Full text of "The assay of tin and antimony"

See other formats

The Assay of Tin
and Antimony.

L. PARRY, A.R.S.M

Assayer and Consulting Metallurgist.

PRICE 5/- NET.

SECOND EDITION.

LONDON :

MINING JOURNAL" OFFICE,

15 GEORGE STREET, MANSION HOUSE, E.G.
[ALL RIGHTS RESERVED.-]

~r/V 3

%«i*v. ;

y*>

Introduction.

regard to the high price of the metal, it is a most
FJJ remarkable fact that the methods usually described in text
® books, taught in school and classes, and in common use, for
the separation and analytical determination of tin, are as a rule
both unpractical and unreliable.

In commercial work it is, in general, of far greater importance
to employ methods which can be relied upon to yield results accurate
to within a quarter or a half per cent, in every case, and in a few
hours at most, or in which the possibilities of error are definitely
limited to a known minimum, rather than to spend three or four
days in finicking with methods which are, properly speaking, only
suitable for atomic weight determinations or other academic research,
in order to obtain results which may possibly be exact to two or three
decimal places and possibly inexact to the extent of 10 per cent.
Gravimetric methods possess the single advantage over volumetric
that one can weigh with greater precision than one can measure ;
the difference is absolutely immaterial in most cases, as far as
buying and selling are concerned, and in every other respect volumetric
methods are far more reliable and are much more rapid than gravi-
metric, involving fewer and less complicated separations ; speaking
generally, and having regard to differences in the kind of error in-
volved in gravimetric as against volumetric assays, it is not going too
far to say that in commercial work gravimetric assays should be
avoided as far as possible— they leave too much to chance unless used
with great discrimination. In an atomic weight determination one
starts with pure chemicals, in assaying there are too many unknowns,
involving too many assumptions, for rapid and accurate gravimetric
work. It may as well be pointed out here that very few combining
weights are known with certainty beyond the first decimal place, or
nearer than i part in 1,000. In the majority of analytical methods
greater accuracy than i part in 1,000 is a physical and chemical
impossibility ; in very many the limit may be put at i part in 500 —
indeed, a method of analysis which is properly systematised and
capable of this degree of accuracy is a most excellent method, and,
what is more, a rara avis. For instance, a really good method may
enable us to report 50-1 per cent., or even 50-15 in very rare cases,
in preference to 50-0 per cent, or 50-2 per cent. ; but such a result
as 50-12 per cent, is a scientific absurdity in the second decimal place.

A 2

I{ Hrjyon^doubtS tMs4et him study care fully Ostwald's "Foundations
k of Analytical Chemistry."

As regards tin, the commonest and most glaring source of error
is undoubtedly the indiscriminate and unreasoning abuse of the estima-
tion by weighing as Stannic Oxide (SnO2). It is, or should be,
perfectly well known to chemists that the usual methods employed
in order to obtain tin as SnO2 involve at the same time the con-
tamination of the stannic oxide with SbO2, As2O5, P2O5, Fe2O3,
PbO, and even with biO2 and WO3, and occasionally CuO, unless
special methods are adopted which ensure the quantitative separation
of these elements. The separation of tin for volumetric estimation
is easy and certain ; on the other hand, its separation for gravimetric
estimation as SnO2 is complicated, tedious, and unreliable. The
result is that the presence of impurities in stannic oxide is very
commonly ignored unless they are in very obvious quantities. Thus,
it is usual to assume that solder is composed of tin and lead, whilst
as a matter of fact nearly all the tinman's and plumber's solder which
is made to-day contains from i£ to 3 per cent, antimony; in con-
sequence, determinations of tin in solder are very generally too high,
either because the antimony is simply allowed to score fully as tin,
or because an imperfect separation of tin and antimony is employed ;
moreover, such determinations are usually too high on account of the
presence of PbO in the SnO2, to the extent of about i per cent.,
as it is a common error to assume that treatment of solder with
HNO3 is a quantitative separation of tin from lead, or of tin and
antimony from lead. When worked without a due appreciation of
its limitations, there is probably not a more unreliable analytical
method in common use than the gravimetric estimation of tin as
SnO2 ; yet in certain special cases it is undoubtedly useful. No wet
method of assaying tin can be considered justifiable which does not
exclude the interference of Sb, As, Fe, Pb, Cu, W, and bi, to mention
the most usual sources of error, and weighing as SnO2 does not permit
of this with certainty. It is probably in consequence of the general
use of this method that errors of 5, 10, and even 20 per cent, in tin
and antimony assays (reported to two decimal places) are of common
every-day occurrence.

Tin ores are very frequently sold on a dry assay basis. Any
dry assay of tin is, however, rather misleading as a valuation basis
unless it is combined with a wet assay for Sw, of the button of
metal obtained in the crucible. Mere oxidation with HNO3 and
weighing the residue as SnO2 will not do — common sense should
teach that ; it is necessary to use a method which will show the
percentage of Su and exclude the possibility of any possible impurity
scoring as Sn. If this is done, the result obtained (as in any
crucible assay) must of necessity be lower than the actual tin contents
of the ore. If it is not done— if the metal button is simply weighed
as tin, or if the percentage of Sn is determined negatively by assaying
the button for one or two impurities, or positively by oxidation with
HNO3 and weighing the residue as SnO2 — the result will in many cases

be too high. It is a common practice to report the percentage of
tin in tin ore as so many per cent, "fine tin." The use of this
term appears to be sanctioned by buyers and sellers alike, though
it is scientifically an indefensible expression with a far too indefinite
meaning. One man's idea of "fine tin" may include tin of 5 per
cent, of impurities, while another man may think of " fine tin " as
something containing ggf per cent, or over of the definite entity 5w,
and altogether it is scarcely advisable to expect concordance of results
between two assayers, one of whom reports the percentage of " fine
tin," and the other the percentage of Sn. As regards tin ores, the only
method of assay which would, in the opinion of the writer, receive
the mutual assent of buyers and sellers, if they could be brought
to see the technical points involved, would be reduction to metal in
hydrogen combined with a volumetric or electrolytic determination
of the tin. There can be no certainty about any method not in-
volving complete solution of the ore.

The originator of the Ferric Chloride titration for estimating tin
is unknown to the writer, who became acquainted with the process
many years ago ; at that time its application was rather limited, and
the greater number of assay methods involving this titration, and
described in Chapter V., as well as many important points about
the titration itself, were elaborated in detail by the writer in con-
junction with Mr. H. Hocking.

Chapter I.

Miscellaneous Facts bearing on the Assay of
Tin and Antimony.

Sn = 1 19. Sb = 120. (O = 1 6.)

(i.) OXIDES. — Oxides of tin formed in the wet way are soluble
in HC1. Oxides of antimony are much less easily soluble. The
oxides of both metals are soluble in alkalies (soda and potash, not
ammonia), and in alkaline sulphides.

SnO2 obtained in the wet way by the action of HNO3 on metals
or sulphides is liable to hold Sb, As, P, as oxides, also PbO, Fe3O3,
CuO, WO3, SiO2, and Bi2O3. Oxides of tin and antimony are almost
insoluble in acids after ignition. Oxide of tin exhibits many modi-
fications according to the method of preparation. Thus the rapid
action of dilute HNO3 on tin (with the aid of heat) yields a product
different from that obtained by the slow action of dilute HNO3 in
the cold, which yields an unstable compound containing nitric acid.
The action of HNO3 on SnS2, the addition of KOH to SnCl^, the
evaporation of SnCl2, with excess of HNO3, the cautious addition of
HC1 to solutions of alkaline stannates, and the addition of NH4NO3
or Na2SO^ to SnCl4, all yield some form-of stannic oxide. Insoluble
stannic oxide results from the ignition of any of these products or
by burning stannic sulphide or tin, in air or oxygen. The hydroxide
precipitates obtained by adding AmOH or NaOH to solutions of
SnCl2 or SnCl4 are soluble in excess of NaOH, and partly soluble
in acetic acid, but are insoluble in AmOH. Soluble in tartaric acid.
Stannic acid is slightly soluble in HNO3, and therefore to ensure the
complete separation of SnO2 by HNO3 it is sometimes necessary to
evaporate to complete dryness, and is always safer. The dilute nitric
acid extract from this will be quite free from tin.

Two acids derived from stannic oxide are generally recognised ;
ordinary stannic acid, H2SnO3, which can exchange all its hydrogen
for metals, is formed by neutralising solutions of alkaline stannates
with acids, and also by the addition of CaCO3 to SnCl4 ; metastannic
acid, which results by oxidising tin with nitric acid, is H10Sn5O15,
5 H2O at ordinary temperatures and H10Sn5O15 when dried at ioo°C.,
and only one-hfth of its hydrogen is replaceable by metals. Both
these substances are soluble in H2SO4 or in HC1. The H2SO4
solution contains stannic sulphate, and on dilution a hydroxide pre-
cipitate forms ; on boiling the diluted solution, all the tin is said
to be thrown down as metastannic acid. The HC1 solution is not
precipitated by dilution. Ignited and native stannic oxides are only

partly acted on by strong H2SO4 or boiling alkalies, and cassiterite
is not completely opened up even by fusion with potassium bisulphate.
When oxides of tin and antimony are dissolved in fixed alkaline solu-
tions, alkaline salts of the particular oxides are formed. When the
oxides are dissolved in alkaline sulphides, mixtures of sulpho and
oxy- salts are formed.

(2.) SULPHIDES. — Sulphides of tin, arsenic, antimony, are all
soluble in alkalies and alkaline sulphides, from which solutions they
are all precipitated by cautious addition of HC1.

Sulphide of arsenic is insoluble even in strong HC1 and is com-
pletely precipitated by H2S on standing, from a strong HC1 solution
of either As2O3 or As2O5. Sulphides of antimony are precipitated
from fairly concentrated HC1 solutions, but on boiling with strong
HC1 both dissolve with formation of SbCl3.

SnS is an inconvenient form in which to precipitate tin, and
stannic sulphide SnS2 is precipitated completely only from fairly
dilute solutions. These sulphides readily assume a colloidal form
and pass into solution when washed with pure water ; hence, when
washing is necessary,_they should always be washed with some salt
solution (NaCl or NaAc) — see Ostwald, " Foundations of Analytical
Chemistry." Sulphide of tin (SnS2) is not precipitated from solutions
containing excess of oxalic acid whilst sulphides of As and Sb are
precipitated under these circumstances. Sulphides of tin are readily
soluble in HC1 and are oxidised to SnO2 by treatment with HNO3.

(3.) OXYCHLORIDES.— Solutions of BiCl3 and SbCl3 in HC1
are precipitated by dilution with water, the oxychlorides being formed ;
they redissolve on the addition of more HC1. A solution of SnCl4
in HC1 is not precipitated by dilution, but solutions of SnCl2 readily
absorb oxygen from the air, and in a neutral or only faintly acid
solution a deposit of tin oxychloride forms ; if the solution of SnCl2
is freely acid with HC1 no precipitate is formed, but the solution
becomes converted into SnCl4. A solution of SbCl5 in HC1 is pre-
cipitated by dilution with water, some form of antimonic acid (said
to be probably orthoantimonic acid H3SbO4) being thrown down ;
the precipitate is only redissolved by HC1 with difficulty, especially
after standing some time. Oxychloride of tin which has been long
precipitated is also difficult to dissolve in HC1. The presence of
tartaric acid or of alkaline tartrates prevents the precipitation of
oxychlorides of tin and antimony. A solution of SbCl3 in HC1 gives
Sb2O3 with KOH, soluble in large excess, easily soluble in presence
of tartaric acid, forming potassium antimony 1 tartrate, which is tartar
emetic (8HOH;cCo°6(Sbo))- The use of organic acids in inorganic analyses
is much too frequent, and is simply a fad with many chemists. Com-
plications which are not always properly understood are very often
introduced, and it is a safe rule to avoid the use of such compounds
as much as possible — at any rate in commercial work.

(4.) REDUCING ACTION OF STANNOUS OXIDE.— An
alkaline solution of SnO (in potash) reduces a solution of cupro-
potassic tartrate with precipitation of Cu2O. An alkaline solution
of As2O3 acts in the same way, but no other metallic oxide. A
solution of SnO in KOH also reduces Bi(NO3)3 with deposition of
suboxide, according to Tilden. Further, such a solution appears to
reduce itself, to use a rather loose expression, depositing tin and
giving a solution of stannate.

SnO in either alkaline (bicarbonate) or acid (HC1) solution reduces
a solution of iodine, forming a stannate or stannic chloride and HI.

SnCl2 in HCi solution reduces AsCl3 with precipitation of a
brown deposit of arsenic containing 4 to 5 per cent, of Sn ; reduces
solutions of AuCl3 and PtCl4 to metals, solution of HgCl2 to first
Hg2Cl2 and then Hg ; (in presence of KI, SnCl2 does not reduce
HgClJ ; reduces CrO3 and Mn^O7 to Cr2O3 and MnO ; reduces SbCl5,
CuClo, and FeCl3 to SbCl3, Cu2Cl2, and FeCl2 respectively. Also
reduces BiCl3 to metal.

(5.) ACTION OF KMnO± AND K0Cr907.—In acid solution
oxidise Sb2O3, As2O3, SnO, FeO, Cu2O.

