Full text of "Atoms And Molecules"

See other formats

CO >

OU 164544 >m

73 73

CQ CO

OSMANIA UNIVERSITY LIBRARY

Call No. *"- Accession No

37/9 ^> ,,

Author v^VC A -

Tide |^A^ V /Mr/ ccfc.AC-i,
This book should be returned on or before the date last marked below.

ATOMS AND MOLECULES

BLACKIE & SON LIMITED

50 Old Bailey, LONDON
i ; Stanhope Street, GLASGOW

BLACKIE & SON (INDIA) LIMITED
\Van\ick House, Fort Street, BOMBAY

BLACME SON (CANADA) LIMITED
1118 Bay Street, TORONIO

ATOMS
AND MOLECULES

Being Part I and Chapter XII of
The Foundations of Chemical 'Theory

R. M. CAVltt

D.Sc.(Lomton), F.I.C

Professor of Inorganic and Analytical Chemistry
in the Royal Technical College, Glasgow

BLACKIE & SON LIMITED

LONDON AND GLASGOW

1927

Printed in Great Britain by
Blackie & Son, Limited, Glasgow

PREFACE

Tliis work is a separate issue of Part I ami
Chapter XTI of my work, The Foundations of
Chemical Theory.

It is issued in this form in response to a request
for an elementary account of the subjects with
which it deals.

R. M. a

March, 1937.

CONTENTS

CHAP Fane

1. THE OLDER ATOMIC AND MOLECULAR THEOHIKS 1

Composition of Matter The Elements Laws of Chemical Com-
bination The Atomic, Theory The Molecular Theory La\s ut
Volumes Avogadro's Theory.

II. EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 18

Equivalent and Atomic Weights -Standard for Equivalent and
Atomic \\eights -Determination of Equivalent Weights
Methods of Determining Atomic Weights: (a) Method ot
Vapour Density and Avogadro's Theory Vapour Densities
by the Methods of Dumas, Hofmann, Victor Meyer;
(h) Method ot Chemical Displacement; (r) Method ot Speeilie
Heats (Dulong and Pet it's Law); (</) Method of Isomorphism
(Mitscheiiich's Law); (t) Method of the Periodic-, Li\\ - Jllu^tia-
tion : The Atomic Weight of Carbon Molecular Weights in
Solution: Cryoscopie and Ebulliscopic Methods Molecular
Complexity Molecular Compositions of (Compound Cases.

III. OLDKII VIEWS OF VALENCY AND CHEMICAL CONSTITUTION- - - 56

Historical Definition of Valency Bonds and Graphic FurmuLe
Variability of Valency Double Bond in Carbon Compounds
"Chemistry in Space " Criterion of Valency Nature of Valency.

IV. CLASSIFICATION OF THE ELEMENTS 69

The Periodic Law according to Mendelecff Law of Octaves
Development of the Periodic Law Periodicity of Physical
Properties Periodicity of Chemical Properties Periodicity
of Valency Uses of the Periodic Law Correction of Atomic
Weight Values Suggestiveness of the Periodic Law Objec-
tions to the Periodic Law.

V. THE MODERN VIEW OF THE ATOM - 93

Electrolysis and the Electron Radioactivity The Electron
and Valency Effect of Radioactive Change Isobares and
Isotopes Atomic Number The Modern Statement of the
Periodic Law- Positive Ray Analysis and Mass Spectra Iso-
topes of Lead Atomic S true Jure Valency Theories of Lewis
and Langmuir Theory of Bohr.

vii

viii CONTENTS

CHAP Page

VI. TlIK MoDKKN VIEW OF THE MoLEf'ULE 118

Electrochemical Theory of Ber/elms Dualistic and Unitary
Systems --- Electrolytic Dissociation Theory of Arrhenius
Eleetrovalenc} and Co valency The Octet Theory Molecular
Structures 1 Space Lattices Crystal Units and Molecules of
Solids FormuLe, Old and New.

VII. THE COLLOIDAL STATE 133

Solution and Suspension Crystalloids and Colloids Hydrosol
and Hydrogel Typical Colloids -The Ultra-microscope Size
of Colloidal Particles (Gradations between Suspension and
Solution Coagulation of CollouU Cataphoresis Peptization
Protective Colloids Phases and Media Uungo oi Colloidal
Phenomena.

THE FOUNDATIONS OF
CHEMICAL THEORY

PART IATOMS AND MOLECULES

CHAPTER I
THE OLDER ATOMIC AND MOLECULAlf THEORIES

i. The Composition of Matter

The term matter at first suggests to the unsophisticated mind
such qualities as bulk, shape, colour, hardness, weight. It is quickly
recognized, however, that some of these qualities belong only to
some kinds of matter, and are absent from others. Consider, for
example, a log of wood floating on water. It will be admitted that
the wood but riot the water possesses hardness, though both possess
weight; that the log but not the water has a permanent shape,
though both have bulk. Above the water is the air, and the air
is something, for it blows in a man's face, and raises ripples on
water. By a proper instrument it can be shown that air possesses
weight. So air is matter, though it is without form, permanent
bulk, apparent colour, or hardness.

At the conclusions suggested by these thoughts the Ancients
arrived after their own fashion. Their fundamental classification
of natural things included the three categories: earth, water, air,
together with a fourth fire. These four were the dements, accord-
ing to Aristotle, but their names really stood for qualities rather
than separate species of matter. For instance, earth meant dryness
and coldness, water wetness and coldness, and so on. Nevertheless,

first three terms at least suggest an outlook on the world which
Essentially true, since they stand for the three fundamental
forms of matter: solid, liquid, gas.

( D tX) ) 1 2

2 CHEMICAL THEORY

This conclusion as to the threefold constitution of the world
is reached by an extensive outlook upon nature; an intensive, an
introspective view, such as the following illustration furnishes,
leads to another conclusion.

Sea-water is distinguished from fresh water by its saltness, that
is, by its special taste. What proportion of salt water mixed with
fresh water could be BO distinguished depends upon the sensitive-
ness of the human palate; but such a test would assuredly fail
when the salt water was highly diluted. The addition of silver
nitrate to the much diluted salt water would, however, serve to
detect the presence of salt after the test of taste had failed, because
of the turbidity or opalescence which the silver nitrate produces
with even very small quantities of chlorides in solution. Would
this test fail in its turn when the utmost delicacy was required, or
would it detect the minutest quantity of salt? It might fail by
reason of defective human vision, were it not that an instrument
has been made on purpose to detect the slightest cloudiness in
a liquid; but it must fail at last for quite another reason. For the
test depends on the insolubility of silver chloride in water; but
silver chloride is not quite insoluble in water, and on this account
will not be precipitated when the salt solution is excessively
dilute.

So salt may perhaps be present in, and diffused through, water
in quantity too minute to be detected by any test whatsoever.
How far, then, may the dilution be carried; will salt still be present
after infinite dilution? Or, more generally, is matter infinitely
divisible? This is the question which arises directly out of an
experiment which any novice in chemistry can perform. The
same question presented itself to the alert minds of the ancient
peoples of the East; not, it is true, by reason of experimental
investigation, but because of meditation on the nature of the
material world. Thus the question was: Is matter infinitely divis-
ible or not ? To believe the latter is the easier and more satisfying
philosophy; this was the philosophy of Plato and Democritus. So
matter was supposed to consist ultimately of hard, indivisible, and
indestructible particles separated by vacuous interspaces; that is, of
atoms.

Of the two theories of the Ancients, to which the Greeks gave
finished expression the theory of Elements and the theory of
Atoms the former passed through strange vicissitudes, which

THE ATOMIC AND MOLECULAR THEORIES 3

need not here be traced; whilst the latter remained latent until
modern times, when it was found to be in accord with the con-
clusions derived by Dalton from experimental data.

2. The Elements

A theory of the elements should precede a theory of atoms.
So, dismissing the ancient theory of the elements, it may be said
quite briefly that an element, as generally understood, is an ultimate
species of matter; or, to adopt the more usual and explicit definition:

An element is a substance which hitherto has not been resolved
into two or more dissimilar kinds of matter.

We owe this idea of an element first of all to Boyle (1678); it
was Lavoisier (1789), however, who realized its provisional nature;
and, indeed, some of Lavoisier's elements, such ^ lime and the
alkalis, are now proved to be compounds. If this definition merely
marked the present state of progressive human achievement it
would not be a scientific definition. An assurance is necessary that
some kind of finality has been or may be reached in the decom-
position of substances; that at least the "elements" are equally
elementary; this assurance may be given with every confidence.
That the elements are absolutely undecoinposable has, however, never
been a settled belief of the chemist; on the contrary, he has held the
opinion from time to time that they are derived from, and so are
resolvable into, a common primordial substance. The phenomena
of radioactivity now furnish evidence of the spontaneous and
perpetual disintegration, into simpler forms of matter, of the atoms
of certain of the elements; and new theories of matter which are now
taking firm root in chemistry represent the atom as a complex struc-
ture; further, the fact has now been established that the elements
are not truly homogeneous, but consist of atoms of differing relative
weights, which, however, are indistinguishable and inseparable by
ordinary chemical means.

These considerations, however, do not affect the chemist's working
theory of the elements. He knows that the nearly ninety different
kinds of matter into which he has resolved the rrymy substances
found in nature, and out of which he can elaborate a vast number
of Compounds to which nature has no counterpart, pass unchanged
through the crucible of his everyday operations. So he habitu-
ally regards the catalogue of the elements that hangs in his

4 CHEMICAL THEORY

laboratory as a permanent record, not only of human skill, but
of Nature's handiwork as well.

3. The Atomic Theory

The atoms of Greek philosophy were indestructible; indeed, the
indestructibility of matter has probably always been an axiom
of science, notwithstanding the surprising and fantastic changes
matter was supposed to undergo in the hands of the alchemists
of the Middle Ages. This principle was first clearly illustrated,
however, by Lavoisier in his application of quantitative methods to
chemistry, and was subsequently demonstrated, within the limits of
the most accurate experimental research, by Stas, Landolt, and
others. 1 In 1770 Lavoisier gave an account of experiments he had
performed to test the supposition that water is transformed into
earth by boiling. A weighed quantity of water was boiled for
101 days in a weighed and sealed glass vessel; and at the end
of that time it was found that while "earth" appeared in the
vessel the water weighed the same as at first, and the weight
of the "earth" was equal to the loss in weight which the glass
vessel had incurred. Thus it was shown that the "earth" came
from the glass and not from the water, and that water is not
transformed into earth by boiling. In these experiments the use
of the balance played an essential part; but this was a novelty in
chemistry. The scientific achievements of such men as Boyle,
Black, Cavendish, Priestley, Scheele, notwithstanding their great
value, were chiefly of a qualitative nature. Henceforth, however,
chemistry was concerned with weighing things, and a new era
began.

It was now but a step to the quantitative analysis of chemical
substances. Lavoisier took this step in his investigation of mer-
curic oxide, or the calx of mercury, which Priestley and Scheele
had decomposed into mercury and oxygen. Soon there arose an
important question, the answer to which could be found only by
quantitative analysis. This was the question: Is a chemical com-
pound necessarily constant in composition, or may its composition
vary within certain limits according to the way in which it is
prepared ? c

It may appear to be a truism that the same compound 4 " mu^t

1 For an account of these researches see The StJidy of Chemical Composition, by I. Freun._

THE ATOMIC AND MOLECULAR THEORIES 5

always have the same composition, so that it is better to state the
problem in this way: Can the products of different chemical re-
actions, designed to produce the same compound, really differ
slightly in composition? Berthollet was of opinion that they
could; that the composition of a compound might vary within
certain limits according to the way in which it was prepared;
indeed, that the conditions of its genesis are the overruling factors
of its composition. Barium sulphate was cited as an example. All
known specimens of this compound were found to be identical in
composition, but this identity was due, not to any inherent pro-
perty of the constituent elements of the compound, but to the fact
that by uniting in such proportions these elements produced a
compound of maximum insolubility in water. -It was fair to suppose,
therefore, that if the salt could be precipitated from some other
medium than water it would have a different composition accom-
modated to a new requirement of maximum insolubility. Such an
idea was gravely erroneous, and was quite foreign to the principles
on which the atomic theory was soon to be founded. Yet the idea
appeared to have experimental support; and, indeed, it contained
the germ of an important truth. In support of his belief, Berthollet
showed that when nitric acid reacted with mercury or with tin the
composition of the nitrate of mercury or oxide of tin produced
varied within certain limits according to the concentration of the
acid employed. Proust, on the other hand, maintained that " be-
tween pole and pole compounds are identical in composition; their
appearance may vary owing to their manner of aggregation, but
their properties never". After a controversy carried on with
Berthollet over a period of eight years (1800-8), Proust fully
established his proposition, and showed that the variable products
obtained by Berthollet were variable mixtures of invariable com-
pounds. Thus was established the first law of chemical combina-
tion the law of definite or fixed proportions:

The same chemical compound always contains the same elements
united together in the same proportions; or the proportions between
the constituent elements of a chemical compound bear an unalter-
able relation to each other, and to the proportion of compound
formed,
" This was the first foundation of the atomic theory.

Nevertheless, it was a pity that the truth in Berthollet's view
was entirely overlooked in the victory of Proust: the truth that

6 CHEMICAL THEORY

the proportions or concentrations in which reacting substances are
present may determine the proportions subsisting between the
products of a reaction, although the proportions in which elements
or compounds actually react to form these products are quite be-
yond the influence of external and variable conditions. Thus, in
the case of the action of nitric acid on mercury, studied by Ber-
thollet, the concentration of the acid determines whether mercurous
or mercuric nitrate or a mixture of these two salts is produced,
although it can have no influence on the unalterable chemical
composition of either of the two salts.

The existence of two nitrates of mercury is, however, a note-
worthy fact, which appears the more striking when it is discovered
that in one compound the proportion of mercury to nitrate is
exactly twice what it is in the other. Further examples of this
phenomenon were observed by Dalton, who showed that the pro-
portion of hydrogen to a fixed quantity of carbon is twice as
great in methane as in ethylene, and of oxygen to a fixed quan-
tity of carbon, twice as great in carbonic acid gas as in carbonic
oxide.

Other examples of compounds showing analogous relations are
two of the oxides of lead, one of which contains twice as much
oxygen compared with lead as the other, and the five oxides of
nitrogen, in which the quantities of oxygen combined with a fixed
amount of nitrogen are as 1 : 2 : 3 : 4 : 5. Here was an important
generalization, which was formulated by Dalton as the law of
multiple proportions:

When one element combines with another in more than one pro-
portion, these proportions bear a simple relation to one another.

The foregoing facts furnish material enough for the atomic
theory. It is usual, however, to add to the laws of definite and*
multiple proportions a third law, the law of reciprocal proportions,
which, however, follows logically from the other two laws.

It was shown by Richter, about 1780, that the ratio between
the quantities of two acids which neutralize a fixed amount of
alkali is the same whatever the alkali may be; and by Berzelius, in
1810-2, that 38 4 l parts of lead combine separately with 58-73 parts
of sulphur and 29-6 parts of oxygen, whilst 58-73 parts of sulphur
combine in turn with 57 45 parts of oxygen. Now, 57 45 = 29 6 J< 2
within the limits of the experimental error of the time; and thes
facts may be expressed diagrammatically thus:

THE ATOMIC AND MOLECULAR THEORIES

Sulphur -< >- Oxygen

So is illustrated the law of reciprocal proportions:

The proportions of two elements which separately combine with
a fixed proportion of a third element are also the proportions of
these elements which combine with each other, or else in accord-
ance with the law of multiple proportions they bear a simple ratio
to these proportions.

This law has within it, especially in the way in which it was
illustrated by Richter, the idea of chemical equivalents; and so it
may be stated in this axiomatic way:

Quantities of substances which are chemically equivalent to the
same quantity of a third substance, are chemically equivalent to one
another.

Thus it appears that, granted the validity of the idea of chemical
equivalents, which will be examined later, the law of reciprocal
proportions requires no experimental justification.

Although we owe the essence of the modern atomic theory to
Dalton alone, the precise way in which the theory took shape in
the mind of its author has been rather problematical. At the
close of a paper on the absorption of gases by water, Dalton wroto
as follows:

"An inquiry into the relative weights of the ultimate particles of
oodies is a subject, as far as I know, entirely new. I have lately been
prosecuting this inquiry with remarkable success."

No hint is given in the context of the way in which the atomic
values, which follow, were estimated, nor of the precise reason why
such values were believed to exist. The idea that matter consists
of discrete particles was, however, in the air. Apart from the
ancient theory of atoms, a theory of particles had oeen held more
or^less firmly by F. Bacon, Boyle, Higgins, and others; whilst
NewiJbn made the following explicit statement:

" It seems probable to me, that God in the beginning formed matter

8 CHEMICAL THEORY

in solid, massy, hard, impenetrable, movable particles, of such sizes and
figures, and with such other properties, and in such proportion to space,
as most conduced to the end for which He formed them; and that these
primitive particles, being solids, are incomparably harder than any porous
body compounded of them, even so very hard as never to wear or break
in pieces; no ordinary power being able to divide what God Himself
made one in the first creation. . . . The changes of corporeal things are
to be traced only in the various separations and new associations and
motions of these permanent particles."

It would almost appear from such a pronouncement that
Newton and not Dalton was the author of the atomic theory.
Yet this statement is not a chemical theory: it is a cosmic theory
intimately related to Newton's great discovery of universal gravi-
tation. Dalton, however, was greatly indebted to Newton and
the idea of ubiquitous particles which the theory of gravita-
tion involved; and it appears that he conveyed this idea into
chemistry and employed it to explain the laws of chemical com-
bination.

To understand the atomic theory, therefore, is simply to under-
stand how the theory of particles fits the chemical laws. This is
quite easy.

Let there be three elements, A, B, C, and let the areas of the

squares:

represent the combining

weights of these elements on any arbitrary scale,

being; the

quantity of B which is found to combine separately with A

parts of A and

proportions; whilst

parts of C, according to the law of definite

parts of C also combine with

parts of A, according to the law of reciprocal proportions, so that
the following compounds are formed:

THE ATOMIC AND MOLECULAR THEORIES 9

Then, according to the law of multiple proportions, compounds
such as these may be formed:

&c. How else can these experimental facts be interpreted than by
the idea of " permanent particles " ? The elements combine accord-
ing to the laws of definite and multiple proportions because they
combine atom by atom: 1 atom of A with 1 atom of B; 1 atom
of A with 2 atoms of B; 2 atoms of A with 1 atom of B; and
so on. That is Dalton's atomic theory, and the theory is expressed
succinctly in the following statements: *

1. All matter consists of discrete particles called atoms, which
remain unbroken throughout chemical change.

2. Atoms of the same element are ordinarily supposed to be
similar in all respects.

3. Chemical compounds are formed by the union of the atoms of
different elements in simple numerical proportions.

4. The proportions in which elements combine to form compounds
are determined by the atomic weights of the elements.

The transition from the laws to the theory is quickly made: it
is taken, so to speak, in a stride; but the boundary line between
them must not be obliterated. The laws of chemical combination
are statements of experimental facts; the theory is an explanation
of these facts which is very probably true, but it does not stand
Vn the same category as the facts. In science, facts and theory
must always be distinguished as clearly as possible.

When the atomic theory is accepted it at once appears that
the combining weights of the elements represent the combining
weights of their atoms. The atomic theory involves the atomic
weights. No atomic theory previous to that of Dalton involved
atomic weights; these were a novelty, and their introduction
constitutes Dalton's great contribution to chemical science. The
following atomic weights are selected from a list published by
Dalton, the atomic weight of hydrogen being 1.

CHEMICAL THEORY

DALTON'S ATOMIC WEIGHTS

Hydrogen

An atom of water or steam,

Azote

composed of 1 of oxygen

Carbon or Charcoal

+ 1 of hydrogen

Oxygen

Phosphorus
Sulphur
Magnesia

9
13

An atom of ammonia, com-
posed of 1 of azote + 1
of hydrogen

Lime
Soda

1 I'ATI

28
38

An atom of carbonic oxide,
composed of 1 of carbon

JL1 Ull . .

Potash

+ 1 of oxygen

Zinc...

A 4-^ f rt Vi 'A

Copper

xi.n ai/om OL caroonic acid,
1 carbon + 2 oxygen

Silver

100

Gold

140

An atom of sulphuric acid,

Mercury

167

1 sulphur + 3 oxygen

If the student compares these atomic weights with those in
use at the present day, he will see that they differ widely from
the modern figures. Inaccuracies in Dalton's values are to be
expected, but it is not experimental error which attributes, for
example, an atomic weight of 7 to oxygen, instead of 16. As a
matter of fact these combining weights are not atomic weights at
all, but are approximately what we now recognize as equivalent
weights.

For, in Druth, Dalton had no means of determining atomic
weights. The value 7 (or 8) for oxygen is derived from the
analysis of water: 8 parts by weight of oxygen combine with
1 part by weight of hydrogen to form 9 parts by weight of water.
Who shall say from this that the atomic weight of oxygen is 8?
That depends on the number of atoms of each element which.,
combine together to form a unit of water, a fact clearly recognize?
by Dalton.

Thus, we have the ratio O : H = 8 : 1 or 16 : 2 or 24 : 3, &e.,
and if 1 atom of oxygen combines with 1 atom of hydrogen,
then the atomic weight of oxygen is 8; if 1 atom of oxygen
combines with 2 of hydrogen, the atomic weight of oxygen is 16;
if 1 combines % with 3, it is 24; if 2 combine with 1, it is 4; and
so on. There was, however, no evidence on which to base a
decision between these alternatives. Just at this point Dalton
made a regrettable mistake. Instead of recognizing the limits,
tions of his experimental knowledge, he made the assumption

THE ATOMIC AND MOLECULAR THEORIES 11

that since only one compound of hydrogen and oxygen was
known, it necessarily had the simplest possible composition, and
so was formed from 1 atom of each of its constituent elements.
Consequently, the atomic weight of oxygen was thought to be
7 (or 8); and for a similar reason the atomic weight of nitrogen
(azote) was supposed to be 5, and that of carbon also 5.

It is worth while to notice, however, that Dalton applied the
term atom to the ultimate particles of substances known to be
compounds as well as to those of elements; it is noteworthy
also that his numerical values furnish examples of the law of
multiple proportions; for instance, the composition of the two
oxides of carbon.

Dal ton's system of atomic symbols was ingenious: C^ stood
for oxygen, Qj for hydrogen, (j| for carbon, &c.; whilst for

compounds such formulae as yK_ t which stands for sulphuric

acid (S0 3 ), had to be constructed. In these formulae, however,
picturesqueness did not compensate for practical inconvenience;
and the suggestion of Berzelius (1811), that initial letters should
replace Dalton's hieroglyphics, found general acceptance.

4. The Molecular Theory

Dalton made no further advance along the road that he had
traversed. His assumption that the simplest formulae for a com-
pound is the right one was a subterfuge which marked the end
of the road. Advance must therefore be sought in another
.ejection; and it is found in the study of gases; for gases are
the simplest form of matter, and, if atoms exist, the properties of
gases will best elucidate their existence.

In 1805 Gay-Lussac and Humboldt studied the volume pro-
portions in which oxygen and hydrogen combine to form water;
and announced that " 100 volumes of oxygen required for com-
plete saturation 199-89 volumes of hydrogen, for which 200 may
be put without error ". This is a single example of a law, Oay-
Lu&ac's law of volumes, which is thus expressed:
^ The* volumes in which gases combine are simply related to each
other, and to the volume of the compound gas which is formed.

12 CHEMICAL THEORY

For example:

2 volumes of hydrogen combine with 1 volume of oxygen to form

2 volumes of steam.

1 volume of hydrogen combines with 1 volume of chlorine to form
2 volumes of hydrogen chloride.

3 volumes of hydrogen combine with 1 volume of nitrogen to form

2 volumes of ammonia.

A necessary corollary of this law is the statement that: the
densities, i.e. the masses of unit volumes, of the elementary gases
are simply related to their combining weights.

Thus, since 1 volume of hydrogen combines with 1 volume
of chlorine, and also 1 grm. of hydrogen combines with about
35-5 grm. of chlorine, the density of chlorine compared with
that of hydrogen as unity is about 35-5.

It further follows, if gases combine volume by volume, accord-
ing to the law of Gay-Lussac, and also atom by atom, according
to the theory of Dalton, that there is a simple connection between
the volume and the atom; and, indeed, that equal volumes of
hydrogen and chlorine, for example, contain equal numbers of
atoms. This conclusion, which was quite valid so far as it went,
was reached by Gay-Lussac, but was denied by Dalton, on account
of a difficulty which arose when the volume of the product was
considered.

Now when two separate and different elementary atoms com-
bine to form a compound atom, or whatever it may be called,
it is one entity they form, not two. It is impossible, for instance,
that 1 atom of hydrogen combining with 1 atom of chlorine can
produce two compound atoms of hydrogen chloride. And yet 1
volume of hydrogen combining with 1 volume of chlorine forms
2 volumes of hydrogen chloride. That was a dilemma; and it
was met by Dalton by a spirited denial of the law of Gay-Lussac.
"The truth is", said Dalton, "that gases do not combine in simple
proportions by volume; when they appear to do so, it is due to
an error in our experiments" !

Now, Dalton was wrong; and yet what other solution can
be found, unless indeed the "atoms" are torn in pieces in the
process of chemical synthesis, and the pieces are afterwards joined
together again in a different way? " ^

That is precisely the solution of the difficulty suggested by

THE ATOMIC AND MOLECULAR THEORIES

Avogadro, in his celebrated hypothesis. In this hypothesis, which
will now be expounded, two orders of particles were distinguished,
which we now call atoms and molecules. Atoms are indivisible
in ordinary chemical changes; molecules are aggregates of atoms
with a few exceptions which maintain their integrity in
ordinary physical changes, but suffer disruption in the course
of chemical change, so that their constituent atoms may be re-
arranged to form fresh molecules.

Now, when hydrogen chloride is formed from its elements
the volume of the product is twice the volume of the hydrogen
or of the chlorine; therefore it is sufficient to assume that the
molecules of hydrogen and chlorine consist of pairs of atoms,
which break into single atoms, and recombine, thus:

so that 1 volume of hydrogen plus 1 volume of chlorine gives
2 volumes of hydrogen chloride, instead of 1 volume, according
to the scheme:

It might be objected, however, that if the molecules of hydrogen
chloride are intrinsically twice the size of the atoms of hydrogen
and chlorine, out of which they are formed, the volume of the
compound gas might be expected in any case to be twice that of
either of the simple gases. Such an objection, however, is invalid,
since the actual size of the molecules of a gas is very small com-
pared with the molecular interspaces, and consequently the question
of a molecule of hydrogen chloride being intrinsically larger than
an atom of hydrogen or of chlorine does not arise.

The formation of 2 molecules of steam from 2 molecules of
hydrogen and 1 molecule of oxygen is thus represented:

CHEMICAL THEORY

The above processes of combination may be set forth in terms
of volumes, by using Dalton's symbols, thus:

1 vol. hydrogen. 1 vol. chlorine.

2 vois. hydrogen chloride.

2 vola. hydrogen.

1 vol. oxygen.

2 vols. steam.

Or by means of chemical equations:

H 2 + C1 2 = 2 HC1.
2H 2 + O 2 = 2H 2 O.

Thus the ipolecular formula H 2 O for water makes its appear-
ance. The proof of this formula is contained in the preceding
argument, ^vvhich may be thus epitomized:

Hydrogen and chlorine gases consist of diatomic molecules,
since the volume of hydrogen chloride they produce is twice the
volume of either single gas.

Similarly, oxygen gas consists of diatomic molecules, since the
volume of the steam is twice the volume of the oxygen it contains.
The only formula for steam which agrees with the diatomicity
of hydrogen and oxygen, as well as with the volumetric com-
position of steam, is H 2 O. That the density of steam (H 1)
is 9 furnishes no additional evidence, since it is deducible from
the densities of hydrogen and oxygen, and the volume of the
steam. That the atomic weight of oxygen is 16 follows from
the fact that its density is 16, and that, like hydrogen, it is
diatomic. The argument would, however, be invalidated if ft
were shown that these gases are not diatomic, that in the mole-
cules HX and O x , & is greater than 2. Underlying the whole of
this argument is Avogadro's hypothesis, which is stated thus:

Equal volumes of all gases and vapours, under the same conditions
of temperature and pressure, contain equal numbers of molecules.

But why hypothesis 1 This statement is not a law, any more
khan Dalton's atomic theory is a law. When first put forward
it was properly regarded as a hypothesis, which, indeed, suffered
much at the hands of its friends. Now, however, it is firmly
established, and is of fundamental Importance. It ought, therefore,

THE ATOMIC AND MOLECULAR THEORIES

to be dignified with the name of theory. Henceforward we shall
speak of Avogadros theory.

It will be seen that this theory is in accord with Gay-Lussac's
law of volumes, and satisfactorily explains the phenomena of
the combination of gases. Thus, 1 volume of hydrogen combines
with 1 volume of chlorine to form 2 volumes of hydrogen chloride,
because 1 molecule of hydrogen reacts with 1 molecule of chlorine
to form 2 molecules of hydrogen chloride. The language of
volumes may be exchanged for the language of molecules; that
is the significance of Avogadro's theory.

That equal volumes of different gases contain equal numbers
of atoms is true only when the molecules of these gases contain
equal numbers of atoms. It is a statement of limited truth, and
of no permanent importance. The same may be said of the state-
ment that the densities of elementary gases are in the same ratio
as their atomic weights. The important fact is that the densities
of all gases are in the same ratio as their molecular weights;
and further, that since the molecular weight of hydrogen is 2,
and its density, which is taken as the standard, is 1, therefore
the molecular weights of all gases are twice their densities. Thus,
the molecular weight of a gas or vapour is revealed by its density,
as the following approximate figures show:

Elementary Gas or Vapour.

Density.

Molecular
Weight.

Atomic
Weight.

Molecular
Formula.

Hydrogen

H 2

Oxygen

Nitrogen
Chlorine

14
35-5

28
71

14
35-5

cf,

Ozone

* Phosphorus

124

Mercury

100

200

Sulphur

128

256

It may be remarked, incidentally, that the magnitude of the
atomic weight of an element cannot be deduced from its gas or
vapour density unless the number of atoms confined within
the molecule of the element, i.e. its atomicity , is known indepenV-
dentty. As a rule, however, the atomic weight of the element
is kno^fn independently, and then the atomicity is deduced from
the density.

16 CHEMICAL THEORY

The breadth of Avogadro's generalization was not realized in
the time of its originator; and, owing to the persistence of the
volume-atom theory of Gay-Lussac, and its unwarrantable exten-
sion by Berzelius, 1 there was much confusion on the subject until
Cannizzaro, in 1858, reinstated Avogadro's theory on a permanent
basis.

It should be added that Avogadro's theory applies strictly
only to an ideal gas. When a gas deviates from Boyle's law it
deviates to the same extent from Avogadro's theory.

A useful fact to remember in connection with gas densities is
that a litre of hydrogen at C and 760 mm. pressure, i.e. normal
temperature and pressure (N.T.P.), weighs almost exactly 009 grm. r
or that 1 grin, measures 11 125 litres. Thus a gram-molecule (i.e.
the molecular weight in grams) of hydrogen at N.T.P. measures
22-25 litres ; and from Avogadro's theory it follows that the
volume of a gram-molecule of any gas or vapour, reduced to
normal temperature and pressure, is 22-25 litres. To determine
the weight in grams of 22 25 litres of any gas or vapour, reduced
to and 760 mm., is therefore to discover its molecular weight

SUMMARY

AN ELEMENT is a substance which hitherto has not been resolved
chemically into two or more dissimilar kinds of matter.

LAWS OF CHEMICAL COMBINATION. 1. Law of definite or fixed
proportions. The same chemical compound always contains the
same elements united together in the same proportions; or, the
proportions between the constituent elements of a chemical com-
pound are always the same.

2. Law of Multiple Proportions. When one element combines
with another in more than one proportion, these proportions bear
a simple ratio to one another.

3. Laiv of Reciprocal Proportions. The proportions of two
elements which separately combine with a fixed proportion of a
third element are also the proportions of these elements which
combine with each other, or else in accordance with the law of
multiple proportions th&y bear a simple ratio to these proportions.

