]\gt ; IV .
The Hatogen Hydrides as Conducting Sotvents .
Part I.\mdash ; The Vapour Pressures , Densities , Surface Energies and Viscosities of the Pure Sotvents .
Part II.\mdash ; The Conductivity and Molecutar Weights of Dissotved Substances .
Part III.\mdash ; The Tramport Numbers of Certain Dissotved Substances .
Part Abnormal Variation of Molecular Conductivity , etc. B. D. STEEL , D.Sc .
, D. McINTOSH , , D. Sc. , and E. H. ARCHIBALD , M.A. : Ph. D. ( late 1851 Exhibition Scholars ) .
Communicated by Sir WILLIAM RANSAY , K. C.B. , F.R. Received February l , \mdash ; Read February 16 , 1905 .
PART I. Vapour Pressures , Densities , Surface Energies and Viscosities of the Pure Solvents .
D. McINTOSH and B. D. STEEL .
ALTHOUGH our knowledge of the ionising power of non-aqueous solvents has been considerably increased during recent years by the investigations of WALDEN , FRANKLIN , KAHLENBERG , and others , the liquefied halogen and sulphuretted hydrogen have received little or no attention .
GORE ( ' Phil. Mag. ' ( 4 ) , 29 , p. 54 ) , who experimented at ordinary temperatures , found that the hydrides of chlorine , bromine , and iodine were very feeble conductors .
BLECKIODE ( ' Pog .
Ann. ' ( 2 ) , 23 , p. 101 ) stated that hydrogen bromide conducts slightly ; while HITTORF ( ' Pog .
Ann. ' ( 2 ) , 3 , p. 161 , 4 , p. 374 , considered these substances to be non-conductors .
With regard to their behaviour as solvents , SKILLING ( Amer .
Ch. Jl 1901 , 26 , p. 383 ) found that at ordinary temperatures sulphuretted hydrogen dissolves potassium chloride freely ; but that the solution is a non-conductor of electricity .
HELBIG and FAUSTI ( Zeit .
fiir angewandte Chemie , ' 1904 , 17 ) state that stannic chloride is soluble in hydrogen chloride , but that this solution also is a non-conductor .
As it seemed highly improbable to us that sulphuretted hydrogen , which is analogous to water in so many ways , should be devoid of dissociating power , we decided to investigate its solvent action systematicaUy , and at the same time to examine.the hydrides of chlorine , bromine , iodine , and 390 .
) .
23.8.06 100 DR. B. D. STEEL , DR. D. McINTOSH AND DR. E. H. ARCHIBALD It has been found , as a result of our preliminary experiments , that water and the ordinary metallic salts which were tried are insoluble , or very sparingly soluble , in any of the solvents .
Hydrogen chloride and bromide are freely soluble in hydrogen sulphide , and hydrogen sulphide in hydrogen bromide .
The saIts of the organic ammonium bases are soluble in hydrogen chloride , bromide , iodide , and sulphide , and the resulting solutions conduct the current .
Certain ammonium salts also yield very feebly conducting soIutions .
Two metallic salts , namely , sodium acetate and potassium cyanide , were , at first , thought to be soluble , as their addition to the solvent greatly increased its conductivity .
This has since been found to be due to decomposition of these salts into acetic acid and hydrocyanic acid respectively .
Both of these acids are soluble in the foregoing solvents .
No substance has yet been found which will dissolve in phosphine and yield a conducting solution .
A few preliminary measurements of the conductivity were made , and in every case the moIecular conductivity diminished considerably with dilution , instead of increasing as it does in aqueous solutions .
The results of these measurements are given in Part II .
, which contains a detailed account of the measurements of solubility and of conductivity .
After we had ascertained that the hydrides of chlorine , bromine , iodine , and sulphur can act as conducting solvents , we proceeded to the measurement of the following physica ] constants of each of the pure substances:\mdash ; ( 1 ) The vapour-pressure curve ; ( 2 ) The density and its temperature coefficient ; ( 3 ) The surface energy and its temperature coefficient ; ( 4 ) The viscosity and its temperature coefficient .
The resuIts of these measurements are described in the following pages .
of Laquefed Gases .
Hydrogen chloride was prepared by the action of sulphuric acid on pure sodium chloride .
The gas was dried by passing it through two wash bottles containing sulphuric acid , and afterwards through a tube containing phosphoric anhydride .
It was then led into a receiver which was maintained at , by means of carbon dioxide and ether , under diminished pressure .
At this temperature the liquefied rapidly , forming a colourless mobile liquid .
This was -distilled before being used for the measurements .
ON THffl HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
The hydrogen bromide was prepared by the action of bromine on red phosphorus suspended in water .
Traces of bromine were removed by passing the gas through a thin paste of amorphous phosphorus and a saturated soIution of hydrogen bromide .
The gas was then dried by passing it over about 40 centims. of phosphoric anhydride , and , in order to remove impurities other than water-vapour , it was passed through two -tubes surrounded by solid carbon dioxide , in each of which a small quantity of liquefied gas soon coUected .
Tho gas bubbled through this liquid , which was thus submitted to a process of fractional distillation .
It was finally condensed in a vessel surrounded by alnixture of carbon dioxide and ether .
The hydrogen iodide was made by the action of iodine and water on amorphous phosphorus , in a similar manner to that employed for the preparation of hydrogen bromide , and similar means were used to purify it .
The liquid was invariably ooloured , and it could not be obtained quite colourless even by repeated distillation .
The hydrogen sulphide was prepared by the action of dilute sulphuric acid on ferrous sulphide .
The gas was washed by passing it through water , dried by passage over phosphoric anhydride , and condensed by means of carbon dioxide and ether .
It was purified by distillation .
Phosphuretted hydrogen was prepared by the action of a solution of potassium hydroxide on phosphonium iodide .
It was dried by means of phosphoric anhydride and condensed in a receiver which was immersed in liquid air .
