sample is [(0.1)(15)]/4.8 = 0.31 g.

6. Indirect Calculation: The FeO equivalent of bornite is

Molecular weight FeO 71.84

Molecular weight FeS.2Cu 2 S.CuS ~" 501.93 "

The Cu 2 O equivalent of bornite is

Molecular weight 5Cu 2 S 715.70 ?

Molecular weight 2(FeS.2Cu 2 S.CuS) 1003.86

The monosilicate factor for FeO is 2.4, and for Cu 2 O is 4.8.
Hence the monosilicate silica equivalent of the FeO derived from
the 10 per cent of bornite in 15 g. of ore is

THE CRUCIBLE ASSAY 129

[(0.1)(15)(0.14)]/2.4 = 0.09 g.

Si0 2 for FeO, and the SiO 2 for the Cii2O from the bornite is
[(0.1)(15)(0.71)]/4.8 = 0.22 g. SiO 2 for Cu 2 O. Hence the total
SiO 2 for the bornite is 0.09 + 0.22 = 0.31 g. SiO 2 , which checks
the value obtained by method a.

CONTROL OF LEAD REDUCTION

Control of lead reduction involves consideration of the appro-
priate size of lead button, the estimation of the oxidizing or
reducing power of ore minerals and of added reagents, and the
proportioning of the assay charge to the end that buttons of
suitable size are reduced in the crucible.

Size of Button. The optimum size of the lead button obtained
in the crucible assay is a compromise between convenience and
the relative losses of precious metals in the slag and in the subse-
quent cupellation of the button. For convenience and accuracy
of cupellation, the cupels in a single row in the furnace should
be of uniform size, and the buttons should be of approximately
the same weight. Nearly all assayers use cupels that are either
IK or 1% in. in diameter. These hold a maximum of 32 and
45 g. of lead, respectively, and impose a practical upper limit to
the button size. Larger cupels can be purchased or made but
are rarely used. Buttons smaller than 10 g. usually fail to collect
all the precious metals, so that 15 g. should be considered a safe
minimum amount of lead. For ordinary work, most assayers aim
for a 25-g. button and accept for cupellation all buttons larger than
15 g. and smaller than the maximum capacity of the cupels in use.

General Methods of Controlling Lead Reduction. In the
crucible assay, control of lead reduction is effected by one of the
following general methods:

1. Controlled reduction, involving the use of litharge in
excess of button requirements, and the addition of a controlled
amount of a reducing or an oxidizing agent.

2. Uncontrolled reduction, involving the use of litharge (plus
the PbO equivalent of lead in the original sample) for button
requirements only, in the presence of an excess of reducing
agents.

3. Oxidation of reducing agents in the ore by a preliminary
treatment, followed by controlled reduction during the crucible
fusion.

130 FIRE ASSAYING

For the controlled reduction method, a reasonably accurate
knowledge of the reducing power (R.P.) and the oxidizing power
(O.P.) of ore minerals is required. For the two other methods,
only a general approximation of the composite state of oxidation
of the ore sample is needed. In all methods the reducing or
oxidizing power of any added reducing or oxidizing agents must
be known.

Oxidation-reduction Units. The R.P. of a reducing substance
is defined as the amount of lead reduced by a unit amount of the
reducer. The reducing effect (R.E.) of a reducing substance
is defined as the total amount of lead reduced by a given amount
of the reducer.

The O.P. of an oxidizing substance is defined as the lead equiva-
lent of a unit amount of the oxidizer, and the oxidizing effect
(O.E.) is the total lead equivalent of a given amount of the
oxidizer. It should be noted (page 109) that most oxidizing
agents in crucible assaying do not react directly with lead but do
react with reducing agents that would otherwise reduce lead from
litharge in the charge.

Reducing Agents. The effective R.P. of various reducing
agents commonly used in assaying is given in Table XIII.

TABLE XIII. EFFECTIVE REDUCING POWER (K.P.) OF REDUCING AGENTS

Effective R.P.,
Grams Pb per Gram
Substance Reducer*

Charcoal 25-30

Sulfur 13-19

Sugar 11-13

Flour 10-12

Starch 10-12

Argols 8-12

Cream of tartar 5-6

Iron filings 4-6

* The lower values apply to impure reducing agents and, to highly acid slags; the higher
values apply to pure reducing agents and to slags having a silicate degree not greatly in
excess of the monosilicate, except that impure argols have increased R.P.

The selection of a suitable reducing agent is largely a matter
of convenience and cost. Flour is generally the cheapest reagent
available and is widely used. If reducers are added separately
to the charge, reagents having a distinctive color such as argol
(burgundy color), charcoal (black), or sulfur (yellow) have the
advantage that they indicate at a glance whether or not the

THE CRUCIBLE ASSAY 131

reducer has been added to a given crucible, thus minimizing the
danger of adding it twice to one charge or omitting it entirely
from another. Reducers having a high R.P., such as charcoal
or sulfur, should be diluted with a neutral substance to permit
rapid measuring of the required amount without the necessity for
accurate weighing. Charcoal may be diluted with an equal or
greater volume of silica or sodium carbonate. If sulfur is diluted
with 3.34 parts of sodium carbonate by weight, the soda required
for the formation of sodium sulfate is automatically included
in the charge. It is obvious that diluted reducers should be
thoroughly mixed, and their R.P. determined before use.

Iron nails are used as the reducing agent in the uncontrolled
reduction method, and iron filings are used in the oxidation-collec-
tion method. Iron has the advantage over other reducing agents
that no incompletely oxidized gases that might escape from the
charge and give erratic results are formed. For this reason,
iron deserves wider use as a reducer in the controlled reduction
method with nonreducing ores, as the R.P. is not affected by the
composition of the slag. Iron is undesirable as a reducer only
in charges high in hematite or magnetite, which are not completely
reduced to the ferrous oxide by metallic iron.

Determination of the Reducing Power of a Reducing Agent.
The R.P. of the common reducing agents has been so well estab-
lished that the assayer seldom needs to determine it unless he
proposes to use an unknown material to suit the convenience of
local conditions. Since the R.P. varies to a certain extent with
the composition of the slag and with the rate of heating, the R.P.
determination should be made under conditions that correspond
to normal assaying practice.

The following charge is similar to assay charges for silicate ores
and may be used to determine the R.P. of a reducing agent:

Sodium carbonate 30 g.

Litharge 60 g.

Silica 15 g.

Borax glass 5 g.

Reducing agent 2 g. or as required.

Mix the reducing-power charge in a 15- or 20-g. crucible and fuse
in a hot furnace at 1000 to 1150C. As soon as the fusion is
complete, which should be in from 15 to 20 min., pour the charge
into a mold. When cool, separate the button from the slag and

132

FIRE ASSAYING

weigh it to the nearest 0.1 g. The R.P. is the weight of the
button divided by the weight of the reducing agent used.

In R.P. determinations the litharge entering the slag may vary
from the possible maximum of all litharge present, as in the case
of a material having no reducing power, to the possible minimum
of no litharge entering the slag, as in the case of pure carbon.
The extremes give slag compositions, with the above R.P. charge,
ranging from a monosilicate to a bisilicate. This variation in
slag composition would not change the R.P. of most reducers
more than 10 per cent from the mean.

TABLE XIV. REDUCING POWER OF CERTAIN SULFIDE MINERALS

Mineral

Formula

S, per
cent

Normal slag-
forming
oxides

R.P., grams per gram
of mineral

Com-
puted
R.P.

Actual R.P.

S to SO 3

Pre-
liminary

assays *

Niter
charges f

Galena ....

PbS

Cu 2 S
FeAsS
Sb 2 S 8
CuFeS 2
ZnS
FeS 2

13.4
20.2
19.7
28.6
34.9
32.9
53.4

PbO
Cu 2 O
FeO, As 2 O 6
Sb 2 O 8
Cu 2 O, FeO
ZnO
FeO

3.46
5.20
8.25
7.35
8.44
8.51
12.07

2.9
4.5
8.1
5.9
8.2
8.2
11.6

2.9

4.7
7.4
5.8
8.2
8.1
11.0

Chalcocite
Arsenopyrite . . .
Stibnite

Chalcopyrite . . .
Sphalerite

Pyrite

* The charge in all cases was 3 g. sulfide, 45 g. PbO, 15 g. Na2CO 8 , 5 g. SiO2.

t The values in this column were determined with monosilicate slags under normal
assaying conditions with niter and excess litharge. U.S.P. niter was used which had an
O.P. of 4.0.

Reducing Power of Sulfide Minerals. The most important-
group of minerals with high R.P. are the sulfides, although some
ores contain organic matter that may make them reducing. The
theoretical reactions and the method of calculating the R.P. of
sulfide minerals are given in Chap. VI in which it is pointed out
that the actual R.P. varies with the conditions of the fusion,
and that the R.P. of all sulfides decreases sharply as the silicate
degree increases beyond the monosilicate.

THE CRUCIBLE ASSAY 133

The theoretical R.P. of sulfides is a maximum value that is
seldom attained in actual fusions. Table XIV gives theoretical
and actual R.P. values of the minerals as determined in the
preliminary assay described in the following section, without
control of button size by the addition of an oxidizing agent.
The column headed " Actual R.P. Niter charges" gives the
values of reducing power normally obtained in controlled reduc-
tion fusions with mono- or subsilicate slags in which niter
(KN0 3 ) is used to control button size.

With the aid of Table XIV the reducing effect (R.E.) of a
given sample of ore can be estimated if the approximate mineral-
ogical composition has been determined. Many assay ers prefer
to make a preliminary fusion to determine the reducing power
of unknown sulfide ore samples, if the controlled reduction method
of assay is to be used.

Determination of the Reducing Power of Sulfide Ores. For
the direct determination of R.P. it is convenient to prepare a uni-
versal charge that may be used on any materials that are sus-
pected of having an appreciable R.P. Such a charge requires a
high ratio of fluxes to ore, and sufficient litharge to provide an
excess for the slag even with ores having the maximum possible
R.P. Moreover the silica and borax glass content should be
such that the slag composition, except for ores comparatively
high in silica, should not be more acid than a monosilicate, as
the R.P. of sulfides decreases rapidly with more acid slags, and
correlation with the conditions to be used in the final assay would
be impracticable.

A general flux recommended for the R.P. determination of ores
is as follows :

Ore 3g.

Sodium carbonate 10 g.

Litharge 45 g.

Silica 3 g.

Borax glass 1 g.

Use a 10-g. crucible and melt at normal operating temperatures.
In general the fusion will be complete in 15 min. To save time
the fusion may be poured on an iron plate so that the button may
be recovered and weighed almost at once after pouring. The
weight of button divided by the weight of ore taken is the R.P.,

134 FIRE ASSAYING

in grams of lead reduced per gram of ore used. Reference to
Table XIV may indicate the necessity for a slight correction in the
estimated R.P. in calculating niter charges for a larger quantity
of ore.

The above charge will yield buttons ranging from 10 g. in
weight for pure galena to 33 g. for pure pyrite. Allowing for soda
to form sodium sulfate and litharge for the button the calculated
silicate degree will be 1.1 for pure pyrite, 0.5 for pure galena,
0.7 for pure sphalerite, 0.4 for pure limestone, and 0.8 for pure
silica. In view of the relatively large proportion of fluxes to
ore, any of the slags will be satisfactory, although in the regular
assay a higher silicate degree would be preferred for some ore
types. Crucible corrosion is an automatic corrective factor,
as the more basic slags will increase their silica content at the
expense of the crucible yet will not greatly exceed the mono-
silicate by that process.

Preliminary R.P. assays are unreliable if the button weight
is too small. If the ore sample is suspected to have a R.P. less
than 2 it is advisable to double the quantity of ore and to add
approximately one-third more of the flux mixture.

Oxidizing Agents. In two of the methods of crucible assaying
to be described, an oxidizing agent is added to the charge in
order to counteract the reducing effect of the ore. The only
oxidizing agent commonly used by assayers for this purpose is
niter (KNO 3 ). Sodium nitrate could be used, but is too deliques-
cent for convenience. Sodium peroxide is open to the same
objection, although it has the advantage of minimizing the boiling
of the charge, by not giving off inert gases, and of providing a
concentrated source of Na 2 O for slag formation.

One gram of niter should furnish sufficient oxygen to oxidize
the equivalent of 5.12 g. of lead, provided all the oxygen was used
in the process and nitrogen gas evolved. Actually the practical
oxidizing power (O.P.) of commercial niter is 4.0, corresponding
closely to a value of 4.1 for the theoretical decomposition of
niter to N 2 0, K 2 0, and oxygen.

Oxidizing Minerals. The higher oxides of certain multi-
valent metals are capable of reduction to a lower state of oxida-
tion in an assay fusion, hence they possess an O.P. that must be
taken into account in calculating the amount of reducing agent
to be added for the desired lead fall.

THE CRUCIBLE ASSAY 135

The commonest ores of this type are those containing hematite
(Fe 2 3 ), magnetite (Fe 3 O 4 ), or the higher oxides of manganese,
as pyrolusite (Mn0 2 ). More rarely, ores containing the higher
oxides of copper, nickel, cobalt, arsenic, or antimony are encoun-
tered in practice; but the O.P. of these is seldom of importance.
The theoretical O.P. of Fe^Oa is 1.3 g. of lead per gram, of FesC^
is 0.9, and of MnO 2 is 2.4. The actual O.P. of these oxides is
somewhat less, and for the purpose of calculating assay charges
the O.P. of pure hematite may be taken at 1.0, that of magnetite
at 0.5, and that of pyrolusite at 2.0.

Determination of Oxidizing Power. The O.P. of any mineral
is not large, so that the oxidizing effect of an ore in an assay can
usually be calculated with sufficient accuracy from a visual
estimate of the amount of oxidizing minerals present. The
estimate is easily made when a lump sample is available for
inspection; pulverized ores should be vanned for observation.

An O.P. determination of an unknown ore sample or of an
oxidizing agent can be made by fusion. The following charge
is recommended:

Ore or oxidizer 5 g.

Sodium carbonate 10 g.

Litharge 45 g.

Silica 6 g.

Borax glass 2 g.

Reducer To give reducing effect of 30 g. *

* For strong oxidizers such as niter, increase the 'reducing effect to 40 g.

With hematite or magnetite the above charge will be approxi-
mately a monosilicate. The oxidizing effect of the 5 g. of mate-
rial investigated is equal to the reducing effect of the reducer
minus the weight of lead actually obtained. The O.P. is found
by dividing the oxidizing effect result by the amount of material
tested, that is, 5 g.

METHODS OF CRUCIBLE ASSAYING

A. Single-stage processes.

1. Controlled reduction.

a. Ores with deficient R.P.
6. Ores with excess R.P.

2. Uncontrolled reduction. The soda-iron method.

B. Two-stage processes.

3. Oxidation-collection method.

4. Roasting-fusion method.

5. Acid-treatment-fusion methods (combination methods).

136 FIRE ASSAYING

Before proceeding with any of the methods it is necessary to
estimate the approximate mineralogical composition of the sam-
ple in order to decide upon the weight of sample to be used
and to select a crucible of suitable size. Before actually adding
ore and fluxes to the crucible and proceeding with the fusion,
see Manipulative Procedures, pages 151ff.

A . Single-stage Processes

Single-stage processes of crucible assaying are those methods
in which the charge components are adjusted so that fusion, lead
reduction, and precious-metal collection take place entirely
within the crucible with no further reagent additions prior to
pouring.

1. Controlled Reduction Methods. Controlled reduction
methods are those in which the size of the lead button is regulated
by adjusting the amount of reducing or oxidizing agents that are
added to the charge. In these methods more litharge than that
required for the lead button is always added to the charge.

Before proceeding with this method it is necessary not only to
estimate the mineralogical composition of the sample but also
to estimate the net reducing or oxidizing power, either by calcula-
tion or by direct determination.

a. Ores with Deficient Reducing Power

CHARGE CALCULATION FOR SILICEOUS ORES. If the net R.P.
of an ore sample is less than that required for producing a lead
button of suitable size, an added reducing agent is required. If
the sample is dominantly siliceous, use the following procedure in
making the charge calculation:

1. Add sodium, carbonate equal to the weight of ore.

2. Calculate the bisilicate silica equivalent of the sodium
carbonate. From the slag-factor table, 1 it is seen that 1 part
of silica combines with 1.8 parts of Na2C03 to form a bisilicate;
therefore the weight of Na 2 CO 3 divided by 1.8 gives the bisilicate
silica equivalent of the sodium carbonate.

3. Deduct step 2 from the estimated silica in the ore.

4. Calculate litharge to form bisilicate with the remaining
silica: from the slag-factor table, 3.7 parts of litharge are required

1 Table XII, p. 126; also inside rear cover.

THE CRUCIBLE ASSAY

137

for each part of silica; therefore, step 3 multiplied by 3.7 gives the
litharge required for a bisilicate. *

5. Add additional litharge for the button. For convenience,
30 g. is a suitable quantity.

6. Add reducer for the button: Multiply the R.P. or O.P. of
the sample by the weight of sample in grams, to give the total
reducing or oxidizing effect of the sample used. If the sample is
reducing, subtract the total R.E. from the desired button weight
and divide by the R.P. of the reducing agent to be used, 1 in order
to get the weight of reducing agent required. If the sample is
oxidizing, add the total O.P. to the desired button weight, and
divide by the R.P. of the reducing agent.

7. Add from 3 to 5 g. of borax glass to the charge before mixing,
or approximately twice the amount of borax as a cover.

Example: Assume that a J^-A.T. (nearly 15-g.) portion of a
quartz-pyrite ore with a R.P. of 1.0 is to be assayed. Since the
amount of pyrite present is small (less than 10 per cent), the base

Step

Item

Calcu-
lation,
grams

Recommended
charge in round
numbers

Weight of ore (14.58 g.)
Na2COs equal to weight of ore

)4 A.T. ore
15 g. Na 2 CO s

Silica in ore
Bisilicate silica equivalent of Na2COs:

(Factor from slag table): (15) -? (1.8)

15.0
- 8.3

4
5

Silica to be fluxed with PbO

Litharge for bisilicate: (6. 7) (3. 7)
Add litharge for 25-g. lead button

6.7

24.7
26.9

Total litharge required

Desired weight of button
Total R.E. of ore: (15)(1.0)

51.6

25.0
-15.0

50-60 g. PbO

Net R.E. to be supplied
Flour required at R.P. 12: (10) -f- (12)

Borax glass (maximum used to counteract

10.0
0.8

1 g. flour
[
5 g. borax glass

basic effect of pyrite present)

1 Table XIII, p. 130.

138 FIRE ASSAYING

content of the ore may be ignored, and the charge calculated as
though the sample were of pure quartz. The charge calculation
is as shown on page 137.

CHARGE CALCULATION FOR BASIC ORES. It is evident that
samples which are dominantly basic yet deficient in R.P. will con-
tain relatively small proportions of sulfides. Into this category
fall the basic carbonates and sulfates, ores high in alumina, and the
oxidized ores of the heavy metals. If lime, magnesia, alumina,
or the higher oxides of iron arc the principal bases, bisilicate slags
should be used; if copper, nickel, or tellurium is present in
interfering amounts, subsilicates with a large excess of litharge
are required; and monosilicates would be preferred for most
other bases, including antimony, arsenic, ferrous iron, lead, and
zinc.

For basic ores requiring bisilicate slags, the following procedure
is used, which may be modified suitably for sub- or monosilicates.

1. Add sodium carbonate equal to the weight of ore.

2. Add litharge for the slag equal to the weight of ore. If the
ore is high in alumina, double the quantity of litharge.

3. Calculate the bases in the ore, other than alumina.

4. Calculate the bisilicate silica equivalent of the bases in
steps 1, 2, and 3. Deduct the silica in the ore and replace one-
third of the remaining silica with borax glass.

5. Add additional litharge for the button.

6. Add reducer for the button, calculating as in step 6 of the
charge calculation for siliceous ores.

7. If the ore is high in alumina, estimate the A1 2 3 content
and provide (if not already present) an equal weight of CaO,
or double the weight of CaC0 3 . Sodium or calcium fluoride
up to 1^ times the weight of A1 2 O 3 may be used in place of CaO.

Note : If impurities are present that require monosilicate slags, the amount
of litharge for the slag (step 2) should be double the weight of ore. If
copper, nickel, or tellurium is present, refer to Table XI for the mixing
of these impurities, and to the example given on page 143 of a charge calcu-
lation for copper sulfide ores.

Example: Assume that a given ore sample contains approxi-
mately 50 per cent hematite (Fe 2 O 3 ), 40 per cent limestone
(CaCOa), and 10 per cent quartz, and that J^ A.T. is to be used
for assay. The charge calculation is as follows:

THE CRUCIBLE ASSAY

139

Step

Item

Calcu-
lation,
grams

Recommended
charge in round
numbers

Weight of ore (14.58 g.)
NaiiCO 8 equal to weight of ore

15.0

l /4 A.T. ore
15 g. Na 2 CO 8

Litharge equal to weight of ore

15.0

Estimated weight of bases in ore:
Fe 2 3 : 0.5(15)
CaCO 3 : 0.4(15)

7.5
6.0

Bisilicate silica equivalent of bases (refer
to slag table) :
Silica for Na 2 CQ 3 : (15) + (1.8)
Silica for litharge: (15) -f- (3.7)
Silica for Fe 2 O 3 : (7.5) + (1.3)
Silica for CaCO 3 : (6.0) -=- (1.7)
Total silica for bases

8.3
4.1
5.8
3.5

21.7

Deduct silica in ore: 0.1(15)
Net silica required
% of required silica as SiOa
:Hj of required silica as borax glass:
(20.2)(1.7)

-1.5

15 g. SiO*
10 g. borax glass

45 g. PbO

20.2
13.5

11.5

15
26.9

3 n ' 5g -

Litharge for slag (step 2)
Add litharge for 25-g. button
Total litharge added

41.9

Desired weight of button
Estimated O.E. of ore:

25.0

7.5 g. Fe.O 3 at 1.0
Total R.E. required
Flour needed at R.P. 12:

7.5

32.5

(32.5) + (12)

2.7

3 g. flour

Step for alumina not required in this case

b. Ores with Excess Reducing Power

If the net reducing power of an ore sample is greater than
the reducing power required for producing a lead button of
suitable size, an added oxidizing agent is required in the con-
trolled reduction method, in order to prevent the reduction of

140 FIRE ASSAYING

more lead than is needed for the button. It is evident that
ores with deficient and with excess R.P. grade into each other.
In fact a ^-A.T. portion of a given ore may have insufficient
R.P., whereas a 1-A.T. portion of the same ore may have an
excess R.P. As the R.P. of ores increases, the amount of added
reducing agent decreases to zero. With further increase in
R.P. an oxidizing agent, usually niter, is added. In no instance
should both reducing and oxidizing agents be added to the
same charge, as this is not only a waste of reagents but com-
plicates unnecessarily the proper control of button size.

On account of the use of litharge in excess of button require-
ments, and the use of niter as an oxidizer, this method is fre-
quently known as the " litharge-niter " method.

Most of the ores of the excess-R.P. class are sulfides, and the
nature of the particular sulfides present influences the type of
charge to be chosen. For the entire group, except for nonsul-
fides containing organic matter, the slags should not be more acid
than a monosilicate, partly because of variable and uncertain
R.P. in the presence of more acid slags, and partly because of the
difficulty of fluxing some of the heavy metal oxides in highly acid
slags. Ores containing copper, nickel, or tellurium in important
amounts require subsilicate or oxide slags and are discussed
separately. Charges for ores containing organic matter should
be proportioned for the other constituents in the ore with an
adjustment of the oxidizing or reducing agent, to allow for the
reducing effect of the organic matter present in the ore.

CHAKGE CALCULATION FOR ORES WITH EXCESS REDUCING
POWER AND RELATIVELY FREE FROM COPPER, NICKEL, OR
TELLURIUM. The general method of calculating crucible charges
for sulfide ores requiring monosilicate slags is as follows :

1. Add sodium carbonate for the slag equal to the weight of
ore.

2. Add litharge for the slag equal to twice the weight of ore.

3. Calculate the bases in the ore.

4. Calculate the monosilicate silica equivalent of the bases
in steps 1, 2, and 3. Deduct the silica in the ore and replace
one-third of the remaining silica with borax glass.

5. Add additional litharge for the button.

6. Find the total reducing effect of the ore by multiplying the
R.P. (as estimated or as determined in a preliminary fusion) by

THE CRUCIBLE ASSAY

141

the weight of ore in grams. Subtract the desired button weight
from the reducing effect. Divide the result by the O.P. of niter,
usually taken at 4.0. This gives the amount of niter to be added.

7. Add additional sodium carbonate for the sulfate layer equal
to one-fourth of the weight of niter.

Example: Assume that V% A.T. of an ore containing approxi-

Step

Item

Calcu-
lation,
grams

Recommended
charge in round
numbers

Weight of ore (14 58 gO

2^ A T ore

Na 2 CO 3 for slag, equal to wt. of ore

15.0

Litharge for slag, 2 times wt. of ore

30.0

Weight of bases in ore:
FeS 2 : 0.4(15)
ZnS: 0.4(15)

6.0
6.0

Monosilicate silica equiv. of bases:
Silica for Na 2 CO 8 : (15) -f- (3.5)
Silica for litharge: (30) -i- (7.4)
Silica for FeS 2 (to FeO): (6.0) -f- (4.0)
Silica for ZnS (to ZnO): (6.0) -f- (3.2)

4.3
4.1
1.5
1.9

Total silica required
Deduct SiO 2 in ore: (0.2) (15)

11.8
-3.0

Net silica needed
% of silica as SiO 2
Borax glass replacing ^ of silica:

(8.8) (1.3) -f- (3)

8.8
5.9

3.8

6 g. SiOa
4 g. borax glass

Litharge for slag (step 2)
Litharge for 25-g. button

30.0
26.9

Total litharge needed

56.9

60 g. PbO

R.E. of ore: (15)(8.0)
Desired size of button

120.0
-25.0

Excess R.E. of ore
Grams of KNO 3 required: (95) -*- (4)

95.0

23.8

24 g. KNO 3

Na 2 CO 8 for slag (step 1)
Na 2 CO 8 for sulfate: (24 g. KNO 3 ) -5- (4)

15.0
6.0

Total Na 2 CO 8 to be added

21.0

20 g. Na 2 CO 8

142 FIRE ASSAYING

mately 40 per cent pyrite (FeS 2 ), 40 per cent sphalerite (ZnS),
and 20 per cent quartz is to be fluxed. By calculation the com-
posite R.P. of the ore is approximately 8.0, which value will be
used for the purpose of illustration. The charge calculation is as
shown on page 141.

CHARGE CALCULATION FOR ORES WITH EXCESS REDUCING POWER
AND RELATIVELY HIGH IN COPPER, NICKEL, OR TELLURIUM.-
Ores that are high in Cu, Ni ; or Te require subsilicate or oxide
slags and a high ratio of litharge to the interfering element.
To obtain buttons that are sufficiently pure to avoid difficulty
in cupellation the minimum ratio of litharge to copper should
be about 30 to 1 by weight. Nickel requires at least 60 parts of
litharge to 1 of nickel, and, when associated with copper in approx-
imately equal proportions, 100 parts of litharge to 1 part of com-
bined nickel and copper are required to avoid cupel scoriae and
retention of copper and nickel in the bead. The silicate degree
of slags for fluxing copper-bearing samples should not exceed
0.5 (subsilicate), and in difficult cases may be as low as 0.25.
For fluxing nickel-bearing materials the silicate degree may be
increased to 1.0, thereby avoiding excessive crucible corrosion,
but if much copper is also present a much lower silicate degree
should be used. Borax should be omitted entirely if much cop-
per is present and should be used sparingly, if at all, for nickel-
iferous materials. Increased sodium carbonate has no advantages
for fluxing either copper or nickel-bearing materials arid is best
held within the usual limits, that is, equal to the weight of sam-
ple, to minimize boiling and to leave room in the crucible for the
extra litharge.

The above considerations impose a practical maximum content
of nickel or copper that can be fluxed in a 20-g. crucible. Unless
a larger crucible is used, not more than about 7.5 g. of copper,
4 g. of nickel, or 3 g. o/ nickel and copper together should be
present in the charge.

Tellurium seldom occurs in ores in amounts greater than a few
hundredths of 1 per cent but may be present in metallurgical
products in greater proportion. 1 Reliable data are lacking on
the maximum limits of tellurium or selenium in the crucible assay,
but in general not over 0.1 to 0.2 g. of tellurium or selenium should

1 Ores sufficiently high in tellurium to cause assaying difficulty will give
a deep-red solution when heated gently in concentrated sulfuric acid.

THE CRUCIBLE ASSAY 143

be present in a crucible assay charge, and the ratio of litharge
to either of these elements should be at least 500 to 1. Increased
sodium carbonate is desirable. The silicate degree should be
between 0.5 and 1.0, and borax glass is useful. If trial cupel-
lations of buttons from fusions that employed the maximum
fluxes indicated the presence of tellurium in undesirable amounts,
buttons from subsequent fusions should be purified by soaking
under molten litharge for 10 or 15 min. In extreme cases a
combination method may be preferable.

The general basis for charge calculations for ores containing
copper, nickel, or tellurium is as follows :

1. Estimate the approximate percentage of copper, nickel, or
tellurium in the sample. Calculate the weight of sample to be
used, so that there will not be over 7.5 g. of copper, 4 g. of nickel,
3 g. of copper and nickel together, or from 0.1 to 0.2 g. of tellurium
present in the assay. Add sodium carbonate equal to the weight
of ore.

2. Add litharge for the slag equal to 30 times the weight of
copper, 60 times the weight of nickel, 100 times the weight of
nickel plus copper, and 500 times the weight of tellurium and
selenium.

3. Calculate the bases in the ore. Nickel or tellurium need
not be calculated, as the amount present can be omitted from
the slag calculation.

4. Calculate the subsilicate silica equivalent of the bases in
steps 1, 2, arid 3. Deduct the silica in the ore. Omit borax
glass for copper or nickel ores but replace one-third of the silica
by borax glass for telluride ores.

5. Add additional litharge for the button.

6. Calculate the niter (or reducing agent in the case of some
telluride ores) according to step 6 in previous examples.

7. Add additional sodium carbonate for the sulfate layer equal
to one-fourth of the weight of niter required.

Note: It is particularly important in the case of copper and nickel ores
to complete the fusion and pouring in the shortest possible time, not exceed-
ing 20 min. in a 20-g. crucible or 30 min. in a 30-g. crucible. Authorities
generally recommend slower fusions at lower finishing temperatures for
telluride ores.

Example for Copper Sulfide Ores: Assume that a given sample
of sulfide copper ore contains 80 per cent chalcopyrite (CuFeS 2 )

144

FIRE ASSAYING

and 20 per cent quartz, and that the precious-metal content is
so low that the maximum weight of ore must be used, up to %
A.T. Chalcopyrite contains 34.7 per cent copper, therefore the
ore contains 27.8 per cent Cu. The following calculations
apply:

Step

Item

Calcu-
lation,
grams

Recommended

charge in round

numbers

Wt. of sample to contain not over 7.5 g.

Cu.

7.5 g. X 100

- ~- -

rt _ ,

= 27 g., hence

A.T. (14.68 g.) should be used, which
will contain (14.68) (0.278) = 4.1 g. Cu.
NaaCO 3 for slag equal to wt. of ore

Litharge for slag, 30 times wt. of Cu:
30(4.1)

Wt. of bases in ore:
CuFeS*: 0.8(15)

Subsilicate equivalent of bases:
Silica for Na 2 CO 8 : (15) + (7.1)
Silica for litharge: (122) -r- (14.9)
Silica for CuFeS 2 to Cu 2 O.2FeO:

(12) -r (6.1)

Total silica required
Deduct SiO a in ore: 0.2(15)

Net SiO 2 to be added (omit borax
glass)

Litharge for slag (step 2)

Litharge for 25-g. button

Total litharge added

R.E. of ore: 12 g. CuFeS 2 at R.P. 8.0
Desired size of button

Excess R.E. of ore
Grams of KNO 8 required (71) ^4

NaaCOs for slag (step 1)
Na 2 CO 8 for sulfate: (18 g. KNO 3 ) -=- 4
Total Na a COa to be added

15.0

122.0

12.0

2.1

8.2

2.0

12.3
-3.0

9.3

122.0
26.9

148.9

96.0
-25.0

71.0
17.8

15.0
4.5

A.T. ore

19.5

9 g. SiO a

150 g. PbO

18 g. KNO 8

20 g. Na.CO 3

THE CRUCIBLE ASSAY 145

Note : A 30-g. crucible will be required, and the fusion should be completed
in 20 to 30 min. in a hot furnace.

2. Uncontrolled Reduction Methods. In the uncontrolled
reduction methods of crucible assaying the amount of added
litharge plus PbO derived from the ore is limited to that required
for the lead button, and an excess of reducing effect (R.E.) is
permitted or is deliberately provided, so that practically all the
lead is reduced to the button. Hence, soda (or potash) is the
only basic flux available for the formation of low-formation-
temperature slags.

The method has three chief applications: (1) to recover precious
metals from slags or cupels containing litharge; (2) to assay
certain classes of sulfide ores without the necessity for controlling
the reducing effect of the charge with niter or by other means;
and (3) to save litharge on quartz, galena, and other simple ores.

With respect to the first of these applications, many assayers
feel that it is necessary to reduce completely all litharge in assay
slags and cupels when making corrected assays. It is doubtful
if this practice is either necessary or desirable. The evidence at
present available indicates that, if litharge in an amount not
more than double the sodium carbonate is provided for the slag,
the precious-metal recovery is entirely satisfactory from such
materials and particularly in the case of the assay of cupels
it is much less difficult to ensure adequate fluxing when litharge
is permitted as a slag constituent.

The uncontrolled reduction method of assaying sulfide ores
by the soda-iron method had considerable vogue in the past but
is seldom used in modern assay practice. The method consists
of a reducing fusion in an alkaline flux, with an excess of metallic
iron in the form of large nails or spikes as a reducing and desul-
furizing agent. The sulfides react with the PbO present and with
the iron nails. After all the lead in the charge is reduced to
metal, the reactions with the remaining sulfides and iron produce
matte, principally ferrous sulfide (FeS), which is suspended in
the alkaline slag.

The soda-iron method is unsuitable for ores containing appre-
ciable amounts of metals near or below lead in the electromotive
force series, as these will be reduced with the lead and con-
taminate the button. Hence, ores containing antimony, arsenic,
bismuth, cobalt, copper, nickel, tellurium, or tin should not be

146 FIRE ASSAYING

assayed by this method; on the contrary, only those ores in which
the principal sulfides are those of lead, iron, or zinc are amenable
to the method.

Silver losses in the slag are higher with the soda-iron method
than with the niter method, probably owing to the increased
solubility of silver in the matte.

Best results are obtained with subsilicate slags in which
approximately 75 per cent of the silica is replaced by borax glass.
In no case should the slag be more acid than a monosilicate.
The buttons should be larger than is customary with other
methods, up to 35 g. for ores high in pyrite. The charge cannot
be calculated accurately in advance of fusion, as a variable
amount of iron is gained from the iron nails; consequently, the
fluxes are estimated empirically.

Procedure. The following procedure is recommended for the
soda-iron method:

1. For heavy sulfides, the ore charge should not exceed }
A.T. If the principal sulfide is galena, use twice as much sodium
carbonate as ore; if large amounts of pyrite or sphalerite are
present, use .three times as much sodium carbonate as ore.

2. Add one-fourth as much borax glass as sodium carbonate.

3. Add one-fifth as much silica as borax glass, unless the ore
contains sufficient silica.

4. Allow for lead in the ore and add sufficient litharge so that
the total lead in the charge is from 30 to 35 g. (The factor
Pb:PbS is 0.865, and the factor Pb:PbO is 0.928.)