1 The practice of referring all gaseous molecules to 2 volumes, which was a per.dciou&
outcome of the theorizing of Berzelius, appears now, fortunately, to be dying f ut. Why,
indeed, should every molecule be regarded as a microcosm of 2 volumes, as if it cw.ld>
necessarily be dichotomized? fc

THE ATOMIC AND MOLECULAR THEORIES 17

THE ATOMIC THEORY. 1. All matter consists of discrete
particles called atoms, which remain unbroken throughout chemical
change. *

2. Atoms of the same element are ordinarily supposed to be
similar in all respects.

3. Chemical compounds are formed by the union of the atoms
of different elements in simple numerical proportions.

4. The proportions in which elements combine to form com-
pounds are determined by the atomic weights of the elements.

GAY-LUSSAC'S LAW OF VOLUMES. The volumes in which gases
combine are simply related to each other, and to the volume
of the compound gas which is formed.

Corollary. The densities of the elementary gases are simply
related to their combining weights. 1

AVOGADRO'S THEORY. Equal volumes of all gases and vapours
under the same conditions of temperature and pressure contain
equal numbers of molecules.

Corollary. Since the molecule of hydrogen contains 2 atoms,
the molecular weight of any gas or vapour is twice its density
compared with that of hydrogen as unity.

A litre of hydrogen at N.T.P. weighs 0-09 grrn., and 1 gram-
molecule of hydrogen (2 grm.) measures 2225 litres. It follows
from Avogadro's theory that this is also the volume at N.T.P.
of 1 gram-molecule of any gas or vapour.

AN ATOM of an element is the smallest particle of matter
which takes part in a chemical change; it is the unit of chemical
exchange.

A MOLECULE is the smallest particle of matter which exists
independently; it is the physical unit. The molecule of an element
contains similar, that of a compound dissimilar atoms.

The number of atoms contained within the molecule of an
element is called the atomicity of the element.

J The term " combining weight" has sometimes signified equivalent weight, and some-
times atomic weight. Since the term is ambiguous, a use is found for it during the
development of the molecular theory when non-committal language is employed. After-
wards the term should be dropped.

(D60)

CHAPTER II
EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS

i. Equivalent and Atomic Weights

It was shown in the last chapter that Dalton's "atomic weights"
were really equivalent weights, and that the equivalent weight of
an element, when not identical with its atomic weight, is a sub-
multiple of the latter. Thus, whilst the equivalent weight of
oxygen referred to that of hydrogen as unity is approximately 8,
the atomic weight of this element, referred to the same standard, is
approximately 16. In general

Atomic weight = n x equivalent weight,

where n is a small whole number, which indicates the valency of
the element. Valency, or atomic value, is a new idea, necessary to
connect together the ideas of atomic weight and equivalent weight.
It will be more fully developed later.

It will now be useful to define equivalent and atomic weights.

EQUIVALENT WEIGHT. The equivalent weight of an element is
that weight of it which combines with, or displaces from combina-
tion, unit weight of a standard element.

ATOMIC WEIGHT. The atomic weight of an element is the ratio
between the weight of its atom and that of the atom of a standard
element.

When these definitions are considered, it appears that the
equivalent weight of an element is an experimental value, inde-
pendent of theory, whilst the atomic weight is connected with
the atomic theory.

It further appears that since equivalent and atomic weights
are ratios, they are not really weights at all, nor masses, but pure
numbers. That the atomic weight of an element is not the weight
of one of its atoms appears plainly enough when it is considered
that the standard of atomic weights has varied from time to time.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 19

Farther, since equivalent weights are values to be determined
experimentally, their determination may well form the starting-
point in the estimation of atomic weights. As a matter of fact the
accuracy with which the atomic weight of an element is known
depends as a rule on the accuracy with which the quantitative
observation of some chemical transformation has been carried out,
so as to determine its equivalent weight.

In some cases, however, atomic weights have been estimated
accurately by the determination of gas density.

For determining equivalents, comparison between reacting
quantities may be made by combination as well as by displace-
ment, because an element combines with, as well as displaces, what
is equivalent to itself. Thus, if there are two elements, A and B,
the chemical equivalent of B referred to A as standard is found by
estimating the amount of B which combines with a ^nown weight
of A, as well as by causing B to displace A, or A to displace B from
combination with another element or group of elements.

When the equivalent weight of an element is known, it is
necessary to determine the value of n in the above equation before
the atomic weight can be fixed. What multiple of the equivalent
weight the atomic weight may be, has to be decided by reference to
one or more of several distinct principles, which lie chiefly in the
domain of physical chemistry, and will shortly be discussed in
detail.

Standard for Equivalent and Atomic Weights.

The question of a standard needs first to be considered; and,
since hydrogen has the least atomic weight of all the elements, and
as small an atomic value (valency) as any element, it is natural to
^choose hydrogen as the standard both of atomic and equivalent
weights, and so to make its equivalent and its atomic weight both
equal to 1.

Now, although hydrogen combines with non-metals, and a few
metals, and is displaced from its combination in acids by some
metals, its chemical activity is too limited to permit its use as
a general standard of comparison. Oxygen, howevej;, with very
few exceptions, combines with all the elements, metals and non-
rnetefls alike; on this account it was called by Berzelius the "pole
of chenftstry". As a matter of practical experience, therefore,
equivalent and atomic weights 'are more often estimated with

20 CHEMICAL THEORY

reference to oxygen than to hydrogen; the hydrogen equivalent
may then be calculated from the oxygen equivalent by multiplying
the latter by the equivalent weight of oxygen, and thence the
corresponding atomic weight may be found.

Now, although Dalion (1808) chose hydrogen = 1 as the atomic
weight standard, oxygen was soon adopted in preference, so that
Wollaston (1814) used oxygen 10, Thomson (1825) oxygen
= I, Berzelius (1830) oxygen = 100, and Stas (1860-5) oxygen
- 16.

Until recently the two standards H = 1 and O = 16 were in
use, but the latter is now the standard adopted by the International
Union of Pure and Applied Chemistry. Although unity as the
standard is sacrificed by this procedure, the O = 16 hap, at least
two advantages over tho, 11 = 1 standard.

It was pointed out by Stas that the standard atomic weight
should, as far as possible, be directly connected with the atomic
weight to be determined, and this is the case when oxygen rather
than hydrogen furnishes the standard. Otherwise the ratio H :
is involved in the calculation when the data are derived from the
composition of an oxide; and whilst this ratio lias been determined
with great accuracy to be 1 : 15*88, any future modification of the
ratio would involve the recalculation of all atomic weights de-
pendent upon it. If, however, the ratio is written 1*008 : 16, the
atomic weight of oxygen being fixed at 16, any future alteration
will involve only the atomic weight of hydrogen. The advantage
of this is plain.

Another advantage of the modern system is the fact that
when = 16 several other important atomic weights approximate
very closely to whole numbers; e.g. C = 12*00, N = 14*01,
Na = 22*997. The reason for this approximation will appear
later when modern views of the atom are considered.

It is unlikely that any further modification of the standard
will now be made, but an unfortunate confusion even now remains
in the minds of those who have employed several standards. For
example, the atomic weight of chlorine has been variously given as
35*37, 35*1, 35*46; and these differences are due not to different
estimations of the atomic weight of this element but to the adop-
tion of three different standards for oxygen, viz. O = T5*9G
(Dumas), 15*88, and 16*00. *

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 21

2. Determination of Equivalent Weights

The following are the more important methods ordinarily
employed in the laboratory to determine the equivalent weights of
elements.

i. The measurement of the volume of hydrogen displaced from
dilute sulphuric or hydrochloric acid by a weighed amount of
a metal.

ii. The conversion of a weighed quantity of a metal into its
oxide which is weighed, or the reduction of a weighed quantity of
oxide to metal.

iii. The displacement of a metal from a solution of one of its
salts by a weighed quantity of a more chemically powerful metal.

iv. The separation of elements at the electrodes during the
passage of an electric current through a series of electrolytes.
This method yields the electro-chemical equivalent t)f an element;
but this value is numerically identical with the chemical equivalent.

i. The chemical equivalent of magnesium, zinc, or aluminium
may be easily determined by dissolving a weighed quantity of the
metal in the dilute acid contained in a piece of apparatus designed
for collecting the evolved hydrogen. The gas is measured over
water at atmospheric temperature and pressure; it will consequently
be moist, and the pressure of water vapour at the observed tempera-
ture must be subtracted from the atmospheric pressure, before the
volume of the gas is corrected to normal temperature and pressure.

The weight of metal divided by the weight of the evolved
hydrogen gives the hydrogen equivalent of the metal. This must
be multiplied by 1*008 if the equivalent on the modern atomic
weight basis is desired; though in view of the likely experimental
error such a correction is superfluous.

The experiment may easily be carried out on the lecture-table or
by students. The following result has been obtained by a student:

\Veightof magnesium taken = 0*033 grin.

Volume of moist hydrogen measured at) __ . v> r _
12 J C. and 756 mm. j - J--i>c.c.

Pressure of water vapour at 12 = 10-5 mm.

Volume of dry hydrogen at N.T.P. = l**

= 30-6 c. c.
of hydrogen = 30*6 X 0-00009 = 0-002754

0*^"^

Equivalent of magnesium = * ' = 12*0.

U'vOii/t.) *

22 CHEMICAL THEORY

ii. Magnesium may be converted quantitatively into oxide by
the ignition of the metal in the air under suitable conditions, or
by dissolving it m dilute nitric acid, evaporating the solution, and
igniting the nitrate until brown fumes cease to be evolved. These
methods are not without sources of error, but it may be shown that
0-30 grrn. of magnesium yields almost exactly 0-50 grm. of oxide, so
that the equivalent weight of magnesium is

-3x8 = 12

The method of conversion into oxide through the nitrate is
applicable to such metals as zinc and copper, which dissolve in
nitric acid and yield stable oxides by the decomposition of their
nitrates. The equivalent of tin may be determined by the con-
version of the metal into hydrated dioxide by means of nitric acid,
since the ignition of the product yields the pure dioxide.

It would be possible to determine the equivalent of carbon by
burning a weighed quantity of the element in a stream of dry air
or oxygen, and collecting and weighing the carbon dioxide formed;
but the great difficulty of obtaining pure carbon free from hydrogen
under ordinary conditions stands in the way of this determination.

For the determination of an equivalent by the reduction of an
oxide to metal, copper furnishes the usual example, since the
reduction is easily carried out by passing a stream of hydrogen
over oxide of copper contained in a boat in a heated tube. Thus
1-00 grm. of black oxide of copper leaves a residue of 0-799 grm. of
copper; whence the equivalent of copper in this oxide is

x 8 - 31<8 '

There is another oxide of copper, however, the red oxide, whose
equivalent weight is 31*8 x 2 = 63 6. This fact is connected with
the exhibition of a dual valency by copper, which again furnishes
an example of the law of multiple proportions. This phenomenon
will be further dealt with under the subject of valency.

iii. A well-known example of the displacement of a metal
from the solution of one of its salts by another metal is the action
of zinc upon a solution of copper sulphate, when the zinc is sup-
posed to displace from combination its equivalent of copper wfiich
may be collected and weighed. This takes place almost quantita-
tively when a cold concentrated solution of copper sulphate is

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 23

employed; but the method is generally unreliable, because other
reactions occur between the displacing metal and the solution
simultaneously with the main reaction, and these vitiate the results.
The method is not therefore to be recommended.

iv. When a suitable electric current is passed through acidified
water contained in a "voltameter", hydrogen and oxygen are

evolved in the proportion of two
volumes of hydrogen to one
volume of oxygen. Provided the
densities of hydrogen arid oxygen
are known, and the conditions
of temperature under which the
gases were measured have been
observed, the hydrogen equiva-
lent of oxygen i^ight be calcu-
lated from the volume relations

Oxygen
0*08 s

Hydrogen
01008 g

Fig. 1

of the gases. The estimation would not, however, be very accurate,
owing to several sources of experimental error.

If, however, the same current passes in succession through
several salt solutions for example, copper sulphate, silver nitrate,
gold chloride solutions it will liberate at the cathodes or negative
electrodes amounts of the metals chemically equivalent to the
hyclro^en which is liberated in the voltameter. Thus, whilst
0-01008 grm. of hydrogen gas is being evolved, and 0-08 grin, of
oxygen, 0-318 grm. of copper, { 079 grm. of silver, and 0-657 grm.

24 CHEMICAL THEORY

of gold will be deposited in the successive electrolytic cells. The
necessary arrangement is shown in fig. 1. Thus the equivalent
weights of these metals are determined.

3. Determination of Atomic Weights

It has been suggested in the previous pages that two distinct
considerations have to be taken into account in the problem of
atomic weight determination. These are:

i. An exact estimation of the chemical equivalent of the
element must be made, generally by carrying out some suitable
chemical transformation, occasionally by other means.

ii. A decision must be arrived at as to the order of magnitude
of the atomic weight, so as to discover the small whole number by
which the equivalent weight must be multiplied to give the atomic
weight.

The order in which the two parts of the problem are here
placed is that which would naturally occur to the mind. Never-
theless it is not the order of historic sequence in relation to modern
atomic weights. The approximate magnitude of the atomic weights
of all the elements has long since been settled and is not discussed
in modern research upon atomic weights; but the determination of
the exact values of all these atomic weights is a laborious task
which is not yet completed.

The methods for determining chemical equivalents which have
been described above are suitable for demonstration purposes, but
not all of them are equally useful in the actual determination of
atomic weights. Illustrations of the methods that have been
employed in accurate atomic-weight determinations will be given
in the sequel.

The principles which have led to decisions upon the order of
magnitude of the atomic weights of the elements will now be
dealt with.

It has already been seen that Dal ton was in need of some
guiding principle to enable him to fix the magnitude of his atomic
weights; andtthat such a principle came to light in the discovery
by Gay-Lussac of the law of gaseous volumes, and the proper
interpretation of this law by Avogadro. Thus, by means 4 of
Avogadro's theory it was shown that the atomic weight of oxygen
is very probably 16 and not 8; but? clearly this theory is limited in

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 25

its application, since it can only be of use in the case of gaseous
or gasifiable substances. Here may be mentioned the method of
chemical displacement, which is of some value in deciding the
magnitude of atomic weights.

In 1819 two other and quite distinct principles became avail-
able in the law of specific heats of Dulong and Petit, and the law
of isomorphism discovered by Mitscherlich. These laws are espe-
cially valuable in furnishing guidance as to the magnitude of
atomic weights, because they are applicable to solid elements and
their solid compounds.

The former of these two laws is the more important, and has
the wider application. Finally, the periodic law, established by
Mendel^eff in 1869, has been of distinct value in several ways in
fixing the approximate magnitude of atomic weights.

So the five guiding principles that aid in settling the order of
magnitude of atomic weights are:

i. Avogadro's theory,
ii. Chemical displacement,
iii. Dulong and Petit's law of specific heats,
iv. Mitscherlich's law of isomorphism.
v. MendeleefTs periodic law.

i, The Method of Avogadro's Theory.

It will be remembered that according to Avogadro's theory the
molecular weights, not the atomic weights, of gases and vapours are
proportional to their densities. It follows, therefore, that the rela-
tive magnitudes of molecular weights, and not of atomic weights,
are directly deducible from Avogadro's theory. So the question
arises how far a knowledge of the relative weights of molecules
can be of use in fixing the relative weights of any of their con-
stituent atoms. Such knowledge may be employed in two ways.

Consider the following volatile hydrocarbons:

Methane.

Ethylene.

Propane.

Benzene.

Naphthalene.

Approximate Density \

i i\ ^ s*\ f

39*

(0 = 16) (

Approximate Molecular \
Wejght /

128

Molecular Proportion of \
Carbon J

120

26 CHEMICAL THEORY

Approximate estimations of gas or vapour density yield ap-
proximate molecular weights; whilst quantitative analysis shows
the proportion of carbon within the molecular proportion of each
compound. Now, it is evident that all these hydrocarbons, except
the first, contain more than 1 atom of carbon in their molecules.
The molecule of methane might indeed contain more than 1 atom,
though the fact than no submultiple of 12 appears in the pro-
portions of carbon in the other molecules is evidence, so far as it
goes, that the figure 12 represents an indivisible unit, or in other
words that 12 is approximately the atomic weight of carbon. And
since by the examination of the very large number of hydrocarbons
that exist, every molecular proportion has been found to contain
12, or a multiple of 12 parts of carbon, the probability that 12 is
the atomic weight of carbon reaches a practical certainty.

The principle thus illustrated may be put in the following words:

The least proportion of an element found within the molecular propor-
tion of any of its volatile compounds is likely to be the atomic weight of
the element; and if the number of compounds which have been examined
is large, the value indicated is very probably the atomic weight.

The question may be asked, however, whether atomic weights
can be determined exactly by the method of Avogadro's theory, i.e.
by the determination of gas density, and the answer is in the
affirmative, provided an ideal gas density can be determined and
the molecular composition of the gas is known.

The density of a gas is determined by weighing a large glass
globe of about 10 litres capacity, first evacuated, and then filled
with the gas, at known temperature and pressure, corrections
being applied for the air displaced by the globe, and for the
slight shrinkage which the glass undergoes when the globe is
evacuated. Thus it has been found, as the mean result of the
experiments of Rayleigh, Morley, and Leduc, that 1 litre of
oxygen at and 760 mm. at the latitude of Paris weighs 1*42895
grm., whilst 1 litre of hydrogen, under similar conditions, accord-
ing to the experiments of Morley and Leduc, weighs 0-08985 gnu.

To conclude, however, that the atomic weights of oxygen and

hydrogen are in the ratio ' ^, although we know that the

08985

molecules of both gases are diatomic, would be erroneous, becau'se it
would be to assume that the gases are ideal gases which behave
in perfect accord with the gas laws (q.v.), and so with Avogadro's

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 27

theory. Such, however, is not the case, and the deviation of these
gases from the ideal must be discovered, and allowed for. This
may best be done, in the present case, by determining the effect
of the deviation upon the volume relations in which oxygen and
hydrogen combine to form water.

Now it has been estimated that 2-00268 litres of hydrogen
combine with 1 litre of oxygen, at and 760 mm. at the latitude
of Paris. This complex ratio is due, not to any discrepancy
between the simple proportions in which the molecules of these
two gases interact, but to the fact that equal volumes do not
contain quite equal numbers of molecules because oxygen is a
little more compressible than hydrogen. But since the densities
relate to equal volumes it may be concluded that

2-00268 x 0-08985 grin.

of hydrogen combine with 1*42895 grm. of oxygen, and therefore
that the hydrogen equivalent of oxygen is

1-42895

2-00268 x 0-08985

= 7-9412;

and its atomic weight 15-88 when H = 1; so that H = 1-0076
when O = 16. A similar method may be applied to determine
the atomic weight of a constituent element of a compound gas.
Thus, by the calculation of the ideal densities of carbon monoxide,
carbon dioxide, methane, and acetylene, by applying a correction
for compressibility to the estimated densities, several observers
have accurately determined the molecular weights of these gases,
and thence the atomic weight of carbon.

METHODS OF DETERMINING VAPOUR DENSITY.

The determination of gas density always consists in weighing a
certain volume of the gas; but for determining the vapour density
of a volatile liquid or solid, an alternative procedure may be
adopted: the volume of the vapour produced by a weighed quantity
of the liquid or solid may be measured under known conditions.

There are three well -recognized methods of vapour density
determination: the methods of Dumas, Hoftnann^ and Victor
Meyer. In the first of these three methods the weight of a
knflwn volume of the vapour is ascertained; in the two latter the
volume* of a weighed quantity of the substance is measured. The
method of Victor Meyer is the easiest and most often employed.

28 CHEMICAL THEORY

(a) Dumas s Method of Vapour-density Determination

A glass globe of the shape shown in fig. 2, and capable of
holding from 50 to 100 c. c. or more, is weighed, and then filled
with the vapour of the substance in the following manner.

A few cubic centimetres of the liquid are introduced into the
globe, which is then immersed in a bath of another liquid whose
temperature is kept constant, and from 20 to 30 above the
boiling-point of the liquid in the globe. As the latter liquid
boils it displaces the air from the globe, and vapour issues from
the neck as long as any liquid remains within the globe. When
the stream of vapour ceases, the globe is filled with
the vapour at atmospheric pressure, and at the tem-
perature of the bath in which it is immersed. The
neck is then sealed by means of a blowpipe; and
the temperature of the bath, and the pressure of
the atmosphere at the time of sealing are recorded.
After being cleansed, the sealed globe is weighed,
and the temperature and pressure of the air in the
vicinity of the balance are also observed.
Since the true weight of the sealed globe with its contents
is equal to its apparent weight plus the weight of the air which
it displaces whilst it is being weighed, the weight of this air must
be calculated and added to the apparent weight. For this calcu-
lation, as well as to ascertain the volume of the vapour at the
time of sealing, the cubical capacity of the globe must be de-
termined. This is done by breaking off the end of the neck of
the globe under water, which should then enter and fill the globe.
The quantity of water in the globe is determined by another
weighing, the weight of the air displaced being in this case
negligible; then the weight of the water in grams shows the
volume of the globe in cubic centimetres with sufficient accuracy.
From these data the weight of the known volume of the vapour
contained by the globe at the temperature and pressure at which
it was sealed is calculated. The volume is then reduced to N.T.P.,
and the weight of hydrogen or air corresponding to it is calculated.
The ratio of the weight of the vapour to that of the hydrogen
is the vapour density of the substance. *

EXAMPLE. Calculate the density of ether vapour fiftm the
following data:

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS

Weight of open globe in air = 22-549 grm.

Temperature of bath at time of sealing = 60 C.
Atmospheric pressure at times of sealing and weighing -'-= 760 mm.

Apparent weight of sealed globe + vapour in air = 22-662 grm.

Temperature of air at time of weighing =15

Capacity of globe, indicated by weight of water itl _ * r

can contain J ~ *'* c ' "

Weight of 1 c, c. of air at C. and 760 mm. = 0-001293 grm.

Weight of 1 c. c. of hydrogen at C. and 760 mm. = 0-0000899 grin.

Calculation
Weight of air displaced when sealed globe is weighed

= O '1293 X 75 X 273 =

288

Weight of vapour in globe = 22-662 + 0-092 22-549 = 205 grm
Weight of an equal volume of hydrogen at 60 C. and 760 mm.

=- Q ' QQQQ899 X 75 X 273 _
333
0-205

So density of ether ((C 2 H 6 ) 2 O) vapour =

0-00553

= 37-1.

(6) Hofmanns Method of Vapour-density Determination

A weighed quantity (about
005 grm.) of the liquid con-
tained in a small, drawn-out
bulb or stoppered bottle is in-
troduced into the Toricellian
vacuum of a graduated barom-
eter tube surrounded by the
vapour of a liquid boiling at
a suitable temperature, which
may be below the boiling-point
of the liquid whose vapour den-
sity is being determined. As
the liquid is vaporized it de-
presses the mercury in the
barometer tube; and when the
volume has become constant it
is read off, and the temperature
of the vapour jacket is observed.
The pressure of the vapour is
equal to atftiospheric pressure
less the height of the mercury
in the tube above its level in lg *

the vessel in which the tube stands. Strictly speaking, the height of

CHEMICAL THEORY

the mercury column should be corrected for expansion by heat; but
this need not be considered. From these data the vapour density
of the liquid can be calculated; as the following example shows:

Weight of stannic chloride (B. P. 114') 1 taken = 0-0445 grm.

Volume of vapour

Temperature of vapour jacket

Barometric pressure

Height of morcury column

Whence pressure of vapour

Volume of vapour reduced to N.T.P.
Weight of 3-75 c.c. of hydrogen at N.T.P.

= 162c. c.
99

752 mm.
512 mm.
240 mm.

3-75 X 0-00009
0-0003375 grin.

Vapour density of stannic chloride (SnCl 4 ) = -

0-0445

'_ = 131-8.

Fig. 4

0-0003375

In this method a weighed quantity of the
substance is made to evaporate into a space sur-
rounded with the vapour of a boiling liquid whose
lx)iling-point is at least 2*5 higher than that of
the substance. The volume of the vapour is not
directly measured, but the air displaced by it is
collected and measured at atmospheric tempera-
ture and pressure; while all the displacing vapour
remains in the locality of its production. The
\\ eight of an equal volume of hydrogen is then
calculated, and the weight of substance taken
divided by this weight of hydrogen gives the
vapour density of the substance, since the vapour
of the substance, if it could be obtained at atmo-
spheric temperature and pressure without conden-
sation, would occupy the same volume as the air.
The tube A (tig. 4) is closed at the lower end, and
is furnished with a bent delivery tube B which
dips under water in the dish C. The upper end
of A is closed by a rubber stopper. The lower
part of the tube is heated by the vapour of a
' D liquid, e.g. water, boiling in the outer jacket D,
and, owing to expansion, air escapes by the side

1 Since the compound ^is vaporized into a vacuum the temperature
of its vapour may be lower than the B.P. of the compound.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 31

tube through the water. When no more bubbles of air are seen,
the graduated tube E is placed over the end of B, and a little tube
or loosely-stoppered bottle, containing a weighed quantity of the
substance under investigation, is dropped to the bottom of A, being-
received on a pad of asbestos or glass wool, which prevents frac-
ture. For the introduction of the little vessel containing the sub-

stance the rubber stopper is removed, and quickly replaced, or the
vessel may be held by a mechanical contrivance at the top of the
tube A, where the air is nearly cold, 1 and then allowed to fall at
the right moment, without opening the tube.

EXAMPLE. 0-144 grm. of chloroform displaced 28-6 cu. cm. of
moist air measured at 14 and 756 mm. pressure. Pressure of
water vapour at 14 = 12 mm.

Vol. of air at N.T.P. = 286 X 273 X (750~JL2) = ^ . Q

287 X 760

Weight of an equal vol. of hydrogen = 26-6 X 0-0000899 grm.

= 0-00239 grm.

Vapour density of chloroform \_ Ovl44 _ rn o
CHC1 3 J ~ 6-00239 ~ -^

The method of Victor Meyer is more easily carried out than
either of the other methods. It employs very little of the substance
and is sufficiently accurate for most purposes. Consequently, it
is the method usually employed.

] It might be supposed that since the air is colder in the upper part of the tube,
which extends beyond the vapour jacket, than in the lower part, which is within it, too
much air will be displaced, and a high result obtained. This, however, is not the case,
because of the contraction of the air that rises in the body of the tube to take the place
of the air driven out. The following proof of this statement has been given by Dr. E. B. K.
Prideaux.

First, suppose temperature constant in the V. Meyer tube, so that the heated vapour
immediately displaces its own volume of heated air, which is then cooled. Let T = abs.
temperature of vapour and air when first expelled, and T abs. temperature of cooled air
leaving the end of delivery tube under water ; let V = vol. of vapour formed and therefore

of air expelled, and V vol. of air collected. Then V = --

Second, let there be two temperatures T and Tj within the tube, with corresponding
volumes of equal masses of air V and V^

T V

Then let V be expelled into the T x region, and thereby become V^ so that V t = 11.

Vj, not V, will now displace its own volume of air, which will be cooled so as to become,

TV T T V TV

say, V t at the end of the delivery tube. Then V 2 = *Jfl = 4il- = H- = V ; and

LI lil 1

similarly w~"h any number of temperature zones.

Thus a temperature gradient within the V. Meyer tube docs not affect the volume of
air displaced from the end of the delivery tube.

32 CHEMICAL THEORY

ii. The Method of Chemical Displacement.

Somewhat related to the above principle is another by which
the molecular formula of a compound may be determined, and so
the atomic weight of a constituent element.

Consider methane. The hydrogen in this compound can be
displaced by chlorine in four distinct stages, the following sub-
stitution products being formed : methyl chloride, methylenc
chloride, chloroform, and carbon tetrachloride. The carbon, how-
ever, cannot be displaced fractionally. From these facts the
inference is drawn that the molecule of methane contains 4
atoms of hydrogen and only 1 atom of carbon; but if methane
is CH 4 , Dalton's problem of the number of atoms in the molecule
is solved, and the atomic weight of carbon is 12.

A similar argument may be applied to water. The composition
of sodium hydroxide proves that half the hydrogen of the water
molecule has been displaced by sodium. By no means, however,
can any fraction of the oxygen of the water molecule be displaced.
Thence it is concluded that water is H 2 0, and that the atomic
weight of oxygen is 16.

The principle of this method of fixing the magnitude of atomic
weights may be stated thus:

When th of the proportion of a constituent element in a chemical
n

compound can be displaced by another element, a molecule of the
compound contains at least n atoms of that element.

iii. The Method of Dulong and Petit's Law.

The specific heat of a substance is the ratio of the amount of
heat required to raise unit weight of it through one degree of
temperature to the amount of heat required to raise unit weight
of a standard substance through the same temperature interval.
The standard substance is water.

In 1819 Dulong and Petit published the specific heats of
thirteen elements, and showed that the product of specific heat
into atomic weight is approximately a constant quantity, the
average of which, on our modern atomic weight basis, i 64.
In the following table, containing the elements studied by Dulong
and Petit, modern values are given throughout.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 33

Element.

Specific Heat.

Atomic Weight.

Atomic Weight
x Specific Heat
= Atomic Heat.

Bismuth

0-0305

208-0

6-34

Lead . . .

0-0315

207-2

6-53

Gold ...

0-03035

197-2

5-99

Platinum

0-03147

195-2

6-14

Tin ...

0-0559

119-0

6-65

Silver ...

0-0559

107-88

6-03

Zinc

0-0939

65-37

6-14

Tellurium

0-0475

127-5

6-06

Copper . . .

0-09232

63-57

5-81

Nickel ...

0-10842

58-68

6-K)

Iron

0-10983

55-84

6-13

Cobalt ...

0-10303

58-97

6-08

Sulphur

0-1712

32-07

5-49

The law of Dido ng and Petit may therefore bo stated thus:

The specific heats of the solid elements are in the inverse ratio
of their atomic weights.

The product of specific heat and atomic weight, which is ap-
proximately a constant, is called the atomic heat because it is
the heat capacity of a quantity of an element proportional to
its atomic weight. Thus, for example, 55*84 parts by weight
of iron require the same amount of heat to raise them through
one degree of temperature as, say, 208-0 parts of bismuth. But
these quantities of the elements contain equal numbers of atoms.
So, in the words of Dulong and Petit, " the atoms of all substances
have exactly the same capacity for heat".

In order to reach this result, however, Dulong and Petit made
some drastic changes in the accepted atomic weight values, which
aroused the opposition of Berzelius, their author. Thus, taking
the atomic weight of sulphur as a true magnitude, they halved
the atomic weights of the metals in relation thereto. This pro-
cedure was, however, justified, even in the opinion of Berzelius,
after Mitscherlich, his pupil, had arrived at similar conclusions
by an application of the law of isomorphism.

Now, since

Specific heat X atomic weight = 6-4 (approx.)

or atomic weight = ^^ T (approx.),

* specific heat

here is a valuable method for fixing the magnitude of the atomic

(D60) 4

34 CHEMICAL THEORY

weight of an element. All that it is necessary to do is to deter-
mine the specific heat of the element, and divide 6-4 by this value.

It must be clearly understood, however, that the value thus
obtained is only approximate, for the atomic heat value, 6-4, is
only approximate, since it is a mean value, even if the specific heat
is accurately known. The method serves to indicate what multiple
of an accurately determined equivalent weight is the atomic weight.
To divide 6-4 by the given specific heat of an element, and report
the quotient as its atomic weight, is a gross error.

The following illustration will make plain the use of Dulong
and Petit's law: Marignac 1 found that 100 grm. of lead yielded
134*201. grm. of the chloride. The specific heat of the metal is
0-0315; find its atomic weight; Cl = 35-46.

The equivalent weight of lead is found from the proportion:

Wt. of chlorine : wt. of lead : : equivalent Cl : equivalent Pb,
so 34-201 : 100 :: 35-46 : 103-68.

The approximate atomic weight of lead, as indicated by its

specific heat, is: A - 6 ^ - 203-2.
-

Therefore the atomic weight of lead ivS twice its equivalent weight ;
so Pb = 103-68 X 2 = 207-36.