The The constant-temperature bath consisted of ether which was contained in a vacuum vessel and cooled by liquid air .
The temperature was measured by a constant-volume hydrogen thermometer , similar to that described by TRAVERS , SENTER , and JAQUEROD Phil. .
The arrangement of the apparatus is shown in fig. 1 , in which represents the hydrogen thermometer , A the large vacuum vessel containing the ether , and a large vacuum flask containing liquid air .
The bulb , , of the thermometer was connected to the dead space of the manometer by a fine capillary tube .
A mercury reservoir was attached to the stop-cock by rubber tubing , and by raising or lowering this reservoir the reury in the dead space couId be adjusted to the level of the glass point The volume of the thermometer bulb and dead space was carefully determined by calibration with mercury .
The constants were volume of\mdash ; ( 1 ) Bulb and portion of stem within the liquid .
centims. at ; ( 2 ) Stem from to surface of ether .
centim. ; ( 3 ) Dead space and stem to mark .
centim. It has been assumed , in making our calculations , that the average temperature of the section ( 2 ) was midway between that of the bath and that of the atmosphere ; an error of a few degrees in the temperature of this section is without influence on the bath temperatures , which are given only to the nearest tenth of a degree .
DR. B. D. STEEL , DR. D. McINTOSH AND DR E. H. ARCHIBALD The vacuum vessel A was closed by a large indiarubber stopper , through which holes had been cut to allow the passage of the stem of the thermometer , the apparatus containing the liquefied gas , and the tubes , and The closed tube was about 7 millims. in diameter and long enough to reach nearly to the bottom of the vessel .
The tube was placed so that its open end came immediately under the tube The large vacuum flask was fitted with tubes as shown in the diagram , so that by blowing into liquid air could be forced into the tube In order to obtain any desired temperature between that of the room and the melting-point of ether the vessel A was filled with ether , and the tube connected to an air blast , by means of which the liquid was continuously and uniformly stirred , the air escaping through the tube , which was provided for the purpose .
After mercury had been taken out of the manometer through , the ether was cooled by blowing liquid air from into , where it rapidly boiled away .
When the temperature of the bath had been adjusted , it could be kept constant for as long as was desired by blowing liquid air in very small quantities into The deposition of dew on the walls of the vacuun ] vessel A was prevented by placing it inside a wider cylindrical glass vessel containing osphoric anhydrida ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
Vapour.pressure Curves .
In order to measure the vapour-pressure , a tube containing the liquid was immersed in the bath and simultaneous observations of the temperature and corresponding were taken .
This simple arrangement could not be used with hydrogen bromide and iodide on account of the action of these gases on the mercury of the manometer .
The errors due to this action were avoided by the use of a special form of apparatus which is shown in fig. 2 .
To use this apparatus , the tube was attached to the pump , and the bulb placed in the low-temperaturs bath , after which the whole apparatus was exhausted to a pressure of about 60 1mlhms .
and the stop-cock closed .
The outer portion of the apparatus was completely exhausted and the stop-cock closed .
A vessel tiontaining the liquefied gas was then attached to by rubber tubing , the point broken within the tube , and as soon as a sufficient quantity of liquid had distiUed into the bulb the apparatus was sealed off at Before making any measurements the stop-cock was opened for a few moments , and all traces of air were displaced from the tube by owing a small quantity ofliquid to evaporate into the pump .
The bath was cooled to the lowest temperature at which observations were to be taken , and the stopcock opened .
As the vapour-pressure of hydrogen bromide and of hydrogen iodide , even at the lowest temperatures employed , was greater than 60 miUims .
, a flow of gas ffom into owed , and continued until the pressure in the manometer became equal to the vapour-pressure of the liquid .
As soon as the pressure ceased to rise , the temperature and pressure were read and the stop-cock was immediately closed .
The temperature was then raised to the next of observation , and the stop-cock again opened , until the pressure in the manometer became constant , when readings were again taken and the stop-cock closed .
In this way a succession of readings was obtained without the hydrogen generated by the action of the on the mercury of the manometer its way into the bulb .
AB a precaution against diffision , the bulbs and the capillary tube were introduced , the stop-cock eing 'opened affer each observation and the contents of the bulbs and withdrawn through the pump .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. E. H. iIBALD in which represents the change of vapour-pressure with temperature , the pressure , the absolute temperature , the constant of the gas equation , and the Iatent heat of evaporation of one gram-molecule of the liquid .
The values of at a pressure of 760 millims. , as calculated from our vapour- pressure curves , are for hydrogen chloride , ergs , bromide , iodide , sulphide , , , phosphuretted hydrogen , We can find no account of any direct determination of for these substances .
The Measurement of Density .
The apparatus ( fig. 4 ) employed for these measurements consisted of a bulb with a graduated capillary stem , to which a two-way stop-cock was attached .
The bulb had a capacity of about .
centim. , and its volume and that of each division of the stem was accurately determined by calibration with mercury .
After the tube had been sealed to the pump , and the apparatus exhausted , it was immersed in the constant-temperature bath .
The tube was then connected to a vessel con- taining the liquefied gas , which was distilled into the bulb until both bulb and stem were completely ffiled .
The stop-cock was turned and the liquid allowed to evaporate into the pump until the meniscus had come to a definite position on the stem , when the stop-cock was turned so as to disconnect all the tubes .
To obtain the volumes occupied by a constant weight of liquid , it was only necessary to read the position of the meniscus at different temperatures .
The weight of liquid was obtained by attaching to the tube a weighed set of GEISSLER'S bulbs containing potassium hydrate solution .
On opening the stop-cock and raising the temperature of the bath the liquid evaporated and the gas was absorbed in the bulbs and weighed .
To prevent the potash solution sucking back , a little mercury was plaoed in the first bulb .
The small amount of gas remaining In the apparatus was finally pumped out through and measured .
ON TIE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
The numbers obtained by the foregoing method have been checked in the case of hydrogen chloride , hydrogen bromide , and hydrogen iodide by distilling each of them into a thick-walled bulb of known volume , and sealing it .