5. Mix the charge in the crucible, and insert, point downward,
four 20-penny nails or one 3^-in. track spike.

6. Heat gradually in a reducing atmosphere and finish in not
less than 30 min. Slow heating aids sulfur elimination.

7. Fusion is complete when the nails can be freed from lead
by gentle tapping and washing in the slag. Remove the nails
and pour the fusion as usual.

B. Two-stage Processes

Two-stage processes of assaying are those in which lead reduc-
tion and precious-metal collection are effected in a separate
stage following a preliminary treatment to destroy the reducing
effect of the ore or to remove impurities that are difficult to

THE CRUCIBLE ASSAY 147

flux. The preliminary treatment may take place in the fusion
crucible or may involve a separate roasting or acid treatment.

In all the two-stage processes the actual lead reduction and
metal collection are accomplished by controlled reduction with
added reducing agents. For this reason litharge is available as a
flux, and the principles of slag formation are the same as those
already discussed for the single-stage controlled reduction
processes.

3. Oxidation-collection Method. The oxidation-collection
method 1 eliminates the necessity for niter estimation in the assay
of sulfide ores yet obtains buttons of nearly uniform siz. Com-
plete oxidation of sulfur and other reducing agents is obtained
during fusion by adding an excess of niter. After oxidation is
complete, briquettes of iron filings and litharge are dropped into
the crucible, in order to effect lead reduction and the consequent
collection of gold and silver.

From the reaction between metallic iron and litharge,

Fe + PbO = FeO + Pb

the theoretical R.P. of iron is 3.7; hence 6.75 g. of iron are
required for a 25-g. button. The actual R.P. is slightly higher
than the theoretical R.P., probably owing to the formation of
some Fe2Os: Consequently, not more than 6.0 g. of pure iron
need be used for 25-g. buttons. Commercial grades of iron
filings are of cast iron, which contains up to 4.5 per cent carbon
and other impurities that increase the R.P. of the iron so that
only 5 to 5.5 g. may be required instead of 6 g.

Present data on the method indicate that the litharge for the
button should be added with the iron, although further investiga-
tion may prove this to be unnecessary. With some types of
ores it is necessary to use two stages of collection in order to
ensure complete recovery of gold and silver. It is convenient to
use briquettes containing 2.5 g. of iron filings and 15 g. of litharge,
bonded with sodium silicate.

For the oxidation-collection method the same slags may bo
made as in the single-stage litharge-niter controlled reduction
method. If litharge for the button is to be added with the

1 SHEPARD, O. C., and DIETRICH, W. F., The Oxidation-collection Method
of Assaying Sulphide Ores for Gold and Silver, A.I.M.E., Tech. Pub. 997,
1938.

148 FIRE ASSAYING

collector the amount added to the charge is decreased accord-
ingly. For J4-A.T. charges, 30 g. of niter is sufficient for com-
plete or nearly complete oxidation of heavy sulfide ores. In
most cases, 25 g. is sufficient. With pure pyrite a small lead
button is formed with 30 g. of niter, but the standard amount
of collector may be used without increasing the size of button
beyond acceptable limits. An excess of niter does no harm, as
it is decomposed by heat according to the reaction:

2KNO 3 + heat = K 2 O + N 2 + 20 2

and the K 2 formed enters the slag as a desirable base. For
economy, and to minimize boiling, less than 25 or 30 g. of niter
may be used for J^ A.T. of ore if it is known that the reducing
power of the ore is less than 8.

In view of the compensating effect of K 2 derived from that
portion of the niter which is not used in oxidizing reactions to
form potassium sulfate, it is possible to prepare a standardized
charge for J^-A.T. samples, which, although wasteful of fluxes
in some cases, can be applied to practically any combination
of sulfides and siliceous gangue. With the following charge :

Ore M A.T.

Sodium carbonate 30 g.

Litharge 30 g.

Silica 10 g.

Borax glass 5 g.

Niter 30 g.

the calculated silicate degree of the slag is 0.70 for pure galena,
0.73 for chalcocite, 0.79 for sphalerite, 0.99 for pyrite, and 1.34
for pure quartz, all of which are satisfactory.

It is evident that if the character of the ore is known in advance,
considerable saving in fluxes may be made and better slags can be
produced if the fluxes are modified to favor the substances that
are difficult to flux, so that in so far as is practicable the sugges-
tions given in the section on fluxing should be applied to specifi-
cally known substances. For example, for best results with a
high copper sulfide ore, an additional 30 g. of litharge should be
added to the above standardized charge, and it is desirable to
decrease the silica to 5 or 8 g. and the borax glass to 3 g. or less.
Pure quartz or other neutral or oxidizing ores, if their character

THE CRUCIBLE ASSAY 149

is recognized in advance, would be assayed by the regular flour
reduction method. Ores high in lime or magnesia will not be
satisfactorily fluxed with the standard charge, as the silicate
degree of the slag would be too low.

In order to minimize boiling and for convenience in preparing
charges, large batches of the flux mixture, without niter, may be
mixed and fritted in a suitable furnace, such as a bullion melting
furnace.

General Procedure for Oxidation-collection Method. After pre-
paring the charge as described above, mixing and placing it in
the crucible, heat for at least 15 min. in a hot furnace. Then add
one collector briquette containing 2.5 g. of iron and 15 g. of
litharge. After not less than 1 min., add a second collector
briquette. Continue heating for at least 3 min. and pour as
usual.

At the time of adding the first collector briquette, most of the
oxidation reactions should be completed, although the charge
does not need to be thoroughly liquid. Before pouring, the fusion
should be quiet, and the slag fluid. If the temperature is suffi-
ciently high and the flux charge is approximately correct the
time factors given above will be sufficient. Shotting of the lead
and other defective fusions result from the same causes as in the
regular litharge-niter method and are corrected in the same
manner.

4. Roasting Method. The roasting method consists of heating
a weighed amount of the material to be assayed in a shallow fire-
clay dish in the presence of air in order to oxidize the metals
present and to eliminate certain impurities.

The impurities that can be eliminated by this oxidizing roast
are carbon, sulfur, arsenic, antimony, and tellurium. After
roasting, the roasted product (calcine) is treated in a crucible
fusion in the same manner as an ore with no reducing power.

Roasting is a slow method of eliminating impurities, and there
is considerable danger of mechanical or volatilization loss.
Except for low-grade materials consisting mostly of carbon or
hydrocarbon, such as charcoal cyanide precipitate, the roasting
method has little to recommend its use. Other impurities can
be handled either in the niter assay or, preferably, because they
are better eliminated, by acid treatment. Should 2-assay-ton
or larger fusions be desired for arsenical or antimonial sulfide ores,

150 FIRE ASSAYING

without using acid treatment, the roasting method might prove
useful.

Procedure. Weigh out the ore and spread in a thin layer in a
roasting dish that has been previously coated with chalk or
silica. (Materials such as a charcoal cyanide precipitate, that
leave a rich residue, should be diluted with an equal weight of
silica before roasting.) Place the dish in a muffle at a barely
perceptible dull-red heat. Keep covered until danger of decrepi-
tation has passed. Gradually increase the temperature until
fumes appear from the ore, then hold at that temperature with
frequent stirring until reaction has nearly ceased. Gradually
raise the temperature to a maximum of 650 to 700C., stirring
at 20-min. intervals, and hold at the final temperature for
30 min.

Sulfur, carbon, or tellurium can be eliminated in a straight
oxidizing roast with an oxidizing atmosphere at the roasting
dish. The only precautions are to avoid loss by violent reactions
and to avoid fusion, which prevents complete reaction and may
cause the material to stick to the dish. If at any stage of roasting
the charge shows signs of becoming sticky from incipient fusion
it should be removed from the furnace and mixed with silica.
This is particularly necessary with galena ores, and silica, equal
to the weight of the ore, should preferably be added at the outset.
Sphalerite will require a higher roasting temperature than other
sulfides, and a longer roasting time, but the maximum roasting
temperature should not greatly exce*ed 700 C. in any case.

In order to secure the removal of arsenic and antimony, roast-
ing must be modified to avoid the formation of arsenates and
antimonates, which may form with strongly oxidizing conditions.
The oxides of antimony boil above the range of roasting tempera-
tures, so that reducing conditions must be maintained in order
to drive off the volatile Sb2S 3 before it is oxidized. Strongly
reducing conditions should be maintained by adding sulfur to the
ore and by heating in a reducing atmosphere. After sufficient
antimony has been removed, finish with an oxidizing roast to
remove carbon and sulfur. Arsenie is volatile as the sulfide and
as the oxide As 2 0s. It is best removed by alternate oxidation
and reduction with sulfur. Finish with an oxidizing roast.

An oxidizing roast converts most metals remaining in the cal-
cine to their higher oxides. Thus the calcine from thoroughly

THE CRUCIBLE ASSAY 151

roasted pyritic ores will contain a high proportion of ferric oxide,
the oxidizing power of which must be considered in the crucible
fusion.

6. Acid -treatment Method. The acid-treatment method con-
sists of leaching a weighed amount of the material to be assayed
with acids and other reagents, in order to dissolve impurities
that cause difficulty in assaying. This process is commonly
used for the assay of high copper materials such as blister and
refined copper, but it can be used for the separation of any acid-
soluble impurity. In most cases, acid combinations that might
dissolve gold are avoided. If silver dissolves, it is precipitated
as the chloride just before separation of the insoluble by filtration.
The precious metals in the precipitate and insoluble are recovered
by scorification, if the total insoluble is small; or by a crucible
assay if it is large. Detailed procedures for the acid-treatment
methods are given in Chap. X.

MANIPULATIVE PROCEDURES IN CRUCIBLE ASSAYING

In routine assaying practice it is important that the assayer
strive for a high degree of efficiency so that the maximum number
of reliable assays may be completed in the minimum time. Fire
assaying is particularly amenable to systematization and the
use of labor-saving devices where the volume of work is suffi-
ciently great to justify the expense of the special equipment.
The following paragraphs illustrate current manipulative
practices.

Crucibles. Most assay crucibles are made of fire clay. Cruci-
ble sizes are now generally designated in grams, which is an
approximate index of the number of grams of a simple quartz
ore that can be fluxed in the crucible. For example, a 15-g.
crucible will hold the charge for K A.T. (nearly 15 g.) of an easily
fluxed ore, but in most cases a 20-g. crucible would actually be
required.

In any given laboratory it is desirable to use the minimum
number of different crucible sizes. In general, to avoid boiling
over, crucibles should not be more than two-thirds to three-
fourths full before melting. For reducing-power determinations
and for 0.1- to 0.2-A.T. ore charges, 10-g. crucibles are suitable.
For most charges of J^ A.T., 20-g. crucibles are used. For
larger charges, 30-g. crucibles are required; for fluxing 2 A.T.

152 FIRE ASSAYING

of ores requiring more than the minimum fluxes it may be neces-
sary to use 40-g. crucibles.

Fire-clay crucibles larger than 40 g. are impracticable for fire
assaying because the severe thermal shock to which assay
crucibles are subjected may cause failure of large crucibles, even
during the first fusion. Dixon sand crucibles are satisfactory
where larger crucibles are needed. When assaying low-grade
tailings, it is sometimes necessary to use 3 or 4 A.T. of the sample.
If it is not desired to use large crucibles, two separate fusions
of 1% or 2 A.T. each can be made, combining the buttons by scori-
fication, by pouring them into a single mold, or by cupeling them
together in a cupel of suitable size.

Crucibles should be kept dry in order to prevent cracking when
placed in the furnace. In moist atmospheres it is advisable to
store crucibles before use on a drying rack located close to the
furnace.

Both fire-clay and sand crucibles are acid refractories and are
quite resistant to corrosion by highly acid slags, such as the sesqui-
or higher silicates. With such slags a crucible may be reused
a number of times, until it finally cracks or becomes salted.
Even with very basic slags the crucible wall is usually thick
enough for more than one use. Basic refractories, such as
magnesite, would be very suitable for highly basic slags but aro
thus far not available at a cost to compete with fire-clay crucibles.

A crucible that shows serious corrosion or a deep crack should
never be reused because the loss in time, materials, and possible
damage to the furnace caused by crucible failure greatly exceeds
the cost of the crucible. Precious metals not recovered in a
crucible fusion are partly left in the crucible, where they may be
picked up to salt a subsequent charge. At some laboratories
every crucible is discarded after a single use, thus avoiding the
possibility of this source of salting.

When crucibles are to be reused, care should be taken to
segregate those in which rich samples have been fused. These
crucibles should be discarded or used only for the assay of simi-
larly rich materials. Every crucible that has contained a viscoxis,
lumpy, or shotted slag, or a fusion in which a small lead button
was obtained, should be discarded.

Assay Sequence. It is of extreme importance that the assayer
devise, and strictly adhere to, a good system of arranging and

THE CRUCIBLE ASSAY 153

handling the daily assays in a predetermined order throughout
the entire process from the time the samples are received until the
final report is made. Errors caused by misplacement of a sample,
crucible, cupel, or parting cup are undetectable by ordinary
means, with the result that high-grade samples might be reported
in place of low-grade samples, and vice versa. To avoid such
errors every assayer should follow a system by which the order
of samples i^ automatically maintained. If one man carries
through all operations himself, his system may be devised to suit
his own peculiarities; but, if helpers are employed, the system
must be such that neither the assayer nor the helpers can confuse
the sequence of assays without detection.

It is impracticable to use permanent marks on crucibles,
because this practice is time consuming and becomes confusing
when crucibles are reused. Cupels cannot be marked satis-
factorily. Even if crucibles and cupels were marked, this
would be no guaranty in itself that the buttons would be placed
in the proper cupel. Parting cups can be marked permanently,
but time is lost in arranging them in the proper sequence. Thus,
a satisfactory routine should be developed that is independent of
numbered marks on the assay receptacles.

When the day's set of pulverized and mixed samples has been
assembled, the assayer arranges them in suitable order, then
records the sample numbers or description on his report form or
notebook. From then on, until the final report is prepared,
the assays are handled in the same sequence and may be assigned
serial numbers from 1 upward.

Ruled lines may be provided in the notebook to separate
each furnace load of crucibles, each set of cupellations, and each
tray of parting cups. The number of assays in each of these
units is usually different, but if special parting cup trays are
made they should hold the same number of parting cups and be
in the same arrangement as a set of cupels in the furnace.

154

FIRE ASSAYING

ver and gold content in a key position in each furnace set. If
this assay appears in its proper place upon weighing the dore*
beads it is reasonable assurance that no misplacements have
been made in the previous operations. For controlling the fusion
and cupellation sequence a small quantity of some element may
be added to one or more of the assays, which will give a stain on
the cupel that is not present in any of the other assays. Copper
or nickel is suggested for this purpose and may be added to any

of the regular assays if the
amount is just sufficient to give a
distinctive cupel stain but not
enough to cause cupellation loss
or freezing of the button.

Handling Crucibles. A carry-
ing box or tray should be con-
structed with the same floor area
as the rmiffle, to hold a complete
set of crucibles for a furnace.
A typical crucible tray for a
muffle furnace, which holds five
rows of three 20- or 30-g. cruci-
bles, is illustrated in Fig. 12.
The crucibles are arranged on the
tray as shown and carried to the
fluxing bench. When ready for
the furnace, the tray is carried
to a bench alongside the furnace,

Front - open

I Inside width of muff le \
r* *1

Make of %'or s / f $ plywood
FIG. 12. Crucible tray.

and the crucibles are transferred to the furnace by placing
the front row (13-14-15, Fig. 12) of the tray in the rear of the
muffle. When the furnace is loaded, the crucibles will be in
reverse order, front to back, but in the same order left to right,
compared with the original position on the crucible tray. In
pouring from left to right in each successive row, the No. 1
crucible will be poured first, and the rest in numerical sequence.
Crucible Tongs. For moving crucibles into and out of a pot
furnace, tongs are used that grip a point on the upper rim of the
crucible. For muffle furnaces the commonest form of tongs con-
sists of a U-shaped fork on the lower arm that partly encircles
the crucible near its mid-height, and a slightly longer tongue
on the upper arm that is held down on top of the crucible.

THE CRUCIBLE ASSAY* 155

Multiple forks as illustrated in Fig. 20, Chap XVI, permit more
rapid handling of the crucibles than the usual type of implement.
These are not on the market at present but can easily be con-
structed. The part holding the crucible is made by tapering the
inside of an iron pipe of suitable size until it closely fits the cruci-
ble of given size about half-way up from its base. Then a ring
M in. long is sawed from the pipe, and enough of the ring is
sawed out to give an open end that will slip around the base of
the crucible.

To make a pouring tongs, one of these rings is brazed to a
Jg-in. pipe. A short crossarm may be brazed or fitted to the
other end of the pipe, to form a handle. In pouring, the edge
of the crucible is rested on the edge of the mold, which prevents
the crucible from falling out of the tongs when inverted. The
charging tongs are built to hold as many crucibles as will fit in a
row across the muffle, by brazing the desired number of tapered
rings to a reinforcing bar and to spreader blocks. At the Selby
Smelter, of the American Smelting and Refining Company, six
crucibles are charged into the furnace at a time by this implement.

Addition of Fluxes. In most commercial assay offices it is
common practice to prepare stock flux batches containing all
the principal ingredients for fluxing ores of a given general type.
By using an excess of fluxes on all samples it is possible to prepare
a flux that will serve for practically all purposes, requiring only
the addition of a reducer or oxidizer in the controlled reduction
method, and minor additions of special reagents for difficult ores.
Some assayers prepare one flux for nonreducing silicate ores in
which the required reducer is incorporated, and another flux
for sulfide ores and concentrates to which niter is added sepa-
rately for each assay. Assayers having a daily routine of mine
samples, concentrates, and tailings may prepare all-inclusive
stock fluxes for each class of material.

Flux batches may be conveniently prepared by weighing out
the various ingredients and mixing them in a suitable mixing
box. A simple form of mixer is a small barrel with lifters on the
inner surface. After weighing out the fluxes the barrel is closed
and is rolled back and forth until the contents are thoroughly
mixed.

Whether or not stock fluxes or individual fluxes are used it is
entirely unnecessary to weigh out all the charge components for

156 FIRE ASSAYING

each assay. Except for the ore and niter, a volume measurement
gives sufficient accuracy. Even the weighing of niter is elimi-
nated in the oxidation-collection method. Although special
measuring devices can be made for any requirements it is possible
to assemble and calibrate a set of measuring receptacles by using,
for example, half-teaspoons, teaspoons, tablespoons, parting
cups, 35- and 50-ml. iron crucibles, scorifying dishes, assay cruci-
bles either whole or sawed off, and various other articles that arc
easily obtainable. The smaller measures may be used "level"
or "heaping," as required. For example, a level half -teaspoon
of flour is close to 1 g. ; a level teaspoon is nearly 2 g. The usual
amount desired, 2% g., is easily obtained by scraping level about
% of the top from a heaping teaspoon.

If batch fluxes are used, it is convenient to store them in bins,
barrels, or boxes at a bench where the crucibles are assembled
on the trays. The assayer or his assistant adds a measure of the
stock flux to each crucible, then carries the tray to the fluxing
bench where the ore and additional reagents are added, and the
charges are mixed.

At the fluxing bench a supply of all individual fluxes should be
kept in a partitioned flux box, which is provided with a lid
that can be closed while ore samples are being weighed out and
the charges are being mixed. This precaution is advisable to
prevent accidental salting of the fluxes. A metal strip mounted
across the top of the compartments is a useful scraper for produc-
ing quickly leveled spoons of flux. In the controlled reduction
method with niter the niter is weighed out to the nearest gram
on a pulp balance. If desired, a set of 10-, 5-, 2-, and 1-g. meas-
ures can be used for niter, but their use will usually take more
time than weighing.

Open-type pulp balances, accurate to 5 mg., are suitable for
weighing out ore samples, except for very small charges (0.1
or 0.05 assay ton) of high-grade ores, in which case an analytical
balance is preferable.

When preparing crucible charges it is customary to add a part
or all of the fluxes before adding the ore, so that there will be less
danger of particles of ore adhering to the crucible walls and
possibly escaping the action of the fluxes. If the charges are to
be mixed in the crucible, and if fluxes are added separately, the
lighter fluxes are added first, and the heavier ones are placed on top.

THE CRUCIBLE ASSAY 157

Mixing. After the ore and fluxes are placed in the crucible,
they are thoroughly mixed with a spatula, in a cocktail shaker,
or on a mixing cloth.

Mixing with a spatula is the commonest procedure. The
spatula is inserted to the bottom of the crucible along the inside
surface and then lifted with a motion that crowds the spatula tip
toward the opposite side of the crucible, the blade finally being
turned on edge at the completion of the stroke as the blade
emerges from the charge. The crucible is rotated a few degrees
with the holding hand, and the spatula stroke repeated The
process is continued until the batch is homogeneous.

A 1-pt. aluminum cocktail shaker can be transformed into a
suitable mixer by sealing the pouring cap (and sieve if any) from
the inside so as to present a smooth interior surface to the charge
batch. From 12 to 16 vigorous shakes are sufficient for mixing
a crucible charge. There is some danger of salting if the charge
is not tapped out carefully, but results are generally satisfactory,
and the process is considerably faster than any other method of
mixing charges.

Mixing the charge on a mixing cloth is slow and cumbersome
and, unless carefully performed, may introduce mechanical
losses.

Furnace Operations. Although slow firing is necessary for
the soda-iron method, best results will be obtained with other
methods by charging the crucibles into a furnace or muffle
that is maintained at a finishing heat, which ranges from 1050
to 1150C. Upon loading the furnaces with cold crucibles the
furnace temperature will drop but will soon increase and, in a
properly fired furnace, the finishing temperature will again be
reached in from 10 to 15 min. In general, if the fluxes are cor-
rectly proportioned, assays in 10-g. crucibles should be ready to
pour in 15 min.; 20-g. crucibles require 20 min., and 30-g. cruci-
bles require from 25 to 30 min. Excessive temperatures or pro-
longed heating cycles should be avoided, as such practices
decrease crucible life and tend to cause abnormalities in button
size and purity. A nearly neutral atmosphere during fusion
also helps to prevent undue variation in the size of the lead
button. This is accomplished by avoiding a large excess of air
in the combustion chamber of a pot furnace, or by closing the
muffle draft in a muffle furnace. If necessary a large piece of coke

158 FIRE ASSAYING

may be used in the front of a muffle furnace, or crucible covers
may be used in a pot furnace.

For the soda-iron method the furnace temperature should be
below 600 to 700C. immediately after charging, then should
increase to a maximum of approximately 1050C. in 30 to
45 min.

It is true that there is an optimum firing cycle for each distinc-
tive type of crucible charge, but in practice it is not convenient
to treat each crucible as an individual, and the above suggestions
provide a standardized firing procedure that is generally satis-
factory. Charges that are known in advance to require higher
melting temperatures than others in the same set may be favored
by arranging these so that they will be placed in the hottest part
of the furnace.

Pouring. Before pouring, a crucible may be removed from the
furnace and inspected to see that fusion is complete, as shown by
the homogeneity and fluidity of the slag, and the absence of
floating pellets of lead or of scoriae of undecomposed ore.

Completed fusions are usually poured into conical iron molds.
A few assayers prefer to pour on a flat iron plate because the
lead cools rapidly and is more easily separated from slag. When
molds are used, it is customary to coat them with chalk or ruddle
and to warm them slightly before using. Molds must be dry,
but, unless warming is required for drying, heating the molds is
unnecessary. The chalk or ruddle coating is used to prevent the
lead button and slag from sticking in the mold. If the inside
surface of the mold is rusted, additional coating is not required.
Some assayers deliberately rust their molds with a salt or acid
solution. Such a coating lasts much longer than chalk or ruddle.

Just before pouring a crucible, many assayers swirl the con-
tents slightly and gently tap the crucible down against a solid
surface. This is supposed to aid in collecting all lead in a single
pool at the bottom of the crucible. Such an operation may have
some beneficial effect in rare instances, but in most cases equal
results are obtained if it is omitted.

Crucibles should be poured in a smooth unbroken motion, turn-
ing them upside down and entirely over in a complete revolution.
With this practice, the last slag drop runs back inside the crucible
instead of down the outside. Slag on the outside of crucibles is
a nuisance when they are reused. As the crucible is tipped, the

THE CRUCIBLE ASSAY 159

slag that floats on top of the lead pours out first. The stream
of slag should be directed at a point in the mold slightly off-center
toward the crucible so that the lead with its greater trajectory will
fall as nearly as possible in the center of the mold. A skilled
assayer pours the slag very rapidly and slows the motion slightly
just before and during the time the lead pours. Assay charges
larger than 1 assay ton usually have more slag than the mold will
hold. No harm is done if the mold is filled to overflowing with
slag before the lead starts to pour, but this can be avoided by
pouring part of the slag to one side before pouring the rest of the
fusion into the mold.

If two buttons are to be combined together in a single mold, a
part of the slag from each crucible should be poured off into a
separate mold or to one side, in order to provide room in the but-
ton mold for both buttons and a part of both slags. The second
pour must follow quickly after the first, before the slag has become
so viscous as to prevent the second button from settling.

After pouring into molds it is necessary to wait until certain
that the lead has solidified bef ore attempting to recover the but-
tons. Acid slags decrepitate on cooling, and with most types of
molds, unless originally overheated, the lead will have solidified
by the time the decrepitation of the slag begins.

Most of the slag can usually be separated from the button by
tilting the block of slag and lead on edge, so that the tip of the
button rests against the upper edge of the mold, and then striking
a sharp blow with the peen of a machinist's hammer at the
junction of the slag and lead. If the slag adheres to the mold,
it is necessary to shatter it with the peen until the lead is exposed,
then pry out the lead with a sharp-pointed tool.

To complete the removal of slag from the button, and to
facilitate charging into the cupel, pound it into an approximate
cube and, if necessary, brush off adhered slag with a stiff brush.

Figure 13 illustrates a type of button mold that is favored by
some assayers to obviate the necessity of separating slag from the
button and hammering the button into shape for cupellation. 1
Each mold is an inverted truncated cone and is large enough to
hold 50 to 55 g. of lead. As many molds as desired may be drilled
in a bar of suitable length of 1-in. square iron or steel, leaving

1 Details kindly supplied by H. H. Bern, assayer, Lava Cap Gold Mining
Corporation, Grass Valley, California.

160

FIRE ASSAYING

about 2 in. at one end as a handle. When pouring a crucible its
lip is held just over the near edge of the mold, and a slow steady
pouring motion is used so that the slag runs down the side of the
crucible and collects on an iron table top. When the lead appears,
the higher trajectory of the molten lead carries it into the button
mold with little or no slag. Buttons are ready for cupellation
immediately upon removal from the mold. Although the pour-
ing operation is slightly slower than when regular molds are used,
this is more than offset by the time saved in preparing buttons
for cupellation. The contamination of the exterior of the crucible
with slag is a disadvantage and tends to shorten the life of
crucibles.

Note : With dimensions as above, each mold holds 50-55 g. of lead

Fio. 13. Longitudinal cross-section of button mold.

Pouring on a flat plate is preferred by some assayers to pouring
into molds. The plate should be of steel of sufficient thickness
to avoid buckling and of large enough area to permit all the
charges from a given furnace load to be poured in a continuous
sequence. A plate % in. or J^ in. thick and 28 by 36 in. is suitable
for pouring 12 assays. Preheating of the plate is unnecessary if
it is dry. In pouring, the stream of slag is directed on the plate
so as to form an elongated pool. When the lead appears, a slight
deflection of the stream will produce a bay into which the lead
can be poured in a compact globule, after which the remaining
slag is poured in a continuation of the original pool. With
practice this method is as rapid as mold pouring and saves much
time in the subsequent cooling of the buttons and separation from
slag. It is particularly convenient for reducing-power deter-
minations, where at least 15 min. can be saved in obtaining a set
of buttons to be weighed.

THE CRUCIBLE ASSAY 161

Trouble Shooting in Crucible Assaying. When imperfect
fusions, outsize or impure buttons, or other irregularities appear
in crucible assaying the assay er is concerned with the correction
of the defect when the repeat determination is made. The
diagnosis of the source of trouble may sometimes require a more
careful identification of interfering minerals than was made
prior to the original assay. Frequently the cause of the diffi-
culty can be detected from some peculiarity exhibited by the
slag or button from the first fusion, or by the appearance of the
cupel and bead during and after cupellation.

The distinction between acid and basic slags can be made by
observing the viscosity of the slag as it cools, and by observing
the appearance of the cooled slag. If an iron rod, or crucible
tongs, is dipped into the molten slag and drawn slowly away from
it, acid slags may be drawn out in thin threads, whereas basic
slags will drip from the end of the rod. During cooling in the
mold, acid slags tend to decrepitate with considerable violence
as room temperature is approached. Fresh fractures of acid
slags are vitreous, whereas basic slags are stony or crystalline.

With respect to slag colors, only the copper green or red, the
cobalt blue, the manganese purple, and, in some cases, the anti-
mony yellow, are sufficiently distinctive to be certain criteria
of the presence of specific elements in the slags. 1 Other colors
are less unique or are masked or altered by litharge or by iron,
which alone imparts a wide range of colors including yellow-brown,
brown, black, and yellow-green.

One of the commonest sources of error with beginners is the
failure to estimate correctly the required amount of fluxes and
oxidizing or reducing agents, and the failure actually to include
the desired reagents in the charge. Before applying other
remedies, the assayer should be certain that the difficulty is not
due to a personal error of this type.

Table XV is presented as a guide to crucible assay difficulties
and their correction. Good judgment is required to decide to
just what degree a given defect may be permitted without
influencing the final result. The term "excessive viscosity "
implies that the slag is so viscous that it will not pour readily
from the crucible, and that lead will not settle through it. If
the slag is homogeneous, a relatively high viscosity may be

1 See slag colors on p. 123.

162 FIRE ASSAYING

TABLE XV. CRUCIBLE ASSAY ABNORMALITIES AND THEIR CORRECTION

Type of defect

Probable cause

Suggested remedy

Excessive viscosity

1. Insufficient finishing tem-
perature

2. Excess of acids (slags are
vitreous)

3. Insufficient total flux for
difficult ores, especially
for CaO, MgO, and
AhOs

1. Increase furnace tempera-
ture, sometimes requir-
ing longer fusion time
2. Be sure that slag is not
more acid than a bisili-
cate. Decrease acids or
increase bases
3. Verify presence of interfer-
ing substance, and adjust
charge accordingly

Undecomposed or
insoluble slag com-
ponents

4. Insufficient fusion time

5. Deficiency of acid fluxes,
especially with CaO,
MgO, FezOs, FeO4, and
AhOs (slags are nonvitre-
ous or crystalline)
6. Lack of suitable flux for a
specific impurity, as
AlaOs, bone ash, and
others

4. Heat until visible action
ceases
5. Add additional acid fluxes
to ensure sesqui- or bisili-
cate slag

6. Identify the impurity and
provide required flux

Shotting

7. Excessive slag viscosity
8. Unfused particles of FesO4
in a mushy layer between
slag and lead, (may be
derived from sulfides as
well as from oxides of Fe)
9. Spattering of lead in mold
caused by pouring before
fusion reactions are com-
pleted

7. See 1, 2, and 3
8a. Repeat with bisilicate slag,
or
86. Add approximately 1 A.T.
PbO to crucible before
pouring.
9. Leave in furnace until
fusion is quiet

Excessive adhesion
of slag to button

10. Slag too acid, generally
with borax glass

10. If other conditions permit,
decrease the acid fluxes
or increase basic fluxes

Excessive crucible
corrosion or break-
age

11. Excessively basic slag
12. Furnace temperature too
high
13. Poor crucible quality

14. Crucibles contain moisture
15. Slag on furnace floor

11. Increase acid fluxes
12. Decrease furnace tempera-
ture
13. Try other crucible brands
under comparable condi-
tions
14. Dry crucibles before using
15. Keep furnace floor free from
fusible slags. Use bone
ash to absorb spills of
slag, lead, or fluxes

Matte or speiss
layer

16. Insufficient PbO or Na 2 CO 3
or both

16. Increase PbO or NaaCOs or
decrease weight of sam-
ple

THE CRUCIBLE ASSAY

163

TABLE XV. CRUCIBLE ASSAY ABNORMALITIES AND THEIR CORRECTION.

(Continued)

Type of defect

Hard or brittle

Outsize

Probable cause

17. Base metals, such as Cu,
A.S, or Sb, or matte in
button not caused by in-
correct furnace opera-
tion

18. PbO in button

19. Rejection of base metal,

especially copper, from
slag by change in com-
position by crucible
attack

20. Very high gold in button

21. Erroneous estimate or

weighing of reducer or
oxidizer

22. Insufficient PbO for button

23. Excessive oxidizing (small

button) or reducing
(large button) atmos-
phere in furnace

24. Slag much more acid than

in reducing-power deter-
mination (small button
from niter assay charge)

Suggested remedy

17. a. Increased PbO and
Na 2 CO 3 usually re-
quired, or charge re-
proportioned for spe-
cific impurity, or

17. 6. Repeat with smaller ore
charge, or

17. c. Treat slag as in 86.

18. Seldom serious, but usually

corrected by higher tem-
per*ature, or decreased
PbO in slag

19. Use a quick, hot fusion

cycle

20. Decrease weight of ore or
increase size of button or
both

21. Recalculate reducer or oxi-

dizer and ensure correct
addition

22. Add more PbO

23. Standardize furnace opera-

tions to obtain nearly
neutral atmosphere

24. Decrease silica or niter, and

increase PbO for addi-
tional button weight ex-
pected

acceptable although it is more convenient, and generally more
reliable, to use slags that are sufficiently fluid to pour cleanly
from the crucible. " Shotting" refers to the presence of small
shots or globules of lead distributed throughout the slag. The
assayer should make a practice of inspecting the interior of the
crucible just after pouring, to detect the existence of shotting.
All shotted assays, as well as those containing undecomposed or
insoluble slag components, should be rejected. The crucibles in
which such assays were made should be discarded unless cleaned
out by the use of a blank charge in which a normal button and
slag are produced.

164 FIRE ASSAYING

Hard or brittle buttons are unacceptable only if they indicate
the presence of too much copper, antimony, arsenic, or other
impurity that might cause loss or freezing during cupellation, or
if they are so brittle as to cause loss when separated from the
slag and hammered into shape for the cupel. Buttons containing
copper usually have an exceptionally bright luster and separate
easily from the slag, but unless they are notably hardened they
may ordinarily be used without excessive cupellation loss. But-
tons that show a faint copper color, especially on the upper
surface, are generally discarded, as they contain too much
copper for reliable results in cupellation. Antimony has a
pronounced hardening effect on lead but does not cause brighten-
ing of the luster and hence may be distinguished from buttons
hardened by copper. Arsenic, zinc, and a few other metallic
impurities cause brittleness. For superior accuracy, buttons
rendered appreciably hard or brittle by antimony, arsenic, or
zinc, as well as by copper, should be discarded. Under certain
conditions, particularly with low fusion temperatures and with
excess litharge in the slag, some PbO may become mixed in the
button, causing a form of brittleness that is seldom sufficient
to cause loss in handling arid does not cause cupellation loss.

Buttons from fusions in which a detectable layer of matte or
speiss appears should be discarded.

The addition of litharge just before pouring helps to prevent
shotting and also helps to produce soft lead buttons. Some
assayers treat all important slags with approximately 1 assay
ton of litharge, to which a small quantity of a reducer may be
added in order to obtain a slag-washing effect.

CHAPTER VIII
THE SCORIFICATION ASSAY

The scorification assay is an oxidizing fusion in a shallow fire-
clay dish, known as a "scorifier," of a relatively small quantity
of ore and a minimum of acid fluxes with a comparatively large
amount of granulated lead. Oxidizing conditions are maintained
by admitting air to the muffle, which oxidizes a part of the lead
and roasts the ore. The fused litharge combines with silica,
borax glass, and the oxidized base metals of the ore to form a slag.
Precious metals are collected by the molten lead and remain
alloyed with the lead button. The buttons from the scorification
are cupeled, weighed, and parted as in the crucible assay.

