1892 .
] Voltaic Cells with Fused Electrolytes .
III .
" Voltaic Cells with Fused Electrolytes ; " By J. Brown .
Communicated by Professor Eyerett , F.R.S. Received May 27 , 1892 .
In attempts to compare the observed electromotive forces of voltaic combinations with the theoretical values found by Thomson's* law , the discrepancies observed in many cases , and the difficulties experienced in ascertaining their nature and origin , are well known .
It is evident that , in order to compare observations with theory , we must know exactly the facts on which the theory is to be founded , i.e. , the true nature of the chemical reaction in the cell , and its relation to the electric phenomena to be observed ; and for this purpose simplicity in the construction of the cell is , in the present state of knowledge , almost a necessity .
In many of the cells that have been examined heretofore , this condition of simplicity is not realised ; and the nature of the chemical action in them is not definitely ascertainable because of the complicated nature of the substances employed .
I have referred to this point in the introduction to a paper on the " Role of the Cation in Voltaic Combinations ; " f and in a foot-note on p. 462 , of that paper I have again referred to what is a very frequent cause of uncertainty , viz. , the action of the solvent.\#151 ; usually of course water\#151 ; as distinct from the dissolved electrolys-able substances .
This uncertainty is still more marked in the case of some of Wright and Thompson 's experiments , J where insoluble salts suspended in solutions were supposed to be alone active to the exclusion of both the water and the dissolved substances .
The solvent may act chemically on the metals , and affect the electromotive force , either directly as an electrolyte , or indirectly by the formation of films insoluble in itself , which act differently from the metals or protect them from chemical action .
LaurieS and Braun|| have respectively shown that the lowering of the forces of aluminium and magnesium in aqueous solutions is probably due to such protective films ; Julius M. Werner^f considers the cause of the comparative inaction of pure zinc in acids to be that , at the moment of immersion , it becomes enclosed in a continuous film of hydrogen which protects it from further action ; while with impure zinc the hydrogen forms on the impurities only , leaving the zinc exposed .
* * * S * ' Phil. Mag. , ' ser. 4 , vol. 2 , p. 434 , 1851 .
f ' Phil. Mag. , ' ser. 5 , vol. 31 , p. 449 , 1891 .
J ' Phil. Mag. , ' ser. 5 , vol. 19 , p. 211 , 1885 .
S ' Phil. Mag. , ' ser. 5 , vol. 22 , p. 213 , 1886 .
|| ' Phil. Mag. , ' ser. 5 , vol. 27 , p. 209 , 1889 .
' Ber .
Deut .
Chem. G-esell .
, ' vol. 24 , p. 1785 .
Mr. J. Brown .
[ June 16 , The probability of such actions does not , I think , usually receive sufficient consideration in electrochemical investigations .
Again there is , in connection with the solvent , the uncertainty whether , in calculating the mechanical equivalent of the action in the cell , we should include the heats of solution of substances formed by the action .
Where the product is easily soluble , the heat of solution is probably adjuvant ; where insoluble , there is of course no heat of solution adjuvant ; where it is soluble with difficulty , the heat of solution may be partly or to some unknown extent ad-juvant .
The consideration of these difficulties , and especially the complication probably caused by the action of the water on magnesium and aluminium , as mentioned above , led me to the consideration of a much simpler form of cell , in which the solvent with all its complications is abolished , and the simple electrolyte in the fused state used alone .
The number of possible combinations of this kind is limited by the considerations that the electrolytes must be fusible at reasonable temperatures ; and that the boiling point of any component in a cell must not be below the fusing point of the electrolyte in contact with it .
A few of the haloid compounds of the metals are suitable ; and the chlorides were chosen to begin with , as in some respects best known .
The results obtained show at all events the direction in which to look for an exact accordance between theory and experiment .
I. Cells with Simple Chlorides .
The cells were of course of the two-fluid type , each metal being in contact with its own chloride only , while the two chlorides were in good contact with each other .
The type may be represented for instance as Zinc | ZnCl2 | MgCl2 | Magnesium .
We cannot say the fused electrolyte is always free from the complication of dissolved oxygen , for in some cases this makes itself evident by the formation of oxychlorides .
This very formation of oxychlorides may , however , remove the oxygen from the sphere of action .
We must admit also that in some of the combinations tried , action goes on with open circuit , since several of the metals , notably iron ; and copper , were somewhat corroded after immersion in the fused salt .