After the bulb containin the liquid had been weighed it was cooled , the stem broken , and the empty bulb and stem again weighed .
The densities so obtained agreed to the 3rd decimal place with those obtained at the same temperature by the first method .
In the case of phosphine the density was determined by the second method and the temperature coefficient by the first method .
The results of the measurements are given in Table III .
, in which represents the expelimental and the smoothed value of the density .
The density in each case is a linear function of the temperature and is given by the reIation where and represent the boiling-point and the temperature of observation respectively , both on the absolute scale ; and being the corresponding densities .
values of the coefficient for the different substances are contained in the foUowing table:\mdash ; TIffl HYDRIDES AS CONDUCTING KOPP has shown that the molecular volume of a liquid at its boiling-point is an additive property , being equal to the sum of the atomic volumes of the component elements .
Certain elements , however , suoh as oxygen , appear to possess two values for the atomic volume , depending on the nature of the linking of the oxygen to the other atoms in the molecule .
It has also been shown that , in the case of the elements , bromine , and sulphur , the atomic volumes calculated from the density of compounds containing them are the same as those obtained from the densities of the pure elements .
We have calculated the atomic volumes of the elements chlorine , bromine , iodine , sulphur , and phosphorus from the densities of their respective hydrides , in order to see ho the values so obtained agree with those given by KOPP and others .
The results of these calculations are given in Table , which contains , the densities of the compounds at their respective boiling-points , the molecular volume , and the atomic volume of the halogen elements .
The values of these are invariably higher than those of KOPP , which are given under , in the fifth column .
It is possible that this discrepancy is due to a variation in the atomic volume of hydrogen , which has accordingly been calculated from each compound by subtracting the figures in the fifth column from those in the third .
The values for , so obtained , are given in the last column , and are uniformly higher than , which is the number found from the study of organic compounds .
TABLE IV.\mdash ; Molecular Volumes at Boiling-point .
Voe Molecular Surface Energies .
The molecular surface energies were measured by RAnSAY and SHIELDS ' method , slightly modffied in order that the measurements might be made at low temperatures .
The apparatus ( fig. 5 ) consisted of a tube , 6 centims. long and centims. in diameter , which was provided with a small side tube , and joined to a two-way stop-cock by a long tube A glass scale , which had been very carefuly cahbrated , was securely fixed inside .
The capiliary was sealed to a long tube , in the manner described by DR. B. D. 8TEELE , DR. D. MoINToeH AND H. AROflIBALD RAXSAY and SHIEbDS .
The tube , which enclosed a pieoe of soffi iron , , was so as to slide easily and smoothly in .
It was held in poeition by the two gbae hooks and , which were placed so that when was resting on the bottom of the capillary was a few millimetres below the scale .
A mark had been previously etched on the capillary , and when making the measurements the position of the tube was adjusted by means of an electro-magnet , so that this mark always coincided with the meniscus inside the capillary .
radius of the capillary near the etched mark was determined by introducing a quantity of pure ether into the apparatus , and measuring the height of the column of liquid when the capillary was in different .sitions .
The radius could then be calculated from RAMSAY and SHIELDS ' values for the surface energy of ether .
The following values were found:\mdash ; ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
To carry out the experiments , the apparatus was placed in the constant-temperature bath and exhausted through the tube .
The stop-cock was then turned , and an exoess of the liquid to be measured was introduced through the tube by distillation .
The stop-cock was again turned , and all traces of air were displaced from the apparatus by some of the liquid to evaporate into the pump , after which the stop-cock was cloeed .
The bath was then maintained successively at different temperatures , and the mark on the capillary having first been brought into coincidence with the memiscus , the height of the column of liquid was accurately measured .
The tube was attached to a manometer , and measurements of the vapour-pressure of the liquid were made during the experiment .
This tube was removed during the measureme1lts of hydrogen bromide and hydrogen iodide .
The results of the measurements are given in Tables V. to IX .
, in the letters employed have the owing meaning:\mdash ; absolute temperature ; density of the liquid ; density of the vapour ; specific volume of the liquid ; V.P. vapour-pressure of the liquid ; height of the column of the liquid ; surface tension in dynes per centimetre ; 7 ( MV ) : molecular surface energy in ergs ; constant of gravity ; molecular weight of the liquid .
TABLE V.\mdash ; Hydrochloric Acid .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. H. ARCHIBALD The average value of this coefficient is , according to EoTvos , .
From the experiments of RAMSAY and SHIELDS it is , while BALY and DONNAN have found that the liquefied gases oxygen , nitrogen and carbon monoxide give values very near to 2 , and this number has also been found by us for the three substanoes hydrogen bromide , iodide and sulphide .
RAMSAY and SHIHiLDS have shown that for normal liquids the relation between molecular surface energy and temperature is given by the equation , in which represents the temperature measured ffom the critical point and is a small constant .
From this equation it follows that the surface energy disappears at a temperature degrees below the critical point , and therefore the curve for a normal liquid , if produced , should cut the temperature axis at this point .
This is the case for hydrogen bromide , iodide and sulphide , for which , as will be seen from Table XI .
, the value of and respectively .
The curves are shown in fig. 7 , in which the critioel temperature of each liquid is indicated by an arrow .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. E. H. ARCHIBALD It will be seen that the curve for hydrogen chloride cuts the temperature axis at , or above the critical point .
We have not been able to ffid any record of measurements of the criticaI temperature of phosphuretted hydrogen .
surement of Visoosity .
The viscosity apparatus ( fig. 9 ) was of the usual form , but with some slight additions , which were designed to prevent access of water vapour .
For this purpose the two ends of the apparatus were joined to the stop-cock , and the two tubes and , containing phosphoric anhydride , were attached .
After the apparatus had been placed in the bath , a definite quantity of the liquid was distilled into it through In order to make the measurements , the stopcock was closed and , by blowing into , the liquid was forced into until it reached a position about one centimetre above the mark The stop-cock was then opened and the time required for the liquid to fall from to measured with a stop-watch .