Applications and Limitations. The scorification method is best
adapted to the assay of ores (1) that are sufficiently rich in gold
or silver so that weighable beads are obtained with ore samples
not exceeding 0.2 assay ton, and usually 0.1 assay ton, in weight;
(2) that are so homogeneous that a small sample is sufficiently
reliable; (3) that are relatively free from basic oxides and from
metals, other than mercury, below lead in the electromotive
force series of elements ; and (4) that contain oxidizable impurities
but not any appreciable quantity of decrepitating compounds.

The above limitations restrict the ideal applications of the
method to high-grade gangue-free or siliceous ores or metallurgic
products containing oxidizable forms of zinc, tin, manganese,
copper-free nickel, cobalt, or lead, particularly those in which
silver, rather than gold, is the principal precious metal. For
such materials, scorification is an acceptable, but not a necessary,
alternative to the crucible assay.

Ores containing large proportions of basic oxides are unsuited
to scorification because of the excessive quantity of acid fluxes
required and because the method, being essentially an oxidizing
process, does not favor the reduction of higher oxides to the lower
forms that are more readily slagged. Hence, unfused scoriae of
iron and manganese may be formed when hematite, magnetite,

165

166 FIRE ASSAYING

pyrolusite, and similar minerals are scorified. Calcium and
magnesium oxides are not fluxed in scorification without excessive
additions of acid fluxes, as the normal scorification slags are
too basic for the formation of fusible slags high in lime or
magnesia.

Ores that decrepitate violently are troublesome in scorification
because of "spitting," which is the violent projection of small
particles of lead from the scorifier. Products such as zinc
cyanide precipitates, containing considerable metallic zinc, may
suffer loss during the intense combustion of zinc.

Ores containing antimony, arsenic, bismuth, copper, nickel-
copper, tellurium, and other metals below lead in the electro-
motive force series can be more readily handled in the crucible
assay, as a greater excess of litharge is available to help oxidize
the impurities.

One of the useful applications of scorification is the retreatment
of crucible assay buttons, either to remove certain impurities,
to decrease their size, or to combine two or more crucible -buttons
in case the crucible charge is so large as to require distribution
between two or more crucibles. ; s The beginner is apt to overdo
this application of scorification in the endeavor to avoid repeti-
tion of crucible assays in which a large or impure button was
obtained. It must be kept in mind that the addition of a
scorification step following the crucible fusion increases the slag
loss of precious metals, and, furthermore, that scorification is
less adapted to the removal of nonvolatile impurities than the
crucible assay.

The principal advantages of scorification over crucible assaying
are that the control of button size is independent of the nature
of the ore, the cost of fluxes is low, less time is required to prepare
the charges for melting, and the muffle capacity for scorifiers can
be made greater by stacking a second layer of scorifiers on the
rims of the first layer. These advantages are largely offset by
the narrow limitations of the method with respect to the size
and character of the sample, and by a longer heating cycle in the
furnace. The cost of fire-clay ware is nearly equal for the two
processes, because scorifying dishes, though cheaper, are ordi-
narily usable^ only once, whereas the average life of crucibles is
3 to 4 fusions. On ores of types suited to ideal scorification,
the accuracy of scorification compares favorably with the crucible

THE SCORIFICATION ASSAY 167

assay for equal sample weights, but in general the error is greater
since smaller samples are used.

Scorifieation has no chemical advantages over the crucible
method for ores that are free from oxidizable compounds. Sco'ri-
fication has an advantage over the crucible assay only hi the
separation of impurities having volatile oxides that escape as a
vapor more readily in scorification than in the crucible assay.
On account of the simplicity of the scorification method, it is a
suitable alternative for high-grade silver ores associated with
lead.

Slag Characteristics. The chief distinction between the slags
formed in scorification and those formed in the crucible assay
is that the scorification slags are free from alkalies unless present
in the ore. The silica and borax-glass content of scorifier slags
is kept at a minimum by adding very little of these reagents.
Normally, only 1 to 3 g. of borax glass is added for 0.1 assay
ton of ore charge. No silica is added unless it is entirely lacking
in the ore, when 1 or 2 g. may be added to prevent excessive cor-
rosion of the scorifier. Although most authorities 1 classify
scorifier slags as oxide slags high in litharge, in which other
metallic oxides are dissolved, actual scorifier slags increase their
silica content at the expense of the scorifier and are distinctively
glassy, corresponding closely in composition to lead sub- to
monoborosilicates. Within this range the litharge of the slag
is able to act as an active carrier of oxygen to aid in oxidizing
metallic sulfides and to retain metallic oxides in the slag. Care
must be taken to avoid excessive additions of acid fluxes, because
in slags more acid than the monosilicate the litharge is so firmly
combined in the slag that it fails to act as an oxidizer.

When it is necessary to add excessive quantities of silica, as in
fluxing CaO or MgO, the operation becomes a miniature crucible
assay with little or no true scorification, and an actual crucible
assay is greatly to be preferred as the fluxes can be balanced more
satisfactorily.

Control of Button Size. The control of button size in scorifica-
tion is fixed by the geometry of the scorifier and the surface

1 BUGBEE, E. E., "Textbook of Fire Assaying," 2ded., p. 127, and note 15,
p. 133, John Wiley & Sons, Inc., New York, 1933. FULTON, C. H. and
SHARWOOD, W. S., "Manual of Fire Assaying," 3d ed., p. 151, McGraw-Hill
Book Company, Inc., New York, 1929.

168

FIRE ASSAYING

tension of the lead and slag. Molten lead has a higher surface
tension than molten slag. When the charge is first melted, the
unfused ore floats on a bath of lead. As roasting and slagging
reactions proceed, the slag forms a ring around the outside of the
high meniscus of the lead. As more slag is produced at the
expense of the lead the slag ring gradually rises over the lead
meniscus until finally the slag completely covers the remaining
lead, and the assay is ready to be poured.

Commercial scorifiers have an inside shape corresponding
approximately to a spherical segment, and are available in two
general shapes, shallow and deep, and in various sizes ranging
from 2 to 4 in. inside diameter at the top. The sizes most com-
monly used are 2 in., 2}/% in., or 3 in. diameter. For most
routine work, scorification practice can be standardized so that
only one size, or at the most, two sizes, of scorifier need be carried
in stock.

For each scorifier size of a given shape, the size of button is
fixed, for all practical purposes, by the amount of granulated
lead used in the assay. Table XVI gives, for the deep form of
scorifier used in American practice, the relation between scorifier
size and granulated lead used to obtain buttons weighing 20
to 25 g.

TABLE XVI. SIZE OF DEEP SCORIFIER vs. WEIGHT OF GRANULATED LEAD

Scorifier diameter, inches
Granulated lead required, grams . . .

2
30

2M
35

2?*
45

3> 2 '
75

Weight of Ore Sample. For mixtures composed largely of
galena with quartz gangue, 0.5-assay-ton charges may be assayed
in a 3- or 3J^-in. scorifier with 75 g. of lead and 2.5 g. of borax
glass, but this does not represent common practice, and the
maximum weight of ore that is ordinarily assayed by scorification
is 0.2 assay ton. This requires a 2j^-in. deep-form scorifier and
45 g. of lead for si m ple galena or pyrite ores with quartz gangue.
For more difficult ores, either a larger scorifier or a smaller sample
weight or both must be used. For example, 0.1-assay-ton sam-
ples of difficult ores may require 60 g. of lead and 3-in. scorifiers.
Samples as small as 0.05 assay ton may be used for very rich
materials, such as cyanide precipitates.

THE SCORIFICATION ARRAY 169*

On account of the small weight of ore taken for assay, excep-
tional care must be taken to ensure that the pulp is ground and
mixed properly and that the portion taken for assay is weighed
accurately. Samples of rich materials from which 0.05 or 0.1-
assay-ton portions are to be obtained should be ground to
200-mesh or finer, and weighings should be accurate to 1 mg. or
less. See Chap. Ill for the principles involved.

Scorification Procedure. The general procedure for scorifica-
tion is embodied in the following steps:

1. Preparation of Charge. Decide upon the weight of ore and
size of scorifier to be used, thus determining the weight of granu-
lated load required. Measure out one-half of the required
amount of granulated lead and place it in a dry scorifier. Weigh
out the ore and mix with the lead in the scorifier, cover with the
second half of the lead, and sprinkle the required amount of
borax glass over the surface.

2. Fusion. Place the scorifier into a muffle at 500 to 600C.,
close the door and muffle^draft, and heat for 2 or 3 min. until the
lead is molted and danger of decrepitation or spitting is past.

3. Roasting. Open the door and draft to admit air and grad-
ually raise the muffle temperature to 900C. In a few minutes
roast reactions are completed and patches of unfused ore should
disappear, leaving a smooth-surfaced ring of slag surrounding
the lead bath.

4. Scorification. Continue heating in an oxidizing atmosphere
until the slag covers the lead. The temperature may be raised
to 1000C. or more if necessary to scorify ores containing large
amounts of base metals other than iron, lead, or copper, but
excessive temperature should be avoided as it increases the
amount of silica in the slag by scorifier attack and may increase
the loss of silver by volatilization.

5. Liquefaction. Toward the end of Scorification, increase
the muffle temperature to 1000 or 1050C. to decrease the viscosity
of the slag for clean pouring. This may be done conveniently
by closing the muffle door and draft for a few minutes after the
Scorification stage is complete.

6. Pouring. Pour the fusion in the usual manner, allow to
cool, and separate the button from the slag. If the button is soft
and malleable it is ready for cupellation in the usual manner.
Buttons containing impurities could be rescorified with additional

170 FIRE ASSAYING

TABLE XVII. SCORIFICATION DIFFICULTIES AND THEIR CORRECTION

Phenomenon

Cause

Remedy

Spitting

1. Moisture in scorifier
2. Volatile matter in scorifier
due to impure clays or
underfiring in manu-
facture. (Ignition loss
at 1000C. should not
exceed 2 per cent)
3. Decrepitating minerals in
ore

4. Excessive rate of heating
or oxidation in early
stages

1 . Dry thoroughly before use
2. Try other brands under
comparable conditions

3. Assay by crucible method,
or heat very slowly at
first
4. Decrease heating rate or
air admission or both
during melting and
roasting stages

Cracked or
leaky
scorifiers

5. See 1, 2, and 4
6. Excessive amounts of cor-
rosive constituents pres-
ent, for example, cop-
per

5. See 1, 2, and 4
6. Crucible assay preferred

Scoriae

7. Deficiency of acid flux

8. Insufficient temperature
9. Large amounts of Sb, As,
Ni, Sn, or higher oxides
of Fe or Mn

7. Increase acid fluxes, but if
more than 3-4 g. are
needed, ore should be
assayed by crucible
method
8. Increase temperature
9. Crucible assay preferred

Shotting

1 0. Insufficient finishing tem-
perature
11. Deficiency of PbO in slag

10. Finish at higher tempera-
ture
11. Decrease acid fluxes, or in-
crease Pb in charge

Impure
buttons

12. Presence of impurities not
readily removed by
scorification, as Sb, As,
Cu, Ni, Se, Te

12. Crucible or combination
assay preferred

Outsize
buttons

13. Incorrect proportions of
lead to scorifier size

13. Adjust quantity of granu-
lated lead (Table XVI)

THE SCORIFICATION ASSAY 171

lead, but this treatment is not entirely successful when much
copper, nickel, or tellurium is present, and in any event the
additional slag loss from the second scorification must be taken
into consideration.

Trouble Shooting in Scorification. As in the crucible assay,
scorification slag colors often indicate the presence of specific
metals in the ore. The slag colors may best be observed by the
glaze on the cooled scorifier after pouring. In addition, scoriae
and fumes may aid in the identification of impurities, as in
cupellation.

The principal abnormalities that may occur in scorification
are spitting, cracked or leaky scorifiers, scoriae, shotting, impure
buttons, and outsize buttons. If any of these phenomena are
pronounced, the accuracy of the assay is adversely affected.

The causes and remedies of scorification difficulties are given
in Table XVII.

CHAPTER IX
THE ASSAY OF BULLION FOR GOLD AND SILVER

In assaying and metallurgy, "bullion" is a generic term refer-
ring to alloys containing sufficient of the precious metals so that
their recovery or estimation is of economic significance.

Lead Bullion. Lead bullion is a bullion in which lead is the
dominant metal. At lead smelters the blast-furnace product is
known as "base bullion " and usually contains in excess of 90 per
cent of lead together with most of the gold and silver charged to
the furnace and minor amounts of various base-metal impurities,
chief among which are antimony, arsenic, and copper. The term
"base bullion" is also used to refer to impure silver or gold bul-
lion produced from zinc precipitates in the cyanide process,
in which case the chief impurities are zinc, lead and copper, but
since such bullions are predominantly silver and gold they should
be classed as silver or gold bullions rather than as lead bullions.

At lead smelters, in addition to the blast-furnace lead, routine
assays include the assay of refined lead and various intermediate
products. The general methods described under the heading of
lead bullion may be adapted to the assay of any such materials,
provided that lead is the dominant metal present.

Copper bullion is defined as a bullion containing in excess of
50 percent of copper, and not more than 10 per cent of base metals
other than lead and copper. The most common copper bullions
occurring in assay practice are the products of various stages of
the copper smelting and refining processes. The product of the
copper converting process is referred to as "blister copper"
because the evolution of gases during the solidification of copper
causes the surface of the cast copper to become blistered. Blister
copper usually contains in excess of 98 per cent of copper, a small
part of which is present as sulfide and as oxide. It commonly
contains in excess of 50 oz. of silver and 0.5 oz. of gold per ton,
but the content of precious metals depends upon the amount
present with the copper in the materials being smelted. Blister

172

THE ASSAY OF BULLION FOR GOLD AND SILVER 173

copper and other sources of crude copper to be refined electro-
lytically are given a preliminary furnace refining, then cast
into anodes, in which form the copper is known as anode copper,
and contains slightly more copper, gold, and silver than the origi-
nal blister copper. The final product of the electrolytic refinery
is called " refined copper " and usually contains less than 0.6 oz.
of silver and 0.01 oz. of gold per ton.

Silver bullion is a bullion containing more than 50 per cent of
precious metals of which more than half is silver, and with less
than 10 per cent of base metals other than lead and copper. The
term "dor 6 bullion" literally means gold-bearing and is generally
used to refer to bullion containing more than 90 per cent of
combined silver and gold, in which silver is dominant. "Fine
silver bullion" refers to bullion containing more than 99 per cent
of silver and practically no gold.

Gold bullion is a bullion containing more than 50 per cent of
precious metals of which more than half is gold, and with less than
10 per cent of base metals other than lead and copper. "Fine
gold bullion" is bullion containing more than 99 per cent of
gold and practically no silver. Free gold or retorted amalgam
from lode or placer mining is a common source of gold bullion.
Dental and jeweler's scrap gold is another source. Some of these
materials may be so impure as to merit the classification of base
bullion.

Fineness. The assay of gold and silver bullion and of some
of the richer base bullions is expressed in parts per thousand,
known as the fineness. For example, a gold bullion containing
990 parts of gold per 1000 is 990 fine. A dor 6 bullion containing
800 parts of silver and 150 parts of gold is 800 fine silver and 150
fine gold, and the total fineness is 950.

For convenience in bullion assaying the milli&me system of
assay weights is frequently used. The millieme is 0.5 mg. and
1 ,000 milliemes (0.5 g.) is the amount of bullion usually taken for
assay. All weights are stamped in milli&rnes, and if a 5-mg.
rider is used on the assay balance, the weights are read directly
in fineness.

.Sampling. The sampling of bullion requires exceptional pre-
cautions, which are fully dealt with in Chap. II. Samples are
obtained as chips, borings, granules, filings, or as sheet metal,
depending upon the method of sampling employed.

174 FIRE ASSAYING

Weight of Sample. The weight of bullion to be taken for
assay should be such that the combined weight of silver and gold
in the assay is not greatly in excess of 1J^ g. Larger beads are
difficult to part, and no increase in accuracy is obtained. A
0.5-g. (l,000-millime) sample is satisfactory for all silver or gold
bullions, that is, those that have a total fineness in excess of
500, unless the gold content is so low as to require a larger weight
for accurate weighing of the gold, in which case a 1.0-g. sample
may be used. For base bullions and for lead or copper bullion,
if the assay is to be reported in ounces per ton, the assay-ton
system of weights is used, using 0.05, 0.1, 0.2, or other suitable
weight up to 1 assay ton for low-grade bullions.

On account of the difficulty of weighing out a predetermined
quantity of bullion in the form of sheet, turnings, or chips, an
approximate quantity is weighed out and the exact weight is
determined and recorded. Experienced assayers, however, often
learn to weigh out the exact desired quantity without undue loss
of time, and it is not uncommon to make an average of one
weighing per minute. If the sample is in the form of filings, the
difficulty of obtaining a predetermined sample weight is decreased.
An accuracy of 0.05 mg. is sufficient for bullion weighings as this
is within the limits of variation of duplicate assays.

For control assays in buying and selling bullions it is best
to make all assays in triplicate or quadruplicate.

LEAD-BULLION ASSAY

For ordinary lead-smelter bullions, comparatively free from
interfering impurities, a simple cupellation of 0.5- or 1.0-assay-ton
samples, wrapped in 5 to 8 g. of sheet lead, is satisfactory. A
cupel correction may be used if a high degree of accuracy is
required. If the bullion contains impurities that interfere with
cupellation, then scorification, crucible fusion, or acid treatment
should be used prior to cupellation.

COPPER-BULLION ASSAY

The assay of copper bullion or blister copper for gold and silver
involves the removal of most of the copper prior to cupellation.
This may be done by scorification, crucible assay, or by acid
treatment. Acid-treatment methods, described in Chap. X,
give the best results.

THE ASSAY OF BULLION FOR GOLD AND SILVER 175

The scorification method, though at one time accepted as
standard by smelters, is wasteful of time and materials and gives
high silver losses unless slag and cupel corrections are made,
because several successive scorifications are needed before the
buttons are pure enough for cupellation.

Crucible Method. The crucible method is rapid and is suffi-
ciently accurate for preliminary estimates or for plant-control
purposes. It is carried out by placing the assay portion of cop-
per bullion, mixed with powdered sulfur, in the bottom of a
30-g. assay crucible. The copper-sulfur mixture is covered with
a high litharge flux, and the charge is then melted. The sulfur
reacts first with the copper to form a copper sulfide matte. Then
the litharge reacts with the matte to oxidize both sulfur and
copper and is itself partly reduced to form a lead button, in
accordance with the principles discussed in Chaps. VI and VII.
The copper-sulfur mixture is not mixed with the flux, in order to
avoid reduction of lead before the copper is oxidized.

A maximum of J^ assay ton of copper can be fluxed in a 30-g.
crucible, and, with this amount of copper, 1.5 g. of sulfur will
produce a lead button of approximately 25 g. The slag should
be very high in litharge a silicate degree of 0.25 gives better
results than the subsilicate slag ordinarily recommended for ores
high in copper. Using the method of charge calculation given
in Chap. VII, the complete charge is as follows:

Copper bullion . 25 A.T.

Sulfur 1.5 g.

Flux (Mix separately and place on top of the
copper-sulfur mixture) :

Sodium carbonate 8 g.

Litharge 250 g.

Silica 9 g.

Fuse in a hot furnace and pour immediately upon completion,
in 15 to 20 min. The buttons from two J^-assay-ton charges of
the same bullion may be scorified together to remove additional
copper prior to cupellation and to increase the accuracy of weigh-
ing the beads.

SILVER-BULLION ASSAY

The assay for silver in silver or gold bullion may be carried
out by either cupellation or volumetric methods. Cupellation

176 FIRE ASSAYING

is used to obtain approximate results. For greater accuracy the
volumetric methods are preferred.

Cupellation Method. The errors in the cupellation assay for
silver in silver-rich bullions are corrected by means of a check or
" proof " assay of a synthetic sample of approximately the same
weight and composition as the bullion sample. The check sam-
ple is cupeled alongside the bullion sample, and it is assumed that
the losses as determined for the check apply to the bullion. Thus
the cupellation method involves a preliminary assay to deter-
mine the approximate composition of the bullion before the check
can be prepared. The entire procedure is as follows:

1. Preliminary Cupellation. Weigh out nearly or exactly
500 mg. of bullion and record the exact weight. If it is known
that extra silver is needed for parting, the estimated amount of
proof silver should be weighed out and added to the bullion,
recording the exact weight added. Fold a small square or
rectangular strip of lead foil into a cone or box, add sufficient
silver-free granulated lead 1 to make a total weight of 20 g., bury
the weighed-out bullion sample and the inquarting silver, if any,
in the granulated lead, and wrap into a compact shape. Cupel
carefully with feathers and with precautions to prevent freezing
or sprouting. Weigh the bead, part, and weigh the gold in
the usual manner. If the bead does not part, the preliminary
cupellation should be repeated with sufficient . added silver, the
exact weight of which is recorded.

2. Preparation of the Check. a. Assume that approximately
1 per cent of the silver was lost in the preliminary assay but that
the gold is approximately correct. The original weight of the
preliminary bullion sample, plus the added silver, if any, minus
101 per cent of the weight of the bead is the estimated weight of
base metal present. The usual base metal is copper and if this
is verified by the cupel stain, use c.p. copper foil to make up the
base-metal content of the check. If the base metal is not
dominantly copper, ascertain by qualitative tests, spectrographic
analysis, or other means what metals are present, and use those
metals in preparing the check.

If the silver-gold ratio is so high as to cause serious disintegra-
tion of the gold in parting, it is desirable to add proof gold in

1 It is advisable to use as little lead foil as possible, as it generally contains
silver, whereas granulated lead may be obtained practically silver free.
A correction should be applied if necessary.

THE ASSAY OF BULLION FOR GOLD AND SILVER 177

sufficient amount so that the bead will not contain more than 3
or 4 parts of silver to 1 part of gold. With bullions low in
gold, this procedure is impracticable, and the assayer must rely
upon manipulative skill to recover all the gold in the floured
partings.

b. Weigh out the ingredients for the check assay in the same
proportions as were indicated in the preliminary bullion cupel-
lation, using proof silver, proof gold, and c.p. base metal. If the
final assay is to be made in duplicate, one check is sufficient; if
in triplicate, two checks are required.

c. Wrap the ingredients of the check in the same amount of
lead as for the bullion assay, as per section 3a following.

3a. Lead Required. If the base impurity is copper, weigh out
the amount of sheet and granulated lead for the cupellation of
500 mg. of bullion and for the synthetic checks in accordance
with the recommendations in Table XVIII.

TABLE XVIII. LEAD REQUIRED FOR CUPELLATION OF COPPER-BEARING

SILVER BULLION
(500-mg. Bullion Sample)

Fineness of
bullion,
An + Ag

Weight of copper

Weight of
lead, grams

Ratio,
Pb:Cu

Milligrams

Milliemes

1,000

950

200

900

100

140

850

150

120

800

100

200

110

700

150

300

600

200

400

500

250

500

36. Preliminary Treatment of Bullions Containing Sb, Se, Te
(and Bi). If the principal impurity is antimony, first scorify
the bullion and checks in 2j^-in..scorifiers with 30 g. of lead and
make up the scorifier buttons to within 1 g. of the same weight
by adding sheet lead to the lighter ones if necessary. It is not
essential that the checks contain the same amount of antimony
as the bullion because the antimony is removed by scorification,
and the scorification losses are practically the same with or with-

178 FIRE ASSAYING

out the small amount of antimony that might be introduced with
a 500-mg. sample of bullion.

If the bullion contains selenium or tellurium in notable amounts,
first scorify the bullion and checks in a 2^-in. scorifier with 40 g.
of lead, then flatten the buttons into sheets about 3 in. square and
proceed to remove the selenium and tellurium by the acid-treat-
ment and scorification method given in Chap. X. It is not
necessary that the checks contain the same amounts of selenium
and tellurium as the bullion, but it is essential that no selenium or
tellurium appear in the cupellation.

Bismuth may be removed by the same, treatment as for selenium
and tellurium, but if the amount of bismuth is known, and the
checks are made up accordingly, direct cupellation will give
good results, as the losses are not greatly different than when
lead is the cupeling medium, and the results will be consistent
between the bullion and the checks. Some bismuth may be
retained in the beads.

4. Final Assay. Arrange the cupels in the furnace so that the
bullion buttons and the check buttons alternate in the same row.
Cupel carefully with feathers with particular attention to uni-
formity of temperature across a row so that the checks and the
bullion buttons they control receive equal treatment. Raise
the temperature toward the finish to avoid freezing, and be
careful to avoid sprouting. Weigh and part the beads in the usual
manner. The preparation of a cornet for parting, as described
for the assay of gold bullion, is desirable, especially if the silver-
gold ratio approaches the minimum (2:1) for complete parting.

5. Calculation of Results. The principle that applies to the
calculation of the silver and gold is that the percentage of loss
in the checks is applicable to the percentage of loss in the bullion
samples adjoining the check in the muffle. If two or more checks
are used in the same row, the average percentage of loss in the
checks is applied to all the bullion assays in the row. The
results are expressed in fineness (milliemes) of gold and silver
and should check within 0.1 gold fineness and 0.5 silver fineness.

Volumetric Method. 1 Chemical methods for the determination
of silver are based on the insolubility on dilute nitric acid solutions

1 For a more extended discussion, see standard treatises on quantitative
analysis, such as W. F. Hillebrand and G. E. F. Lundell, " Applied Inorganic
Analysis," John Wiley & Sons, Inc., New York, 1929.

THE ASSAY OF BULLION FOR GOLD AND SILVER 179

of either the chloride or the thioeyanate. The volumetric thio-
cyanate, or Volhard's, method is generally the most convenient
and is the one described here. Good results are obtained on
ordinary silver bullions in which copper is the principal impurity.
For impure alloys, ores, and metallurgical products, fire assay
methods are generally more suitable, on account of the difficulty
of removing interfering elements.

Principles. The reactions on which Volhard's method is based
arc as follows:

AgNO 3 + KSCN = AgSCN + KNO 3 (1)

3KSCN + Fe(N0 3 ) 3 = Fe(SCN) 3 + 3KNO 3 (2)

Reaction (2) does not occur until (1) is completed, hence the
addition of a small amount of a soluble ferric salt serves as an
indicator by the formation of red ferric thioeyanate in the pres-
ence of an excess of thioeyanate.

If more than 70 per cent of copper is present, silver in known
quantity must be added to prevent interference of copper in
titration.

Chlorides interfere by reaction with the silver, hence all rea-
gents and glassware must be free from chlorides.

Mercury reacts with thioeyanate in a manner similar to silver
and hence must be absent. Amalgams and retort sponge must
be given a preliminary fusion.

Palladium interferes but is rarely present. Other interfering
elements and compounds are generally eliminated by the acid
treatment of the sample.

Standard Thioeyanate Solution. Either potassium or ammo-
nium thioeyanate may be used as a standard solution. A 0.05 N
solution is satisfactory, but many chemists prefer to use an empiri-
cal standard that is adjusted so that 1 ml. on a 0.5-g. sample of
bullion is exactly equal to 1 or 2 per cent of silver. A solution
equivalent to 1 per cent of silver per milliliter when a 0.5-g.
sample is taken requires 9.01 g. of KSCN or 7.06 g. of NH 4 SCN
per liter. Standardize the solution against a weighed amount of
pure silver, treated as in the procedure given below.

Ferric Alum Indicator. Prepare a saturated solution of ferric
alum in distilled water, then add sufficient (usually about 10 per
cent) pure colorless nitric acid to bleach the brown color of the
solution.

180 FIRE ASSAYING

Procedure. 1. Weigh out 0.5 g. of bullion and transfer to a
200-ml. Erlenmeyer flask or a 250-ml. beaker.

2. Dissolve in 15 ml. of 1 : 1 nitric acid and boil until nitrous
compounds (brown fumes) are expelled.

3. Cool, dilute to 50 to 75 ml.

4. Add 5 ml. of ferric alum indicator and titrate with standard
thiocyanate solution to the appearance of a faint red color that
persists on shaking or stirring.

5. Calculate the percentage (or fineness) of the bullion from
the amount of thiocyanate used and its known strength as deter-
mined by standardization.

/
GOLD-BULLION ASSAY v/

The general method of assaying gold bullion for gold at mints
and assay offices is by cupellation and parting accompanied by
check assays on synthetic alloys corresponding in composition
to the bullions. To ensure uniform parting without disintegra-
tion of the gold, the ratio of silver to gold must be adjusted within
narrow limits, the bead must be rolled thin, and the parting
procedure standardized carefully. To aid in the removal of
lead during cupellation and to toughen the alloy to minimize
cracking in rolling, approximately 3 per cent of copper should be
present prior to cupellation. Some copper is retained in the
bead but dissolves in the parting acid and does not affect the
gold assay. The removal of lead is necessary to avoid error and
to prevent breaking up of the gold in parting.

In the gold assay, small losses of gold occur by volatilization
(approximately 0.0001 per cent), absorption (0.04 to 0.05 per
cent), and solution (0.0005 per cent), but in assays exceeding
700 to 800 fine gold plus silver, these losses are more than offset
by the retention of silver in the gold, so that the final weight of
gold in such cases is usually from 0.02 to 0.1 per cent high. The
algebraic sum of the errors is known as the " surcharge," and
its exact value is determined by calculation from the check assay.
A negative surcharge will usually occur in bullions less than 700
to 800 fine.

The procedure given herein is suitable for the determination
of gold in copper-bearing bullions containing upward of 500 gold
plus silver fineness. If the base metal is not copper or lead and
is present in interfering amounts, see the suggestions given in

THE ASSAY OF BULLION FOR GOLD AND SILVER 181

paragraph 36 of the cupellation method for the silver-bullion
assay.

The method to be chosen for the determination of silver in
gold bullion depends on the composition of the bullion and the
degree of accuracy required. For most purposes the silver
fineness as calculated from the data obtained from the gold assay
is sufficiently accurate, but many assayers prefer to use a volu-
metric method.

1. Preliminary Assay. Weigh out 500 mg. (1,000 milliemes)
of bullion and cupel as in the preliminary cupellation of silver
bullion. Weigh the resulting bead and subtract from the
original weight to estimate the base metals present.

Experienced assayers judge the approximate fineness of the
bead by color, but the touchstone method is more reliable. In the
touchstone method the streak of the sample on a black jasper
slab is compared with the streak of alloy strips ("needles")
of known composition. 1 If no other means of estimating the
silver-gold ratio is available the bead is inquarted with enough
silver to make the silver-gold ratio from 2:1 to 3:1 and
parted. Estimation of the gold fineness within 5 per cent by
the preliminary assay is sufficiently accurate unless silver is to
be determined in the same assay, when a closer estimate is desir-
able. In the gold assay, variations in composition and weight
of the bead and in the conditions of cupellation do not affect
the cupellation losses as much as with silver.

2. Final Assay. Duplicate assays and a single check are usu-
ally sufficient. The check is made up as nearly like the bullion
in weight and composition as possible. Since all beads will be
calculated to contain a fixed ratio of silver to gold, and since all
will contain a fixed minimum amount of copper, it is convenient
to prepare a proof alloy of silver, gold, and copper that can be
used as the foundation for preparing the checks, to which addi-
tional copper is added as needed. Pure or "proof " gold for mak-
ing check assays can be purchased or may be prepared in the assay
office.

The proper silver-gold ratio is established by each assayer
within the limits of 2 : 1 to 3 : 1 and is coordinated with the parting

1 Jeweler's test needles are used for standard jewelry alloys, in which
copper is the principal alloying element with gold. Such needles read in
carats, and 24 carats is 1,000 fine.

182 FIRE ASSAYING

procedure. Once established, the same ratio is maintained on all
subsequent assays. If less than 2:1 silver to gold is used, too
much silver is retained in the gold when parted, unless the parting
treatment is inconveniently prolonged. With more than 3 parts
of silver to 1 of gold, the gold may break up in parting. The
Sari Francisco mint uses a ratio of 3 silver to 1 gold, the Denver
mint uses a 2J^ to 1 ratio, and the New York mint uses a 2 : 1
ratio.

The lead for cupellation is ascertained from Table XVIII.
Some authorities recommend slightly different ratios of lead for
gold bullion than for silver bullion, but the distinction is unim-
portant. Some assayers prepare uniform squares of sheet lead
weighing 5 or 6 g. and use as many as necessary to approximate
the desired ratios, keeping the same proportions for the duplicates
and checks of a given sample.

The description of the silver-bullion cupellation should be
referred to for the manner in which the ingredients of the assay
are wrapped together and cupeled. In the gold-bullion assay it is
not so important to use strictly silver-free lead, hence all the lead
may be supplied as lead foil if desired.

Since the melting point of a silver-gold alloy is higher than
that of silver, the cupellation temperature at the finish should
be higher than with silver bullions. Moreover, if gold only is
being determined, more complete removal of lead and copper is
obtained at higher cupellation temperatures without increased
loss of gold, provided that temperatures in excess of 1050C.
are not employed. Hence, the entire cupellation may be con-
ducted at a muffle temperature of 900 to 950C. Any variation
in the gold loss due to cupellation temperature will be com-
pensated by the check if the temperature is uniform for the assays
and checks in a* set. At the finish of cupellation the blick will
occur in the absence of platinum. It is desirable to leave the
cupels in the furnace for 10 to 20 min. after finishing, to remove
the last traces of lead, but many assayers remove the cupels at
once, even while still fluid, as there is no danger of sprouting if
the silver-gold ratio is less than 3 to 1, and if copper is present.

3. Parting. The beads are removed from the cupels, flattened
with a hammer, rolled into a fillet approximately 0.01 in. thick,
and shaped into a spiral or cornet. Parting is commonly done
in parting flasks. Two acid treatments of 15 to 20 min. each
at or near the boiling point are generally necessary for com-

THE ASSAY OF BULLION FOR GOLD AND SILVER 183

plete parting, using nitric acid of 1.17 to 1.22 sp. gr. for the first
treatment, which is decanted off and replaced by strong acid,
for example, of sp. gr. 1.2. Approximately 30 ml. of acid are used
for each treatment. After the acid treatments the acid is
decanted off, and the cornet is washed at least three times with
hot distilled water, transferred to an annealing cup, annealed,
and weighed. The gold is reported in fineness, and the correction
from the check applied. Duplicate results should check within
0.1 fineness.

It is essential that the preparation of cornets and all subsequent
parting operations be standardized because variations in the
thickness reduction at each stage of hammering and rolling, in
acid strength and time of contact, and in annealing practice
affect the amount of surcharge.

* PREPARATION OF FINE GOLD

Assayers sometimes have occasion to prepare small lots of fine
gold. Several methods are available, 1 but the oxalic acid pre-
cipitation is usually the most convenient in modestly equipped
laboratories. The procedure is as follows:

Dissolve the purest gold available, as from assay partings, in
one part of c.p. nitric acid and 4.7 parts of c.p. hydrochloric
acid by volume. Drive off the excess acid, cool, dilute to con-
tain about 30 g. of gold per liter. Add a small excess of hydro-
chloric acid or salt to precipitate silver, let settle, and siphon
off the clear solution. Dilute further, settle until clear, siphon
the? clear solution into a flask. Warm the solution, and precipi-
tate gold with a solution of oxalic acid, using approximately
as much acid as gold by weight. Shake vigorously and keep
warm for several days until precipitation is nearly complete.
Siphon off the clear solution, wash thoroughly with hot water,
transfer to a porcelain dish, and evaporate to dryness. The gold
may be used as a loose powder or melted and rolled into strips.
For melting, transfer to a clay crucible that has been glazed with
molten borax, using scraps of filter paper to clean the dish. Melt
under a borax cover and then cast into an iron mold. After
casting, clean the bar with fine sandpaper or with assay silica
moistened with water, roll, and clean again. For highest purity,
polish and heat to redness just before using.

