]\gt ; 147 : Ionic Size in to Molecular Physics , together Law Relating to the of of Solid , Liquid ; and Ionic Molecules .
By W. R. BOUSFIELD , M.A. , K.C. ( Communicated by Sir J. Larmor , Sec. .
Received October 18 , \mdash ; Read Vovember21 , 1912 .
) CONTENTS .
PAGE 1 .
Introduction 147 2 .
Components of Heats of Formation of Solid , Liquid , and Ionic Molecules 149 3 .
ljerimental Work 156 4 .
Deduction of Ionic Yolumes and Empirical Volume lations 158 5 .
Comparative Values of Ionic Volumes 160 6 .
AnalyticaI Treatment of the Empilical Equations 161 7 .
The Empirical Relation between )oint Depression and Ionic Volume , and Deduction of Constants 165 8 .
Application to Ionic Volumes at Infinite Dilution 1 .
Introduction.\mdash ; The greater part of this communication is concerned with obtaining the volumes and densities of the ions and ionic nuclei of three salts from conductivity , viscosity , and -point data , a process which is of limited interest and is therefore reserved for later sections of this communication .
The deductions which are drawn from the values so obtained , in conjunction with other data already available , lead to certain conclusions as to the heats of formation both in the solid and ionic state which are of much wider interest and are dealt with in the next section .
In two former communications* there was developed a theory with reference to the ascertainment of the relative sizes of ions which was based upon a consideration of various physical data relating to KC1 and .
In the present paper the necessary data are given for extending the results to .
In the former papers an empirical linear relationship was found between the solution volumes calculated from density measurements and the ionic volumes deduced from conductivity and viscosity measurements , which is now extended to .
In the present paper a new empirical relation will be found between the ionic volumes and the reciprocals of the effective * I shall have necessity to refer from time to time to these papers , and will do so under the short names " " Ionic Sizes I\ldquo ; and ' ' Ionic Sizes II An abstract of the former is to be found iu 'Roy .
Soc. Proc 1905 , p. 563 , but the references herein given are to the full paper , which was published , with some additions , in the 'Zeit .
fur Phys. Chem 1905 , vol. 53 , p. 257 .
The second paper was published in ' Phil. Trans 1906 , vol. 206 , p. 101 .
For the understanding of the later sections of the present communication it is necessary that the reader should have the second paper before him .
VOL. 148 Mr. W. R. Bousfield .
ombination interesting information aoint dhree Sations aound t relative volumes and densities of the ions and of the ionic nuclei . !
The three salts above mentioned are compounded of some of the most : compressible of the elements , and are therefore admirably suited for a consideration of volume changes in relation to heats of formation .
Now that we are coming to regard an atom as being itself composed of hundreds of more minute bodies ( which we will designate as corpuscles , in order to adopt a neutral term ) there is no a priori improbability in the theory of the compressibility of atoms .
We shall therefore express our results in terms of this hypothesis , to which they appear to give considerable support .
In ordel to correlate changes of atomic volume with heaCs of formation , it was necessary to consider a wider erroup of compounds of the same class as the three salts above mentioned .
For this purpose certain available data exist within a sufficient area to enable a useful generalisation to be made .
These data lead us to a new law relating to the heats of formation of the solid and liquid compounds formed from the elemerits and radicles of the ubstances considered , which may be enunciated as follows:\mdash ; The heat of formatio due to the combination in the liquid ' solid state of any of electro-positive elements with any of the electro-negative elements or radicles is approximately equal to the sum of certain calorific constants of the two elements or radicles , together with , where is the change of atomic volume produced by the combination .
This leads us to the conception of the internal energy of an atom as the sum of the kinetic energy of the corpuscles and of their potential energy under their mutually attractive forces .
On this view , the heat component , , represents the change in the internal energy of the atoms due to their volume changes on combination .
The heats of ionisation of , and are apparently irregular and have not hitherto been regarded as periodic functions of the atomic volumes .