Dulong and Petit's law applies strictly only to solid elements,
generally metals, whose atomic weights exceed 30. The specific
heats of other solid elements vary with temperature, but become
approximately constant at high temperatures, when they give an
atomic heat value of about 5-5.

iv. The Method of the Law of Isomorphism.

Isomorphism is similarity of crystalline form. It was supposed
by the earlier mineralogists that identity of crystalline form
generally indicated identity of chemical composition; but it was
shown by Mitscherlich in 1819 that compounds of analogous as
well as identical composition crystallize in similar forms belonging
to the same crystal systems. Thus di-sodium hydrogen phosphate
and di-sodi\im hydrogen arsenate, which are now represented by
the formulae Na 2 HPO 4 -12H 2 O and Na 2 HAsO 4 -12H 2 O, were found
to be isomorphous. Careful measurements of the crystal angles
of isomorphous salts show that these angles are not quite equal,

* Marignac, (Euvrt* Gamplite*, 1846, I, 186.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 35

but the following criteria of isomorphism have been established:
(i) great similarity of crystalline form, (ii) analogous composition,
(iii) power to form mixed crystals by simultaneous* crystallization,
(iv) power of crystal overgrowth, so that a crystal of one com-
pound may form the matrix on which the growth of the crystal
may be continued by the deposition of another substance.

With regard to the second criterion, it must be noted that
isomorphism is sometimes observed in pairs of compounds which
are not chemically analogous, but have the same numbers of atoms
within their molecules. Thus calc-spar (CaCO 3 ) is isomorphous
with Chili saltpetre (NaNO 3 ), and aragonite (CaCO 3 ) with nitre
(KN0 3 ).

Mitscherlich stated the law of isomoiy hism as follows:

4 'The same number of atoms combined in the same manner
produce the same crystalline form; the crystalline form is independent
of the chemical nature of the atoms, and is determined solely by
their number and mode of combination."

Nevertheless isomorphism such as that illustrated by the case
of sodium phosphate and arsenate is the rule; that is to say, not
only do the molecules of isomorphous compounds contain the same
number of atoms similarly combined, but these atoms themselves
are analogous, as, for instance, are phosphorus and arsenic. Indeed,
isomorphism is taken to be a sign of chemical analogy.

Therefore, for practical purposes, the law of isomorphism may
be stated more briefly:

The molecules of isomorphous substances contain equal numbers
of atoms, which when not of identical are of analogous elements.

The consequence of this law, when applied to the case already
mentioned, is that the atomic weights of phosphorus and arsenic
can be directly compared, and if one atomic weight is known the
other is derivable from the results of chemical analysis.

A simple numerical example is furnished by the following
results of the analysis of the isomorphous salts potassium sulphate
and potassium selenate, carried out by Mitscherlich:

K 2 SO 4 K 2 Se0 4

100 parts contain 100 parts contain. 127*01 pSrts contain

K 44-83 K 35-29 44-83

* O 36-78 O 28-96 36-78

SV 1 ^!^ Se 35 ' 75 45-40

100-00 ICfe-OO 127-01

36 CHEMICAL THEORY

In the third column is shown the proportion of selenium in an
amount of the selenate which contains the same amounts of potas-
sium and oxygen as are shown in the percentage analysis of the
sulphate. Whence it follows that 4540 parts of selenium take
the place of 18*39 parts of sulphur.

Now, the law of isomorphism declares that the ratio between
these quantities is the ratio between the atomic weights of the two
elements. Therefore, if the atomic weight of sulphur is 32-0 that

of selenium is * = 79-0.
18-39

The phenomena of isomorphism are somewhat confused by those
of dimorphism and polymorphism. Thus calcium carbonate, as
shown above, is dimorphous in calc-spar and aragonite; ammonium
nitrate, NH 4 NO 3 , is tetramorphous, crystallizing in four distinct
forms; arsenious and antimonious oxides, As 4 O 6 and Sb 4 O 6 are
isodimorphous, that is to say, they are both similarly dimorphous.
Nevertheless, the phenomena of isomorphism have been of value,
not only in confirming atomic -weight magnitudes derived from
other considerations, but in correcting erroneous magnitudes.

For example, previous to the recognition of isomorphism,
Berzelius regarded various metallic monoxides MO as dioxides
M0 2 ; similarly, Fe 2 O 3 was written FeO 3 , Cr 2 O 3 was CrO 3 , and CrO 3
was Cr0 6 . But when this chemist recognized the isomorphism of
chromates with sulphates he altered Cr0 6 to Cr0 3 to agree with
S0 8 , the oxide known to be present in sulphates. Consequently,
the former CrO 8 became Cr 2 O 3 ; and since chromic and ferric alums
were isomorphous, what was formerly FeO 3 became Fe 2 3 , and so
Fe0 2 became FeO. But compounds of copper, nickel, cobalt, man-
ganese, zinc, and magnesium are isomorphous with corresponding
iron compounds, and so if FeO 2 should be FeO the corresponding
dioxides of all these metals should really be monoxides.

This sweeping change would involve the halving of a number
of accepted atomic weights; nevertheless, the change was made
by Berzelius in accordance with the principles of isomorphism; and
it was at once ratified by the law of specific heat, which required
the same atomic weight magnitudes for the elements concerned.

The alums which conform to the general formula

M 2 S0 4 .X 2 (S0 4 ) 3 -24H 3
are amongst the best -known isomorphous compounds; and the

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 37

atomic weight of the element X can be determined by an analysis
of its alum. For this purpose it is best to ignite the ammonium
alum, which leaves a residue of the oxide X 2 O 3 .

Thus, if a grm. of the alum leaves b grm. of oxide, the value of
X is calculated from the expression:

a : b = (NH 4 ) 2 S0 4 .X 2 (S0 4 ) 3 -24H,O : X 2 O 3

= [132 + 2 X + 288 + 432] : [2 X + 48].

By this means Lecoq de Boisbaudran, who discovered gallium,
found that 3-1044 grm. of its ammonium alum left on ignition a
residue of 0-5885 grm. of the sesquioxide; whence Ga = 70-1.

v. The Method of the Periodic Law. *

An adequate account of the periodic law is necessary to an
appreciation of its value as a guide to the magnitudes of the
atomic weights of the elements; but this must be deferred to a
later chapter.

It will be sufficient to state here that a natural connection
exists between the properties of an element and its atomic weight,
and therefore that the order of magnitude of the atomic weight of
an element may be judged from a study of the properties of the
element and its compounds. Examples of this use of the periodic
law will be given later.

The application of the foregoing methods of atomic - weight
determination is well illustrated by the case of carbon.

4. The Atomic Weight of Carbon

The determination of the atomic weight of carbon has consisted
of two parts:

i. The determination of the order of magnitude,
ii. The determination of the exact value.

i. Determination of the Order of Magnitude of the Atomic Weight.

Dalton and his contemporaries attributed the value 6 to the
atomic weight of carbon, but this was really only an equivalent
weight. The following is the evidence that the atomic weight is
about 12:

(a) Avogadro's Theory. Never fewer than 12 parts by weight of
carbon are present in a molecular proportion of any of the gaseous
or volatile compounds of this element.

(b) Chemical Displacement. Use might be made of the argu-

38 CHEMICAL THEORY

inent that, for example, hydrogen is displaceable from methane in
four equal fractions, but carbon not fractionally; whence it follows
that the formula for methane is CH 4 and the atomic weight of
carbon 12.

(c) The Law of Specific Heat. Although Dulong and Petit's law
does not apply strictly to an element whose atomic weight is less
than 30, and the specific heats of diamond and graphite differ
widely from each other at ordinary temperatures, at 600 the
specific heats of these two allotropic forms of carbon, which vary
with temperature, become almost constant and equal, and give an
atomic heat of 5-5, if G = 12, a value which is comparable witli
the atomic heats"of analogous elements.

(d) The Law of Isomorphism. The iodides of carbon and
silicon are isomorphous; therefore they are similarly composed, and
the atomic weights of carbon and silicon are in the ratio 12 : 28.

(e) The Periodic Law. With an atomic weight of 12, carbon is
appropriately placed in the periodic table between boron (10-82)
and nitrogen (1401); and is thus the first or " typical" element of
the fourth group. If carbon forfeited its place owing to an altera-
tion in the magnitude of its atomic weight, there is no vacant place
in the periodic table which this element could fill, nor is any element
known which could occupy the place of carbon.

ii. Determination of the Exact Value of the Atomic Weight.

There are two ways in which the atomic weight of carbon has
been determined exactly:

(a) By estimating the densities of its gaseous compounds.

(b) By the combustion of carbon or the analysis of its compounds.

(a) It has already been pointed out that gas or vapour density
is simply related to molecular weight only when Avogadro's theory
is rigidly true. This, however, is never the case; but an "ideal"
density can sometimes be calculated from carefully ascertained
data. This has been done 1 for the three gases: carbon monoxide
(CO), carbon dioxide (CO 2 ), and acetylene (C 2 H 2 ).

CO C0 2 C 2 H 2

Experimental density (O 2 = 1) 0-87495 1-38324 0-8194

" Ideal" density 0-87516 1-37516 0-81331

Molecular weight 28-005 44-005 26-026*

Atomic weight of carbon ... 12-005 12-005 li-005

'Leduc, Ann. Cktm. Phys., 1898'[vii.], 15, 5; 1910 [viii.], 19, 441.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 39

It appears from these figures that the method leaves nothing to
be desired from the point of view of accuracy.

(6) The atomic weight of carbon has been 'determined by
several chemists by burning diamond or carefully purified graphite,
weighing the carbon dioxide produced, and then calculating the
result from the proportion:
Weight of CO 2 : weight of C : : 32 + atomic weight C : atomic weight C.

The following are the results, as originally given, and as cor-
rected by Scott:

Uncorrected. Corrected by Scott. 6

Dumas and Stas 1 11-9975 11-9938

Erdmann and Marchand 2 12- 0093 12 - 0054

Roscoe 3 12-0029 11-9973

Friedel 4 12-0112 12-0056

VanderPlaats 5 12-0031 12-0017

The ignition of organic silver salts, such as the acetate and
tartrate, which leave a residue of pure silver, serves as a means
of estimating the atomic weight of carbon; or the silver may be
estimated electrolytically, as was done by Hardin, 7 with the follow-
ing results, obtained with silver acetate and benzoate respectively:

(1) C 2 H 3 O 2 Ag : Ag - 100 : 64-637
whence atomic weight of carbon 12-000.

(2) C 7 H 6 O 2 Ag : Ag = 100 : 47-125
whence atomic weight of carbon = 12-001.

[Ag = 107-880, H = 1-00762,0 = 16-00.]

The above results are selected from amongst others as typical;
they serve to show the degree of Accuracy which has been attained
in the determination of the atomic weight of carbon. This value
lies between 12-000 and 12-005, and may be taken to be 12-003.

5. Determination of Molecular Weights (in Solution)

The establishment of molecular weights by the determination
of gas and vapour densities has been fully considered in the pre-
ceding pages. By the study, however, of the influence of dissolved

1 Dumas, Pogg. Annalen, 1838, 44, 110.

* Erdmann and Marchand, /. prakl. Chem., 1841, S3, 159.
Roscoe, Compt. rend., 1882, 94, 1180.

* Friedel, Bull. Soc. Ckim., 1884 [ii.], 41, 100.
8 Van der Floats, Compt. rend., 1885, 100, 52.
6 Scott, Trans. Chem. 804., 1897, 71, 550.
'Hardin, J., Amer. Chem. Soc., 1896, 18, 990.

40 CHEMICAL THEORY

substances on the solidifying- and boiling-points of liquids, the
molecular weights of substances in solution in these liquids may be
determined; and it will be appropriate to consider here these newer
methods of molecular- weight determination.

It is well known that salt water freezes at a lower temperature
than fresh water, and that sea ice when melted yields fresh water.
Thus, when a dilute solution of salt in water is cooled, crystals of
pure ice begin to separate from the solution at a temperature a
little below 0. Blagden, in 1788, showed that the depression of the
freezing-point of water by a dissolved salt is directly proportional
to the amount of salt present. The boiling-point of water, on the
other hand, is raised by salt in solution, and the elevation of boiling-
point is directly proportional to the amount of salt dissolved. In
1883-4 F. M. Raoult discovered that not only are the depression
of freezing-point and rise of boiling-point of a solvent proportional
to the number of molecules of a particular substance in solution,
but that equimolecular proportions of different substances have the
same influence on the freezing- and boiling-points.

Raoult's law, which applies equally to freezing- and to boiling-
points of solvents, may be stated thus:

The depression of freezing-point and elevation of boiling-point of
a solvent by a quantity of dissolved substance are directly proportional
to the number of molecules of the substance in solution, and con-
sequently inversely proportional to its molecular weight.

Or, otherwise:

Equimolecular solutions, with the same solvent, have the same
freezing- and boiling-points.

Evidently these facts provide a means of comparing molecular
weights, or of determining them if a substance of known molecular
weight is chosen as a standard of comparison. It should be added
that the extent to which a freezing- or boiling-point is affected
depends also upon the solvent; consequently the first procedure
is to determine the freezing- or boiling -constant (K) for a particular
solvent by the use of a substance of known molecular weight. This
jconstant is tjie number of degrees the freezing-point is lowered or
(boiling-point raised by 1 grm.-molecule of the substance dissolved
in 100 grm. of the solvent.

For instance, 2 grm. of cane sugar dissolved in 100 grm. of
water cause a depression of the freezing-point, A = 011. Since

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 41

the molecular weight of cane sugar (C 12 H 22 11 ) is 342, the freezing-
constant, K, for water, sometimes called the molecular depression, is

0-11 X 342 _ 19

The same quantity of sugar dissolved in the same amount of
water raises the boiling-point of the water 0-030. Therefore
the boiling-constant or molecular elevation, K, for water is:

0-030 X 342 __ R o
2 5 ^-

When the freezing- or boiling -constant K for a solvent is
known, an unknown molecular weight is calculated from observed
data as follows:

Let K = depression or rise caused by 1 grm.-mol. of a substance

in 100 grm. of solvent (known constant).
S = weight of substance taken.
L = weight of solvent taken.
A = observed depression or rise.
M = required molecular weight.

Then, since the observed depression of freezing-point or rise of

boiling-point is directly proportional to the amount of substance

taken, and inversely proportional to the amount of solvent,

A K X S X 100 ,, 100 K8

A = M "XL- r M = "Air"

Practical Methods.

The prime necessity for the experimental determination of
molecular weights of substances in solution is a thermometer which
will indicate accurately hundredth^ of a degree. If this ther-
mometer is to be used both for freezing- and boiling-points, it
would appear necessary for it to have a long range in addition.
Real temperatures, however, have not to be read; only temperature
differences. Consequently a thermometer has been devised by
Beckmann with a range of about six degrees, the scale being
divided into hundredths of a degree, and furnished with a reser-
voir of mercury from which mercury can be added if low tem-
peratures are to be recorded, and into which mercury can be
driven when the instrument is to be used for higher temperatures.
By* the use of this device the same thermometer can be employed
for temperatures near the freezing- as well as the boiling-point
of water or other solvent.

CHEMICAL THEORY

The Cryoscopic Method,

The determination of molecular weights by the cryoseopic
method, that is, by observing the depression of freezing-point, is
carried out in the apparatus of Beckmann shown in the figure.
The tube (A), furnished with a side limb for the introduction of

the substance, is fitted with a cork through
which the thermometer (T) and platinum
stirrer (S) pass. The lower part of this tube
is surrounded by a wider tube (B) which
provides an air jacket between the tube (A)
and the freezing-mixture contained in the
outer vessel (C). This freezing - mixture,
whose temperature should be about 5 below
the freezing-point of the solvent employed, is
also furnished with a stirrer (S 1 ). A weighed
quantity of water, or other solvent, is placed
in the tube (A) and then frozen. Owing to
under-cooling the temperature indicated by
the thermometer falls below the freezing-
point, and then quickly rises again, and be-
comes stationary at that point as soon as ice
separates. When the freezing-point of the
solvent has been indicated on the arbitrary
scale of the thermometer, a weighed quantity
of the substance is introduced and the freez-
ing-point of the solution determined. The
amount of substance added should produce a
depression of about 0-5. The determination
may be repeated after the addition of a
further quantity of substance. The reading
should, however, be taken when a minimum
quantity of the pure solid solvent has separated, so that the
concentration of the solution may not be appreciably increased.
The following are important freezing-constants (K):

Water 18 6; acetic acid 39 ; benzene 50; phenol 73.

EXAMPLE. Successive quantities of 0*317, 0-394, and 0-5152
grm. of a substance were dissolved in 18-054 grm. of benzene,
the depressions of freezing-point being 0-278, 0-348, and 0-452
respectively; what is the molecular weight of the substance? The

Fig. 6

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 43

molecular lowering of the freezing-point of benzene (K) is 50.
(Institute of Chemistry, July, 1902.)

100 KS

M =

i. M
ii. M

iii. M

= 1QQ X 50 X 0-317 =
0-278 X 18-054

= 100 X 50 X 0-394 _
0-348 X 18-054

= 10Q X 5Q X ' 5152
0-452 X 18-054

315-8.
313-6.
315-7.

The Ebulliscopic Method Beckmann's Apparatus.

The tube (A) (fig. 6) em-
ployed in the Beckmann appa-
ratus for determining elevation
of boiling-point resembles that
in which freezing-point deter-
minations are carried out; but,
in addition to the side tube
for the introduction of the sub-
stance, it is provided with an-
other tube (B) fitted with a
reflux condenser for the con-
densation of the vapour arising
from the boiling liquid. In
order to prevent super-heating
of the liquid, and consequent
irregular boiling, a short piece
of stout platinum wire (C) is
fused into the bottom of the
tube, which also contains some
small beads which surround the
lower part of the thermometer
bulb, and serve to break and
distribute the bubbles of vapour
as they rise. In addition to
this, the boiling-tube is sur-
rounded with a wider vessel (D)
packed with some non-conduct-
ing material to prevent loss of
heat by radiation, or sometimes Fig. e

44 CHEMICAL THEORY

with a glass envelope containing the vapour of the boiling solvent
The whole apparatus stands upon a sheet of asbestos (E), below
which the burner for heating is placed.

In carrying out an experiment a weighed quantity of the
solvent is heated until it boils briskly, and its temperature has
become constant. If the condenser is acting efficiently the solvent
should not lose in weight; but about 0-3 grm. should be subtracted
from its weight to allow for the quantity required to wet the
internal walls of the tube and condenser. After the boiling-point
of the solvent has been recorded, the weighed quantity of the
substance is introduced, and a reading again taken when the
temperature has become constant. As in the case of freezing-point
determinations, successive quantities of substance may be added to
the same quantity of solvent, and corresponding readings taken.
If much time elapses between the observations of the boiling-points
of solvent and solution, it is necessary to read the barometer, and
make a correction for change of atmospheric pressure during the
interval.

The Modified Landsberger Apparatus.

A method of determining elevation of boiling-point, introduced
by Sakurai, 1 modified by Landsberger, 2 Walker and Lumsden, 3 and
others, ancl more recently by Turner and Pollard, 4 consists in raising
the solvent to its boiling-point by passing into it the vapour of the
same liquid boiling in another vessel. The vapour condenses, and
its latent heat eventually causes the solvent to boil, although the
boiling-point after the addition of the substance is above that of
the pure solvent. As the amount of the solvent continuously in-
creases by condensation of vapour, it is estimated by weighing
or measuring after condensation has been arrested instead of
before heating is begun.

By this method all possibility of superheating is avoided, and
accurate results are rapidly obtained.

The construction of the apparatus is shown in fig. 7. The vessel
(A), about 16 cm. high and 3 cm. in diameter, is fitted with a two-
holed cork through which pass the thermometer (T) and the delivery

tube (B) by which vapour is conveyed to the bottom of the vessel

i Trans. Chem. Soc., 1802, 61, 994. 2 Btr., 1898, 31, 461.

3 Tram. Chem. Soc., 1898, 73, 502.

* Trans. Chtm. Soc. y 1910, 97, 1184, Proc. Chem. Soc., 1913, 29, 349.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 45

from the liquid boiling in the flask (F).
upper part of the vessel
allows uncondensed va-
pour to pass into the
outer vessel (D), where
it forms a vapour jacket
and is then condensed,
either here or by sub-
sequent cooling after
escaping by the side
tube (E).

The boiling-constants
(K) of important liquids
are: Water 5-2; ethyl
alcohol 11-7; chloroform
39; benzene 27.

EXAMPLE. Turner
found that 1-150 grm.
of diphenylamine
(C 6 H 6 ) 2 NH, dissolved in
42-82 grm. of chloro-
form caused the boiling-
point of the solvent to rise 0-618.
weight of diphenylamine.

A small hole (C) in tho
ft

Fig. 7

K = 39. Find the molecular

M = !2 =
AL

100 X 39 X 1-160 _.

J y. = 169-5.

0-618 X 42-82"
Theory for (C 6 H 5 ) 2 NH = 169-1.

6. Molecular Complexity

The methods and results of determining the molecular weights
of gases and vapours and of substances in solution have been
reviewed in the preceding pages; and it appears that the molecules
of substances in solution are often of the same order of magnitude
as those of the same substances in the state of vapour. For example,
ferric chloride in a state of vapour at about 750 consists of mole-
cules represented by the formula FeCl 3 , and the elevation of the
bofiing-point of ether or alcohol by dissolved ferric chloride points
to the same molecular formula. The reason for this identity of
molecular state is to be found in the fact that the vaporous state

46 CHEMICAL THEORY

and the state of solution are analogous to each other, and that the
process of vaporization of a solid or liquid, with the consequent
distribution of its molecules through space, resembles the process of
solution of the same substance, and the distribution of its molecules
throughout the solvent.

Yet the molecular state of a dissolved substance depends some-
times upon the liquid in which it is dissolved. Hydrogen chloride,
"for example, forms molecules when dissolved in benzene and nitro-
benzene which may be as much as five times as great as the gaseous
molecule; its molecules are then said to be associated. With regard
to liquids themselves, there is good reason to believe that their
molecules are often associated. Consider water, for example.
Water is the first of the series of four hydrides: H 2 O, H 2 S, H 2 Se,
H 2 Te; three of these are gases; why, therefore, is water a liquid?
Since the atomic w r eight of oxygen is the least of the atomic weights
of the four elements combined with hydrogen in this series, and
volatility diminishes from H 2 S to H 2 Te, water would be expected
to be the most instead of the least volatile of the four hydrides.
The reason water is a liquid at atmospheric temperature must be
that it forms complex molecules (H 2 O) n . Steam even appears to
contain a very small proportion of molecules, which are regarded as
double, since its density is a little greater near the point of con-
densation than corresponds with the simple formula H 2 0. Liquid
water undoubtedly consists of associated molecules, e.g. (H 2 0) 2 and
(H 2 O) 3 , whilst ice is believed to be (H 2 O) 3 only. It is noteworthy
that hydrogen fluoride, which follows water in the periodic classifi-
cation, also contains associated molecules, and has an anomalous
boiling-point.

Regarding benzene, C 6 H 6 , there is evidence that near its point of
condensation the saturated vapour begins to contain double mole-
cules, (C 6 H 6 ) 2 , and that liquid benzene consists entirely of these
molecules, until near its freezing-point, when (C 6 H 6 ) 4 molecules
begin to appear, and increase in number until solid separates,
which consists wholly of (C 6 H 6 ) 4 molecules.

It is remarkable that the association of various molecules is
promoted by- the entire absence of water. Thus benzene, which
usually boils at 80, was found by H. B. Baker 1 to boil at 106 after
being dried over phosphoric oxide for eight years. This rLse of
boiling-point must, no doubt, be attributed to molecular association.

1 Chcm. Soc. Trans., 1922, ll t 570.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 47

Concerning the molecular state of solids little or nothing has
been known until recently; but the examination of the X-ray
spectra produced by solids is throwing much lighi on their mole-
cular structure. The nature of the problem presented by solids
may be realized by considering a simple example. Carbon and
silicon are related elements; yet carbon when it burns forms a gas
and silicon a solid. Thus the molecules of C0 2 are not associated,
but those of SiO 2 are, if indeed Si0 2 molecules can be said to exist
at all. What is the reason for this difference? The difference
nuiy be expressed, [though not explained, by saying that any field
of influence outside the C0 2 molecule, by which other molecules of
the same kind might be attracted, and associated, is very limited,
or else the attractive force exerted therein is very small; whereas
in the case of the Si0 2 molecule the external field of attractive
force is considerable in extent or strength or both.

The difference between the two cases is referable to the difference
between the carbon and silicon atoms, and must depend ultimately
upon the different structures of these atoms. Meanwhile it may be
stated that the " crystal unit ", i.e. the smallest unit that takes part
in crystal growth of silica, is (SiO 2 ) 3 .

With regard to the formula) to be applied to liquid and solid
compounds there is much difficulty if real molecules are to be
represented. Indeed it can hardly be asserted that molecules,
NaCl, of common salt exist either in the solid state or in solution.
It is not necessary, however, to have molecular formulae in order to
represent chemical reactions, since the simplest formulas represent-
ing the inherent properties of compounds are sufficient for all
ordinary purposes. The formula CaC0 3l for example, serves for
chalk; it would be interesting to know what is the molecular or
crystal unit of this compound, but such knowledge is not necessary
for the representation of its common reactions.

In some cases, however, the smallest empirical formula would
be untrue; e.g. hyponitrous acid is H 2 N 2 O 2 and not HNO; and
benzene is C 6 H 6 and not CH nor C 2 H 2 .

7. The Molecular Compositions of Compound Gases

The evidence on which the molecular formula H 2 for water is
based has already been considered. There are a number of com-
pound gases whose molecular formulae may be established by the

48 CHEMICAL THEORY

application of the principles set forth in this chapter; and these
will now be dealt with.

It has already been shown that the molecules of hydrogen,
chlorine, and oxygen are diatomic. This follows, it will be re-
membered, from the fact that the hydrogen chloride formed from
equal volumes of hydrogen and chlorine occupies twice the volume
of each separate gas, and that steam occupies twice the volume of
its constituent oxygen at the same temperature and pressure. By
an extension of the principle here employed the number of atoms
of a gaseous element within the molecule of a compound gas may
always be determined.

Thus, since it can be shown that 2 volumes of ammonia gas
yield when decomposed 3 volumes of hydrogen and 1 volume
of nitrogen, it follows, provided the nitrogen molecule is diatomic,
that ammonia must be represented by the formula NH 3 ; for the
molecular change on the decomposition of ammonia is:

2 niols. ammonia yield 3 H 2 + N.j,
consequently 2 NH 3 = 3 H.> + N 2 .

The argument may be put in another way. Since the volume
of the ammonia is to that of the hydrogen as 2 : 3, the atomic con-
centration of hydrogen in ammonia is to that in free hydrogen as
3:2; and since the volume of ammonia is to that of nitrogen as
2:1, the atomic concentration of nitrogen in ammonia is to that
in free nitrogen as 1:2; whence the formula NH 3 follows.

In the case of a gas containing a solid element, such as sul-
phurous anhydride, the additional estimation of the density of the
gas suffices to show how many atoms of the solid element it
contains, provided the atomic weight of this element is known.
Thus, (a) the gas produced by burning sulphur in oxygen measures
the same volume as the oxygen; therefore the molecule of this gas
contains 2 atoms of oxygen; (6) the density of the gas is 32, and
its molecular weight consequently 04, whilst the weight of oxygen
within its molecule is 32, and the atomic weight of sulphur is 32;
therefore it follows that the molecule of the gas contains 1 atom
of sulphur, and that its molecular formula is SO 2 .

The follchving statements epitomize the evidence for the mole-
cular formula of a number of the best-known gases:

Hydrogen Chloride.

That 1 volume hydrogen + i volume chlorine give 2 volumes

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 49

hydrogen chloride is fundamental to the molecular theory. The
following facts suffice to prove this relation:

(a) Electrolysis of an aqueous solution of hydrogen chloride
under suitable conditions yields equal volumes of hydrogen and
chlorine.

(6) Sodium amalgam removes the chlorine from hydrogen
chloride gas, and the remaining hydrogen occupies half the volume
of the hydrogen chloride.

Water and Steam.

(a) Electrolysis of acidified water yields hydrogen and oxygen
in the proportion of 2 volumes of the former to 1 of the latter.

(6) When a volume of electrolytic gas, i.e. a mixture of 2
volumes of hydrogen with 1 volume of oxygen is exploded in
a eudiometer kept at a temperature above the boiling-point of
water, the volume of the resulting steam is two-thirds the volume
of the mixed gases. Therefore

2 vol hydrogen + 1 vol. oxygen yield 2 vol. steam.
Carbonic Anhydride.

When carbon is burnt in oxygen gas the volume of the gas
remains unaltered. Therefore a molecule of the gaseous product
contains 2 atoms of oxygen (O 2 = 32).

The density of carbonic anhydride is 22 ; therefore its molecular
weight is 44. Within this molecular proportion are 32 parts
(O 2 ) of oxygen, and therefore 12 of carbon. But 12 is the atomic
weight of carbon. Therefore carbonic anhydride is CO 2 , and is
rightly called carbon dioxide.

Sulphurous Anhydride.

When sulphur is burnt in oxygen the volume of the gaseous
product is the same as that of the oxygen. The density of sul-
phurous anhydride is 32, and its molecular weight 64. The atomic
weight of sulphur is 32; therefore, by the same argument as applies
to carbon dioxide, sulphurous anhydride is sulphur dioxide, SO 2 .

Hydrogen Sulphide.

When hydrogen sulphide gas, confined over mercury, is decom-
posed by electric sparks, or when its sulphur is removed by means
of tin heated in the gas and so converted into sulphide, the volume

(D60) 5

50 CHEMICAL THEORY

of the remaining hydrogen is equal to the volume of the original
hydrogen sulphide, whose formula is consequently H 2 S n . That
n = 1 is proved by the fact that the gas density is 17 and
molecular weight 34; for of this 32 parts must be sulphur, and 32
is the atomic weight of sulphur. Thus the formula for hydrogen
sulphide is proved to be ILS.

Nitrous Oxide.

Potassium, sodium, copper, and other metals remove the oxygen
from nitrous oxide when heated in the gas, leaving nitrogen.
Thero is some risk of nitrite being produced if the two former
metals are heated too strongly in the gas, but strongly heated
copper removes only the oxygen, and loaves all the nitrogen in a
pure state. By this means it may be shown that nitrous oxide
contains its own volume of nitrogen, and therefore that its mole-
cule contains 2 atoms of this element. The density of nitrous
oxide is 22, and its molecular weight is 44, and this weight contains
28 parts (Nj) of nitrogen, and therefore 16 parts of oxygen. Since
16 is the atomic weight of oxygen the molecule of nitrous oxide
contains 1 atom of this element, and therefore the molecular
formula for the gas is N 2 0.

The same conclusion is reached by mixing nitrous oxide with
its own volume of hydrogen and exploding the mixture. After
condensation of the steam pure nitrogen remains equal in volume
to the nitrous oxide. Thus it is shown, not only that nitrous
oxide contains its own volume of nitrogen, but that the oxygen it
contains would occupy half that volume, since it combines with
a volume of hydrogen equal to that of the nitrous oxide. These
facts are sufficient to establish the formula N 2 for nitrous oxide.

Nitric Oxide.

If potassium is heated in nitric oxide the vigorous combustion
which takes place results in the formation of nitrite and nitrate;
but a spiral of iron wire heated electrically removes all the oxygen
from the gas without combining with the nitrogen, and the residual
nitrogen then occupies half the volume of the original nitric oxide.
This proves that a molecule of nitric oxide contains 1 atom o
nitrogen (N = 14). The density of nitric oxide is 15, and, Miice
its molecular weight is 30, the molecule contains 1 atom of
oxygen (O = 16), and the molecular formula is NO.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 51

Ammonia.