Further , while the difficulty about the heat of solution of the nascent salts is avoided , there is a question as to whether the heat of solution of one fused chloride in the other , the heat of combina1802 .
] Voltaic Cells with Fused Electrolytes .
77 tion of the two chlorides with , each other affects the electromotive force of the cell .
It is impossible to answer this question without further data of various sorts .
The small irregular variations from theory shown in the tables may possibly be due to some action of this kind .
The results of experiment show , however , that these few complications have but a small effect ; very much less than those of the aqueous form of cell with its metals coated with oxide and hydrogen films , and its mixture of two or more electrolytes and dissolved oxygen , such as might perhaps be represented as Zinc + ZnO + H | ZnCl2 + H20 + 0 | MgCl2 + H20 + 0 | Magnesium + MgH202 + H , which seems to be too complicated for exact calculation .
In arranging the practical details , care was taken to keep the metals as clean as possible , the contacts good , the electrolytes fused " but not boiling , and the temperatures fairly equal at the two metals ; also to prevent , as far as possible , the two chlorides from diffusing into one another .
After trying various arrangements , that of which fig. 1 represents about half the actual size was adopted .
Fig. 1 A V-tnbe A of hard glass , formed as shown , and supported by its longer limb , contained one of the chlorides , in which dipped the wire or strip of appropriate metal C. The bend of this tube was heated as required by one or two Bunsen burners .
Into the shorter limb was inserted the other tube B , its lower end contracted somewhat , and filled with a ping of asbestos , to act as a porous partition .
This smaller tube contained the second chloride , with its appropriate metal D. When the metal represented by 0 was fusible at nearly Mr. J. Brown .
[ June 16 , the same temperature as its chloride , it was first fused in the bend of the tube A , and then its chloride fused on it in the shorter limb .
The melted metal was connected to the Thomson electrometer by a wire of either copper or iron dipping in it .
When this wire was of copper , long contact with the fused metal seemed to dissolve it somewhat ; and a slight error may thereby be introduced in the cases of tin , zinc , and cadmium ; but this is probably not large in any case .
For these fusible metals , a clay tobacco pipe was sometimes substituted for the tube A , the connecting wire passing through its stem .
When the metal in the small tube was of this easily fusible kind , it was found best to add another asbestos plug , as at e , fig. 2 , which was permeated by the chloride below it at , and supported the fused metal g with its connecting wire/ .
While all mixture of the two chlorides was to be avoided as far as possible , it was considered especially necessary to keep the chloride of higher combining heat , which to save repetition we may call H , from being permeated by that of lower combining heat L , for the obvious reason that , when L mixed with H , its metal became at once reduced on the metal in H , and the conditions were thereby altered .
If , for example , in the cell Copper | Cu2Cl2 | ZnCl2 | Zinc the copper chloride ( 1/ diffuses into the zinc chloride ( H ) , copper becomes reduced on the zinc , and the electromotive force is lowered .
To obviate or mitigate this difficulty , the following precautions were taken:\#151 ; The \/ -tu1ae A was first set up , and the chloride fused in it .
When the arrangement was such that chloride H was in the small tube B , this tube was first held in the Bunsen flame till the chloride fused and saturated the porous plug to the end or nearly so .
The tube B was then inserted in its place , any small mixture of H with L not being considered of importance .
If , on the other hand , H was in A , and L in B , then B was simply heated till L was about fusing point , and was immediately inserted in A , so that H had time to soak up into the plug before L soaked down .
The following was the usual plan of observing the electromotive force :_As soon as the electrolytes were in order , and the metals , previously connected to the electrometer , were inserted , the deflexion was noted immediately , and again after about five minutes .
The metals were then , if solid , taken out , c'eaned , and re-inserted , and the observations repeated .
In these sets of observations the variation from the mean of any given experiment was less with fused metals than with solid metals , the latter being more liable to surface alterations .
The adopted mean of each set was derived from those experiments in which the theoretical conditions were most strictly fulfilled .
1892 .
] Voltaic Celle with Fused Electrolytes .
79 As regards the quality of the substance used , the zinc , lead , iron , tin , and copper , and the chlorides of magnesium and tin , were purchased as pure .
The other metals and chlorides were the substances as supplied by the chemical manufacturers under their respective labels .