The application of the factor to the contraction of volume of the nuclei will be found to show the heats of ionisation as periodic functions .
It appears probable that we can briug he heat of ionisation into complete subordination to the same general law as that above enunciated for the solid and liquid compounds of the same class .
* In a former paper ( Ionic Sizes II , p. 142 ) I used this expression to indicate the so-called molecular depression divided by , where is the ionisation coefficient .
The effective molecular depression of the freezing point may be shortly indicated by the .
letters E.M.D. , so that the E.M.D. is .
* .
1912 .
] Ionic Size in Relation to Molecular Physics .
149 2 .
Components of Eeats of Formation of'Solid , Liquid , and Ionic Molecules.\mdash ; Sufficiently complete data are available for most of the compounds of Na , and Li , with the radicles Cl , Br , I , , to enable us to arrive at a generalisation within this limited area as to the relation between heats of formation and atomic volumes .
The values of the atomic volumes of the elements in the uncombined state are indicated by the letter , and are set out at the foot of the table .
The values are those given by Mendele'eff , with the exception of those for , Na , and Li , for which the values are those given by Richards based on his own density determinations .
* The atomic volumes of the positive elements of each compound are set out in the column headed and those of the negative radicles in the column headed .
The values given for the compound radicles , such as , are merely the sums of the atomic volumes of their components .
The column headed gives the sums of the atomio volumes of the elements in each compound .
Under are given densities of the solid and liquid compounds , and under their atomic weights .
The quotient .
is therefore the actual ular volume of each compound .
The difference is the contraction of volume for each molecule which takes place on combination , and is set out in the column .
The actual heats of formation in -calories from Thomsen 's data are given in the column headed H.F. It will now be seen from the table that each heat of formation is the sum of three quantities , viz. , a part which is proportional to the total contraction , and is expressed by a remainder which is the sum of two components and , which have approximately the same value for each adicls in every combination .
To split the additive components into parts and we require a definite starting-point , which is wanting .
It is obvious , however , from the heats of formation of HC1 , , and that the calorific constant for the hydrogen atom is comparatively small .
If we place it at 1 then the calorific constants for the other radicles are those given under ' Values of \ldquo ; at the foot of the table .
If the value of for hydrogen were 3 instead of 1 we have only to diminish the heat components for each of the negative radicles by 2 , and increase the heat components for the positive radicles by the same amount .
It must , therefore , be understood that the value of 1 for the heat component for hydrogen is arbitrary , but it makes no difference to the additive scheme .
The is that all the heats of formation may be approximately expressed as separate components are set out in three columns of table and the 'Journ .
Amer .
Chem. Soc 1909 , vol. , 156 W. R. Bousfield .
[ Oet .
18 , 1912 .
] Ionic Size in Relation to Molecular Physics .
151 sum in the column headed 2 .
The column headed " " Diff shows the differences between the observed and calculated heats of formation .
The principle which is involved in the relationship which thus comes to light is probably capable of wider application , but data for extending the generalisation are very difficult to obtain .
At the suggestion of Prof. Sir J. Larmor data have been collected for another group .
They are set out in Table II .
The errors between observed and calculated values are about * twice as great as in the first group , but the heats of formation of the bivalent .
salts are also about twice as great .
Within the area of both these groups we can say that the heat of formation due to the combination of one of the electro-positive atoms with one of the electro-negative atoms or radicles is the sum of three components , two of which are constants to the several combining atoms or radicles , and the other of which is approximately , where is the contraction which takes place on combination .
The last component possesses reat interest in connection with the theory of the compressibility of the atom .
It appears probable that the component represents the change of internal energy due to the contraction of the atoms S on combination .
Let us conceive the corpuscles of a free or uncombined atom as vibrating in a certain maximum space with a maximum average amplitude at a given temperature , the volume of the atom being determined by an equilibrium between attractive forces , which tend to approximate the corpuscles and to reduce the volume , and the energy of motion of the corpuscles , which tends to increase the volume .