(6.) ACTION OF A SOLUTION OF FeCls IN HCI.—
'Liberates iodine from a solution of KI, converts Cu2Cl2 into CuCl2,
converts SnCl2 into SnCl4 and dissolves ppted Sb, As, and Cu, with
formation of SbCl3, AsCl3, CuCl2. Does not oxidise HCI solutions of
Sb2O3 or As2O3, but on the other hand FeCl2 reduces HCI solutions
of Sb2O5 and As2O5 under some conditions. FeCl2 does not reduce
HgCl2. SbCl5 — FeCl3 CuCl2 are in the order of reducibility by
SnCl2, which in a hot HCI solution containing all three of the above
chlorides, reduces first the SbCl5, then the FeCl3, and lastly the CuCl2.
FeCl3 does not under similar conditions oxidise HCI solutions of
SbCl3 or As2O3, but oxidises solutions of Cu2Cl2 or SnCl2. CuCl2
in such solutions does not oxidise SbCl3 or FeCl2, but oxidises SnCl2.
Cu2Cl2 reduces SbCl- before FeCl3, and has very little, if any, action
on HCI solutions of As2O5. In alkaline solution As2O3 reduces CuO
as does SnO also. In acid solution As2O3 does not reduce CuO. A
solution of SbCl3 in HCI seems to be permanent, a solution of FeCl2
or As2O3 gradually oxidises in the air, and solutions of Cu2Cl2 or SnCl2
rapidly oxidise — we should expect from the above that SnCl2 would
be less stable than Cu2Cl2, but it is not ; we should also expect that
FeCl2 would reduce SbCl5 and As2O5, and to a certain extent and
under some conditions this appears to be the case. KI reduces SbCl5,
FeCl3, CuCl2, and As2O5 in HCI solution, but not SnCl4, whilst iodine
oxidises Sb263, As2O3, SnO, in alkaline solution. The behaviour of
arsenic in HCI solutions is very peculiar, depending much on the
amount and concentration of the HCI present. FeCl3 does not oxidise
As2O3 in HCI solution, and excessof FeCl2 reduces As2O5 in a strong
(^ saturated) HCI solution (Fischer-Hufschmidt distillation process), so
that the place, of arsenic would appear to be before iron in order of

reducibility. SbCl5 is certainly more rapidly and completely reduced
by KI in HC1 solution than is As2O5, the reaction in the case of
As2O5 only moving to completion as fast as the liberated-kxiine is
removed, and being reversible unless the iodine is removed ;
further, HC1 solutions of As2O3 (in which, by the way, the arsenic
is supposed to exist as As2O3 unless the liquid is saturated with
HC1, when it exists as AsCl3) show a strong oxidation tendency, so
that we should arrange the order of reducibility thus : SbCl5 — As2O5
(in HC1) — FeCl3 — CuCl2 — SnCl4, and the order of permanency or
stability of the lower chlorides would, of course, be SbCl3 — AsCl3 —
FeCl2— Cu2Clo— SnCl2. But, on the other hand, if a little FeCl3 or
CuCl2 be added to a solution of arsenic acid in fairly strong HC1, and
the solution titrated with SnCl2, the FeCl3 or CuCl2are at once reduced,
and would therefore appear to be under these conditions more readily
reduced than As2O5 ; it also follows that FeCl2 does not reduce As2O5
under these conditions. Again, the reducing effect of SnCl2 on HC1
solutions of arsenic acid is very doubtful, and, further, As2O3 reduces
even CuO in alkaline solution. From this we should arrange the
order of reducibility thus: SbCl5 — FeCl3 — CuCL— As2O5 — SnCl4.
SO2 reduces Sb, As, Fe, from -ic to -ous in HC1 solutions, but neither
Cu (unless in presence of KCNS) nor Sn. As2O5 may be also reduced
by PC13. Also, it would appear from the foregoing that if we have
SbCl5, As2O5, CuCl2 in hot strong HC1 solution and titrate with SnCL,
the SbCl5 is first reduced, then any trace of FeCl3, then CuCl2, at
which point the solution becomes colourless ; this suggests an obvious
and very easy estimation of Sb in alloys containing Cu and As, when
the Cu is known, the As making no difference. The strength of solu-
tions of SnCl2 may be checked with FeCl3, CuCl2, K2Cr2O7, KMnO4
or with iodine. Some of the foregoing reactions are applied in the
"Weil" assays of Fahl Ores, described in " Sutton's Volumetric
Analysis." These oxidising and reducing actions are extremely com-
plicated, and for the elucidation of a satisfactory theory of them much
investigation is needed, but the subject is well worthy of it.

t (7.) METALLIC PRECIPITATION.— From HC1 solutions of
— ic chlorides.

Copper precipitates Hg, Ag, As, Sb, not Sn.

Iron „ Hg, Bi, Sb, Cu, and (in presence of SnCl4) As,

and reduces FeCl, and SnCL to Fed., and

SnCL.

Tin „ As, Sb, Cu, Hg.

Lead ,, Cu, Hg, Sb, Bi.

Aluminium ,, Cu, Sb, and most of the common metals.
Zinc „ Sn, Sb, Cu, Pb, As, Cd, Hg.

Magnesium ,, Fe, Zn, Co, and Ni, and most of the common

metals.

Zinc evolves a considerable proportion of the As and Sb as
hydrides, as does iron, though to a less extent, very little SbH3 being

formed with iron. Even magnesium does not appear to completely
evolve As and Sb as hydrides, though Crookes states that it does.

(8.) SOLUTION OF METALS.— Arsenic, antimony, copper,
bismuth, mercury, are insoluble in HC1 either strong or dilute, hot
or cold, when in a coherent form ; from alloys a certain amount of
As is evolved as AsH3, and certain tin copper alloys are completely
soluble in HC1 giving a solution of SnCL, and Cu2Cl2 ; in presence
of oxidising agents such as free Cl the above metals are all soluble.
Finely divided Sb and Cu readily oxidise in contact with air and
moisture, and the oxides are, of course, then readily dissolved by any
HC1 present, thus often giving rise to the appearance of solution of
the metals themselves in HC1. The above metals are all soluble in
aqua regia.

Tin, zinc, cobalt, nickel, iron, cadmium, aluminium, are readily
soluble in strong HC1, and lead is also completely soluble, though
slowly. When antimony is dissolved in HC1 and KC1O3 it gives
SbCU ; when it is dissolved in HC1 and iodine, SbCl3 is formed.

Zinc, iron, cobalt, nickel, cadmium, readily dissolve in dilute
H2SO4. Tin, copper, mercury, and finely divided arsenic and an-
timony are soluble in hot strong H2SO4, forming in the case of arsenic
and antimony solutions of the lower oxides. Sulphides of As, Sb, Sn,
are also soluble, forming in the case of Sb and As solutions of Sb2Oa
and As2O3. In the case of Sn a solution of stannic sulphate is pro-
duced unless the metal is in excess.

Copper, lead, zinc, iron, cobalt, nickel, bismuth, mercury, cadmium,
are dissolved by HNO3 with formation of nitrates; tin and antimony
(and in presence of tin) phosphorus and arsenic remain as insoluble
oxides, though a little tin may go into solution if the liquid is not
evaporated to dryness, and the residue is also liable to hold varying
amounts of other metallic oxides, especially those of lead and iron.
In absence of tin, arsenic and phosphorus go into solution as arsenic
and phosphoric acids.

Aluminium is not easily soluble in H2SO4 or HNO3.

Cg.) PRESENCE OF PHOSPHORUS AND ARSENIC IN
TIN ASSAYS.— When SnTv and Asv or Pv in solution in HC1 are
treated with iron, the precipitate holds tin. In the case of As, the
ppted metal holds 3 or 4 per cent. Sn. When Sn and As, or Sn and
P, in any soluble form, or as metal or sulphide, are evaporated with
HNO3 insoluble 2 SnO2, As2O5, and 2 SnO2P2O5 are formed, soluble
in strong HC1.

(10.) Arsenic acid is not reduced by HC1, as is sometimes asserted,
at any rate not under ordinary circumstances, and HC1 solutions
of arsenic acid may be boiled freely without any loss of arsenic by
volatilisation.

Ferric chloride does not convert precipitated Sb into SbCl.-, but
into SbCl3.

The low boiling point of SnCl4 (i 14° C.) is no bar to the boiling or
evaporation (unless carried very low) of HC1 solutions of SnO2 — as
experiment has shown that any such loss is entirely negligible under
all ordinary circumstances, and the possibility of such loss can always
be checked. Crookes states that when a Sn — Sb — As alloy is boiled
with HC1 the residue contains only Sb, that all the As is evolved as
AsH3, but the statement requires considerable qualification.

(n.) PURITY OF CHEMICALS used in the ensuing tin and
antimony assays : — Sheet zinc must be absolutely free from tin, and
must not contain more than small traces of arsenic or antimony.

Hydrochloric acid should be as concentrated as possible and
free from As, HNO3, or Cl.

Ferric chloride should be free from Cl, HNO3, As, or FeCl3.

Soda ash and caustic soda used for fusions should be free from
chlorine.

Water used for diluting should be boiled and free from oxygen.
Iron used should be either best piano wire or soft iron rod.

(12.) METHODS OF SEPARATING TIN— ANTIMONY-
ARSENIC from one another.

Separation of Tin.

i. — In a moderately strong warm HC1 solution, containing excess
of oxalic acid (20 grms. oxalic acid for each grm. of tin)
H2S pptes. As and Sb, — Sn remains in solution. This
method, which was devised by Mr. F. W. Clarke, is tedious
and unsuitable for commercial work, though accurate. For
Carnot's modification, using thio-sulphate as a precipitant,
see " Crookes' Select Methods."

2. — Iron wire in an HC1 solution of As, Sb, Sn, precipitates As
and Sb as metals, though a certain amount of SbH3 and
AsH3 are formed. The arsenic precipitated contains 3 or
4 per cent. Sn. Except where more than 10 per cent, of
As is present, this is the best practical method of separating
tin from arsenic and antimony.

3. — Strong boiling HC1 dissolves tin and lead, but not arsenic or
antimony. This is a good rough separation, provided not
more than 5 per cent. As and Sb are present.

4. — Electrolytic separation (see " Classen ").

5. — Fractional separation with H2S. As is ppted in a concen-
trated HC1 solution and Sb on slight dilution. Sn not
ppted in a strong solution. (Doviton's Method ?)

6. Winkler's separation with CaCO3 and KCN (see Menschutkn).

Method unpractical and not suitable for commercial work.
7- — Fusion of oxides of As, Sb, Sn, with caustic soda. Stannate

and arsenate soluble in water, antimonate insoluble. This
is a thoroughly unpractical and unreliable method.

In practice, methods 2 and 3 are the most useful, methods i and
4 rarely.

Separation of Antimony and Arsenic.

N.B. — Antimony can be readily estimated in presence of tin.

i. — When an alloy or an arsenious compound is distilled with
a solution of FeCl3 and CaCl2 in HC1, the arsenic is all
evolved as AsCl3 at a temperature of i25°C. By adding
a solution of ZnCl2 the antimony can then be all evolved
as SbCl3 at a temperature of about i9o°C. (Gibbs' method).
In "distilling off arsenic there is no loss of SnCl4.

(a) When a solution of arsenic acid saturated with HC1

is distilled with FeCl2 the arsenic is quantitatively
evolved as AsCl3 (Fischer-Hufschmidt process). By
distilling with a mixture of FeCl3 and FeCl2 in con-
centrated HC1 and CaCl2 solution the whole of the
arsenic may be obtained as AsCl3, not only from alloys
but from any arsenic compound. These distillation
separations are chiefly used in the assay of arsenic.

(b) Arsenic may also be separated from HC1 solutions with

metallic copper, and the copper-arsenic then distilled
with ferric-chloride mixture.

•2. — HC1 dissolves sulphides of antimony and tin on boiling, whilst
sulphide of arsenic is left undissolved. An imperfect method,
though very convenient sometimes.

3. — As2S3 is completely precipitable by H2S in a strong HC1
solution, leaving Sb and Sn in solution. This is a good
and useful method in special cases. (Rose's method.)

4. — By addition of magnesia mixture to an ammoniacal tartarate
solution of Sn, Sb, As (ic), the As is precipitated as Mg,
NH4, AsO4. It may also be precipitated from an alkaline
sulphide solution, with MgO mixture.

5. — Sulphide of arsenic is soluble in Am2CO3 solution. Sul-
phides of tin and antimony are insoluble. A rough practical
method, though imperfect.

6. — The gallic acid separation of antimony is unsuitable for com-
mercial work except in special cases, but is a good method.

7. — Electrolytic separation (see "Classen ").

8. — Strong H2SO4 dissolves As, Sb, Sn — the Sb compound
separates on cooling.

9. — KHSO3 dissolves freshly precipitated As2S3 or As2S5. Sul-
phides of Sn and Sb are not dissolved. Method is un-
practical.

I0. — AsH3 and SbH3 when passed into AgNO3 act differently,
AsH3 forms H3AsO3, and Ag is precipitated. SbH3 gives
a deposit of SbAg3. (Houzeau's method.1) This reaction
is of doubtful value except as a qualitative test, as it is
seldom possible to depend upon the quantitative evolution
of arsenic and antimony as hydrides.

In practice, the first four are the most useful methods.

We might add here that most of the text-book methods of
separating arsenic are quite useless in presence of tin, as they
usually involve solution of the arsenic in HNO3, which is impossible
from a tin arsenic mixture.

Chapter II.

The Assay of Tin.

SUMMARY OF SEPARATIONS.

When necessary, tin is best separated from accompanying elements
as follows : —

i. — From Chlorine —

(a) Evaporation with excess of HNO3.

(b) Boiling with Na3CO8.

2. — From Sulphur —

(a) Evaporation with HNO3 (tin ore).

(b) Solution in aqua regia.

3. — From Tungsten —

(a] Reduction of HC1 solution with iron.

(b) Solution of unigmted WO3 in Am^COg or AmOH» •

4. — From Antimony, Copper, Bismuth—

(a) Boiling the metals with HC1 (rough).

(b) Reduction of HC1 solution with iron.

5. — From Cobalt, Nickel, Iron, Phosphorus —

By pption with H2S in dilute HC1 solution.

6. — From Arsenic —

(a) Iron wire in HC1 solution.

(b) HoS in HC1 — oxalic solution.

(d) MgO in tartaric solution.

(*) Distillation of metals with FeCL.

7. — From Lead, Zinc —

No need to separate for volumetric assay, but Sn and Pb
are best separated by alkaline sulphide pption, and
Sn and Zn by HNO8 in cases where it is necessary to
estimate Pb or Zn.

8. — From Silica — -

By evaporation with HF.

9.— From PbO, Fe3O8, CuO—

Stannic oxide may be roughly separated by evaporation to
dryness with nitric acid and extraction with dilute nitric.
All the SnO2 is in the residue, but is very impure.

SUMMARY OF CHIEF METHODS OF DETERMINATION.

i. Dry Assays —

(a) Ferrocyanide assay.

(b) Cyanide assay.

3. — Gravimetric assay by electrolysis.

4. — Volumetric assay by titration of SnCl2 with ferric chloride.

5. — Volumetric assay by titration of SnCL with iodine in acid
solution.

6 and 7. — Volumetric assay by titration of SnCl2 with KMnO4
or K2Cr2O7, with or without addition of FeCl3.

8. — Titration with iodine in alkaline solution.

9. — Solution of metallic tin in FeCl3 and titration with KMnO4
or K2Cr2O7.

10. — Titration of stannic chloride with ferrocyanide.

The dry assay should, of course, in every case be combined
with a wet assay of the button of metal obtained, otherwise it is not
an assay at all, but a mere guess of a somewhat greater degree of
approximation than a vanner's test for "black tin." la is a safe
buyer's assay for tin ashes and tin slags, is quick, convenient, and
fairly reliable within certain limits. In some cases Ib is more
accurate for tin ashes — e.g., in the case of irony tin dross. Ib and
Ic are fairly good assays for tin ore if the button is assayed for tin
by a volumetric method. The dry assay of tin ore is rapid, and the
results are fairly consistent, but of course always rather lower than
the actual percentages of tin present. If the buttons are not assayed
for tin, the results are quite unreliable, and generally too high.

Method No. 2, though occasionally useful, is a thoroughly un-
reliable method for general use, for reasons given in the introduction.
Method No. 3 is accurate, but tedious, and more suitable for a clean,
quiet research laboratory than for ordinary use in commercial work.
Method No. 4 is at once the most accurate and practical method of
estimating tin, involving in general only one filtration, and that from
a slight precipitate, or in some cases two filtrations, is rapid and for
large numbers of assays is cheap and convenient. Its accuracy is
inherent, and is probably almost absolute when the assay is worked
under proper conditions, the titration figures being strictly propor-
tional to the amounts of Sn present.

Method No. 5 is a very good one, but it is unreliable to the
extent that an assumption not justifiable a priori has to be made,
and it is one that is not readily capable of direct verification — viz.,
that all the tin tetra:ch]oride in every assay is reduced to stannous
chloride by 20 to 30 minutes' reduction with iron in warm dilute
HC1 solution. It is a most elegant method requiring but little
apparatus and few chemicals and no filtrations, is rapid and in practice
it is found to yield results which as a rule agree closely with those
got by ferric chloride assay, though it does not possess the inherent
accuracy of that method.

Methods 6 and 7 are workably accurate but not as convenient as
the two previous methods, though titration with KMnO4 is useful
for checking the working strength of standard solutions of stannous
chloride.