When ammonia solution is dropped into chlorine gas, hydrogen
chloride is formed, and nitrogen set free. The experiment may be
carried out in a long graduated tube, sealed at one end and pro-
vided at the other end with a cork furnished with a tap funnel.
Ammonia solution is passed through the funnel into the chlorine,
and the reaction is accompanied by a greenish flame and fumes
of ammonium chloride. After the ammonia has been added in
excess, dilute sulphuric acid is introduced to combine with the
excess of ammonia, after which water is allowed to enter until
the gas in the tube is at atmospheric pressure, when the flow of
water ceases. Then it is found that the gas, which is nitrogen,
fills one-third of the tube. Since hydrogen and chlorine combine
in equal volumes to form hydrogen chloride, the hydrogen of the
ammonia from which the hydrogen has been liberated would have
occupied three times the volume of this nitrogen. This shows
that ammonia, when decomposed, yields 1 volume of nitrogen
to 3 of hydrogen; but since the volume of ammonia gas which
is thus decomposed is unknown, all that this experiment reveals
is that the molecule of ammonia is (NH 3 ) n .

The relation between the volume of ammonia and the volumes
of its decomposition products may be determined by confining a
measured volume of the gas over mercury and passing electric
sparks through it until expansion ceases. The gas will then
have been decomposed into a mixture of hydrogen and nitrogen
which will occupy twice the volume of the ammonia. That this
mixture consists of 3 volumes of hydrogen and 1 volume of
nitrogen may be shown by adding excess of oxygen and ex-
ploding the mixture.

Thus for example:

Volume of ammonia = 10*0 c. c.

Volume of nitrogen + hydrogen after sparking = 20-0

Volume after addition of oxygen = 75

Volume after explosion =52-5

Thus 22-5 c. c. of gas have disappeared, of which 15*0 c. c. must
have been hydrogen. So it follows that 10-0 c. c. of ammonia were
decomposed by electric sparks into 15 c. c. of hydrogen and 5 c. c.
of nitrogen; and, as shown before, this proves the molecular
formula NH 3 for ammonia.

52 CHEMICAL THEORY

Phosphine.

The case of phosphine (.litters from that of ammonia because,
when the gas is decomposed by electric sparks, the liberated phos-
phorus remains as a solid whose volume is negligible. Conse-
quently, the proportion of phosphorus in the molecule must be
discovered by density determination as in the case of sulphur
dioxide, hydrogen sulphide, &c.

Two volumes of phosphine, decomposed by electric sparks,
yield 3 volumes of hydrogen. Therefore a molecule of the gas
contains 3 atoms of hydrogen.

The density of phosphine is 17, and its molecular weight 34.
Consequently, the proportion of phosphorus within the molecular
proportion of phosphine is 31. But 31 is the atomic weight of
phosphorus. Therefore the molecule of phosphine contains 1
atom of phosphorus, and so its formula is PH 3 .

Carbon Monoxide.

Carbon monoxide can be converted into carbon dioxide by
exploding it with oxygen, when it is found that 2 volumes of the
gas combine with 1 volume of oxygen to form 2 volumes of carbon
dioxide; or, since the molecular formula) C0 2 and O 2 are known, in
the equation,

2 C x O y + O 2 = 2 C0 2 ,

x and y both = 1, so that the molecular formula CO is proved.
This conclusion is confirmed by the density of the gas, which is
14, whence the molecular weight is 28; and C = 12, O = 16, so
that CO = 28.

Methane, Ethylene, and other Hydrocarbons.

Tf a certain volume of a hydrocarbon is exploded with a
known volume of oxygen used in excess, the resulting moist gas,
measured at atmospheric temperature and pressure, consists of
carbon dioxide mixed with unused oxygen. The volume of
carbon dioxide formed is estimated by absorbing this gas in
sodium hydroxide solution, and the total volume of oxygen used,
part of which has produced carbon dioxide, and part water, is
shown by the difference between the original and the remaining
volume of oxygen. These data are sufficient to establish the
molecular formula of the hydrocarbon.

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 53

For, consider the gaseous hydrocarbon C x Hj. The result of
its explosion with oxygen is represented by the equation

C,H 7 + (x + f )0 2 = xCO. + ?H 2 0.

N 4 / ^

The volume of steam formed and condensed is not measured;
but when the volume of carbon dioxide, referred to that of the
hydrocarbon as unity, which is x 9 has been ascertained, the value
of y is found by subtracting this from the total volume of oxygen
used, referred to the same standard, and multiplying the remainder
by 4.

When x and y are found, the formula of the hydrocarbon
is settled. Vapour density will confirm the formula, but is not
necessary to establish it.

Methane.

When a mixture of 10 c. c. of methane with 30 c. c. of oxygen
is exploded, the resulting gas, measured at the same temperature
and pressure, is a mixture of 10 c. c. of carbon dioxide and 10 c. c.
oi* oxygen.

Thus 1 volume methane requires for combustion 2 volumes
oxygen, and yields 1 volume carbon dioxide.

So in the equation

C x H y + (x + 1)0 2 = arC0 2 + |H,O, x = 1 and | = 1 ;

therefore the formula for methane is CH 4 ; or otherwise, because
the volume of the carbon dioxide produced is equal to the volume
of the methane, a moleculQ of the latter contains 1 atom of carbon;
and, because the volume of the oxygen required to burn the
'hydrogen of methane is equal to the volume of the methane,
the atomic concentration of hydrogen in the methane molecule
is twice what it is in the free hydrogen molecule; i.e. there are
4 atoms of hydrogen in methane. Thus the molecular formula
for methane is CH 4 .

Ethylene. t

When a mixture of 10 c. c. of ethylene with 40 c. c. of oxygen
is exploded, the resulting gas, measured at the same temperature
and pressure, is a mixture of 20 c. c. of carbon dioxide and 10 c. c.
of oxygen.

54 CHEMICAL THEORY

Thus 1 volume ethylene requires for combustion 3 volumes
oxygen and yields 2 volumes carbon dioxide.
So in the equation

C\H, + (* + ^)0 2 = *C0 2 + |H 2 0, x = 2 and J = 1 ;

consequently the formula for ethylene is C 2 H 4 .

Or, to employ the alternative argument, since the volume of
the carbon dioxide produced is twice the volume of the ethylene,
a molecule of this hydrocarbon contains 2 atoms of carbon; and
since the volume of oxygen required to burn the hydrogen of
ethylene is equal to the volume of the ethylene, this hydrocarbon
contains 4 atoms of hydrogen. Thus, again, the molecular formula
for ethylene is C, 2 H 4 .

In a similar way the molecular formula of any gaseous hydro-
carbon may be established.

SUMMARY

EQUIVALENT WEIGHT. The equivalent weight of an element
is that weight of it which combines with, or displaces from com-
bination, unit weight of a standard element.

ATOMIC WEIGHT. The atomic weight of an element is the
ratio between the weight of its atom and that of the atom of
a standard element. The standard is: = 1600.

DETERMINATION OF ATOMIC WEIGHT:

(a) Exact estimation of chemical equivalent.

(b) Decision as to order of magnitude.

Guiding principles: i. Avogadro's theory.

ii. Chemical displacement,
iii. Law of specific heats,
iv. Law of isomorphism.
v. Periodic law.

PRINCIPLE OF CHEMICAL DISPLACEMENT. When l/n th of the
proportion 6f a constituent element in a chemical compound can
be displaced by another element, a molecule of the compound
contains at least n atoms of that element.

LAW OF SPECIFIC HEATS: DULONG AND PETIT'S LAW. Tho
specific heats of the solid elements are in the inverse ratio of

EQUIVALENT, ATOMIC, AND MOLECULAR WEIGHTS 55

their atomic weights; or the atoms of the solid elements have
the same capacity for heat.

Specific heat X atomic weight = 6-4(approx.)

or atomic weight = * . -.
specific heat

LAW OF ISOMORPHISM: MITSCHERLICH'S LAW. The molecules
of isomorphous substances contain equal numbers of atoms, which
when not identical are analogous.

DETERMINATION OF MOLECULAR WEIGHTS: RAOULT'S LAW.
The depression of freezing-point, and elevation of boiling-point of a
solvent caused by any dissolved substance are directly proportional
to the number of molecules of the substance in solution, and
consequently, inversely proportional to its molecular weight; or
equimolecular solutions, with the same solvent, have the same
freezing- and boiling-points.

CHAPTER III

OLDER VIEWS OF VALENCY AND CHEMICAL
CONSTITUTION

In the preceding pages the experimental foundations of the
atomic and molecular theories, as these were laid by the chemists
and physicists of the nineteenth century, have been considered;
and it has been found possible to exhibit and expound these
without reference to modern conceptions of the atom; this is
because the phenomena concerned have been superficial and have
not dealt with the inter-relations of the atoms themselves in
chemical compounds.

When, however, the subjects of valency and chemical constitu-
tion are approached the case is otherwise. It is impossible at the
present time to consider these subjects adequately without bringing
into view the atom as it appears in the light of to-day's knowledge.
Moreover, the tide of this new knowledge is so powerful that
much that was considered sound and stable has been broken by its
flood; so that the first task of the chemical philosopher is to
strengthen what remains of the harbour of his thought, whilst the
flotsam disappears.

Consequently this chapter on " Older Views of Valency and
Chemical Constitution" is historical; touching lightly the great
subjects with which it deals, it brings chemical knowledge up to
the boundary of the new domain, and leaves for a further chapter
the task of exploration. It will be enough for the present purpose
if what was temporary and must disappear can be distinguished
from what is permanently useful. If this purpose is achieved, if
impedimenta are dropped, and only useful tools and weapons are
retained, there is hope that in the new field valuable possessions
may be acquired.

It has already been seen, with regard to an element, that
Atomic weight = n X equivalent weight ;

and that n is the valency or atomic value of the element.

VALENCY AND CHEMICAL CONSTITUTION 57

The doctrine of valency, in the form in which it was held during
the latter part of the nineteenth century, was a matter of slow
development. It arose during the growth of organic chemistry,
because of the need of a theory of structure in systematizing this
branch of the science. Chemists developed the habit of referring
various organic compounds to a few simple inorganic types, and this
idea proved fertile, for it stimulated research, and led to important
discoveries. These simple types were:

H H H) U H

Hf 01 f Hf v ttr H

O.J H

-C.

It cannot be said that the conception of valency was definitely
contained in these types, yet it was not far off.

The chief exponent of the theory of types was Gerhardt; but
it was E. Frankland who first introduced the idea of saturation
capacity or valency. Frankland showed that whilst 1 atom of tin
was capable of combining with two atoms of oxygen to form the
dioxide Sn(X> a molecule of the compound tin diethyl, Sn(C 2 H 5 ) 2 ,
or SnEt. 2 , could combine with only 1 oxygen atom forming the
compound SnEt^O. Thus it appeared that the tin atom had a
certain saturation capacity, that it could combine with not more
than 2 atoms of oxygen or their equivalent; and that its power of
combining with oxygen was diminished by the extent to which it
was already combined with other atoms or groups of atoms.

The principle was further illustrated by Frankland by reference
to such compounds as

in which tho atoms of nitrogen and phosphorus combined with
.3 atoms of hydrogen or halogen; and by Kekule, who showed that
the carbon atom could combine with four other atoms, as in the
compounds CH CHC

The phenomenon here illustrated is now called valency, about
which the following statement may be made:

The valency of an element indicates the number of other atoms
with which one of its atoms can directly combine.

. An atom may be uni-, bi-, ter-, quadri-, quinque-, sex-, sept-,
or even octa-valent; 1 equivalent terms are monad, dyad, triad,

1 The Greek prefixes mono-, di-, tri>, &c.j which, when attached to valent, make hybrid
words, are now being abandoned.

68 CHEMICAL THEORY

tetrad, &c. In the compounds cited above the nitrogen and phos-
phorus atoms are tervalent, and the carbon atom is quadrivalent;
whilst the hydrogen, chlorine, and iodine atoms are univalent.
Hydrogen is never more than univalent, and therefore its atom
is chosen as the standard of valency; chlorine is univalent with
regard to hydrogen and metals, and, indeed, probably to all ele-
ments except oxygen; it may therefore replace hydrogen as a
standard.

The following hydrides exhibit the valency of a number of
elements:

Valency 1

OH 2

NH 3

CH 4

C1H

SH 2

PH 3

SiH 4

BrH

AsH 3

SbH 3

and the following halides 1 illustrate valency more extensively:

Valency 1

NaCl

OC1 2

BC1,

CC1,

PF 5

SF B

Zn01 2

PC1 3

SiCl 4

AsF 6

TeF e

AgCl

HgCl 2

A1C1 3

SnCl 4

SbF 6

UF,

8
OsF,

Oxygen is here shown to be bivalent. It is seldom other than
this; and if oxygen is bivalent, the large number of oxides that
exist may be classified to show valency, on the assumption that the
valency of an element is equal to twice the number of oxygen
atoms with which one of its atoms combines.
Valency

T 1

Na,>O

MgO

B 2 3

CO 2

N 2 6

SO,

C1 2 O 7

OsO 4

KoO

CaO

A1 2 3

SiO 2

P 2 6

Cr0 3

dA)

RuO 4

Ag a O

ZnO

Fe a O s

PbO 2

Bi 2 6

UO 3

Mn 2 O r

These oxides are in a different category from that of the
foregoing hydrides and halides. In those the molecular formulae
have in all cases been established by vapour density or other
measurements, and the valency of the element concerned is directly
indicated by the number of attached hydrogen or halogen atoms.
The formulae for these oxides, however, are seldom truly molecular.
Phosphoric oxide, for example, is (P 2 O 6 ) 2 even as vapour, and the
crystal unit of silica is (SiO 2 ) 3 ; but little, as a rule, can be said
about the molecules of solids. Moreover, according to the above
statement, valency as a property cannot strictly be judged from
oxides at all, for these compounds do not exhibit, attached to a nuclear

1 Halite = fluoride, chloride, bronVide, or iodide.

VALENCY AND CHEMICAL CONSTITUTION 5<J

atom, a number of peripheral atoms corresponding to its valency.
Nevertheless, there is good reason to regard the valencies indicated
by oxides such as those in the table to be correct.

The establishment of the idea of valency was soon followed by
a device by which the facts of atomic union were represented
graphically.

Bonds were introduced by Couper to show the joining together
of the atoms in the following way:

H H H

Cl H, H-O H, N^ H-C-H.

I I

H H

Thus graphic or constitutional formulae were constructed, with
bonds to show units of valency, or units of affinity, which they
might be called, if they are thought of as standing for the forces
by which the atoms are united.

Oxides containing bivalent oxygen have been represented by
graphic formulae, such as the following:

Mg=O, B O~F>, OCO, ^N-O-Nrf , O *S= O,

\ n / 0^ N ^O ||

or
O-B-0-B-O,

00 O

O-C1- 0-Cl=0, O=0s=0.

II II li

GO O

Formulae such as these are chiefly of historic interest, for they
have to be reconsidered carefully in the light of modern knowledge
and theory concerning the atom. It will appear later that Couper's
bonds ought not to be used indiscriminately or similarly for all
these compounds; e.g. while they are appropriate in the case of
carbon dioxide they are hardly proper in the case of magnesium
oxide, since the mode of chemical union in this case seems to be
different from that in the case of the gaseous oxide*

When bonds were first employed it was thought that the atoms
in all compounds were united together in a similar way; and all
that has been done in constructing these formulae has been to
arrange the atomic symbols in relation to one another so as to

60 CHEMICAL THEORY

represent known or supposed i'acts of chemical constitution, and
then to join these symbols by bonds to represent the supposed
acting valencies of each atom. How artificial such formula) are is
seen by comparing the two formulas given for B 2 O 3 . Each satisfies
the requirement that boron be tervalent and oxygen bivalent,
yet both cannot truly represent the constitution of this oxide.

Is, then, the writing of graphic formulae merely an interesting
geometrical exercise based on the facts of valency alone? Con-
sider, for example, a substance with the formula C 3 H 6 0. Since
carbon is quadrivalent, oxygen bivalent, and hydrogen univalent,
two graphic formula) are possible for this substance:

H O H H H H

H C-C-C-H and H-C-C-O-O.

I I II

H (i) H H H (ii)

Does it matter which formula is adopted? The answer is that
two quite different substances are known, both of which are
C 3 H 6 O; and one of which, acetone, certainly has the constitution
(i), whilst the other, propaldehyde, as certainly possesses the con-
stitution represented by (ii).

Thus graphic formulae are constitutional formulae, and only so
far as they represent the ascertained constitution of compounds are
they valid; therefore the construction of graphic formulae for com-
pounds which have not been definitely proved to have a certain
constitution is to be deprecated.

A fundamental question connected with this subject was raised
by comparing magnesium oxide with carbon dioxide. This question
has lately assumed great prominence, and may be illustrated more
clearly by considering the two chlorides NaCl and CC1 4 . Until
recently these compounds have been formulated thus:

Na Cl; Cl C Cl;
Cl

there are, however, great differences between the two chlorides,
both in physical and chemical properties, which suggest different
modes of union of their constituent atoms. Sodium chloride is a
solid whose separate atoms 1 of sodium and chlorine are arranged

1 Or, more accurately, ions.

VALENCY AND CHEMICAL CONSTITUTION

Fig. 8

like cubes 1 packed close together (fig. 8),
until cubic crystals of various sizes are
produced; whilst carbon tetrachloride con-
sists of self-contained CC1 4 molecules which
easily separate from one another, and exist
apart in the state of vapour. Further,
when sodium chloride dissolves in water
its chlorine becomes reactive in a manner
in which the chlorine of carbon tetrachloride
never becomes reactive; i.e. it shows the reaction of chloride with
silver nitrate which carbon tetrachloride fails to show. Can it
reasonably be maintained that the same kind of bond unites chlorine
with sodium as unites this element with carbon? It cannot; and
consequently there appear to be two kinds of valency; the kind of
valency which exists in carbon tetrachloride can be properly repre-
sented by "bonds", whilst the kind which unites sodium with
chlorine cannot. Therefore the use of Couper's bonds in expressing
the constitution of many inorganic compounds is being discontinued.
The modern view regarding the mode of union of the elements in
these compounds will appear later.

Variability of Valency.

Early in the development of the theory of valency the question
arose whether valency is a fixed and inherent property of an atom,
like its mass, or whether it can vary under varying circumstances.
Kekul^, who showed the quadrivalency of carbon, believed valency
to be unalterable; and the study of carbon compounds alone ap-
peared to justify Kekule's opinion. The following compounds were
cited by Kekule to illustrate the constant quadrivalency of carbon:

H H Cl Cl

'III Cl

H C-TT, H-C Cl, Cl C Cl, II C Cl,

i ! I I

H H Cl Cl

O=C=O, S=0=S, H-C-N.

Frarikland, on the other hand, observed that nitrogen formed not
only NH 3 , in which the element is evidently tervalent, but also
NH 4 C1, in which it was apparently quinquevalenb Thus was
expressed the idea of a maximum potential valency, and an actual
valency, exercised in specific compounds, which might be less than
this. And it was observed that the actual valency frequently fell

1 It must not, however, be supposed that the atoms are cubical in shape.

62 CHEMICAL THEORY

short of the potential valency by two units, as, for example, in the
pairs of compounds

NH 3 , NH 4 Ci; P 2 O 3 , P 2 O 6 ; SO 2 , SO 3 ; SnCl* SnCl 4 ;

so it was supposed that when valency decreased from the maximum
it was always by two units, and that consequently the valency
of an element remained either odd or even. Then it was thought
that the two valencies which remained disengaged in the lower
compounds satisfied each other, so that no valencies remained free.

There are, however, some notable exceptions to this supposed
rule, and it cannot be regarded as a natural law. Examples of
these exceptions are shown in the sets of compounds

NO, N 2 3 , N0 2 , N 2 6 ; CIO* C1 2 O 7 ; IO 2 , I 2 O 6 ; FeCl 25 FeCl 3 ;
InCl, InCL, lnC! 3 ; WC1 5 , WC1 6 .

The lower compounds are unsaturated, and combine with oxygen,
chlorine, &c., to form higher compounds.

When rise of temperature causes dissociation it thereby causes
the acting valency of the nuclear atom or atoms of a compound
to diminish. Thus when ammonium chloride, NH 4 C1, in which the
nitrogen atom is regarded as quinquevalent, dissociates into ammonia
and hydrogen chloride, the nitrogen atom becomes in consequence
tervalent. Similarly tungsten hexachloride, WC1 6 , dissociates at high
temperature into the pentachloride WC1 6 and chlorine. Occasion-
ally dissociation involves the halving of molecules, as the following

examples show :

N 2 4 N0 2 + N0 2

Fe 2 Cl 4 FeCL + FeClo

A1 2 C1 6 A1C1 3 + A1C1 3 .

Presumably this dissociation involves a reduction in operative
valency, since such valency is necessary to hold together the two
parts of the double molecule, but becomes inoperative on dissociation.
It may be added that, unless there is reason to the contrary,
association into double molecules, such as those represented above,
is supposed to be effected by means of 1 unit of valency. Mole-
cular association in liquids and solids must also be accounted for
by the exercise of additional valencies. Thus liquid water contains
double moledules, or molecules of even higher complexity; and the
existence of these complex molecules is accounted for by assuming
oxygen to be quadrivalent, thus:

2>0=0<g;

VALENCY AND CHEMICAL CONSTITUTION

since oxygen is known to be quadrivalent in some other compounds.
The existence of double salts / and salts with water of crystal-
lization cannot be explained by the narrower conceptions of valency.
Consider, for example, potassium alum, K 2 S0 4 'A1 2 (S0 4 ) 3 24H 2 O.
The constitutional formulae for potassium and aluminium sulphates
have been constructed thus:

/O SOj-C

Al 0-S0 2 -O-

A1 ;

and

but it is difficult to see how these formula are to be united together,
and 24 molecules of water to be incorporated in the scheme as well.

At one time it was customary to describe such compounds as
" molecular " rather than " atomic ", but such a distinction is no
longer regarded as valid, and several theories have been proposed
to account for the constitution of these compounds on the ground
that auxiliary or latent valencies come into play in their formation.
These theories cannot, however, be considered here.

The variation of valency with the kind of compound formed
has been illustrated in the lists of hydrides, halides, and oxides
already given. Thus it appears that whilst the valency of an
element towards oxygen and the halogens 1 may rise as high as
8, valency towards hydrogen is never greater than 4; no single
atom is known to combine with more than 4 hydrogen atoms.
Now hydrogen and oxygen are reciprocally related, and it is a
noteworthy fact that as the valency for hydrogen diminishes in
a series of elements with increasing atomic weight, the valency
for oxygen correspondingly increases, and the sum of the oxygen
and hydrogen valencies remains equal to 8. This is shown in the
following compounds, although fluorine and bromine fail to form
oxides, and iodine is not known to form the oxide I 2 O r ; moreover,
tin has recently been shown to form an unstable hydride, though
this has not been proved to be SnH 4 .

CH 4 CO,
NH 3 NA
OH 2
FH -

SiH 4 SiO 2
PH S P,0 6
SH. SO 3
01 H CIA

OeH 4 GeO 2
AsH 3 As 2 O 6
SeH, ScO 3
BrH

8nHt SnO,
SbH 3 Sb 2 O 5
TeH, TeO,
IH (1,0,)

These phenomena have a deep significance, which will appear later.
With regard to valency for the halogens, it must be noted that

i The halogen elements are fluorine, chlorine, bromine, iodine.

64 CHEMICAL THEORY

as a rule halides are not so stable as the corresponding oxides. For
example, NC1 3 is so unstable as to be highly explosive, whilst N 2 O S
does not split off oxygen; PC1 6 dissociates into PC1 3 and C1 2 , whilst
P 2 O 6 is stable; S0 2 may be united with oxygen to form S0 3 , whilst
SC1 4 , formed below C., easily loses chlorine.

Fluorides, however, are much more stable than the other halides:
PF 6 and SF 6 are stable gases, and the existence of OsFg, 1 in addition
to OsF 6 and OsF 4 , shows a valency of 8 towards a halogen.

The Double Bond in Carbon Compounds.

Consider the two hydrocarbons ethane, C 2 H 6 , and ethylene, C 2 H 4 .
The former is saturated, the latter is unsaturated; that is to say, it
is capable of combining with 2 more hydrogen atoms or their
equivalent. This state of unsaturation of ethylene is represented
by a double bond, the graphic formulae for the two compounds
being

Ethane. Ethylene.

H H H H

H C C H and H-C=C H.

The question may be asked whether the double bond is simply
employed to keep up the appearance of the quadrivalency of carbon,
or whether it has any real meaning; whether, indeed, carbon is not
really tervalent in ethylene, so that the formula might as well be

H H

H-C C H.

This question may be answered in the negative for several reasons.

First, no such compound as CH 3 CH 2 is known, in which
one atom of carbon is quadrivalent, whilst the other is tervalent; so
that both atoms must be either saturated or unsaturated. Here, at
least, the idea that the two unsaturated atoms in ethylene satisfy
one another appears justified; and the double bond expresses their
mutual dependence.

Further, the double bond between carbon atoms, the ethylene
linkage, as ft is called, expresses something more than unsaturation;
for the nature of this union differs from that represented by the
single linkage. It is weaker than the single linkage, for when

Tschirch, Jfer., 1913, 46, 929.

VALENCY AND CHEMICAL CONSTITUTION C5

a compound contains a chain of carbon atoms in which there is
a double linkage, this is the point at which the chain breaks when
the compound comes under disruptive influence. The fact that
the double is weaker than the single linkage shows that no
mechanical significance must be attached to bonds.

There is still a third characteristic of the double bond, which,
however, can only be made clear by the study of the stereo-
chemistry of carbon compounds.

" Chemistry in Space. 1 *

How far, it may be asked, is the graphic formula

H C H
II

supposed to represent the real configuration of the molecule of this
simple hydrocarbon, methane? The answer may at once be given
that it is probably an imperfect representation of the truth, because
it is a flat formula, a formula in two dimensions, whereas matter
exists in three dimensions; the formula has length and breadth,
but the molecule of methane has thickness as well as length and
breadth. Moreover, the adequacy of the formula may be tested
in a very 'simple way. The formula suggests that there might be
two methylene chlorides, CH 2 C1 2 ,

Cl Cl

I I

H C II and H C Cl,

A A

in which the two chlorine atoms are respectively opposite and

adjacent to each other. Two such chlorides

do not, however, exist; therefore a method

of formulation must be found which does

not suggest their existence. Only when

the valencies of the carbon atom are

equally distributed in tridimensional space

is this ^requirement met; that is to say, ,

when they are directed from the centre to

the angular points of a regular tetrahedron,

thus: '

(D60)

66 CHEMICAL THEORY

Since this figure is symmetrical, the positions of the 2 hydrogen
and 2 chlorine atoms in methylene chloride shown upon it may
be interchanged in any way without causing a difference in the
relative positions of these 4 atoms. This conception of the
disposition in space of the valencies of the carbon atom, which
is due chiefly to van 't Hoff, has been very fruitful in organic
chemistry. The aspect of the science thus suggested has been
called "chemistry in apace", or stereochemistry. Space-formulae
should, of course, be applied to all chemical compounds, and some
progress has been made with elements other than carbon; but these
formulae are mainly of use in elucidating the structures of carbon
compounds, where the question of constitution is of such vital
importance.

It may be added that double and triple bonds are represented
stereocheraically by the joining of two tetrahedra along their
edges and adjacent surfaces respectively. For example, ethylerie,
CH 2 CH 2 , and acetylene, CH~~~CH, are thus represented:

Fig. 10

The Criterion of Valency.

The facts recorded in the foregoing pages suggest that valency
might furnish a means of chemical classification of the elements,
were it not that the exercise of this property varies somewhat
irregularly. On the other hand, an independent classification of
the elements might be expected to furnish information regarding
valency. Such information is supplied by the Periodic Classifica-
tion, which will shortly be studied. It will be sufficient to state
here that in this classification the elements fall into nine groups
Groups O to VIII; and that the maximum valency of each element
appears to be identical with the number of the group which
contains it. Thus, the no-valency elements of the argon family
are in Group O, the univalent metals of the alkalis in Group I, the
bivalent metals of the alkaline earths in Group II, and so on.

VALENCY AND CHEMICAL CONSTITUTION 67

Very seldom does the acting valency of an element exceed that
indicated by the group to which it belongs; nevertheless in
Group IB copper forms CuCl 2 and gold AuCl 3 ; more often, how-
ever, it falls below it. For example, the halogens belong to the
seventh group, and should therefore have a maximum valency of
seven. This is realized by chlorine in C1 2 7 , and by iodine in
H 6 I0 6 ; but not by fluorine or bromine. Iron, nickel and cobalt,
as well as osmium, ruthenium, &c., belong to the eighth group; but
whilst the two latter metals realize octa valency in OsO 4 and RuO 4 ,
tho three former metals appear never to be octavalent.

Nevertheless the Periodic Law is the true criterion of the
valency of an element. This will appear later when atomic struc-
ture is considered.

The Nature of Valency.

A study of the operation of valency, however detailed, or the
graphic representation of the union of elements in chemical com-
pounds by the use of bonds or solid geometrical figures, leaves the
nature of valency itself quite unexplained. It may be said that
the force which binds the atoms together is chemical affinity; but
this explains nothing, arid, moreover, the term " chemical affinity "
has received a meaning in physical chemistry which is not closely
associated with the idea of units of valency acting in specific
directions through space.

More than a century ago H. Davy l expressed the opinion that
" electrical effects are exhibited by the same bodies, when acting on
masses, which produce chemical phenomena when acting by their
particles". Berzelius extended this idea in his electro-chemical
theory, whence is derived the method of classifying the elements
as electro-positive and electro-negative. Faraday, later, showed
that during electrolysis a definite quantity of matter is always
associated with a definite quantity of electricity, a fact which
suggests that electricity as well as matter is atomic. This sug-
gestion starts a trail which might be followed into all the intri-
cacies of modern knowledge and theory concerning the structure of
the atom. The purpose of this chapter, however, has now been
fulfilled; but when the earlier development of the periodic law

has been considered in the next chapter the way will have been

i H. Davy, Phil. Trans., 1807, 1.

68 CHEMICAL THEORY

fully prepared for an excursion into this new domain; and the
promise may be made that in the course of this adventure the
" nature of valency " will become illuminated in such a remarkable
way that an entirely new and impressive conception will be gained
concerning it.

SUMMARY

VALENCY. The valency of an element indicates the number of
other atoms with which one of its atoms can directly combine.

CHAPTER IV
CLASSIFICATION OF THE ELEMENTS

The Periodic Law according to Mendeteeff

When the elements are regarded collectively, and in view of
their ascertained atomic weights and properties, two considera-
tions present themselves: (i) How may the elements be classified?
(ii) What is their origin? These considerations are related, for the
classification of material species is likely to lead to questions
regarding the origin of such species.

Probably the first systematic classification of the elements was
derived from the electro-chemical theory of Berzelius, to which
reference has already been made. This theory grew out of the
facts of electrolysis. Thus, if, for example, an electric current
passes through an aqueous solution of sodium chloride, the sodium
appears at the cathode or negative electrode, and the chlorine at
the anode or positive electrode. Consequently sodium was regarded
as electro-positive, being attracted to the electrode of opposite sign,
while chlorine was, for a similar reason, electro-negative. Or, more
generally, metals were considered to be electro-positive and non-
metals electro-negative. Further, it was recognized that some
rnetals are more electro-positive than others, power of metallic
replacement being regarded as a criterion of electro-positiveness.
For example, since zinc displaces copper from copper sulphate in
solution, zinc is more electro-positive than copper; and, conversely,
since chlorine displaces iodine from potassium iodide in solution,
chlorine is more electro-negative than iodine.

So this method of classification served not only for the distinc-
tion of metals from non-metals, but also for the recognition of
metallic and non-metallic intensity.

When the atomic weights of a sufficient number of the elements
had been established with some degree of accuracy, it was perhaps
inevitable that numerical relationships should be sought for between

70 CHEMICAL THEORY

them, and that attempts should be made to discover a connection
between the properties of an element and its atomic weight.