If we suppose such an atom to contract or to be compressed isothermally , the internal kinetic energy would be diminished , owing to the reduction of the amplitude of the corpuscular vibrations with reduction of volume , and the internal potential energy would be diminished , owing to the approximation of mutually attractive corpuscles .
This reduction of internal energy appears to be expressed by , or at all events included in , the component , which would thus represent the heat so developed .
This , in the case of , amounts to more than a quarter of the total heat of formation , and in the case of the nitrates and iodates to a still larger fraction .
We have spoken of a reduction of internal energy due to compression , but it must be observed that in the cases of HC1 and , to which the law also extends , there is an increase of atomic volume on combination , and a resulting increase of internal energy .
We may now apply the result obtained as regards to consideration of the heats of formation of ions .
The ession of the atoms in the ionic state is even greater than in the solid state , as we may see the values of Mr. W. R. Bousfield .
[ Oct. 18 , 1912 .
] Ionic Size in Relcition to Physics .
set out below , which deduced from the values of obtained in Section 7 .
: This is probably due to the superiority of the attractive foroes which come toms themselves oteinto petween ttoms a ter , over the mutual forces between The heats of formation of the solid salts , which are in the order of their $ atomic weights and volumes , being 105 .
may be regarded as periodic functions .
On the other hand , the heats of ionisation appear to be irregular , It seems clear that the heat component due to atomic compression , which probably expresses the reduction of the internal energy of the atom due to the approximation of its corpuscles , should equally be a component of the heat of ionisation .
The application of this consideration at once the heats of ionisation .
If , 13 , 21 , respectively be the numbers of water molecules combined with the solute see the last section of this paper ) , and if we attribute to each molecule of combined water a calorific constant , we may express the heats of ionisation of the three salts by one formula .
Heat of ionisation It will be noticed that in this scheme the idiosyncrasies of the ions Na , Li have disappeared except so far as they affect contraction and water combination , no place having been found for their individual calorific constants and , as to whose values it may be observed that they are so qmnu a ?
1 uaqA AttiI xapun I \mdash ; : oIlOfl buaqos ?
uauod -moi ) S1 } xatBI SIItIA 4 uo 8 . ? . . . . . . . . . . .
quIOO 30 eJOJj -uoo SI( UBO eq lsnm A )'JJOO $ ?
a{I ) $uuodmoo ?
addle $I AtOU SUOO onoJQ q a $deoxBJJUOO IIOS SI ?
SUOO 9 gsnog The last column shows the value of which results from these figures on the assumption that the whole of the water is in combination with the acid .
Thus up to a value for is yielded which is very nearly equal to the value deduced from the ions of the three salts above mentioned .
For 'Principles of Chemistry , ' Ed. 1905 , vol. 2 , p. 271 .
' ' Ionic Sizes I p. 304 .
Wiss .
Abhand .
der Phys. Tech .
Reichsanstalt , ' 1905 , vol. 4 , p. 246 .
uo pestq Jadtd pam$ ?
tqo dapuI Oenba S oq sqns pue $ U10J ; }Isuap a aas S1 sJaded : Relations between and KC1 . . . . . . . .
The molecular solution volumes are obtained by multiplication by the equivalent weights , which are ( old atomic weights ) 58 .
These lead to the empirioal volume relations in the form in which they are required for KC1 . . .
These equations are of the form the constants and for each salt being as shown .
from the viscosities it was that , both in the case of KC1 and , we were led to the result that the ctive molecular freezing-point depression , over a certain range of concentration , came out as a constant when reckoned upon the free water in } solution , that is , after deducting from the total water the amount of combined water calculated from the ionic volumes .
In the present paper this is taken as starting point for the determination of .
A new and more convenient empirical relation between ionic volume .
and freezing-point depression enables the factors to be worked out upon a systematic basis .
This relation is of the form , where is the effective molecular freezing-point depression , is the number of cules of solute per 1000 .
water , is the ionic volume reckoned in the old units , and is the volume of the ionic nuclei in the same units , and being constants .