Methods 8, 9, and 10 are not to be recommended in practice.

GENERAL SUMMARY OF WET ASSAY OF TIN.

Tin occurs as (a) metal, (b) sulphide or arsenide, (c) soluble
oxide or soluble salts, (d) silicate, (e) insoluble oxide.

SOLUTION can always be effected as follows : —

(a) Removal of combined silica by evaporation with hydro-

fluoric acid (seldom necessary except with slags).

(b) Treatment of residue by boiling with HC1 and HNO3.

All the tin not present as insoluble oxide is then ob-
tained in solution as stannic chloride.

a stream of hydrogen or coal gas, at a low red heat.

(d) Solution of reduced metal in HC1 and HNO3.

(e) Fusion of any slight siliceous residue with KNaCOs

or soda ash, and borax, in a platinum crucible, and
solution of the melt in HC1.

The method of effecting solution of insoluble tin oxide by fusion
with caustic alkalies or caustics, alkalies and sulphur in porcelain
vessels is a dirty, tedious, and unreliable method.

SEPARATION.

1. — As, Sb, Cu, VV are separated by heating the acid solution

with iron, when As, Sb, Cu are precipitated in the metallic
form and tungsten as blue oxide.

2. — When it is desired to separate tin from Fe, P, Co, and Ni,

precipitate with H2S in dilute HC1 solution. The sulphides
of Sn, As, Sb, Cu are redissolved in HC1 and KC1O8, and
the As, Sb, Cu separated with iron.

ESTIMATION.

(i/) For titration with ferric chloride the filtrate from the
iron reduction is precipitated with strip zinc, the pre-
cipitated tin and lead and undissolved zinc are dis-
solved in pure HC1, and the solution containing the
tin as SnClo is' titrated at the boiling point with ferric
chloride.

(b) For titration with iodine, the solution is not filtered
from the iron deposit, but is cooled, the iron rod with-
drawn, and the solution containing the tin as SnCl2
titrated in the cold with iodine, using starch paste as
indicator.

Chapter III.

The Dry Tin Assay.

A. — TIN ASHES AND SOLDER ASHES.

The fire assay gives low results with ashes which contain chlorine,
grease, sulphur, or zinc, or much lime, slag, or fine carbonaceous
matter.

The general method is to fuse.

50 — 100 grms. ashes with 5 — 30 per cent, potassium ferrocyanide,

10 — 20 per cent, soda ash,
5 — 10 per cent, borax.

in a G Cornish crucible, adding a little fluor and white arsenic if
necessary. A high temperature is necessary, and the best fuel is
gas carbon. When the fusion is complete the assay is allowed to
cool in the crucible, which is broken up when cold. The button of
metal usually consists of a soft or non-irony portion and of a hard or
irony portion It is weighed, and the soft portion is melted away
in a hand ladle and poured into a small hemispherical mould ; the
iron button is weighed and the weight of the soft portion calculated

by difference. Both portions are assayed for tin by the direct
ferric chloride assay (see Wet Assay of Tin, Chapter V., A.), the iron
portion being pounded in a mortar and the non-irony portion broken
in two and filed across the centre. As a rule the soft portion holds
about twice the percentage of tin that the iron holds, but if the iron
button is very arsenical, it is always very poor in tin, and if only
small in amount may be sometimes neglected altogether.

Speaking generally, the iron portion is an iron — tin — antimony
— arsenic alloy, while the soft metal is a lead — tin — copper — anti-
mony alloy, which, however, may contain 10 to 20 per cent. Fe in
the absence of lead. If lead is present in any quantity the soft portion
is as a rule free from iron. If sulphur is present a matte-speise may
be formed as well as a soft portion. In ashes containing chlorine,
sulphur, zinc, there is usually considerable loss of tin by volatilisation.
If thought desirable, chlorine may be removed by boiling the ashes
with carbonate of soda, and the effect of sulphur removed by evapora-
tion with dilute HNO3 previous to the fusion in the crucible. The
poorer the ashes and the greater the amount of lime and silica they
contain, the greater the loss of tin in the slag. Very carbonaceous
ashes also give low results, and if free from zinc and chlorine may
with advantage be calcined previous to reduction. The best results
are obtained from ashes which are mainly composed of metal and
metallic oxides. Ashes which are practically pure oxide of tin, and
irony tin dross containing S and As may be with advantage assayed
by the cyanide method as in the case of tin ores.

This combination of fusion with ferrocyanide and direct ferric
chloride assay of the button is at best a rough method, though it
is a safe buyer's assay ; the result will always average i or 2 per cent,
lower than a complete wet assay both on account of loss by volatilisa-
tion and in slag, and to a slight extent on account of the rough nature
of the separation of the tin from Sb, As, Cu, adopted in the wet assay
of such buttons. It is a purely commercial matter, whether in any
particular instance it is worth while to adopt a complete wet assay
of the ashes in place of the dry assay, depending mainly on the size
of the lot. It is certainly not worth while from any buyer's point
of view to adopt a complete wet assay for a lot of two or three hundred-
weight of solder ashes. In ashes poor in lead the iron button, being
the heavier, is the bottom portion ; but if the ashes are leady the
iron button is lighter and is found above the soft. The irony portion
does not, as a rule, contain copper unless much arsenic or sulphur, cr
much copper, is present. Copper gives a white crystalline fracture
to the buttons from tin ashes, easily recognisable after a little ex-
perience. Antimony gives a white lustrous, crystalline fracture in
rich tin buttons rather different in appearance from the copper fracture.
Zinc gives a bluish crystalline fracture to the metal, and a bluish
flame during the progress of the assay. Arsenic gives a graphitic-
looking fracture to solder buttons. In the absence of Cu and Sb.
the fracture of a rich tin button is dull grey and holocrystalline.

B. TIN SLAGS.

50 grms. slag. 5 grms. white arsenic.

30 grms. ferrocyanide. 5 grms. fluor.

5 grms. soda ash.

Mix, transfer to a G crucible, add a little borax, and fuse in the
hottest fire possible. When in a state of tranquil fusion, remove
from the fire and allow to cool in the crucible. When cold, break
the crucible, knock off the slag, weigh the button of iron — tin-
arsenic alloy and assay for tin. This is an excellent comparative
method for works purposes, but the results are of course low, as it
is quite impossible to reduce all the tin to metal.

C. TIN ORE.

The dry assay of tin ore is only to be recommended for works
purposes, mines, or for prospectors. It is not accurate enough for
buying and selling on equitable terms, having regard to the amount
of money hanging on a single assay and the increasing keenness of
competition.

THE CYANIDE ASSAY is of the most general applicability, but

(a) where the ore is poor and contains much Fe2O3 it should be
digested with strong HC1 provided the tin is present as cassiterite.

(b) If much pyrites is present it should be evaporated to dryness with
HNO3 before extraction with HC1. These operations may be con-
veniently performed in a 6-in. evaporating basin with a clock glass
cover. Oxide of lead, if occurring in tin ore, is very difficult to
extract completely even with strong hot HC1, whilst HNO3 often
removes only half of it. Take 20 grms. dried ore, 20 c.c. of strong
HNO3, and about the same amount of water, and evaporate cautiously
to complete dryness. Add 100 c.c. strong HC1 and digest for half-
an-hour just below the boiling point. Dilute with an equal bulk of
water, filter and wash by decantation until the washings are free
from HC1. Ignite the filter paper and add the ash to the cleaned
residue of ore in the dish. The HC1 extract will usually contain
most of the iron, arsenic, antimony, &c., and will in general be free
from tin, but should always be tested or assayed for tin (see Wet
Assay). It is never safe to assume that the cleaned residue is free
from metallic oxides other than SnO2, though the assumption is
frequently made. Dry the cleaned ore and mix it with an equal
weight of cyanide (98 per cent. Au — not commercial cyanide). Take
a small dry crucible and charge in 10 grms. cyanide, then the above
mixture, and finally 10 grms. cyanide as a cover. Place in a fire
at a low red heat and fuse gradually, increasing the heat to bright
redness at the finish of the fusion, which should not take more than
ten minutes. Allow the assay to become quite cold in the crucible,
and when cold break out the button and either remelt it by dropping
it into a crucible containing melted borax, or under palm oil in a
ladle, or cut it in two and boil out the adhering cyanide with

water. The button should in every case be assayed by a wet method
for tin, otherwise the result is a mere guess and may be very mis-
leading, no matter how much the ore has been cleaned, or however
pure the ore is supposed to be, or however clean the metal appears
to the eye. The use of the term " fine tin " has been already re-
ferred to in the introduction, and it is only necessary to add here
that Sn is a definite entity, while " fine tin " is an expression which
awaits definition. Further, it is quite as bad to assay such a button
by oxidation with HNO3 and weighing the oxide residue as SnO2 —
that would be merely making the same unjustifiable assumptions (in
another form) as to absence of certain impurities, which are involved
by weighing the prill as tin. The button must be assayed positively
for tin by a method which shall ensure the elimination of the inter-
ference of every possible impurity, and that, in practice, means assay-
ing the button for tin by a volumetric method. Of course, if this
procedure is systematically followed (and any other is illogical as an
assay, and so uncertain as a valuation basis as to be inexcusable on
account of the high price of tin), the results will always favour the
buyer. The remedy is. not to attempt to counteract this " low "
tendency, which is inherent in any dry assay, by balancing an un-
known " high tendency " (impurities) against it (the net result of which
is, in practice, to favour the seller), but to use a wet assay throughout.

Of other dry methods of assaying tin ore, the Cornish tin assay
is no doubt useful on mines as a comparative test, where the quality
of the ore remains fairly uniform, whilst the vanning test is also
most useful to prospectors and on mines.

The German assay of tin ore by mixing with oxide of copper and
fusion for white metal, does not appear to possess any advantages
over the cyanide assay as regards accuracy, and is an exceedingly
complicated method.

The method of fusing tin ore witlj Na2CO3 and borax in a luted
carbon lined crucible, in the muffle, is said to give very perfect re-
duction of the tin.

Hallet's method— fusion with KHF2, solution in H2SO4 and pre-
cipitation of the tin as metastannic acid on dilution and boiling —
seems to be a good assay, but is in reality a wet method.

Chapter IV.

The Wet Assay of Tin.

GRAVIMETRIC ASSAY BY WEIGHING AS SNO2.

Bi, Pb, As, Sb, Fe, W, Si, should be absent.

The tin from i grm. of material, separated either as metal or
sulphide, is treated with 20 c.c. of dilute HNO3 (i : i) in an evapo-
rating dish, and evaporated almost to dryness. It is diluted with

B 2

50 c.c. hot water and boiled, then filtered, the residue well washed
with hot water, dried, and ignited in the muffle in a small porcelain
dish. When cold, weigh the SnO2. It contains 78.7 per cent. Sn.

The tin in bronze coins and tin copper alloys free from Sb,
Pb, As, may be estimated this way, but its application to solder,
metal from crucible assays of tin ore, and the sulphide precipitate
from tin slags is inaccurate.

If solder is treated in the above manner the residue consists of
SnO2 and SbO2 and some PbO. Multiply the weight of residue
from i grm. of solder by 78.7, and the result, less i per cent, de-
duction for lead, may be taken as the sum of the percentages of
tin and antimony in the solder.

GRAVIMETRIC ASSAY BY ELECTROLYSIS.

This assay is fully described in Classen's "Chemical Analysis
by Electrolysis."

VOLUMETRIC ESTIMATION WITH FERRIC CHLORIDE.

When ferric chloride is added to a strong hot HC1 solution of
stannous chloride it is immediately reduced to ferrous chloride, and
stannic chloride is formed at the same time. One drop in excess of
the ferric chloride gives a decided yellow colour to the previously
colourless solution, provided the solution is hot and strongly acid.
The nearer the boiling point and the greater the concentration of
HC1 in the solution the more rapid is the completion of the reaction.
A solution of FeCl3 in dilute HC1, of which 100 c.c. = 2 grms. Sn,
is employed. In the assay the addition of FeCl3 from the burette
cools the solution somewhat, so that the finish is rather slower than
the commencement of the reaction, both owing to dilution and con-
sequent lowering of the temperature, and to the presence, in increasing
concentration, of ferrous chloride in the solution, but in any case the
titration should never take more than a minute, if worked as directed.
FeCl3 gives a far stronger colouration in a hot, strongly acid solution
than in a cold, faintly acid solution. The titrations cannot be done
by gas light or electric light, and should always be effected in the
daytime. In an emergency they may be done by magnesium light.

The equation representing the chemical change is 2 FeCl3 -f-
SnCl2 = 2 FeCl2 + SnCl.t. The presence of chlorides of lead, zinc,
aluminium, iron (ous), cobalt, nickel, antimony (ous), copper (ous),
cadmium, does not affect the quantity of FeCl3 required ; the presence
of FeCl2 in quantity somewhat retards the finish and lessens the
delicacy of the colour indication ; Cu2Cl2 reduces FeCl3 with forma-
tion of CuCl2, but SnCl2 reduces CuCl2, and the net result of this is
that not a trace of CuCl2 is formed until all the SnCU is converted
into SnCl4 — the next drop of FeCl3 forms a trace of CuCl2, which
gives a similar colour indication to that of FeCl3 itself ; CoCl2 and
NiCL, give highly coloured solutions which render the recognition
of the end point difficult — one way of remedying this is to dilute

the solution somewhat, with boiling water, which removes the blue
colour, but of course renders the reaction slower and lessens the
colour intensity of the drop or two excess of FeCl3 ; SbCl3 is not
converted into SbCl5 by FeCl3, and under the conditions of the
a^say neither SbCls nor SbCl5 ever occur in the solution ; Cu0Cl2,
CoCl2, NiCl2 are rarely present, also CdCl2 and A12C16 ; FeCl2 is
often present to begin with, and ZnCl2 and PbCl2 are generally present
in greater or less amounts. BiCl3 and HgCl2 are reduced to metal
by SnCl2, but Bi and Hg would be separated with iron. Precipitated
Sb, Cu, As, are attacked by hot acid ferric chloride and blue oxide
of tungsten is affected by it, but the assay method excludes the
presence of these substances during titration. Acid solutions of
SnCl2 very readily oxidise by exposure to air ; the method of dis-
solving the tin from the state of metal by boiling with HC1 in an
atmosphere free from oxygen excludes the formation of SnCl4, if the
operation is carried out as subsequently directed, and if the solutions
are titrated as soon as ready and at the boiling point the oxidation
tendencies are completely eliminated. Briefly, the best conditions
are : (i) Solution from the state of metal as rapidly as possible in a
non-oxidising atmosphere, the solution being brought to the boiling
point before the solution is complete. (2) Use of strongest and
purest HC1. (3) Bulk 150 — 250 c.c. (4) Titration rapid and at the
B. Ft. (5) Strength of FeCl3 100 c.c. = 2 grms. (6) Absence of
precipitated Sb, As, Cu in the solution.