The first attempt to establish numerical relations between the
atomic weights was made in 1815-6 by an Edinburgh physician
named Prout, who tried to prove that all the elements are con-
densations of hydrogen as the primordial substance, by affirming
that all the atomic weights are whole numbers when that of
hydrogen is unity. This affirmation was unjustified at the time,
for Berzelius subsequently showed that a number of atomic weights,
determined with accuracy by the use of material ordinarily avail-
able, were far removed from whole numbers. Nevertheless the
fact remained that when the atomic weight of oxygen is made
equal to 16*00 " the atomic weights tend to approximate to whole
numbers far more closely than can reasonably be accounted for by
any accidental coincidence''; 1 and therefore it appeared, even a
quarter of a century ago, that the complete rejection of Prout's
hypothesis was unwarranted. Kecent work, of which an account
will be given in the sequel, has gone far to re-establish Prout's
hypothesis; which, however, considering the time of its promul-
gation, must be regarded as a philosophic guess rather than a
conclusion of inductive science.

Another attempt was made by Dobereiner, in 181 V and 1829,
who showed that in various triads of related elements the central
member of each group possesses properties and an atomic weight
which are approximately the mean of the properties and atomic
weights of the extreme members of the triad. These triads are:
lithium, sodium, potassium; calcium, strontium, barium; phosphorus,
arsenic, antimony; sulphur, selenium, tellurium; chlorine, bromine,
iodine.

It will be sufficient to give numerical details for the first and
last of these triads.

-P..^ Mean of Extreme

Differences. 4 , . , ir . , ,

Atomic Weights.

' 23-02

Element.

Atomic Weight.

Lithium

6*04

Sodium

23-00

Potassium

39-10

Chlorine

35-46

Bromine

79-92

Iodine

126-93

47-or ...:.: 8i - 19

It will be observed that the atomic weight of sodium is almost

1 K. J. fctrutt, Phil. May. [vi], 1, 311 (1901).

CLASSIFICATION OF THE ELEMENTS 71

exactly the mean of the atomic weights of lithium and potassium,
but that the atomic weight of bromine is considerably leas than the
mean of the atomic weights of chlorine and iodine. The relations
suggested by these triads are therefore approximate only. It has
been objected, moreover, that triads should not be made up to the
exclusion of other related elements; thai, for example, there are
four halogens, and that it is arbitrary to exclude fluorine by form-
ing a triad with the other three. But when it is recognized that
fluorine differs from the other halogens, not only in atomic weight
relationship, but also widely in the properties of its compounds,
this objection loses force. So that without doubt the relationships
shown by Dobereiner's triads are remarkable; nevertheless their
value is historic only, for they are now merged in the generalization
known as the periodic law.

Another kind of triad was, however, observed by Dobereiner,
in which the three related elements have nearly identical atomic

weights. These triads are:

Iron 55*84
Cobalt 58-94
Nickel 58-69

Ruthenium 101-7
Rhodium 10:2 -9
Palladium 100 -7

Osmium 190-8
Iridium 193-1
Platinum 195-2

They also find a place in the periodic classification. Dobereiner's
observations were limited to the elements cited above. These
observations could not give rise to a generalization, since they were
concerned with only a minority of the elements; the majority did
not form triads; and therefore it is difficult to see what significance
could have been attached at the time to the existence of these
triads.

Strecker, in 1859, initiated the idea of seeking relations between
the elements placed in atomic weight sequence; whilst de Chan-
courtois, in 1862, placed the elements in sequence in a spiral round
a cylinder divided into sixteen equal sectors to represent atomic
weight magnitudes. Thus analogous elements of low atomic weights
fell into places in vertical columns because properties recur in such
elements after atomic weight differences of 16.

In 1863-6 J. A. R. Newlands arranged the elements in ascend-
ing order of their atomic weights, commencing witli hydrogen,
thus:

H Li Be B C N O

F Na Mg Al Si P S

01 K Ca Cr Ti Mn Fe, &c.

72 CHEMICAL THEORY

In this way he discovered that the eighth element is " a kind of
repetition of the first ", the ninth a repetition of the second, and so
on; Na, for example, is a repetition of Li, Si of C, Cl of F. This
discovery he called the

Law of Octaves. " Members of the same group of elements
stand to each other in the same relation as the extremities of
one or more octaves in music."

This simple "law" did not apply to the elements of higher atomic
weight; even in the above table manganese is wrongly classified
with phosphorus; and it was suggested by a contemporary of
Newlands that it would be as useful to arrange the elements in
alphabetical order as in the order of their atomic weights! Never-
theless, the law of octaves is valid as an introduction to the
periodic law. In the year 1869 Mendeleeft* arranged all the
elements in the order of their atomic weights, and discovered
a periodicity in their properties. The fact of this periodicity he
enunciated in the following statements:

1. The elements arranged according to the magnitudes of their
atomic weights show a periodic change of properties.

2. Chemically analogous elements have atomic weights either in
agreement (Pt, Ir, ()s), or increasing by equal amounts (K, Kb, Cs).
(Of. Dobereiner's triads.)

3. The arrangement according to atomic weights corresponds
with the valencies of the elements, and to a certain extent the
difference in chemical behaviour, for example: Li, Be, B, C, N, O, F.

4. The elements most widely distributed in nature have small
atomic weights, and all such elements are distinguished by their
characteristic behaviour. They are thus typical elements, and the
lightest element, hydrogen, is therefore rightly chosen as the typical
unit of mass.

5. The magnitude of the atomic weight determines the properties
of the element, whence in the study of compounds regard is to be
paid not only to the number and properties of the elements and
their mutual action, but to the atomic weights of the elements.
Hence the compounds of S and Te, Cl and I, show, beside many
analogies, y et striking differences.

6. It allows the discovery of many new elements to be foreseen;
for example, analogues of Si and Al, with atomic weights between
65 and 75.

7. Some atomic weights will presumably experience a correction;

CLASSIFICATION OF THE ELEMENTS 73

for example, Te cannot have the atomic weight 128, but 123 to 126.

8. From the table new analogies between elements become
apparent. . . .

Some of these statements are open to criticism or require modi-
fication. Thus regarding statement 4, lithium and beryllium are
not so widely distributed as the heavier metals sodium and potas-
sium, and magnesium and calcium respectively, and conversely
among the heavier metals tin and lead are more widely distributed
than is the lighter germanium. Also the difficulty in statement 7
lias been overcome recently, but not in the way suggested by
Mendeleeft*.

Nevertheless these generalizations marked a great advance on the
position of earlier chemists, and Mendeleeff, whilst acknowledging
the work of his predecessors, rightly claimed that he was the
first " to foretell the properties of undiscovered elements, or to alter
the accepted atomic weights ".

The Periodic Law, according to Mendeleeff, may therefore be
stated thus:

The physical and chemical properties of the elements and their
compounds are periodic functions of the atomic weights; or

If the elements are arranged in the order of increasing atomic
weight, their properties vary definitely from member to member of
the series, but return to a more or less similar value at fixed points
in the series.

The periodic system, according to Mendeleeff, may now be
developed.

Hydrogen, the element of lowest atomic weight, became the
sole member of Series 1 in Mendel^efTs system. Series (2) and (3)
were the same as in the octaves of Newlands, thus:

(2) Li Be B C N O V

(3) Na Mg Al Si P S 01.

The next two series were:

(4) K Ca Sc Ti V Cr Mn F Co Ni

(5) Cu Zn Ga Go As Se Br,

being linked together by the triad Fe, Co, Ni; for to place these
three elements in the consecutive positions occupied by Cu, Zn, Ga,
thus displacing all that follow them, would have been not only to
obliterate periodicity from the scheme, but also to ignore the

LONG PERIODS

c5 1 1

^ ^ 1 1

w - 1 M
< (/) 1 W

_J CO

< o

'* "

C 1 *f>
O CO ' i^

H 0.

o 5 1 H

I 1

u. u

C -O I bO

N U 1 E

o </>

3*11

2 CU

2 1

,. __j

U OJ 1 A

j <

& OS O

4) bo
CQ g

1 1 1 1

.-, nj

*J 2

U | 1 D

^ o

> 1 1 fh 1

<tf
U

nf |

CQ I

O
(E

O
i

CO
D
CO

CO
CO

CLASSIFICATION OF THE ELEMENTS

peculiar relations these three elements bear to one other as members
of a triad.

Series (6) and (7) connected by another triad were:

(6)
(7)

Sr
Cd

Cb
Sb

Mo
Te I.

Ru Rk Pd.

Now it will be seen that Series (4) and (6) begin with the alkali
metals K and Rb, whilst Series (5) and (7) begin with the metals
Cu and Ag, which, whilst allied to each other, differ widely from
the alkali metals. Similar differences exist between subsequent
members of odd and even scries. Elements in vertical columns
constitute groups; of which, according to Mendeldeff, there
were eight: seven groups corresponding with Newlands's
octaves, and an eighth group in which Dobereiner's triads of
nearly equal atomic weight were placed. When the inert gases
were discovered, these were placed in a group by themselves:
Group O, which preceded the other groups. Except with regard
to the elements of Series (2) and (3), Groups I to VII were sub-
divided into A and B Sub-groups, to show the above-mentioned
differences between consecutive members of the same group. Thus
the complete periodic system took the following form.

Groups

III

VII

VIII

Sub-

A B

A 15

A B

A 13

A B

groups

Series 1

Cr Mn KeCONi

Br,

I! 11 IMi Pd

' ' 7

Cs Ba

Rare

Earth

Os Ir Pt

Oxides

X 2

XO 2

X0 3

XO,

The above arrangement was improved upon, and a clearer
view obtained, by recognizing the existence of short and long
periods. Thus Series (2) and (3^ constituted short periods; Series

70 CHEMICAL THEORY

(4) and (5), with the linking elements of the eighth group, formed
one long period. Other long periods followed, and the whole scheme
shown iii the chart on p. 74 resulted.

The great advantage of this mode of presenting the Periodic
System was that the A and B Sub-groups were separated, so that
elements which have little resemblance to one another were not
classified together. For instance, it may well be objected that
Cu, Ag, and Au, being very unlike the alkali metals, should not
be placed with them in Group I. This objection is sufficiently
answered when it is shown that these metals occupy positions
near the centres of the long periods, whilst the alkali metals
are quite differently situated at the beginning of these periods.
Similar remarks apply to the relation between manganese and
the halogens.

The arrangement of elements in any group now takes this
form, illustrated by Group [:

(A) Na (B)

K Cu

Kb Ag

The table on p. 74 represents the final and most useful form of
the Periodic System according to Mendel^eff, but before proceeding
further it is desirable to point out some of its shortcomings, ami
thus to give a hint of the modification the system has necessarily
undergone on account of recent knowledge.

Mendeleett* did not classify the metals of the rare earths. For
one reason the number of these was unknown, and for another their
properties did not progress from member to member as did the
properties of elements in the recognized periods. Therefore the
position of the rare-earth metals in the scheme could not be given
in detail; but it was indicated that they intervened between Ce in
Group IV and Ta in Group V.

Whilst the rare-earth metals could not be spaced, there remained,
nevertheless a large number of blank spaces following these elements;
and in course of time it became increasingly improbable that these
spaces ought to be reserved for elements hitherto undiscovered. It
was scarcely credible that if 18 elements indicated by blank spaces
existed not one of these should have been discovered. So it was

CLASSIFICATION OF THE ELEMENTS 77

proposed to fill these blank spaces with Ta and the elements that
follow it, moved up from the series below, so making one very lon<^
period including the rare-earth elements.

Thus A. Werner 1 suggested a long period of 33 elements from
caesium to the higher analogue of xenon, now know as radon; and
except that his estimate of the number of existing rare-earth
metals was one too many, time has proved that Werner was
right.

It must be observed, however, that on account of the rare-earth
metals, progression of properties from member to member is not
shown throughout this long period; but that these metals may be
regarded as functioning as a single element in the same sense as
the individual members of the triads in the eighth group function
together as a single element.

Atomic Weight Differences in the Periodic System.

Attention may now be drawn to atomic-weight differences
between analogous elements in consecutive short and long periods.
These differences are shown for a number of the elements in the
following tables:

THE Two SHORT PKRIOJDS

He Li Be B NO F

Ne Na Mg Al Si P 8 Cl

Differences 16-2 16-06 15-30 16-15 16-06 17-02 16-06 16-46

THE FIRST Two LONG PERIODS

Ar K Ca Sc Ti V Cr Mn Fe

Kr Rb Sr Y Zr Cb Mo llu

Differences 42-99 46-34 47-56 43-8 42-9 42-14 43-99 45-80

Co Ni Cu Zn Ga Ge As Se Br
llh Pd Ag Cd Tu Sn Sb Te 1
Differences 43-97 48-01 44-31 47-03 45-08 46-1 46-81 48-3 47-02

It will be observed that the differences in the short periods
are approximately 16, and in the long periods about 45; in the
short periods 8 elements intervene before a recurrence ^f properties,
and in the long periods 18 elements. The differences are by no
means constant, for no mathematical relations exist between the
atomic weights; but anomalies are seen in the differences between

i Ber.. 1905. 3S. 914.

78 CHEMICAL THEORY

krypton and argon, palladium and nickel, tellurium and selenium,
in accordance with the anomalies in the atomic weights of argon,
nickel and tellurium, to which attention will be drawn. It must
be confessed, however, that there are other anomalies which are
not pronounced enough to affect the order of the atomic weights
of the elements.

Stress, however, must not be laid upon atomic- weight differences,
because atomic-weight values themselves are now known to be of
only secondary importance in matters of theory and classification of
the elements. Indeed all the anomalies in atomic-weight relations,
whether they affect the relative positions of the elements in series
or not, are now removed, because the conception of atomic number
(q.v.) has displaced that of atomic weight as of primary impor-
tance.

The elements of the short, or so-called typical periods, may be
allied to those either of the A or the B Sub-groups. In the case of
Group I, Li and Na are plainly related to the other alkali metals
K, Rb, Cs, in the A Sub-group, rather than to Cu, Ag, and Au
in the B Sub-group, but in Group VII, F and Cl are related to
Br and I in the B Sub-group, rather than to Mu in the A Sub-
group. This latter relationship obtains in all groups from II to
VII.

The periodic law, according to Mendeleoff, states that the physical
and chemical properties of the elements and their compounds are
periodic functions of their atomic weights. This statement must
now be illustrated.

Periodicity of Physical Properties.

Perhaps the most obvious property of a solid element is its
density. It was shown by Lothar Meyer, in 1870, that the
densities of the elements vary periodically. Instead, however,
of using the densities of the elements directly, L. Meyer calculated
from them the atomic volumes, and plotted these values on a curve
as ordinates, together with the atomic weights as abscissae. The
atomic volivne of an element is related to its density in the
following manner.

The reciprocal of the density is the specific volume,

Specific volume = ^ -, ' = volume of unit mass :
density

CLASSIFICATION OF THE ELEMENTS

S 3 VM

o
pi (At

80 CHEMICAL THEORY

the atomic volume is this value multiplied by the atomic weight
thus:

Atomic volume = atomicweight
density

For example, the atomic weight of copper is 63*6, and its
density 8-9; consequently

Atomic volume Cu = 63 ~ = 7-15.

O 7

This figure stands for the relative volume of a mass of copper
proportional to the atomic weight of the element; it does not
express the relative size of the copper atoms themselves; it could
only do this if the atoms were packed without interspaces, or if
the interspaces were constantly related in volume to the atomic
material of the elements. What it does express is the relative
volume of the atom plus its share of atomic interspace. The
atomic volume curve shows a remarkable periodicity; for it is
like a series of waves consisting of crests and hollows; moreover,
the crests of successive waves increase in height with increasing
atomic weight. The most important fact connected with the curve,
however, is that related elements occupy analogous positions upon
it. For example, the alkali metals, potassium, rubidium, and
caesium, are at the apices of successive curves, the halogens,
chlorine, bromine, and iodine, are on ascending, and the alkaline
earth metals, calcium, strontium and barium, on descending parts
of the curves.

The following other physical properties of the elements and
their compounds are periodic.

Melting-point, malleability, coefficient of expansion, atomic
refraction, conductivity for heat and electricity, colours of salts
in solution. Consequently there are certain regions, which are
similar on successive curves, where these properties are manifested,
or reach their maxima. The student may test this statement with
reference to the melting-points of the elements.

The periodic occurrence of colour in compounds is very striking.
Thus all the metals whose salts give coloured solutions are included
in the following series:

Ti V Cr Mn Fe Co Ni Cu

Mo Ru Rh Pd

W Os Ir Pt Au
U

CLASSIFICATION OF THE ELEMENTS 81

These series consist of metals in atomic weight sequence, arid
they occupy the lowest portions of successive parts of the atomic
volume curve. There are other coloured compounds, however,
which do not give coloured solutions; e.g. various sulphides and
iodides. In the case of these compounds there is generally a
deepening of colour with rise of atomic weight in a group, as in
the sulphides of zinc, cadmium, and mercury. It is noteworthy
that the colours of these compounds belong only to the solids;
for when scarlet mercuric iodide is dissolved in alcohol, and
yellow lead iodide in water, colourless solutions are obtained.
This is to be expected, since colour is not associated with mercuric,
lead, or iodide ions; thus the nitrates of mercury and lead are
colourless, and so are the iodides of the alkali and alkaline earth
metals.

Periodicity of Chemical Properties.

The fundamental chemical division of the elements is into
metals and non-metals; and, according to the classification of
Berzelius, metals are electro - positive, and non-metals electro-
negative.

As the elements are traversed in the order of ascending atomic
weights the variation of metallic and electro -chemical properties
is periodic. Thus in the two short periods from lithium to
fluorine, and from sodium to chlorine, there is continuous and
regular transition from great metallic and electro - positive to
extreme non-metallic and electro - negative characters. In the
long periods which follow, for example the period from potassium
to bromine, there are two phases; the first phase is from potassium
through manganese to the eighth-group metals iron, cobalt, arid
nictfel; the second phase is from copper to bromine. The transition
from potassium to bromine is similar in degree to that from sodium
to chlorine, but the period contains more than twice as many
elements; and the stages of this transition present an interesting
phenomenon.

The elements of the first phase (K to Fe, Co, Ni) are all
metals, but there is a continuous diminution of electro-positiveness
throughout them; the elements of the second phase begin with
the comparatively inert and electro-negative metal, copper, and
there is actually a rise in metallic strength to zinc, followed by
a regular fall to the non-metallic and electro-negative bromine.

(D60) 7

82 CHEMICAL THEORY

Similar relations exist in the subsequent long periods; but the
inertness of the central elements, i.e. those of the eighth group
and of Group IB, becomes more pronounced with elements of
higher atomic weight.

When the transition of properties within the separate groups,
i.e. the elements in vertical columns, is considered, an increase of
metallic nature or decrease of non-metallic nature is found to be
the rule. Thus, for example, the alkali metals increase in electro-
positiveness with rise of atomic weight; and the halogens similarly
show a diminution of electro-negativeness with rise of atomic
weight. Within the region of chemical inertness and metallic
electro-negativeness, i.e. the eighth group, Group IB, and to a less
extent Group II B, an opposite state of things, however, exists;
there'is a diminution of electro-positiveness and chemical reactivity
with rise of atomic weight. Thus the inert metals, platinum,
gold, and mercury, occur consecutively as the last members of
Groups VIII, IB, and II B.

From all this it follows that the most powerful metals are to
be found at the extreme left of the periodic diagram; caesium,
the most electro-positive metal, being in the lower left-hand corner;
whilst the non-metals occupy the upper right-hand portion of
the diagram; fluorine, the most powerful non-metal, being in the
upper right-hand corner. The dotted line in the diagram on p. 74
delimits the region of non-metals.

Periodicity of Valency.

The following statement is generally true.

The maximum valency of an element corresponds with the
number of the periodic group to which it belongs. The statement
is illustrated by the formulae of the typical oxides appended to the
table on p. 75. In Chapter III valency was illustrated by lists of
hydrides, halides, and oxides, and in most of the formulae for
the halides and oxides, but not the hydrides, the numerical
value of the valency indicates the group to which the element
belongs. A valency of seven is not always realized in the seventh
group; lean often is a valency of eight seen in the eighth group.
On the other hand, copper and gold in Group IB show bi- and
ter-valency respectively in CuCl 2 and AuCl s , but the elements
of this group are in any case somewhat anomalous in their
relationships. A more striking exception is shown in the case of

CLASSIFICATION OF THE ELEMENTS 83

boron, which forms the hydride B 2 H 6 and other hydrides, in which
the element can hardly be less than quadrivalent.

It was seen in the chapter on valency that the sum of the
oxygen valencies and hydrogen valencies in volatile hydrides of
an element is equal to eight; and this is true irrespective of the
periodic group to which the element belongs. The elements of
Groups I, II, and III, however, excepting boron, form no volatile
hydrides, and exhibit only the lower valencies in the oxides. Thus
in the first and second short periods the oxides and hydrides show
valencies as follows:

I II III IV V VI VII

LLO ?>eO B 2 O 3 CO 2 NA Oxides.

- B 2 H C , &c. CH 4 NH 3 OH 2 FH Hydrides.

Na,,O MgO Al a O s SiO 2 P 2 O 6 SO 3 Cl 2 O r Oxides.
- - SiH 4 PH 3 SH 2 C1H Hydrides.

Metals which form non-volatile hydrides exhibit the same
valencies in these compounds as in the oxides; for example,
K 2 O ? KH; CaO, CaH 2 .

The periodic law constitutes a valuable criterion of valency,
because the periodic group to which an element belongs indicates
almost invariably the valency of the element in the highest oxide
which it can form. The existence of super-oxides, such as NaoO. 2
and Ba0 2 , constitutes no real exception to this rule, because these
compounds are constituted thus:

Na-O-O-Na, Ba/ | , or ^OrO, Ba=O:O;

and so the valencies of the metals are the same as in the corre-
sponding basic oxides. The elements of Groups VI, VII, and VIII
often fail to realize their maximum valency; and indeed some,
e.g. iron, never exhibit the group valency. Since the chemical
character of a compound depends largely upon the active valency
of its nuclear element, the elements of these higher groups show
great variety in the properties of their compounds, because they
exhibit highly variable valency. The highest oxide is to be
regarded as the typical oxide, provided it exhibits tiie valency
of the group to which the element belongs; it is then found
that lower oxides and their derivatives show relationships to
oxides and their derivatives of similar type, but belonging to
elements in other groups.

84 CHEMICAL THEORY

For example:

Derivatives of

Mn 2 O 7 in Group VII are isomorphous with those of C1 2 O 7 in Group VII.

MnO, VII S0 3 VI.

Mn a 8 VII Fe 2 8 VIII,

andA! 2 O 3 III.

MnO VII FeO VIII,

and ZnO II.

Other examples might be given, all of which show that poly-
valent elements, forming several classes of compounds, exhibit
several relationships corresponding to these classes, and therefore
that the type is the determining factor in chemical relationship.
Consequently manganese, which can be septavalent, is not dis-
qualified from appearing in Group VII by reason of relationships
to metals in Groups VIII, VI, III, and II.

Uses of the Periodic Law.

Prediction of Unknown Elements. In the periodic scheme,
as first formulated by Mendel^eff, there were some significant
omissions. The positions now occupied by scandium, gallium,
and germanium were left blank, since no elements were known
qualified to fill them. If every available space had been filled
with the known elements, placed in the order of their atomic
weights, there would have been no periodic system, or but a
distorted one, because every element which now follows a
space that should have been left unoccupied would thereby have
been moved one space forward, and the arrangement of analogous
elements in groups would have been interfered with. On the
other hand, deliberately to leave certain spaces blank, so as to
preserve the desirod periodicity, was to suggest that elements
remained to be discovered to fill these spaces, arid so to provoke
a severe test of the truth of the periodic law.

The latter alternative was chosen by Mendel^eff, and in par-
ticular the existence of three elements was foretold, which were
named provisionally eka-boron, 1 eka-aluminium, and eka-silicon.
The first o these lay between calcium and titanium in the periodic
table; the other two were placed consecutively to fill two blank
spaces between zinc and arsenic. Moreover, by reference to the
properties of neighbouring elements in series and in group, it

1 Eka is Sanskrit for one.

CLASSIFICATION OF THE ELEMENTS

was possible to foretell with considerable accuracy the properties
of these undiscovered elements.

This prophetic use of the periodic system by its discoverer
has been rightly compared with the employment by Adams and
Le Verrier of mathematical calculation to foretell the existence of
the planet Neptune from observed irregularities in the movements
of Uranus, and it has had an equally satisfactory vindication. For
the elements scandium, gallium, and germanium, subsequently dis-
covered, have been found to possess properties closely agreeing
with those foretold by Mendeteeff. This is illustrated in the follow-
ing comparison of eka-aluminium with gallium.

EKA-ALUMINIUM.

GALLIUM

Atomic weight, dr. (58.

Metal of density 59 and low
melting -point; not volatile; un-
affected by air ; should decompose
steam at a red heat and dissolve
slowly in acids and alkalis.

Oxide should have formula E1 2 O 3 ,
density 5 5, and dissolve in acids to
form salts of the type E1X 3 . The
hydroxide should dissolve in acids
and alkalis.

There should be a tendency
towards the formation of basic salts.
The sulphate should form alums.
The sulpnide should be precipitated
by H ? S or (NHAjS. The anhydrous
chloride snoula be more volatile
than zinc chloride.

The element will probably be dis-
covered by spectrum analysis.

Atomic weight, 69-9.

Metal of density 5-94; melting at
30-15; not volatile; unchanged in
air; action on steam not known;
dissolves slowly in acids and alkalis.

Oxide, Ga 2 O 3 ; density not known;
dissolves in acids, forming salts
GaX 3 . The hydroxide dissolves in
acids and alkalis.

Salts readily hydrolyze and form
basic salts. Alums are known. The
sulphide can be precipitated by
H 2 S or (NH 4 ) 2 S, but only under
special circumstances. The anhy-
drous chloride is more volatile than
zinc chloride.

Was discovered by spectrum
analysis.

There are other blank spaces in the periodic system which pre-
sumably correspond with hitherto undiscovered elements. Modern
research, however, has shown that probably only three spaces
remain unfilled, since most of the spaces shown in the table on
p. 74 are obliterated when the talkie is rearranged according to
recent knowledge.

The three elements required to fill these spaces *y*e: another
alkali metal to precede radium, one rare-earth metal, and a halogen
element to precede radon. A fruitless search has been made for
the analogue of caesium, but the discovery of the two missing
analogues of manganese has recently been announced.

86 CHEMICAL THEORY

Correction of Atomic Weight Values.

Since the periodic law requires sequence of atomic weight
values and sequence of properties to be in accord, grossly erroneous
atomic weight value placed in sequence must disturb the sequence
of properties; or, conversely, if sequence of properties is maintained
it will necessitate a departure from atomic-weight sequence. In
cither case the erroneous value is revealed when the element in
question is considered in the light of the periodic law. Indeed
the erroneous atomic -weight value must cause a position to be
claimed for the element, which, according to its properties, should
be occupied by another element, and must consequently leave
vacant a place suited to the element and in accord with its true
atomic weight. Therefore the periodic law is of value, not only
for detecting false atomic- weight values, bat also for suggesting
true ones. For example, the atomic weight of caesium was at
first erroneously thought to be 123-4. This value would place
caesium after antimony, and, of course, cause the displacement
of tellurium, iodine, and other elements one place to the right.
Such a condition cannot be thought of; therefore the value
123-4 is condemned. On the other hand, since caesium is an alknli
metal it should follow rubidium in group, and consequently have an
atomic weight of about 131*8, so that Cs Rb = Rb K = 45-35.
The atomic weight of crosium is now known to be 132-81, and so
this metal occupies its proper place in the scheme. In the cases
of beryllium, indium, and uranium the periodic law has furnished
the means of deciding what multiple of the equivalent is the atomic
weight. The equivalent weight of beryllium is 4-55, and the
atomic weight of this element was at first thought to be 4-55 x 3
= 13-65. This value would place beryllium in an impossible posi-
tion between carbon and nitrogen, whereas 4^55 X 2 = 9-1 weuld
give it a place in harmony with the periodic law. Subsequent
considerations have confirmed the value Be(Gl) = 9-1.

Indium with the equivalent weight 38-27 was thought to have
an oxide InO, and atomic weight 76-54, which would place this
element between arsenic and selenium, where it cannot stand. An
atomic weight of 38-27 x 3 = 114-8, with the corresponding
oxide In 2 O 3 , would satisfy the periodic law; and this value has
subsequently been accepted on the grounds of specific heat.

The atomic weight of uranium was originally thought to be
about 60, or else 120; but neither of these values enables the

CLASSIFICATION OF THE ELEMENTS 87

element to be placed suitably in the periodic scheme. The value
240 was required by Mendeleeff, so that the element might become
the last member of the sixth group, following tungsten. This high
value, or more accurately 238 2, has been supported by the vapour-
density method applied to the halides, and by the fact that
uranium is radio-active, since radio-activity is characteristic of the
heaviest atoms.

The criticism of the atomic weights of the elements by means
of the periodic law may be carried further. The accepted atomic
weight of argon is greater than that of potassium, that of cobalt is
greater than that of nickel, and that of tellurium is greater than
that of iodine; yet the individual members of these three pairs of
elements are placed in the reverse order of their atomic weights
in the periodic scheme, because their properties do not permit of
any other arrangement. Repeated attempts were made in the case
of tellurium to reduce the value of its atomic weight below that of
iodine, but without avail, and it was supposed that the relationships
of these three pairs of elements constituted exceptions to the periodic
law. Such a conclusion was unsatisfactory; but the difficulty has
been removed by the recognition of atomic number as the criterion
which decides the position of an element in the periodic scheme;
and the atomic numbers of the elements in question place them in
a sequence which accords with their properties.

The Suggestiveness of the Periodic Law.

In spite of the apparent imperfections and anomalies it contains,
the periodic law is true in principle. Indeed it cannot be doubted
that the truth, beauty, and value of this law as an index to the
material world have been enhanced by the discoveries of recent
yeurs. This fact is a challenge to the scientitic imagination; it
must provoke questionings and research. For example, in a group
of allied elements, such as the alkali mrtals, Li, Na, K, Eb, Cs,
there are series of compounds such as oxides, hydroxides,
chlorides, sulphates, carbonates, and so forth, which may be ex-
pected to be related to one another somewhat as the metals
themselves are related. The. examination of the pjiysical and
chemical properties of these compounds may therefore be under-
taken with a view to discovering the gradations which exist
between them. Interesting relations will thus be established^
and this fact will become apparent: that there is a break in the

83 CHEMICAL THEORY

gradation of properties between Na and K; in other words, that
K, Rb, and Cs and their compounds are closely related, while Na
and its compounds, as well as Li and its compounds, stand apart
from them. The periodic classification affords an explanation of
this phenomenon; it is that Na is situated in the second short
period, whilst K occupies a different kind of position near the
beginning of the first long period, and Rb and Cs follow K in
quite analogous positions in subsequent long periods. Having
observed this, the student may then remember that although
caustic soda and caustic potash are thought of as very similar
substances, sodium salts are after all not very similar to potassium
salts, for they do not crystallize with the same amounts of water
of crystallization as the latter, and frequently they are not iso-
morphous with them, while their solubilities in water are so
different from those of potassium compounds that solutions of
sodium salts are used to precipitate potassium, and vice versa.

At the other extremity of the periodic table the halogens pre-
sent another interesting subject for study. The fact that the
affinity for hydrogen diminishes from F to I in the hydrides H F,
HC1, HBr, HI is well known, and is quite in accord with what occurs
in other groups; e.g. in the hydrides OH 2 , SH, 2 , SeH.,, TeH, 2 in
Group VI, or NH 3 , PH 3 AsH 3 , SbH 8 , (BiH 8 ) in Group V; but fluorine
is widely different from the other halogens. Why is this'

This is a sort of question that must be answered ultimately by
reference to the constitution of the atom; but consideration elicits
this remarkable fact: that all the elements of the first short period
are unique, being widely separated in properties from those in the
same groups which follow them. It suffices to draw attention to
carbon, nitrogen, and oxygen, which cannot be properly classified
with the elements succeeding them. Again, hydrogen fluoride differs
remarkably in condensibility from the other halogen hydrides; is
there any analogy to this phenomenon in neighbouring groups?
Assuredly there is; if water were no more condensible than hydro-
gen sulphide, the world would be a very different place to live in!