In the case of metals forming more than one chloride , that one was chosen in which a chlorine molecule Cl2 gave the highest heat of combination .
The corresponding formulae are stated in the tables of results .
The heats of combination employed are Thomsen 's .
Table I gives the results obtained with the chlorides of the metals as stated , together with the theoretical values ; the unit employed being the electromotive force of a Daniell cell , which for the Daniell I used may be taken as about 1*13 .
In calculating the theoretical value , it is assumed that the reaction is simply the substitution , in the double fluid electrolyte4 , of the metal with higher combining heat for that with lower ; and , therefore , that the theoretical electromotive force is , according to Thomson 's law , simply proportional to the difference of their combining heats with chlorine , which is the anion of the electrolyte .
This difference , divided by 50,130 , the heat equivalent of a Daniell cell , gives the theoretical electromotive force for each pair , in terms of the Daniell .
In practice , it is more convenient to divide , once for all , the heat of combination of each metal by the heat equivalent of the Daniel ! .
The quotients thus found are placed under their respective metals in the table .
It seems possible that these numbers may represent the true differences of potential at contact of metal and liquid .
The theoretical electromotive force for any pair of metals is then given by the difference of their respective numbers .
In the table , this will be found in heavy type , at the intersection of the vertical and horizontal lines passing through any given pair of metals ; the metal to the right forming the negative pole of the arrangement .
Immediately below each theoretical electromotive force , is placed that found by experiment , being the mean of the number of sets of experiments indicated by the figure in brackets below it .
The highest and lowest observed values are stated to the right and left of it .
In this table it will be observed that , while the mean experimental result for any combination of the metals zinc , lead , and tin agrees fairly well with Thomson 's law ; those combinations in which silver , copper , or cadmium occur do not exhibit such good agreement ; and closer examination shows that much more consistent results could be obtained by applying a constant correction for each metal .
I have determined , by a tentative process , the best values for these corrections , and the results are shown in Table II .
It is only a question of differences ; and if the correction for cadmium be taken as zero the | ' corrections for all the other metals will be positive .
A justification for this selection of cadmium will be found in a later paragraph .
Silver in Ag-2C12 .
1-172 oo o Table I. Copper in Cu..Cl2 .
1-312 0-14 0 08 , 0 10 , O il ( 2 ) Zinc in ZnCl2 .
1-939 Cadmium in CdCl . ? .
1-86 0079 012 , 013 , 0-13 ( 3 ) Lead in PbCL .
1-651 0-209 O il , 0 14 , 018 ( 4 ) 0-288 017 , 0 22 , 0 25 ( 6 ) Tin in SnCL\gt ; .
1-612 0-039 0 05 , 0-05 , 0-06 ( 3 ) 0-248 015 , 016 , 017 ( 3 ) 0-327 0-28 , 0-31 , 0-35 ( 5 ) 0-3 014 , 016 , 019 ( 2 ) 0-339 018 , 0 21 , 0 24 ( 5 ) 0548 0-26 , 0-31 , 0*36 ( 4 ) 0-627 0-40 , 0-46 , 0-51 ( 4 ) 0-44 0-28 , 0 30 , 0 32 ( 2 ) 0-479 0 30 , 0 31 , 0 32 ( 3 ) 0*688 0 37 , 0-39 , 0 41 ( 3 ) 0-767 0 51 , 0 52 , 0 53 ( 2 ) dr .
J. Brown .
[ June 10 1892 .
] Voltaic Celh with Fused Electrolytes .
Table II.\#151 ; Table I modified .
Zinc in ZnCl2 .
1-939 + 0-049 \#151 ; 1-988 Cadmium in CdCl2 .
1*86 0*128 013 Xiead in PbCl2 .
1*651 + 0*099 = 1*750 0*110 014 0*238 022 Tin in SnCl2 .
1*612+ 0*093 = 1*705 0*045 005 0*155 016 0*283 031 Copper in Cu2Cl2 .
1*312 + 0*23 = 1*542 0*163 016 0*208 0*21 0*318 0*31 0*446 046 Silver in Ag ; 2Cl2 .
1*172+ 0*273 = 1*445 0*097 0*10 0*260 0*30 0*305 0*31 0*415 0-39 0*543 0*52 Table II shows these corrections applied to each metal , and a comparison of these results with the mean experimental results , the order of arrangement being the same as in Table I. It will be seen that , with two exceptions , lead-cadmium 0*03 , silver-tin 0*04 , the discrepancy is always below 0*03 .