By the use of this empirical relation in combination with that between ionic volume and solution volume it is possible to evaluate and to obtain the volumes both of the ions and of the ionic nuclei .
One theoretical defect in this process must not be overlooked .
" " Ionic Sizes II p. 156 .
" " fonic Sizes II p. 139 .
' fouic Sizes II p. 149 .
1912 .
] Ionic Size in Relation to Molecular Physics .
161 This is that we are using ionic volumes based on conductivities measured at C. as applicable to the solutions in the hbourhood of the freezing point .
To make our deductions strictly accurate , the ionic volumes should be based on conductivities also measured in the neighbourhood of the freezing point .
But no sufficiently accurate data of this kind are in existence .
The general character of the deductions of this paper will not be seriously modified by this defect .
It should be remembered that though a different factor is required for each salt to make the ionio volumes derived from the mobilities comparable as between the different salts , yet for any given salt the ionic volumes are strictly comparable through a certain rang of concentration .
Each set of ionic volumes , calculated as in the former papers from conductivity and viscosity data , must be considered as being given in its own arbi rary unit , the factor in each case reducing the arbitrary unit to ordinary units .
It should be noted that though we use the separate radions and of the positive and negative ions to arrive at the joint ionic volume of the solute , our factor , being applied to the joint volumes , cannot be used to obtain the absolute values of the radions .
These are not wanted for our present purpose , but to indicate that and require different factors to reduce them to absolute values it is better to term and the " " mobility radions 6 .
Analytical Treatment of the Equations.\mdash ; Before we are in a position to evaluate the required constants of the new empirical relation it is necessary to consider its physical meaning in combination with that of the other empirical relation It will be conyenient at this stage to collect together the symbols used in A tabulated form for reference .
Symbols Solution volume of 1 grm. of solute .
Volume which 1 .
of solute occupies in solution .
Weight of 1gramme-molecule of solute .
Weight of 1 gramme-molecule of water .
Density of free or uncombined water .
Average density of water which is in combination with the solute .
Volume of 1 gramme-molecule of free water .
Volume of 1 gramme-molecule of combined water .
We found formerly*that the E.M.D. , when reckoned upon the free water of the solution only , came out as a constant .
The E.M.D. is when this is reckoned upon the total water .
We shall still keep express this fraction .
The fact that the E.M.D. when reckoned upon the free water is a constant will then be expressed by , ( 1 ) where is a constant , and is equal to , or whatever number we decide to adopt as best for the limiting value of the E.M.D. The volume of water combined with 1 gramme-molecule of solute is , where is the volume of the ionic nuclei .
If is the factor required to reduce ionic volumes to litres and is the average density of the combined water , the weight ( in grammes ) of water combined with gramme-molecule of solute is 1000 , so that Hence equation ( 1 ) may be expressed as .
( 2 ) ' ' Ionic Sixes II p. 149 .
1912 .
] lonic Size in Relation to Molecular Physics .
Our empirical relation is Hence we have and S and we thus get for the required value of But there is still an unknown quantity , the density of the combined water , involved in this equation and an unknown quantity is required for the deduction of and .
We can obtain these by a concurrent consideration of the other set of empirical relations To interpret this equation we must remember that , the solution volume of a molecule of solute , is the volume which the molecule of solute occupies in solution less the contraction of the water which combines with the solute , that is to say , where is the volume which 1 .
of solute occupies in the solution , is the average contraction which 1 molecule of water undergoes in combination with the solute , and is the number of molecules of water combined with 1 molecule of solute .
It must further be noted that the volume of 1 molecule of the hydratedt solute , that is of 1 molecule of the ions and partly ionised molecules , is the volume which 1 molecule of solute occupies in solution plus the volume of the combined water , that is to say , where is the weight of a molecule of water .
By elimination of between ( 6 ) and ( 7 ) and noting that where is the density of the uncombined water , we get If we assume that at high dilutions , the density of the combined water , .
and , the density of the free water , are constants we may now compare our empirical equation ( 5 ) with equation 8 ) , and equate coetflcients .