From the dilute peroxidised HC1 solution of Sn, Sb, Hg, Bi, As,
Cu, Pb, Cd, Zn, Co, Ni, Fe, P, the Sn, Sb, As, Cu, Hg, Bi, and
some Pb and Cd are separated as sulphide by H2S if it is desired to
separate from Co, Ni, Fe, P. The sulphide precipitate is redissolved
in HC1 and KC1O3, and the solution reduced by heating with iron
wire. The As, Sb, Cu, Hg, Bi are precipitated in the metallic form,
and the solution (which must be strongly acid to avoid precipitation
of SnOCl2) is filtered and neutralised with thin strips of zinc. The
action finished, the mother liquor, after testing for tin with H3S water,
is poured off as completely as possible, and the residue of spongy
metallic tin and lead and undissolved zinc, is dissolved in the same
flask in about 200 c.c. of pure HC1, the flask being provided with
a rubber cork and leading tube, and the liquid is brought to a boil
as rapidly as possible ; a piece of pure zinc about the size of a
pea is added to assist in preserving a non-oxidising atmosphere of
hydrogen and hydrochloric acid in a flask until the liquid clears
and boils. As soon as everything is in solution and the liquid is
boiling, the flask is removed from sandbath or plate and titrated
immediately with ferric chloride. The ferric chloride should be free
from ferrous chloride, nitric acid, chlorine and arsenic, and the solution
should contain 300 — 500 c.c. HC1 in two litres. It is best made up
from a concentrated stock solution in HC1, made by dissolving piano
wire as directed subsequently. If the assays turn dark greenish after
titration the FeCl3 solution is contaminated with HNO3. The FeCl3
may be made up also by dissolving 180 grms. of the yellow com-

mercial lump salt, which is Fe.,Cl6 12 H2O, in about 200 c.c. HC1
and evaporating it to dryness. The residue is dissolved in 300 c.c.
HC1 and diluted to two litres. The solution is standardised against
i grm. of the purest tin obtainable, filed with a fine file. This is
weighed into an 8 oz. flask, and the flask is about three parts filled
with pure HC1, rubber cork, and leading tube inserted, and boiled (but
not too rapidly) until solution is complete ; then titrated at once.

Pure tin is more readily obtained from smelters of tin ore than
from dealers in chemicals — the writer o?ice ordered some " pure tin
for standardising purposes " from a firm of wholesale chemists and
received metal holding 3 per cent. Sb.

It is rarely necessary to complicate the assay by separating the
tin as sulphide — having once got everything in solution, reduce with
iron wire, filter, and precipitate on zinc. This method has been re-
peatedly checked on made up metals of known composition containing
varying amounts of Sn, Pb, Cu, Sb, and the results are in every case
so close as to leave no doubt whatever that the method is extremely
accurate ; indeed, it is much more accurate than the electrolytic assay
on account of the complicated separations which the latter involves,
and is incomparably quicker.

The favourite objection to the ferric chloride assay is the oxida-
tion tendency of solutions of SnCl2 — the method if properly worked
overcomes this completely. It has also been objected that five or
six drops of ferric chloride solution are necessary to give a perceptible
colour indication. This is quite incorrect — one drop in excess of
FeCl3 is ample if the operator possesses normal colour vision. It is
also stated th,at Sb dissolves in HC1 — this is not the case. It is true
that finely divided Sb in contact with air and HC1 slowly dissolves,
but even then, FeCl3 does not oxidise SbCl3, and, further, under the
conditions of the assay the absence of Sb is ensured by the iron wire
separation. In direct ferric chloride assays on solutions from metal
filings, the assays so far from being too high through Sb dissolving,
are too low because of Sn retained with the black powder (Chapter V.,
A.), and it should also be remembered that metallic tin precipitates
antimony from solution.

VOLUMETRIC ESTIMATION WITH IODINE IN ACID SOLUTION.

The dilute acid solution of metallic chlorides, which should not
be more than about 50 c.c. in bulk, and should be contained in a
4-inch beaker, is reducted by heating with a clean piece of iron rod
resting in the solution against the side of the beaker, which is
covered by a watch glass. The whole is heated to 80° or 90° C. (not
to boiling) over a Bunsen flame ; five or six assays may be conveniently
heated in a small frying-pan sandbath. The assays are heated for
20 — 30 minutes after they have lost their original red, yellow, or
greenish colour. The Sb, As, Cu, are precipitated, and the SnCl4
is assumed to be reduced to SnCl2 ; in practice this assumption is
found to be justified, though it is really one of the weak points of

the assay. The assays are cooled in a basin of cold water, and when
cold the watch glass and rod are rapidly washed with a little cold
boiled water, starch paste added, and the solution titrated rapidly
with iodine. It is not necessary to remove the black precipitate of
Sb, As, Cu, as the finishing point in the case of SnCl2 and iodine is
so sharp ; but the finely-divided metallic precipitate sometimes seems
to slowly remove the blue colour. The iodine solution is made up by
dissolving 21.32 grms. iodine and 45 grms. pure KI in about an inch
of water in a small beaker and diluting to one litre. 100 c.c. =
i grm. Sn. Not more than .5 grm. Sn should be present in the assay.
It has been proposed to increase the accuracy of this assay by
titrating in an atmosphere of CO3.

Mr. A. H. Low states ("Technical Methods of Ore Analysis,"
2nd Edition, page 185) that copper and iron in contact precipitate
tin. He does not, however, state the conditions which obtained when
he made the observation, and the writer wishes to emphasise the fact
that in a freely acid (HC1) solution (a sine qua non in tin assays), no
such precipitation of tin ever takes place.

Titanium and tungsten both interfere with the ferric chloride and
tungsten with the iodine titrations for tin, as titanium sesquichloride,
TioCl6, and the blue oxide of tungsten, WO2, are peroxidised in HC1
solution by ferric chloride, and WO2 by iodine. Titanium sesqui-
chloride is not affected by iodine, and titanium does not interfere with
the iodine titration. Hence Ti and W, if present, must be separated
for the FeCl3 assay, and W must be separated for the iodine assay.
WO2 is precipitated with antimony by iron, and Ti2Cl6 remains in
solution when tin is precipitated on zinc provided the liquid is dis-
tinctly acid with HC1 when poured off from- the precipitated tin.
Uranium does not interfere with either iodine or ferric chloride assays.
Uranic salts in HC1 solution are reduced by iron, but uranous chloride
is not oxidised by iodine. Uranous chloride which makes boiling
HC1 solution greenish, is peroxidised slowly by FeCl3 after the tin
has been peroxidised ; as in the case of Cu2Cl2, the colour change
to yellow takes place as soon as the tin is peroxidised, so that uranium,
if present, does not affect the titration reading. In this case, how-
ever, the colour due to FeCl3 slowly goes as the uranium is per-
oxidised, and is replaced by the much fainter yellow of uranic
chloride ; so that the exercise of considerable care is required in noting
the finish of the tin titration. Further, zinc precipitation leaves the
uranium in the mother liquor if this is kept sufficiently acidulated.
Molybdenum interferes with both iodine and ferric chloride assays,
and tin requires to be separated from it by zinc precipitation, the
mother liquor, if sufficiently acid after separation of the tin, containing
the molybdenum. Molybdous compounds colour a boiling HC1 solu-
tion reddish ; on titration with ferric chloride the colour changes to
green, finishing a strong yellow.

Chapter V.

Specific Tin Assays.

A. — DIRECT FERRIC CHLORIDE ASSAY. — Applicable to alloys of
lead, tin, zinc, aluminium, iron, arsenic, antimony, copper, cobalt, and
nickel, with less than 5 per cent. Sb, As, Cu, and to certain tin copper
alloys.

SOLDER, TERNE, PEWTER, CAPSULE METAL, &c. BUTTONS OBTAINED

IN DRY ASSAY OF TlN AsHES.

The sampling of alloys of lead and tin and antimony is quite as
important as the assay, on account of the pronounced liquation
phenomena which these alloys exhibit. Wherever possible the sample
should be taken from the molten kettle of metal. Where there is no
option but to sample from the pigs, chips may be taken from the centre
of opposite top and bottom longitudinal edges, and the chips melted in
a ladle. The sample for assay should be cast in a. small hemispherical
or rectangular mould (not in long thin strips) and cut transversely in
two with a chisel. Filings are taken with a clean and fairly fine cut
file across the surface of fracture, and the filing should be done gently.
Very brittle alloys such as irony arsenides may be pounded in a steel
mortar. The objection sometimes raised against filing, that the use
of a file involves the contamination of the sample with particles from
the file, is more imaginary than real, and in practice may be ignored
quite safely.

THE ASSAY. — Weigh up i grm. of filings into a clean 8-oz. flask
and add about 200 c.c. of pure HC1 (s.g. 1.16). Close the flask
with a rubber cork with leading tube attached as shown, and boil
the assay on a sandbath enclosed in a fume
cupboard, over a moderate fire. As the object
is to obtain the tin in solution as stannous
chloride, precautions are necessary in order to
prevent oxidation. The assay should not boil
too rapidly, or the HC1 becomes weakened in
strength before it has had time to thoroughly
attack the metal, and any black residue is
broken up so finely that it settles afterwards
only with great difficulty. Neither should the

assay come to a boil too slowly, as in that case the metal may be com-
pletely attacked while there is still air in the flask, with attendant
formation of SnCl4. The tin goes into solution as SnCl2 — tin,
lead, zinc, aluminium, iron (forming FeCU), cobalt, nickel, all dis-
solve, and in some cases copper (forming Cu2Cl2) whilst arsenic,
antimony, copper (also bismuth and mercury) are left undissolved as
a black powder, also often considerable quantities of iron, and a
certain amount of unattacked tin and lead occluded by the powder.

Some As is evolved as AsH3, but in general the As remains as a
brownish flocculent residue, while only faint traces of SbH3 are formed.
Certain tin copper alloys are completely soluble in HC1, giving a
colourless solution of SnCL and Cu2Cl2. Cn2Cl2 is oxidised to
CuCL by FeCl3, but its presence does not affect the tin titration, as
was explained in the previous chapter. The presence of SbCl3, PbCl2,
FeCl2, CoCljj, NiClo, ZnCl2, has no effect on the titration figures.
Alloys containing only lead, zinc, tin, and aluminium go completely
into solution in fifteen to twenty minutes, and may then be immediately
titrated with ferric chloride ; such alloys yield perfectly accurate results
by this method.

In the presence of a residual black powder the results will always
be too low, though when the percentage of antimony and copper
is low the results are near enough for many purposes, as, for example,
in the assay of buttons got in the fire from tin ashes and solder
ashes.

Thus solders with 20 — 60 per cent. Sn and i| — 4 per cent. Sb
yield results for tin by this method which are ^ — i per cent, too low.
Alloys of iron and tin are also difficult to attack completely, and
further the presence of much FeCl2 in the solution retards and ob-
scures the finish somewhat. CoCl2 and NiCl2 give coloured solutions
in HC1 which also obscure the colour finish ; this may be oxercome
by adding hot boiled water before titration.

In all cases where solution in boiling HC1 is incomplete, the assay
is allowed to boil until the metal is completely attacked. This
is a matter for the judgment of the operator, and requires experience
to determine, but in general the fine stream of gas bubbles is replaced
by a slower stream of larger bubbles as solution proceeds. The assays
must not be allowed to boil below 150 c.c. or the FeCl3 used in
titration may cool the assay so much as to impair the sharpness of
the colour indication at the finish of the titration. It is an improve-
ment to allow the leading tubes to dip under water — this may be
effected by attaching a straight vertical tube to the ordinary leading
tube, with a piece of rubber tubing, or the leading tubes may be made
longer and with a second right angle bend ; or a rubber tube valve
may be fitted on the glass leading tube to prevent air finding its way
into the flask.

As soon as the metal is completely attacked, remove the assay
to the titrating bench and allow it to stand for a minute or two until
the black powder has settled. A slight rotatory jerk, given to the
flask when it is removed from the sandbath, assists the powder to
collect in the centre of the flask. Remove the cork and leading tube
and carefully pour the liquid off into another clean flask, from any
black powder — pour oft as closely as possible, but so as not to carry
over any powder into the other flask. (The black residue of Sb, Cu,
As is soluble in FeCl3). Titrate the hot acid solution of tin rapidly
with ferric chloride solution from a fast running burette with glass
stopcock. The assays must not be allowed to stand off the fire before
titrating, for more than two or three minutes. After five minutes the

effect of oxidation begins to be perceptible, and the oxidation increases
very rapidly on further standing.

The ferric chloride solution should be made up two litres at a
.time from a concentrated stock solution made by dissolving piano
wire in HC1, peroxidising with HNO3 and evaporating twice with
HC1, to dryness, and then dissolving in HC1, and is standardised
.against i grm. of purest tin dissolved by boiling with HC1 as usual.
The standard solution may also be prepared by dissolving 180—190
grams of the yellow lump salt (which is, roughly, Fe.,Cl6, 12 H2O)
in 200 c.c. HC1 and evaporating the solution just to dryness. The
residue is dissolved cautiously, and in a fume cupboard, in 300 c.c.
HC1, and the solution diluted to two litres. Evaporation with HC1
is necessary to free the solution from HNO3 and arsenic. Only the
purest and strongest HC1 should be used in these assays.

As previously stated, in the case of alloys which dissolve com-
pletely in HC1 the above method is perfectly accurate, but where
.solution is incomplete (i) the black powder holds Sn, (2) oxidation
possibilities are introduced in pouring off into the second flask,
(3) a little powder may get over into the second flask, (4) it is im-
possible to pour off quite clean— though the error from this cause is
much less than might be supposed, and may be practically eliminated
by pouring a little hot boiled water on to the residue in the first flask
immediately after decantation, and decanting off the washings into the
second flask at the end of the titration. Admitting that in cases of
incomplete solution the results are too low, they are in all cases
sufficiently close to render the method very reliable, especially when
'One considers the extent to which antimony and tin are allowed to act
as understudies to one another in so many assays. Briefly, in cases of
complete solution, this direct assay is quite accurate, while with in-
complete solution the results are low, but provided not more than
5 per cent. Sb and Cu are present they are reliable and within i per
cent, of the Sn percentage. In the case of plumber's solder and
tinman's solder it is a fairly safe rule to add ^ per cent, in the case
of plumber's and f per cent, in the case of tinman's; by doing this
it is possible to arrive more accurately at the percentage of Sn in solder
in half-an-hour than is otherwise possible (except by the next assay
to be described) in a week. Tinman's solder very commonly contains
46 per cent. Sn and 3 per cent. Sb. This metal assays 50 per cent.
Sn by oxidation with HNO3 and weighing residue as SnO.,. Plumber's
.solder often holds 30! per cent. Sn and 2 per cent. Sb.

All tin assays for estimation with ferric chloride are finished in
the manner just described. The precipitated tin and excess zinc are
dissolved up together, over a good fire, in a flask with leading tube,
and a granule of zinc is added to assist in keeping a non-oxidising
atmosphere in the flask until the assay comes to a boil. In the
general tin assay, As, Cu, Sb are always absent, being separated in
the iron wire stage, so there is no error through decantation.

Alloys containing zinc or aluminium effervesce violently in the
cold, with strong HC1. Tungsten gives a bluish tint to the solution.

Cobalt gives a blue colour much weakened by dilution, and nickel a
weaker green colour which goes on dilution.

B. ALLOYS OF LEAD — TIN — ANTIMONY holding muchSb and but

little Cu and As.

TYPE-METAL — ANTI-FRICTION METALS.