This is interesting so far as it goes ; and the next step is to inquire if a physical cause can be assigned for these constants .
That they have their chief origin in the high temperature at which the experiments were made seems very probable .
It has been shown by Gladstone and Tribe* that in the cells Silver , j Ag2Cl2 | Silver Copper | Cu2Cl2 | Copper if one of the metal-electrolyte junctions be heated more than the other , a difference of potential between the two metals is established , and a current may be produced from hot to cold through the electrolyte .
I have made similar cells , using \/ -tubes containing the fused salts , with wires of the appropriate metals in each limb .
In the copper cell , with one junction a little above the fusing point of the double chloride Cu2C12KC1 ( whose low fusing point allows a greater difference of temperature to be employed ) , and the other junction at a red heat , the observed electromotive force was about 0*1 Daniell , and gave a strong current through a low resistance galvanoscope , which could be reversed very prettily by simply moving the Bunsen VOL. LII .
* 1 Phil. Mag. , ' feer .
5 .
, vol. 11 , p. 508 , 1881 .
G Mr. J. Brown .
[ June 16 , lamp from one side of the cell to the other .
A silver chloride cell of the same form gave a similar result .
Lead was tested in the fused state in two asbestos-plugged tubes , like B , fig. 1 , immersed in a \/ -tube containing fused lead chloride .
The other easily fusible metals , tin , cadmium , and zinc , were tested in \/ \/ -shaped tubes , open at the apex in the middle , for introduction of the chloride after the metal had been fused in the two lower bends .
With all the easily fusible metals , except zinc , the tubes were made sufficiently long to allow the portion of metal surrounding the copper connecting wire to remain solid , so as to avoid risk of alloying from long fusion in contact with the copper .
In all cases , except cadmium , with which in three carefully-made experiments , I found no decided effect , the hotter metal was found to be the negative pole of the arrangement .
Zinc was somewhat doubtful , and may have been influenced by alloying with the copper connection .
To give some idea of the magnitudes of the electromotive forces of these thermal cells , the results of these rough experiments are given in column A below .
In column B are the corrections employed in Table II for comparison .
A. B. \#187 ; Silver ... ... ... 0-07 Daniell 0'273 Copper ... ... ... 0T0 " 0*23 Tin ... ... ... ... .
0-01 " 0-093 Lead ... ... ... ... 0-015 " 0'099 Cadmium ... ... . .
0"00 " O'OO Zinc ... ... ... .
002 ?
" 0-049 There is a kind of rough correspondence between the numbers in the two columns , but the inexact nature of the experiments of column A precludes a strict comparison .
I had no means of measuring , otherwise than by rough estimation , the difference of temperature between the two sides of the cells ; but as a mere approximation it might be taken for tin as rather less than the difference between the freezing and boiling points of stannous chloride , or say 300 ' C. ; for zinc and copper it was rather more than this , and for lead and silver less , as their chlorides are less easily fusible , and the higher temperature was limited to the softening point of the glass tube .
Besides the experiments of Gladstone and Tribe mentioned above , others bearing less definitely on the subject have been made by Andrews , * Hankel , j " and more recently by Mr. T. Andrews , J of Sheffield .
* ' Phil. Mag. , ' ser. 3 , vol. 10 , p. 433 , 1837 .
t ' Pogg .
Ann. , ' vol. 103 , p. 612 , 1858 .
% ' Boy .
Soc. Proc. , ' vol. 38 , p. 216 , 1885 .
1892 .
] Voltaic Cells with Fused Electrolytes .
As regards these currents due to difference of temperatui'es , Andrews concluded , from the non-corrosion of the platinum wires used by him , that the action was in his experiments thermo-electric .
Gladstone and Tribe state that in their experiments " it is difficult to imagine that chemical action in any way initiates the current .
" Their experiments show clearly , however , that the phenomenon is of electrolytic nature , since the current formed is accompanied by the solution of metal at the hot side , and its deposit in crystalline form at the cooler electrode .
The same effect was also observed in my experiments with copper in its fused double chloride ; a large deposit of crystalline metal was found on the cooler wire after the current had passed for an hour .