The apparatus was calibrated with distilled water at In the tables , refers to the density of the liquefied gas ; is the viscosity compared with that of ater at , and is the temperature coefficient of viscosity .
TABLE XII.\mdash ; Hydrogen Chloride .
Apparatus B. Time of flow for water seconds at DR. B. D. STEEL , DR. D. McINTOSH AND DR E. H. ARCHIBALD PART II .
The Conductivity and Weights of Dissolved Substances .
D. McINTOSH E. H. ARCHIBALD .
THE second part of this investigation deals with the solubilities of substances in the liquefied halogen hydrides and sulphuretted hydrogen , and with the conductivities of the resuIting solutions ; the molecuIar weights of a few substances , when dissolved in each of these solvents , have also been determined .
The Temperature Bath .
As liquid air was not available in sufficient quantities to make use of the bath described in Part I. , a mixture.of carbon dioxide and ether , which under atmospheric pressure gives a very constant temperature , was used for the measurements of soIutions in hydrogen bromide and sulphide .
The same mixture , under reduced pressure , was used for the measurement of solutions in hydrogen chloride .
By carefully regulating the pressure over the mixture the temperature was maintained at For the hydrogen iodide solutions a temperature of was obtained by slowly running cold ether into the vacuum vessel and syphoning off the warmer upper layer .
The Deterrnination of Solubilities .
The solubilities were measured by means of an apparatus ( fig. 1 ) which consisted of a test-tube , to the bottom of which a delivery tube was sealed .
The bottom of A was covered with a thick layer of asbestos which acted as a filter , and the whole was immersed in the ature bath .
The liquefied gas and the substance of which the solubility was under investigation were introduced into , where they were vigorously with a platinum rod .
A portion of the liquid was then blown through the delivery tube into a weighed and graduated test-tube .
The volume of liquid in the test-tube was observed , the liquid allowed to evaporate , and the residue weighed .
The lity was calculated ffi the data thus obtained .
ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVFNTS .
The results of the measurements may be summarised as foUows:\mdash ; ( 1 ) Inorganic substances insoluble , or soluble only in traces , in any of the solvents\mdash ; The chlorides , bromides , and iodides of the alkalis and alkaline earths : salts of nickel , iron , lead , and mercury ; stannous chloride , potassium permangamate , and potassium bichromate .
( 2 ) Inorganic substances soluble\mdash ; in some cases with decomposition:\mdash ; ( a ) In hydrogen chloride\mdash ; *Stannic chloride , phosphorus nide , phosphorus pentachloride , and phosphorus oxychloride ; In hydrogen bromide\mdash ; Phosphorus oxychloride , bromide , and sulphuretted hydrogen ; ( c ) In hydrogen iodide\mdash ; Iodine and phosphorus oxychloride ; ( d ) In sulphuretted hydrogen\mdash ; Sulphur , phosphorus oxychloride , hydrogen bromide , and hydrogen chloride .
( 3 ) Inorganic substances soluble with decomposition in hydrogen chloride , potassium cyanide , ammonium sulpho-cyanate , sodium acetate .
( 4 ) Organic substances .
In addition to the organic ammonium bases , which , as stated in Part I. , dissolve somewhat freely in all the solvents , we have found that a very large number of organic compounds are soluble , as , for example , the aldehydes , ketones , alcohols , ethereal salts of fatty and of aromatic acids , cyanides , and sulphocyanates , hydrocarbons , and nitro-compounds .
Hydrogen sulphide is an excellent solvent for such bodies , but the solutions , as a rule , are non-conductors .
The solutions in the halogen hydrides , on the other hand , usually conduct the current .
Measurement of Conductivity .
Although the investigation of the solubility of inorganic salts failed to indicate that these were soluble in more than traces in any of the solvents , we have tested the conductivity of the different solvents after the addition of certain inorganic substances .
We find that an increase of conductivity was produced by adding the following substances to : \mdash ; * HELBIG and FAUSTI ( Atti R. Accad .
Lincei , ' 1904 ( V ) , 13 , p. 30 ) found that stannic chloride was soluble in hydrogen chloride .
Wo that by an oversight contradicted this statement in our Note , and we wish now to make the necessary correction .
VOL. CCV .
122 DR. B. D. STEEL , DR. D. McINTOSH AND DR. H. ARCHIBALD ( a ) Hydrogen chloride\mdash ; Bromine , potassium iodide , thionyl chloride , sulphuryl chloride , and uranium nitrate ( very slight increase ) , phosphorus pentachloride , pentabromide , and oxychloride considerable increase ) ; ( b ) Hydrogen bromide\mdash ; Phosphorus oxychloride ( considerable increase ) ; ( c ) Hydrogen iodide\mdash ; Iodine , sulphuric acid , carbon disulphide , and phosphorus oxychloride ( slight increase ) ; ( d ) Sulphuretted hydrogen\mdash ; Phosphorus pentachloride , and sulphuryl chloride ( slight increase ) .
The following substances did not cause an increase in the conductivity of either of the solvents:\mdash ; Sodium , sodium sulphide , sodium biborate , sodium acid pbosphate , sodium nitrate , sodium sulphide , sodium thiosulphate , sodium arsenate , chromic acid , the following salts of potassium : the mtrate , hydroxide , chromate , sulphide , acid sulphate , ferrocyanide , ferricyanide ; ammonium fluoride and carbonate ; rubidium and caesium chlorides ; Magnesium sulphate , calcium fluoride , strontium chloride , barium chloride , oxide , nitrate and chromate ; copper sulphate , mercuric chloride , zinc sulphate , boron tl.ichloride , aluminium chloride , and sulphate ; carbon dioxide , stannous chloride , lead peroxide , nitrate , and cyanide ; phosphorus tribromide , bismuth nitrate , tartar emetic , manganese chloride , ferric chloride , ferrous sulphate , nickel sulphate , and cadmium sulphate .