But it is likely that there may be a change in the density both of the free watel and VOL. Mr. W. R. Bousfleld .
[ Oct. 1@ , .
the combined water in passing from high to low dilutions , and this limitation must , therefore , be borne in mind .
Equating coefficients we get , ( 9 ) .
( lO ) Now is the volume of the ionic nuclei in the same units as , and , therefore , from ( 9 ) and ( 10 ) we get .
( 11 ) From ( 4 ) and ( 10 ) , by eliminating , we get .
Whence also .
( 12 ) is thus expressed in terms of the constants of the empirical equations .
Also from and ( 10 ) by eliminating we get ' ( 13 ) and from ( 4 ) , ( 9 ) , and 13 ) ( 14 ) is thus expressed in terms of the constants .
By reason of 12 ) the last equation may be written in the interesting form .
( 15 ) Finally , the density of the combined water is given by ( 9 ) , which may be written .
( 16 ) For the simple calculation of from the experimental data we obtain from equations ( 3 ) and ( 12 ) the relation .
( 17 ) It will be noticed that we have treated and as if our conductivity data were taken at the same temperature as the freezing-point data .
This may in some equations produce an error of nearly two parts per thousand .
but the error produced on the value of is inappreciable .
It should also Ionic Size in Relation to Molecutar Physics .
be notieed that by treating and as constants in the region of high dilution we bring out as a constant .
This would introduce a small error at low dilution , as must probably be larger at such dilutions , owing to the reduced pressure on the nuclei when the ions are smaller and the water combination is less .
We have also neglected the difference in the value as between C. and C. 7 .
The Empirieol Bptation between Freezing-Point Depress and Volume , and of Constants.\mdash ; We are now in a position to obtain the empirical relations between freezing-point depression and ionic volume for the three salts under consideration .
Let us first take the identical set of figures set out iu the former paper*for KCL These were there shown to lead to the relation within the limits of experimental error .
* They equally lead to the new relation within the same limits .
The choice of the new relation instead of the old one is determined by the fact that it admits of a precise theoretical interpretation as was shown in the preceding section .
In the following table , the first , second , third , and fourth columns are set out the same figures as were formerly given for .
The values for the molecular depression of the freezing point are those given by Jahn .
The figures given for the calculated values of are obtained by taking Table VIII.\mdash ; Empirical Belation between Freezing-point Depression and Ionic Volume for * Ionic Sizes II p. 143 .
' Zeit .
Phys. Chem 1904 , vol. 60 , p. 136 .
166 Mr. W. R. Bousfield .
ahn 's [ Oct. 18 , If we look at the values of derived from Jahn 's duplicate differences oated vuesof lrder adifferences , aywithin timits oerror Aoncentration , which the , refore sithout doubt tation i is the most important for determining the limiting value of , Jahn 's results give a difference of on .
But such experimental differences at low concentration cannot be avoided , and , as a consequence , we find that Jahn 's figures for the three salts , yield three different values for the constant , which should theoretically be the same for each .
It seems clear that these differences in the value of , which is the limiting value of at infinite dilution , can only be due to experimental errors , and we shall , therefore , take a fixed value for which is the same in all three cases .
The reciprocal of the value which is yielded by the above figures is .
This is a higher value than that which is usually accepted .
Bedford 's figures*seem to be the most recent and the most reliable .
The highest figures which he obtained for the molecular depression of the freezing point with KC1 solutions were and , corresponding to effective molecular depressions of and .
We shall assume as the of our calculation the right figure , instead of .
This gives the value of our constant , instead of .
Taking this value of we might deduce from Jahn 's figures for each salt a value of for the salt which would be the mean value yielded by all the figures .
Since , however , the probable error in the value of given by any particular ation is inversely proportional to we shall deduce the value of from the mean of the two highest concentrations given in each case by Jahn .
As we are dealing with a straight line law this will be likely to give us the best values .