The action must , therefore , be of electrolytic or voltaic character ; and the application of Thomson 's law to these unequally heated cells wrould suggest that , with the apparent exception of cadmium , there is an increase with temperature of the heats of combination of the metals with chlorine , conjoined with a possible Peltier effect .
Such a variation is , in fact , well ascertained in other cases ; and Thomsen* gives a formula for its calculation from the specific heats of the bodies involved .
The want of data for the fused salts prevents its further discussion here .
Assuming this variation of the heats of combination to exist , and assuming , also , that the chemical energy is all adjuvant at the various temperatures , !
we could directly account for the additive corrections empirically deduced from Table I and applied in Table II , the numbers in Table I being founded on Thomsen 's heats of combination obtained at 18 ' to 20 ' , while my experiments were made at temperatures several hundred degrees higher .
It will thus be important to determine the temperature co-efficients of these cells , and I hope to do so at some future time .
Another point demanding careful attention is the extent to which these cells are reversible .
The following experiments bear on this subject:\#151 ; Two clay tobacco pipes were clipped by their stems in wooden supports , one in each , and placed with their bowls touching and facing upwards .
One contained some melted zinc , and the other melted tin , with iron connecting wires through the stems .
The respective chlorides were fused over these metals and connected together by a bunch of asbestos , a notch having been cut at the point * ' Thermochem .
Untersuchungen , ' vol. 2 , p. 54 .
f [ This assumption is of course provisional .
It was necessary because of the absence of any investigation here as to what part , if any , of the energy is non-adjuvant .
Helmholtz and other well-known authors have discussed this question in connexion with the ordinary aqueous form of cell , but I have not attempted its investigation as yet in the present case .
When the temperature coefficients and other data have been obtained this may be undertaken.\#151 ; 2 , 1892 .
] G 2 Mr. J. Brown .
[ June 16 , of contact of the bowls to receive this .
The bowls were covered by a porcelain lid .
The cell in this form gave a deflexion on the electrometer of 43 divisions .
It was then subjected to the following consecutive tests .
The entries are to be read consecutively from left to right , the first observation in the second column being immediately subsequent to the last observation in the first column , and so on .
" Short circuited " means that the current of the cell was passed through a low resistance astatic galvanoscope , sending the needle to the stops ; " reversed " means that a nearly equal current , from a Daniel !
cell , was sent in the opposite direction through the cell under examination and the galvanoscope .
After the duration of either of these currents for the times given below , the force of the cell was again measured on the electrometer , with the results stated :\#151 ; Zinc | ZnCl2 | SnCl2 | Tin .
Short circuited 15 seca .
30 secs .
5 min. Means .
Deflections Reversed 41 30 secs .
40-5 30 secs .
42 5 min. 41 -2 Deflections 41 41 42 41-8 A similarly arranged zinc-lead cell gave , when set up , a deflection of 22 , and then the following deflections were noted , immediately after the cell had been either short circuited , reversed , or rested on open circuit , for a period of four minutes in each case , except the first two observations , which were for one minute each .
Zinc | ZnCl2 | PbCl2 | Lead .
Means .
Short circuited. . . .
22-5 20-0 ISO 18-5 19-75 Reversed 20-5 18-5 18 5 18 -0 18 -87 Rested . .
18 -0 17-5 18 -0 18-87 Zinc-silver and zinc-copper cells were tested in the form adopted for the general investigation .
The silver chloride was in the tube B , fig. 1 , with an asbestos plug 3 cm .
long , to prevent mixing of the chlorides in the separate tubes during the experiment .
The copper chloride was in a similar tube , corked to exclude air .
The durations of short circuit , rest , and reversal were each four minutes , except in the second and third observations with the zinc-copper cell , in which they were each one minute .
1892.J Voltaic Cells with Fused Electrolytes .
85 Zinc I ZnCl2 | Ag2Cl2 | Silver .
Means .
Short circuited. .
" R.ArPffiflfl . .
a . . .
85-5 87-5 85 87 85-5 87 84 87 78-5 81 83 *7 85*8 86 -0 85-5 86 84 810 \#151 ; 84 *6 XVv 2 ?
L vVA * . . . .
f Zinc | ZnCl2 [ CuCl2 | Copper , Short circuited ... ... 49 55 44 46-5 Means .