In addition to the organic ammonium bases , we have , in conjunction with Dr. J. W. WALKER , *examined the conductivity of solutions of about 80 organic substances in each of the foregoing solvents .
The only substances which form conducting solutions in are the ammonium bases and a few alkaloids such as nicotine and pyridine .
On the other hand , many ethers , ketones , esters , nitrites , and , generally , substances containing oxygen or nitrogen , form conducting solutions in hydrogen chloride , bromide , and iodide .
The hydrocarbons , although in some cases soluble in all proportions , conduct .
We have noticed that the solution of those substances which conduct is accompanied by a considerable evoIution of heat , while little or no heat is evolved in the case of other substances .
This indicates chemical interaction between the conducting solute and the solvent , and many of the resulting compoun have been isolated and analysed .
It has been * Journal of the Chemical Society , ' 1904 , 85 , p. 1098 .
ARCIIIBALD and , ' J. C. 1904 , , p. 919 .
ON THE HALOGEN AS CONDUCTING SOLVENTS .
uffi , for example , that ether enters into combination with the three halogen acids , compounds which have the following : , HBr , M.P. , HI , M.P. We have explained the formation of these and similar compounds by assumin the existence of tetrad oxygen at these low temperatnres .
The compound of ether and hydrogen bromide would thus have the formula , and might be expected to undergo electrolytic dissociation .
Quantitative Measurements of Conductivity .
The pure solvents are extremely poor conductors of electricity , their specific conductivity being as follows : Hydrogen chloride about bromide , , iodide , , sulphide , , that of the purest water being The resistances are thus much greater than that of an ordinary sample of distilled water .
The majority of the measurements of conductivity were made in an apparatus ( fig. 2 ) consisting of a graduated test-tube with fixed electrodes .
DR. B. D. STEEL , DR. D. McINTOSH AND DR E. H. ARCHIBALD cooled platinum spoon .
The mixture was then stirred until the conductivity remained constant , after which the volume of the solution was observed .
A ffirther weighed quantity of the substance was then introduced and dissolved , and the conductivity again measured .
The same series of operations was frequently repeated until a sufficient number of measurements had been made .
Other measurements were made in the apparatus shown in fig. 3 .
This consisted of a graduated test-tube , provided with fixed electrodes , and with a delivery tube attached .
A saturated solution of the substance was made in the apparatus shown in fig. 1 , portion removed for analysis , and a sufficient amount put into the conductivity vessel , where its conductivity was measured and the volume noted .
More of the solvent was then added , and the liqui\amp ; were well mixed by blowing air through the delivery tube B. The volume was again read and the conductivity measure\amp ; This succession of operations was repeated until the vessel became full of liquid , after which a measured volume of the solution was removed , and the operations were continued until a sufficient number of measurements had been obtained .
In all the measurements the electrodes were sufficiently immersed to give the maximum conductivity of the apparatus .
Our results are given in the following tables , which also contain the temperature coefficient of conductivity for those solutions which are marked with an asterisk .
The dilutions , which are given under V , represent the number of litres of soIution which contain 1 gram-molecule of solute , and the molecular conductivities in reciprocal ohms are given under .
The numbers are thus expressed in the same units as the molecular conductivity of aqueous solutions as given by KOHLRAUSCH and HOLBORN ( ' Leitvermogen der Elektrolyte DR. B. D. STEEL , DR. D. McINTOSH AND DR. H. ARCmBALD As a general rule , the same substances conduct better when dissolved in hydrogen chloride than in the other solvents .
Next , in the order given , come solutions in hydrogen bromide , iodide , and sulphide .
An exception to this rule is found in the case of ether , which conducts best in hydrogen chloride , and worst in hydrogen bromide .
The conductivity temperature coefficients do not appear to have anything in common with the viscosity temperature coefficient , so that the ions cannot be looked upon as being surrounded with an atmosphere of the solvent ( KOHLRAUSCH , ' Roy .
Soc. Proc 71 , 338 , 1903 ) .
The coefficients are for the most part positive , the conductivity increasing with rise of temperature , an exception occurring in the case of hydrocyanic acid dissolved in hydrochloric acid .
The of Molecular Weights .
The molecular weights were determined by measuring the rise in boiling-point which was brought about by the addition of known quantities of the dissolved substance .
Considerable experimental difficulty has been experienced during the progress of the work , which has also proved expensive on account of the very large quantities of carbon dioxide which were required , and consequently only a few determinations have been made .
The accurate measurement of small differences of temperature at low temperatures has been successfully accomplished by the use of a differential method , in which two platinum resistance thermometers were employed to measure the temperatures , one of the thermometers being immersed in the pure boiling solvent and the other in the boiling solution .
* The thermometers were each made from about 2 metres of 6 mil wire , and had exactly the same resistance , which was of such a magnitude that a difference of 1o in the temperature of the two coils produced a displacement of minims .
the balance point on the bridge .
The thermometers were supplied with compensation leads in the usual way .
The bridge was of the Carey-Foster type .
With the galvanometer used a difference in temperature of could be detected with certainty .
The apparatus was tested by immersing the two coils in ( a ) boiling water , ( b ) melting ice , ( c ) boiling hydrogen sulphide , boiling hydrogen bromide , and ( e ) boiling hydrogen chloride ; the same balance point was obtained in each case .
The measurements were made in the two pieces of apparatus shown in fig. 8 , bridge and resistance thermometers were lent to us by Dr. H. L. BARNES , through whose and assistance many difficulties have beon avoided .
We take this opportunity of expressing our thanks to him for his kindness and help .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. E. H. ARCfflBALD vessel it weighed .
The condensers were then filed with a mixture of solid carbon dioxide and ether , maintained at atmospheric pressure for the with hydrogen bromide and sulphide , and at reduced pressure for the with hydrogen chloride .