1 ROSE, T. K., and NEWMAN, W. A. C., "Metallurgy of Gold," 7th ed.,
p. 524, ,T. B. Lippincott Company, Philadelphia, 1937.

CHAPTER X

THE,ASSAY OF MATERIALS REQUIRING PRELIMINARY
ACID TREATMENT

Certain impurities in ores and metallurgical products are not
readily eliminated by fusion and cupellation, and certain others,
even though removable, introduce errors at one stage or another
of the assay process. When present in harmful proportions it
may be necessary to remove the greater part of such impurities
by preliminary acid treatment before proceeding with the fire
assay. The principal impurities that may require acid treatment
for their removal are bismuth, copper, nickel, selenium, tellurium,
and zinc.

Bismuth is rarely present in critical amounts and in general
affects only the silver assay, by retention of some bismuth in the
silver bead. The bismuth error is usually corrected by deducting
the bismuth in the silver bead (determined by wet analysis). A
correct silver assay could be made by volumetric determination
of silver in the parting acid or by preliminary acid treatment to
separate the bismuth before the assay is made. In the latter
case, silver is also extracted and must be precipitated and added
to the remainder of the ore before assaying.

The maximum tolerance for copper in an all-fire method is
approximately 7.5 g. per charge, and the tolerance for nickel is
4 g. per charge. Beyond these limits it becomes necessary to
remove copper and nickel by wet chemical methods. Preliminary
acid treatment is the preferred method of assaying copper bullion
for gold and silver.

Selenium and tellurium are the most potent of the interfering
elements in their harmful effect on the recovery of gold in fire
assaying, and as little as 0.1 g. of these elements in an assay charge
may leave enough in the lead button to cause serious loss of gold
in cupellation. Somewhat larger amounts of these elements may
be removed in fire assaying by special methods, such as an oxidiz-
ing roast prior to fusion to form iron tellurite, which is removed in

184

MATERIALS REQUIRING ACID TREATMENT 185

the slag, or by a prolonged " soaking" of the lead buttons under a
litharge flux, but in extreme cases it may become necessary to
remove selenium and tellurium by acid treatment prior to fusion
methods.

Zinc interferes only when present in metallic form, as in zinc
box precipitates, especially "zinc shorts," consisting of undis-
solved zinc shavings containing more or less gold and silver.
Although it is possible with such materials to obtain good results
directly by fire assaying, it is sometimes more convenient to
remove the zinc by acid treatment prior to scorification and
cupellation.

The essential conditions to be observed in planning a combina-
tion assay method are to ensure reasonably complete attack and
solution of the undesired impurities without dissolving gold and
then to precipitate the silver with a slight excess of sodium
chloride, followed by a prolonged period of settling. After
filtering and washing, the residues are subjected to a crucible or
scorification fusion, followed by cupellation of the resulting
button.

Choice of Acids. When considering the choice of an acid for
the removal of impurities prior to a fusion assay, it should be
remembered that it is not essential to dissolve and remove all the
interfering elements, as the subsequent fire-assay step is still
capable of removing small amounts of undesired impurities,
within the usual limits of tolerance. Hence it is more important
to avoid solution of gold than it is to secure complete decomposi-
tion of the sample.

Pure hydrochloric acid has no solvent action on gold, but in the
presence of nitric or sulfuric acid, iron sulfates, cupric chloride
and various other base-metal salts, the gold solubility in hydro-
chloric acid is such as to prohibit the use of this acid in most cases
that occur in the practice of assaying, because of the danger of
forming one or more of the above-mentioned compounds during
the treatment of the sample. Hydrochloric acid is more effective
than other acids in dissolving oxidized ores, such as those con-
taining copper, nickel, and iron, and if either nitric or sulfuric acid
fails to give satisfactory results, hydrochloric acid may sometimes
be used successfully if care is taken to avoid high concentrations
of acid at any stage of the procedure "and to avoid a great excess
of the acid over that required for the solution of impurities

186 FIRE ASSAYING

and the conversion of silver to chloride. Hydrochloric acid is
not effective in decomposing sulfides, hence incompletely oxidized
ores may require a subsequent treatment with nitric acid after
filtering and washing the residues free from hydrochloric acid.
Under no circumstances should hydrochloric acid be used with
either nitric or sulfuric acid simultaneously in contact with the
sample, otherwise excessive solution of gold may result. One
of the few safe applications of hydrochloric acid is for the treat-
ment of metallurgical products containing metallic zinc. The
zinc is readily dissolved in dilute hydrochloric acid and silver and
gold are not attacked until all the zinc is dissolved. By avoiding
an excess of hydrochloric acid and prolonged treatment, the zinc
may be removed without danger of loss of gold. In this case, as
in practically all others not amenable to the use of nitric acid, it is
preferable to use sulfuric acid in spite of the prolonged treatment
necessary in some cases to ensure adequate solution of impurities.

Hydrofluoric acid, which is sometimes used in chemical analysis
to decompose silicates, has no place in the preliminary treatment
of gold and silver ores, because its solvent action on gold is too
pronounced.

Sulfuric acid has no action on gold but is not a rapid or reliable
solvent for all the associated base-metal compounds that may be
present. The most important applications of sulfuric acid in the
present connection are in the assay of copper bullions, in the
treatment of products containing metallic zinc, in the treatment
of oxidized ores, and for the removal of selenium and tellurium.
Sulfides are not readily attacked by sulfuric acid, hence, if the
sample is composed largely of sulfides, nitric acid should be used.

Nitric acid attacks practically all the common minerals of the
elements in the interfering group, except silicates and certain
oxides. It is the best general solvent in combination assays for
gold and silver and is particularly valuable for the decomposition
of sulfides. Gold is slightly soluble in concentrated nitric acid,
but its solubility in pure dilute acid of 1.26 sp. gr. or less is so
slight that it may be disregarded in the assay of all materials
except gold bullion. It is essential that the nitric acid be free
from hydrochloric acid and chlorine to avoid the formation of
aqua regia, in which gold is soluble. Complete data are lacking
on the solubility of gold in nitric acid in the presence of various
metal salts such as might be formed in the decomposition of

MATERIALS REQUIRING ACID TREATMENT 187

heterogeneous ore samples. That appreciable solubility may
occur under certain conditions is indicated by the fact that the
nitric acid methods for the assay of copper bullion are known to
give low results on gold, which has been attributed by various
authorities to the possible presence of sulfuric or hydrochloric
acid or the chlorides or nitrates of base metals, particularly those
of iron or copper. 1

General Procedure for Combination Assays. The following
procedure will serve as a guide to the assay of materials, other
than copper bullions, that require preliminary removal of
impurities by acid treatment prior to a fusion process. Following
this section specific procedures are given for copper bullion. The
nitric acid treatment is suitable for the removal of copper, nickel,
and other impurities from sulfide ores and products, and the sul-
furic acid treatment is preferred for the decomposition of mate-
rials containing selenium or tellurium, for the oxidized ores of
copper and nickel, and for products containing metallic zinc.

la. Nitric Acid Treatment. For sulfide ores and products the
initial attack is expedited by using small increments of concen-
trated nitric acid and heating cautiously until the sulfides are
decomposed. A maximum of approximately 60 to 75 ml. of acid
is required for K-A.T. samples, and 90 to 100 ml. for ^-A.T.
samples. Avoid prolonged heating and heat just sufficiently to
decompose the sulfides and drive off the brown fumes of nitrous
oxide, usually evaporating to about half of the original volume.
Dilute nitric acid (1 : 1 to 1 : 3) is preferred for materials containing
metals in metallic form or if considerable proportions of oxides
are present, and may be used for sulfides to reduce the violence
of the reaction that occurs when concentrated acid is used.

16. Sulfuric Acid Treatment. When sulfuric acid is used it is
usually advisable to start the treatment with 1:1 rather than
with concentrated acid, using up to a maximum of 200 ml. of
the 1 : 1 acid for a 1-A.T. sample. Heat until visible dissolving
action ceases, usually 1 to 2 hr., then evaporate to strong fumes
of S0 3 .

2. Precipitation of Silver.* Cool the solution obtained from
la or Ib and dilute with at least four times its volume of water.

1 FULTON, C. H. and SHARWOOD, W. J., " A* Manual of Fire Assaying,"
3d ed., p. 160, McGraw-Hill Book Company, Inc., New York, 1929.

2 See note a, p. 189.

188 FIRE ASSAYING

Stir to break up cakes of precipitated salts. Heat to boiling
to dissolve precipitated salts, adding additional water if necessary.

Precipitate the silver by adding a calculated small excess of
sodium chloride solution, such that the resulting solution will
approximate O.OlAf chloride. 1 A preliminary assay of a small
quantity of ore may be necessary to estimate the approximate
silver content to permit the proper adjustment of the quantity
of sodium chloride to be added. After adding the sodium
chloride, heat to boiling and stir vigorously to agglomerate the
precipitate, let stand for 1 hr., stir again if the solution is cloudy,
then let settle thoroughly, preferably overnight. If much silver
is present, keep in a dark place to avoid photodecomposition of
the silver chloride.

Filter on double medium-textured filter papers or a single close-
textured paper. Be* careful to scrub the sides of the beaker
thoroughly and to transfer all residues to the paper. If adherent
crusts have formed, dissolve them in a small quantity of a hot
solution of sodium hydroxide, acidulate with the same acid used
in step 1, and wash into the filter. Wash three or four times with
cold water or with 0.01 N NaCl and discard the filtrate.

3a. Scarification Fusion. If the acid treatment resulted in
little or no siliceous residue the precipitate from step 2 can be
treated by scorification, otherwise use the crucible method
described in step 36.

Sprinkle a few grams of granulated lead on the inside of the
filter paper containing the gold and precipitated silver. Remove
the paper, fold it carefully, and place it in a 2J^-in. scorifier con-
taining 15 to 20 g. of silver-free granulated lead. Dry on a hot
plate or in front of the muffle and then place in a dull-red muffle
to incinerate the paper and decompose silver chloride. Remove
from the muffle, add 15 to 20 g. of granulated lead and J^ g. each
of borax glass and silica, and scorify in a hot muffle for 5 to 10 min.
Pour, cupel, weigh and part as usual.

36. Crucible Fusion. If a large siliceous residue was left in
step 1, or if considerable amounts of selenium or tellurium are
known to be present, the residue and precipitate from step 2
should be treated by a crucible fusion instead of by scorification.

Sprinkle a few grams gf litharge on the inside of the filter paper
containing the gold and precipitated silver, transfer to a 20-g.

1 See note 6, p. 189.

MATERIALS REQUIRING ACID TREATMENT 189

crucible in which has been placed 1 or 2 A.T. of litharge, heat the
crucible at a dull-red heat to incinerate the paper. Remove from
the furnace, cover to prevent loss or contamination, and allow
to cool. Calculate the required fluxes in the usual manner,
provide for the reduction of a 20- to 25-g. button, and proceed
with fusion, cupellation, weighing and parting in the usual
manner.

Notes: a. Some assayers prefer to filter the gold and insoluble residue
prior to the precipitation of silver in the belief that some gold may be dis-
solved in the presence of chloride. If excessive boiling is avoided after
adding sodium chloride there is little danger of dissolving gold in the dilute
acid at that stage.

b. The solubility of silver chloride is least in 0.01 N chloride solutions 1
at normal temperature and progressively increases with increased tempera-
ture, or chloride strength, or in the presence of excessive amounts of mineral
acids, alkaline or alkaline-earth chlorides or nitrates, alkaline cyanides, or
ammonium hydroxide.

If a solution containing 5.45 g. of NaCl per liter (0.0933 N} is prepared,
1 ml. will be equivalent to 10 mg. of silver. To obtain a 0.01 N chloride
solution after adding sufficient chloride for the silver, add 12 ml. additional
NaCl solution per 100 ml. of solution.

Lead acetate is used by some assayers in conjunction with sodium chloride
for silver precipitation. If the sample was dissolved in nitric acid, a few
milliliters of sulfuric acid arc added with the lead salt, in order to form lead
sulfate to aid in settling the silver chloride precipitate. This procedure is
unnecessary unless the silver content is very low and is of doubtful value in
any case.

Acid -treatment Methods for Copper Bullion. Numerous acid-
treatment methods have been described for the preliminary
removal of copper from copper bullion. Nitric acid methods were
developed by Greenwood 2 and Van Liew 3 while sulfuric acid
methods were developed by Hunt, 4 Flinn, 5 and Keller. 6 Since

1 FORBES, G. S., The Solubility of Silver Chloride in Chloride Solutions,
Jour. Am. Chem. Soc., vol. 33, p. 1937, 1911.

2 BUGBEE, E. E., " Textbook of Fire Assaying," 2d ed., p. 224, John Wiley
& Sons, Inc., New York, 1933.

3 VAN LIEW, W. R., Losses in the Determination of Gold and Silver in
Copper Bullion, Eng. Min. Jour., vol. 69, p. 469, 1900.

4 HUNT, F. F., Determination of Gold in Copper Bullion, Eng. Min. Jour.,
vol. 87, p. 465, 1909.

6 FLINN, F. B., Assay of Gold in Copper Bullion, Eng. Min. Jour., vol. 87,
p. 569, 1909.

6 KELLER, *E., Recent American Progress in the Assay of Copper Bullion,
Trans. A.I.M.E., vol. 46, p. 772, 1913.

190 FIRE ASSAYING

any method of dissolving copper also allows silver to dissolve,
provision must be made for the precipitation of silver from the
solution used to extract the copper.

Nitric Acid Method. The nitric acid method gives low results
for gold, probably owing to the slight solubility of gold during
the prolonged treatment. The results for silver are also low but
can be corrected by check assay. The method is now used by
some assayers for the assay of refined copper in which the
amount of gold is so small as to be relatively unimportant.
The nitric acid procedure is considerably shorter than the
sulfuric acid method.

A typical procedure 1 for the assay of refined copper by the
nitric acid method is given in the following steps. 2

1. Weigh out 2 A.T. of the refined copper sample and place
in a 1,500-ml. beaker. Assays are usually made in duplicate, and
a check assay is prepared of c.p. copper and silver in approxi-
mately the same amounts as in the refined copper assay portions,
in order to correct the silver result.

2. Add 200 ml. of water and 220 ml. of concentrated ni trie
acid. Cover and place on a steam plate and heat until action
ceases.

3. Remove cover and evaporate on a hot plate to low volume
in order to remove nearly all the nitric acid. Remove from hot
plate and immediately add 4 or 5 drops of concentrated hydro-
chloric acid and a few milliliters of a kaolin suspension to aid the
collection and settling of the precipitated silver chloride. Dilute
to 1 1., stir, and allow to settle overnight.

4. Filter on double medium-textured filter papers. Wash two
or three times with cold water. Excessive washing or the use of
hot water will cause increased silver loss.

5. Carefully fold the paper into a roll and place in a 2J^-in.
scorifier. Cover with 30 g. of silver-free granulated lead.

6. Dry on a hot plate or in front of the muffle and then incin-
erate in a dull-red muffle until most or all of the paper is burned.
Remove from the muffle.

1 Adapted from a procedure given in a personal communication by M. A.
Jackson, chief chemist of the Great Falls refinery of the Anaconda Copper
Mining Co., Great Falls, Mont.

2 The end of each numbered paragraph marks a convenient stopping
place.

MATERIALS REQUIRING ACID TREATMENT 191

7. Add 30 g. of additional granulated lead and J^ g. each of
borax glass and silica and then scorify in a hot muffle. Pour
and separate the button from the slag.

8. Cupel the button and calculate the silver in ounces per ton,
applying the correction as found in the check assay.

9. Wrap the beads from 20 or more samples in lead foil and
cupel, part the bead, and calculate the gold in ounces per ton.

Notes : a. In step 3, a solution of sodium chloride may be used instead of
hydrochloric acid. In either case, an excess of chlorine must be avoided.
The amount suggested is sufficient for most refined copper samples. If in
doubt, calculate the required amount stoichiometrically and add a slight
excess.

b. In step 3, settling overnight or longer is necessary to secure complete
precipitation of the silver.

c. A crucible fusion may be used in place of the scorification method
described in steps 5 to 7.

d. In the solution as recommended, from 0.2 to 0.3 mg. of silver are dis-
solved. Hence the check should be of approximately that order of
magnitude.

Mercury-sulfuric Acid Method. Copper does not dissolve
readily in sulfuric acid alone, and solution is incomplete if sulfide
copper is present. However, if the copper is first amalgamated
with mercury, complete solution is obts^ned without difficulty,
and hence the assay methods for copper bullion based on sulfuric
acid attack of the sample involve a preliminary amalgamation
of the copper.

The method is widely used on blister and anode copper and
with suitable modifications is successful for the metallic portion
of refinery slags and for other products within the classification
of copper bullions. A typical procedure follows: 1

1. Weigh out 1 A.T. of the copper into an 800- or 1,000-ml.
beaker.

2. Add 30 ml. of water, 10 ml. of a mercuric nitrate solution
containing 25 g. of mercury per liter, and 10 ml. of concentrated
sulfuric acid. Shake well, let stand until amalgamated, and then
add 100 ml. of concentrated sulfuric acid.

1 This procedure is a composite of information kindly supplied by the
Anaconda Copper Mining Company, at Anaconda and Great Falls, Mont.;
the International Smelting and Refining Company, Perth Amboy, N.J.;
and the American Smelting and Refining Company, El Paso, Tex.

192 FIRE ASSAYING

3. Cover and place on a medium hot plate, preferably over-
night but in any case until all the copper is dissolved and SO 3
fumes are evolved. Near the finish the supernatant liquid
becomes dark green and then changes to a light grayish blue at
the finishing point.

4. Remove from the hot plate and cool. Some of the copper
sulfate will crystallize as a sludge.

5. Add, with constant stirring, 100 ml. of cold water, then
complete the dilution to 400 ml. with hot water with continued
stirring. Bring to boiling and continue boiling if necessary to
complete the solution of copper sulfate. Add a slight excess of
hydrochloric acid or sodium chloride solution to precipitate silver
and mercury and continue boiling for 5 to 10 min. until the silver
chloride is coagulated, but avoid prolonged boiling. Remove
from the hot plate, wash down the sides of the beaker, dilute to
500 to 900 ml. with cold water, and allow to settle overnight.

6. Filter on double medium-textured filter papers, such as
Whatman No. 1, Munktell No. 100, or S. & S. No. 597. Wash
two or three times with cold water but avoid excessive washing.
Wipe out beaker with a small piece of filter or tissue paper and
add to the original filter paper with a pinch of granulated lead.

7. Carefully fold the paper and place in a 2J^-in. scorifier
containing 15 to 20 g. ^ silver-free granulated lead.

8. Dry on a hot plate or in front of a muffle and then place
in a dull-red muffle to incinerate the paper and decompose
silver chloride. Remove from the muffle.

9. Add 15 to 20 g. of granulated lead and }/% g. each of borax
glass and silica and scorify in a hot muffle for 5 to 10 min.

10. Cupel the button and detach the bead from the cupel.

11. Grind the slag and stained part of the cupel together,
make a crucible fusion, cupel the bead, and weigh with the original
bead.

12. Part both beads together and wash, dry, anneal, and weigh
the gold.

Notes: a. The same procedure may be used for ^[-A.T., J^-A.T., or other
sample weights by properly proportioning the critical reagents. The
amount of mercuric nitrate suggested in step 2 is sufficient for 1 A.T. of pure
copper. The mercuric nitrate solution may be prepared by dissolving 25 g.
of mercury in 50 ml. of concentrated nitric acid and then diluting to 1 1.

b. The amount of chlorine required for the precipitation of silver and
mercury should be calculated stoichiornetrically, and a large excess avoided.

MATERIALS REQUIRING ACID TREATMENT 193

For example, a solution of sodium chloride containing 27.2 g. per liter will
precipitate 50 mg. of silver or 45.6 mg. of mercury per milliliter. Some
assayers use lead acetate with the salt, but this does not seem to be advan-
tageous. After precipitation of the silver a long period of settling is neces-
sary to ensure completion.

c. In step 5 the procedure of adding only part of the water as cold water,
followed by dilution with hot water saves time, but all the dilution may be
made with cold water if desired, provided the resultant solution is boiled
long enough to dissolve all copper sulfate before precipitating the silver and
mercury.

d. A crucible fusion may be used instead of the scorification procedure
outlined in steps 7 to 9.

e. Step 1 1 is merely a simplified method of making a combined slag and
cupel correction (see Chap. XIV for further details). If the silver and gold
content of the copper bullion is relatively low, this step may be omitted.

CHAPTER XI
ASSAY OF SOLUTIONS FOR GOLD AND SILVER

The testing and control of processes that make use of aqueous
solutions of gold and silver require frequent assays of these solu-
tions. Cyanide solutions are widely used for the recovery of
gold and silver from ores, and solution assays are regularly made
at cyanide mills. Industrial electroplating establishments also
use gold and silver cyanide solutions and occasionally require
assays. The use of thiosulfate ("hyposulfite") solutions for
the recovery of silver from ores has declined, but the hyposulfite
process is still used for treating a few unusual materials. 1 The
determination of silver in hyposulfite solutions is also encoun-
tered in connection with the recovery of silver from the hypo-
sulfite solutions used to fix motion-picture film. The chlorination
process for the recovery of gold from ores is no longer used, but
gold chloride solutions are used in electrolytic gold refining
(Wohlwill process).

The general methods used in assaying solutions for gold and
silver are :

1. Evaporation to dryness with subsequent recovery of gold
and silver from the residue by crucible assay, by scorification,
or by direct cupellation with lead.

2. Precipitation of gold and silver sufficiently pure to weigh
directly or in an impure form from which the gold and silver are
recovered by crucible assay, by scorification, by cupellation with
lead, or by colorimetric methods.

Evaporation, with crucible or scorification assay of the residue,
is suitable for the assay of any solution, but more rapid and con-
venient methods have been developed to suit individual solutions.

Solution Assay Portions. Assay portions of solutions are
measured by volume with a pipette or graduated cylinder. In
electroplating work the assay portion is measured in milliliters,

1 POCHON, M., Radium from the Canadian Arctic, Eng. Min. Jour., vol.
138, No. 9, p. 40, 1937; BABB, P. A., Refractory Patio Tailing Responds to
Leaching, Eng. Min. Jour., vol. 126, No. 21, p. 832, 1928.

194

ASSAY OF SOLUTIONS FOR GOLD AND SILVER 195

and the assay result is expressed in grams per liter. In cyanide
mills the assay portion is measured volumetrically in assay tons,
assuming that 1 ml. of solution weighs 1 g. Pipettes, calibrated
in assay tons, are available for measuring out assay-ton volumes
of solution.

The assay portion of a homogeneous solution need be only
large enough to produce a bead that can be weighed with the
desired accuracy. Electroplating solutions are rich, and a 10-ml.
assay portion of solution is ordinarily used. In cyanide mills the
solutions vary in grade from the pregnant solution about $3
per ton to the barren solution that contains very little gold or
silver. Assay portions of at least 30 assay tons are used for
barren-solution assays, at either gold or silver cyanide mills.
For pregnant-solution assays, assay portions of about 5 assay
tons are used at gold mills, but at silver mills the pregnant solu-
tion contains a larger amount of metal, and a 1-assay-ton portion
is sufficient.

ASSAY OF CYANIDE SOLUTIONS

The principal methods used for the assay of cyanide solutions
are:

1. Litharge evaporation.

2. Lead-tray evaporation.

3. Zinc-lead precipitation.

4. Cuprous chloride precipitation.

5. Lead-acid precipitation.

6. Electrolytic precipitation.

7. Sulfuric acid precipitation.

8. Colorimetric test for gold.

9. Colorimetric test for silver.

Litharge Evaporation. In the litharge-evaporation method an
assay portion of solution is evaporated to dryness in a porcelain
dish containing about 60 g. of litharge. The temperature should
be hold below 100C., to avoid loss by spattering or by baking
the residue. After evaporation the litharge containing the
residue is removed with a spatula and placed in an assay crucible.
Residue clinging to the dish may be wiped out with filter paper
that has been moistened with dilute acid. The filter paper is
then placed in the crucible with the litharge. Assay silica, soda,
and reducing agent are mixed with the material in the crucible,

196 FIRE ASSAYING

and the charge is fused to collect the gold and silver in a lead
button, from which the gold and silver are recovered and weighed
by the usual methods.

The litharge-evaporation method can be used for any solution,
but it requires an unusually long time, for which reason it is
seldom used.

Lead -tray Evaporation. Lead trays for the lead-tray evapora-
tion method are formed from square 4-in. sheets of lead foil,
which are conveniently cut from a 4-in. roll of assay lead foil.
A wooden block 2J^ by 2J^ in. is placed in the center of a sheet
of lead foil, and tray sides of about % in. are bent up around the
wooden form. Trays of this size will just hold 2 assay tons of
solution; assay portions of only 1 assay ton are generally used,
however, in order that the evaporation may be completed in a
short time.

Evaporation is carried out by placing the lead tray on a hot
plate and adding the assay portion of solution. If the plate is
hot at the start a 1 -assay-ton portion of solution can be evapo-
rated in 35 min. As soon as the solution has evaporated, the
trays are removed from the hot plate and allowed to cool for
easy handling. Each tray is then folded into a compact bundle,
care being taken in enclosing the evaporated residue in the bundle
that none be lost.

The bundle of lead foil containing the residue is frequently
cupeled directly as a lead button from an ore fusion. Spitting,
that is, the ejection of small droplets of lead into the air, is likely
to take place soon after cupellation starts. This is caused by
the action of the salts in the residue. Loss due to spitting is not
so serious as one might expect; nevertheless, it should be avoided
when high accuracy is desired. Spitting can be avoided by a
brief preliminary scorification.

For scorification, add 30 g. of granulated lead and a few grams
of borax glass to the lead-foil bundle, which has been put in a
scorifying dish. Scorify for a few minutes or until a quiet fusion
is formed and then pour and cupel the button as usual.

The lead-tray evaporation method is convenient and requires
but little of the operator's time. It is widely used for rich
solutions where an assay portion of 1 assay ton is sufficient.
Other methods are preferable for low-grade solutions that require
large assay portions.

ASSAY OF SOLUTIONS FOR GOLD AND SILVER 197

Zinc-lead Precipitation. In the zinc-lead precipitation (Chid-
dey) method, the gold and silver are precipitated in a lead sponge
that can be cupeled.

Place the assay portion of solution in a beaker, add 20 ml. of a
10 per cent lead acetate solution and from ]/^ to 1 g. of zinc dust.
Stir and heat nearly to boiling. Slowly add 30 ml. of dilute
HC1 (1:1) and continue heating for about % hr. or until the
cessation of bubbling indicates that the zinc has dissolved.
Decant the solution with care, that particles of lead may not be
allowed to become detached and lost. Wash the lead sponge
several times with water. Then press it into a ball with the
fingers. Wrap the collected lead in lead foil, leaving a vent for
the escape of steam, and cupel as usual.

If zinc is not entirely dissolved by the acid treatment, it burns
vigorously on cupellation, causing loss and forming a scoria in
the cupel, which may occlude globules of the precious metals.

The zinc-lead precipitation method is suitable for large or
small assay portions, and it is one of the most popular methods
for low-grade solutions. It gives good results with ordinary
solutions, but impure solutions, particularly those containing
copper, will not produce a coherent lead sponge. If the lead
sponge is not coherent, filtration is required to recover all the
precipitate. This makes the method inconvenient, as the filter
paper should then be burned by scorification before cupellation.
Fair results can be obtained by tearing off the clean top part of the
filter paper, wrapping the precipitate with the rest of the filter
paper in lead foil, and cupeling directly.

Cuprous Chloride Precipitation. Addition of an excess of
cuprous chloride to a cyanide solution containing gold and silver
causes the precipitation of gold and silver cyanide, together with
cuprous cyanide. The precipitate is finely divided, and, to
facilitate collection by filtration, a copious cuprous ferrocyanide
precipitate is produced by the addition of a small amount of
potassium ferrocyanide. An excess of potassium ferrocyanide,
over that which reacts with the remaining cuprous chloride, has
a solvent action on gold and must be carefully avoided. King
and Wolfe 1 have improved the method by substituting paper

1 KING, J. T., and WOLFE, S. E., An Improved Cuprous Chloride Method
of Assay for Gold in Cyanide Solutions, Canadian Min. Jour., vol. 59, p. 6,
January, 1938.

198 FIRE ASSAYING

pulp for the f errocyanide precipitate. The procedure given below
is substantially that recommended by King and Wolfe.

A stock cuprous chloride solution is prepared by placing copper
sulfate and strong hydrochloric acid (!CuSO4:3HCl) in a flask
with a considerable amount of copper wire. A reflux condenser
is attached to the flask, and the mixture is boiled gently until
the solution becomes clear. Then the condenser is detached, a
layer of liquid petrolatum is introduced to prevent oxidation
from the air, and a siphon is connected for dispensing.

Paper pulp for use in the process is prepared by moistening
shredded filter paper with concentrated hydrochloric acid. After
the paper has digested for 5 or 10 min., dilute with water and shake
to form a suspension. The paper pulp is used to coat an. ordinary
filter paper in a filter funnel. This is done by filtering well-
diluted paper pulp through the ordinary paper filter.

King and Wolfe recommend a stock flux of litharge 40 parts,
sodium carbonate 36 parts, silica 8 parts, borax glass 4 parts, and
flour 1 part for fusion of the cuprous cyanide precipitate. Only
a small amount of reducing agent is required, because of the
reducing effect of the filter paper that is introduced with the
precipitate.

The assay is started by placing the assay portion in a beaker
and adding 10 ml. of the completely reduced cuprous chloride
solution. This is sufficient for an assay portion up to 20 assay
tons. For every additional 10 assay tons in the assay portion,
the amount of cuprous chloride should be increased by 2 ml. The
solutions are mixed by stirring, and then a few minutes is allowed
for the precipitation to take place. The precipitate is collected
by filtration through a close-textured paper that has been coated
with paper pulp. All precipitate must be policed out of the
beaker and washed into the filter. After the solution has drained
through the filter, the paper and precipitate are lifted out of the
funnel and placed, apex down, in a 20-g. crucible containing about
10 g. of stock flux. It is then covered with 80 g. of stock flux.
The charge is fused, and the resulting lead button cupeled as usual.

Cuprous chloride precipitation is sometimes also brought about
by the addition of copper sulfate with sodium sulfide, which
reduces the cupric copper to the cuprous form.

The cuprous chloride precipitation method is used mainly
for gold solutions, where it gives equally good results on large or

ASSAY OF SOLUTIONS FOR GOLD AND SILVER 199

small assay portions. Copper already in solution does not cause
trouble. If much sulfocyanide is present, low results are
obtained. Consequently, either the zinc-lead or lead-acid pre-
cipitation method is preferred for solutions from the treatment
of sulfide ores.

Lead-acid Precipitation. The lead-acid precipitation method
was devised to avoid the difficulty encountered in the zinc-lead
precipitation method when copper is present in the solution.

The assay portion is placed in a beaker and heated nearly to
boiling. Then 8 to 10 g. of granulated lead are sprinkled over
the bottom of the beaker. The solution is acidified with HC1,
and an excess of about 10 ml. HC1 added. (Caution! Poisonous
HCN gas is formed upon acidifying a cyanide solution. The
operation should be carried out under a well-ventilated hood.)
A cover glass is placed on the beaker, and the solution boiled
for from 10 to 15 min. The longer period should be used for
large assay portions. The solution is decanted with care, to
avoid loss of solid particles, or it may be poured rapidly through
a coarse-textured filter paper. The remaining lead is washed
twice by decantation and then dried on the hot plate. After
drying, the lead cake is brushed onto a sheet of lead foil in which
it is wrapped for cupellation.

If the solution has been decanted through a filter paper,
the paper is examined for particles of lead. When particles
of lead can be seen, the parts of the filter paper containing
them are torn out and included in the lead-foil bundle bef ore
cupellation.

The lead-acid precipitation method is more convenient than the
zinc-lead precipitation method, particularly when the solutions
contain copper. It has given satisfactory results on solutions
from laboratory cyanide tests of many kinds of ores; 1 it has not,
however, as yet been tried with mill solutions.

Electrolytic Precipitation. The electrolytic precipitation
method of separating gold and silver from cyanide solutions is
rarely used for assaying. It has given good results at the
Kleinfontein Group Central Administration Assay Offices in
South Africa. The method is described by Crichton. 2

1 Unpublished work of O. C. Shepard and A. K. Schellinger.

2 CBICHTON, C., The Assay of Gold-bearing Cyanide Solutions by Elec-
trolysis, Jour. Chem. Met. Min. Soc. South Africa, vol. 12, p. 90, 1911.

200 FIRE ASSAYING

Ten-assay-ton portions of solution are placed in beakers that
are held in a frame. Anodes of iK6~ m - arc-light carbons are
clamped to a copper bar so that they can be lifted above the
beakers or lowered to a position that places one carbon vertically
in the center of each beaker. Cathodes consist of 2^ by 9-in.
sheets of lead foil. One edge of each sheet is coarsely serrated,
and at one end a strip Y in. wide is cut and bent up to connect
with the negative source of potential. The lead-foil sheets are
then formed into a cylinder around a bottle or wooden form 2)^
in. in diameter. The ends of each lead-foil sheet are folded
together to maintain the cylindrical shape after the form is
removed.

One lead-foil cylinder is placed in each beaker, and the J^-i n -
connecting strip is fastened to a copper bar carrying the negative
potential. The anodes are lowered into the beakers, and elec-
trolysis is allowed to proceed under a potential of 6 volts for a
period of 4 hr. With very dilute solutions some additional
cyanide should be added to lower the solution resistance, until
a current of about 0.1 amp. is produced.

An adherent deposit of metallic gold is obtained on the lead
cathode. At the end of the 4-hr, period the carbon anodes are
raised. The lead cathodes are then taken out immediately and
dried on a hot plate. After drying, the lead is folded into a
compact bundle and cupelled.

Electrolytic precipitation is slower than other precipita-
tion methods; it has an advantage, however, in that while pre-
cipitation is in progress the assayer is free to carry on other
duties.

Sulfuric Acid Precipitation. Sulfuric acid precipitation is
sometimes used for the determination of gold in electroplating
solutions, particularly where assay-furnace equipment is not
available. The procedure is described by Kushner. 1

An assay portion containing from 5 to 200 mg. of gold (10 ml.
of a plating solution is usually sufficient) is placed in a 500-ml.
Erlenmeyer flask with 50 ml. of water. A few drops of a dilute
potassium iodide solution are added, after which the silver nitrate
solution is run in from a burette until all the free cyanide is
combined with silver, and yellow silver iodide starts to precipitate.

1 KUSHNER, J. B., A Rapid Method for Gold in Cyanide Plating Solutions,
Ind. Eng. Chem., Anal. Ed., vol. 30, No. 11, p. 641, Nov. 15, 1938.

ASSAY OF SOLUTIONS FOR GOLD AND SILVER 201

The flask is then placed in a well-ventilated hood, and H 2 S04
is cautiously added until vigorous action ceases. Then 50 ml.
additional H 2 SO 4 is added, and the solution is boiled until the
precipitate of gold turns light brown and the solution becomes
clear.

The precipitate is allowed to settle, and the acid is decanted.
An additional 50 ml. of concentrated H 2 SO 4 is again boiled with
the gold and decanted as completely as possible. After cooling,
the remaining acid is carefully diluted with 200 ml. H 2 O, and the
gold is recovered by filtration through an asbestos mat in a tared
Gooch crucible. The gold is washed, dried, and weighed in the
Gooch crucible.

Colorimetric Methods. Colorimetric methods are widely used
for testing barren solution from cyanide-mill precipitation.
Tests that can be quickly carried out and that indicate the
presence of even a small amount of gold or silver are desired.