48 -6 Reversed \#151 ; 46 46 41 \#151 ; 50 45 *8 Rested 48 \#151 ; 44 43 49-0 46'0 Except in the zinc-silver cell , and to a minute extent in the zinc-tin cell , there is no evidence of polarisation by reversal .
The other two cells show the remarkable property of possessing a greater electromotive force after short circuiting than before it ; and since this recurs repeatedly in the observations on both cells , it would seem to be more than accidental .
II .
Cells with Double Chlorides .
As I have already stated , the present investigation was suggested by the behaviour of magnesium and aluminium in aqueous electrolytes ; and as a matter of fact these metals were experimented on in the more complicated cells described below before the simpler cells of Table I were tried .
Owing to the high fusing point of magnesium chloride , and the volatility of the aluminium compound , it was inconvenient to work with these chlorides in the fused state ; and this difficulty had to be surmounted by employing the double chlorides of magnesium and potassium , and of aluminium and sodium .
A similai method was adopted with ferrous chloride , which is infusible by itself .
It may be objected that this method introduces a solvent , the very complication which it is desired to avoid , since potassium chloride may be looked upon as the solvent of the other chlorides .
But in comparing fused potassium chloride with water as a solvent of chlorides , there are two important differences , ( i ) With potassium chloride , the anion is the same as that of the dissolved substance , while with water it is not .
( ii ) The heat of combination K , Cl is greater than that of any of the other chlorides X , Cl dissolved in it in these experiments ; therefore the metal X when immersed Mr. J. Brown .
[ June 16 , in the double salt does not decompose the potassium chloride , but only reacts with its own chloride .
Hence we have grounds for assuming that potassium chloride remains practically inert , so far as concerns any direct chemical action between it and the metal immersed .
The potassium chloride is , however , active in another way , viz. , in combining with the various metallic chlorides to form their doable salts ; and the heat of this combination should no doubt be taken into account .
Unfortunately , the heats of formation of these double salts appear to be unknown .
If they are nearly equal , so that their differences are small in comparison with the whole heat equivalent of the cell , their effect will be negligible .
A comparison of Tables I and II shows that this is probably so , but naturally the results of these experiments with double salts cannot be compared with theory quite so satisfactorily as those with simple chlorides .
In any given experiment the two chlorides employed should be either both double or both single , to avoid the complication which would otherwise be introduced by the formation of a double salt on one side and not on the other .
In the preliminary stage of this work , some experiments were made without observance of this rule , and the discordant results obtained showed that it cannot be disregarded .
It was therefore necessary to form double chlorides for all the metals employed ; and the combinations whose formulas appear in Table III were found , either by reference to the books or by experiment , to be suitable .
The experiments were carried out in the same way as the first set , and the results are given in Table III , where for magnesium , iron , and aluminium , they are stated in the same form as in Table I. For the other metals , only single observations , as a rule , were made , for comparison with Table I. The corresponding numbers in both tables aie at least of the same order of magnitude , the difference only in one case appearing in the first decimal place .
Table III .
ifSl I .
Magnesium in \#166 ; MgCl2XCl .
3 012 Aluminium in Al2ClBNao012 .
0*871 Zinc in 2*141 0*78 , 0*75 , 0 64 ZnCLE^CL .
1*939 0*202 1*073 0*14 , 0*17 , 0*22 0*80 , 0*87 , 0*90 Cadmium in ( 4 ) ( 3 ) Cd2Cl4KCl .
1*86 0*079 0*281 1*152 019 0*40 , 0*44 , 0 48 1*04 , 117,1*21 Lead in PbCl2K2Cl2 .
( 4 ) ( 3 ) 1*651 0*209 0*288 0*490 1*361 ( 0*13 ) ( 0*28 ) 0*43 , 0*45 , 0-47 1*17 , 1*19 , 1*21 Iron in ( 3 ) ( 3 ) FeCLjJCjCLj .
1*637 0*014 0*223 0*302 0*504 1*375 -0*012 , -0 02 , -0*03 0*02 , 0*09 , 0*14 0*33 , 0 33 , 0 33 0*40 , 0*45 , 0*48 1*17,117,117 Tin in ( 2 ) ( 3 ) ( 3 ) \lt ; 3J ( 3 ) SnCl2KCl .
1*612 0*025 0*039 0*248 0*327 0*529 1*400 -0*02,0-03,0*08 007 0*14 033 0*46 , 0 56 , 0*61 1*19 , 124,1*27 Copper in ( 2 ) ( 4 ) ( 3 ) Cu2Cl2K2Cl2 .