After the liquid in the vessels had commenced to boil , the vessels were wrapped in natural wool , and the balance point on the bridge was determined .
Weighed quantities of the substances whose molecular weight was to be determined were then successively introduced , and the displacement of the balance point was determined after each addition .
From these displacements the corresponding rise in boiling-point was calculated .
The loss by evaporation , due to the high vapour pressure of the solvents at the temperature of the condenser , was corrected for by means of a blank experiment .
Evaporation also occurred when the substances were introduced into the apparatus , but as this evaporation was proportional to the amount of substance added , a correction was easily applied .
As a check on these corrections , the boiling-point apparatus was removed after each two or three determinations , and when liquid air was available , the apparatus was cooled and weighed .
When liquid air could not be obtained , the tube was cooled to a definite temperature , the volume of solution measured , and the amount of solvent calculated on the assumption that no volume change occurred on mixing .
From the data thus obtained the molecular weight constant was calculated by means of the formula in which the molecular weight of the dissolved substanoe is expressed in terms its weight in grammes dissolved in -gramme of the solvent , and of the corresponding rise in boiling-point being a constant in the case of a solute which is neither associated nor dissociated .
The values of , and for the various substances investigated are given in the first four columns of Tables , and VII .
The following example will show the method of making the calculations:\mdash ; Toluene in Hydrochloric Acid .
Apparent volume of liquid toluene .
cub. centims. beads . . . . .
, , , Real volume ofliquid . . . . . .
, , Volume of toluene . . . . . .
, , hydrochloric acid Weight ( Part I. ) . . . .
Rise of boiling-point Constant . . . . .
ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
VAN'T HOFF has shown that the molecular rise of boiling-point can be calculated by means of the formula where is the absolute temperature and is the latent heat of evaporation of 1 gramme of the solvent .
We have calculated the values for the molecular rise of boiling-point of the various lvents by means of the latent heats which are given in Part I. The moleqular weight of the dissolved substaIlce has been calculated from the values so obtained for the molecular rise .
The molecular weights are given under in the sixth column of the tables , and in the fflh column the concentrations of the solutions are expressed in grammes of solute dissolved in 100 grammes of solvent .
TABLE V.\mdash ; Molecular Weights in Hydrogen Chloride , { \eth ; ]y ; ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
l37 These results may be briefly summarised as follows:\mdash ; Toluene , which is a non-conductor in each of the solvents , has an average molecular weight of about 30 in hydrogen chloride , 78 in hydrogen bromide , and about 90 in sulphuretted hydrogen , and therefore appears to be dissociated when dissolved in hydrogen chloride and hydrogen bromide and to a greater extent in the former solvent .
KAHLENBERG Jour .
Phys. Chem has noticed a similar anomaly in the case of a solution of diphenylamine in methyl cyanide .
Ether in hydrogen chloride and hydrogen bromide , and acetone in hydrogen bromide have molecular weights which indicate association in the more dilute solvents and dissociation in the more concentrated .
Triethylamine hydrochloride appears to be dissociated when dissolved in sulphuretted hydrogen , the dissociation being greater in the more concentrated solutions .
VOL COW .
PART The Transport Numbers of Certain Dissolved .
By B. D. THE strikingly abnormal variation 01 molecular conductivity with dilution that we have found to occur in solutions in the liquefied halogen hydrides finds a possible explanation in the assumption that it is the solvent and not the solute which is ionised .
As the transport number of the dissolved substance might be expected to yield information not only as to the correctness of this assumption , but ako as to the constitution of the electrolyte , the transport numbers of a few substances have been measured , and the resuIts are given in the following pages .
The only measurements of the migration ratio which have hitherto been made in solvents other than water are those of a few salts in methyl and ethyl alcohol , and silver nitrate in pyridine and in acetonitrile .
Direct measurements of the velocities of certain ions in liquefied ammonia have recently been made by FRANKLIN and CADY ( ' Journal of Amer .
Chem. Soc 1904 , vol. 26 , p. 499 ) , who used a modification of method ( ' Phil. Trans , 1902 , vol. 192 , p. 331 ) .
Method of Measuremoent .
It has been shown by the author STEEL , Phil. Trans that the direct method of measurement gives trustworthy results only when the salt under examination is of the simplest type .
Now HITTORF has shown that in alcoholic solution cadmium iodide and certain other salts are dissociated into ions which are much more complicated than those occurring in aqueous solutions of the same concentration .
The only substances which we have found to be capable of forming conducting solutions in any of the solvents which we have been investigating are certain organic compounds , and although the nature of the ions into which these dissociate is entirely unknown , it is probable that the ionisation is even more complicated than that of cadmium iodide dissolved in alcoholic solution .
From these considerations it was decided to use method , notwithstanding the fact that it is much more tedious and presents greater experimental difficulties than the alternative method of direct measurement .
HITTORF'S method consists in the analysis , afber electrolysis , of the solution which surrounds one of the electrodes .
The original concentration being known , the actual .
ON THE HALOGPN AS CO. NDUCTING SOLVENTS .
annunP of substance which has been carried to the electrode by the current can then be lated .
The current is usually measured by a silver voltameter placed in the it me ruit as the electrolytic cell .
The calculation , neglecting certain small oonections , is as follows:\mdash ; If grammes of the substance , whose equivalent weight is , be transported by : the ourrent which deposits grammes of silver on the cathode of the voltameter , then the transport number of the cation is given by represents the ffiction of the total current which is carried by the cation , on the assumption that one unit d1arge of electricity is associated with one equivalent of the dissolved substance .
It is probable that this condition is ffilffiled only in solutions of salts of the simplest type ( STEEL , loc. cit Preparation of Solutions .
The most convenient refrigerant which was available was a mixture of carbon dioxide and acetone , and as at the temperature of this mixture hydrogen chloride is a gas and hydrogen iodide a solid , the choice of solvent was limited to hydrogen sulphide and hydrogen bromide .