In the region of the periodic chart where volatile hydrides
occur the fallowing compounds are found:

OIL

NH 3

OH.,

SiHL
GeH 4

PH :? J
AsH,

SH;

SeH.,

C1H
BrH

(SnH 4 )

SUH 3

TeH^

III

CLASSIFICATION OF THE ELEMENTS 89

The periodic law suggests a comparison between them in series
and in group; and thus the following gradations of properties are
discovered.

The hydrides diminish in stability with rise of atomic weight in
every group. Thus, for example, in the fifth group ammonia is very
stable, and is decomposed only slowly by the passage of electric
sparks: phosphine, PH 3 , is less stable than ammonia, and is rapidly
decomposed by the same agency; AsH 3 is broken up into its
elements when passed through a tube heated to 230, SblI 3 is
similarly decomposed at 150, and BiH 3 is too unstable to be
isolated.

In series, i.e. in the hydrides standing in horizontal lines, there
is an increase of stability with rise of atomic weight, corresponding
with the increase of non-metallic characters, and also the diminution
of hydrogen valency, so that there is less hydrogen to be retained.
Thus hydrogen fluoride is the most stable volatile hydride, and ger-
manium and bismuth hydrides the least stable. It may be observed
that Ge, As, Sb, and Bi are metalloids, that is, almost metals. No
true metal forms a volatile hydride. The power to form alkyl
compounds, i.e. compounds with radicles, such as methyl, 'CH 3 , arid
ethyl, *C 2 H 5 , is more extensive than that to form hydrides; so that
some metals in the B sub-groups preceding in series the above non-
metals form these so-called organo-metallic compounds. Perhaps
the best known of these substances is zinc ethyl, Zn(C 2 H 5 ) 2 ; but,
in addition to zinc, cadmium, mercury, tin, lead, and bismuth form
them, and thus come into line with the above non-metals, all of
which form alkyl compounds as well as volatile hydrides.

Another interesting but rather difficult question is that of the
relative acidic or base-producing power of these volatile hydrides.
Consider the four hydrides:

CH 4 , NH 3 , OH 2 , FH.

Methane is inert; ammonia is base-producing, for its solution in
water is alkaline owing to the reaction: NH 3 -f H" + OH' ^-^
NH 4 " + OH'; water is neutral, and hydrogen fluoride is'acid. Why
is not methane, CH 4 , more base-producing than NH 3 ? the grada-
tion of properties seems to require it to be. The answer is that in
CII 4 carbon is already saturated .with hydrogen, so that this sub-
stance cannot form an additive compound with water or an acid as

90 CHEMICAL THEORY

ammonia does; for the peculiar base-producing power of ammonia
is an additive property, viz.: NH 3 + H* = NH 4 V
Consider again the hydrides:

NH 3 OH 2
PH3 SrI 2 .

There is a loss of base -producing power from NH 3 to PH 3 , and
an apparently analogous increase in acidity from OH 2 to SH 2 ; but
it is difficult to generalize here, for ammonia is unique in base-
producing power, just as nitrogen is unique as an element; aud
water, again, like oxygen, is unique in its properties. Moreover,
it must not be concluded that increase in acidity of hydrides with
rise of atomic weight in a group is general, for C1H, BrH, and IH
are acids of about equal strength.

The comparison of properties of the oxides of elements in the
various groups of the periodic system is a simpler and more satis-
factory exercise. For there is in general a loss of acidic and a
corresponding gain of basic properties with rise of atomic weight
in a group. This is shown, for example, in the oxides

N 2 O 3 P 2 O 3 As 2 O 3 Sb a Os Bi 2 O 3 ,
and N 2 O 6 P 2 6 AsA Sb 2 O 6 Bi 2 O 6 .

In the trioxides there is a gradual transition from wholly
acidic, through amphoteric 2 to purely basic properties, and in the
pentoxides from powerfully to very feebly acidic properties.

Again, the trioxides of Group VI A,

CrO 3 , MoO 3 , WO 3 , UO 3 ,

form an interesting series; for, in accordance with the above
generalization, basic properties actually appear, together \YJth
acidic properties, in the oxide UO 8 , which is basic with regard to
one oxygen atom only, forming basic salts, such as UO 2 (NO 3 ) 2 , the
uranyl salts.

Objections to the Periodic Law.

A consideration of the criticisms to which the periodic system
has been submitted is valuable. If the criticisms are baseless, as
some of them are, the process of their refutation will be illumi-
nating; if they are valid, their consideration may exhibit the

1 The meaning of this will appear when the subject of ionization is considered.
3 Both basic and acidic, a/m^orepo? = both.

CLASSIFICATION OF THE ELEMENTS 91

relations of the elements from a new point of view, and so
increase our knowledge concerning them.

The most sweeping accusation which has been brought against
the periodic system is that it places together dissimilar elements,
whilst separating similar ones. It brings together the alkali metals
and copper, silver, and gold in Group I, it is said a most unnatural
alliance. This objection has already been met by a denial of the
statement that these dissimilar metals are brought together. It is
further objected that the periodic classification separates copper
from mercury and barium from lead. But it may be maintained
that such separation is proper; for the similarities between the
metals in these several pairs are superficial rather than funda-
mental, for copper and mercury are widely different in physical
properties and in oxidizability; and, in spite of the fact that both
metals form two series of salts, and that their lower chlorides
are insoluble in water, there is little further resemblance between
their corresponding salts. The differences between barium and
lead are even more fundamental, so that to regard the elements
as similar on account of the insolubilities of their sulphates, and
the isomorphism of some other salts, is a grave error of judgment.

The discovery of argon, and the determination of its atomic
weight, furnished material for adverse criticism of the periodic law.
For not only was it supposed that no room could be found in the
scheme for an element with such extraordinary properties as argon
possessed, but the atomic weight of argon was found to be greater
than that of potassium; and it was manifestly impossible to place
this element between potassium and calcium. Then other inert
elements were discovered helium, neon, krypton, xenon, the
companions of argon; and these have atomic weights less than
tho^e of the neighbouring alkali metals. Thus the atomic weight
of argon is recognized as anomalous, like that of tellurium, and the
inert gases therefore form a new group, which is like a buffer
between the extremely different halogen elements and alkali metals;
just as the metals of the eighth group intervene between manganese
in Group VIlA and copper, silver, and gold in Group IB. So it
is recognized that the elements of the argon family ere properly
placed as Group O, the periodic law is vindicated, and, in recog-
tion of their analogy with the noble metals, the elements concerned
are sometimes called the noble gases.

The periodic law, however,' needs no vindication. Modern

92 CHEMICAL THEORY

research, it is true, has modified it by causing the conception of
atomic number to displace that of atomic weight; but this has
served only to strengthen the law by removing its anomalies, so
that it has now become the supreme generalization concerning the
origin and constitution of matter as revealed by the inter-relations
of the elements.

SUMMARY

PERIODIC LAW ACCORDING TO MENDEL^EFF. The physical and
chemical properties of the elements and their compounds are
periodic functions of the atomic weights; or

If the elements are arranged in the order of increasing atomic
weight, their properties vary definitely from member to member
of the series, but return to a more or less similar value at fixed
points in the series.

USES OF THE PERIODIC LAW. Prediction of unknown elements.
Correction of atomic weight values. Stimulation of thought and
research regarding the elements.

CHAPTER V
THE MODERN VIEW OF THE ATOM

When the atom was introduced into science by Daltori it
appeared in a theory brought forward to account for the laws of
chemical combination; and since it was supposed to be an ultimate
particle of matter the question of its structure did not arise, for
structure involves parts. Newton had supposed that matter con-
sisted of "solid, hard, impenetrable particles"; and although the
atoms of different elements differed in weight, and presumably
therefore in size, the reason for this was no more an active question
than is the reason for the difference in size of the marbles in a
bag to the boy who plays with them.

The question what the atoms were made of was, nevertheless,
soon raised by Prout; and if it had been conceded that the atoms
were made of hydrogen the great diversity of properties between
the elements must have caused inquiry as to how one primordial
material could give rise to such diversity. This inquiry might
have become more urgent when periodicity of properties amongst
the elements was discovered; but, owing to the suppression of
Prout's idea, no inquiry regarding atomic constitution appears to
have been made until recently. The generalization of Mendeldeff,
however, that the properties of the elements are periodic functions
of their atomic weights, appears inadequate apart from some idea
regarding atomic constitution. For, consider the two short periods:

Li Be B C N F Ne

Na Mg Al Si P S Cl Ar.

Continuous increase of atomic weight is connected with progressive
change of properties from Li to Ne; but why should .ibis progres-
sive change stop at Ne; why should the addition of about 3 units
of atomic weight to Ne produce an element (Na) which with some
modification reproduces the properties of Li? This fact was long-
ago represented by de Chancourtois by means of the " telluric

94 CHEMICAL THEORY

screw ", a spiral curve on which the elements were marked; but
the representation of a fact is a very different thing from its
explanation. There is no explanation, unless the elucidation of
atomic constitution can provide it. So the periodic system
demands a theory of atomic constitution to give it meaning.

The facts of electrolysis investigated by Davy, the electro-
chemical theory of Berzelius, and the laws of electrolysis estab-
lished by Faraday, have some bearing on the constitution of
the atom, though this was not realized by these chemists.
Metals were elements whose atoms could carry positive charges
and travel to the negative electrode or cathode during electro-
lysis; non-metals were elements whose atoms carried negative
charges and in electrolysis travelled to the positive electrode or
anode. Thus elements were distinguished as electropositive or
electronegative, and electricity and chemical affinity were seen to
be closely allied; but these were forms or components of energy
rather than of matter; and that electricity itself could form part
of a material atom was an idea not entertained.

Nevertheless Faraday showed that an ion during electrolysis
was always associated with a fixed quantity of electricity, a
bivalent ion being associated with twice as much electricity as a
univalent ion. This fact is now interpreted as signifying that
electricity, like matter, is atomic, but such a conclusion was not
reached by Faraday. The smallest quantity of electricity associated
with an atom of matter in electrolysis was called by Johnstone
Stoney, in 1874, an electron, and so at that date was recognized as
an atom of electricity.

No real beginning, however, was made towards any knowledge
regarding the constitution of the atoms of matter until these atoms
themselves furnished evidence regarding their internal contents
and structure. The first evidence of this kind was the outcome of
the work of Crookes on high vacua. Crookes found that when an
electric discharge took place through a high vacuum rays travelled
from the cathode in straight lines, that these rays caused the o-lass
of the containing vessel to fluoresce, but that they were intercepted
by a material object which thus caused a shadow. These rays were
considered by Crookes to consist of matter in an ultra-gaseous
state, and they were subsequently called " cathode rays " or
"cathode particles". Sir J. J. Thomson, in 1897, investigated
these particles, found that they travelled with a velocity about

THE MODERN VIEW OF THE ATOM 95

one-tenth that of light, and proved that their mass was l/1850th
part of the mass of a hydrogen atom. The most significant dis-
covery concerning them, however, was that their nature was in-
dependent of the gas originally present in the vacuum tube, and
of the metal used as cathode. Consequently they were judged to
be not only disintegration products of material atoms, but in-
variable constituents of those atoms. This was the first piece of
evidence regarding the constitution of the atoms of matter.

The next evidence was furnished by the facts of radioactivity,
which began to be discovered after attention had been dtawn by
Kontgen to the fluorescence of the glass of the Crookes tube as the
source of those peculiar rays called X-rays.

The radioactivity of uranium, radium, and thorium was found
to be caused by the emission of two kinds of particles known
respectively as a- and /3-particles; and after a time these particles
were recognized to be actually disintegration products of the atoms
of those heavy metals, and were likewise identified. Thus an
a-particle was found to be an atom of helium carrying a double
charge of positive electricity, and a /3-particle to be the same as a
cathode particle which was now also identified with the electron,
the atom of negative electricity.

Yet a-particles, i.e. positively charged atoms of helium, are not
the smallest known particles of matter, for there are ions of
hydrogen, or hydrogen nuclei, i.e. positively charged atoms of
hydrogen which o\ve their charge to the loss of an electron. It
has been inferred, without direct experimental evidence it is true,
that the atom of helium has been formed by the condensation of
four hydrogen atoms; i.e. four hydrogen nuclei plus four electrons
have produced one helium atom; and further, that two of the
electrons in the helium atom are detachable from that atom so as
to leave a helium ion or doubly charged helium atom, which is the
a-particle derived from radioactive matter. Incidentally it must
be noted, however, that helium ions are unknown in chemistry,
although they have been recognized in work upon positive rays;
we have no power of removing two electrons by chemical means
from the helium atom so as to produce a helium ion <jr a-particle.
Moreover, a-particles, ejected from the atoms of radioactive ele-
ments, soon take to themselves electrons, and become helium atoms,
as was show r n by Ramsay and Soddy.

A helium atom or ion has never been known to yield hydrogen

96 CHEMICAL THEORY

atoms or ions by disruption, and hydrogen atoms or ions have
never been observed as the products of spontaneous .radioactive
change. Nevertheless there is direct evidence that some of the
lighter atoms of matter contain hydrogen nuclei as integral parts
of their structure. This evidence has been furnished by the ex-
periments of Sir Ernest Rutherford, 1 who has shown that hydrogen
nuclei are discharged from the atoms of boron, nitrogen, fluorine,
sodium, aluminium, and phosphorus under bombardment by
a-particles; and it is significant that elements whose atomic
weights are multiples of four, i.e. carbon and oxygen, do not
yield hydrogen nuclei under such treatment. Therefore it is
concluded that the massive parts of those atoms which contain
hydrogen nuclei contain them as such in addition to the requisite
number of helium nuclei. For example: N = 14 = 3 He + 2H;
and F = 19 = 4 He + 3H.

Thus a clear idea has been reached concerning the different
parts of which all atoms are composed. They are composed of
hydrogen nuclei, helium nuclei (a-particles), and electrons (/3-par-
tides); and if each helium nucleus is regarded as reducible to four
hydrogen nuclei, then the atoms of matter consist of hydrogen
nuclei and electrons alone.

Now the electrons are atoms of negative electricity, and in an
electrically neutral atom these must be balanced, whatever their
number, by an equal number of atoms of positive electricity. These
atoms of positive electricity must be the hydrogen nuclei, which
are the only other constituents of a material atom; they are called
protons] so that every neutral atom is composed of protons and
electrons in equal numbers.

It is now desirable to discover something concerning the manner
of distribution of the protons and electrons in an atom; and*. the
tirst insight into this manner of distribution is gained by consider-
ing and placing in contrast two kinds of change which some atoms
can undergo: chemical change, and radioactive change.

The main characteristic of chemical change in general is that it
can be initiated by man and is reversible, whilst radioactive change
is beyond ijian's control and, as far as we know, is irreversible. It
is concluded from this and the conclusion is now supported by
strong evidence that chemical change touches only the surface of
the atom, whilst radioactive change affects its internal parts.
1 Trans. CJiem. Soc., 1922, 121, 400.

THE MODERN VIEW OF THE ATOM 97

That chemical change affects even the surface of an atom is an
idea which would not have been acceptable to Newton or Dalton,
who regarded the atoms of matter as unchangeable. Indeed, a
generation ago this idea would have been thought revolutionary.
Chemical affinity, manifested through valency, was a force exerted
by atoms, but exerted outside themselves; the atoms came un-
scathed through chemical change; they bore no superficial wounds
to show that they had been in action. That chemical change
actually affects and alters the surface of an atom is the idea which
underlies the present-day electronic theory of valency; and in
developing this theory it is well to begin with electrolysis.

It will be remembered that the electric charges upon the ions in
an electrolytic solution are due to definite quantities of electricity
which are the electrons. The question may be asked: whence do
the electrons come which are associated with the atoms of matter in
electrolysis? They are not brought into existence by the current;
they must therefore be derived from the compounds in solution.
Sodium chloride, for example, must contain electrons, which become
available as electric charges when this compound is dissolved in
water. Such a view is consistent with the theory of Arrhenius,
which supposes that when a salt or other electrolyte dissolves in
water it breaks up spontaneously into charged ions, which are
ready to carry or be carried by the current when it comes.

The late Sir William Ramsay represented the electron in a
sodium chloride molecule, and the behaviour of the molecule when
it dissolved in water, in the following manner:

NaECl ^ Na' + ECl'.

Thus an electron, as an atom of the chemical element electricity,
was*the binding material between the atoms of Na and Cl, but
when the salt was dissolved in water this electron which was previ-
ously shared in common by both atoms became attached solely to
the chlorine atom, with the consequence that the sodium atom by
the loss of negative electricity became positively charged and
functioned as a cation, whilst the chlorine atom, by the gain of
negative electricity consequent on having the electrorf to itself,
became negatively charged and was the chloride ion. This idea
has now been developed so that the neutral sodium atom is believed
to have on its surface one loosely .attached electron which it easily
parts with so as to become a univalent positive ion, whilst the

(D60)

98 CHEMICAL THEORY

neutral chlorine atom is believed to receive easily and accommodate
an electron, thus becoming a univalent negative ion, the chloride
ion. So the chemical union between sodium and chlorine is believed
to consist in the transfer of an electron from each sodium atom to
each chlorine atom.

Such a view, however, necessarily modifies the electrolytic dis-
sociation theory of Arrhenius. Ions are formed, according to the
most recent view, when the compound is produced; for the transfer
of electrons converts neutral atoms into ions. Sodium chloride is
thus always ionized, it would not otherwise be sodium chloride;
but in the solid state the positive and negative ions are held together
by electrostatic attraction, just as two oppositely charged pith balls
are attracted together. When, however, the salt is dissolved in
water, the electrostatic attraction gradually gives way, and the
ions become dissociated. We must not now speak of ionization as
a consequence of solution, for that has occurred already in the
formation of the salt; but electrolytic or ionic dissociation is a suit-
able term to describe the separation, through the medium of the
solvent, of the already existing ions.

The idea of chemical change thus briefly outlined, by which a
salt is produced by the transfer of electrons from metal to non-
metal, clearly suggests that in the act of union the metallic atom
loses part of its substance and the non-metallic atom receives an
addition to its substance; that is to say, the atoms themselves
suffer change in their substance. Thus in contrast with the old
doctrine of the uiichangeableness of the atoms of matter we have
the new doctrine that in every chemical change the atoms suffer
change. Nevertheless this change is superficial, and its reversi-
bility depends upon its superficiality.

It is quite different with regard to radioactive change. 'This
change is believed to affect the innermost recesses of the atom;
that is to say, the nucleus where are situated the protons which,
being hydrogen nuclei, constitute its effective mass. Such a change
is profound; it has not been initiated by man; and when it occurs
it is irrevocable, and so radically alters the properties of the atom
that elemental transmutation is said to take place. Sometimes
only electrons, or /3-particles, are ejected from the atoms of an
element by radioactive change; then the mass of an atom is un-
affected though the properties of the element are altered. If,
however, an a-particle, i.e. a helium nucleus, with an atomic weight

THE MODERN VIEW OF THE ATOM 99

of 4, is cast forth, the atom changes not only its chemical properties
but also its mass, for it becomes an atom having an atomic weight
4 units less.

It is possible, however, to show a little more clearly what is
the effect of the loss of a- and /3-particles by the atoms of an
element through radioactive change.

When a /3-particle is ejected from the nucleus of an element the
predominating positive charge there is increased by one unit, and
the atom as a whole will consequently carry one positive charge
if it was previously neutral; it would therefore become a univalent
positive ion if it did not at once take to itself an electron from
outside. When such an electron is assimilated it does not enter
the nucleus, or the radioactive change would be reversed; it remains
on the exterior of the atom as a valency electron. The negative
valency of the neutral atom, i.e. its power of appropriating electrons,
will consequently be reduced by one. With regard to the periodic
system, then, the loss of an electron from the nucleus of an atom
transfers that atom one place to the right in the table. Examples
of such change, with a corresponding effect on chemical properties,
are known; e.g. radium-B in Group IVB, by losing a /3-particle
from the nucleus of its atom, becomes radium-C in Group VB.
The fact, however, must be emphasized that there is no loss of
mass in this radioactive change, for even the loss of an electron
from the nucleus is compensated for by the gain of an electron at
the exterior. Therefore it appears that two separate elements, as
radium-B and radium-C, judging from the chemical properties,
seem to be, may have the same atomic weights. Such elements
are called isobares. Stress, however, must be laid upon the differ-
ence between the loss of an electron from the nucleus and from the
exterior or sheath of an atom. The latter is the accompaniment
of any change by which a neutral atom becomes a univalent posi-
tive ion; and it has no further significance.

When an a-particle is ejected from the nucleus of an atom two
kinds of loss are sustained by the atom. The first kind of loss is
that the atom loses two units of positive charge, since the a-particle
is a helium atom carrying two units of positive charge, #ris particle
being composed ultimately of four protons and two electrons, which
are associated together inseparably, so far as experience goes. Such
a change, when it has been compensated for by the eventual escape
of two electrons from the ^atomic surface, transfers an element two

100 CHEMICAL THEORY

places to the left in the periodic system, because the positive
valency of the neutral atom, i.e. its power of losing electrons, has
thereby been reduced by two. Various examples of such radio-
active change are known; e.g. the atom of radium in Group HA,
by losing an a-particle, becomes an atom of radium-emanation or
radon in Group O. The second kind of loss sustained by an atom
which ejects an a-particle is a loss of mass. Since an a-particle
is a helium nucleus with atomic weight of 4, 4 units of mass dis-
appear. So whilst the atomic weight of radium is 225*95, that of
radon is 222.

Now the atoms of the heaviest elements are capable of suc-
cessive radioactive changes in which both a- and ^-particles are
ejected. Suppose that an atom loses first an a-particle and then
successively two /3-particles. The loss of an a-particle moves the
element two places to the left in the periodic table, and the loss
of two /3-particles brings it back again two places to the right,
leaving it in the same group it occupied originally. This is the
kind of change uranium, for example, undergoes. Ui in Group VlB-
loses an a-particle, with 4 units of mass, and becomes UXj^ in
Group IVfi; UX 3 loses a /3-particle, becoming UX 2 in Group VB;
and then UX 2 also loses a /3-particle, becoming Un, which again is
in Group Vlfi. Thus Ui and Un are both in Group V!B, having
atoms which differ by 4 units of mass, but are chemically indis-
tinguishable. Such elements have been called by Soddy isotopes.

The following conclusion regarding the manner of distribution
of the protons and electrons in an atom has now been reached.

All the protons with some of the electrons are situated in the
nucleus of the atom; the remainder of the electrons are external
to the nucleus, and some of them are at the extreme superficial
limit of the atom in its sheath, that is to say. These external
electrons are removable by chemical change, but the nucleus remains
intact in all chemical changes, and is affected only in radioactive
change or when submitted to intense bombardment by a-particles
(Rutherford). In a neutral atom the number of protons must be
equal to the number of electrons; therefore the number of electrons
external tothe nucleus must be equal to the excess of protons over
electrons in the nucleus.

It thus appears that there is a number which is both the excess
of protons over electrons in the nucleus of an atom and the number
of electrons external to the nucleus when the atom is uncharged.

THE MODERN VIEW OF THE ATOM 101

This number characterizes the atom as regards its chemical pro-
per ties; it is called the atomic member. It does not depend upon
the number of protons in the nucleus; for the loss from the nucleus
of four protons and two electrons, constituting an u-particle, together
with two more electrons, leaves the atomic number the same and
the chemical properties the same, as, for example, was shown in the
case of the uraniums cited above. Nevertheless diminution in the
number of protons involves diminution in atomic weight. Since,
therefore, atomic weight may alter while chemical properties remain
identical, the generalization that the properties of the elements are
periodic functions of their atomic weights no longer appears strictly
true; for it is apparent the atomic number takes precedence over
atomic weight so far as chemical properties are concerned.

Now it has already been seen, in the discussion of the periodic
law according to Mendeleeff, that the question of atomic weight is
not paramount in deciding the position which an element is to
occupy in the system. The cases of argon and potassium, cobalt
and nickel, and tellurium and iodine will be recalled. In each of
these pairs of elements the first-named member has a greater atomic
weight than the second. Yet it was agreed to place the elements
in the order named because their properties demanded this in spite
of their atomic weights.

If every element is to have an atomic number, these numbers
cannot all be decided without reference to the total number of
existing elements from hydrogen to uranium. The idea of number-
ing the elements is not new; it was entertained by Newlands, who,
however, could not carry his law of octaves very far because he did
not recognize gaps in the procession of the elements. Mendeleeflf
recognized the gaps but did not lay stress upon atomic number.
NoMifficulty now arises in numbering the elements until the rare-
earth metals are reached; for the only vacant place previous to
these metals is that following molybdenum, where an undiscovered
analogue of manganese should be. Decision as to the number of
rare-earth metals existing, however, has always been a difficulty.

This difficulty has now been overcome by the discovery of an
experimental method of determining the atomic number of an
element. This method is the result of the work of Moseley on
X-ray spectra. 1 It was discovered by Moseley that these spectra

1 For an account of this work a special tt&tbook should be consulted, e.g. The Structure
of Matter, by Dr. J. A. Cranston. ^

102 CHEMICAL THEORY

are much simpler and more regular than the luminous radiation
spectra of the elements, since the principal lines of the X-ray
spectra of successive elements follow one another in regular grada-
tion like a flight of steps, so that a missing element would be
revealed by a gap in the series of spectra. Moreover, the vibration
frequency V of the principal line in the X-ray spectrum of an
element is connected with the atomic number N by the following
formula, where A is a constant:

V = A(N - I) 2 .

Moseley's results may be summarized thus:

1. Every element is characterized by an integer N, which
determines its X-ray spectrum.

2. This integer N, the atomic number of the element, is
identified with the numerical value of the charge of positive
electricity on the atomic nucleus.

3. The order of the atomic numbers is the same as that of the
atomic weights, except where the latter disagrees with the order
of the chemical properties.

It is interesting to report that the atomic number of uranium,
the last of the elements, is 92; thus there are 92 elements in all,
from hydrogen to uranium. Of these 87 have already been dis-
covered, and five remain to be discovered; these are two analogues
of manganese, numbers 43 and 75 ;* a rare-earth, number 61; a
halogen to follow iodine in group, number 85; and an alkali metal
to follow caesium in group, number 87. With regard to element
number 72, it is not yet quite decided whether this should be called
celtium or hafnium.

Further, the placing of argon and potassium, cobalt and nickel,
tellurium and iodine in their accepted order according to cherqjical
properties rather than atomic weights has been justified; for the
atomic numbers of these elements are: 18 and 19, 27 and 28, 52 and
53 respectively.

Since the atomic weight of an element must now be subordinated
to its atomic number, it follows that the statement of the periodic
law according to Mendel^efF must be modified.

Thus the statement that the properties of the elements are
periodic functions of their atomic weights becomes:

1 These appear recently to have been discovered, and it is proposed to name them
masurium (Ma) and rhenium (Re) respectively.

THE MODERN VIEW OF THE ATOM 103

The properties of the elements are periodic functions of their atomic
numbers.

This is the periodic law in its modern form.

Further, since the order o sequence and total number of the
elements are now definitely known, it is possible to develop the
periodic system itself in a form likely to be permanent. These are
the questions to be decided: (i) how many periods are there; (ii)
how many elements are there in each period?

Already a partial answer has been given to these questions; for
Mendeleeff showed short and long periods containing, if the inert
gases are included, 8 and 18 elements respectively; and, as was
stated on p. 77, Werner, in 1905, proposed a long period of
33 elements to include the rare earths. Rydberg, in 1897, 1 had at-
tempted a classification of the elements according to numbers derived
from their atomic weights, and Rydberg's system, corrected by means
of Moseley's atomic numbers, now furnishes a very simple formula
to express the numbers of the elements in successive periods.

The atomic numbers of the inert gases are these:

He Ne Ar Kr Xe Rd

2 10 18 36 54 86;

therefore the successive periods, which are completed by these
elements, contain the following numbers of elements:

2, 8, 8, 18, 18, 32
= 2 [I 2 , 2 2 , 2 2 , 3 2 , 3 2 , 4 2 ].

The successive periods or series represented by these numbers
are called the Rydberg series; and although no ultimate explana-
tion of these numerical relationships has yet been given, they are
tak^n to express the manner of arrangement of the elements in
series in the periodic system. There are thus six series or periods,
with the beginning of a seventh, containing the elements from
87 to 92. So the modern periodic system, according to the
Rydberg series, takes the form on p. 104.

If a clear idea of atomic number has now been gained, this will
furnish a more complete conception of the nature of isotopes. It
will be remembered that isotopes, according to the observations
and definition of Soddy, were elements which, having made

excursions into different groups of the periodic system, on account

*Zeit.jhy*. Chern., 1897, 14, 66.

PH
W
P5

cc;
O
O

W
H

PM
O

f-H

I S5
PHOO

Qoo

nS Sc3

pqo

i I

g
1

o
3s

-S

ST 1

ii
cq

(I
00

II
oo

C<l

II
8

104

THE MODERN VIEW OF THE ATOM 105

of radioactive changes in their atoms, were accommodated in the
same place in the system, either permanently, or else only tem-
porarily because further radioactive changes removed them from
that place. It will now be seen that since isotopy relates to
elements having the same atomic number irrespective of their
atomic weights, it is not necessarily limited to radioactive elements.
Indeed there is no a priori reason why the phenomenon should not
occur widely throughout the whole range of the elements. If, how-
ever, it did so occur, the phenomenon would result from different
atoms of what is chemically the same element with the same
atomic number having different atomic weights because of different
numbers of protons in their nuclei.

The idea that different atoms of the same element may have
slightly differing relative weights is not new. It was put forward
by Crookes in 1888, with reference to yttrium, in the following
words: "The atomic weight which we ascribe to yttrium therefore
merely represents a mean value around which the actual weights
of the individual atoms of the 'element 7 range within certain limits.
But if my conjecture is tenable, could we separate atom from atom,
we should find them varying within narrow limits on each side of
the mean."

If this possibility is admitted for yttrium, we cannot refuse to
consider it for other elements, as indeed Crookes realized. That
the atomic weight of any of the elements represents not the weight
of every atom of that element, but the mean weight of an un-
numbered host of those atoms, is an idea which evidently has some
connection with the subject of isotopes; but how can such an idea
be put to the test, and how can it be regarded as any other than
an unprofitable speculation? A banker in pre-war days would
weigh a hundred sovereigns instead of counting them, because he
knew the average weight of such a number to be constant. Yet
the sovereigns might have been weighed one by one on a delicate
balance, and differences in their weights detected. So the chemist
weighs many atoms of an element together, and finds the average
weight of the same number always the same. To detect differences
in individual weights, however, he would need to weigh the atoms
separately; but that he cannot do.

Yet an instrument called a mass-spectrograph has been devised
by which the atoms of an element are separated in such a manner,
when charged electrically, that tfiey register themselves on a photo-

106 CHEMICAL THEORY

graphic plate in positions which depend only on their individual
masses.

Sir J. J. Thomson began work upon this subject in 1912, carry-
ing out what was called positive-ray-analysis, because the " rays ",
now called " mass rays ", which produced the effects were positively
charged particles or ions. Thus Thomson separated gaseous neon,
with an atomic weight of 20-2, into atoms, most of which were
shown to have a relative weight of 20, and a much smaller number
a relative weight of 22. So it was demonstrated that the element
neon is a mixture of isotopes, its accepted atomic weight being the
mean of the atomic weights of the separate isotopes present in
the requisite numerical proportions, viz. 90 per cent of Ne 20 and
10 per cent of Ne 22 .

After the war Dr. F. W. Aston developed the method of
Thomson and elaborated the instrument, and thus has been able
to show by means of " mass spectra " that a large proportion of the
chemical elements are mixtures of isotopes. Up to the end of 1924,
56 elements had been examined by Aston and others, and of these
25 were found to consist of identical atoms, and 31 of mixtures
of isotopes.

Now the elements, all of whose atoms are identical in weight,
are also elements whose accepted atomic weights approximate very
closely to whole numbers, whilst among the elements which are
mixtures of isotopes are those whose atomic weights are far
removed from whole numbers. Examples of the former are:
= 12-00, N = 14-01, O = 16, S = 32-06, P = 31-02, Cr = 52-00;
and of the latter: Mg = 24-32, 01 = 35-46, Cu = 63-57, Zn = 65-38,
Se = 79-20, Kr = 82-92, Hg = 200-60.