1*312 0*300 0*325 0*339 0*548 0*627 0*829 1*700 019 0*23 , 0 27 , 0 30 0*20 0*33 0*54 0*60 , 0 66 , 0*80 1*40 , 1*44 , 1*52 Silver in ( 3 ) ( 3 ) ( 4 ) AftOlsKjCls .
1172 0*140 0*440 0*465 0*479 0*688 0*767 0*969 1*840 0*10 0*36 0*35 , 0*40 , 0*50 0*35 0*50 059 0*78 , 0*80 , 0*81 1*55,1*57,1*60 ( 3 ) ( 3 ) ( 3 ) 00 1892 , ] Voltaic Cells with Fused Ele* 88 Mr. J. Brown .
[ June 16 , While the results in Table III are naturally not so regular as those in Table I , most of them are nearer the theoretical values than those of the same combinations in aqueous solutions .
In this respect the results for cells in which magnesium or aluminium is concerned are particularly striking .
For comparison I take the following numbers ( in volts ) for metals in aqueous solutions of their chlorides , from Wright and Thompson ; * Braun , t has found a similar discrepancy for magnesium in aqueous cells , and attributes it to hydrogen deposited on the magnesium .
Wright and Thompson .
Metals in aqueous solutions of their chlorides .
Metals in fused chlorides .
Table III .
Calc .
9 Obs. Diff. Calc .
Obs. Diff. Magnesium | zinc 1*634 0*702 0*932 1*07 0*87 0 *20 Magnesium | cadmium . .
2*000 1*030 0 -970 115 1 *17 -0*03 Aluminium | zinc 1*008 -0*280 1*288 0 *20 0*17 0*03 Aluminium | cadmium ... 1*374 0*05 1*324 0*28 0 *44 -0*16 A still closer agreement between the observed and theoretical numbers for the metals in Table III is produced by the application of empirical corrections , as in Table IV ; and these are of a similar order of magnitude in the two tables , except in the case of zinc .
On testing iron , aluminium , and magnesium for electromotive force due to difference of temperature , iron was found to behave like copper , the hot metal being the negative pole .
Aluminium in its double chloride with sodium gave , at first , a strong similar effect ; but after some twenty minutes\#151 ; when a dark deposit , which appeared to contain lead , carbon and iron reduced by the immersed metal , had ceased to separate , and the liquid had become colourless and clear\#151 ; the effect ceased .
It may have been due to impurity in the chloride .
With magnesium I could not detect any electromotive force due to difference of temperature .
Therefore , considering the uncertainty introduced by these irregularities , and the complication introduced by the use of double salts , cadmium has been allowed to remain as the zero of the scale of corrections , as in Table II .
My experiments were carried out in the laboratory of Queen 's College , Belfast , and I am much indebted to Professor Everett for permission to work there and for the use of his electrometer and other apparatus .
I am also indebted to Professors Everett and * 'Phil .
Mag. , ' ser. 5 , vol. 19 , pp 211 , 213 , 1885 .
f ' Wied .
Ann. , ' vol. 16 , p. 578 , 1882 .
Table IV , Mag-nesium .
3*012-0*04 = 2*972 Aluminium .
2-141 + 0-095 = 2-236 0-74 075 Zinc .
1 939 + 0-124 = 2-063 0-17 017 0-91 0-87 Cadmium .
1-86 0-20 019 0-38 044 1-11 117 Iron .
1-637 + 0-143 = 1-780 008 0-09 0-28 033 0-46 045 1-19 117 Lead .
1-651 + 0-115 = 1-766 001 0-02 0-09 013 0-30 0-23 0-47 045 1-21 119 Tin .
1-612+ 0-113 = 1-725 0-04 0-07 005 003 0-14 0-14 0-34 033 0-51 0-56 1-25 1-24 Copper .
1 312-^0-223 = 1 535 0*19 0-19 0-23 0-20 0-25 017 0-33 033 0-53 054 0-7 066 1-44 144 Silver 1-172 + 0-2S6 = 1 408 0 13 010 0-32 0-36 0-36 0-35 0-38 0-40 0-45 0-50 0-66 059 0-83 0-80 1-56 1-57 GO Voltaic Cells with Fused .