Colorimetric Test for Silver. A portion of barren solution is
placed in a test tube or beaker, and a few milliliters of a 10 per
cent solution of sodium or potassium sulfide is added. In the
absence of silver, only a white precipitate of zinc sulfide forms.
When silver is present, dark silver sulfide precipitates. The
amount of darkening of the solution indicates roughly the amount
of silver present.

At cyanide mills treating silver ores, the barren solution is
tested every few hours by this method, as a check on the precip-
itation operation. Precipitation is normally so perfect that no
silver can be detected by the test.

Colorimetric Test for Gold. One liter of barren solution is
placed in a bottle or large beaker and treated with 2 drops of a
saturated solution of lead acetate and J^ g. of sodium sulfite.
About 2 g. of zinc dust are stirred into the solution. The zinc
dust should be kept suspended in the solution by swirling or
stirring for 2 min. The precipitate is then allowed to settle,
and most of the solution is removed by decantation. The
precipitate is washed into a small beaker, and the wash solution
is decanted. The precipitated gold is dissolved by heating with
10 ml. aqua regia (3HCl:lHNOs) and evaporated nearly to dry-
ness. The residue is taken up in 2 ml. HC1 and transferred to a
small test tube. After cooling, a few drops of saturated stannous
chloride solution are addled.

202 FIRE ASSAYING

The presence of gold is indicated by a purplish ring at the con-
tact of the solutions or by a purplish tinge throughout, if the
solutions are mixed. As little as 0.001 oz. of gold per ton of
solution can be seen plainly. Some operators become so profi-
cient in the colorimetric test that they can estimate with con-
siderable accuracy the amount of gold in solution.

ASSAY OF T HIOSULFATE SOLUTIONS

Thiosulfate leaching solutions are used only for the extraction
of silver, but in treating ores a small amount of gold may also
be dissolved. When both gold and silver are to be determined,
the assay is best carried out by one of the evaporation methods
as described for cyanide solutions.

Silver can be determined more rapidly in thiosulfate solutions
by sodium sulfide precipitation than by evaporation, particu-
larly when a large assay portion of solution is used for the assay.
Sodium sulfide precipitation is not suitable for determining gold,
as an excess of sodium sulfide holds some gold in solution.

In the sodium sulfide precipitation method the assay portion
of solution is treated with a solution of sodium sulfide until no
more precipitate forms. The silver sulfide precipitate is sepa-
rated by filtration through a close-textured filter paper. After
the solution has drained from the filter the paper containing the
precipitate is removed from the filter funnel and placed in a
scorifying dish on top of 40 g. of granulated lead and a few grams
of borax glass. A little granulated lead is sprinkled on top
of the paper, and the scorifier is placed at the front of a hot assay
muffle where the filter paper will slowly burn. After the filter
paper has burned, a short period of scorification is carried out as
usual, and the resulting lead button is cupeled to recover the
silver.

ASSAY OF GOLD CHLORIDE SOLUTIONS

Gold chloride solutions can be assayed by the lead-acid method
as described for cyanide solutions, by a modified evaporation
method, or by precipitation.

If the gold chloride solution is nearly neutral, either of the
evaporation methods described for cyanide solutions may be
used. Acid solutions should be evaporated with assay silica
containing charcoal, which will precipitate the gold from the

ASSAY OF SOLUTIONS FOR GOLD AND SILVER 203

chloride compound. The dry residue is mixed with litharge and
soda and then treated as an ordinary crucible assay.

Gold is easily precipitated from chloride solutions, and many
different methods of precipitation could be used. The following
scheme has been used for assaying solutions from the chlorination
process :

The solution to be assayed is acidified with a few drops of HC1.
An excess of a ferrous sulfate solution is added and the mixture
is stirred, warmed, and allowed to stand for 1 hr. A few addi-
tional drops of ferrous sulfate solution may be added to test for
complete precipitation. The gold precipitate is collected on a
paper filter, which is subsequently thoroughly washed with
water. After filtration, the paper is dried and burned over a
sheet of lead foil in which the ash containing the gold is wrapped
for cupellatioii. Ordinarily, the cupeled bead is free from silver
and may be weighed directly for gold.

CHAPTER XII
THE FIRE ASSAY FOR THE PLATINUM METALS

The platinum group of metals consists of platinum (Pt),
palladium (Pd), indium Ir), rhodium (Rh), osmium (Os), and
ruthenium (Ru). Platinum is the most important metal of the
group and occurs in greatest quantity. The other metals of
the group are found associated with platinum arid are produced
as valuable by-products of platinum production.

Many copper and gold ores contain traces of platinum metals,
and small amounts are usually recovered in refining copper and
lead bullion. The main production of platinum metals comes
from (1) the Sudbury nickel-copper deposit of Ontario, Canada;
(2) the placer deposits of Colombia, the Ural Mountains, and
Alaska; and (3) the igneous deposits of the Lydenberg, Potgie-
tersrust, and Rustenburg districts of South Africa.

In placer deposits the platinum metals occur as water-worn
grains and flakes of native metallic alloys, which frequently are
associated with placer gold. The most abundant of the platinum
grains contain from 60 to 90 per cent platinum alloyed with a small
amount of the other platinum metals and also with gold, copper,
and iron. A small amount of the grains found in placer deposits
consists mainly of an alloy of osmium and iridium, called "osmi-
ridium" or "iridosmine." Ruthenium, the least abundant of the
platinum metals, comes mainly from osmiridium, which fre-
quently contains about 5 per cent of ruthenium.

In primary deposits, platinum metals are found as native
metals and as various compounds such as sperrylite (PtAs2),
cooperite [Pt(AsS) 2 ], and stibiopalladinite, containing mostly
antimony and palladium. Usually in primary deposits, as in
placer deposits, platinum is the predominant metal of the plati-
num group. In the primary sulfide ores of South Africa, how-
ever, palladium is sometimes found in greater abundance than is
platinum. The platinum metals other than platinum and
palladium are present only in relatively small amounts.

204

THE FIRE ASSAY FOR THE PLATINUM METALS 205

Determinations of the platinum metals are required of (1)
ores and concentrates from primary deposits, (2) concentrates
and bullion from placer deposits, (3) platiniferous products from
the refining of copper and lead bullions, and (4) purified metals
and alloys of the platinum metals. Fire assay methods are used
to collect the platinum metals from ores, concentrates, and other
low-grade materials into a rich bead, which is used for the
separation and determination of the individual metals by wet
chemical methods. Platinum and palladium are usually present
in so much greater proportion than the other platinum metals
that 20 mg. or more of platinum metals must be used to obtain
weighable amounts of the rarer platinum metals. The separation
and individual determination of the metals are long and expensive
operations, which can be avoided on most routine mine samples
by determining only the total of platinum metals and gold.

The proportion between the valuable metals can be determined
by occasional complete assays of the collected beads. Separate
determinations should be made on beads from different parts
of a deposit, because the ratio between the platinum metals
may not be uniformly constant. In the Rustenburg district, for
example, the ratio of palladium to platinum is lower in the oxi-
dized ore than in the sulfide ore. 1

Fusions for the Platinum Metals. The platinum metals
present are collected in the lead button of an ordinary fire assay
crucible fusion or scorification. When the assay is made particu-
larly for the platinum metals, the fusion is customarily finished at
a higher temperature than generally used for gold or silver work.
A finishing temperature of 1200C. is used for crucible fusions
by Seath and Beamish, 2 who found no advantage in a higher
temperature or in heating the charge for an unusually long time.

Crucible charges for platinum-bearing materials should be
proportioned according to the material to be fluxed and accord-
ing to the oxidizing or reducing effect of the ore, in exactly the
same manner as for a gold or silver ore. Adam 3 found that when

1 TYLER, PAUL M., and SANTMEYERS, R. M., Platinum, U.S. Bur. Mines
I.C., 6389, p. 56, 1931.

2 SEATH, J., and BEAMISH, F. E., Assay for Platinum Metals in Ore Con-
centrates, Ind. Eng. Chem., Anal Ed., vol. 12, No. 3, p. 169, 1940.

8 ADAM, H. R., Determination of the Platinum Metals in Ore and Con-
centrates, Jour. Chem., Met., Min. Soc. South Africa, vol. 29, p. 106, 1928.

206 FIRE ASSAYING

good fusions are obtained, no serious slag loss occurs, but that,
when poor fusions are obtained or when very rich samples are
assayed, platinum metals can be recovered from the slag. A
considerable slag loss of platinum metals was encountered by
Seath and Beamish 1 in crucible fusions of a roasted platinum
concentrate that contained copper, iron, and nickel oxides. They
attributed the high slag loss to the presence of nickel oxide
(18.6 per cent), and used a retreatment of the slag with extra
litharge and flour to collect the lost platinum metals.

The lead buttons from fusions of a material containing con-
siderable copper or nickel are likely to contain sufficient impurities
to interfere with cupellation. In order to avoid this difficulty the
lead buttons should be purified by melting and swirling with a
small amount of litharge or with an excess litharge slag. After
swirling, the charge should be allowed to stand molten for a
short time, to allow settlement and collection of osmiridium, which
does not alloy with lead.

Osmium, iridium, and ruthenium do not alloy with lead, but
are "wetted" by it in an assay fusion, and because of their high
density they sink into the lead and become mechanically held
in the button. Davis 2 recommends cooling a fusion in the cruci-
ble, instead of pouring it into a mold, when platinum metals are
to be determined. After the lead and slag have solidified the
crucible can be broken, and the solid button separated from the
slag and crucible with less danger of loss of the loosely held
metals. There is no particular danger of loss of platinum,
palladium, or rhodium during fusions, as these platinum metals
alloy with lead and are collected the same as gold and silver.

Cupellation of Buttons Containing Platinum Metals. Cupella-
tion of a lead button containing more than fifteen times as much
gold and silver as platinum metals proceeds normally. Platinum,
palladium, and rhodium alloy with the gold and silver and are
satisfactorily recovered in the final bead. Platinum causes a
roughening of the surface of silver beads, which gives them a
characteristic frosted appearance.

Iridium does not alloy with gold or silver; when present during
cupellation it forms a black deposit that clings to the bottom of
the bead.

1 Op. ait.

2 DAVIS, C. W., The Detection and Estimation of Platinum in Ores, U.S.
Bur. Mines, Tech. Paper 270, 1921.

THE F1KE ASSAY FOR THE PLATINUM METALS 207

Osmium and ruthenium form volatile oxides when heated in
air. Consequently they suffer a large loss during cupellation.
Remaining ruthenium dioxide forms a blue-black coating on the
bead. When an accurate determination of osmium and ruthe-
nium is required, the lead button should be treated by wet
chemical methods without cupellation.

Cupellation of lead containing platinum metals without a large
proportion of gold or silver will not proceed to completion at
ordinary cupellation temperatures. * The high melting points of
platinum-rich alloys cause the alloy to solidify when the lead has
been reduced to form 40 to 60 per cent of the bead. Unabsorbed
litharge, characteristic of the usual frozen cupel, is not exhibited
with a finishing temperature of about 900C, but a bead that
solidifies with considerable lead, owing to its high platinum metal
content, will not " brighten" and can usually be recognized by its
dull and flattened appearance.

Many platinum-bearing placer concentrates contain sufficient
gold for successful cupellation, but some placer concentrates
and most platinum ores are deficient in gold and contain a
negligible amount of silver. Cupellation of platinum-bearing
lead buttons is carried out by :

1 Adding from ten to fifteen times as much silver as platinum
metals and finishing at 900C.

2. Adding about twenty times as much gold as platinum metals
and finishing at 900C.

3. Cupeling without the addition of gold or silver, and then
giving the platinum-frozen bead a post treatment at 1300C.
for 1 hr.

When either the gold-addition or the post-treatment cupella-
tion method is used on silver-free materials the weight of platinum
metals plus gold can be obtained by weighing the bead directly
(deducting the weight of gold added in the gold-addition method).
Direct weighing is not feasible for beads from the silver-addition
cupellation method, because of the considerable loss of silver
during cupellation. The silver-addition method of cupellation
is usually preferred when the platinum metals are to be separated
and accurately determined. A modified sulfuric acid parting
method can be used to make a preliminary separation of silver,
in order to obtain quickly an approximate value for the platinum
metals plus gold.

208 FIRE ASSAYING

The post-treatment method of cupellation has been described
by Adam. 1 It developed from a method used by O'Neill, at
Potgietersrust, of adding neither gold nor silver to aid cupellation
but of finishing cupellation at the highest temperature obtainable
in his coal-fired muffle. The weight of beads from this treatment
was used as a rough value for gold and the platinum metals.
Correction factors were developed for the metals in the bead
from periodic analyses of the collected beads. This plan worked
reasonably well, and proved particularly attractive for assaying
materials containing a large proportion of palladium. The
Government Areas Laboratory in South Africa improved the
method by cupeling at about the temperature previously used
for silver-addition cupellation and then transferring the beads
to small depressions in a bone-ash cupel for post treatment in
an electric furnace at 1300C. for 1 hr.

Adam 2 gives 6.69 dwt. per ton for the average of 110 assays by
silver-addition cupellation and modified parting, while the post-
treatment cupellation results on the same ores averaged 6.55 dwt.
per ton. Analysis of the post-treatment beads gave 91.2 per cent
precious metals. Seath and Beamish 3 found that, by introducing
oxygen into the post-treatment furnace, lead was eliminated more
rapidly and the time of treatment could be reduced to J^ hr. with
a consequent reduction in the loss of platinum metals.

By the post-treatment method, lead can be almost completely
eliminated from beads high in platinum. Elimination of lead
is less complete when the bead contains a considerable proportion
of other platinum-group metals. Seath and Beamish report that
the presence of a considerable proportion of rhodium or iridium
causes the bead to lose coherence, which increases the loss of
metals in the process. It seems likely that osmium and ruthe-
nium are completely converted to their volatile oxides and lost
in the post-treatment process.

Parting Beads Containing Platinum Metals. The nitric acid
parting treatment, given dor6 silver beads for the assay of gold
and silver, will dissolve palladium and part of the platinum that
may be present. To avoid reporting these metals as silver, the
spent acid from parting ordinary gold and silver assays should be

1 ADAM, op. cit., p. 108.

2 Ibid.

3 SEATH and BEAMISH, loc. cit.

THE FIHE ASSAY FOR THE PLATINUM METALS 209

watched for indications of platinum and palladium. A small
amount of palladium produces a distinct yellow color in the acid,
and the color becomes darker with increase in the amount of
palladium. Platinum dissolves only when the bead contains at
least ten times as much silver as platinum, and the amount that
dissolves increases with increased proportion of silver. A con-
siderable amount of platinum gives the solution a dark-brown
color like a colloidal solution of carbon.

Rhodium, iridium, osmiridium, and part of any platinum
present in a dore* bead remain with the gold and give it a "not
completely parted" appearance. The monetary error in report-
ing gold for rhodium and iridium is not so great as that in report-
ing silver for platinum; nevertheless the presence of platinum
metals in ores should be detected and reported, so that the ore-
treatment process may be adapted to their recovery.

Assay beads from platiniferous materials contain the gold,
platinum metals, and silver from the ore, together with a small
amount of lead, which is usually left from cupellation, and any
gold or silver added to aid cupellation. The quantitative sepa-
ration of these metals is a long and tedious process, requiring
reagents not available in the ordinary assay office. Various
methods of parting in acids have been devised for treating assay
beads containing metals of the platinum group, in an effort to
determine the most important metals by a simple process and
with reagents ordinarily on hand.

Prior to 1920, practically all the platinum-metal production
came from placer deposits in which gold and platinum were of
predominant value. The determination of these metals was
considered important, and the other platinum-group metals
were neglected except by specialists at refineries and platinum
works. A parting method for the determination of gold and
platinum was developed and was commonly used by commercial
assayers.

In the parting method for gold and platinum the material to
be assayed was put through an assay fusion to which silver, for
silver-addition cupellation, had been added. After cupellation
the silver-platinum-gold bead was flattened and parted in boiling
or nearly boiling sulfuric acid for a period of about 30 min.
This treatment dissolved silver and most of the palladium present
in the bead. The residue was weighed and recorded as gold plus

210 FIRE ASSAYING

platinum. It was then inquarted with silver equal to ten times
its weight and parted in nitric acid to dissolve platinum. Usu-
ally the inquarting and parting had to be repeated at least three
times to remove all the platinum. The nitric acid residue was
weighed and recorded as gold. Platinum was reported equal to
the difference in weight between the sulfuric acid residue and the
nitric acid residue.

The parting method for the separate determination of gold and
platinum gives only approximate results. It furthermore has
the serious fault of overlooking palladium. Complicated elabora-
tions of the method have been devised, both to improve its
accuracy and to allow for the presence of other platinum group
metals, but other methods are better adapted for the determina-
tion of the separate platinum metals. For the rapid approximate
determination of the value of an ore, methods are available that
lump together gold and the platinum metals including palladium
and require less time than the parting method for the separation
of platinum and gold.

In the absence of silver the sum of the platinum metals and
gold is given by the weight of the bead from either the gold-
addition or the post- treatment method of cupellation. When the
material to be assayed contains silver, or when silver is added to
aid cupellation, modified sulfuric acid parting should be used
in order to dissolve silver without palladium. 1

Modified sulfuric acid parting consists of parting the bead
(which should contain from 10 to 15 times as much silver as
insoluble metals) in hot, but not boiling, slightly diluted sulfuric
acid. The innovation of the modified process consists of remov-
ing the flask or parting cup from the source of heat as soon as the
silver is dissolved, as indicated by the termination of bubbling.
After the acid has cooled somewhat by standing for 10 or 15
min. it is carefully decanted from the residue by pouring down a
glass rod into a pyrcx beaker. The beaker should be clean and
fshould be placed on white paper so that loss of particles of residue
can be readily detected.

After decanting the strong acid the parting vessel should be
allowed to cool before adding distilled water for washing. The
first small amount of wash water is better if cold and must be

1 GRAHAM, K. L., Platinum Assays, South African Min. Enq, Jour., vol.
38, p. 57, Mar. 19, 1927.

THE FIRE ASSAY FOR THE PLATINUM METALS 211

very carefully added, as violent spattering is caused by adding
water to hot concentrated sulfuric acid. As soon as the acid
that has remained in the cup has been diluted with several vol-
umes of water the danger of spattering is past and hot wash water
can be used. The parting vessel is then filled about half full of
wash water and heated nearly to boiling. Care must be taken
to decant the hot wash water into a beaker separate from that
containing the strong parting acid. Usually the residue contains
some lead sulfate, which can be removed by treatment with a
hot, strong solution of ammonium acetate followed by three
washes with hot water. The washed residue is dried in air at
50 to 60C. and weighed for platinum metals and gold.

A small amount of palladium is dissolved by the modified
parting treatment, but this loss is usually overbalanced by the
retention of silver in the residue. The amount of palladium lost,
and the amount of silver retained, depend upon the size of the
bead and the proportion of metals in the bead, as well as upon the
temperature and time of acid treatment. It is advisable to
obtain correction factors by means of check assays, made with
known amounts of the platinum metals in approximately the pro-
portion they occur in the ore, or by means of occasional analysis
of the residue and parting acids.

THE DETERMINATION OF GOLD, SILVER, AND METALS OF THE
PLATINUM GROUP

Many methods of analysis have been proposed for the separa-
tion and determination of gold, silver, and the individual plati-
num metals. The treatment given here represents an attempt
to combine selected parts of the various processes into a system-
atic general scheme. It can be simplified when the determina-
tion of only a few of the metals is required or when negligible
amounts of some of the metals are present.

The determination of platinum, palladium, rhodium, iridium,
and gold is commonly required for platiniferous materials.
Either silver-addition or post-treatment cupellation beads are
used for this purpose. When silver is present it must be sepa-
rated, but on account of the loss of silver in high-temperature
cupellation its determination is best made on the lead button
from a separate fusion. Osmium and ruthenium should also

212 FIRE ASSAYING

be determined without cupellation, and their determinations
may be made on the lead button used for the silver determination.

Determinations of osmium and ruthenium are seldom required,
since these metals are of relatively little economic importance.
In ordinary analytical methods they separate in the acid-insoluble
residue that is usually attacked by a fusion with alkaline oxidizing
fluxes. Sodium osmate (Na2Os()4) and sodium ruthenate
(Na 2 Ru0 4 ) are formed and can be dissolved in water. When
considerable iridium and rhodium are present, the fusion may
have to be repeated to extract the osmium and ruthenium com-
pletely. From solution, osmium and ruthenium are nearly always
separated from other metals and from each other by distillation
of their volatile tetr oxides. Gilchrist has investigated the
conditions of distillation, absorption, and determination of
osmium 1 and ruthenium. 2 His methods of analysis for these
metals are substantially followed in the procedure given below.

Procedure to Determine Silver, Osmium, and Ruthenium.
For the determination of silver, osmium, and ruthenium in ores
and concentrates, collect the precious metals in a lead button
by a crucible fusion. If the material to be assayed is low in
these metals, use 2-A.T. samples or combine by scorification
the buttons from several 1-A.T. fusions. Roll the final lead
button into a thin strip with the bullion rolls, so that its rate of
solution will be rapid. Then proceed with the analytical method
outlined in Fig. 14 and described below.

Place the lead strip in a 250-ml. beaker and boil with dilute
HNOa to dissolve lead, silver, palladium, and platinum. Filter
through a small asbestos pad in a Gooch crucible and wash with
hot H 2 O. The filtrate is used for silver and the residue for
osmium and ruthenium.

Silver. Add a small amount of NaCl to the HNO 3 solution,
to precipitate AgCl and some PbCl 2 . Cool, filter, and wash with
a hot NaCl solution to remove lead. Wash once with H^O and
then sprinkle granulated lead in the filter paper. Place the paper
in a scorifying dish, cover with about 40 g. of granulated lead

1 GILCHRIST, R., A Method for the Separation and Gravimetric Deter-
mination of Osmium, Bur. Standards Jour. Research, vol. 6, p. 421, 1931.

2 GILCHRIST, R., A Method for the Separation of Ruthenium from Plati-
num, Palladium, Rhodium, and Iridium, Bur. Standards Jour. Research,
vol. 12, p. 283, 1934.

THE FIRE ASSAY FOR THE PLATINUM METALS 213

and add about 1 g. of borax glass. Scorify a few minutes, pour,
and cupel the button at a temperature suitable for silver.

Osmium. Osmium and ruthenium are left in the insoluble
residue from the HN0 3 treatment of the lead button. The

Lead Button- from Assay Fusion

Roll into Strip
Nitric Acid Treatment
Gooch Filter

Residue
Au.Os.Ru.Rh.lr
Sodium Peroxide Fusion
HN0 3 Distillation

Solu

Vapor Os0 4
Absorb HC1-S0 2 Solution

Hydrolytic Precipitation
NaHC0 3 topH6

GoochVilter

Filtrate
Pb,Ag,Pt,Pd

Chloride Precipitation of Silver
Filter and Wash Hot Nad Solution

Filtrate Discard Precipitate AgCl
Scorify
Cupel

Weigh Silver

Filtrate Discard

Eliminate HN0 3

Fume H 2 S0 4
Add Sodium Bromate

Precipitate Hydroitcd
Osmium Oxide

Impregnate NH 4 CI

Ignite in Hydrogen

Weigh Osmium

Distill

Solution
Discard

Vapor RuC*4

Absorb HCt-S0 2
Solution

HydrolyKc Precipitation
NaHCOjtopHe

Gooch Filter

Filtrate
Discard

Precipitate Hydrwted
Ruthenium Oxide

Ignite in Hydrogen
Weigh Ruthenium

FIG. 14. Outline of the method for the determination of silver, osmium, and

ruthenium.

asbestos pad containing the residue is picked out from the Gooch
and mixed with sodium hydroxide and sodium peroxide (3NaOH :-
lNa 2 O 2 ) in an iron, nickel, or silver crucible. The material is
then fused at red heat, cooled, and leached with 100 ml. H 2 O
in a pyrex beaker. Osmium is nearly all dissolved by this treat-
ment while rhodium, iridium, and some ruthenium may remain

214 FIRE ASSAYING

insoluble. Filter out the residue for retreatment and add the
solution to a distilling flask attached by glass tubes and ground-
glass joints to a series of three absorption flasks. The last flask
should be attached to suction, so that a current of air can be
bubbled through the distilling flask and the vapor bubbled
through the liquid in the absorption flasks. Rubber connections
and stop-cock lubrication are avoided, since organic compounds
cause osmium dioxide to deposit from the tetroxide vapor.
Add dilute hydrochloric acid solution (1HC1:1H 2 0) saturated
with SO 2 to the absorption flasks, start a slow current of air
through the train and add 100 ml. H 2 O and 30 ml. HN0 3 to
the distilling flask. Boil gently for 1 hr. During this time the
osmium is carried over into the absorption flasks. Save the
liquid in the distilling flask for the determination of ruthenium
and combine the liquid from the absorption flasks.

Evaporate the liquid from the absorption flasks on the steam
plate four times with HC1, that the sulfite compounds of osmium
may be decomposed. Then add 150 ml. H 2 O, heat to boiling,
add a few drops of bromophenol blue indicator and add slowly
a solution of NaHCO 3 until a faint blue color (pH 6.5) is reached.
Boil for 5 min., filter through a weighed Gooch crucible and wash
with a hot 1 per cent NH 4 C1 solution. Impregnate the residue
with a saturated NHUCl solution, ignite in hydrogen for 10 min.,
cool, displace hydrogen with CO 2 , and weigh as metallic osmium.

Ruthenium. The HNO 3 solution left in the distilling flask,
after the separation of osmium, is evaporated several times on a
steam bath with HC1 to eliminate HNO 3 . Add 20 ml. dilute
sulfuric acid (1H 2 SO 4 :1H 2 O) and fume. Cool, carefully wash
the solution into the distilling flask, and add 100 ml. H 2 O.
Place dilute HC1 saturated with SO 2 in the absorption flasks
and start a slow current of air through the system by suction.
Add 100 ml. of a 10 per cent solution of NaBrO 3 to the distilling
flask, and boil gently for 2 hr. Then combine the liquid from
the absorption flasks and evaporate to a moist residue on a
steam bath. Digest with a small amount of HC1, dilute with
50 ml. H 2 0, and boil until the residue is dissolved. Dilute and
neutralize the solution with NaHC0 3 to cause the hydrolytic
precipitation of ruthenium in a manner similar to that used for
osmium. Filter through a weighed Gooch crucible and wash
with a hot 1 per cent (NH 4)280 4 solution. Ignite, reduce in

THE FIRE ASSAY FOR THE PLATINUM METALS 215

hydrogen, wash out soluble salts with hot H 2 O, dry, and weigh
metallic ruthenium.

Procedure to Determine Platinum, Palladium, Rhodium,
Iridium, and Gold. The determination of platinum, palladium,
rhodium, indium, and gold in ores and concentrates is made by
collecting the precious metals in a lead button that is cupeled by
either the silver-addition or the post-treatment method. Trie
bead of precious metals is then analyzed by wet chemical meth-
ods. In Fig. 15 a method of analysis suitable for silver-addition
cupellation beads is outlined. This scheme closely follows that
recommended by Beamish and Scott 1 for silver beads. When
post-treatment cupellation has been used, sulfuric acid parting
may be omitted and the treatment started with aqua regia.

Separation of Silver and Recovery of Palladium from Sulfuric
Acid Parting Solution. Do not clean or flatten the cupeled bead
but part it directly in 30 to 40 ml. of hot, slightly diluted EUSO^
Keep the temperature below boiling and remove the parting
vessel from the heat as soon as bubbling ceases. Let cool for
15 min. and then pour the acid into a beaker containing 150 ml.
cold H 2 O. Filter through a tight paper and wash with hot
H2O. Set aside the filtrate for the recovery of the small amount
of palladium that dissolves with the silver and wash the residue
with an ammonium acetate solution and hot H 2 O to remove lead.
Discard the washings.

Precipitate palladium from the filtrate by neutralizing the
solution carefully to a pH of 6 with sodium bicarbonate in the
presence of sodium bromate. 2 Boil to coagulate the brown
palladium dioxide precipitate, and filter. Wash with hot dis-
tilled H 2 O with pH of 6. Discard the filtrate and wash the
palladium dioxide through the filter with hot dilute hydrochloric
acid (1HC1:1H 2 0). Dilute the palladium solution that has
passed the filter so that it contains only about 3 per cent by
volume of HC1. Cool and add a 1 per cent solution of dimethyl-
glyoximc in 95 per cent ethyl alcohol (1 ml. of dimethylglyoxime
solution is required for every 10 mg. of palladium). Let stand

1 BEAMISH, F. E., and SCOTT, M., Analysis of Platinum Metals Silver
Assay Bead, Ind. Eng. Chem., Anal Ed., vol. 9, No. 10, p. 460, 1937.

2 GILCHRIST, R., Methods for the Separation of Platinum, Palladium,
Rhodium, and Iridium from One Another and for Their Gravimetric Deter-
mination, Bur. Standards Jour. Research, vol. 12, p. 296, 1934.

216

FIRE ASSAYING

. E 3
g.a-

SZLfc-

!" c

s'3

4 g

"S
"E

U^JJ!

THE FIRE ASSAY FOR THE PLATINUM METALS 217

for 1 hr. and filter on a weighed Gooch crucible. Wash with
cold, followed by hot, H 2 0. Dry at 110C. and weigh. The
weight of palladium glyoxime multiplied by 0.3167 gives the
weight of palladium. The palladium recovered here is only a
small part of that present in the bead. The remainder of the
palladium is dissolved by aqua regia parting and is precipitated
by dimethylglyoxime after the separation of gold.

Aqua Regia Parting. Place the residue from H 2 SO 4 parting
or the cupeled bead, if post-treatment cupellation was used, in a
beaker and digest for 2 hr. on a steam bath with 30 ml. of aqua
regia. Filter and wash with hot H 2 0. The residue contains
rhodium and iridium and is reserved for their determination.

Evaporate the aqua regia solution containing platinum,
palladium, and gold nearly to dryness three times with HC1 on a
steam bath, to eliminate HNOs. Then digest with 25 ml. of
hot H 2 and filter to remove silver chloride. Wash and discard
the residue. The filtrate is treated to recover gold, palladium,
and platinum.

Hydroquinone Precipitation of Gold. The solution from aqua
regia parting and the subsequent silver chloride separation
amount to about 50 ml. Add 5 ml. concentrated HC1 and a
solution of hydroquinone. The weight of hydroquinorie used
should be about equal to the amount of gold to be precipitated.
Boil 5 min., filter, and wash the precipitate. Set the filtrate
aside for the precipitation of palladium. Ignite the gold precipi-
tate and weigh metallic gold.

Dimethylglyoxime Precipitation of Palladium. Dilute the
filtrate from the gold precipitation to 350 ml. Add 5 ml. HC1
and a 1 per cent alcoholic solution of dimethylglyoxime (1 ml. for
every 10 mg. of palladium). Let stand for 1 hr., filter on a
weighed Gooch crucible, and wash the precipitate with cold and
then hot H 2 O. Reserve the filtrate for the sodium formate
precipitation of platinum. Dry the precipitate of palladium
glyoxime at 110C. and weigh. Multiply the weight of palladium
glyoxime by 0.3167 to obtain the weight of palladium. If H 2 S0 4
parting was used, combine the weight of palladium recovered
here with that from the H 2 SC>4 parting solution.

Sodium Formate Precipitation of Platinum. Evaporate the
filtrate from palladium precipitation to a sirup on the steam bath.
Dilute to 100 ml. Add 3 g. of sodium acetate and 1 ml. of formic

218 FIRE ASSAYING

acid. Heat nearly to boiling for 3 hr. Filter and wash free from
chlorides. The filtrate may be given an additional sodium
formate treatment, to be sure that precipitation is complete.
Place the filter paper with the platinum precipitate in a porcelain
crucible. Ignite and weigh metallic platinum.

Treatment of the Aqua Regia Residue. The residue from aqua
regia parting contains rhodium and iridium. Place it in a silver
crucible with about 1.5 g. of sodium peroxide and fuse just below
the melting point of silver for 5 or 10 min. Cool, and leach the
melt in a beaker with 100 ml. H^O. Wash out the crucible with
a small amount of dilute HNO 3 . Evaporate to about 10 ml. on
a steam bath, add 10 ml. of 1:1H 2 SO4 arid heat to fumes.
Cover the beaker and swirl over a burner until the condensing
sulfuric acid cleans the sides of the beaker. Cool, add 100 ml.
of H2(), and boil. (There should be no insoluble material
remaining at this point.) Neutralize the solution with NaHCOs
to exactly pH 6, add 10 ml. of a 10 per cent solution of sodium
bromate, and boil for the hydrolytic precipitation of rhodium
and iridium. Filter through an asbestos mat in a Gooch crucible
and wash with hot H 2 O of pH 6. Discard the filtrate. Dissolve
the precipitated dioxides of rhodium and iridium by digesting
on a steam bath with dilute HC1 (1HC1:1H 2 O). Dilute and filter-
to remove asbestos and silver chloride. The filtrate is treated
by the methods described by Gilchrist 1 for the determination of
rhodium and iridium. A blank determination should be made,
starting with the sodium peroxide fusion, to correct for metals
such as copper, which are picked up from the silver crucible.

Determination of Rhodium and Iridium. Dilute the HC1 solu-
tion containing rhodium and iridium to exactly 100 ml. and
pipette out 50 ml. for the determination of the sum of rhodium
and iridium from one-half the sample. Reserve the remaining
50 ml. of solution for the determination of rhodium by titanous
chloride precipitation.

Precipitate rhodium and iridium on the first 50-ml. portion of
solution by the hydrolytic process with NaHCO 3 and sodium
bromate at pH 6, as before. Filter, ignite the residue, reduce
in hydrogen and weigh one-half of the total iridium and
rhodium.

1 GILCHRIST, R., A Method for the Separation of Rhodium from Iridium,
Bur. Standards Jour. Research, vol. 9, p. 547, 1932.

THE FIRE ASSAY FOR THE PLATINUM METALS 219

Determination of Rhodium. Add 10 ml. H 2 S0 4 and a few
milliliters of HNOa to the 50-ml. portion of solution reserved
for the rhodium determination. Heat to fumes, cool, add 20 ml.
H 2 0, and again fume. Cool, dilute to 200 ml., and boil. Care-
fully add drop by drop a 20 per cent solution of titanous chloride
until the solution assumes a slight purple tinge. Boil, filter,
and wash with a cold 2 per cent H 2 SO4 solution. Place paper
and precipitate in a 500-ml. Erlenmeyer flask, add 10 ml. H 2 SO4,
and heat to fumes. Destroy organic matter with fuming HNOg.
Cool, dilute, and repeat the precipitation and solution. Add
20 ml. H 2 O, 10 ml. HC1, and boil for 15 min. Dilute to 500 ml.
and precipitate rhodium with H 2 S passed through the boiling
solution. Filter and wash with a 2 per cent H 2 S04 solution.
Ignite, reduce in hydrogen, and weigh metallic rhodium from
one-half the sample.

Iridium is found by subtracting the rhodium result from the
sum of the rhodium and iridium previously determined.

CHAPTER XIII ' v
FIRE ASSAY METHODS FOR BASE METALS

Although in modern practice, fire assaying is almost exclu-
sively confined to the assay of precious metals, approximate
assays of certain base metals can be obtained by fusion methods,
and a brief resume* of the procedures that have been used in the
past will serve to indicate the potentialities and limitations of
such methods.

In order that a fire assay method of a base metal be feasible
the metal must be at least below iron in the electromotive force
series and be present in some form that can be reduced to metal
at reasonable temperatures, usually not exceeding 1100C. in a
fusion with iron or carbon as a reducing agent. Furthermore the
metal must be the only one present that is reducible under the
required conditions.