Weigh up i or 2 grms., according to the probable amount of
tin present, into an 8-oz. flask, and boil gently with 50 — 75 c.c. HC1
until action ceases, when most of the Pb and Sn are in solution.
Complete solution is effected by the cautious addition of a saturated
solution of KC1O3. Boil off the excess of chlorine, remove the
assay from the bath, and add at once two bunches of fine piano
wire. The action in the hot, strongly acid liquid is very rapid, and
one minute after the solution turns colourless add a thin strip of
copper foil. As soon as a small piece of copper remains bright
(alter a minute or two) dilute with 30 or 40 c.c. of hot boiled water
and boil for a minute. Filter very rapidly into another similar flask,
keeping the precipitate as far as possible out of contact with the air,
and wash the flask and paper once, with hot dilute HC1. Neutralise
the solution (which should be freely acid) with thin strips of zinc —
use excess of zinc straight away, as if the neutralisation is effected
too slowly tin is apt to remain unprecipitated.

When the action ceases pour a little of the liquor off into a
beaker containing a little H3S water; if the precipitate is white, all
the tin has been precipitated. Pour away the mother liquor as closely
as possible after allowing any floating particles of metal to settle.
The best plan is to empty the liquid from the flask into a beaker,
which is then rapidly filled with water from the tap. Larger pieces of
spongy tin are pressed against the side of the beaker with a glass
rod, after which they readily settle ; the addition of a drop or two of
ammonia or pouring the liquid to and fro from one beaker to another
will generally ensure the settling of the lighter particles. The metal
in the beaker is washed back into the flask with 180 c.c. HC1 and the
liquid brought to a boil as rapidly as possible with the addition of a
granule of zinc and using a cork and leading tube. When solution
is complete titrate at once with ferric chloride.

Black powder should be absent, as all the As, Sb, Cu are pre-
viously separated. This method has been repeatedly checked on
made up alloys of known composition, and is perfectly accurate.
From start to finish it need not take more than one hour, but the
assays require unremitting attention. Copper is best precipitated
on iron in a dilute HC1 solution, whilst antimony comes down im-
mediately in a strong HC1 solution. The arsenic is not really pre-
cipitated by iron but by the SnCL, formed by the action of Fe on
SnCl4 ; iron precipitates Cu and Sb. Copper precipitates As and
Sb ; tin precipitates As, Sb, and Cu. Phosphorus is said to bring
down some tin on iron wire, and in the case of arsenic the pre-
cipitated metal holds 3 or 4 per cent. Sn. It may be asked, why

reduce the tin to metal ; why not stop at the half-way stage when
the SnCl4 is reduced to SnCl2 ? This is actually done in the iodine
assay, but the procedure described affords a far surer way of obtaining
the tin in solution as SnCl2, and further it is necessary to obtain
the SnCl2 in strong boiling HC1 solution free from precipitated As,
Sb, and Cu, and this can only be ensured by the above method.

c- In alloys which contain much Cu, and with non-ferruginous

mattes, a little HNO8 may be used instead of KC1O3 to effect complete
solution, and the reduction by iron wire prolonged in dilute solution.

D- ALLOYS AND MATTES rich in Fe, Co, Ni, P.

Weigh up i grm. and dissolve in 20—30 c.c. HC1 and i c.c.
HNO3. Heat until solution is complete, boil off red fumes, dilute to-
200, and precipitate with H2'S. Allow to stand for an hour or two,
filter, but do not wash, as SnS2 is decomposed by water. (In cases
where the filtrate is required wash with NaCl solution.) Test the
filtrate with a further stream of H2S. Wash the sulphide precipitate
back into the flask, add 30 — 40 c.c. HC1 and some KC1O3, boil, add
iron wire, and heat to 70° or 80° for half-an-hour to two hours, or
until reduction is complete. Filter, precipitate the tin on zinc, dis-
solve and titrate with ferric chloride.

E- ALLOYS, MATTES, AND SPEISES rich in arsenic and iron.

i. — In many cases methods C or D are quite satisfactory provided
the assay is well boiled with plenty of iron wire, and is kept
freely acid. The assay is too low for tin by 5 per cent, of
the arsenic percentage.

2. — Method D is adopted, but instead of dissolving the mixed
sulphides in HC1 and KC1O3 they are boiled with 50 c.c.
HC1 alone, down to about 30 c.c., diluted and filtered from
the insoluble sulphide of arsenic. The filtrate is heated
with iron wire to precipitate the Sb, filtered and precipitated
with zinc as usual. This is a rough and ready separation,
as the As2S3 always holds some SnS2, and if much copper
is present the loss is increased by the Cu2S which is un-
dissolved in HC1.

3. — The arsenic may be distilled off with a solution of FeCl3 and
CaCl2, and the tin separated as sulphide from the diluted
solution. This is sometimes convenient where the arsenic
has to be estimated, as there is no loss of SnCl4.

4. — One grm. 25 c.c. HC1 and 2 c.c. HNO3. Heat till dissolved
and boil off the fumes. Add sodium sulphide solution very
cautiously until solution is alkaline and precipitate is black,
and pass H2S. Warm, dilute, warm again, filter and wash
well with hot dilute Na2S. (The precipitate if very bulky
should be redissolved and reprecipitated.) The alkaline
filtrates are cautiously acidified with HC1, the liquid allowed

to stand for half-an-hour in a warm place, and the sulphides
of Sn, Sb, As are filtered, washed well with NaCl, and
washed back into an 8-oz. flask, and boiled down with
50 c.c. HC1 to 30 c.c., then diluted and filtered from the
As3S3. Boil with iron wire, filter and precipitate on zinc.

5. — Or separate with H2S in acid solution first, and digest the
precipitate with Na2S. Then proceed as before.

^.—Dissolve in 30 c.c. HC1 and a little KC1O3. Boil off the
chlorine, and precipitate with H2S in strong HC1 solution.
Allow the solution to stand for some time in a warm place
and again pass H2S. As2S3 alone is precipitated (and
Cu2S). Filter through asbestos, dilute and precipitate the
SnS2, &c., with H2S.

7. — Or having obtained the sulphides of As, Sb, Sn, separate the
antimony by electrolysis, the arsenic with magnesia mixture,
and the tin by electrolysis. (See electrolytic assay.)

8. — The oxalic acid separation may also be used with arsenical
material. One grm. is dissolved in 20 — 30 c.c. HC1 and
i c.c. HNO3. Add 20 grms. oxalic acid dissolved in 100 c.c.
water and gas the solution with H2S, while at the boiling
temperature. Two separate gassings are necessary. Filter
the sulphides of As and Sb on the water bath and wash
with a hot concentrated solution of oxalic acid. Neutralise
the filtrate with ammonia, acidify with acetic acid, dilute
to i litre, warm, and precipitate the tin completely with
H2S. Filter, dissolve in HC1 and KC1O3, and precipitate
with zinc as usual.

F. GENERAL ASSAY applicable to everything not containing the

tin as silicate ; suitable for calcined matte and calcined speise, bronze
ashes, many tin copper alloys, chloriny and leady tin ashes, sulphide
•ores, metallic copper.

i to 5 grms. is evaporated to complete dryness with 10 — 20 c.c.
HNO3 and some water, in a covered porcelain dish. It is then
digested with dilute HNO3 and filtered. The residue will contain
•all the tin, and is dried and ignited with the filter paper in the same
dish, the ignition being best done in the muffle. The calcined residue
is carefully transferred to a porcelain boat and reduced at a low red
heat in a current of coal gas for an hour and a half or two hours (see
Wet Assay of Tin Ore). The boat and its contents are allowed to
cool in the tube in a current of coal gas, and when cold are trans-
ferred to an 8-inch beaker, and the metal dissolved in HC1 and a little
HNO3, the nitrous fumes boiled off, and the solution filtered. The
residue of carbon and silica is dried, ignited, and examined for Sn
as a matter of precaution by fusing with KNaCO3 and borax in a
platinum crucible, dissolving the melt in HC1, and precipitating with
zinc, &c.

The main solution if from 5 grms. is made up to 500 c.c., and 100
c.c. are pipetted into an 8-oz. flask.

(a) In presence of much arsenic, see Section E. Most of

the arsenic is, however, volatilised in the reduction tube.

(b) In presence of a little arsenic, precipitate the As and

Sb on iron, and the tin on zinc as usual.

SnS3 with H2S in acid solution. (See Section D.)

(d) In presence of tungsten, the uncalcined residue obtained

by evaporation with HNO3 is extracted with dilute
AmOH or Am2CO3, when WO3 dissolves.

(e) Copper, cobalt, nickel, zinc, mostly go into the HNO3

solution, also much of the lead, whilst usually most of
the iron and sometimes half the lead remain with the
SnO2. Some Co and Ni may remain with the tin, in
which case a separation with H2S is advisable as the
coloured chlorides of Co and Ni interfere with the
appearance of the colour finish in the ferric chloride
titration.
In any tin compound or mixture of substances containing tin

(except tin slag), the tin may be brought into solution by the above

method.

G. ASSAY OF MISCELLANEOUS ALLOYS.

i. — TIN PLATE — SPELTER — TEA LEAD. — 5 grms. is boiled with
300 — 400 c.c. HC1 in a i6-oz. flask provided with cork and
leading tube. When the action is complete pour off from
any residue and titrate with ferric chloride. In the case of
spelter use less acid or more spelter, and bring rapidly to
a boil. See article by R. Job, " The Mining Journal,"
December 12, 1908, p. 743.

2. — TIN FOIL— CAPSULES — BORINGS. — 50 or 100 grms. is dropped
into a crucible three parts full of melted borax with a little
charcoal sprinkled on top. The button of metal is assayed
for tin by method A or C, according to composition. The
presence of zinc or aluminium in borings is easily recognised
by the ease with which they burn in the crucible, and by
the violent effervescence with cold HC1.

3. TIN DRILLINGS AND SAWINGS AND FINE BORINGS. — 5, 10, 20,
or 40 grms. is dissolved in a large beaker in 100 — 500 c.c.
HC1 and the minimum amount of HNO3. When solution
is complete, dilute to 500, 1,000, or 2,000 c.c. and measure
out the equivalent of i grm. Reduce with iron, filter, and
precipitate on zinc.

4. — ALLOYS OF TIN AND ALUMINIUM. — May be assayed by
method A.

5. — METALLIC COPPER. — Use method £.4. — separation with Na2S,
or, better, method F. — treatment with HNO3 and reduction
of residue. Take 5 to 10 grms. for assay.

6. — TIN AMALGAM AND FUSIBLE ALLOYS. — Weigh up to 5 grms.
Treat with dilute HNO3 and evaporate to dry ness- -extract
with dilute HNO3 — filter — calcine and reduce in coal gas.
Dissolve as usual in HC1 and a little HNO3, separate Sb,
&c., with iron, and precipitate tin on zinc.

H TIN ASHES.

Samples of tin ashes should in general be tried for moisture and
the tin assay done on the dried sample. In sampling material of
this kind a sample weighing from 14 to 28 Ibs. is broken up and
made to pass through a quarter inch sieve. Large pieces of iron are
picked out and weighed and allowed for in making up final calcula-
tions. 100 grms. of the dry ashes are pounded in an iron mortar
and passed through a sieve with 16 meshes to the linear inch to
separate the metallics. The metallics are weighed and the difference
from 100 is the weight of the " fine " which is not separately weighed
owing to slight loss of dust in pounding. 5 to 10 grms. of the ashes
(with fine and metallic in proportion) is weighed out and treated by
one of the following methods. Most tin ashes are best assayed by
method i. Method 2 is suitable for chloriny or leady ashes (and for
tin ore). Finely divided ashes, if free from chlorine, grease, sulphur,
and metallic— e.g. tin oxides, may be tried by method 3. Siliceous
ashes in which silicate of tin is present are treated by method 4.

i. — GENERAL METHOD. — 5 to 10 grms. ashes is weighed into a
400 c.c. beaker, treated with 100 c.c. HC1 and 5 or 10 c.c. HNO3,
and allowed to stand in a moderately warm place until the action
abates. It is then boiled until the further action is completed.
Dilute with an equal bulk of water and filter. If the ashes are
leady and PbCl2 clogs the filter, wash with hot sodium acetate,
when the PbCl2 dissolves (alternative to the use of method 2). Wash
the residue well, firstly with dilute HC1, then with hot water, dry,
and calcine with the paper in a porcelain dish in the muffle. Transfer
the calcined oxides to a porcelain boat and reduce to metal in a
current of coal gas at a low red heat. Time of reduction, i| — 2
hours. Transfer boat and contents to the same beaker when cold,
and add 50—100 c.c. HC1 and 2 or 3 c.c. HNO3 ; boil off the nitrous
fumes and dilute. Filter into the first HC1 solution and well wash
the residue which is dried, calcined, and fused with four or five times
its bulk of a mixture of KNaCO* and borax, the fusion being done
in a platinum crucible in the muffle. The melt is dissolved in HC1
and the residual tin precipitated on zinc as usual. The solution
is made up to 500 or 1,000 c.c. and the equivalent of i gram is
measured out with a pipette and reduced with iron wire in an 8-oz.
flask, filtered and precipitated with zinc as usual, prior to solution
in HC1 and titration with FeCl3.

2.— EVAPORATION with dilute HNO3 and reduction of the residue
in coal gas (see method F). Leady or chloriny ashes.

3. — FINE ASHES free from Cl, S, grease, moisture, and metallics.
5 grms. is weighed into a porcelain boat and reduced for 2 hours in

a current of coal gas. The reduced meta's are dissolved as in
method i.

4. — SLAGGY ASHES. — 5 grams is weighed into a platinum basin
and evaporated just to dryness with HF. (If the ashes are also
chloriny they must be first boiled in a porcelain dish with Na2Co3,
and the liquid filtered off and, of course, examined for tin.) To
the residue from HF add 40 — 50 c.c. HC1, heat to loosen from the
basin and transfer to a 400 c.c. beaker. Add 50 c.c. HC1 and 5 c.c.
HNO3 and proceed as in i.

5. — TIN CHALK, TIN PASTE, TIN MUDS, AND PRECIPITATES in
which the tin all exists as soluble oxide, sulphide, oxychloride, or
as metal.

10 grams is dissolved in 100 c.c. HC1 and a little HNO3 diluted
and filtered. The residue which is small in amount is fused with
KNaCO;3 and borax. The solution is treated as in i unless Sb, Cu,
As are known to be absent when the reduction with iron wire is
omitted, the measured portion being at once treated with zinc.

6. — TIN GREASE.— 5 grms. is washed with benzene in a dish,
then filtered, dried, and treated like ordinary tin ashes. Slightly
greasy ashes should be treated by method 2.

7.— IRONY SULPHARSENIDE TIN DROSS 20 grms of the dross is

treated with 300 c.c. HC1 and 20 c.c. HNO3 as in the case of ordinary
tin ashes. The solution is diluted and filtered and the residue well
washed. In this case the residue is washed into a dish without
opening out the paper and evaporated to dryness with a little HNO3,
extracted with dilute HNO3, and filtered through the same filter
(not, of course, into the main solution). The residue is then dried
and ignited as usual, and may be then reduced in coal gas, or if
small in amount may be fused at once with KNaCO3 and borax. The
tin is determined in the usual manner — reduction with iron — precipi-
tation on zinc — solution in HC1, and titration with ferric chloride.