The figures which Jahn gives for these concentrations are as follows:\mdash ; Table IX.\mdash ; Jahn 's Values for Depression of Freezing Point .
Using the values of the hydration numbers and ionisation constants before given we arrive at the following figures by taking the mean of each pair of observations .
' Roy .
Soc. Proc 1910 , , vol. 83 , p. 464 .
is the number of molecules of vater per gramme-molecule of solute .
From these figures , taking , we arrive at the values for and * which are given in the following table:\mdash ; Table X.\mdash ; Values of the Constants and The values of are easily obtained by the use of equation 17 ) , taking the given data and putting , and taking the value of furnished by the empirical volume relation .
Our complete table of the constants of the empirical equations and of the values deduced from them by means of equations ( 4 ) , ( 12 ) , 13 ) , ( 14 ) , 15 ) , and ( 16 ) is as follows:\mdash ; Table XI.\mdash ; Constants Derived from the Empirical Equations .
8 .
Application to Ionic Votumes at Infinite Dilution.\mdash ; We are now in a position to complete the data required , in order to at the analysis of the ionic heats which is contained in Section 2 of this paper .
For this purpose it is only necessary to apply the figures obtained to deduce approximately the number of molecules of water combined with each pair of ions .
The toral volume of a gramme-molecule of ions in cubic centimetres is 1000 From this we must deduct the volume of the ionic nuclei to get the volume of the combined water , and multiply by the mean density of this water to get 168 Mr. W. R. Bouffield .
[ Oct. 18 , the weight .
Hence the weight of water combined with a gramme-molecule of ions of the so.lute at infinite dilution is Using the data given above the weights of combined water deduced are as follows:\mdash ; which expressed in molecules of water are 21.3 These numbers should be integers at infinite dilution , and taking the nearest integers we get for the molecules of water combined with the pairs of ions at infinite dilution , It has , of course , been observed that our analysis has not given us any means of separating the volumes of the several ions , or of the amounts of their separate water combination .
Our faotor is a factor which deals with the sum of the ionic volumes , and our quantity is the sum of the volumes of the nuclei .
But having regard to the mobilities of and Cl and to the transference numbers , we may reasonably conclude that the and Cl ions are of nearly equal size .
we took the Cl ion to be combined with five molecules of'water , we should have for the water combination of the other ions .
Na .
This is a very attractive looking series and there is something to be said for it .
Let us compare certain corresponding properties of the and Cl atoms and ions\mdash ; Now the joint contracted volume in the ionic state is 30 .
This is of the original joint volume 72 .
If the compressibilities were equal and the internal pressures in the ions equal , the respective contracted volumes would be ; nearly This would allow the Cl ion to have one more molecule of water than the ion , and yet have a less volume and greater mobility .
Unfortunately , Washburn 's transference experiments in order to determine the relative hydration*indicate that the numbers should be the other way , and that the ion has one more molecule than the Cl ion , in which the numbers for the separate ions yielded by our figures would be .
Cl. Na .
5 9 17 Whether Washburn 's method , which rests on the mixture with the solution a non-electrolyte , such as sucrose or raffinose , is completely reliable ia possibly open to some question .
If the Cl ion be taken to be combined with four molecules of water , the figures which his experiments yield are which may be compared with those above .
His figures are only relative , as * his method does not.yield absolute results , and the figures are for concentra- .
: tions of about normal instead of infinite dilution .
The division of the combined water between the separated ions is a matter of great interest , but fortunately it is quite unnecessary for the main purpose of this paper , which was deyeloped in the second section .
In conclusion , we may again refer to the fact that the precise figures .
depend , according to this method , on what is to be taken as the true value of the freezing-point constant .
If the low value be taken instead of the value , the figures come out as follows:\mdash ; but the general results of the second section of this paper are not affected thereby , though the value of the calorific constant for the water molecule in combination would be slightly altered .
* Technological Quarterly , ' 1908 , vol. 21 , p. 288 .