The metals that are amenable to the fire assay are, in the order
of increasing difficulty and decreasing reliability of assay, lead,
bismuth, tin, antimony, and copper.

Lead Assay. The fire assay for lead was at one time used as
the basis of settlement between buyers and sellers of ores. Even
today a vestige of the former prestige of the fire assay remains
in most smelter contracts, in which a flat deduction from the
volumetric assay is made on the assumption that an equivalent
amount of lead is lost in the smelting process.

The principal sources of error in the fire assay for lead are
low results due to the volatility of lead oxide and lead sulfidc and
to the slagging of lead, and high results from the reduction of
antimony, copper and certain other metals with the lead.

On rich ores relatively free from other base metals the fire
assay is capable of giving results within 0.5 per cent of the true
lead percentage and duplicate assays should check within 0.2 per
cent lead, but when interfering elements are present the error
may be 2 per cent or more. On account of excessive slag losses
the fire method is unsuited to the assay of materials containing

220

FIRE ASSAY METHODS FOR BASE METALS 221

less than 1 or 2 per cent of lead, unless it is possible to concentrate
the lead-bearing minerals by panning a large sample.

The most common type of slag used for the lead assay is a sub-
silicate of soda (or potash), silica, and borax glass. A cyanide
fusion is also feasible, as in the fire assay for bismuth, antimony,
or tin, which see for details. The subsilicate slag for the lead
assay is similar to the slags used in the crucible assay for gold
and silver except that litharge is absent. The strongly alkaline
slag has a solvent action on sulfides. The assay procedure is
analogous to the soda-iron method described on page 146 in that
metallic iron is used as a reducing agent. Iron is necessary not
only to decompose sulfides according to the following reactions:

4Na 2 C0 3 + 7PbS = 4Pb + 3(Na 2 S,PbS) + Na 2 SO 4 + 4CO 2 ) ,

(Na 2 S,PbS) + Fe = (Na 2 S,FeS) + Pb ) ( '

PbS + Fe = FeS + Pb (2)

but also to secure a sufficiently strong reducing action to prevent
the formation of lead silicate, or to reduce lead from silicates
already present, as in the following reaction

2PbO.Si0 2 + 2Fe = 2FeO.SiO 2 + 2Pb (3)

The iron is provided as nails or spikes or by the use of an iron
crucible. In addition to iron, flour or other carbonaceous agent
is added to intensify the reducing action, particularly during the
early stages of the fusion. Because argols were generally used
as a source of carbon the method was generally known as the
soda-argol method.

If much arsenic is present, part of it is volatilized as arsenic
trioxide, and some enters the slag, probably as sodium arsenate,
but in the presence of iron considerable speiss 1 may form, which
will be found as a brittle layer on top of the lead button after
pouring and cooling the fusion. Very little arsenic is reduced to
the button, and if the speiss is removed carefully, the reliability
of the lead assay is not affected.

Antimony is easily reduced, and most of it will be present in
the lead button, unless the amount present is in excess of about
half of the lead, when a speiss will form.

Any bismuth present will be reduced and practically all of it
will be found in the lead button.

1 See pp. HOjf.

222 FIRE ASSAYING

Copper in sulfide form dissolves in the alkaline slag, but, if
there is a deficiency of slag, matte will form. If copper is
present in oxidized form, metallic copper is reduced and con-
taminates the button. Hence, sulfur should be added if neces-
sary, in sufficient amount to form sulfide with all copper present.

Zinc is partly reduced and volatilised and partly remains in the
slag. Zinc sulfide is not fusible at the temperatures reached
in the lead assay nor is it so soluble in the slag as are the sulfides
of copper and iron; hence if much zinc is present as sulfide or
with other sulfides, the slags may be pasty, and considerable
shotting of lead may result, causing inaccurate assays.

Procedure. 1. For rich ores and concentrates, use a 5- or 10-g.
sample and a 15- or 20-g. crucible; for medium- and low-grade
ores, use 20 g. of ore and a 30-g. crucible.

2. Add 2^2 to 3 times as much sodium carbonate as ore, or a
mixture of sodium and potassium carbonate, which has a slightly
better carrying capacity for sulfides than either alkali alone.
Use an equal weight of potassium and sodium carbonate if much
zinc, iron sulfide, or basic gangue is present.

3. Add sufficient borax glass to form the equivalent of a sub-
silicate, but provide a minimum of one-fourth as much borax
glass as ore, even on siliceous ores. For rich ores largely com-
posed of lead sulfide, oxides, or carbonates, about half as much
borax glass as ore will be required. For zincky ores it may be
necessary to increase the acidity of the slag to a monosilicate,
and part of the acidity should then be provided by silica. See
examples of slag calculations in Chap. VII.

4. Add 4 to 6 g. of flour or other carbonaceous reducing agent.
If copper is present in the absence of sulfides, add sufficient sulfur
to form Cu2S with the copper present.

5. Mix the charge and insert three tenpenny or two twenty-
penny nails or other suitable form of iron into the crucible. An
iron crucible is very suitable for high-grade ores instead of nails
but is rapidly corroded with impure ores.

6. Place crucible in furnace at a dull-red heat (550 to 600C).
A reducing atmosphere is desirable; hence a muffle furnace gives
better results than a pot furnace. A piece of coke or coal in
the front of the muffle aids the maintenance of reducing condi-
tions. Gradually raise the temperature over a period of 20 to
30 min. to a bright-yellow heat (1050 to 1150C.), and hold at

FIRE ASSAY METHODS FOR BASE METALS 223

that temperature for 20 to 30 min. A rapid fusion in a hot
furnace is undesirable. The slower heating schedule favors the
reduction of lead to metal before appreciable amounts are lost
by volatilization and the prolonged heating at fusion tempera-
ture favors complete reduction of lead from the slag. The entire
heating cycle should range from 45 to 60 min.

7. Test for completeness of fusion by removing a nail for inspec-
tion. If drops of lead adhere to the nail and are not removed
by stirring the nail in the slag and then tapping it against the
edge of the crucible, either the slag composition is incorrect or
fusjon is incomplete.

When completely fused, remove the crucible from the furnace
with the crucible tongs held in one hand and, by means of short
tongs in the other hand, remove the nails as quickly as possible
one by one, washing them in the slag and tapping them on the
sides of the crucible to remove adhering lead.

Pour into a conical mold, cool, detach slag, and speiss if any,
and weigh the button.

Calculate as percentage of lead, to the nearest 0.1 per cent.

Notes : a. Slags should be glassy. If they are dull, insufficient acid flux
was added.

6. Brittle buttons are an indication of the presence of antimony. Hard
buttons indicate the presence of antimony or copper. If impurities are
suspected, the buttons may be cupeled and the cupels examined for charac-
teristic indications of copper, antimony, or bismuth (see Chap. IV).

c. If much silver and gold are present the amount should be deducted
from the lead assay.

d. A corrected lead assay may be made by reassay of the slag. If the
original slag was glassy, make the new slag more basic by the addition of
5 to 10 g. of sodium carbonate; if originally basic, add more borax glass.
Use the same nails as in the original fusion. Heat to 700 to 800C., drop in
a lump of approximately 5 g. of potassium cyanide, heat to approximately
1100C., pour (do not inhale fumes) , recover the button and add to the lead
from the original fusion.

Tin Assay. The fire assay of tin can be applied successfully
only to cassiterite (Sn0 2 ) concentrates as free as possible from
other substances, particularly silica and silicates, because tin
acts as either an acid or a base, forming silicates with silica and
stannates with iron and other bases. Ores and gravels containing
cassiterite must first be concentrated by panning as far as
possible without loss of tin. Then the concentrates are roasted to

224 FIRE ASSAYING

remove sulfur and are treated with aqua regia to remove most
of the impurities except quartz, some silicates, and tungsten
minerals. If much quartz and silicates are left, a treatment with
hydrofluoric acid in a platinum dish may be necessary. The
final concentrate will then consist almost exclusively of cassiterito
plus not more than a small percentage of silicates and a part
of any tungsten minerals originally present. The fire assay
itself is therefore largely a means of estimating the tin content
of the concentrate and of verifying that the concentrated mineral
is cassiterite rather than wolframite.

The best flux for cassiterite is potassium cyanide, which com-
bines the functions of a flux and a reducing agent according to
the following reaction:

2KCN + SnO a = Sn + 2KCNO

The KCN should be free from carbonates, sulfates, and sul-
fides, as these cause serious loss of tin in the slag. Cyanide is a
powerful poison, and every precaution must be taken to avoid
inhaling its dust or fumes.

Alkaline fluxes, such as sodium or potassium carbonate, with
carbon reduction are not so satisfactory, mainly on account of
the formation of stannates, especially in the presence of iron.

Procedure. 1. The amount of concentrates used should be
close to 10 g. Use a 20-g. crucible and tamp in a layer of 2 to
3 g. of KCN. Mix the concentrate with three or four times its
weight of KCN, place the mixture in the crucible and cover with
5 g. additional KCN.

2. Fire at 750 to 800C. for 15 to 20 min. in a well-vented
furnace. A longer fusion may be required if the concentrates
are impure.

3. Remove from the furnace and cool under a hood.

4. Recover the button by breaking the crucible. The tin
button should be soft and white if free from foreign metals.

Copper Assay. The fire assay of copper possesses considerable
merit for the determination of copper in materials rich in coarse
metallic copper. Such materials are difficult to sample accu-
rately, and larger samples are required than can be assayed by wet
methods. The fire assay for copper was formerly used in the Lake
Superior district for mine rock, mill concentrates, gravel and

FIRE ASSA Y METHODS FOR BASE METALS 225

tailings, and slags from reverberatories and remelting cupolas, 1
but according to data supplied by Professor N. H. Manderfield 2
the method is now used in that district only for coarse mill
concentrates. Professor Manderfield presents the following
typical crucible charge used by the amygdaloidal mills:

Weight of sample 1000 g.

Flux batch:

Cream of tartar* 160 g.

Na 2 00 3 100 g.

Borax glass 80 g.

Total 340 g

* The use of cream of tartar follows customary practice. Any other
sulfur-free reducing agent would be suitable.

Sampling. The concentrate is sampled with a shovel as it is
being loaded into a car or with pipe sampler after the car is
loaded. A sample weighing approximately 100 Ib. is obtained
in this manner, and in turn it is cut to about 5,000 g. with a
Jones riffle.

Preparation of Samples. Moisture is run on the total sample.
Two 1,000-g. samples are riffled out, and each is mixed thoroughly
with 340 g. of flux, when assaying finisher-jig concentrates con-
taining 70 to 85 per cent copper, or with 255 g. for products con-
taining 80 to 90 per cent copper. The samples are placed in
Denver size L fire-clay crucibles and covered with a small amount
of borax glass.

Fusion. The crucibles are placed in the furnace and capped
with clay covers and allowed to fuse for approximately 1 hr.,
finishing at 1100C. or higher. After fusion the crucibles are
removed and allowed to cool and set. The crucibles are then
broken and the buttons are freed from slag, cleaned with a ham-
mer, and weighed.

The percentage of copper in the original sample is computed
from the weight of the copper buttons obtained.

Antimony and Bismuth Assay, Approximate assays for anti-
mony and bismuth in the absence of each other and of lead, cop-
per, and tin can be made by a fusion of 5 or 10 g. of ore with four

1 FULTON, C. H., and SHARWOOD, W. J., "A Manual of Fire Assaying,"
3d ed., p. 251, McGraw-Hill Book Company, Inc., 1929.

2 Personal communication, Apr. 11, 1940.

226 FIRE ASSAYING

or five times its weight of potassium cyanide plus a cover of
potassium cyanide, following the procedure given for the assay
of cassiterite concentrates. The buttons will be brittle and care
must be taken to avoid loss in separating them from slag. Better
results are obtained with bismuth than with antimony, as
antimony has a stronger tendency to enter the slag.

CHAPTER XIV

THE ACCURACY OF THE FIRE ASSAY
FOR GOLD AND SILVER

The precision with which an assay result represents the average
grade of the material from which the assay sample was taken
depends upon the error introduced in each of the steps of the
assay operation. There are errors peculiar to each of the usual
steps in the assay process: (1) sampling, (2) fusion, (3) cupella-
tion, (4) parting, and (5) weighing. In addition, serious error
may be introduced at almost any stage of the process by salting.

Salting. Salting consists of accidental or fraudulent introduc-
tion of extraneous gold or silver into the assay. Fraudulent
salting occasionally occurs during sampling and assaying for
valuation purposes. It is carried out by introducing metallic
grains, metal filings, amalgam, cyanide precipitate, rich ore,
concentrate, or solutions such as gold chloride, silver nitrate,
or gold and silver cyanide into the sample or into the assay fluxes.
The assayer's principal responsibility starts when he receives
the samples, but he should subject important valuation samples
to an examination for the purpose of detecting the presence of
gold or silver in unnatural particles, and to a leaching test for
detecting water-soluble gold or silver that may have been added
before the samples reached the assay office.

In a sample salted with gold metal particles to about the grade
of an average ore the salting particles are too few to be seen
directly in the sample, and a concentrate should be panned out
for examination. It is advisable to use a hand lens or a micro-
scope for studying the concentrate, and any unusual character-
istic of the valuable particles should be looked upon with
suspicion.

A leaching test for soluble gold or silver is made by agitating a
portion of the sample with H 2 O for about *^j hr. Then filter or
decant to separate the solution. Evaporate the solution to
about 25 ml. in a beaker, and complete the evaporation in a small
rectangular , tray, made by bending up the edges of a sheet of

227

228 FIRE ASSAYING

lead foil. After the solution has evaporated, roll the lead tray
into a compact bundle and cupel. The presence of even a small
amount of water-soluble gold or silver almost conclusively proves
salting. A considerable part of the soluble gold or silver added
to salt the s.ample may not redissolve in water, particularly if the
sample has been heated to dry it for the ordinary assay.

In the assay office, valuation samples should be kept in a
locked compartment, except when they are in process, and
unauthorized persons should not be admitted to the sampling,
fluxing, and furnace rooms. The secret introduction of a few
samples of barren material in a series of valuation samples is a
useful aid in detecting either accidental or fraudulent salting.

Common causes of accidental salting are:

1. Contamination of low-grade samples with rich material,
owing to the careless handling of rich and low-grade samples dur-
ing their preparation for assay.

2. Use of reagents such as mercury, litharge, and lead foil
containing undetermined amounts of gold and silver.

3. Reuse of a crucible after an unsatisfactory previous fusion.
Contamination during sample preparation can be avoided by

careful cleaning between samples. Grinding a small amount of
barren rock between samples is desirable in order to clean the
rubbing surfaces of the pulverizer, which may have become
impregnated with gold from rich samples.

The best reagents obtainable for assaying contain some gold
and silver. As a rule, the amount is so small that it may be
neglected, but this point should always be determined by " blank "
assays of the reagents. Each new lot of reagents should be
tested by a blank assay, made with assay silica that is treated
as a quartz ore. When running a blank for gold a small amount
of silver should be added before cupellation, as a minute gold
bead cannot be located and recovered from the cupel. Several
assay-ton charges may be required to obtain a weighable amount
of gold. Many assayers use commercial litharge and correct
for its gold and silver content, rather than pay nearly twice as
much for the "silver-free" variety of litharge. Silver-free
litharge is not entirely barren and should be tested by a blank
assay.

A fusion producing a lumpy slag, a shotted slag, or an insuffi-
cient lead button, leaves precious metals in the crucible to con-

THE ACCURACY OF THE FIRE ASSAY 229

t animate subsequent fusions. Such crucibles could be cleaned
by a blank fusion in which a lead button is reduced, but it is
generally cheaper to discard them. It is good practice either
to discard crucibles that have been used on rich samples or to set
them aside for use only with materials known to be rich. At a
Western custom smelter, all assay crucibles are discarded after
a single use, that the possibility of using salted crucibles may be
avoided.

Sampling. Two kinds of errors are encountered in sampling:
(1) the normal error, due to chance variation in the number of
rich or low-grade pieces that enter the sample; and (2) the error
due to bias in the sampling method.

The error due to chance variation in sampling is discussed in
Chap. II. In general the reliability of a sample is improved by
increasing the weight of the sample, by increasing the number
of units it contains, or by crushing and mixing the material
before sampling to decrease the dispersion per unit of weight.
Greatest reliability of sample is required at custom smelters
where the sample is the basis of settlement between buyer and
seller. Care is generally taken to secure samples that are accurate
to within 1 per cent of the assay.

When single assays are used only as a general guide to opera-
tions, such as mining and milling, and when calculations from
the assays are based upon the average of a large number of single
assays, considerable chance variation may be allowed in each
sample. The reliability of the average of a number of samples
increases as the square root of the number of samples included
in the average.

For example, the practical limit of error in daily mill-head
samples at a gold mill could be as much as 0.05 oz. of gold per
ton, yet the monthly metal-balance calculation would not have
an error of more than 0.01 oz. of gold per ton due to chance
variation in sampling.

Bias in the sampling method produces samples that consistently
give either high or low results. It can be avoided by making sure
that the difference in behavior of particles that vary in density
or size does not affect their selection or rejection from the sample.

Fusions. Two sources of error are encountered in assay fusion
processes: (1) error due to disregarding the gold and silver con-
tent of the fluxes (salting), and (2) error due to lack of complete

230 FIRE ASSAYING

recovery of the precious metals in the lead button. There is
always some loss of precious metal, even in a normal fusion,
but the loss should be less than 0.5 per cent of the precious metals.
The metal lost is practically all in either the slag or crucible,
and it can be recovered by remelting the slag with litharge and a
reducer in the original crucible, which will produce a second lead
button as in the corrected assay. Some assayers regularly
"wash" the slag with lead by adding litharge with some addi-
tional reducing agent shortly before removing the crucible from
the furnace. The purity of the lead button can be improved
at the same time by adding more litharge than will be reduced.
The extra litharge sinks through the slag and forms an excess
litharge layer just above the lead button.

Flour or charcoal is generally used with loose litharge for
washing the slag. In developing the oxidation-collection method
of assay it was found that litharge with iron filings was much
more effective than litharge with either flour or charcoal, in col-
lecting the precious metals from a molten slag. It was also
found that the best results were obtained by briquetting the
litharge and iron filings. Similarly the addition of a collector
briquette of litharge and iron filings should give the best slag
washing.

Materials are sometimes encountered that are claimed to con-
tain precious metals in a colloidal or volatile form that are not
recovered by standard assay methods. Experiments have shown
that gold added as the volatile chloride or as a colloid is invariably
recovered in ordinary assay fusions even from mysterious ores
that are reputed to cause volatilization or other loss. Claims of
recovering more gold than revealed by the fire assay are usually
based upon experiments in which mercury that is salted with
gold has been used.

Cupellation. The cupellation process is subject to (1) loss of
precious metals by volatilization, (2) absorption or loss of
precious metals that enter the body of the cupel, and (3) reten-
tion of base metals, causing a gain in weight of the silver bead.

Loss of precious metals by volatilization during cupellation is
inappreciable with normal conditions.

Loss of precious metals by cupel absorption accounts for the
greatest normal loss in the fire assaying process. Roughly the
loss amounts to from 0.1 to 0.5 per cent of gold beads, and from

THE ACCURACY OF THE FIRE ASSAY 231

1 to 3 per cent of silver beads. Cupel absorption increases with
(1) the temperature of cupellation, (2) amount of lead in the
button, (3) presence of impurities in the button, and (4) size
of the bead. The type of cupel, and the ratio of gold to silver in
the bead, also affect cupel absorption.

Temperature. The loss of precious metals increases rapidly
with increase in temperature of the cupeling alloy; consequently
the temperature should be maintained at the lowest point at
which the process will proceed. This is slightly above the melting
point of litharge 888 C. The muffle temperature must reach
this point for cupellation to start. After the button opens, the
heat furnished by the rapidly oxidizing lead raises the tempera-
ture of the alloy despite the lowered muffle temperature. Meas-
urement of the temperature of the cupeling alloy is difficult, and
as a result cupellation temperatures are generally given as
muffle temperatures. At any muffle temperature the tempera-
ture of the cupeling alloy varies with the amount of lead oxidizing
and the rate at which heat is removed from the alloy.

For example, the temperature of the cupeling alloy in a dense
magnesia cupel does not rise so far above muffle temperature as it
does in a bone-ash cupel, because the magnesia cupel is a better
conductor of heat. This is the principal reason for good magnesia
cupels giving lower losses than bone-ash cupels.

Amount of Lead. The loss of silver increases slightly with
increase in the amount of lead to be cupeled, when other condi-
tions are held as nearly constant as possible.

For example, if cupeling 20 mg. of silver with 10 g. of lead pro-
duces a loss of 2 per cent, cupeling with 15, 20, and 30 g. of lead
will produce losses of about 2.8, 4.5, and 5 per cent, respectively.
Gold and large amounts of silver do not show so marked an
increase in loss with increase in amount of lead.

Impurities. Practically all impurities in the lead button
increase the loss of precious metals during cupellation. The
effects of harmful impurities in the lead button are to (1) lower
the surface tension of the alloy, allowing increased cupel absorp-
tion; (2) cause the bead to freeze unless an unduly high cupel-
lation temperature is used; and (3) form high-melting-point
oxides that occlude small globules of precious metals in a solid
scoria. Serious loss can be expected if sufficient impurities are
present to require an unusually high cupellation temperature, or

232

FIRE ASSAYING

if solid scoriae are left in the bowl of the cupel. Such assays
should be repeated with fusion conditions adjusted to reduce the
amount of impurities in the lead button.

Size of Bead. The loss of precious metals during cupellation
increases with the amount present, but the percentage loss
decreases slightly with the size of the bead. Sharwood 1 proposed
the rule that, other conditions being equal, the loss on cupellation

10 . , - rw _

-p --

/ X

-Xx

'' b

a ' 3 > '

^ ^

__|--

...

s ,v

r,'

/ J *"~

E .0.8
0.6

feO.4
0.2
0.1

J^r

^x^

x*f^

1
1

1
I

1
i

2 3 4 6 8 10 20 3040 6080100

200 300400 600 1000

Milligrams of silver cupellecf
FIG. 16. Curve representing silver losses in cupellation, plotted on logarithmic
cross-section paper. (After Fulton and Sharwood, Manual of Fire Assaying.

is proportional to the surface of the finished bead. The loss
would thus be proportional to the two-thirds power of the bead
weight. This relation between milligrams of bead weight and
milligrams of weight lost is a straight line on logarithmic cross-
section paper, parallel to a line representing y x**. By the
aid of a logarithmic chart, such as Fig. 16, the cupellation loss of
beads of different sizes can be estimated from loss obtained by
cupeling a proof with a known amount of gold and silver.

For example, point a in Fig. 16 represents the loss of 1.8
mg. of silver. Then, from the diagram, silver beads 6, weigh-
ing 250 mg.; c, weighing 35 mg.; and d, weighing 15 mg.; cupeled
under the same conditions, should be expected to have lost

1 SHAKWOOD, W. J., A Rule Governing Cupellation Losses, Trans.
A.I.M.E., vol. 52, p. 179, 1915.

THE ACCURACY OF THE FIRE ASSAY 233

2.85 mg., 0.72 mg., and 0.40 mg., respectively, and may be cor-
rected accordingly.

Ratio of Silver to Gold. Gold loss during cupellation can be
estimated on the logarithmic chart, provided the ratio of silver
to gold during cupellation is the same in the proof as in the assays
to be corrected. The ratio of silver to gold is important, as the
loss of gold during cupellation decreases with increase in the pro-
portion of silver to gold. Rose 1 found a loss of 1.2 per cent of
the gold in cupeling 1 mg. of gold with 4 mg. of silver and 25 g. of
lead at 900C. A similar cupellation of 1 mg. of gold with 6 mg.
of silver gave a gold loss of 1.05 per cent, and an increase of silver
to 10 mg. reduced the loss still further to 0.80 per cent of the gold.

Retention of Base Metals.- The retention of base metals in the
silver bead increases its weight, but ordinarily the error is inap-
preciable. In any event the gold weight is not affected, because
the base metals are separated by nitric acid parting.

A minute amount of lead is generally retained in the bead,
particularly when the finishing cupellation temperature is low.
The amount of retained lead is too small to warrant correction,
but it tends to cause large beads to break up during parting.

Copper aids in the elimination of lead from the bead and, for
this reason, some copper is added in the gold-bullion assay if it is
not already present. When the lead button in a silver assay
contains insufficient copper to interfere with the cupellation proc-
ess, the weight of copper retained in the bead may be neglected.

When bismuth is present in an ore being assayed, it concen-
trates in the lead button, and an appreciable amount is retained in
the bead, which gives a plus error for silver. The amount,
retained varies with the silver and bismuth content of the ore.
A correction to the silver assay of bismuth-bearing ores is com-
monly made at smelters, by deducting the bismuth content of the
silver beads (determined by wet analysis of the silver beads).
Bugbee 2 gives a typical example, showing a correction amount-
ing to 2 per cent of the silver for a concentrate containing 0.3 per-
cent bismuth and giving an unconnected silver assay of 78 oz.
per ton.

1 ROSE, T. K., Cupellation and Parting, Eng. Min. Jour., vol. 79, p. 708,
1905.

2 BUGBEE, E. E., "A Textbook of Fire Assaying," 2d ed., p. 211, John
Wiley & Sons, Inc., New York, 1933.

234 FIRE ASSAYING

Parting. The error in parting ordinary gold assays should be
negligible. A small amount of gold may dissolve in the parting
acid, 1 but under ordinary conditions a larger amount of silver
stubbornly remains with the gold, resulting in a net gain in weight
of about 0.1 per cent.

Mechanical loss of particles of gold in the parting acid or wash
water can be seen by the operator. Such loss, however, can be
avoided by careful manipulation.

Weighing. Errors may be introduced in the weighing opera-
tion by (1) mistakenly recording a different weight than the
balance indicates, (2) using a faulty balance or weights, or (3)
the natural error of the balance due to lack of perfect sensitivity.

Mistakes in reading and recording the weight indication of the
balance can be avoided. The operator should check the balance
reading for each bead, by recording the weight and then reading
the weight again before proceeding with the next bead.

The assayer should make certain that his assay balance and
weights are in good order. The rider can be checked against the
1-mg. weight, and the total of several small weights checked with
a larger weight. It is desirable to reserve a precision set of assay
weights in the office for occasional comparison with the weights
in use.

The assay balance may cause error in weighing if the knife-
edges are damaged, or if the arms are of unequal length. Dam-
aged knife-edges are readily detected by the sluggish and erratic
behavior of the balance. Unequal length of arms is a more
serious fault, as it produces results that are consistently either
high or low. This difficulty is seldom encountered but should
be checked frequently by weighing a large bead in both the left-
hand and right-hand pans. If the two weights obtained are not
equal, the length of the balance arms is probably at fault.

Every assay balance has a definite natural error equal to the
smallest unbalanced load that noticeably affects equilibrium.
Assay balances can be purchased with a rated sensitivity of 0.002
mg. Gold beads are commonly weighed to the nearest 0.005 mg.
The error in any one weighing can be held within 0.003 mg.,
but this represents an appreciable percentage of small gold
beads.

1 FULTON, C. H., and SHAKWOOD, W. J., "A Manual of Fire Assaying/'
3d ed., p. 208, McGraw-Hill Book Company, Inc., New York, 1929.

THE ACCURACY OF THE FIRE ASSAY

235

For example, a weighing error of 0.003 mg. represents an error
of 5 per cent in a 1 -assay- ton assay of a low-grade ore containing
0.06 oz. of gold per ton. The error can be reduced by taking a
larger sample.

Composite Samples. The composite method of checking is
frequently employed as a safeguard against abnormal errors in a
sequence of single assays. As a check on assaying alone a simple
arithmetical average can be used, in which case an equal weight
is taken from each individual sample in the series, and the com-
posite sample is then thoroughly mixed and assayed. The assay
of the composite is compared with the arithmetical mean of the
individual assays.

Weighted composite samples, prepared by taking for the com-
posite a weight in proportion to the tonnage applying to each
sample, are used to check the cumulative value of concentrate or
ore over a period of time. Table XIX shows an example of a
weighted composite sample used to check daily mine samples for
a 6-day week.

TABLE XIX. EXAMPLE OF WEIGHTED COMPOSITE ASSAY

Weight taken

Daily gold

Total gold

Day

Tons mined

in composite,

assay, ounces

(tons X assay),

grams

per ton

ounces

375.2

37.5

0.32

120.0

401.8

40.2

0.28

112 5

386.6

38.7

0.41

158 5

260.4

26.0

0.29

75.5

348.1

34.8

0.50

174.0

416.3

41.6

0.37

154.0

Total

2,188.4

794.5

Computed mean is 794.5/2,188.4 = 0.363 oz. per ton. Assay of com-
posite is 0.35 oz. per ton.

The assay error between the composite (0.35 oz. per ton) and
the weighted mean of the daily assays (0.363 oz. per ton) in
Table XIX is only 0.013 oz. per ton. This is within the limits of
error of routine gold assay practice.

Metal Balances. Metal balances are the best means of check-
ing the accuracy of weighing, sampling, and assaying operations.
They should be made for all metallurgical testing as well as for
mill and smelter operations. The balance most commonly used

236

FIRE ASSAYING

is that between the calculated recovery of precious metals (cal-
culated by subtracting the precious metals in the tailing from that
in the feed) and the actual recovery of the precious metals
(determined from weights and assays of concentrate and bullion,
or from mint and smelter returns). Table XX shows a typical
yearly metal balance at a small gold mill producing bullion from
amalgamation, which is sold to the mint, and concentrate from
flotation, which is sold to a smelter. In Table XX there is a
discrepancy of 125 oz. between the calculated and actual recovery
of gold for the year. This error is less than half of 1 per cent of
the total gold involved and represents particularly good work.
Careful weighing of ore and concentrate, and corrected assays,
are required to obtain so close a check.

TABLE XX. TYPICAL YEARLY GOLD-MILL METAL BALANCE

Material

Dry weight

Gold assay

Calculated
gold content
(weight X
assay), ounces

Feed

49,253 tons

0.675 oz/ton

33,246

Tailings

48,047* tons

0.035 oz/ton

1,681

Calculated gold recovered .
Flotation concentrate
Amalgamation bullion

1,206 tons
31,155oz.

5.07 oz/ton
821 fine

31,565
6,112
25,578

Actual gold recovered

31 690

Actual gold recovered, 31,690 oz.; calculated gold recovered, 31,565 oz. =
125 oz. discrepancy.

* Feed and concentrate are usually weighed, and the weight of the tailings ia obtained by
difference.

In addition to the metal balance shown in Table XX a further
check of the gold recovered is made with the mint and smelter
returns. Usually another check is made by calculating the mine
production of metal separately, for comparison with the precious
metal in the mill feed.

Monthly and annual metal balances are made at mills and
smelters. Metal balances for shorter terms are troublesome and
unreliable, because of the variable amount and grade of ore in
transit, storage, and process.

Control Assays and Splitting Limits. Four samples are gener-
ally taken by mills and smelters purchasing ore, as shown in

THE ACCURACY OF THE FIRE ASSAY 237

Fig. 5. One sample is furnished the seller for assay, and one
sample is assayed by the purchaser. These are known as " con-
trol assays. " If the control assays check within previously
specified limits, called " splitting limits/' the difference is ordi-
narily split and settlement is made on the average. Some pur-
chasers, however, pay only on their own assay, except when an
umpire is called for.

A difference between control assays of 0.02 oz. per ton is usually
split on gold assays of less than 2 oz. of gold per ton. The split-
ting limit is increased to 1 per cent of the assay for richer ores.
Silver assays, on ores containing less than 50 oz. of silver per ton,
are allowed splitting limits of from 0.2 to 0.5 oz. per ton. The
splitting limit for higher grade ores is increased to approximately
1 per cent of the silver assay.

If the discrepancy between the control assays is greater than
the splitting limit, the third sample is sent to a disinterested com-
mercial assay office for an " umpire assay." Umpire assays are
usually made in quadruplicate and the results averaged.

When the umpire assay falls between the two controls, settle-
ment is, as a rule, made on the umpire assay, and the cost of the
umpire is shared by buyer and seller. If the umpire assay falls
outside the controls, the nearest control is used for settlement
and the one whose assay is farthest from the umpire pays its cost.

Control and umpire assays of ores and concentrates are made
without correction for normal fusion and cupellation losses; con-
sequently they are likely to be from 0.1 to 1.0 per cent low for
gold and from 1 to 4 per cent low for silver.

CORRECTED-ASSAY METHODS

An assay that has been corrected for errors is known as a "cor-
rected assay." Corrected assays are used to obtain accurate
metal balances and to evaluate rich products such as bullion and
cyanide precipitate. Smelters seldom accept corrected assays
as a basis of payment for ores and concentrates.

Corrected-assay methods may be classified as direct or indirect,
according to whether the correction is determined directly on the
assay in question, or a separate " check" assay, or some other
indirect means is used to estimate the correction.

Indirect Corrected Assays. The indirect method is generally
made by correcting an assay for the error found in treating a

238 FIRE ASSAYING

weighed amount of pure gold and silver in a synthetic " check 7 '
assay of approximately the same composition as the regular assay
and run under the same conditions. Check-assay corrections
give highest accuracy and should be used whenever the approxi-
mate composition of the material to be assayed is known or can
be easily determined. It is the standard method for assaying
gold bullion and is also used in the fire assay method for assaying
silver bullion.

In the assay of mill feed and products the composition of each
material and the conditions of assay frequently remain fairly
constant. Rough corrections, based upon the usual errors
encountered in assaying these products, are of great advantage
in obtaining accurate metal balances.

For example, at a mill treating a silver ore by the cyanide proc-
ess the calculated silver recovery obtained by subtracting the
silver in the tailing from the silver in the mill feed was generally
found to be about 2 per cent below the silver actually recovered
as bullion. An investigation revealed that uncorrected assays
had been used in the calculations and that the usual greater loss
of silver in the feed assay compared with that in the tailing assay
accounted for most of the discrepancy. It was found by experi-
ment that the average feed assay of about 25 oz. per ton suffered
a normal loss of 2.3 per cent or 0.57 oz. per ton, while the average
tailing assay of about 3 oz. per ton suffered a loss of 4.5 per cent
or 0.14 oz. per ton. The difference in loss, 0.43 oz. per ton,
amounted to 1.7 per cent of the average feed assay. Subse-
quently the feed assays were arbitrarily increased by 1.7 per cent,
and yearly metal balances were obtained that checked within
0.5 per cent. This method of correction could be refined by
varying the correction according to the size of the bead, with the
aid of a chart such as Fig. 16.

Direct Corrected Assays. When a direct corrected assay is to
be made, the customary assay is carried out, and the slag, cruci-
ble, and cupel are saved. In order to recover the lost precious
metals for the correction the slag and cupel are then assayed,
using the original crucible, in a manner similar to an ore assay.
The corrected result is likely to be slightly high, particularly with
large beads, because the plus errors (due to retention of base
metals in the silver bead and of silver in the parted gold) are
neglected. Neglected losses by volatilization and losses in the

THE ACCURACY OF THE FIRE ASSAY 239

second slags and cupels aid in reducing the plus error, but the
neglected losses seldom exactly balance the weight of retained
impurities. The appreciable plus error, which may be introduced
by correcting for the major losses and neglecting gain in weight
due to retention can be seen from the check-assay correction,
or surcharge, for gold bullion.

A minus surcharge is generally obtained in assaying impure
gold bullion that is less than 700 fine. With bullions between
700 and 800 fine, the plus errors about balance the total losses,
and the surcharge is nearly zero. In assaying gold bullion that
is more than 800 fine the plus errors are usually greater than the
total losses, and a plus surcharge is obtained that must be sub-
tracted in order to correct the assay. Seriously high assays of
high-grade gold bullion would be obtained if retention were
neglected, and if the cupellation loss were determined and added
to the assay.