Solutions in the former solvent are very much more difficult to analyse than those in the latter , and accordingly hydrogen bromide only has been used as solvent during the investigation .
The hydrogen bromide was prepared and purified by the method described in Part L In order to make the solutions for electrolysis , the gas was condensed in a graduated vessel in which a sufficient quantity of the substance under examination had .
even placed , the condensation being stopped as soon as the desired volume of solution had accumulate\amp ; A quantity was usually made sufficient for two experiments , and by placing the receiver in a good silvered vacuum vessel with a stiff paste of the carbon dioxide and ether , the solution could be kept for a period of twenty hours without renewal of the reffigerant .
The apparatus in which the electrolysis was carried out was immersed in a bath of solid dioxide and acetone contained in a large cylindrical silvered vacuum vessel .
This mixture can be maintained at a practically constant temperature by blowing a steady stream of air through it ; the temperature , moreover , may be varied within certain limits by altering the rapidity of the air current .
The Validity of FARADAY'S Law .
The measurement of the transport number depends on FARADAY'S law , and although is known to hold rigidly for aqueous solutions , there is no evidence as to its validity for solutions such as those under investigation .
Experiments were therefore undertaken with the object of testing the law .
This head by oomparing the weight of silver deposited in a voltameter DR. B. D. STEEL , DR. D. MoINTOSH AND DR. E. H. ARCBIBALD with the volume of the hydrogen evolved at the cathode during the electro of solutions in hydrogen bromide .
The apparatus used ( fig. 1 ) consisted of a tube with a coiled platinum wire sealed through the bottom and projecting about an inch into the tube , the stem of the projecting part bning covered with blue enamel glass .
Electrical contact with this electrode , which was used as cathode , was made .
by means of mercury contained in the tube .
The .
inner cell was provided with a long capillary , which passed through the rubber cork , and served for delivering the hydrogen into a measuring tube .
The anode consisted of a ring of platinum wire , which was attached to the tube In carrying out the experiments the tube A was first immersed in a bath of carbon dioxide and ether ; the solution to be electrolysed was then run in , and , which had been previously cooled , placed in position .
A silver voltameter was then placed in the circuit , and current from a battery of about 60 volts was passed through the cel The hydrogen evolved in A escaped through and was collected and measured .
Two experiments of this nature were carried out , the details of which are as follows:\mdash ; Experiment 1 .
Solution of diethylamine in hydrogen bromide .
E.M.F. volts .
Current am.pire .
Silver deposited in voltameter gramme , equivalent to cub. centims. hydrogen at and 760 millims. Hydrogen evolved cub. centims. at and 760 milhms .
Experiment 2 .
Solution of acetophenone in hydrogen bromide .
E.M.F. volts .
Current ampere .
Silver deposited in voltameter gramme .
Hydrogen equivalent cub. centims. at and 760 millims. Hydrogen evolved cub. centims. at and 760 millims. These experiments were considered sufficient to show that FARADAY'B law is valid for solutions of organic substances in hydrogen bromide .
The Method of As the total inorease in concentration which had to be measured amounted to only a few centigrammes , it was necessary to carry out the analysis with a high of accuracy .
This was found to be extremely difficult on account of the .
high vapour-pressure of hydrogen bromide even at temperatures near ik ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
only after months of failure that an apparatus was designed , by means of sufficiently accurate analyses were obtained .
apparatus consisted of two parts , the transferrer and the absorber .
transferrer consisted of a wide -shaped tube , with a tube passing through three of its branches , the two tubes being sealed together at and , as shown in the figure .
By filling the space between the tubes with a mixture of carbon dioxide and ether , the capillary between and could be oobledto .
In fig. 5 the transferrer is shown when placed in the electrolytic cell .
The absorber ( fig. 2 ) consisted of a stoppered tube , connected by with the bubbler ; this bubbler was so constructed that it was impossible for water to be either ejected from the apparatus or sucked back into A. A smaller vessel of the same type was contained within the apparatus .
The calcium chloride tube , which was provided with a stop-cock , was attached to by a ground joint .
The method of using the apparatus was as foUows : quantity of moist garnets vrere first placed in the tube , and a quantity of glass beads in the section of the .
The requisite amount of distilled water was then placed in the bubblers and , and the tube and stop-cock were replaced .
The apparatus was then , a glass counterpoise of approximately the same size , shape , and weight being The tube A was next immersed in a mixture of solid carbon dioxide and ether , and affir the cap , fig. 3 , had been placed in position by means of the rubber cork DR B. D. STEEL , DR. D. McINTOSH AND DR. .
ABOHIBALI ) the limb of a transferrer was passed through , and the whole apparatus made tight by means of a piece of rubber tubing , The object of the cap was to prevent the limb of the transferrer from coming into contact with the rubber grease , with which the stoppers were lubricated , and it was constructed so that when in position its narrow portion was exactly in the axis of the tube The transferrer was then packed with the carbon dioxide and ether mixture , and the hydrogen bromide blown into A by means of a small indiarubber bellows .
The rubber tube was then cut away , after which the transferrer and then the cap were removed , and the stop-cock was re-inserted .
The tube A was finally removed from the cold bath and the hydrogen bromide allowed to boil off .
This always took place steadily , provided the vessel contained moist beads or garnets ; in the absence of these , or if they were dry , the violent bumping which resulted was liable to blow out one or other of the stoppers .
The hydrogen bromide as it boiled off passed through , and was almost completely absorbed at the surface of the water in the outer portion of B ; a.small quantity of gas bubbled through the hole and was absorbed inside ; very occasionally a few bubbles passed through , where any traces of acid which might have passed through the larger portion of the apparatus were absorbed .
After the ])ydrogen bromide had evaporated the stop-cock was closed , apparatus immersed in distilled water , carefully wiped and again weighed , the neCessary correction being made for the increase of of the iquid contained in it .
The increase in weight gave the amount of solution that had been use The contents of the absorber were next washed into a large beaker and the hydrogen nide determined by titration with a twice normal alkali solution , which .
had been carefully standardised and was ffee from carbonate .