It is obvious, however, that the possession by an element of an
atomic weight which is approximately a whole number is no proof
that the element does not consist of isotopes; for the mean of a
number of isotopes might happen to be nearly or exactly a whole
number. Such is the case not only with Kr = 82-92 but
also with Br = 79-92, which is a mixture of the isotopes Br rr>
and Br 81 .

The question here arises how many isotopes of an element may
there be, and what range of atomic weight, or mass number as it is
now called, is possible. The answer seems to be that 8 is the maxi-
mum difference in mass number, and therefore 9 the maximum
number of isotopes possessed by any element. Thus the following

THE MODERN VIEW OF THE ATOM

107

data for tin and xenon, as well as for potassium and copper, are
given by Aston (Cliem. Soc. Ann. Report, 1924):

Element.

Atomic
Number.

Atomic
Weight.

Minimum
Number
of Isotopes.

Mass Numbers of
Isotopes in order
of Intensity.

118-70

7(8)

(120, 118, 116, 124, 119,
\ 117, 122, (121)

130-20

7(9)

(129, 132, 131, 134, 136,
\ 128, 130, (126), (124)

3D -10

39, 41

63-57

63, 65

The greatest numbers of isotopes are possessed by elements of
even atomic number; indeed elements of odd atomic number, e.g.
potassium and copper, seem to consist of not more than two isotopes
whose mass numbers differ by two units.

It appears that isobares are present amongst the above isotopes;
e.g. Sn 124 and Xe 124 , provided the latter value is substantiated. Yet
although some of their atoms have equal masses, tin and xenon are
entirely distinct elements; for whilst the number of protons in the
nuclei of their atoms may be the same, the numbers of electrons,
and hence of electric charges therein, must differ. Other examples
of isobares are furnished by Ar 40 , Ca 40 and Ge 74 , Se 74 . A striking
fact in connection with the subject of isobares is that tellurium
with mass numbers 120, 130, 126, and atomic weight 12*7-5, or
more probably 127-8, shares all those numbers with xenon, with
which it is thus trebly isobaric. Iodine, however, with atomic
weight estimated to be 126-92, consists of I 127 only.

The mass numbers of atomic isotopes are always given as whole
numbers, whilst the estimated atomic weights of the elements are
oft&i fractional. Thus is raised a question which it is necessary to
discuss. The mass of an atom is due to its protons, and a proton is
a hydrogen atom minus an electron, which thus has a mass of
1-008 when = 16-00.

The oxygen atom contains 16 protons, yet its mass is not quite
16 times the mass of a proton. This loss of mass is attributed to
a " packing effect " in the nucleus, where the additive J$iw of mass
is not obeyed. On the basis of O = 16-00, however, the atomic
masses of the isotopes are known to conform to the whole number
rule except for a few small variations.

Thus since the masses of the individual atoms of all the elements

108 CHEMICAL THEORY

are, within a close approximation, whole numbers when O = 16*00,
it is clear that the fractional atomic weights with which we have
been long familiar, and which we are still compelled to employ in
accurate analytical work, are averages due to mixtures of isotopic
atoms; and so we understand why the atomic weight of an element
which does not exhibit isotopy is very nearly, if not precisely, a
whole number.

A further question connected with isotopy is this: If the same
element is found in different parts of the world, will it always
have the same estimated atomic weight? The idea that the same
element, obtained from different sources, may have differing atomic
weights is a disturbing one which strikes at the foundation of all
accurate chemical work. For example, a redetermination of the
atomic weight of antimony has resulted in a drastic change from
120-2 to 121-76. Can it be that different isotopic mixtures of
antimony atoms have yielded those discordant results? This is
believed not to be the case; but it is thought, rather, that the new
figure is a correction of the old. Indeed there is abundant evidence
that the atomic weights of naturally occurring elements which
are not of radioactive origin are always constant whatever the
sources of the elements. Thus cobalt and nickel of meteoric origin
have the same atomic weights as the terrestrial elements, and the
same is true of silicons from cosmic and terrestrial sources. Never-
theless it is highly desirable for those who undertake the redeter-
mination of atomic weights to state the source of their material.

So far as present knowledge goes, the experimental atomic
weights of normal inactive elements are still to be regarded as
constants of nature, since nature lias effectively mixed her isotopes,
and never sorts them out again. Nevertheless the proved existence
of isotopes is a challenge to man to separate them. Chemical
methods of separation are unavailing, since isotopes do not differ
in chemical properties; but physical methods, such as fractional
diffusion and distillation, have been attempted with some success
in the case of chlorine, mercury, and perhaps zinc (Cliem. Soc.
Ann. Report, 1922).

The ca$e is very different with an element which is the residue
of radioactive change. The uranium atom, for example, with atomic
weight 238-17, passes through a succession of radioactive changes,
in which it loses eight a-particles as well as /3-particles, the final
product being lead. This lead, therefore, should have an atomic

THE MODERN VIEW OF THE ATOM 10!)

weight of 238-17 32 = 206-17. Now uranium minerals arc
found to contain a small proportion of lead, and the lead extracted
from such minerals has been estimated to have an atomic weight
of 206*46, whilst that of ordinary lead, not associated with radio-
active material, is invariably found to be 207*20.

Again, the thorium atom, with atomic weight 232*15, loses six
a-particles in the series of radioactive changes which end in lead;
therefore the lead derived from thorium would be expected to have
an atomic weight of 232*15 24 = 208-15. Lead obtained from
thorite has been found experimentally to have an atomic weight
of 207*77; which, although a little lower than that it would be if
the lead were derived from thorium alone, is considerably higher
than the atomic weight of ordinary lead.

Uranium-lead and thorium-lead are indistinguishable from
ordinary lead in chemical properties and in all physical properties
except density. Thus three leads have been named which, chemi-
cally speaking, are one lead. These, with their experimental
atomic weights and densities, are:

Uranium-lead.

Thorium-lead.

Ordinary Lead.

Atomic weight
Density
Atomic volume

206-08
11-213

18-28

207-77
11-376
18-26

207-20
11-352
18-25

It is seen, moreover, that the densities of these leads vary as their
atomic weights, so that their atomic volumes are constant.

There is more to tell about the internal structure of an atom.
Since an atom consists of a nucleus and surrounding electrons, it is
desirable to gain some idea of the size of the nucleus as compared
with that of the atom as a whole. The experiments of Rutherford
on the scattering of a-particles yield the desired information. These
experiments have already been referred to because they furnish
information regarding the description of the atoms of some of the
lighter elements; but if attention is concentrated on the tracks of
the a-particles themselves rather than on the havoc they work by
their bombardment, some quite different information is obtained.

When the a-particles from a radioactive source traverse a gas,
their tracks can be made visible by the condensation of super-
saturated aqueous vapour which occurs along them. Thus it is
discovered that whilst some of the a-particles undergo sharp

110 CHEMICAL THEORY

deflections, these deflections are many times fewer than they would
be if collision with an atom as a whole caused deflection. So it is
concluded that a large proportion of the a-particles pass through
the atoms as through empty space, and that only when an a-
particle collides with or comes very near to a nucleus is it deflected.
Thus it has been calculated, owing to the rarity of these collisions,
that the diameter of the nucleus of an atom is about one ten-
thousandth part of the diameter of the entire atom.

On account of this relation of the nucleus to the surrounding
electrons, an atom of matter has been compared with the solar
system, the nucleus being analogous to the sun, and the electrons to
the planets.

The different planets of the solar system have different orbits,
and they revolve round the sun. Have the planetary electrons of
an atom different orbits, and do they revolve round the nucleus in
their several orbits? This is a question to which some answer
must now be given, though the details of the answer are not yet
beyond the region of controversy.

The number of electrons external to the nucleus of an atom of
an element is the same as the atomic number of the element; and
thus with elements of high atomic number these electrons are
numerous. In the atom of uranium there are 92 electrons external
to the nucleus, and it is not to be supposed that these are all
situated or revolve on the surface of a single shell or envelope.
Rather must it be supposed that the electrons are distributed in a
number of shells which succeed one another like the layers of an
onion. How many electrons are present in each layer is a question
on which in some cases there is difference of opinion. It must be
remembered, however, that since the electrons on the outer layer
of the atom, which is called the sheath, are the valency elections,
their number will be related to the valency of the element; and
further, that since the atom of an inert gas has no valency, the
sheath of such an atom will presumably consist of a completed
layer of electrons, to or from which no electron can be added or
removed.

To form a mental picture of the structure of the atoms of
matter it will be best to begin at the beginning, with hydrogen.
The neutral hydrogen atom consists of 1 proton + 1 electron, and
the helium atom of 4 protons + 4 electrons, 2 of these electrons
being bound up in the nucleus with the 4 protons, and the other

THE MODERN VIEW OF THE ATOM

111

two being in a sheath which is complete since the helium atom
manifests no valency, although it can exist momentarily without
these two electrons as an a-particle ejected from a heavy atom
during radioactive change. A consideration of the hydrogen atom
under different circumstances will illuminate the subject of valency.
When a hydrogen atom becomes a cation, i.e. the hydrion, as in the
formation, say, of an acid in aqueous solution, this atom parts
with its solitary electron and becomes reduced to a naked proton.
It is possible, however, for the hydrogen atom to assimilate
a second electron and so become an anion. This is shown by the
fact that lithium hydride, LiH, yields on electrolysis hydrogen at
the cinode. Thus the hydrogen atom shows a tendency either to
lose its single valency electron, or more rarely to gain another, by
which means it would assume an external configuration character-
istic of the uncharged helium atom.

Now consider the short period Li to Ne.

Atomic number

Number of electrons in)
sheath of neutral atom/

Valency \ normal
(Abegg) /contra

+ 2
1>

+3
-5

-2
+ 6

The atomic number of each atom in this period exceeds the number
of electrons in the sheath of the neutral atom by two, because there
are two electrons in the under layer which correspond with the
two electrons in the complete sheath of the helium atom. Now
when the significance of the numbers in the above table is under-
stood, the nature of valency will stand revealed.

It is the great merit of Abegg 1 to have drawn attention to the
fact that the sum of the hydrogen and oxygen valencies of a
number of elements is equal to 8, e.g.

SiH 4 PH 3 SH 2 C1H
SiO 2 P 2 O 6 SO 3 C1 2 O 7 ,

and to have derived therefrom the theory of normal #nd contra-
valencies, an example of which is given in the table above. Ac-
cording to Abegg, the normal valencies of an element are the more
usual and characteristic; the contravalencies are more seldom

i Z. anorg. Chem. (1904), $9, 330.

112 CHEMICAL THEORY

exercised, at any rate with the more extreme members of a period.
Now if, in addition to this, valencies are regarded as positive or
negative according to whether they are exercised towards electro-
negative or electropositive elements respectively, and it is supposed
that the actual exercise of valency implies the loss or gain of
electrons by the sheath of an atom, and further, that every atom
undergoing chemical combination tends to assume the condition of
an inert gas as regards its sheath, then the following ideas regard-
ing the valencies of the elements of the first short period follow.

Lithium, with 1 electron in its sheath, can assume the external
condition of an inert gas either by losing 1 electron, so as to
simulate the helium atom which precedes it, or by gaining 7
electrons, so as to have a sheath identical with that of a neon atom.
It is, however, much easier for an atom to lose 1 electron than gain
7; hence lithium invariably manifests a valency of +1 by becoming
a cation carrying one positive charge, rather than a valency of 7,
that is, an anion carrying seven negative charges.

Similarly beryllium loses 2 electrons on ionization, becoming a
bivalent cation, rather than gaining 6 electrons to become a sexa-
valent anion.

With carbon, however, the chances of losing or gaining electrons
are about equal; and with nitrogen the alternative of the loss or
gain of electrons also exists. Oxygen and fluorine, however, are
too electronegative ever to become cations by losing electrons, the
contravalencies, at any rate in the case of fluorine, being entirely
latent.

The same considerations apply to the next short period from
sodium to argon, the only difference being that the atoms of both
the inert gases, neon and argon, to the external configuration of
which the intervening elements tend to conform when they enter
into chemical union, are both alike in having 8 electrons in their
sheaths.

So far it appears that valency depends on the number of
electrons in the sheath of an atom; and whether that atom exercises
positive or negative valency depends upon whether it more easily
loses or gains electrons, so as to present a completed sheath on its
outer surface.

Langmuir, who, following G. N. Lewis, has developed this idea, 1
carried it further, and applied it to the whole of the periodic

1 J. Anier. Chem. Su., 1919, 41 868.

THE MODERN VIEW OF THE ATOM

113

system, so that whilst neon and argon at the end of the two short
periods have each 8 electrons in their sheaths, krypton and xenon
have each 18, and radon has 32. Thus Langmuir postulates 1 that
" the electrons in atoms tend to surround the nucleus in successive
layers containing 2, 8, 8, 18, 18, and 32 electrons respectively*'.

In the long periods, however, it is not possible to connect valency
with the electronic content of the atomic sheath in the same simple
fashion as in the short periods. Consider the period of 18 elements
from K to Kr. This period begins and ends like the preceding
short period; i.e. K, Ca, Sc resemble Na, Mg, Al in valency, and
As, Se, Br similarly resemble P, S, 01; but with the intermediate
elements, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, more compli-
cated relations appear. In one sense the elements K to Mn resemble
a short period ; e.g. with regard to the oxides

K 2 O, CaO, Sc 2 3 , Ti0 2 , V 2 O 6 , CrO 3 , Mn 2 O 7 .

Thus Mn 2 O 7 resembles C1 2 O 7 , permanganates being isomorphous

with perchlorates, so that a similar arrangement of electrons on

the atomic sheaths of Cl and Mn may be inferred. The oxides

Cu 2 O, ZnO, Ga 2 O 3 , GeO 2 , As 2 O 5 , SeO 3 , ,

again, less perfectly reproduce the relations of the oxides of the
first short period. Moreover, a new phenomenon occurs in the
centre of this long period; this is reducibility of the higher com-
pounds with the loss of single units of valency, and the simul-
taneous appearance of coloured ions.

Thus salts corresponding with the following oxides have coloured

ions:

TiO

CrO

MnO

FeO

CoO

TiA

CrA

MnA

Fe 2 3

CoA

(Ti0 2 )

MnOj

NiO

(Cu 2 O)
CuO

VA - -

CrO, MnO, FeO,

In view of these considerations, MendeleefTs division of the long
periods into the elements of the A and B sub-groups may be brought
forward again, thus:

Elements of A Sub-groups

Sc ' Ti

Fe Co Ni

Elements of B Sub-groups

Ga Ge

(D60)

* Science, July, 1921.

114

CHEMICAL THEORY

and so it may be pointed out that the triad (Fe Co Ni) appears to
function like a single element, i.e. like an inert gas at the end of a
period; and although Ni cannot bo compared with an inert gas,
Pt may, because it is so inert. Thus if the non-valent, inert gases
are regarded as " standards of atomic stability ", Ni, Pd, Pt, or more
accurately, imaginary inert forms of these elements, are sometimes
regarded as " sub-standards of atomic stability ".

Now if Langmuir's postulate is examined in the light of all
these considerations, it will appear difficult to account for valency
by reference to the electrons supposed to be contained in the atomic
sheaths of the neutral atoms. Nickel, for example, would have
10 electrons, and require 8 to complete the sheath; copper would
have 11 electrons, and require 7 to complete the sheath, and so on.
And whilst the existence of Ni(CO) 4 would seem to justify a valency
of 8 for nickel, there are no properties of copper or the succeeding
elements to show such high valencies as would thus be attributed
to them.

Consequently attention may be drawn to other views regarding
the distribution of the electrons in an atom.

In 1921 Bury 1 modified Langmuir's theory by assuming that
the number of electrons in a completed sheath of an atom never
exceeds 8; and in the same year Bohr, from a consideration of the
spectra of the elements (see later), adopted a similar view.

The following table sets forth the atomic structures of the nou-
valent gases according to Bohr. The numbers in brackets are the
atomic numbers of the elements, and therefore the numbers of
electrons external to the nucleus.

Orbits, numbered outwards
from Nucleus.

Helium (2)

Neon (10)

Argon (18)

Krypton (36)

Xenon (54)
Kadon (86)

2
2

8
8

18
18

18
82

8
18

Aii advantage of this view is that it provides for the addition
of electrons either in the sheath of an atom or in an orbit or shell
below it. When the latter occurs there need be no change of

i/. Amer. Chem. Soc., 1921, #, 1602.

THE MODERN VIEW OF THE ATOM 115

valency in passing from one element to the next, as is shown, for
example, in the chlorides

VC1 2 , CrCla, MnCljj, FeCl., CoCl 2 , NiCl 2 , CuCL 2 , ZnCl 2 .
The idea is specially helpful, however, in accounting for the 14
elements of the rare earths, all of which have the same valency.
Successive additions of electrons are here supposed to be made to
the electrons in the fourth orbit, so increasing these from 18 to 32.

A final question, so far as the present study of atomic structure
is concerned, is that of the activities of the electrons within the
atom. Regarding these activities the views of the physicist and
the chemist appear to be at variance. The physicist believes the
electrons to be revolving round the nucleus in their several orbits
as the planets revolve round the sun. To him an electron at rest
is as unthinkable as a planet in such a condition. The chemist,
however, is well content to think of a stationary electron; indeed
he seems to demand it by his ideas of valency and the constitution
of compounds. How can chemical compounds be formed without
points of attachment between the atoms, and how can points of
attachment be provided by swiftly revolving electrons? It is true
that an electrolyte like sodium chloride might exist; for the chemist
has learned to regard its atoms as held together not by bonds, but
by electrostatic attraction between oppositely charged ions. But
what is to be said about such compounds as methane and the host
of organic substances, concerning whose structure and stereo-
chemistry the chemist has such elaborate and satisfying ideas,
based upon the doctrine of bonds?

The physicist, however, needs to account for those beautiful
phenomena, the luminous spectra of the elements. It used to be
asked: how can the atom of iron vibrate in hundreds of ways at
once so as to give rise to the hundreds of lines in its luminous
spectrum? It is sufficient now to ask how the hydrogen atom,
consisting of one proton and one electron, can vibrate in various
ways so as to produce the various lines in its spectrum. We are
indebted to Bohr 1 for an explanation of this phenomenon, based
on Planck's Quantum Theory of Energy, which now finds general
acceptance.

Imagine an atom with revolving electrons which are radiating
energy into space. If this radiation were continuous, the electrons

1 Vide The Theory of Spectra and Atomic Cdnstitution, by Niels Bohr : Cambridge University
Press, 1924,

116 CHEMICAL THEORY

would be continually losing energy, and in consequence continually
approaching the nucleus in a spiral path. Moreover, such continu-
ous radiation could not produce a discontinuous line spectrum. To
avoid the nemesis of the atom by the collision of planetary electrons
and nucleus, it is assumed that a revolving electron loses no energy
so long as it remains in a single orbit; that it is only change of
orbit which is accompanied by change of energy; a loss of a definite
amount of energy, the so-called quantum, will thus accompany the
fall of an electron from one orbit to that beneath it, i.e. nearer to
the nucleus; and a corresponding gain of energy will accompany
the restoration of the fallen electron to its former state. It is now
easy to understand, if there are numerous possible orbits, that each
kind of fall gives rise to a particular radiation which produces its
own line in the spectrum; and that the several lines occurring
simultaneously in the hydrogen spectrum are produced by corre-
sponding simultaneous falls from several different orbits in the
peripheries of hydrogen atoms with their electrons in several
different states, although each atom contains only one electron.

The apparently irreconcilable views of the physicist and chemist
may be expressed thus: to the physicist an atom is a hive of
activity, a home of swarming electrons; to the chemist it is an
abode, if not of rest, then of nothing more than vibratory motion of
electrons about their mean positions. Can these views be recon-
ciled? It is possible that they may be if a revolving electron can
be considered to be more in one place than any other, if there is
any point through which it passes very frequently whilst other-
wise tracing out divergent paths. This is impossible if an electron
describes circles in the same plane round the nucleus as centre, or
if its path is a simple ellipse, like the path of a planet, with the
nucleus at one of the foci of the ellipse.

If, however, the motion of an electron is compounded of a cir-
cular or elliptical motion, and a circular motion at right angles to
it, the path travelled will be precessional upon the surface of an
ellipsoid, 1 i.e. it will be represented by a series of curved lines
whose directions are constantly altering so as to cover the whole
surface of tfre ellipsoid very much as the coloured and twisted lines
on an ornamental glass marble cover its surface. The consequence
of such a motion will be that the rotating electron will pass, during

1 An ellipsoid is the solid figure formed by rotating an ellipse about its major axis, just as
a sphere is the solid figure formed by rotating a circle about its diameter.

THE MODERN VIEW OF THE ATOM 117

one cycle, many times through two points, which are at the ex-
tremities of the major axis of the ellipsoid, but only once through
every other point.

Such a conception, which is due to J. D. Main Smith, provides
for the localization of an electron, as well as satisfying some re-
quirements of the physicist. Whether, however, it will suffice to
account for both the physical and the chemical properties of the
atom cannot yet be said. Meanwhile the idea of stationary elec-
trons within or upon the surface of an atom is so very valuable
a contribution to the theory of chemical structure that it will be
adopted and developed in the next chapter, which deals with the
joodern view of the molecule.

CHAPTER VI
THE MODERN VIEW OF THE MOLECULE

The student of chemical history is aware that two views have
been held regarding the structure of chemical compounds. The
first view was expressed in the electrochemical theory of Berzelius,
which postulated electricity as the binding force between atoms, so
that a molecule consisted of atoms held in electrical equilibrium by
mutual attractions.

There were two kinds of electricity, and each atom in a com-
pound possessed some of both kinds, but in unequal quantities, so
that a positive or negative charge preponderated, according to
whether the atom was metallic and electropositive, with a larger
positive than negative charge, or non-metallic and electronegative,
with a larger negative than positive charge. Thus it followed that
every molecule consisted of two parts, a positive and a negative
part; and these parts in turn might consist each of two smaller
positive and negative parts, and so on, down to the individual
atoms. For example, the double salt potassium-alum, apart from
its water of crystallization, would be accounted for somewhat in

this way:

+
KOS0 3 A1 2 O 3 3SO 3

+ +

KO SO 3 A1A 3S0 3

+ - + - + - f + - 1

K O S 3 A1 2 3 3[S Oj.

This was the dualistic system] and it was successful in account-
ing for the structure of electrolytes, which are polar compounds,
but failed when applied to organic compounds, which are non~
electrolytes or non-polar compounds. Thus if every compound is
composed of positive and negative parts, in equilibrium, what, it
may be asked, are these parts in such a compound as CH 4 ; and
again, if + and parts are balanced in CH 3 COOH, how is it

118

THE MODERN VIEW OF THE MOLECULE 119

possible for electropositive hydrogen to be replaced by electro-
negative chlorine, so as to produce CC1 3 COOH?

In view of questions like these, Dumas propounded a second
view in his unitary system of chemical compounds, in which every
compound formed a complete whole, and did not therefore consist
of two opposite and balanced parts. He thus referred the pro-
perties of a compound to its type rather than to the properties of
its constituent atoms. The consequence was that unitary views
prevailed and dualism was discredited. When, therefore, the
doctrine of valency was developed, graphic formulae with " bonds "
were employed indiscriminately to represent the structure both
of electrolytes and non-electrolytes.

The electrolytic dissociation theory of Arrhenius, however,
marked a return in part to dualism; and that theory, together with
the properties of solutions to which it was related, emphasized the
real difference which exists between electrolytes such as sodium
chloride, and non-electrolytos such as chloroform. Since, however,
" bonds " were supposed equally to join the atoms of sodium and
chlorine in sodium chloride, and carbon, hydrogen, and chlorine in
chloroform, simple solution in water involved the breaking of bonds
in one case, but not in the other; and it was difficult to find a valid
reason for such an extraordinary difference of behaviour of differ-
ent compounds as they dissolved in water.

Now it has been seen that the electronic theory of valency, so
far as it was developed in the last chapter, appears to deal with
valency in electrolytes alone. This is true of Ramsay's idea of
valency, and also of the theory of transference of electrons during
the combination, say, of sodium and chlorine to form sodium
chloride; so that a conception of the molecule is reached resembling
thAt of Berzelius, because it represents atoms, or, more strictly,
charged ions, as held together by electrostatic attraction. There is
difficulty, however, in applying this simple theory to all molecules;
and to realize the difficulty it is only necessary to consider the
molecule C1 2 . Sodium and chlorine combine because of an electro-
chemical difference between these elements; the sodium atom loses
an electron which the chlorine atom, because of its different
chemical nature, readily takes up. No such reason can account
for the union of two chemically identical chlorine atoms to form
a molecule, so that there cannot be transfer of electrons in such
a case.

120

CHEMICAL THEORY

This difficulty is met by a conception due to G. N. Lewis and
developed by Langmuir: the conception of covalency, as distinct
from electrovalency, which is the kind of valency hitherto con-
sidered. Now an atom of chlorine has 7 electrons in its sheath,
and requires 1 to complete the octet characteristic of the sheath of
an inert gas. Such an atom, however, cannot gain its required
electron from a similar neighbouring atom, and even if it did it
would become a chloride ion such as does not exist in chlorine gas.
It is possible, however, for two chlorine atoms, with identical
requirements to satisfy these requirements, mutually, by the sharing
of a pair of electrons, each chlorine atom providing one electron of
the pair. The accompanying figure makes this plain.

Thus two octets containing only 14 electrons between them, are
possible because two of these electrons are common to both octets.
The shared electrons held in common by both chlorine atoms,
which are shown in the figure within the rectangle, constitute a
duplet. This duplet is a unit of covalency, and is equivalent to a
single valency bond. According to Langmuir, it is the only sort of
bond, and it represents the kind of union which exists between
the atoms of compounds which are not electrolytes.

Thus, by this view, there are two kinds of valency: electro-
valency and covalency. The theory of electrovalency is the
modern equivalent of the dualistic theory of Berzelius; that of
covalency corresponds with the unitary theory of Dumas.

Covalency may now be further illustrated. The molecule O 2
consists of a pair of atoms, each of which separately has six
electrons in its sheath, and therefore requires two electrons to
complete the octet. Two oxygen atoms can combine together to
produce a pair of octets if each atom shares two electrons with its
neighbour, producing a pair of duplets representing a double bond,
O = O, thus:

^37^

Similarly, carbon dioxide, = C = O, can be represented by

THE MODERN VIEW OF THE MOLECULJS

121

showing the carbon atom sharing two duplets with each oxygen
atom thus:

The combination of hydrogen with oxygen, and with chlorine, to
form water and hydrogen chloride respectively, may now be con-
sidered. Since water is a non-electrolyte, the two hydrogen atoms
in each molecule are supposed to unite with the oxygen atom by
covalency, in a manner which may be represented thus:

O + II

IIOH

Thus each hydrogen atom completes its sheath of two electrons,
and the oxygen atom its sheath of eight.

As regards hydrogen chloride, since in the anhydrous state this
compound is a non-electrolyte, its molecule may be represented
thus, the atoms of hydrogen and chlorine being united by covalency :

In presence of water, however, the covalency bond is broken, and
the hydrogen chloride is ionized, becoming hydrochloric acid. It is
now believed, however, that the hydrogen ion of hydrochloric acid
is hydrated, that, indeed, it is an unstable oxonium ion (OH 3 )"
analogous to the ammonium ion (NH 4 )', and formed thus: 1

OH 2 + HCI = (OH 3 )' + or.

This change would be represented structurally thus:

OH,

HCI

[OH

[CI]'

The explanation of this change of hydrogen chloride from a
neutral, non-electrolytic substance to a powerful acid through the
assimilation of water is to be found in the principle of symmetry.

1 Vide Lowry, Chemistry and Industry, 1923, 46.

122

CHEMICAL THEORY

The inolecuiar condition of non-ionized hydrogen chloride is an
unsymmetrical and therefore relatively unstable one, whilst the water
molecule requires two hydrogen nuclei to be attached to the four
remaining electrons to create perfect symmetry. Having attached
one hydrogen nucleus, however, the water molecule, which has now
become a positive [OH 3 ]' ion, is unable to take up another because,
being positively charged, it repels similarly charged hydrogen nuclei.
Hence the condition of things shown in the figure is arrived at.

Now when hydrogen chloride solution is mixed with ammonia,
neutralization takes place thus:

[OHJ- + [01]' + NH 3 [NHJ- + [C1J + H 2 0,

the neutralization being due to the fact that the ammonia molecule
appropriates another hydrogen nucleus to form the highly sym-
metrical ammonium ion with greater force than the water molecule
retains it. Nevertheless, just as [OH 3 ]' is unstable, easily losing a
hydrogen nucleus in presence of the hydroxidion of an alkali to
leave a stable OH 2 molecule, so [NHJ* is also unstable, and in
presence of much hydroxidion similarly loses a hydrogen nucleus
to provide hydrion and form water. That is why ammonium salts
evolve ammonia in presence of alkalis.

The student of organic chemistry is conversant with the theory
that the four valencies of the carbon atom are directed towards the
angular points of a regular tetrahedron; and he knows that this
theory has been most fruitful in elucidating the structure and
stereochemistry of carbon compounds. It is therefore necessary to
inquire how the stereochernical theory is related to the electronic
theory of valency, so far as this applies to the carbon atom.

In the preceding figures the atoms of all the elements except
hydrogen have been represented as having a cubical structure, or
rather, the eight electrons in the completed sheath of an atom have
been placet! at the angular points of a cube. Now a regular tetra-
hedron is the heinihedral form of the cube; if, therefore, the eight
electrons draw together into four pairs, two pairs being produced
by movements at right angles to the movements of the other two,
a tetrahedral figure will be produced (see fig., p. 123). Thus it is

THE MODERN VIEW OF THE MOLECUL1J

123

believed that whilst the cubical form of the atom as regards the

distribution of its electrons is preserved when chemical union is by
electrovalency as in sodium chloride, union by
covalency involves the distortion of the cubical
form into the tetrahedral So the structure of
carbon dioxide is represented by the following
figure. Thus the tetrahedral model of the
carbon atom is preserved, and union by single,
double, and triple bonds becomes union at an

angle, a side, and a face of the tetrahedron respectively, by one,

two, or three pairs of electrons.

It is troublesome though picturesque to represent atoms by
cubes, and molecules by numbers of united cubes. A simpler plan
is to use the ordinary atomic symbol surrounded by dots to re-
present electrons. Thus the molecule of chlorine, instead of being
represented as in fig. on p. 120, becomes

and other formulao are:
H 2 NaCl

H:H Na :C1:

O 2

o
H:C:::C:H,

which are equivalent to

H H Na- 01' O=O

0=0=0 H 0==C H.

So it appears that a pair of electrons, acting as a unit of covalency,
is equivalent to a single chemical bond. It is to be noted, more-
over, that with the exception of H, which has two electrons, and
Na, which is an ion with eight electrons in a lower layer, every
atom is represented as having or sharing eight electron^, since all
the electrons which bind two atoms together are shared by both
atoms concerned.

The question may be asked whether there is an absolute
distinction between electrovalency and covalency, or otherwise

12 1 CHEMICAL THEORY

between polar and non-polar compounds; or whether one kind of
valency merges into the other in a series of compounds, which thus
show transition from polarity to non-polarity. Thus, while electro-
valency is the mode of union in sodium chloride, it may be asked
whether all chlorides, metallic and non- metallic, are constituted
similarly. Now electrovalency must be the mode of union in the
case of the chloride of a univalent atom, for such an atom, e.g. Na,
cannot share the two electrons which are necessary to a covalent
bond; and covalency must be the mode of union when two similar
atoms unite, as in the case of C1 , for there is no reason for
electronic transfer.

It may be questioned, however, whether electrovalency is the
mode of union of the atoms in CC1 4 . Rather may it be supposed
that this compound, which resembles methane in inertness, is simi-
larly constituted, in which case the two molecules may be repre-
sented thus:

TI :C1:

H:C:H :C1:C:C1:.