8. — VERY CHLORINY ASHES which often contain much ZnCL and
large fragments of metal (" tin scruff") should not be assayed on the
dry material, as this is extremely hygroscopic. Weigh out 100 or
200 grams of the ashes, wet, on sampling, boil with NA.,CO3 and filter.
The washings are made acid with HC1 and a measured fraction pre-
cipitated on zinc. The residue is dried and separated into fine and
metallic. The fine is dried by method i or 2. The metallic is melted
with reducing agents in a clay crucible and the button assayed
for tin by A or C. If the ashes are chloriny but without large
fragments of metal, weigh up 10 — 20 grms. wet and treat with
200 c.c. HC1 and 100 c.c. HNO3 in a large beaker.

K.— ASSAY OF TIN ORE.

Tin ores are often exceedingly complex ; they may contain in
addition to stannic oxide, ferric oxide, and silica, some or all of the
following substances — bismuth, copper, pyrites, iron pyrites, mis-

pickel, wolfram, titanic acid, lead oxide, antimony oxide. Antimony
and arsenic are common impurities, especially in South American ores.
Contrary to a statement in " Crookes' Select Methods," the writer's
experience is that antimony is almost invariably associated in small
quantities with tin ore. It should be apparent that complete solution
of the ore is absolutely essential in every case, in order to systematically
ensure the complete extraction of the tin and its quantitative deter-
mination. Any method which does not involve complete solution of
the ore is quite unreliable as a method, although it may often yield
correct results. Further, the final determination of the tin should
always be effected volumetrically ; if a gravimetric estimation is adopted
the assay develops into an academic research, owing to the number
and complicated nature of the separations necessary to ensure that no
possible impurity may score as tin ; while if such thorough separa-
tion is neglected, the results obtained are quite unreliable.

The following methods have been proposed and used for the wet
assay of tin ore : —

i. — Fusion with alkalies or alkalies and sulphur in nickel, silver,
platinum, or porcelain crucible.

2. — Continued agitation with zinc and HC1.

3. — Reduction of the cleaned stannic oxide in hydrogen and cal-
culation of the tin from the loss in weight, which is assumed
to represent only the oxygen of the stannic oxide.

4. — Cleaning with HC1 and reduction with cyanide in a porcelain
crucible.

5. — Fusion with potassium hydrogen fluoride, solution in sulphuric
acid and precipitation as metastannic acid by dilution and
boiling.

6. — Reduction in coal gas or hydrogen and extraction with HC1 and
HNO3, combined with fusion of the siliceous residue with
Na2CO3 and borax in a platinum crucible and solution of
the melt in HC1.

7. — Fusion with caustic soda in an iron crucible.

The first method is now hardly ever used, being tedious and un-
certain. The second method, agitation with zinc and HC1, is very
slow ; it may occasionally give all the tin, but there can be no
certainty whatever about such a method, and the same remark applies
to the third method, calculation of the tin from the weight of oxygen
lost by reducing the stannic oxide in hydrogen after presumably
purifying it from other metallic oxides by cleaning the ore with acids ;
rapidity is claimed for this method, but the assumptions as to com-
plete reduction of the tin and absence of other metallic oxides reducible
by hydrogen, are vital objections, and render the method unsound
from either a scientific or commercial standpoint. Method four,
reduction of cleaned ore with cyanide, is in reality a dry assay, not-
withstanding the use of the porcelain crucible, and when the metal

obtained is assayed for tin, must necessarily give results which are
lower than the tin contents. It is not denied that in many cases
the results will only be very slightly under the actual percentage,
but one can never be certain of this ; often, indeed, results obtained
in this way are 2 or 3 per cent, too low. If the button of metal
is not assayed for tin, the results obtained may be either a little too
low, correct, or up to as much as 4 or 5 per cent, too high. The
method is not sufficiently certain for the commercial valuation of tin
ores. Method 5 (Hallet's method) is logically admissible provided
that the precipitated metastannic acid is not merely ignited and weighed
as stannic oxide, but is redissolved and the tin carefully separated.
This, however, complicates the method somewhat. Method 6 is the
most practical method of assaying tin ore, which ensures at the same
time accuracy and reliability as a method, and it alone will be con-
sidered here in detail. Method 7 (the Pearce-Low method), which
consists of fusion with caustic soda in an iron crucible, solution in HC1,
reduction with iron, and titration with iodine, is a workable and an
exceedingly rapid method for isolated assays; one assay can be done
in an hour easily (see Low, " Technical Methods of Ore Analysis,"
2nd Edition). The above italics are mine, and I consider it necessary
to emphasise this point, as Mr. Low claims that the method is at
once the quickest and most accurate assay of tin in tin ores. It is
undeniably quicker for isolated assays, though it is not as economical
of time as the reduction method, since one can do seven or eight re-
duction assays in conjunction with other work in the course of a
working day, which is scarcely possible with the above fusion method.
With regard to accuracy, the method compares favourably with most
other assays, but is scarcely as accurate as the reduction assay.

The foregoing remarks have reference to the assay of dressed
tin ores in the main ; in previous editions it was omitted to clearly
point this out. It is for dressed tin ores that fusion assays in general
are so objectionable, owing to the possibilities of slight losses by
spurting, creeping, volatilisation, &c., which on rich material intro-
duce errors of commercial importance ; on poor material this objec-
tion does not exist, and the writer must certainly admit that for poor
mine stuff up to 5 or 6 per cent, of tin the Pearce-Low method is cer-
tainly the method, and as slight losses in fusing do not here appre-
ciably affect the result, a batch of such assays may be worked off
very rapidly. On the other hand, the writer must still maintain his
opinion that for dressed tin ores, the reduction assay, with fusion of
the siliceous residue, as herewith described, is the only admissible
method from a commercial standpoint, as .it favours neither buyer
nor seller.

DESCRIPTION OF METHOD.

Tin ores may be either pyritic or non-pyritic. Pyritic ores must
be, and non-pyritic ores may be, treated with NHO3 before reduc-
tion, as sulphide of tin is volatile at a red heat. 5 grms. ore, ground
as finely as possible in an agate mortar, is treated in an evaporating

basin with clock glass cover, with 20 c.c. dilute HNO3 and carefully
evaporated to complete dryness. The residue is digested with dilute
HNO3 and filtered. The washed residue is dried, ignited in the
dish, transferred to a porcelain boat and heated to a low red heat for
2 hours in a current of hydrogen or coal gas. Coal gas is much
more convenient to use and quite as effective as hydrogen. The boat
is i\ inches by \ inch, and two at a time are placed in a porcelain or
glass tube 12 inches long and f inch bore, which is then placed in
the reduction furnace. A very convenient form of gas reduction
furnace with clay body, brass gas jets, and asbestos rings to fit over
the ends of the tube against the clay covers, is made in 6-inch lengths
by Messrs. Fletcher, Russell, & Co., of Warrington. The ends of
the tube should project about 3 inches from each end of the furnace
and should be closed with rubber corks fitted with glass tubes as
shown ; the escaping gas (about 2 bubbles per second) is passed
through dilute HC1.

The water 'through which the gas escapes should not be thrown
away, but saved, and every now and then the tube should be washed
out with HC1 and KC1O3, and the two solutions tested for tin as a
check, against, for instance, loss by volatilisation through sulphur in
the coal gas. A two-way gas branch is used; one jet supplies gas to the
tube, the other supplies the gas for heating to the jets of the furnace.
The boats are allowed to cool in the furnace, and when cold each boat
and its contents is transferred to a 400 c.c. beaker and treated with
100 c.c. HC1 and 5 c.c. HNO3, the assay being allowed to stand in a
warm place until the action abates, when it is boiled for a few
minutes, diluted with an equal bulk of water and filtered. The residue
is well washed with hot acid water, then with hot water, is dried,
ignited, and fused with four or five parts of a mixture of fusion mix-
ture or sodium carbonate (free from chlorine) and borax, in a platinum
crucible, and the melt dissolved in HC1 and precipitated with zinc
as usual. The residue rarely holds more than \ per cent, of the
total tin in the ore.

Instead of fusing the residue with KNaCO3 and borax in a
platinum crucible, it may be fused with about 2 in. of stick
caustic soda in an iron crucible over a bunsen burner. The melt is
dissolved up in HC1 and precipitated on zinc as usual. (See K i,
paragraph 2.)

The main solution is made up to 500 c.c. and the equivalent
of i grm. is pipetted into an 8-oz. flask, reduced with iron wire and

c 2

filtered ; the filtrate is precipitated with strips of sheet zinc as usual,
and the metallic sponge dissolved in HC1 and titrated with ferric
chloride.

Except with impure ores the reduction with iron wire may be
omitted, as it is always an advantage to save a filtration when
possible. In this case the aliquot part of the solution is at once
precipitated with zinc. The boiling HC1 solution for titration with
FeCl3 must, however, be free from black powder of Cu, Sb, &c., and
from tungsten blue in suspension or solution. (N.B. — Lower oxides
of tungsten in solution give a brownish pink or light claret -coloured
solution.) During precipitation with zinc the solution should be
freely acid to start with, and in presence of Ti, or Mo, should be
distinctly acid when poured off. (See Section B, p. 27.)

Any copper in the tin ore is found in the HNO3 solution, though
traces may remain with the SnO2. In the iron wire stage, the
arsenic which escaped extraction with HNO3 and volatilisation in
the reduction tube, is partly evolved as AsH3 and partly precipitated
with antimony in the metallic form. It comes down as a brown
flocculent deposit which contains 3 or 4 per cent, of its weight of tin.
As there is generally only a few per cent., at most, of As in a tin
ore, the loss of tin from this cause is quite negligible, but as a check
one should save the iron wire and precipitates and filter papers and
examine them from time to time for tin ; it will be found, as in the
case of the deposit in the tube and the dilute HC1 through which
the escaping gas bubbles, that only the merest traces of tin are lost
in these operations. Further, the HC1 solution of the reduced metal
may be done in a conical flask with rubber cork and leading tube
dipping under water, to assure oneself that there is no appreciable
lose by volatilisation of SnCl4.

If the ore contains wolfram, the tungsten is mostly found as WO3
in the residue from HC1 and HNO3 extraction of the reduced metal,
from which it may be removed before fusion, with AmOH. Any
tungsten which gets into the main solution comes down as blue oxide
with the iron wire precipitate, and any which is fused with KNaCO3
and borax should be removed by reducing the HC1 extract of the
melt with iron wire, before precipitating with zinc. In general, all
the antimony and some of the lead in the ore will be found in the
main HC1 solution, whilst some of the lead will be obtained in the
HNO3 extract.

Note. — After the HNO3 evaporation the residue may be boiled
with HC1 (40 or 50 c.c.), diluted and filtered, though in this case the
extract must be tried for tin as a matter of precaution. It will in
general hold all the copper and most of the arsenic, antimony,
lead, and iron, though one can never be sure that the residue is
free from the oxides of these metals. Occasionally this HC1
extract will hold a little tin. The residue is reduced in the usual
manner.

K 1.— THE ASSAY OF POOR SILICEOUS ORES, BATTERY TAILINGS,
AND SILICEOUS RESIDUES HOLDING LESS THAN 2 PER CENT.
OF TIN.

Siliceous material very poor in tin is best assayed by a fusion
method of which either of the following will serve : —

i. — .5 — i grm. of finely-powdered ore is fused in a platinum crucible
with soda ash (free from chlorine) or NaHCO3, and borax.
The melt is dissolved in dilute HCl, transferred to an 8-oz.
flask, and heated with ten 2|-in. iron nails until the colour
goes ; it is then allowed to stand in a moderately cool place
for 20 minutes to half-an-hour, poured off into a clean flask
containing CO2, and titrated with one-tenth strength iodine.

la.— Or, the HCl solution of the melt is precipitated with zinc, for
the usual ferric chloride titration.

2. — Mr. A. H. Low's method is a very useful one. .5 grms. fine
ore is fused over a bunsen burner in an iron crucible with
3 or 4 inches of stick caustic soda. . The melt is extracted
with water and solution effected with HCl. The solution is
reduced with iron nails as above, and titrated with iodine
as above.

A blank assay with iron nails and dilute HCl must be
done, as it has been found in practice that when using
ten or a dozen 2|-in. iron nails and digesting these with
HCl as above, the liquid requires .1 to .2 c.c. iodine, even
in absence of tin.

The solution of the melt may also be precipitated with
zinc and the metallic residue dissolved in HCl and titrated
with FeCl3.

K 2. — REVISED ASSAY OF DRESSED TIN ORES.

In order to avoid the preliminary evaporation of pyritic black
tin with nitric acid, the following procedure may be employed :
The ore is weighed direct into the porcelain boat and reduced as
usual. After reduction the deposit in the tube is washed out with
HCl and KClO3 and mixed with the water (which should have been
acidified with HCl before starting the assay), through which the gas
bubbled, and the joint solution reduced with iron or zinc and assayed
for tin by either the iodine or ferric chloride assay ; the amount
so found being calculated on to the percentage determined from the
solution of the reduced ore.

This procedure applies to all dressed tin ores holding not more
than 2 per cent, sulphur ; that is to say, it applies to ninety-nine out
of every hundred buying samples of black tin or tin barilla. In such
exceptional samples as may hold more sulphur than this, evaporation
with nitric acid must precede reduction.

As we have already mentioned in these notes, the reduction of the
main assay solution with iron wire may in most cases be omitted, and

it will be found that in this amended procedure we attain with the
gas reduction and ferric chloride assay the following conditions :—

1. The minimum of manipulation of the main portion of the

assay both before and after solution.

2. Complete solution of the ore.

3. The use of a method of determination by which only tin can

score as tin.

The first two of these conditions which ensure the elimination of all
sources of loss, protect the seller's interests; while the third, which is
at the same time absolutely fair to the seller, protects the buyer.

L. — ASSAY OF TIN SLAGS.

i. — HYDROFLUORIC ASSAY. — The slag is pounded up finely in a
steel mortar ; the more siliceous the slag the finer it should
be powdered. 2 grms. slag is weighed into a platinum
basin and about 20 c.c. HF (pure and strong) added in a
fume cupboard and the assay evaporated to bare dryness.
Most of the SiO2 is removed and the tungsten remains as
blue oxide, other metallic oxides being left in a form easily
soluble in acids. The residue is treated in the basin with
30 or 40 c.c. HC1, and after warming, the contents of the
basin are transferred to a 400 c.c. beaker, the basin being
cleaned out with a rubbered glass rod and as small a
quantity as possible of hot water. Add / c.c. of HNO3 and
boil for 2 or 3 minutes, dilute to about 350 c.c. and pass a
rapid stream of H3S for about 10 minutes. The SnS2 should
come down granular if these instructions are adhered to.
Allow to stand for a few minutes and filter, but do not wash.
(If the SnS2 comes down in a very finely divided form, the
assay must be allowed to stand for two hours and again
gassed.) Tin should never be separated as SnS. Wash
back the precipitate into the beaker and dissolve by boiling
with 30 c.c. HC1 and a little KC1O3. Boil with iron wire,
filter and precipitate on zinc as usual. Dissolve in HC1,
adding a granule of pure zinc and bringing rapidly to a boil,
and titrate with ferric chloride.