The assay of ores and concentrates differs in several respects
from the assay of gold bullion, and the result of a properly con-
ducted direct corrected assay of most ores and concentrates is
considerably nearer the true assay than the result of an ordinary
uncorrected assay.

The direct corrected assay is the usual method of reducing the
error of a normal fire assay of materials whose approximate
composition cannot readily be duplicated in a check assay.

Assay of Slags. Many authorities postulate that it is neces-
sary to alter the composition of the original slag and, in particu-
lar, to reduce all the litharge from the slag, replacing it with soda
in order to obtain good recovery of its gold and silver content.
Extensive experiments have shown that there is no advantage in
altering the slag composition, and that replacement of the
litharge with soda is in many instances detrimental to good
recovery, on account of the failure of the new slags to retain some
of the impurities in the original slag. Furthermore the addition
of sodium carbonate makes the charge bulky and increases the
danger of loss by boiling.

An assay of a slag that was of good composition in the first
assay can be made as follows : Grind the slag to 35-mesh or finer.
Mix the ground slag with 30 g. of litharge and sufficient reducing
agent to reduce a 25-g. button. Fuse in the original crucible and
proceed as with any other crucible assay.

240 FIRE ASSAYING

Assay of Cupels. Cupels may be composed of bone ash, port-
land cement, mixtures of cement and bone ash in varying propor-
tions, or magnesia. Typical compositions are given below:

Type of Cupel Typical Composition

Bone ash Ca 8 (PO 4 ) 2 , 80-90 per cent

Cement CaO, 60 per cent; SiCh, 25 per cent

Magnesia MgO, 80 per cent

The cupels ordinarily are of bone ash, bone ash-cement (1:1), or
magnesia; straight portland-cement cupels are rarely used.

It is desirable to keep to a minimum the amount of cupel
material which must be handled in the assay fusion. This can
be accomplished by limiting the size of the original lead button
to about 25 g. and by discarding all but the litharge-saturated
portion of the cupel. A 25-g. button produces 27 g. of PbO ; but
approximately 5 g. volatilizes during cupellation, leaving 22 g. of
PbO absorbed in the cupel. Bone-ash cupels absorb approxi-
mately their own weight of litharge, bone-ash-cement cupels
(1:1) absorb slightly more than their weight, and magnesia
cupels absorb about 75 per cent of their weight. Thus a 25-g.
button will saturate roughly 22 g. of a bone-ash cupel, 22 g. of
a bone-ash-cement cupel, or 30 g. of a magnesia cupel.

Magnesia and portland cement from cupels are fluxed in the
same manner as similar constituents in an ore (Table XI, page
125). Bone ash disperses in a slag as inert particles and does
not affect the silicate degree. The addition of fluorspar equal
to one-half the amount of bone ash is recommended by some
authorities. Fluorspar acts as a slag diluent, increasing the slag
fluidity, but it can just as well be omitted.

Cupel materials, with the exception of cement, are not readily
attacked by fluxes. Consequently, cupels should be ground to
at least 65-mesh before mixing in the assay charge. Litharge
contained in the cupels tends to smear on the rubbing surfaces
of the grinding machine and should be cleaned off with the silica
that is to be used in the cupel fusion.

The crucible-charge compositions as given in Table XXI are
satisfactory for the assay of from 40 to 60 g. of litharge-saturated
cupel. Usually the slag and cupel corrections need not be
separated, and the slag and cupel may be ground and assayed
together, with consequent saving in time and materials.

THE ACCURACY OF THE FIRE ASSAY
TABLE XXI. CRUCIBLE CHARGES FOR CUPELS

241

Bone ash,
grams

Bone-ash-
cement,
grams

Magnesia,
grams

Cupel material (approximate).
PbO.

( Bone ash 25
}PbO 25

30
30
15
10
(Sufficier

Bone ash 12
Cement 12
PbO 25
30
30
15
10
it for a 25-g. b

Magnesia 30
PbO 25

30
30
20
30

utton)

Na 2 CO 3

Borax glass

Silica

Reducing agent

Examples of Direct Corrected Assays. Corrected assays of a
zinc box precipitate from the cyanide process (containing approxi-
mately 60 per cent zinc), and of a pyrargyrite-galena-pyrite-
quartz ore, are given in Table XXII.

TABLE XXII. EXAMPLES OF DIRECT CORRECTED ASSAYS

Zinc box precipitate,
milligrams per 0.1 A.T.

Pyrargyrite ore,
milligrams per 0.25 A.T.

Assay 1

Assay 2

Assay 3

Assay 1

Assay 2

Assay 3

Ore fusion:
Dore*

400.20
13.30
386.90

1.24
0.06
1.18
(Calrnix
4.77
0.02
4.75

13.38
392.83

399.72
13.27
386.45

1.19
0.08
1.09
magnesif
4.54
0.02
4.52

13.37
392.06

399.16
13.31
385.85

0.96
0.06
0.90
i cupels)
4.72
0.02
4.70

13.39
391.45

512.42

0.92

(Bone-a*
10.26

523.60

510.65

2.01

jh-cement
10.22

522.88

511.41

1.16

cupels)
10.57

523.14

Au *

Slag refusion:
Dore"

Cupel fusion:
Dore"

Total:
Au

Corrected assay
(ounces per ton)
Au

133.8
3,928.3

133.7
3,920.6

133.9
3,914.5

2,094.4

2,091.5

2,092.6

CHAPTER XV

THE PREPARATION OF GOLD AND SILVER

BULLION FROM AMALGAM AND

CYANIDE PRECIPITATE

At many gold and silver mines, particularly those having a
small production, the assayer is responsible for the preparation
of bullion from rich mill products such as amalgam, free-gold
concentrates, and cyanide precipitates. All these products con-
tain more or less silver, but gold is the dominant precious metal
except at silver mines using the cyanide process, when the
cyanide precipitate will consist largely of silver.

This chapter takes up only the bullion-production phase of
milling. For other metallurgical information, standard metal-
lurgical texts should be consulted. 1

Amalgam Retorting. 2 The amalgam obtained from milling
operations is seldom sufficiently free from extraneous substances
to permit retorting without cleaning. At large plants, various
mechanical devices are used to separate the amalgam from
clean-up products, but in small operations the final cleaning of
the amalgam is done by hand either by panning or by a small
hydraulic classifier such as that illustrated in Fig. 17, which is in
use at a California mine.

At many small gold mines the mercury is distilled from amal-
gam, and the retort metal is sold without melting and casting into
bars. The metal is acceptable by the mint in this form, and the

1 ROSE, T. K., and NEWMAN, W. A. C., "The Metallurgy of Gold," 7th ed.,
J. B. Lippincott Company, Philadelphia, 1937.

HAMILTON, E. M., "Manual of Cyanidation," McGraw-Hill Book Com-
pany, Inc., New York, 1920.

DORR, J. V. N., "Cyanidation and Concentration of Gold and Silver
Ores/' McGraw-Hill Book Company, Inc., New York, 1936.

2 In the preparation of this section, valuable assistance was given by E. M.
Smith, assistant mill superintendent and metallurgist, Argonaut Gold Min-
ing Company, Jackson, Calif., and D. L. Wooster, superintendent, Western
Quartz Mining Company, Murphys, Calif.

242

THE PREPARATION OF GOLD AND SILVER BULLION 243

producer is saved the expense of melting and, in the United
States, the cost of a gold melter's license.

Amalgam retorts, furnaces, and condensers of various sizes
and designs can be purchased from equipment dealers or can be
made locally. Retorts are usually made of cast iron, and if only

Discharge
lip

,-Cone of/6ga.
iron welded
to 1" pipe
at bottom

Hose to
water supply

FIG. 17. Simple hydraulic cone classifier for cleaning mercury and amalgam.

the retort is purchased the condenser and furnace can be made at
moderate cost with the tools and supplies ordinarily available at
an operating mine. The condenser can be made from standard
pipe and fittings, using an inner pipe whose diameter fits the
retort outlet and an outer pipe about 30 in. long and four or five
times the diameter of the delivery pipe, fitted with a small inlet
and outlet tube for circulating water. The inlet should enter
the lower end of the water jacket. The condenser delivers into

244 FIRE ASSAYING

a pail partly filled with water, and the delivery pipe should always
be about 2 in. above the surface of the water. Before use the
water circulation through the condenser should be tested, and the
delivery tube checked to see that it is free from obstructions.
Care should be taken to avoid inhalation of mercury vapor.

A simple and economical method of heating small retorts is to
suspend the retort and condenser assembly on an iron rod or pipe
over a carbide can or an oil drum with the top removed. Holes
are cut near the bottom of the can for the admission of air, and a
wood fire is used. Any other suitable fuel may be used, in
appropriate furnaces.

To prevent adhesion of the metal to the retort the inside
should be coated with lime, fire clay, or iron oxide wash or lined
with several thicknesses of paper, prior to charging with amalgam.

The entire lot of amalgam should be weighed just before retort-
ing, and the recovered gold sponge and mercury should also be
weighed. Any abnormal losses indicate leakage of the retort and
condenser system.

After charging the amalgam the cover of the retort must be
carefully luted on with stiff mud or other suitable seal and prefer-
ably should be clamped with an asbestos joint. A stiff wire
should be available for clearing obstructions in the delivery pipe,
particularly if the amalgam contains base metals such as anti-
mony or zinc,

The entire retort should be heated gradually to redness. A
temperature of about 800C., slightly below normal cupellation
heat, is sufficient, provided the entire retort is so heated. The
mercury 1 distills over and condenses and is delivered in a series
of drops. After the steady drip of mercury ceases to come over,
the heating is continued for J^ hr. to 1 hr. longer at full tempera-
ture, after which the firing is stopped, and the retort is allowed
to cool well below the boiling point of mercury (357C.) before
the condenser is disconnected and the retort is opened. The
condenser is swabbed out with a rag tied to a rod to collect all the
mercury.

The normal amount of mercury remaining in the bullion is
from 3 to 5 per cent of the weight of the retorted metal. The
weight of the retorted metal plus the recovered mercury is usually
from 1 to 2 per cent less than the original weight of the amalgam.

1 Mercury is colloquially known as "quick," from quicksilver.

THE PREPARATION OF GOLD AND SILVER BULLION 245

The loss is caused by free moisture in the amalgam, vapor-pres-
sure loss of mercury, slight leakage of mercury, and slight absorp-
tion of mercury in the condenser. Higher losses should be
investigated.

The total time for retorting varies from 2 hr. for small lots of a
few pounds of amalgam up to 7 or 8 hr. for large lots.

Bullion Melting. Bullion, whether from amalgam retorts,
gravity concentration, or other sources, may be put into more
acceptable form for marketing by melting and partial refining
followed by casting into bars. One reason for the practice of
bullion melting at the mine is that the danger of theft of gold is
less when readily identifiable bars of gold are produced than when
loose gold or retort sponge is shipped. Furthermore, bullion
melting removes some of the impurities, increasing the market
value of the product and decreasing express charges, and also
provides a product that is amenable to accurate sampling and
assaying.

Bullion melting is usually conducted in graphite crucibles,
which are actually a mixture of graphite and fire clay. If oxidiz-
ing fluxes are to be used, a fire-clay lining or a fire-clay crucible
fitted into the graphite crucible is preferable. Graphite crucibles
should be thoroughly dried and annealed at a low temperature
before using for the first time, otherwise they may spall badly
due to the sudden expulsion of water vapor.

In melting retort sponge, or clean free-gold concentrates, the
only flux needed is borax. The amount of flux to use can be
ascertained by small-scale trials in an assay crucible and is usu-
ally from 5 to 15 per cent of the weight of the bullion. If much
silica is present, sodium carbonate should be added. Some
operators charge and melt the flux before charging the bullion,
others mix and charge the flux and bullion together. The tem-
perature required for melting is in excess of 1100C., and the
operation should be completed in about 45 min. after charging
into a hot furnace. If a longer time is required, either the tem-
perature was too low or incorrect fluxes were used.

A clean borax slag is very fluid and tends to follow the bullion
into the molds on pouring. This does no harm, as the slag can
be hammered free after solidification, but some operators prefer
to pour off part of the slag into a separate mold, then thicken the
remaining slag with silica or bone ash so that it can be held back,

246 FIRE ASSAYING

or skimmed off separately. The second slag is remelted with the
next charge, and if silica was used to thicken it instead of bone
ash, it is more easily fluxed, using sodium carbonate in proper
proportions. Before discarding slags they should be crushed and
panned to remove prills and shots of gold. It is good practice
to return the crushed slag, after panning, to the mill feed, unless
it interferes with cyanidation or flotation, in which case the slags
should be accumulated for shipment to a smelter or for separate
treatment by gravity concentration or in an amalgamating
barrel.

Bullion molds should be coated with chalk or iron oxide or
swabbed with oily waste before use, to prevent the bullion from
adhering to the mold.

The melting of impure materials, as, for example, free-gold
concentrates containing arsenic, antimony, or other base metals,
requires special oxidizing fluxes. The general principles are dis-
cussed in connection with the melting of cyanide precipitates.

Melting of Cyanide Precipitates. Appropriate melting prac-
tices for cyanide precipitates and impure free-gold concentrates
depend upon the nature and amount of impurities present. The
precipitates from efficient modern plants are often quite clean
and may require only 2 to 5 per cent of borax flux, by weight of
precipitate. If the cyanide solution prior to precipitation is
fouled with base metals, or is imperfectly clarified, the precipi-
tates will be dirty, and a considerable quantity of fluxes, together
with an oxidizing agent, is required. Small-scale tests in an assay
crucible may be necessary to ascertain the best flux combination
for melting. Preliminary sulfuric acid treatment is sometimes
needed to remove excessive zinc, as from zinc-box precipitates or
zinc shorts. (Caution! If the precipitates contain arsenic, take
care to vent fumes, which may contain arsine, a powerful poison.)
With ordinary precipitates, the material may be roasted in the
drying pan to effect sufficient oxidation of zinc to facilitate direct
melting without acid treatment.

The flux for precipitates should make an acid slag in sufficient
volume to hold the oxidized base-metal impurities. A combina-
tion of borax glass, sodium carbonate, and silica sand is generally
used. The grade of the bullion can be raised in the melting
operation by adding an oxidizing agent. Potassium nitrate
or manganese dioxide is used for this purpose. Manganese

THE PREPARATION OF GOLD AND SILVER BULLION 247

dioxide is generally avoided with silver precipitates because of
the belief that it increases the silver loss in the slag. Typical
flux proportions are given in Table XXIII.

TABLE XXIII. TYPICAL FLUXES FOR CYANIDE PRECIPITATES

Pounds of flu

x per 100 Ib. c

f precipitate

Constituent

Yellow
Aster*
(gold)

Mountain
Copper f
(gold)

Typical
silver (over
800 fine)

Borax glass

23-35

5-10

Borax

Silica . .

3-5

Sodium carbonate
Fluorspar

15
1

3-5

Niter

10-20

Manganese dioxide

Total

87-109

105

11-20

* FROLLI, A. W., Open-pit Mining and Milling Methods and Costs at the Yellow Aster
Mine, U.S. Bur. Mines I.C. 7096, p. 41, 1940.

t YOUNG, G. J., Cyaniding Low-grade Gold Ore, Eng. Min. Jour., vol. 131, No. 12, p. 563,
1931.

Gold is melted in crucible furnaces. Graphite crucibles are
used except when a considerable amount of an oxidizing agent is
present in the charge. When oxidizing agents are added, a fire-
clay-lined crucible should be used. Various other types of fur-
naces, designed to handle large amounts of material with greater
ease and fuel economy, are used for silver.

The precipitate and fluxes are mixed together, and the mixture
is commonly placed in paper bags and charged to the hot crucible
with a pair of tongs. After melting, the bullion is cast and the
slag treated as described under bullion melting.

CHAPTER XVI
ASSAY EQUIPMENT AND SUPPLIES

Factors affecting the selection and operation of assay equip-
ment and supplies are dealt with in this chapter. Detailed
descriptions and illustrations of manufactured equipment are
omitted, because up-to-date information can be obtained from
catalogues issued by assay supply houses.

The first step in planning the installation of an assay office is to
request catalogues and price lists from the distributors of assay
equipment and supplies serving the territory in which the assay
office is to be established. A representative list of the leading
assay supply houses of the world, as of 1940, is given in Appendix
C. To aid in ordering equipment and supplies for the establish-
ment of an assay office, the minimum items of equipment for a
small assay office and their approximate costs are listed in
Appendix A. In Appendix B the minimum quantity lots and
costs of assay supplies arid reagents for 1,000 assays are listed.

ASSAY FURNACES AND ACCESSORIES

Assay furnaces are of two general types: (1) crucible or pot
furnaces for melting only, in which the crucible is in contact with
the fuel or the flame; and (2) muffle furnaces, needed primarily
for cupellation but also used for melting, in which the contents
are heated in an enclosed muffle out of contact with the flame.
One of the popular types of furnaces for small assay offices com-
bines a melting chamber for crucibles and a small muffle for cupel-
lation and scorification. Another modification is the use of -plate
muffles, which have some of the advantages of a full muffle at
some saving in maintenance cost and a slightly greater capacity.

Large muffle furnaces are preferred by most assayers for both
fusion and cupellation where the volume of work is sufficient to
justify their use. To conserve floor space and to secure fuel
economy, two muffles may be built into a single furnace, either
side by side or superimposed, or three muffles may be arranged

248

ASSAY EQUIPMENT AND SUPPLIES 249

with two below and one above. These designs have been used
frequently with furnaces that are fired with coal or wood, but
separately heated muffles are preferred with gas- or oil-fired fur-
naces, to secure greater operating flexibility and more accurate
temperature control.

Fuel and Furnaces. Assay furnaces may be designed to
utilize any desired fuel, or may be heated by electricity. Fur-
naces heated by any of the solid fuels wood, coal, charcoal, or
coke are usually built in place, whereas furnaces fired with oil,
gasoline, or gas are usually purchased ready-made, complete
with burners and other accessories, but may be made locally or
converted from a furnace formerly fired with solid fuel. Electri-
cally heated furnaces may be obtained on special order from
various manufacturers, or may be made locally and the resistor
units and electrical control purchased separately.

The relative cost of different fuels is often compared on the
basis of the cost per million B.t.u. For example, if a certain
grade of bituminous coal contains 11,000 B.t.u. per pound and
costs $12 per short ton delivered at the assay office, the cost per
million B.t.u. is $0.55. A light Diesel or stove oil at 8 cts. per
gallon having a specific gravity of 0.85 and containing 19,500
B.t.u. per pound costs $0.58 per million B.t.u. Electricity at
1 ct. per kw.-hr. (1 kw.-hr. = 3413 B.t.u.) costs $2.93 per million
B.t.u. Since the efficiency of the various heating media varies
widely, a cost comparison based on theoretical heating value is
misleading.

Solid fuels are the least efficient, because much heat is lost in the
waste gases, ash, unburned fuel, and in starting and stopping the
furnace. From the operating standpoint, furnaces fired with
solid fuel require more attention and respond more slowly to
temperature adjustment than other types. Hence, such furnaces
are advantageous only where the unit cost of fuel per million
B.t.u. is much less than that of other available fuels, and espe-
cially where they are to be used" for a large volume of work and
are kept under continuous fire for relatively long periods of
time.

Bituminous coal or wood produces a long flame and is preferred
to other types of solid fuel for muffle furnaces. Anthracite coal,
coke, or charcoal produces a short flame, requiring that the muffle
be surrounded with burning fuel.

250

FIRE ASSAYING

A typical design of a two-muffle soft-coal-fired furnace is shown
in Fig. 18. The coal consumption is approximately 45 Ib. per
hour. Each muffle holds fifteen 20-g. crucibles, allowing room
for a heater brick or a row of empty crucibles in the front. The
temperature of the upper muffle will be considerably less than
that of the lower muffle, which disadvantage is eliminated in
designs providing for placing the muffles side by side. For wood

FIG. 18. Two-muffle coal-fired assay furnace. (After Fulton and Sharwood,
Manual of Fire Assaying.)

firing the grate should be 8 to 10 in. below the bottom of the fire
door to provide a firebox sufficiently deep for good combustion
and uniform heating.

Pulverized coal has been used in a few installations and is
nearly as satisfactory as a gaseous fuel, except for the accumula-
tion of ash in the furnace and the comparative complexity of the
pulverizing and blowing equipment.

Liquid or gaseous fuels have a much higher combustion effi-
ciency than solid fuels and are much more satisfactory for assay
service because of ease and cleanliness of operation, quick adjust-
ment of temperature, and saving of fuel when not in actual use.

ASSAY EQUIPMENT AND SUPPLIES

251

Gasoline is used only for small furnaces, which are usually of the
combination type. Light Diesel or other fuel oil in the range of
24 to 35A.P.I. (American Petroleum Institute) 1 gravity is
widely used throughout the Western United States and is gener-
ally preferred to heavy fuel oil of 12 *to 14A.P.I. which
requires preheating to secure atomization and tends to clog
burner orifices. Natural gas, where available, is an excellent

AjiiB
i fil

FIG. 19. Braun gas-fired combination furnace. (Photograph by W. Vernon
Smith, Sacramento Junior College.)

fuel. Manufactured city gas, butane, and compressed gas in
cylinders are used in some localities, but the cost is high relative
to other fuels.

A popular type of gas- or gasoline-fired combination furnace
widely used in small assay offices is illustrated in Fig. 19. This
furnace is made in two sizes. The smaller furnace (Braun Type
42) holds six 20-g. crucibles and the larger one holds ten 20-g.
crucibles, and the muffle sizes are 4% by 8 by 3 in. and 6 by 10
by 4 in., respectively. The average gasoline consumption is
K to Y gal. per hour, and the consumption of natural gas (1150
B.t.u. per cubic foot) is about 50 cu. ft. per hour. The same

1 Specific gravity 141.5/(131.5 + A.P.L).

252

FIRE ASSAYING

furnace is available for firing with kerosene or light fuel oil, or
with natural, city, or cylinder gas.

Several successful types of oil-fired muffle furnaces have been
placed on the market, one of which is illustrated in Fig. 20. This

F 1G . 20. Denver oil-fired muffle furnaces. (Photograph by Eric Nordman,

Stanford University.)

furnace uses light Diesel or fuel oil, atomized with air. Two
sizes are available, the smaller of which has a 12- by 20- by 7%-in.
(size LF) muffle, and the larger a 14- by 20- by 6K-i n - (size LG)
muffle. The effective capacities are fifteen and twenty 20-g.
crucibles, respectively. Oil consumption under full fire is
-approximately 2.5 gal. per hour, or 1.5 gal. per hour at operating
temperature.

Electric assay furnaces are ideal from the standpoint of con-
venience, cleanliness, absence of noise, accuracy of temperature
control, and uniformity of heating throughout the heating cham-
ber. They also give up to 40 per cent longer life to crucibles.
Although the first cost of the installation is more than three times

ASSAY EQUIPMENT AND SUPPLIES

253

that of other furnaces, the maintenance costs are reasonably low.
In regions where electric power can be obtained for let. per kilo-
watt-hour or less, the actual cost of heat compares favorably
with that of fuel-fired furnaces because the thermal efficiency of
the electric furnace is from five to six times as great as that of a

FIG. 21. Hevi-Duty electric assay furnace. (Photograph by D. G. McAllister,
Pacific Scientific Co., San Francisco.)

fuel-fired furnace owing to the absence of products of combustion
and the compactness of design, which permits good insulation
and close proximity of the working chamber to the source of heat.
In 4 or 5 years' operation at full capacity for 8 hr. per day, the
saving in crucible consumption may offset the higher first cost of
the furnace. The effective capacity of an electric furnace of
given hearth dimensions is greater than that of fuel-fired furnaces,
especially for cupellation, because of more uniform heating from

254 FIRE ASSAYING

front to rear of the working chamber. Furthermore, fusions and
cupellations can be made by unskilled operators, as no judgment
is required to control temperatures within the desired limits.
Assaying accuracy is improved by the maintenance of nearly
ideal temperature cycles for cupellation and the ability to repro-
duce the same cycle from day to day. The chief disadvantage
of electric assay furnaces, aside from their first cost, is the greater
amount of time required to bring them to operating temperature.
For example, a 12-kw. furnace with net hearth dimensions of
15J^ by 20% in. requires from 1^ to 2 hr. to reach a temperature
of 1050C., whereas an oil-fired furnace with a muffle 14- by 20-in.
hearth area will reach the same temperature in % to 1 hr. This
time loss is unimportant, as the assayer is usually occupied with
the preparation of charges and other matters while the furnace is
heating, and in most localities the furnace could be started at an
earlier hour by a watchman, janitor, or other employee. A
typical electric furnace is illustrated in Fig. 21.

The Hevi-Duty Electric Co. of Milwaukee, Wis., has recently
installed electric assay furnaces at a number of plants. The
Hevi-Duty furnace is heated with a rod type of metallic heating
element and is equipped with an autotransformer, thermocouple,
and potentiometer-type automatic temperature controller. A
similar furnace has also been designed by the Denver Fire Clay
Company of Denver, Colo. The Carborundum Company of
Niagara Falls, New York, manufactures carborundum heating
elements, known as Globar elements, but does not manufacture
complete furnaces. A number of mining companies have used
Globar elements in furnaces that have been constructed
locally.

Table XXIV gives data on the size, cost, capacity, heating
rate, and average fuel or electric consumption of various types of
assay furnaces. In some cases the effective crucible capacity
given in the table is less than manufacturers' ratings, because the
manufacturers seldom allow for the fact that the front 3 or 4 in.
of a muffle furnace is too cool to be used for melting crucible
assay charges. Most assayers use a brick or a row of empty
crucibles in front of the muffle to act as a warmer for the charged
crucibles. In calculating cupel capacity, ample allowance must
be made for portions of the furnace that are unusable on account
of too high or too low temperatures. If plate muffles are used in

ASSAY EQUIPMENT AND SUPPLIES

255

ll ftl

K op fl *
H > o oo

s.s .
ll ISA

1 ?

'&
O

: S

' t^

- O O i"H >O O <

W <N <N (N <N (

8;8 8888 8

CCO OO>OiOO O

OOCO OO"^<OI>- O

8 88

256 Ft HE ASSAYING

a well-designed gas-fired furnace, almost the entire hearth area
can be used for cupellation, but in most fuel-fired furnaces, if the
highest silver accuracy is desired, it is unwise to attempt to cupel
more than three or four rows of cupels in depth, otherwise the rear
rows will be too hot even though the front row is cupeled at the
minimum possible temperature. If gold only is being assayed, a
wider temperature range is permissible in cupellation, and some
furnaces may be almost completely filled with cupels, except for
the front two or three rows, and a clearance of 2 or 3 in. from the
rear wall of the muffle. Electric furnaces have a lower tempera-
ture gradient from front to rear, with correspondingly greater
cupellation capacity.

Muffles. Manufacturers of assay clay goods produce a num-
ber of stock sizes of fire-clay muffles. The short life of fire-clay
muffles, especially when used for fusions as well as for cupellation,
has led to the extensive adoption of muffles made of silicon car-
bide, known under the trade name of Carbofrax or Crystolon.
These muffles have an indefinitely long life if properly supported
in the furnace, as durable patches can be made with silicon car-
bide cement. Silicon carbide has a higher heat conductivity and
a lower specific heat than fire clay, and hence muffles of this
material heat more rapidly than fire-clay muffles and are more
responsive to heat control. The cost of silicon carbide muffles is
from three to four times that of fire-clay muffles.

An increasing number of large assay offices have adopted plate
muffles, which consist of a hearth with a low flange around the
two sides and rear. The advantages of this muffle over the com-
pletely enclosed type are greater capacity per unit of hearth area
and less maintenance cost. The greater capacity is obtained
through the possibility of setting crucibles close to the outer
edges of the plate, so that the upper part of the crucible, which is
wider than the base, actually projects beyond the edges of the
muffle plate. Furthermore, large crucibles can be loaded with-
out the necessity for using exceptionally high muffles, provided
,that the door and the heating chamber are sufficiently high.
Plate muffle furnaces are fired with gas or oil and are also used in
electrically heated furnaces. One operator using a gas-fired
plate muffle furnace reports that the heat distribution is so uni-
form that the entire furnace, with the exception of about 3 in. at
the front, is used for cupellation.

ASSAY EQUIPMENT AND SUPPLIES 257

Furnace Refractories. Fire-clay refractories are generally
used for furnace linings and muffle supports in fuel-fired furnaces.
The furnace should be designed so that all destructible parts can
be readily replaced. Muffle replacements are the most common
when fire-clay muffles are used. Muffle supports and flame
baffles also deteriorate rapidly. If silicon carbide muffles are
used it is recommended that muffle supports and flame baffles also
be made of the same material, which will well repay their extra
cost in increased life and freedom from the annoyance and expense
of frequent furnace repairs. If prefabricated supports are not
available, they can be formed from Carbofrax or equivalent
cement in plaster molds cast from models of the required shapes,
allowing from 5 to 8 per cent linear drying and firing shrinkage.
After drying, the shapes may be fired by placing them in the
assay furnace and gradually increasing the temperature up to the
maximum operating temperature over a period of 6 or 8 hr., then
holding at maximum temperature for 1 hr. or more.

In a few installations the annoyance of muffle rest failures has
been eliminated by the use of water-cooled pipe supports.

Furnace parts not subjected to bearing loads, corrosive slags,
lead or litharge fumes, or to the direct impingement of a gas or oil
flame require but little repair. The resistance of fire-clay
refractories to corrosive slags or flame action can be increased by
coating them with a wash of Carbofrax or other high-temperature
cement.

The side covers of small combination furnaces are subjected to
severe spalling conditions, on account of the necessity of removing
them when crucibles are loaded into and removed from the fur-
nace. By removing only the front covers on each side of the
furnace, the life of a set of four covers can be nearly doubled, and
old, badly damaged covers may be used in the rear positions, as
is shown in Fig. 19.

Elements of Assay-furnace Design. The special service
requirements of assay furnaces govern their design. The impor-
tant factors to be considered are rapid heating, rapid and accurate
temperature control, manipulative convenience, and easy replace-
ability of destructible parts. Fuel economy is of subordinate
importance as the cost of crucibles and other supplies, the operat-
ing labor cost, and the importance of accuracy and efficiency far
outweigh extreme costs of fuel or electrical heating.

258 FIRE ASSAYING

The typical daily cycle of an assay furnace is rapid heating
from a cold start to a temperature of 1000 to 1100C., followed
by a period of several hours' operation at that temperature until
all fusions are completed and then several hours' operation for
cupellation. Each cupellation cycle should start at approxi-
mately 900C. muffle temperature, drop to 810 to 840C. during
the greater part of the cupellation process, and then rise to 850
to 870C. at the finish. The initial heating period should not be
unduly prolonged, and 1% to 2 hr. is the maximum time allow-
able in most cases, unless someone is delegated to start the furnace
before the assayer starts his day's work. During the fusion
period of the day the furnace is being charged with a load of cold
crucibles and their contents at intervals of 25 to 30 min., requir-
ing a rapid heat input to heat the crucible contents to melting
temperature. 1 The drop in temperature after the start of cupel-
lation should be effected in 4 or 5 min. or less, and the increase at
the finish should be done in the shortest time possible.

To obtain the desired flexibility of temperature control the fur-
nace must not only be provided with a great excess of fuel-burning
(or electrical heating) capacity, but the thermal characteristics
of the furnace walls must be properly designed. The optimum
furnace-wall design is obtained with a relatively thin interior
lining of firebrick or other high-temperature refractory having a
relatively low thermal conductivity, surrounded by one or more
layers of insulating material. If too much firebrick is used, the
heat capacity of the furnace walls becomes excessive, rendering
the furnace sluggish in responding to changes of firing and requir-
ing a prolonged period of firing to reach equilibrium at a given
temperature. At the other extreme, if the refractory lining is
too thin, or has too high a heat conductivity, and no insulation
is used, heat losses are excessive, the furnace does not heat
uniformly, and temperature control is too erratic. An appropri-
ate degree of insulation conserves heat and prevents sudden
fluctuations of temperature. Just enough refractory protection
'should be used on the interior so that the temperature of the
refractory in contact with the insulation does not exceed the safe
working limit of the insulating material.

1 Twenty-four average fusions in 20-g. crucibles require about 10,000
B.t.u. to raise their temperature from the cold to 1050C.

ASSAY EQUIPMENT AND SUPPLIES 259

Most of the manufactured oil- or gas-fired assay furnaces have
firebrick linings 3 or 4 in. thick, surrounded by a K- to 1-in. layer
of insulating material such as diatomite or mineral wool, the
whole enclosed in a sheet-steel shell. Electric furnaces are
usually lined with molded aluminum oxide or silicon carbide,
backed with several inches of insulating material. The insula-
tion is usually applied in layers, using several grades, each of
which is best suited for the particular temperature range to which
it is to be subjected.

Furnace Tools. Manufacturers 7 catalogues should be con-
sulted for illustrations and prices of the common furnace tools
required for the handling of crucibles, scorifiers, and cupels. Any
of these tools can be made by a local blacksmith if desired, often
at considerable saving in cost. The minimum complement of
furnace tools and accessories required for a small assay office is
listed in Appendix A, and photographs showing typical assem-
blies are shown in Figs. 19 and 20.

For handling a large volume of work a number of specially
designed devices have been developed, some of which have
already been described in the chapters on cupellation and the
crucible assay. For keeping a furnace load of crucibles in order
for charging with fluxes and loading into the furnace, a wooden
tray having the same area as the firing chamber of the furnace is
convenient. Such a tray is illustrated in Fig. 12, and its use is
described on page 154. A multiple charging fork such as that
illustrated in Fig. 20 is very efficient for charging an entire row
of crucibles into a muffle furnace, and a single fork of similar
design is often used for withdrawing and pouring crucibles. For
placing a number of cupels in the furnace at a single operation, a
simple tray with pusher is commonly used. Multiple button-
charging devices have been developed at a few custom-smelter
assay offices where an exceptionally large volume of assaying is
involved.

Two common types of button molds are shown in Figs. 19 and
20. A different type of mold is illustrated in Fig. 13, and the
practice of pouring directly on a flat plate is described on page
160.

For pounding buttons, a heavy steel plate, a small black-
smith's anvil, or a short section of rail may be used.

260 FIRE ASSAYING

Equipment for Sample Preparation. Manufacturers' cata-
logues present a wide range of crushing and grinding equipment
for the preparation of assay samples. For small-scale work a
hand-power jaw crusher and bucking board and muller are suf-
ficient, but power-driven equipment is necessary for efficient
operation if more than 15 or 20 samples per day are to be pre-
pared. For coarse crushing, jaw crushers are generally used, but
various types of cone crushers are also available. For fine grind-
ing, disk pulverizers are popular, but cone and ring grinders are
used in some localities. For quantity production, especially
with large samples, crushing rolls are very desirable for inter-
mediate crushing. A bucking board should always be provided
for grinding small lots.

In planning the layout of a sampling room, care should be taken
to provide ample workbench area, so that samples can be mixed
with a mixing cloth with full freedom of action. It is desirable
that the top of the mixing bench be covered with sheet metal,
for convenience of cleaning. The sampling room should be iso-
lated or partitioned off from other parts of the assay office to
avoid salting from dust. A suction fan over the crushing units
is desirable, and a compressed-air nozzle for cleaning the crusher,
benches, pans, scoops, etc., is a sine qua non in a busy assay office.

A drying rack should be provided in a room other than the
crushing room, with sufficient capacity for the daily array of
samples.

Sample pans, scoops, brushes, riffles, and other incidental
equipment are selected in accordance with anticipated needs.

A suggested list of equipment for sample preparation is
included in Appendix A.

ASSAY SUPPLIES AND REAGENTS

Assay crucibles, scorifiers, cupels, and parting cups are
described elsewhere in the text. Manufacturers' catalogues
should be consulted for details of size, shape, and prices. A list
of the essential supplies for 1,000 assays is given in Appendix B.