H ":<JI:"

With SiCl 4 , however, the case is different, for this compound is
distinguished from CC1 4 by its reactivity towards water. Possibly,
therefore, the state of union between the silicon and chlorine atoms
may be represented as something intermediate between covalency
iind electrovalency, thus:

:Cl:

:Ci: JSi :C1: ,

the eight electrons which in covalency the silicon atom would Share
with the chlorine atoms being drawn towards these atoms, so that
the .silicon atom becomes charged positively and the chlorine atoms
negatively, though these charges are not sufficiently free to consti-
tute the compound an electrolyte. This can be understood if it
is remembered that the electrons in the sheath of a silicon atom
are fartheH* from the positively charged nucleus than those in the
sheath of a carbon atom, and so would be held less tenaciously by
the nucleus; or in other words, that the silicon atom is more electro-
positive than the carbon atom, and more ready to shed the electrons
of its sheath.

The accompanying photograph is of a chart which
was exhibited in the British Empire Exhibition
(Wembley, 1924), and is included here by courtesy
of the Royal Society and the National Physical
Laboratory.

Where the crystalline structure of an element has
been determined by the X-ray method, a model of
the space lattice is shown in the appropriate place
in the periodic table. It will be noted that the
majority of these lattices are cubic. The atomic
weight and the atomic number are given for each
element, the latter being in brackets.

D6o A andM.

THE MODERN VIEW OF THE MOLECUL^ 125

The stability of CC1 4 , as well as of CH 4 , as compared with SiCJ
and SiH 4 , is probably to be referred eventually, however, to the
inner structures of the atoms of the two elements. Between the
sheath of the carbon atom and its nucleus there is only the shell
of two electrons which constitute the sheath of the helium atom,
whilst in the case of the silicon atom the completed octet which
constitutes the sheath of the neon atom intervenes. Consequently
tliQ eight electrons of the completed sheath of the carbon atom in
a compound such as CC1 4 can assume the tetrahedral position
without strain, whereas the neon octet in the case of the silicon
atom makes tetrahedral symmetry and resulting stability more
difficult of attainment.

The idea of gradations between polar and non-polar molecules
is due to G. N. Lewis. Thus Lewis accounts for the properties of
hjftirogen chloride by supposing that the two electrons which are
shared between the hydrogen and chlorine atoms in the anhydrous
molecule may be displaced in the direction of the chlorine atom so
as to cause this to become negatively charged, whilst the hydrogen
atom becomes positively charged to the same degree, but that when
displacement is complete hydrion and chloridion result thus:

H:C1: H :C1: -

Langmuir, on the other hand, believes that the electrolytic pro-
perties of hydrogen chloride solution are due to the hydrogen
nucleus uniting with the water molecule to form a hydrated hydro-
gen ion, or oxonium ion, according to the theory already given.

The octet theory that is, the theory that the completed sheath
of an atom contains eight electrons accounts for the frequency
with which four atoms are joined to a central atom, to form a
molecule if the product is electrically neutral, or an ion if it is not.

Examples are the molecules CH 4 and OsO 4 , the cation NH 4 ', and
the anions Si0 4 "", P0 4 '", S0 4 ", Mn0 4 ", MnO/, CIO/, which are
formulated thus:

:0:

~ :0: "

~ :6:

:0:

:0:Si:O:

:O:i':O:

:O:S:():

:O:Mu:O:

:O:.Mn:O:

:O:C1:O:

: -- :

_ : - : _

: - :

:( -- :

: - :

the valencies of the separate atoms, and of the ions, their algebraic
sum, being:

CHEMICAL THEORY

Mn = +7 Cl = +7

= 8

= -4 PO 4 =-3 S0 4 = -2

= 8

4 = -:2 MnO 4 = -1 C1O 4 = -1

It is noteworthy that the covalent bond, i.e. the duplet, is the
same whether the uniting element is univalent like H or bivalent
like O. Consequently the practice of representing O as united
with other elements by means of double bonds, unless these are
double covalent bonds, disappears. *

The way in which the ions represented above are built up may,
however, be elucidated a little further. Consider the S0 4 ion.
The sulphur atom, when neutral, had a valency of +6 with six:
electrons in its sheath, and needed two more to complete its octet;
the four oxygen atoms, when neutral, similarly possessed six
electrons in their sheaths, so that each also required two more
electrons to complete its octet, eight in all being required, of which
the sulphur atom was ready to provide six for covalent union.
Two more electrons were therefore necessary, and these were pro-
vided by two neutral hydrogen atoms, which so became ions.
Thus H 2 SO 4 consists of two hydrogen ions each with one + charge,
because of the loss of an electron, and the S0 4 ion having two
negative charges owing to the gain of two electrons.

Actually, of course, H 2 SO 4 is produced by the union of H 2 O
and S0 3 , both of which are neutral molecules because their com-
ponent atoms mutually satisfy each other in covalent union. In
order that the sulphur octet in SO 3 may be completed, however,
one oxygen atom is represented as united with the sulphur atom

by a double covalent bond, thus: O=S<Vx, so that the reaction

H 2 O + S0 3 -> H a S0 4
becomes: ~

H:O:H + 8:O:
6:

:O:S:O

THE MODERN VIEW OF THE MOLECULE

127

It is thus plainly seen that the two extra electrons provided by
the hydrogen are necessary because, owing to the opening out
of the double bond, an oxygen atom fully furnished with an octet
of electrons must be available to convert the molecule S0 3 into
the ion SO 4 .

It has been usual to show the constitution of sulphuric acid by
its derivation from sulphuryl chloride thus:

Cl HOH OH HCi

I
. O- S () +

OH HCI;

O-IS-0 +

HOH

this reaction now becomes:
:C1: 11:6:11

:O:S:O:

H:0:H

::S::

II 4

Thus the connection between the old and the new way in chemistry
is perceived. In this new way, or something like it, chemical com-
pounds will be formulated in the textbooks of the future.

The methods of representing the constitutions of inorganic salts
according to the older ideas of valency sometimes led to difficulty
on account of the isomorphism of compounds not related chemically.
Thus sodium nitrate, NaN0 3 , and calcspar, CaCO 3 , are isomorphous;
yet they are chemically unrelated, and were given constitutional
formulae to accord with their chemical properties, thus:

Na O

in which nitrogen was shown to be quinquevalent and carbon quadri-
valetit.

Now in the modern method of formulation the valencies of the
individual atoms in a compound radicle disappear when the octets
of the constituent atoms are completed. Thus the nitrate and
carbonate ions are similar in constitution though they differ in the
electric charges they carry, the two salts being formulated thus:

0::N:6:

and

t
'<*'

O::C::

It is thus an argument in favour*of the electronic theory of valency

US CHEMICAL THEORY

these formulae accord with the fact of isomorphism, which is
obscured by the older formulas.

Moreover, it now becomes clear that Mitscherlich, who enunci-
ated the law of isomorphism, was right when he stated that
isomorphism depended primarily on the number and mode of
arrangement of the atoms in the molecule of a compound rather
than upon the chemical nature of those atoms.

This chapter on the modern view of the molecule would bo
incomplete without reference to the question of the existence of
the molecule in the solid state, although this subject has already
been dealt with briefly under the subject of molecular association
(p. 46).

Molecular formulae have long been used in expressing the
reactions of solids without consideration whether they stand for
realities. There is no harm in their use, provided it is understood
that they represent, on an atomic basis, only the quantitative
relations of reacting substances. Formula) such as NaCl and
CaCCX, &c., suggest no more in their common use than the atomic
relations within the compounds they represent; and it is immaterial
for general purposes whether the true molecular formulae of these
compounds should be simple, or, say, (NaCl) m and (CaC0 3 ) n . Never-
theless it is desirable to form a mental picture of a solid, and if
possible obtain a true conception of its molecular state. When
a chemical compound is truly solid it is crystalline, for the amor-
phous state is really the state of super-cooled liquid. The arrange-
ment of the atoms in a crystal is revealed by X-ray spectrography,
according to the researches of Sir W. H. and W. L. Bragg 1 ; and
the method can be applied not only to obvious crystals, but also to
powders, 2 such as precipitated calcium carbonate, which are thus
shown to be essentially crystalline.

In connection with the arrangement of the atoms within the
crystal, the idea of a space-lattice has been introduced. Consider a
piece of wooden trellis-work, which can be opened to show a pattern
of diamond-shaped spaces. Such a pattern, indefinite in extent, but
with units which are similar parallelograms, is a lattice, i.e. a plane
lattice, a lattice in two dimensions; and all that is essential to form
it is two sets of parallel lines, the lines in each set being equi-
distant Now extend the idea to three dimensions, and let there

Crystal Structure (G. Bell & Sons).
2 Hull, /. A our. Chtm. Soc. (1910), 41, 1168.

THE MODERN VIEW" OF THE MOLECULE"

129

be three sets of parallel planes which intersect. Thus a series of
identical units or cells is produced, each cell being a paralleloprped
This is a space-lattice. Moreover, the pattern is preserved, whether
in two or three dimensions, if the lines or pianos are obliterated,
provided the points of intersection of these lines or planes are
preserved. Such a pattern in three dimensions is a picture of the
disposition of the atoms in. a crystal; for the lines, of course, are
only imaginary, and the atoms may be regarded as points in space.
Further, a lattice unit, whether in two or three dimensions, is
a single parallelogram or parallelepiped; similarly there is a crystal
unit or crystal cell, which is the smallest unit in which the essential
properties of the crystal, as regards the space disposition of the
atoms of the substance, are expressed without repetition.

Tin 1 ligure below depicts a crystal unit of sodium chloride as

revealed by X-ray speetrography, white
spheres representing sodium atoms, arid
black spheres chlorine atoms. The space-
lattice, of which this is the smallest es-
sential part, is called the face-centred
cube lattice, because an atom in this
ease a chlorine atom is at the centre
of each face of the cube. Jt is imma-
terial whether a sodium or a chlorine
atom forms the face centre, for by
bisecting the cube parallel to a face
and adding to one of the halves another half cube a sodium face-
centred cube would be formed.

It is to be observed, however, that each sodium atom in such
a structure is surrounded by six equidistant chlorine atoms, as is
seen to be the case with the central atom in the figure; and
similarly, that each chlorine atom is surrounded by six equidistant
sodium atoms. The question may therefore be asked: what has
become of the molecule of sodium chloride? and the answer is that
no such molecule exists in the solid salt. Such an answer, more-
over, is quite in accordance with the electronic theory of valency
as applied to sodium chloride. It is ions, however, and not neutral
atoms of sodium and chlorine, which are packed together in solid
salt; and when the salt disintegrates in water these ions wander
freely in the solvent without existing as NaCl molecules.

It is therefore quite impossible to write a molecular formula for

(D60) 10*

130 CHEMICAL THEORY

solid sodium chloride, and therefore the simple formula NaCl for
the *salt serves every useful purpose.

The electrostatic attraction which binds the ions of sodium and
chlorine together in sodium chloride is the cause why this com-
pound is a solid at ordinary temperature and not a gas.

There are no molecular boundaries, and all the ions in a mass
of the salt are fastened together in one bundle by a pervasive
force or field of electric attraction, which hitherto has been called
cohesion. When, however, such a mass is fused, and so strongly
heated as to be converted into vapour, ions of sodium and chlorine
pair oft' and exist as NaCl molecules. It could hardly be otherwise.
Oppositely charged ions of sodium and chlorine could remain
separate in the state of vapour only if they possessed such high
velocities on account of elevated temperature that the attractive
force between them was rendered ineffective. That would ,^he
thermal dissociation, such as occurs even when pairs of identical
atoms have been united by covalency, as with I 2 .

In view of the conception of the crystal unit as the smallest
portion of the solid necessary to represent completely the properties
of the crystal, it may be asked whether this unit is identical with
the chemical molecule. It has been seen that in the case of
crystallized sodium chloride no chemical molecule can be said to
exist; yet molecules of organic compounds, that is, compounds
whose atoms are united by covalent bonds, are believed to be
present even when those compounds are solid. Nevertheless, ac-
cording to Sir William Bragg, 1 there is 110 reason why the crystal
unit should be identical with the chemical molecule; and X-ray
analysis has shown that this unit generally consists of two, three,
or four molecules; e.g. the crystal unit of naphthalene consists of
(C 10 H 8 ),,, and that of a-naphthol, where the symmetry is reduced by
the introduction of an OH group, of (C 10 H 7 OH) 4 . Similarly two
molecules of benzene, C C H 6 , constitute a crystal unit, but four of
benzoic acid, C 6 H 5 COOH.

Consider the two solid and related elements carbon and silicon.
When carbon burns it forms an oxide which is a gas, but when
silicon burns the oxide formed is a solid. What is the reason for
this difference? The answer is that carbon dioxide consists of
molecules of three atoms compactly joined together by covalency,
each forming a self-satisfied system with very little outside in-

* "The Signiticance of Crystal Structure", Vraiu. Ckem. Soc., 1922, Itl, 2766.

THE MODERN VIEW OF THE MOLECULE 131

fluence, so that only at low temperature and high pressure do the
separate molecules unite to produce liquid and then solid carbon
dioxide; but that it is otherwise with silica. Silica has long been
recognized to. consist of polymerized molecules, and it is now
known that as quartz its crystal unit is (Si(X) 3 . It may well be
doubted, however, whether Si 3 O 6 molecules, existing side by side
with little attractive force between them, would be competent to
produce an inert and non-volatile solid like quartz. Bather it
would seem that a cohesive force of great strength exists between
the Si 3 6 molecules, to overcome which very high temperature
is necessary; or otherwise that the mode of union pertaining
between silicon and oxygen in silica is analogous to that between
the elements in TiO 2 > Sn0 2 , and other related oxides. Indeed,
as compared with other dioxides it is carbon dioxide which has
exceptional properties, not silica; and, as was said earlier in this
chapter, it is the inner structure of the carbon atom that fits it
for covalent union, which then confers volatility on the compounds
carbon forms with other elements.

Nevertheless, it is remarkable that silicon tetrachloride (B.P. 57)
is so volatile compared with silica. This may be because the
silicon atom in the tetrachloride is unable to exercise much attrac-
tive influence on its similar neighbours'Owing to the cordon of four
chlorine atoms with which it is surrounded; whereas the oxygen
atoms of silica are not numerous enough to exercise this influence.
It may be noted, moreover, that S0 3 , C1 2 O 7 , and even OsO 4 are
volatile.

Compounds such as PC1 5 and SF 6 call for comment. If it is
believed that the halogen atoms are attached to the other atom
in these molecules by covalency, then ten and twelve electrons
respectively are concerned in the process; but it seems unlikely
that the octet of electrons is exceeded in the sheaths of atoms of
such low atomic number as phosphorus and sulphur. The alterna-
tive is to regard the halogen atoms as united by electrovalency, in
which case the five and six electrons originally present in the
sheaths of phosphorus and sulphur atoms respectively will have
left these atoms to become attached severally to the halogen atoms.
In this case the stability and volatility of SF 6 , as of other poly-
fluorides, is to be attributed to the simplicity of internal structure
of the fluorine atoms which allows them to come very close to a
sulphur atom; whilst the dissociation of PCl^ vapour into PC1 3 and

132 CHEMICAL THEORY

C1 2 , with subsequent oxidation of PC1 3 , may be represented as
takiiig place in the following way:

:C1: :(fl: :6'l: :O .:C1:

:1 1: ' + :P:Cl:,and :P:(5l: -" :6:P:C1:;

which shows two chlorine atoms returning an electron each to the
phosphorus atom, and then forming a chlorine molecule by co-
valent union; whilst the other three chlorine atoms also enter into
union with the phosphorus atom, so that PCI ^ becomes a covale^it
compound, which may be subsequently oxidized to phosphoryl
chloride in the manner shown.

These illustrations suffice to show the trend of the modern
theory of the molecule; and they leave no doubt that the kinds of
formula) which have embellished our textbooks for a generation
must soon give place to formula) of another kind. The present
position of the theory, however, must be regarded as a phase, for
many of the ideas are speculative, but so long as chemical vscience
remains alive and active, so long must its theories continue to
undergo modification.

CHAPTER VII
THK COLLOIDAL STATE

When a finely -divided solid is mixed with water or other
solvent, either it may dissolve completely or some or all of it
may remain undissolved. These two conditions are easily distin-
guished. If the solid dissolves completely the resulting liquid,
whether coloured or not, is clear or transparent. If the solid does
not dissolve completely, the liquid when shaken will appear turbid
or opaque, and if the mixture is allowed to stand undisturbed
the solid in suspension will in time settle, leaving the supernatant
liquid clear.

Tho distinction between a substance in solution and one in
suspension appears fundamental; for a liquid containing suspended
matter may be filtered to be made clear, but it is necdloss to filter a
solution. Yet suspended solids differ in the fineness of their sub-
division, and in the ease with which they are removed by subsidence
or filtration. Sand, for example, will settle more quickly than pre-
cipitated chalk, and both of these can be filtered more easily than
precipitated barium sulphate or calcium oxalate. And the student
is familiar with substances even more diflioult to filter than the last
named. The sulphur which separates when hydrogen sulphide gas
is passed through an oxidizing solution, or when acid is added to a
polysulphide, cannot be removed completely by means of ordinary
filter-paper, the pores of which are evidently too large to retain the
minute particles of which the precipitate consists.

Indeed, although the solution may become practically trans-
parent, yet the presence of suspended sulphur is revealed by a slight
opalescence. Precipitated silver chloride presents a similar pheno-
menon; whilst copper sulphide, imperfectly precipitated from cold
solution, may yield a brown filtrate, which, although transparent,
contains the sulphide in a very fipe state of subdivision. It appears,
therefore, that a liquid iay contain suspended matter so finely

133

134 CHEMICAL THEOKY

divicbd as not to produce opacity or to be removable by subsidence
or ordinary filtration. It is pertinent to ask, therefore, how such a
suspension differs from a true solution.

Meanwhile the subject may be approached from a different point
of view.

When an aqueous solution of sodium silicate, or soluble glass, is
acidified with dilute hydrochloric acid, silicic acid is liberated, and,
if the solution is in a concentrated state, will separate from it ; n
the form of " gelatinous silica". If, however, the solution is suffi-
ciently dilute, there is no precipitate, the liquid remaining clear. It
might be supposed that the difference between these two conditions
depended simply on the amount of water present, there being enough
water to hold the silicic acid in solution in the one case, but not in
the other. If this were so, the gelatinous silica separated in the
former case would be in equilibrium with a saturated solution of tne
same substance. This, however, is not so; the phenomenon here
exhibited is indeed quite different from an ordinary case of precipi-
tation, as will appear in the sequel.

This and kindred phenomena were first investigated by Graham
in 1849 in connection with experiments on liquid diffusion. Graham
found that the rates of diffusion into pure water of different sub-
stances in aqueous solution were various, and that simple salts and
acids passed rapidly through an animal membrane or parchment
paper, whilst complex substances like gelatine or glue in aqueous
solution did not penetrate these membranes. These latter substances
Graham called colloids (/coXXa, glue), whilst acids and salts, being
crystallizable or related to crystallizable substances, he called crystal-
loids. So crystalloids and colloids can be separated from one
another by aqueous diffusion through a parchment or other suitable
membrane, fixed on a frame like a drum and floating on water,
crystalloids passing through the membrane into the external water,
whilst colloids remain behind in the drum. The process is called
dialysis, because it involves separation of one substance from
another by passing it through a membrane, through which crystal-
loids in solution, but not colloids, can pass.

A diluted acidified solution of sodium silicate may be submitted
to dialysis. The sodium chloride formed in the reaction,

Na,SiO 3
and excess of hydrochloric acid pass through the membrane of the

THE COLLOIDAL STATE 135

dialyser, leaving the silicic acid behind in pure aqueous solution.
Such a solution may be concentrated by evaporation to a strength
of about 14 per cent. After this it changes to a jelly, similar to
that obtained by acidifying a concentrated solution of soluble glass.
Not only does concentration cause dialysed silicic acid to coagulate,
but a trace of hydrochloric acid, or some simple salt, acting
fcatalytically, produces the same effect. It thus appears that there
awe two forms of aqueous colloidal silicic acid; the clear form, which
seems to be a solution, and the gelatinous form, which evidently
is not. These two forms, in which colloids in general may occur,
are called respectively the kydrosol and hydroyel, or simply sol
and gel.

Besides silicic acid and organic substances such as gums and
resins, glue and gelatine, various inorganic substances occur or can
be obtained in the colloidal state.

Graham prepared the sols of ferric, chromic, and aluminium
hydroxides by dialysis ; the sols of arsenious and antimonious
sulphides may be prepared by boiling arsenious oxide and tartar
emetic respectively with water, and adding hydrogen sulphide to
the solutions: the liquids become yellow and red respectively
because of the formation of the sulphides in the sol condition. A
drop of hydrochloric acid added to either solution precipitates the
yellow arsenious or orange-red antimonious sulphide, the sulphides
thus assuming the gel condition.

It will now be understood that the above -mentioned brown
liquid obtained in precipitating copper sulphide contained the sol of
this sulphide, whilst the opalescent liquid containing sulphur held
this element also in the colloidal state.

The sols of certain metals are interesting, and often display
remarkable colours. Thus gold and silver may be separated from
their salts by hydrazine, formaldehyde, &c. Faraday produced
blue, violet, and rose-coloured liquids by reducing gold chloride by
means of an ethereal solution of phosphorus floating on the surface
of the solution; and Bredig obtained sols of gold, silver, platinum,
&c., by an electric discharge through water between poles of the
metal.

It is easy to understand that colloids have been regarded by
chemists with much interest from the time of their discovery to the
present day. They are of practical importance because they embrace
many common non-cryst&llizable organic substances such as gum,

136 CHEMICAL THEORY

resin,, glue, starch paste, egg-albumin, casein, and gelatine; but
they are particularly interesting from the physico-chemical stand-
point because they present a fresh phase of the great subject of
molecular physics. Indeed, a transparent dialysed liquid, consisting
of silicic acid and water for example, which on the addition of a
suitable catalyst becomes a jelly that can be inverted without flow-
ing, presents to the scientific mind a subject for investigation full of
an interest that can scarcely be surpassed. i

What is this transparent liquid? Is it a solution like a mixture
of sodium chloride and water? If it is, why does the silicic acid
remain in the dialyser whilst the sodium chloride passes through it?
It has already been suggested that the process of dialysis is a kind
of filtration, i.e. that the silicic acid molecules are too large to pass
through the pores of the parchment paper. But filtration is applied
to something in suspension. Is the silicic acid in this apparently
clear liquid really in suspension?

An answer to this question as regards colloids in general has
been gained by the use of the ultra - microscope invented by
Siedentopf and Zsigmondy in 1903. The principle underlying
the use of this instrument is that illustrated by the vision of tho
"mote in the sunbeam". It is well known that the moving dust
of the air, which cannot ordinarily be seen, is made visible in a
beam of sunlight entering a darkened room through a chink
in a shutter. At the same time the track of the beam itself is
clearly outlined; but if the air is free from dust the sunbeam dis-
appears.

This effect, studied by Professor Tyndall, is applicable also to
liquids, and will reveal the presence of suspended particles within
them in the same way that it shows aerial dust.

Moreover, the lesson to be learned is that very intense and
localized light, bv increasing the intensity of reflection, greatly en-
hances our powers of vision, And if the dust of the air, otherwise
quite invisible, thus becomes apparent to the naked eye, particles
too small for microscopic vision under ordinary illumination may be
seen under illumination analogous to that of the sunbeam. This is
the principle of the ultra-microscope, in which a beam of sunlight,
or from the electric arc, passes through a slit horizontally, or is
focused into a liquid which is examined by the microscope verti-
cally. Any light which enters the .microscope must then have been
reflected from the surface of particles ^suspended in the liquid.

TfflE COLLOIDAL STATE 137

Particles having a diameter only one -hundredth that of f the
smallest particles visible under ordinary illumination can then bo
seen as spots of light like planets in the darkness. And so colloidal
liquids have * been seen to be suspensions, and the size of the
suspended particles has been evstimated by counting the number of
them in a volume of the liquid containing a known weight of
riiaterial. Thus the particles of platinum, gold, and silver seen
UP colloidal suspensions of these metals have been discovered
to"* have diameters ranging from 2xlO~ 4 to 6xlO~ 4 mm. The
smallest particles detectable by this method, when illuminated
by bright sunlight, have a diameter of 4xlO~ 6 nun., whilst the
individual molecules of substances like chloroform and alcohol
have diameters of 0-4 X 10" 6 to O'SxlO" 6 mm., and of hydrogen
0-lxlO- 6 mm.

Thus the particles of colloidal metals in aqueous suspension
have diameters about a thousand times as great as those of mole-
cules which form mixtures with water regarded as true solutions;
whilst the smallest particles that can be rendered visible have
only about ten times the diameter of gaseous and other simple
molecules.

These metallic suspended particles are not, however, molecules,
but rather minute fragments of solid metal; since molecules of
solid metals, consisting of definite aggregates of atoms, can scarcely
be said to exist. Thus, they differ from silicic acid and complex
organic substances which are known to consist of very large mole-
cules. The molecule of egg-albumin is estimated to have a mole-
cular weight of 17,000, and the molecules of the enzymes emulsin
and invertin have molecular weights of about 45,000 and 54,000
respectively, with molecular diameters of about 6 x 10~ 6 mm.
Such molecules can be seen by the ultra-microscope, but there
seems no hope that the simpler inorganic molecules will ever be
revealed to the eye of man, although they lie but a little way
below his range of vision aided by this powerful instrument.

Finally, although colloids cannot ordinarily be filtered, the fact
that parchment paper retains them suggests that special methods
of filtration might effect their separation as parchment paper does.
Special filters have in fact been prepared by treating ordinary
filter- paper with collodion or gelatine, which have pores varying
in diameter between 930 X 1>~ 6 mm. and 21 X 10~ 6 mm. By
means of these filters various colloids have been differentiated and

138 CHEMICAL THEORY

classified according to the sizes of their particles, with the following

results:

Suspensions of non-colloids.

Colloidal platinum.
Colloidal ferric hydroxide.
Colloidal arsenious sulphide
Colloidal gold.
1 per cent gelatin.
Colloidal silicic acid.
Litmus.
Dextrin.

Solutions of crystalloids.

Thus it is seen that colloids afford a gradation between what
are commonly regarded as suspensions and solutions; and so it is
evident that the idea of a colloid as a glue-like substance has been
extended so as to include all kinds of matter in a state between
that of molecules and that of gross particles which arc ordinarily
perceptible. Therefore such natural products as clay and lime are
regarded as colloids, and many precipitates formed in analysis are
believed also to pass through the colloidal state before separating
as particles that can be filtered. Indeed it appears that there are
two ways in which a precipitate may be formed and pass into a
state of equilibrium with the liquid phase with which it is in
contact. These are either by direct crystallization, as in the case
of magnesium ammonium phosphate, as well as calcium carbonate
at elevated temperature; or the formation and coagulation of a
colloid, as in the case of many sulphide and hydroxide precipitates.
Silver chloride and calcium oxalate are also examples of this kind
of precipitate, and barium sulphate is probably a limiting case
between the two kinds.

The coagulation of colloids is a subject which has received much
attention. It has already been pointed out that a colloidal sub-
stance such as silicic acid can exist in two conditions, those of "sol"
and " gel " ; and that an electrolyte has the power of coagulating a
sol, thus converting it into a gel. It is now recognized that the
particles of a sol are electrically charged; in some cases such as
metallic hydroxides positively, in others such as metallic sulphides,
silver chloride, and silicic acid, negatively. Consequently a col-
loidal sol is capable of a kind of electrolysis, the colloidal matter
travelling either to the cathode or the fcnode when a current is

THE COLLOIDAL STATE , 139

passed through the liquid. This motion of colloidal particles in an
electric field is called cataphoresis.

The coagulation of a sol by means of an electrolyte, as, for
instance, that* of arsenious sulphide sol by means of hydrochloric
acid, with formation of a flocculent precipitate, is probably an
electrical phenomenon, in which the particles lose their charges and
Consequently coalesce. The reverse process may, however, take place
ii* too concentrated solution. For example, precipitated basic ferric
"ST&tate is prone to pass into the sol condition, chiefly as hydroxide
owing to hydrolysis, and so form a slimy mass impossible to filter.
The formation of a sol from a gel, which is the opposite of coagu-
lation, is called peptization.

Connected with peptization is the subject of protective colloids.
Certain organic colloids have the power of hindering the coagula-
tion of inorganic colloids with which they are associated. The
persistence of the sol condition is the result of the mutual repul-
sions of similarly charged particles; and when from any cause
these repulsions cease to operate coalescence of particles may take
place with coagulation of the colloid. If, however, the individual
particles of the sol are protected by an envelope of any material
which hinders their electric discharge, the sol condition will persist
even in presence of an electrolyte which would otherwise cause
coagulation. Thus, as was found by Faraday, gold sol, formed by
the reduction of auric chloride solution by phosphorus, is stabilized
by gelatin, and the precipitation of silver chloride is prevented
by the same colloid, the less stable silver chloride sol being pro-
tected against the electrolyte present by the more stable gelatin sol.

So far the study of colloids in this chapter has been restricted
to phenomena of the liquid state. In the original and narrower
sense in which the term was employed colloids necessarily contained
a liquid phase, but just as the term " solution ", which is generally
applied to liquid mixtures, may be extended to include gaseous
and solid mixtures, so the term colloid is now extended to include
certain phenomena of the gaseous and solid states.

The distinctive property, however, of the colloidal state, which
differentiates it from the state of solution, is heterogeneity; that is
to say, the colloid is dispersed through the medium, there is a
disperse phase and a dispersion medium. So far the disperse
phase has been regarded as sqlid, and the dispersion medium as
liquid. When, however,, the conception is extended to gaseous and

140

CHEMICAL THEORY

solid systems, a comprehensive scheme results in which a variety
of interesting phenomena are included. Thus Wo. Ostwald has
proposed the following classification of colloids:

Disperse
Phase.

Dispersion
Medium.

Examples.

Solid

Carbon particles in iron. Gold in ruby glass.

Solid

Liquid

Colloidal solutions of metals,

gelatin, starch, &c.

Solid

Gas

Smoke. Fine dust. Fumes.

Liquid

Solid

Certain minerals.

Liquid

Emulsions.

Liquid

Uas

Fog. Mist. Clouds.

(las

Solid

Soliditied froths, e.g. pumice.

(Jus

Liquid

Froths and foams.

The only system excluded from this scheme is that of a gas
dispersed in a gas. This system, however, is homogeneous, like
a true solution, there being no distinction as regards molecular
dimensions between its different components; therefore it is not
colloidal.

It is evident from a consideration of the table that a very large
number of phenomena encountered in nature and employed in the
arts come under the category of colloids. Indeed, as Ostwald has
said: " It is simply a fact that colloids constitute the most universal
and the commonest of all things we know. We need only to look
at the sky, at the earth, or at ourselves to discover colloids or sub-
stances closely allied to them. . . . We have only recently come
to learn that every structure assumes special properties and a
special behaviour when its particles are so small that they can no
longer be recognized microscopically, while they are still too large
to be called molecules. Only now has the true significance of Jhis
region of the colloid dimensions The World of Neglected Dimen-
sions become manifest to us."

SUMMARY

DIALYSIS is the separation of substances in solution by the use
of a membrane, through which crystalloids in solution will pass
but not colloids.

COLLOIDS, e.g. silicic acid, can exist in two states, the hydroaol
(or sol) state, and the hydrogel (or gel) state.

Sols are converted into gels by catalysv?.

THp COLLOIDAL STATE 141

By means of the ultra-microscope colloidal liquids have been
seen to be suspensions. >

Colloids have been separated by the use of special filters, and
a gradation Jias been established between colloids in suspension
and crystalloids in solution.

Colloids consist of electrically charged particles which can be
separated by an electric current, the process being called c.ata-
phoresis.

A sol passes into a gel by coagulation, a gel into a sol by
peptization.

More stable colloids can act as protective colloids to less stable
colloids, thus preventing their coagulation by electrolytes.

A colloidal system includes a disperse phase and a dispersion
medium. The system may bo gaseous, liquid, or solid.

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 "Atoms And Molecules"

See other formats