2. — HYDROFLUORIC ACID AND REDUCTION. — In the case of certain
very rich slags the tin may be present either wholly or in
part in a form insoluble in HC1 after removal of SiO3
with HF. In such cases dilute the HC1 extract to 75 c.c.
and filter; reduce the residue in coal gas, dissolve, the re-
duced metal in HC1 and a little HNO3and filter the solution
into the first HC1 extract. If there is still a residue it may
be fused with KNaCO3 and borax. The main solution
(if from 5 grms. slag) is diluted to 250 or 500 and i grm.
pipetted out and precipitated with H2S. Filter, dissolve in
HC1 and KC1O3, reduce on iron and precipitate with zinc
as usual. If the insoluble residue is small in amount it

may be fused at once with KNaCO3 and borax, after removal
of WO3 with dilute ammonia. In" the case of leady slags
the HF residue may be treated with HNO3 according to
general method F. In the case of slags containing prills of
metal the metallics must be sieved out, from a portion of
loogrms. Weigh the fine and metallic in proportion.

3. — FUSION ASSAY. — Melt 5 — -10 grms. NaOH in a nickel (or,
better, iron) basin and dust in 2 grms. of the finely powdered
slag. Cover and heat to bright redness for 10 minutes with
the aid of a foot blowpipe. Cool, loosen the melt, and
transfer to a beaker; dissolve in HC1 and a little HNO3,
precipitate with H2S and proceed as before.

M. — IODINE ASSAYS.

The general method consists in —

(a.) Obtaining the tin in dilute HC1 solution as SnCl4, bulk

50 — 60 c.c. in a small beaker.
(b.) Reduction of this solution to SnCL, by a piece of soft iron

with precipitation of As, Sb, Cu, as metals.
(c.) Titration with iodine and starch indicator in the cold.

The method is reliable, quick, neat, and cheap, requires little
apparatus or room, but has not the certainty of the ferric chloride
assay. The preceding general methods and separations remain un-
altered in most cases, as far as obtaining the tin in HC1 solution
as SnCl4.

i. — In the case of ALLOYS and MATTES, &c., .5 — i grm. filings or
powder is weighed into a 4-in. beaker with watch glass
cover and covered with about an inch of HC1 ; then boil
gently until action ceases, on a frying pan sandbath over a
bunsen burner. Add a crystal of two of KC1O3 to complete
solution, boil off the chlorine, add an equal bulk of water
and a piece of soft iron (3 in. by 3-16 in.), cover with
watch glass and allow to simmer gently for 20 — 30 minutes
after the solution becomes colourless (which is an indication
that all the iron in solution is present as FeCl2). At this
point we assume (a weak point of the assay) that all the tin
is reduced to SnCl3. The Sb, Cu, and As are, of course,
precipitated by the iron. Cool the assay quickly in a basin
of cold water, wash the rod and cover with a rapid stream
of cold boiled acid water saturated with CO2, add a little
starch paste and titrate at once with iodine, made by dis-
solving 21.32 grms. pure iodine, and 45 grms. KI free from
iodate, in a little cold water, and diluting the solution to
i litre .100 c.c. = i grm. Sn. In standardising take .5
grms. tin, and dissolve in i in. of HC1 in a small beaker.
Dilute, insert a piece of iron, reduce gently for five minutes,
cool, wash cover and rod rapidly and titrate at once. The

4°

oxidation tendency is the important source of error in these
assays. Precipitated Sb and Cu are not dissolved by
iodine (except slowly) in a cold dilute solution, and therefore
their presence as precipitated metal does not matter except
that they darken the liquid, if in any quantity, and obscure
the starch blue until the assay has stood for a moment.
The iodine is more safely standardised by comparison with
KMnO4 of known strength, or against pure As3O3.

2. — ASHES AND ORES. — The measured portion of the HC1 solution
should not be too bulky ; it is reduced in a beaker by an
iron rod as in the process just described.

3. — SLAGS, (a.) HF ASSAY. — Tungsten blue should be filtered
off. The HC1 solution of the HF residue may be reduced
by iron, filtered, and again reduced.

(b.) FUSION ASSAY. — The HC1 solution is reduced and filtered
from the tungsten blue, then reduced again and titrated.
H2S separation seems to be unnecessary with iodine assays.

Chapter VI.

The Assay of Antimony.

SUMMARY OF METHODS OF DETERMINATION.

i. — Gravimetric estimation as SbO2 (unreliable).

2. — Gravimetric estimation by electrolysis.

3. — Titration of Sb2O3 with iodine in alkaline solution.

4. — Titration of iodine liberated by action of SbCl5 on KI in
HC1 solution.

5. — Solution of precipitated Sb in H0SO4 and titration with
KMn04.

6. — Titration of precipitated Sb with FeCl3 in a boiling HC1
solution.

7. — Titration of precipitated Sb with bichromate after solution
in FeCl3.

8. — Dry assays (a) iron reduction (b) carbon reduction.

The electrolytic assay is useful in special cases, as are 5, 6, 7.
The dry assays are also useful in their proper place. The ordinary
gravimetric estimation is a bad method. Methods 3 and 4 are of
the most general applicability, and are both accurate and practical.

SOLUTION from speises, mattes, metals, is easily effected by HC1
and KC1O3 or HNO3 ; from oxides by reduction to metal with
KCN or coal gas and solution of the reduced metal in HC1 and
KC1O3, by fusion in a silver basin with caustic alkali and solution

in HC1, by extraction with alkaline sulphides (with or without a
sulphurising fusion), and sometimes with the help of tartaric acid
which dissolves the oxides and oxychloride; also by oxidising metal
or sulphide with HNO3 and solution of the oxide in alkalies or
alkaline sulphide.

SEPARATION. — When necessary, HoS in a not too concentrated
HC1 solution separates from Fe, Zn, Co, and Ni. Extraction with
Na2S dissolves sulphides of As, Sb, Sn. Sulphides of As, Sb, Sn,
are'dissolved in HC1 and KC1O3 and HT, AmOH, and AmCl added ;,
then magnesia mixture, when the arsenic is separated as Mg, NH4r
AsO4. Antimony may be readily estimated in presence of tin.
Arsenic may also be distilled off from a metal, with a solution of
FeCl3, and CaCL, in HC1. As2S3 may be precipitated in a concen-
trated HC1 solution, Sb remains in solution. Sb may be precipitated
in the metallic form by iron.

Chapter VII.

Specific Antimony Assays.

A. — Alloys of Tin — Lead — Antimony — Copper with but little
Arsenic.

TYPE METAL — SOLDER — ANTI-FRICTION METALS.

SbCl5 in HC1 solution oxidises KI. The liberated iodine may be
titrated with stannous chloride — not with Hypo.

In this method the presence of tin (as SnCl4) and lead has no-
effect whatever on the titration ; indeed, the titration is done with
SnCl2. Copper, arsenic, and iron, when present in solution in their
highest state of oxidation, score as antimony, and should be separately
estimated and allowed for, this being preferable to separating the
Sb from them. Fe is rarely present, but in the case of alloys with
more than i per cent. As, it is advisable to dissolve in aqua regia
and separate the Sb and As as sulphides, as it appears that in some
cases when an alloy rich in arsenic and tin is boiled with HC1 there
is a considerable evolution of H3As. This matter is under investiga-
tion. Arsenic, however, is rarely present in the above alloys in
quantities greater than i per cent, i — 3 grms. of fine filings is
weighed into a i6-oz. flask and boiled gently with 150 c.c. HC1 until
action ceases and most of the Pb and Sn are dissolved ; to complete
solution add cautiously a cold saturated solution of KC1O3, which
also peroxidises the metals. After solution is complete, add a little
KC1O3 and make sure that excess of chlorine is present. The solu-
tion is diluted with rather less than an equal bulk of water and boiled
until free from chlorine. In practice half-an-hour is a sufficient
length of time ; the most practical and most delicate test of the

absence of free chlorine in the cold solutions is the sense of smell. The
assays should not be boiled below 150 c.c. as a matter of precaution,
on account of the volatility of chloride of antimony. Allow the assay
to cool; when quite cold, fill the flask with CO2 from a Kipp charged
with marble and HC1, add 20 c.c. of a fresh 20 per cent, solution
of KI, and titrate as rapidly as possible with stannous chloride. This
is made by dissolving 10 grms. of tin, or 20 grms. SnCl2, 2 aq., in
300 c.c. HC1, and diluting to i litre. It should be kept under CO2,
and must be standardised every time it is used, against a standard solu-
tion of bichromate made by dissolving e xact ly 1 6 grms. of pure K.->CraO7
in i litre of distilled water. 100 c.c. Bic. = 2 grms. Sb, and looc.c.
SnClo = about i grm. Sb. The assays, if overdone, are brought
back with the standard bichromate (rapidly), but should only require
a few drops at most, otherwise the accuracy of the assay is impaired.
The assays have a great tendency to absorb oxygen, after titration,
and reoxidise ; hence the necessity for a CO2 atmosphere and rapid
titration, especially in presence of As and Cu.~ Further, the reaction
in the case of As is not instantaneous, and only proceeds to comple-
tion as fast as the liberated iodine is reduced by SnCL, — being to some
extent reversible. Acid solutions of AsCl3, CugCl2, FeClg, and KI,
readily absorb oxygen from the air ; SbCl3 is scarcely affected. The
bichromate remains quite constant, and it may be standardised against
pure iron as a check on the weighing up. In standardising the
stannous chloride, pour a little KI solution and starch paste into an
8-oz. flask, fill the flask with CO2 and run in rapidly, 20 c.c. SnCL
from the burette. Then titrate rapidly with K2Cr2O7.

As = Sb = Cu2 = Fe2 = O.

i per cent. As= 1.6 per cent. Sb. i per cent. Cu = o_945 per cent. Sb.
i per cent. Fe = 1.03 per cent. Sb.

HC1 solutions of Sb (ic) As (ic) Cu (ic) and Fe (ic) oxides are
all reduced by KI to the corresponding — ous compounds, with libera-
tion of an equivalent amount of iodine. These ( — ous) solutions, as
well as a solution of KI in HC1 and of SnCl2, absorb oxygen from
the air, and therefore necessitate a CO2 atmosphere. The solution
requires to be more strongly acid in the case of arsenic. SnCl4 is
not reduced by KI, hence the liberated iodine may be titrated with
SnCLj. In assaying type metal for Sb the assays generally have a
slight yellow colour after boiling off all the chlorine. This may be
due to either Cu or traces of Fe, but is not due to SbCl5. Rich type
metals sometimes contain a little Fe, and a little iron may get in from
the file, but it has been proved that any error arising from the use of
a file is quite negligible. Alloys of the above class which are rich
in As appear, in some cases at any rate, to lose much or most of their
As or AsH3, during the action of HC1. The matter is being in-
vestigated, but the possibility of such loss of As may be guarded
against by adding solution of KC1O3 before the HC1, and by keeping
the solution saturated with chlorine until solution is complete.

The arsenic in such alloys may be always estimated by distillation
with a solution of FeCl3 and CaCl2 in HC1 (see Beringer, " Text-
book of Assaying "). According to Mr. A. Gibb, after removing the
As as AsCl3 in this manner, the solution, if mixed with ZnCl2 and
redistilled yields all Sb (as SbCl3) at 184° C. Since SnCl4 boils at a
much lower temperature than either AcCl3 or SbCl3, it seems strange
that SnCl4 is not distilled off in this assay, but the probability is
that it forms a thick double salt.

The above method is both rapid and accurate, but it requires con-
siderable judgment and experience in performing the titrations. The
titration must be done as rapidly as possible, but not too rapidly ; it
must be borne in mind that the finish is more or less gradual, and
especially so in presence of arsenic. Sometimes in running in a rapid
stream of SnCl2 the blue colour vanishes, but reappears like a flash
the moment after. In such cases, one should go on titrating, as the
assay is still unfinished. If the flask is not properly filled with CO2,
after-bluing will take place in any case, and it requires long experience
to distinguish in every case between a finished and unfinished titration.
Duplicate assays should always be done ; the bichromate solution
itself should be standardised against pure iron wire, which contains
99.6 per cent. Fe. Take the at. wt. of Fe as 56 and that of Sb 120
in calculating the strength of the bichromate.

B. — Alloys of antimony and iron with arsenic and copper.
(a.) The As is estimated by the distillation assay.
(b.) The Cu by the usual means.

i. — i grm. metal. Dissolve in 20 — 30 c.c. HC1 and i or 2 c.c.
HNO3 in an 8-oz. flask. Add sodium sulphide till alkaline
and pass H2S. Warm, dilute, and filter. Redissolve the
Cu2S, PbS, FeS, &c., and reprecipitate. Filter, mix the
filtrates, and precipitate the sulphides of As, Sb, Sn by
cautious addition of HC1. Filter, dissolve the precipitate in
100 c.c. HC1 and KC1O3. Boil until solution is complete,
dilute, filter, add KC1O3, and boil off the excess of chlorine.
Cool, add KI and titrate with SnCl2, making the equivalent
deduction for As.

2. — i grm. metal, 40 — 50 c.c. HC1 and a little HNO3. Dilute to
1 20 when solution is complete, and pass H2S, thus separating
from Fe but not from Cu. Dissolve the precipitate in HC1
and KC1O3, and proceed as in i.

C. — Mattes and sulphurous compounds generally, of Fe, Cu, Pb,
As, Sb, Sn. Antimony ore, &c. (if impure).

In these the As cannot be estimated by the direct distillation
assay, and must be separated. Solution is effected as usual. The
sulphides of As, Sb, Sn, obtained by precipitation of an alkaline
polysulphide solution with HC1, are filtered and dissolved in 20 c.c.
HC1 and a little KC1O3 and the Cl boiled off. 5 grms. AmCl and

5 grms. HT dissolved in 20 c.c. water are added, and then AmOH
until the solution is alkaline. (If a precipitate forms on adding
AmOH, more AmCl and HT are needed.) The solution, which
should not be more than 60 or 70 c.c. in bulk, is mixed with 20 c.c.
of magnesia mixture and allowed to stand for 24 hours. The Mg,
NH4, AsO4, is filtered off and the filtrate acidified with HC1, diluted
to i litre and precipitated with H2S. Allow to stand in a warm
place for an hour or two. Again pass H3S and filter the sulphides
of Sn and Sb. Dissolve the precipitate in HC1 and KClo3 as usual.
Add KI and titrate with SnCl2.

D. — ANTIMONY ORE (with "but little As). — i grm. ore, 30 c.c.
HC1, 2 c.c. HNO3. Add NaOH till alkaline, as soon as
the ore is completely attacked, and pass H2S. Warm, filter,
and wash with hot dilute Na2S. The precipitate if bulky
should be redissolved and reprecipitated. Precipitate the
filtrate with HC1, filter, wash with NaCl, boil the precipitate
well with HC1 alone (As2S3 left undissolved), dilute, filter,
peroxidise with KC1O3, boil off the free chlorine, cool, add
KI and titrate with SnCl2, or boil off the iodine, cool, add
Rochelle salt, neutralise, make alkaline with bicarbonate
and titrate with Iodine.

E- — OXIDES. Calcined antimonial material. Boil with HC1 and
KC1O3, dilute and filter, (a] Dry, ignite, and fuse the residue
with NaOH in a silver dish. Extract with HC1 and mix
with the main solution ; (b) or reduce the insoluble oxide
to metal in coal gas at a very low red heat, or in a small
porcelain crucible with KCy, and dissolve the reduced
metal in HC1 and KC1O3, and mix with the main solution.
Proceed as in C if arsenic is present or as in B2 if no arsenic
is present.