Assay Reagents. The nature and function of all necessary
assay reagents are discussed throughout the text in connection
with their specific applications. The minimum number of
reagents and suitable quantities of each for 1,000 assays are given
in Appendix B.

ASSAY EQUIPMENT AND SUPPLIES 261

The principal assay reagents, especially sodium carbonate,
litharge, borax, borax glass, granulated lead, and sheet lead, are
available in special grades for assaying, the chief requirement of
which is that they be free from, or contain a uniformly distributed
small amount of, silver and gold. Two grades of litharge are
obtainable, commercial and silver-free. The commercial grade
costs about half as much as the silver-free grade and is generally
acceptable if care is taken to thoroughly mix each new lot and
obtain an accurate assay of its silver and gold content to be
applied as a correction to all assays. Corrections of the order of
0.1 to 0.2 mg. of silver and not over 0.0025 mg. of gold per assay
ton are normal, but occasional lots of litharge may contain
greater amounts of precious metals. Assay granulated lead is
usually nearly free from precious metals, but lead foil frequently
contains some silver, and the silver content may vary in different
parts of the same roll. Assay reagents other than lead or lith-
arge are generally quite free from precious metals, but it is advis-
able as a routine precaution against accidental or intentional
salting to assay new lots of all reagents.

Besides the specific reagents mentioned in Appendix B, certain
additional reagents and substitutes are favored by some assay ers.
Red lead, the peroxide of lead, approximately Pb 8 O 4 , is sometimes
used instead of litharge but is usually more costly per unit of lead
and has no advantages compared with litharge.

Sodium bicarbonate was formerly favored by assayers for some
applications instead of the normal carbonate but contains only
63 per cent as much Na 2 O and gives off nearly twice as much gas
per unit weight. Potassium carbonate, replacing part or all of
the sodium carbonate, gives a more powerful fluxing action on
some ores, but its use is seldom warranted on account of greater
cost.

The choice between borax and borax glass is partly a matter of
cost per unit of Na2O.2B 2 O3 and partly a question of avoiding
the undesirable effect of the evolution of water vapor from borax
in assay fusions. Borax is much cheaper and may be used as a
cover, but the glass should be used if it is to be mixed with the
charge.

Lime (CaO) or limestone (CaCO 8 ) is useful in fluxing ores high
in alumina, as noted in Table XI. Smelter-flux limestone or
commercial lime are suitable for assay purposes, after grinding.

262 FIRE ASSAYING

Fluorspar (CaF 2 ) is favored by many assayers to improve the
fluidity of some slags, for example, when bone-ash cupels are
fluxed, but is of doubtful value, as it merely introduces an inert
diluent in the slag, and sufficient fluidity can always be obtained
by proper flux balance without fluorspar.

The selection of reducing agents has been discussed in Chap.
VII, and various reducing agents are listed in Table XIII.
Flour is commonly the cheapest and most generally available car-
bonaceous reducing agent. It is desirable to have a small stock
of powdered sulfur and iron filings on hand for experimental
purposes, as on some ores, one or the other of these reducers may
prove more effective than carbonaceous reducing agents.

Summary of Equipment and Supplies Cost. From Appendix
A, it will be noted that the minimum cost of assay equipment
for a small assay office to handle 30 to 40 assays per day with
hand-power crushing and grinding is approximately $600 f.o.b.
shipping point. With motor-driven crushing and grinding equip-
ment, the minimum cost is $860. These figures do not include
shipping costs, sales tax, if any, or the cost of housing and
installation of equipment. The minimum floor space required
without uncomfortable crowding is about 225 sq. ft., which ghould
preferably be partitioned into a sampling room, a furnace room,
and a balance room. A separate storeroom is also desirable.

Supplies for a minimum of 1,000 assays cost approximately
$150 f.o.b. shipping point. The quantities recommended in
Appendix B are selected on the assumption that there will be
about an equal number of siliceous ore and pyritic concentrates
to be assayed and should be adjusted in accordance with tho
anticipated character of the samples. After the first 1,000 or
more assays have been made, the assayer can readily calculate
the actual average consumption of supplies, and order accord-
ingly in the future.

APPENDIXES

APPENDIX A

MINIMUM EQUIPMENT FOR A SMALL
ASSAY OFFICE

Fire assay for gold and silver only.

Capacity 30 to 40 assays per day, with one assayer and helper, including
sample preparation and cupel making.

Benches, tables, and machinery installation not estimated.

* Articles marked with an asterisk can often be made better and more
cheaply locally.

Approximate

Cost, F.o.b.

Shipping

Point
A-l : Sample grinding by hand power.

1. Hand-power jaw crusher, 6- by 3-in. jaw $ 75.00

2. Bucking board and muller, 18 by 20 in 12 .50

3. Paint brush, 3 in 1.25

Total, hand-grinding equipment $ 88 . 75

A-2: Sample grinding by electric power.

1. Jaw crusher, 6- by 3-in. jaw, tight and loose pulley. . . . $125.00

2. Disk pulverizer, tight and loose pulley 135.00

3. 2-hp. motor, 1,750 r.p.m., with starter and base 65.00

4. Line shaft, pulleys, bearings, coupling, safety collars,

grease cups, and belts 19 . 50

5. Bucking board and muller, 18 by 20 in 12 50

6. Scoop for feeding pulverizer* 1 .25

7. Paint brush, 3-in 1 . 25

Total, power-grinding equipment $359.50

B. Sampling equipment.

1. Jones rime, 8 by 8 in., with four sample pans, scoop, and

cleaning brush (optional) $ 14 . 65

2. Tyler standard sieve, 8-in., 100-mesh 5.25

3. Tyler sieve-bottom j>an, 8-in 2 .40

4. Tyler sieve cover, 8-in. (optional) 1 .25

5. 1 doz. (minimum) sample pans, approximately 1 qt.

capacity and 6 in. in diameter and 1% in. deep. Pie
tins are suitable for samples to be dried ; bread tins are
useful for temporary storage of samples 1 .50

6. 4-in. spatula, stainless-steel blade .40

7. 6-in. gold pan (for general use) 0.40

Total, sampling equipment $ 25 . 85

C. Fluxing and furnace-room equipment.

1. Flux box (locally made) $ 5.00

2. Measuring scoops, see page 156

3. Pulp balance, 150-g. capacity, 2 mg. sensitivity, open

type (or 60-g. balance, 1 mg. sensitivity at slightly

lower cost) 24 .00

264

APPENDIX A 265

Approximate

Cost, F.o.b.

Shipping

Point

4. Set of weights, assay ton, 0.05 to 4 A.T $ 8.00

5. Set of weights, metric, 10 mg. to 60' g 2 .60

6. 1 pt. aluminum cocktail shaker, with sieve removed or

sealed. (Optional, for quick mixing of charges, see

page 167)..... 0.60

7. Braun combination assay furnace, type 42] 4^- by 8-

by 3-in. muffle, gasoline-fired. (See Table XXIV

for alternative choices.) 83 .00

8. Carbofrax muffle for furnace 4 .50

9. Table-model cupel machine, \y- and 1^-in. dies

(suitable table to be designed, also provide drying

rack for cupels. Omit if cupels to be purchased) .... 45 .00

10. Crucible tongs* 1 .25

11. Scorifier tongs for 2K- or 3-in. scorifiers* 1 .25

12. Cupel tongs* 1 .00

13. 6-hole pouring mold .75

14. Button forceps, 6-in . 10

15. Button brush, stiff, double-end .75

16. Cupel tray, for 16 cupels ' .85

17. Ball-peen hammer, 16-oz .65

18. Blacksmith's anvil, 70-lb., or a short length of railroad

rail, or use top of pouring table, if covered with heavy

steel plate* 12.50

Total, fluxing and furnace-room equipment $191 .60

D. Bead handling and parting equipment.

1. Curved pliers, 5-in $ .70

2. Bead anvil, 2J4 by 1% by J in., hardened tool steel*. 0.30

3. Bead hammer, or square-headed tack hammer 0.50

4. Two parting-cup trays, for 12 cupels each* 2 .00

5. Nichrome crucible tongs, 9-in., for handling parting cups 3 .25

6. Electric or other hot plate or stove, 8-in. round or larger. 11 .75

7. Still, heated by gasoline, gas, or electricity, or on a stove.

minimum capacity 1 qt. per hour (omit if distilled

water to be purchased) 30 .00

8. Gasoline blow torch with vertical burner, stand and

nichrome triangles or nichrome wire gauze (the assay
muffle is more convenient for annealing sets of gold
partings, but a good burner is essential lor other pur-
poses, and is useful for annealing when the furnace is

not in operation) 10 .00

Total, bead-handling and parting equipment $ 58 . 60

E. Bead-weighing equipment (should be in separate room).

1. Keller assay balance, type 3B, or equivalent (provide

substantial bench on piers separate from building). . . $200.00

2. Set fractional weights, 1 mg. to 1 g., with 2 1-mg. riders.

"A" grade, having a 0.005-mg. tolerance is preferred,
but "B" grade with a 0.01-mg. tolerance can be sub-
stituted at about two-thirds of the cost 22 .50

3. Curved forceps, brass with ivory tips, 4 in. long 1 .00

Total, bead-weighing equipment $223 .60

Total equipment cost, using hand crushing and grinding $588 .20

'')tal equipment cost, using power crushing and grinding $858.95

APPENDIX B
LIST OF ASSAY SUPPLIES AND REAGENTS

Nearest minimum quantity lots for 1,000 assays, with equipment listed in
Appendix A. Includes replacement parts that should be carried in stock for
emergency repairs.

Approximate

Cost, F.o.b.

Shipping

Point

A. Glassware and porcelainware.

60 porcelain crucibles (parting cups), size 00 $ 7 .00

6 porcelain crucibles, size 1, for bullion parting 1 .20

2 150-ml. beakers, pyrex .38

2 400-ml. beakers, pyrex . 52

2 1,000-ml. beakers, pyrex 1 .00

4 2^-1. bottles. Need not purchase, as temporary bottles
can be used until empty acid bottles are available. Acid
carboys of 5 or 12 gal. capacity are necessary for parting
acid and water storage for quantity production
1 wash bottle, consisting of a 1,000-ml. pyrex flash with
wicker- or asbestos-covered neck and 2-hole No. 6 rubber
stopper. Make blowtube and delivery tube and nozzle
with glass and rubber tubing. For quantity work, an over-
head hot distilled water system is indispensable, for which
order more glass and rubber tubing than listed 1 .60

1 graduated cylinder, 1,000-ml 2 .00

2 glass funnels, 3 in. diameter, 6-in. stems .44

% Ib. glass tubing, 6 mm. and 8 mm. outside diameter,

assorted .50

> Ib. glass rod, 4 to 6 mm. diameter 0.20

Total glassware and porcelainware $ 14 .84

B. Clay goods.

384 (4 cartons) clay crucibles, 20-g., or 384 (3 cartons) 15-g.
crucibles if most assays are on ^-A.T. samples of simple
ores, for which deduct 80 cts. per 100 $ 27 . 85

84 (1 carton) clay crucibles, 30-g 7 .75

126 (1 carton) clay crucibles, 10-g. Omit if preliminary
assays not needed. Increase if found possible to flux %-
A.T. charges in this size crucible 6 .68

25 scorifiers, 2>^-in. Increase and include 3-in. size if many

scorifications are planned . 80

1 set furnace-renewal parts as follows:

Covers $1 .50

Burner-hole boss . 30

Muffle plug 0.50

Dome.. 8.00

End bricks with muffle rests 4 . 50

Deflecting brick 0.50

Total 15.30

(Bottom and side bricks seldom need to be replaced)
266

APPENDIX B 267

Approximate

Cost, F.o.b.

Shipping

Point

10 Ib. Hytempite or other refractory cement $ 1 .00

5 Ib. Carbofrax cement 0.60

Total clay goods 4 $ 59 .98

C. Reagents and consumable materials.

150 Ib. litharge, commercial for assaying. Increase to 200 Ib.

if high-copper ores are to be assayed $ 13 .25

100 Ib. sodium carbonate (soda ash), 58 per cent, light 2.00

50 Ib. borax glass for assaying (Considerable cost saving if

borax is substituted for borax glass when possible) 8.00

25 Ib. silica, powdered. Increase quantity if most ores to be

assayed are basic . 50

25 Ib. potassium nitrate, powdered, U.S. P. Decrease if most

samples are nonreducing 2 . 50

5 Ib. flour 0.30

5 Ib. granulated lead, c.p., assay grade 1 .50

5 Ib. lead foil, c.p., assay grade 1 .35

1 oz. silver foil, c.p. Increase if many bullion assays required 1 . 50
5 g. proof gold foil. This is sufficient for initial stock for
gold-bullion assays. Additional gold may be recovered
and purified from assays, for which add hydrochloric acid

and oxalic acid to this list 10 .00

4 oz. sheet copper, about 22-gage (|<j 2 m -) or thinner. Com-
mercial sheet metal satisfactory if tested for precious

metals. Omit if no bullion assays required .25

3 7-lb. bottles nitric acid, c.p ". 4. 50

1 Ib. calcium chloride, technical, granulated, for desiccation.

Not necessary in dry climate .25

1 pkg. (1,000) Hermann inquarts 2 .50

50 Ib. bone ash, Denver XX or equivalent. Omit if cupels

purchased and allow $20.00 for 1,000 cupels 5 .00

50 Ib. cement. Omit if cupels purchased 0.60

Total, reagents and consumable materials $ 54 00

D. Miscellaneous supplies.

1 pr. asbestos mittens $ 3. 00

1 pkg. ll-cm. filter papers, No. 1 Whatman or others 0.30

1,000 sample envelopes, Kraft paper, with metal fastener

4 by 7 in., or substitute 1-oz. grocery bags at $0.50 per M . . 7 . 50

2 yd. 20 to 24-oz. rubberized mixing cloth. . . . 3 .00

1 ft. rubber tubing, % e m - inside diameter. For wash bottle

connections. Increase if gravity-feed wash-water system

is used for parting . 10

2 (extra) 1- mg. riders, loop in plane of legs, tolerance 0.005

mg 1.60

2 (extra) 1- mg. weights, tolerance 0.005 mg 2 .00

1 (extra) 2- mg. weight, tolerance 0.005 mg 1 .00

Total miscellaneous supplies $ 18.50

Total cost of supplies for approximately 1,000 assays $147 .32

APPENDIX: c

ASSAY SUPPLY HOUSES

United States
California

Colorado

Idaho

Illinois
Missouri
Montana
New York

Oregon
Texas

Utah

Washington

Canada
Alberta
British Columbia

Manitoba
Ontario

Quebec

Los Ati&eles

San Francisco
Denver

Wallace

Chicago

St. Louis "*-.

Butte

New York City

Portland
El Paso

Salt Lake City

Seattle
Spokane

Swastika
Vancouver

Winnipeg

Cobalt

New Liskeard

Toronto

Montreal

268

The Braun Corporation

The Calkins Co.

Braun-Knecht-Heimann Co.

The Denver Fire Clay Company

Mine & Smelter Supply Com-
pany

Coeur d'Alene Hardware &
Foundry Co.

E. H. Sargent & Company

Henry Heil & Company

Montana Hardware Co.

The Denver Fire Clay Company

Eimer and Amend

J. & H. Berge

Scientific Supplies Co.

The Denver Fire Clay Company

Mine & Smelter Supply Com-
pany

The Denver Fire Clay Company

Mine & Smelter Supply Com-
pany

Braun-Knecht-Heimann Co.

Steward and Holmes Drug Co.

C. M. Fassett & Company

George Taylor Hardware, Ltd.
Marshall- Wells, B. C., Ltd.
Cave and Company
Marshall- Wells Company, Ltd.
George Taylor Hardware, Ltd.
George Taylor Hardware, Ltd.
Canadian Laboratory Supplies,

Ltd.

Fisher Scientific Co.
Canadian Laboratory Supplies,

Ltd.
Fisher Scientific Co.

APPENDIX C

269

Philippine Islands Manila

Baguio, Mt. Prob.
Central and South America

Argentina Buenos Aires

Brazil Rio de Janeiro

Chile Santiago

Colombia Bogota

Mexico Mexico D. F.

Peru
Others
Australia
Belgium

China

Monterrey
Lima

Sydney
Brussels

Shanghai

Dutch East Indies Batavia

East Africa Nairobi, Kenya

Colony
England London

France
Japan

Paris
Tokyo

Marsman Trading Corporation
Fisher Scientific Co.
Philippine American Drug Co.
Marsman Trading Corporation

Eduardo Alvarez de Toledo

Adolpho Botelho

W. R. Judson

Almacen Padco

Hoffmann- Pin ther & Bosworth,

S. A.

Sanford Supply Company
A. y F. Wiese, S. A.

H. B. Selby & Co. Pty., Ltd.
Henkart, Van Velsen and Laou-

reux
Harvey, Main & Co.,

Ltd. in all

Anderson, Meyer and principal

Co., Ltd. cities

Schmidt & Co., Ltd.
N. V. Otto Pfeiffer
Gailey & Roberts, Ltd.

Baird and Tatlock
Gallenkamp and Co.
Poulence Freres
Maison Wiesnegg
Suzuki Sohati Company

INDEX

Accuracy of samples, 15

Acid treatment, choice of acids for,

185

removing impurities by, 184-193
Alkaline salts, composition of, 87,

120

in softening lead, 93-94
Amalgam, retorting of, 242-245
Annealing, gold from parting, 82-83
Ant'mony, assay for, 225-226
behavior in assaying, 91, 93
in captation, 68
roasting for the separation of, 150
slag color, 123

Argol, reducing power of, 100
Arsenic, behavior in assaying, 91, 93
in cupellation, 69
roasting for the separation of, 150
Assay, records of, 11-13
sequence, 10, 152-154
Assay office, arrangement of, 6

location of, 6
Assay portion, methods of taking,

27, 40, 195

size of, for bullion assay, 174
for crucible assay, 40, 124
for scorification, 168
for solution assay, 194-195
Assay ton, volumes, 195
"weights, 12, 39
Assayer, duties of, 6-9
Assaying, fire, definition of, 1
equipment and supplies, 5
general methods, 3-5
metals adapted to, 1
objectives of, 1-2
routine, organization of, 9-10
units used in, 39-40

Assays, combination, general pro-
cedure for, 187-189

Balances, assay, 42
pulp, 40

sensitivity of, 42
Beads, cleaning of, 65
Bismuth, behavior in assaying, 91

in cupellation, 69
Bismuth assay, 225-226
Bone ash, grades of, 49
Briquettes, collector for oxidation

collection, 147
Bullion, copper, 172

crucible assay of, 175
mercury-sulfuric acid assay

method, 191-193
nitric acid assay method, 190-

191
gold, 173

assay of, 180-183
lead, 172

assay of, 174
melting of, 245-247
preparation, from amalgam, 242-

245
from cyanide precipitates, 246-

247'

sampling of, 36-37, 173
silver, 173

cupellation assay of, 176-178

volumetric assay of, 178-180

weights of assay portions of, 174

Carat, 40

Carbon, roasting for the separation
of, 150

271

272

FIRE ASSAYING

Charcoal, reducing power of, 99
Cobalt, slag color, 123
Colorimetric assay, cyanide solu-
tions, 201-202
Coning and quartering, 30
Constitutional diagram, Na2O.SiC>2

SiO 2 , 118
PbO SiO 2 , 119
Control assays, 236-237
Copper, assay for, 224-225
behavior in assaying, 91, 93
crucible charges for ores high in,

142-145

in cupellation, 69
in lead buttons, 164
slag color, 123

tolerance of, in fire assay, 184
Copper bullion, 172

crucible assay of, 175
Copper bullion assay, mercury-

sulfuric acid method, 191-193
nitric acid method, 190-191
Corrected assays, direct, 238241
cupel assays, 240-241
examples of, 241
slag assays, 239
indirect, 237-238
Cost summary, assaying equipment

and supplies, 262
Crucible assay, acid-treatment

method, 151

addition of fluxes, 155-157
for antimony, 225-226
for bismuth, 225-226
charge calculation, basic ores,

138-139

siliceous ores, 136-137
sulfide ores, 139-145
controlled reduction methods,

136-144

for copper, 224-225
fluxing for, 124-129
furnace operations, 157-158
for lead, 220-223

manipulative procedures in, 151-
161

Crucible assay, methods, 135
oxidation-collection method, 147-

149
for platinum-group metals, 205-

206
pouring, on flat plate, 160

in molds, 158-160
roasting method, 149-150
for tin, 223-224
trouble shooting in, 161-164
uncontrolled reduction methods,

145-147
Crucibles, 151-152

handling, 154

Cupellation, cleaning beads from, 65
definition of, 46
effect of impurities in, 67-73
evidence of platinum-group met-
als, 70-73
formation of litharge "feathers"

during, 55, 59, 66
freezing during, 59-60
of gold bullion, 182
loss, of gold in, 46, 51

of silver in, 46, 49, 51
platinum metals, gold-addition

method, 207
post-treatment method, 207-

208

silver-addition method, 207
process, 54-67
of silver bullion, 176-178
spitting during, 54, 57
sprouting of beads from, 63, 64
surfusion during, 61
temperature of, 54-55, 66-67
Cupellation errors, effect, of amount

of lead on, 231

of base-metal retention on, 233
of impurities on, 231
of silver-gold ratio on, 233
of size of bead on, 232
of temperature on, 231
Cupels, absorption of litharge by, 47
assays of, 240-241
bone ash, 49
bone-ash-cement, 49

INDEX

273

Cupels, cement, 49
magnesia, 51
making, 52
shape of, 47-48
types of, 46
Cuprous chloride precipitation assay,

cyanide solutions, 197-199
Cyanide precipitates, acid treat-
ment of, 246
melting of, 246-247
Cyanide solutions, assay of, colori-

metric, 201-202
cuprous chloride precipitation,

197-199
electrolytic precipitation, 199-

200

lead-acid precipitation, 199
lead-tray evaporation, 196
litharge evaporation, 195
sulfuric acid precipitation, 200-

201
zinc-lead precipitation, 197

Dore", 74

Electric assay furnaces, 252-256
Electrolytic precipitation assay, cy-
anide solutions, 199-200
Electromotive series of elements, 88
Equipment for assaying, cost sum-
mary of, 262

itemized list and prices, 264-265
Errors in assaying, corrected assays,

237-241

cupellation, 230-233
fusions, 229-230
metal balances for determining,

235-236
parting, 234
salting, 227-229
sampling, 229
weighing, 234-235
Eutectic, 118

Fine gold, preparation of, 183

JTin^paa, 4O, 173

Flour, reducing power of, 98-99
Fluxes, 3

addition of, 155-157

for assaying, 260-262

for melting cyanide precipitates,

246-247

for melting retort sponge, 245
mixing of, 157

Fluxing, for crucible assay, 124-129
requirements of ore constituents,

125

Fuels for assay furnaces, 249
Furnace, operation for crucible

assay, 157-158
Furnace tools, 259
Furnaces, assay, data on typical, 255
design of, 257-259
electric, 252-256
fuels for, 249
muffles, 256
refractories for, 257
types of, 248-249
using liquid or gaseous fuels,

250-252

using solid fuels, 249-250
Fusion errors, 230

Gold, annealing of, 82-83

determination in presence of plati-
num metals, 215-219
fine or proof, preparation of, 183
melting point of, 61
solubility in lead, speiss, matte

and slag, 96
Gold bullion, 173

assay of, 180-183

Gold chloride solutions, assay of,
202-203

Impurities, acid treatment for re-
moval of, 184-193

274

FIRE ASSAYING

Inquartation, 75
Inquarts, Herman, 75
Iridium, determination of, 215-219
Iron, reducer, for lead assay, 221-223
for oxidation-collection method,

147

for soda-iron method, 145-146
reducing power of, 101
slag color, 123

Lead, assay for, 220-223
Lead bullion, 172

assay of, 174
Lead buttons, brittle, 67, 106, 115,

164
collection of precious metals by,

94, 97
control of size of, crucible assay,

129-135

scorification, 167-168
hard, 67, 164
impurities in, 90-94
size of, 96, 97-110, 129
solubility of precious metals in,

96, 206

Lead-acid precipitation assay, cy-
anide solutions, 199
Lead-gold alloys, melting points of,

62
Lead-silver alloys, melting points of,

Lead-tray evaporation assay, cy-
anide solutions, 196
Litharge, elimination of sulfur by,

116

melting point of, 57
separation of impurities by an

excess of, 91-93

Litharge-evaporation assay, cyanide
solutions, 195

Manganese, slag color, 123
Matte, composition of, 86, 112
in controlled reduction, 115

Matte, formation and decomposition
of, 112-116

order of formation with metals,
114

solubility of precious metals in, 96

in uncontrolled reduction, 1 1 4
Mechanical sampling, 31
Metal balances, use of, to check

assays, 235-236

Metal oxides, basic character of, 85
Metallics assay, 32-36
Metals, adapted to fire assay, 1

oxidation states of, 104

sampling of, 36-37
Millieme, 173
Minerals, oxidizing, 109, 134-135

qualitative tests for, 123

reducing, 102, 132
Moisture, determination of, 29
Molds, for crucible fusions, 158-160
Muffle temperature, cupellation, 54-

55, 66
Muffles, 256

Nickel, crucible charges for ores

high in, 142-145
in cupellation, 69
tolerance of, in fire assay, 184
Niter, oxidizer for oxidation-collec-
tion method, 148
oxidizing power of, 108
Nonmetal oxides, acidic character of,
85

Ores, basic, crucible charge calcula-
tion for, 138-139

classification of, 121

determination of constituents of,
121-124

high in copper, nickel, or tel-
lurium, crucible charge calcu-
lation for, 142-145

siliceous, crucible charge calcula-
tion for, 136-137

INDEX

275

Ores, sulfide, crucible charge calcula-
tion for, 139-145

Osmium, determination of, 212-214
Oxidation, explanation of, 88-90
by niter, 108-109
by roasting, 112-114
Oxidation-collection method, 147-

149
Oxidizing agents, oxidizing power of,

., 107-109, 134
Oxidizing minerals, determination of

oxidizing power of, 135
jDxidizing power of, 109, 134-135
Oxidizing power, determination of,

135

of oxidizing agents, 107-109
of oxidizing minerals, 109, 134

Palladium, determination of, 215-

219
Parting, beads containing platinum

metals, 208-211

breaking up of gold during, 75, 80
decantation and washing, 81-82
definition of, 74
effect, of chlorides on, 78-79

of impurities on, 81
errors in, 234
flattening beads for, 76
of gold bullion, 182-183
indication of platinum-group met-
als in, 81
modified sulfuric acid method,

210-211

procedure for, 83-84
ratio of silver to gold for, 75
receptacles used in, 77-78
recovery of fine gold from wash-
ings, 82
silver recovery from waste liquor,

temperature of acid, 80
time required for, 80-81
Pennyweight, 12, 39
Phases, in assay fusions, 86
Pipe sampling, 31

Placer gravels, assay of, 34
Platinum, determination of, 215-219

melting point of, 61
Platinum-group metals, assay fu-
sions for, 205-206
in cupellation, 70-73
cupellation of buttons containing,

206-208
determination of Ag, Os, and Ru,

212-215
determination of Pt, Pd, Rh, Ir,

and Au, 215-219
in parting, 81

parting beads containing, 208-211
source of, 204

Price list, assay equipment, 264-265
assay supplies and reagents, 261-

265
Proof assay, in cupellation, of gold

bullion, 181
of silver bullion, 176
Pyrite, calculation of reducing power
of, 102-103

Reagents for assaying, 260-262
itemized list and prices, 267
Records, of assays, 11-13
^Reducing agents, determination of

reducing power of, 131
reducing power of, 130
ing power, 98, 130
of agents, determination of, 131
effect of charge composition on,

104-107

of minerals, 102-107, 132
calculation of, 102-107
determination of, 133
of reducing agents, 98-101, 130
Reduction, explanation of, 88-90
Refractories, 257
Reports, of assays, 8, 11-13
Retorting amalgam, 242-245
Rhodium, determination of, 215-219
Riffle, Jones, 31

Roast, elimination of impurities by,
149-150

276

Ft RE ASSAYING

Roast, oxidising, M 2-1 13
Roasting assay^ n^thod, 149-150
Roast-reaction process, 113
Ruthenium, determination of, 212-
215

Salting, accidental, 228-229

fraudulent, 227-228

guarding against, 2, 227-229
Samples, accuracy of, 15

error in, 14

limit of error, 15
calculation of, 22
practical, 16

mine, preparation for assay, 26-28

moisture, 29

smelter, preparation for assay,
28-32

standard deviation of, 16

types of, 14

weight of, for bullion assays, 174
calculation of minimum, 21, 33
for crucible assay, 40, 124
effect of particle size on, 23-25
for scorification, 168
Sampling, broken rock, 28-32

bullion, 36-37

coning and quartering, 30

equipment for, 260

errors, 229

material containing metallics, 32-
36

mechanical, 31

metals, 36-37

ores and brittle materials, 25-32

pipe, 31

shovel, 29

solutions, 37-38

Scorification, separation of impuri-
ties by, 94

Scorification assay, application and
limitations, 165-167

control of button size in, 167168

procedure, 169

slag characteristics, 167

trouble shooting in, 170-171

orification assay, weight of ore
sample, 168
ifiers, 168
Selenium, tolerance of, in fire assay,

184

Sensitivity of balances, 42
Shotting, of crucible fusions, 163
Shovel sampling, 29
Silver, determination in presence of

platinum metals, 212-215
melting point of, 61
solubility in lead, speiss, matte,

and slag, 96
Silver bullion, 173

cupellation assay of, 176-178
lead required for cupellation assay,

177

volumetric assay of, 178-180
Silicate degree, 87

effect on reducing power, 106
Slag factors, 126
Slags, assays of, 239

from bullion melting, disposal of,

246

colors caused by constituents, 123
composition of, 86, 116, 117-120
indicated oxidation state of metals

in, 104

silicate degree classification, 87
smelter, 116-117

solubility of precious metals in, 96
washing of, 164, 230
Soda-iron assay method, matte in,

114

procedure, 146
use of, 145
Sodium carbonate, elimination of

sulfur by, 116
Sodium chloride, cover for crucible

charges, 120
Sodium sulfate, formation in assay,

100-101, 120
Solutions, assay methods for, 195-

203

assay portions of, 194-195
sampling of, 37-38, 194
Speiss, composition of, 86, 111
solubility of precious metals in, 96

INDEX

277

Splitting limits, 236-237

Sprouting, 63

Standard deviation, 16
calculation of, 20

Stibnite, reducing power tinder
various conditions, 106

Sulfide elimination, conditions fav-
oring, 115-116

Sulfide minerals, reducing power of,
132

Sulfur, in crucible assay of copper

bullion, use of, 175
reducing power of, 100
roasting for the separation of, 150

Sulfuric acid precipitation assay,
cyanide solutions, 200-201

Supplies for assaying, cost summary

of, 262
itemized list and prices, 266-267

Supply houses, list of, 268-269

Surcharge, 180

Surfusion, 61

Tellurium, behavior in assaying, 91
in cupellation, 69
roasting for the separation of, 150
tolerance of, in fire assay, 184

Temperature-color scale, 56

Thiosulfate solutions, assay of, 202

Tin assay, 223-224

Umpire assay, 237

Uncontrolled reduction assays,

matte in, 114
Units used in assaying, 39-40

Vanning test, 122

Volhard's method, silver determina-
tion by, 179-180

Weighing, assay balance, 43

assay portion, of bullion, 174

of ores, 40
Weighing, beads, 41-45

deflection method, 45

equal-swings method, 44

errors, 235

pulp, 40
Weights, assay ton, 12, 39

Zinc, metallic, assay of materials

containing, 186
effect of, in scorification, 166

Zinc-lead precipitation assay, cy-
anide solutions, 197

SLAG FACTOBS, EQUIVALENT WEIGHTS PER UNIT or SILICA

Dominant
element

Compound

Original form

Combining
form

Wt. orig, form per
unit wt. SiOt

Silicate degree

Sub

Mono

Sesqui

Aoid Fluxes

Boron
Boron

Borax glass
Borax

Na 8 O.2BiO,
Na 2 O.2B 2 O 3 .10HaO

Na 2 0.2B,Oi*
Na 2 O.2B 8 O3*

1.2
2.2

1.3
2.4

1.5
2.8

1.7
3.1

Basic Fluxes

Lead

Litharge

PbO

14.9

7.4

4.9

3.7

Sodium

Sodium carbonate

NajCO

NaaO

7.1

3.5

2.4

1.8

Potaa-

Potassium carbo-

KaCO.

0.2

4.6

3.1

2.3

eium

nate

Basic Ore Constituents

Antimony

Antimony triox-

SbiOt

SbzOj

6.5

3.2

2.2

1.6

ide

Stibnite

Sb 2 S,

SbaOa

7.5

3.8

2.5

1.9

Calcium

Calcium oxide

CaO

3.7

1.9

1.2

0.9

(lime)

Calcite or lime-

CaCOa

CaO

6.7

3.3

2.2

1.7

stone

Copper

Cuprous oxide

CujO

CusO

9.5

4.8

3.2

2.4

Cupric oxide

CuO

CujO

10.6

5.3

3.5

2.7

Chalcocite

CuiS

CujO

10.6

5.3

3.5

2.7

Chalcopyrite

CuFeSi

CujO.2FeO

6.1

3.1

2.0

1.5

Iron

Ferrous oxide

FeO

4.8

2.4

1.6

1.2

Hematite

FesOa

FeO

5.3

2.7

1.8

Magnetite

FeiO*

FeO

4.9

2.5

1.6

1.2

Pyrite (marca-

FeSi

FeO

8.0

4:0

2.7

2.0

site)

Lead

Lead oxide (lith-

PbO

14.9

7.4

4.9

3.7

arge)

Galena

PbS

PbO

15.9

8.0

5.3

4.0

Magne-

Magnesium oxide

MgO

2.7

1.3

0.9

0.7

sium

(magnesia)

Magnesite

MgCOa

MgO

5.6

2.8

1.9

1.4

Manga-

Manganous oxide

MnO

4.7

2.4

1.6

1.2

nese

Pyrolusite

MnOa

MnO

5.8

2.9

1.9

1.4

Zinc

Zinc oxide

ZnO

5.4

2.7

1 8

1.3

Sphalerite

ZnS

ZnO

6.5

3.2

2.2

1.6

Note: The fluxing of aluminum and arsenic is explained in Table XI.
* The slag factors for borax and borax glass compensate the silicate degree for the base
NajO in the borax or borax glass.

REDUCING POWER OF CERTAIN PULFIDE MINERALS

Mineral

Formula

S,per
cent

Normal slag-
forming
oxides

R.P., grams per gram
of mineral

Com-
puted
R.P.

Actual R.P.

S tp SO,

Pre-
liminary
assays*

Niter
charges f

Galena . ...

PbS
Cu 2 S
FeAsS
Sb 2 S 3
CuFeSa
ZnS
FeS 2

13.4
20.2
19.7
28.6
34.9
32.9
53.4

PbO
Cu 2 O
FeO, As 2 O 6
Sb 2 O 8
Cu 2 O, FeO
ZnO
FeO

3.46
5.20
8.25
7.35
8.44
8.51
12.07

2.9
4.5
8.1
5.9
8.2
8.2
11.6

2.9
4.7
7.4
5.8
8.2
8.1
11.0

Chalcocite

Arsenopyrite . . .
Stibnite

Chalcopyrite . . .
Sphalerite

Pyrite ....

* The charge in all cases was 3 g. sulfide, 45 g. PbO, 15 g. NasCOg, 5 g. SiOj.

t The values in this column were determined with monosilicate slags under normal
assaying conditions with niter and excess litharge. U.S.P. niter was used which had an
O.P. of 4.0.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Fire Assaying"

See other formats