90 Voltaic Cells with Fused Electrolytes .
[ June 16 G. F. Fitzgerald and to Mr. J , Larmor for several important sugges .
| tions .
Since writing the above , my attention has been called to a paper by L. Poincare , * " Sur les Piles a Electrolytes Fondus et snr les Forces 1 Thermo-electriques a la Surface de contact d'un Metal et d'un Sel j Fondn .
" M. Poincare has measured the thermo-electric actions in one or two thermo-cells similar to those which I have described , and finds the following thermo-electromotive forces for a difference of 1 ' between the two metal electrolyte junctions , the sign being that of hotter pole :\#151 ; Silver in silver nitrate ... \#151 ; 0*00027 volt .
Zinc in zinc chloride ... ... + 0*00013 " Tin in stannous chloride. . . .
0*000028 , , The sign for tin does not appear to be given ; that for zinc is the reverse of my result .
M. Poincare has also measured the electromotive force of one voltaic cell with fused electrolytes , Zinc | ZnCl2 | SnCl2 | Tin , and finds it to be 0*335 volt .
My result , in Table I , when reduced to volts , is 0*350 .
He finds for the value just after solidification of the electrolyte , 0*37 volt , and considers that , in order to obtain the true heat equivalent of a cell with fused electrolytes , the latent heat of fusion of the salts used should be deducted from Thomsen 's heats of combination for the solid salts .
There would appear to be , on theoretical grounds , some doubt on this point .
It could be experimentally tested more definitely if data for these heats of fusion were available .
Synopsis .
In instituting a comparison of the observed electromotive forces of cells with their theoretical values , as given by Thomson 's law , the simpler the construction of the cell the more easy is it to ascertain the nature of the chemical action going on in it , upon which the calculation rests .
In cells with aqueous electrolytes , the solvent introduces several irreducible complications , arising from possible action of the solvent itself as an electrolyte , from the formation of insoluble films ( either inactive and protective as oxides , or active , as hydrogen ) , or from the uncertainty of calculations involving heats of solution of the products of voltaic reaction .
When the liquidity of the electrolyte is produced by fusion instead * ' Comptes Rendus , ' vol. 110 , p. 339 , 1890 .
1892 .
] Physiological Action of Nitrites of Paraffin Series .
91 of solution , these complications are , to a large extent , avoided ; and , as a matter of experiment , two fluid cells made up of metals immersed each in its own fused chloride , give results nearer the theoretical values than those obtained with aqueous solutions .
This is specially noticeable in the case of metals with high heats of oxidation as in the case of magnesium and aluminium .
The electromotive forces of cells containing pairs of the metals , tin , lead , and zinc , come out nearest to the theoretical values .
The other metals which were tried did not give so close an agreement , hut can be brought into agreement by applying constant corrections , one for each metal ; and reasons are given for attributing these corrections to the high temperatures of the cells as compared with the temperatures for which the recognised heats of combination are true .
Four of the cells were tested by passing currents through them in both directions alternately , and noting the electromotive force after the passage of each current .
Polarisation was observed to a small extent in one case\#151 ; zinc-silver ; in the others it was practically absent .
IV .
" The Physiological Action of the Nitrites of the Paraffin Series considered in connexion with their Chemical Constitution .
Part II .
Action of the Nitrites on Muscular Tissue and Discussion of Results , " By J. Theodore Cash , M.D. , F.R.S. , Professor of Materia Medica in the University of Aberdeen , and Wyndham R. Dunstan , M.A. , Professor of Chemistry to the Pharmaceutical Society of Great Britain .
Received June 15 , 1892 .
( Abstract .
) Continuing the examination of the physiological action of various pure organic nitrites of the paraffin series ( Part I ; ' Roy .
Soc. Proc. , ' 1891 ) , the authors have studied their effect on striated muscular tissue .
When the vapours of these nitrites come into contact with the muscle a paralysant effect is observed .
All the experiments were made with the triceps and gastrocnemius of Rana temper aria .
The muscle was contained in a specially constructed air-tight chamber , which , whilst it admitted of connexion being made between the muscle and the recording apparatus , rendered it possible to bring into contact with the muscle the vapour of a known quantity of the nitrite without any loss taking place .
A very extensive series of experiments was necessary , as it is unsafe to contrast the gastrocnemius of one frog with that of another , so that the action of every