F. — The electrolytic assay has been already referred to. (See
Chapter IV.)

G. — ROUGH VOLUMETRIC ASSAY BY TITRATING PRECIPITATED
ANTIMONY WITH FECLS.

This method is chiefly useful as a rapid approximate determina-
tion of antimony in solder, and is used as an adjunct of the direct
ferric chloride assay of tin in such alloys. It depends upon the
fact that FeCl3 dissolves finely divided Sb with the formation of
FeCl2 and SbCl3. After pouring off the hot acid solution of SnCl2
from "the black powder (see Chapter V., A.), pour on to the latter,
immediately, a little hot boiled water to prevent aerial oxidisation of
the Sb. When the tin titration is finished, pour back on to the black
powder and retitrate. (Cu if present is also converted into Cu2Cl2 —
as soon as any CuCl2 is formed it gives the colour indication.) The

percentage of Sb is two-thirds of the apparent tin percentage, equiva-
lent to the extra ferric chloride used — in absence of copper.

H. — VOLUMETRIC ASSAY BY DISSOLVING PRECIPITATED SB IN H2SO4

AND TITRATING WITH KMNO4.

Useful for type metal and solder. Cu does not interfere, but As
should be absent. The method is used as an adjunct of the iodine
tin assay. The antimony adhering to the iron rod is washed off,
allowed to settle, and the liquid decanted into another beaker ; the
residue washed by decantation with hot dilute HC1 and the washings
added to the rest of the tin solution, which is again reduced before
titration. The precipitated Sb and Cu are heated with a few c.c. of
strong H2SO4 until solution is complete and white fumes evolved.
Then cooled, diluted, and titrated with KMnO4.
Fe2 = Sb.

K. — Titration of Sb2O3 in alkaline bicarbonate solution with iodine.
(Mohr's method.) This is a useful assay in the absence of
Cu, and is fully described in Sutton's " Volumetric Analysis."

L. — SOLUTION OF PRECIPITATED ANTIMONY IN FfiCL3 AND TITRATION
WITH K2CR3O7 OR KMN<D4.

This method may be useful as an adjunct of the iodine tin assay
as in H, and is useful for type metal and solder in the absence of
copper. The precipitated Sb is dissolved in a small basin in a few
c.c. of FeCl3 and HC1 (free from FeCl2), and the solution titrated with
bichromate until all the FeCl2 is peroxidised, using spots of ferricyanide
indicator on a plate.

3 FeCl3 + Sb = SbCls + 3 FeCl2
3 FeClo + 3 HC1 + Oil = 3 FeCl
SbClg + 2 HC1 + O = SbCl5 +

o 3 i = 3 e3 i

g + 2 HC1 + O = SbCl5 + H20

l3 + ii H2O.

That is to say, when dissolving antimony in FeClg and titrating with
K2Cr2O7, Sb requires altogether 2% O, or is equivalent to 5 Fe in
terms of the strength of the bichromate solution when standardised by
titrating FeO to Fe2O3.

Although FeCl3 does not oxidise SbCl3, yet the SbCl3 is all
oxidised by the Bic. before the FeCl2 ; before the disappearance of
the FeCl2 is shown by the spot indications which is taken as the finish
of the titration. The development of the colour is, however, slow
towards the finish. SbCl3 will reduce the brown solution obtained by
mixing FeCl3 and K3FeCy6 with formation of Prussian blue. Indeed,
SbCl3 may be thus titrated with Bic., using ferric ferricyanide solution
as an outside indicator, or FeCl2 as an inside indicator and K3FeCy6
as outside indicator. When KMnO4 is used, the SbCl3 is also
oxidised in the titration as well as the FeCl2. The above method is not
to be recommended except for rough purposes, as it is hard to wash
pptd Sb (by iron) free from FeCl2 without redissolving some Sb by
aerial oxidation.

M. — DRY ASSAY OF ANTIMONY.

i. — Antimonial litharge.

(a) 50 litharge roughly pounded,

10 — 15 soda ash, 10 — 15 charcoal,

5 borax, and a little fluor.
or (b) 50 litharge,

50 — 100 cyanide,
10 argol.

In either case fuse in a clay crucible at a moderate heat, and
assay the button for Sb.

2. — Sulphide Ore. — This if mixed with much gangue should be
concentrated by liquating1 1,000 grams in a double luted
crucible.
Fuse 25 grms. ore, 5 soda,

12 grms. iron filings, 5 borax,

in an E crucible, and assay the button, which should con-
tain Fe, for antimony.

3. — Type Ashes — usually carbonaceous. — Fuse with 15 to 20 per
cent, soda ash and a little ferrocyanide and borax. Melt the
button away from iron shots in a ladle, and pour into a hemi-
spherical mould. Assay the metal for tin and antimony.

N.— ASSAY OF ANTIMONY IN PRESENCE OF ARSENIC AND TIN (with-
out separation).

We have seen that Sb may be estimated with SnCl2 in presence
of SnCl4 and As2O5 by titrating the iodine liberated from KI in
HC1 solution and deducting the equivalent of the As3O5 present.
It is, however, possible to estimate Sb by direct titration with SnCl3
in a fairly strong and hot, but not too strong HC1 solution ; in such
a solution arsenic acid is not affected at all until all the SbCl5, FeCl3,
and CuCL, have been reduced, and therefore if we titrate until the
colour goes we have the measure of the Sb + Fe + Cu. This method
may be applied to type metals, which contain only traces of Fe, and
usually under i per cent. Cu. The Cu must be known and deducted
for, but it serves the purpose of an internal colour indicator. The
method is beautifully simple and rapid, as the Cu may be estimated
with sufficient accuracy in type metals and most anti-friction metals in
a few minutes. In the case of solders, which are, as a rule, free from
Cu or Fe, a drop of FeCl3 is added to the assay as an indicator. With
regard to the accuracy of the method, there can be no doubt that it
is sufficiently accurate for most commercial purposes.

N1. RAPID DETERMINATION OF ARSENIC, ANTIMONY, AND TIN

in alloys, or when separated together as sulphides.

i. — The tin is determined in the usual way; solution in HC1 and
KC1O3, separation of Sb and As on iron wire, filtration,
precipitation with zinc, solution in HC1, and titration with
ferric chloride. Add 5 per cent, of the arsenic percentage.

2. — The antimony is determined by solution in HC1 and KC1O3,
boiling off the excess of chlorine, and titration of the hot
solution with SnCl2. In absence of Cu (the amount of
which must be known and allowed for), add a drop of
FeCl3 as an indicator.

3. — The arsenic is determined by (a) distillation with FeCl3 and
CaCl2 in HC1, and titration of the evolved arsenic with
iodine in alkaline bicarbonate solution ; (b) sulphides are
dissolved in HC1 and KC1O3, the solution concentrated and
distilled with a ferric chloride mixture containing also ferrous
chloride, the evolved arsenic being estimated as before by
titration with iodine in an alkaline bicarbonate solution ;
(c) or the solution of sulphides in HC1 and KC1O3, from
which the free chlorine has been driven off, is boiled with
copper foil, which, with the precipitated Sb and As is after-
wards distilled with ferric chloride mixture. The deter-
mination of the three metals in an alloy should not take
more than 2 hours by these methods.

30. — Or, better, the fine filings are dissolved by slow digestion with
HC1 and KC1O3 in a warm place, keeping always present a
slight excess of free chlorine. When solution is complete,
add a little more chlorate, dilute, boil off the chlorine, cool,
add KI and titrate with SnCl0 = As + Sb.

APPENDIX.

AT the suggestion of the Cornish Correspondent of "The Mining
Journal," and with the permission of the Editor of that paper, we
subjoin a table showin'g a series of analyses of Cornish tin ores.
Part of the article which accompanied the table in question, and
which appeared in "The Mining Journal" for August 19, 1905, is
also given herewith. The analyses were done by the writer of this
book : —

These analyses were carried out by Mr. L. Parry, A.R.S.M.,
Assayer and Consulting Metallurgist, of Union Bank Yard,
Huddersiield.

The wet assay adopted for tin is the one given in Mr. Parry's
book, " The Assay of Tin and Antimony " — viz., reduction in a
current of coal gas combined with a volumetric estimate of the tin.

In the cyanide assays the ores were extracted with aqua regia,
and the solutions were tested for tin. In one or two cases a little
tin (under .25 per cent.), was found in these solutions, and is
included in the results given. The cleaned residues were reduced
with cyanide (98 per cent. Au*), in clay crucibles. The assays were
done on 20 gram charges, and the buttons of metal obtained were
assayed for tin by the same volumetric method employed in the
wet assays. The net cyanide percentages thus obtained average
ij per cent, under the wet assays.

It is possible, by adopting a different procedure, to arrive at
results by the cyanide assay that will more closely approximate to
the wet assay than those given above. The ores, either cleaned
with aqua regia or not, are reduced with cyanide in porcelain crucibles
in the muffle, 5-gram portions being taken for assay. When the
fusion is complete, the assays are cooled, and the crucibles and con-
tents are boiled in water until the cyanide and cyanate of potassium
are dissolved out. The residue is then extracted with HC1, and the
tin in the solution then determined volumetrically. Although rather
better results are obtained, the assay is more tedious, and there is
little saving of time over the wet assay, whilst there always remains
the possibility of tin being lost as alkaline stannate. In cyanide

* "98 per cent. Au " is the trade mark of the purest cyanide made, which is
used for electro-plating, hence the "98 per cent, gold." This cyanide holds
98 per cent, of pure KCN.

assays the cyanide should be finely powdered and well mixed with
the ore, otherwise some of the ore is liable to escape reduction.

The wet assays were done throughout in duplicate. The arsenic
and antimony were estimated in the aqua regia extract, and the
results thus obtained are likely to be somewhat under the actual
contents. The iron, silica, and tungstic acid were merely determined
approximately. The crop ores having already been roasted, were
not assayed for sulphur, but the unroasted slimes, Nos. 2, 5, 7, 8,
and 1 1, were all rather strongly attacked by HNO3,and were, therefore,
assayed for sulphur.

It has been necessary to detail the tests made for the simple reason
that the majority of our mine managers, and others who control the
industry — apart from the smelters — will now, for the first time, secure
reliable data as to the actual constituents of the black tin they have
so long handled, and sold on an assay (Cornish method), which, to
put it mildly, affords only sufficient indication of values to keep the
buyer from blundering. In this connexion, the necessity for assaying
the buttons obtained, by the cyanide method, for tin — that is, for
purity of metal — was amply demonstrated. Notwithstanding the clean-
ing with aqua regia, only seven out of nineteen buttons — viz., Nos. i,
3, 4, 14, 15, 1 8, 19 — were pure tin. The others varied from g6|
to 99 per cent. Sn, with the exception of Nos. 9, 10, and n, which
came out 90, g2j, and 93! per cent. Sn, respectively. The balance
of the buttons was iron, showing that it is not possible to extract all
the FE2O3 from all ores direct by extraction with aqua regia. The
percentage weight of the buttons was greater than the actual tin
percentages shown by wet assay in the cases of Nos. 6, 8, 10, u, and
12, which held g6J, 90, 92.5, 93.6, and 97.2 per cent. Sn respec-
tively. In Cornwall the purity of the metal produced does not much
trouble the seller of black tin. Later on the importance of paying some
attention to this matter may be deemed worthy of the consideration
of every mine manager in the county, and more especially of those
whose products, as indicated by the above table, call for special
treatment.

Remarks.

Trace of Cobalt.
Trace of Cobalt and Zinc.
Unroasted Slimes.

Trace of Cobalt. Unroasted
Slimes.
A little MnO.
Slimes Unroasted.
Slight trace of Cobalt. Un-

,2 d
D« HHH<>h

Alumina.

'PI *8T 'IT '01 *8

eo os t^ r-t co eo o t-

^H us t^ §5 ^ co ^ us us en co

*L '9 'S ;o raus IBJOJ,

0505 05 0505 050505

feSo^o? SSSSS^o?

« rt d

'S P %

1 "o W

•*

• •«

^S ^«

uiojj pajBjno[B3 '8o^3J

CNO C^C-Jl>C OO 1C US

5§S23 3235^

G ^O 9^

•-* cfl

•ti

•%3J

«ff S3?. CO^C,

«W« «*!?«'-W

*ls

'% ^O'S

H« « *

H^H« HC.

•" C/3

•N

us "o

'% 80AV

oo ooo e^oo

000Tus o-T^-0

iff

C '

<?1

•%s

8 • 88 • 8^

88^88 888888

; ° |

•%qs

•*»*» -»*-w *» •*» •« J3

*»•«, ^^ . . .

• ^ B

% sy

1 1

•%HD

+1 .1* ,33

US 1C US

''8M. PUE apiUB^3

Sooo usooo cooes

(N'^CfSCO'^ -.<N05l2uSO

US "g

u-i

_? ,J .4 ,4 <N ^ (N <N
II III III

1 1 1 1 1 1 1 1 1 1 I

'i jjf

•us JOj suojjng jo
Xpssy J3M sapn[0
-ui— jau apiuBXif) uix

1C US 1C US 1C

t~ CO O IN O «>. O CO

•* 00 US CO O5 O <M 00 1C O S

o'oc<susic e4-He<;t>rcoo

t~ CO CO CO CO t^ t- CO O t^ t~

S*!

" C

•us jo ABSSV J3M

OCO MOW IMCO-*

05 t^ 00 00 00 05 05 00 00 O5 S

00 =. H

CO «S >B9* 0> i-l O US

us us ic

3ssV |»M uiX

i-H 0 CN ,-5 05 US 00 00

«>.<» t~t^us cousus

I-H IN -* t^ t- CO »M «* OO CO i-H

t- co co co co t- t^ co <a t~ t~

s r §

. . . . ...

S "*

Description.

sf Ilj ill

e| ^~r e»S

- = i r"

o' c Z

2 J... i.

i i *-

*°" 1 §

Q > O

~$%~ '' -3. :^? U

2l|g-| IfiSal

f T

Average
N.B.— f mea

„

INDEX.

PAGE

Alloys, Assay of for Tin 24 — 31

„ Assay of for Antimony ... ... ... ... 40 — 47

Ashes, Wet Assay of 31 — 32

„ Dry Assay of 16 — 17 and 24

Actions, Oxidising and Reducing 8—9

Anti-friction Metals 27 — 28 and 41 — 43

Cornish Tin Ores, Analysis of 48

Ferric Chloride Titration for Tin 20—28

Gravimetric Estimation of Tin ... ... ... 3 — 5 and 20

Iodine Titration for Tin 22—23 and 39

Metallic Precipitation 9

Mattes and Speises, Assay of 28 — 29 and 43

Purity of Chemicals ... ... ... ... ... ... u

Rapid Determination of Arsenic, Antimony, and Tin in

Alloys 46

Solder, Assay of ...4,20, 24—26, 41 — 44

Separation of Tin, Antimony, Arsenic n — 13

Solution of Metals 10

Sulphides, Oxides, Oxy chlorides 6 — 7

Summary of Tin Assays ... ... ... ... ... 13 — 16

Type Metal, Assay of 27— 31 and 41 — 43

Tin Ore, Assay of 4—5, 18—20 and 32 — 37

Tin Slags, Assay of 38 and 40

UNIVERSITY OF CALIFORNIA LIBRARY
BERKELEY

Return to desk from which borrowed.
This book is DUE on the last date stamped below.

JAN 4 1949

LD 21-100m-9,'47(A5702sl6)476

YL JJ004

UNIVERSITY OF CALIFORNIA LIBRARY

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 "The assay of tin and antimony"

See other formats