The difference between the amount of acid found in t , his manner and amount of solution actually weighed gave the weight of the dissolved substance .
The alkali was contained in a weighing burette of the pattern ehowIi in In order to deliver \amp ; om this burette , the cap removed and the stop-cock opened , when by blowing through the ei\amp ; liquid was forced through the tube .
The burette was weighed to glass counterpoise of approximately the same volume being use\amp ; Phenolphthalein was used as indicator , and an excess of one or two dPop added , the exact amount of excess being determined by titration with a normal solution of hydrobromic acid .
UN THE HALOGEN AS CONDUCTING SOLVENTS .
In order that the determination of acid by titration should be strictly comparable with the weighing of the solution , the alkali was standardised .
by direct comparison vrith about 60 grammes of pure hydrogen bromide which was weighed in the absorber .
Duplicate standardisations of the same alkali solution gave the following figures for the amount of hydrobromic acid equivalent to 1 gramme of alkali solution : 1 1 gramme solution HBr , : ( 2 ) 1 , , , , HBr .
The Electrolysis .
In designing an electrolytic cell it was necessary to consider the changes of density which were brought about during electrolysis , and to construct an apparatus in which the lighter solution would be formed at the top and the heavier at the bottom .
Preliminary experiments were undertaken to ascertain the influence of an increase in centration of the dissolved substance on the density of the solution , and it was foun ' that in all the cases examined the less concentrated solution was the heavier .
The experiments which had been conducted to test FARADAY'S law having shown that the bromine is carried to the anode , the apparatus was designed to enable the cathode solution to be analysed .
The anode solution was neglected , as its analysis was complicated by the presence of bromine .
The apparatus which has been employed is shown diagrammatically in fig. 5 .
It consisted of a -tube , both arms of which were provided with side tubes .
The anode consisted of a platinum wire , which was sealed through the bottom of the side tube , the other end of the wire projecting into the glass tube , by means of which connection with the battery could be made .
A small side tube was also attached to the same arm of the -tube , and there was a constriction at into which the hollow stopper was ground to fit tightly .
The stopper was sealed to a branch of the transferror , and a hole was bored in its shoulder to allow the free passage of liquid through the transferrer .
It was found in the preliminary experiments that a considerable amount of mixing was occasioned by the escape of hydrogen at the cathode , and in order to reduce this to a minimum , the side tube was attached to the apparatus .
In the centre of a narrow tube was fastened , inside which the cathode was placed , so that the esoaping bubbles of hydrogen were confined to this tube and very little mixing took place outside .
P. A small hole had been blown in the wall of the tube so that the pressure should be equal at all parts of the surface of the liquid .
The tube was made long enough to be held in the clamp outside the vacuum vessel in which the apparatus was placed and its end was closed by a stopper , through which passed the platinum wire which was used as cathode .
One end of the transferrer was made lorg enough snd bent so as to reach to the bottom of the -tube , and both the DR. B. D. STEEL , DR. D. McINTOSH AND DR. H. ARCHIBALD transferrers were attached to the open ends rubber tubing , and In carrying out an experiment the method of procedure.was as \mdash ; The apparatus was placed in the vacuum vessel and held rigidly by a clamp which grasped the tube P. The open ends of the transferrers and were then closed by rubber tubing and glass plugs , and a tube filled with phosphoric oxide was attached to .
The stop-cock with the cathode attached was then removed and a third transferrer ( not shown in the figure ) inserted in the tube P. The aoetone was then placed in the vacuum vessel and solid carbon dioxide added till the temperature had fallen to the desired point .
The outside arm of the hird transferrer was provided with a filter of glass wool and reached to the bottom of a vessel containing 10 centims. more of the solution than was necessary for the electrolysis .
Afler the whole of this liquid had been forced into the apparatus by blowing air from a rubber bellows into the vessel containing the liquid , the transferrer was removed and the stopper replaced .
Before commencing the experiment three absorbers had been filled distilled water and weighed .
The transferrer was next attached to one of these in the : ON THffl HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
manner previously described and 10 cub. centims. of the solution were taken out and analysed .
This analysis gave the concentration of the solution before electroly.sis .
A silver voltameter and a milliampermeter having been placed in the circuit , the electrolysis was started by connecting the electrodes to the terminals of a battery , .
the voltage of which could be varied within wide limits , and the voltage was adjusted .
so as to give a current of not than 12 or 14 milliamperes .
If a larger current .
than this was employed , the heating effect was found to cause mixing of the liquid by convection .
After the electrolysis had been continued for about two hours it was stopped by removing the cathode from the voltameter , and the liquid in the cell was separated into two portions by inserting the stop-cock into its socket .
An absorber .
was then attached to the transferrer and the column of liquid contained within the dotted lines was blown into the absorber , which was then removed .
The third absorber was next attached to the transferrer and the solution contained in the apparatus forced into it by blowing through , the tube acting as a syphon to remove the liquid contained in , and finally the third absorber was removed and eighed .
Three solutions were thus obtained for analysis , namely : ( 1 ) the 10 cub. centims. which had been removed before electrolysis and gave the original concentration ; ( 2 ) the small quantity taken out in the second absorber ( which should be of the same concentration as ( 1 ) ) ; and ( 3 ) the solution surrounding the cathode , which gave the change of concentration brought about by the electrolysis .
In that an experiment should be successful it was necessary that the foregoing procedure should be strictly folowed .
Identical values for the original concentration and that of the middle portion have never been obtained unless the whole solution was first placed in the apparatus and all three portions were taken ffom it .
At least half-a-dozen other methods have been tried without success .
Fortunately it is easy at the close of an experiment to see if any mixing has taken place , from the fact that bromine is liberated at the anode , where it forms a deep red solution in the hydrogen bromide .
If the experiment has been successful , this solution remains as a very clearly defined layer surrounding the anode , and the coloration does not extend more than about centims. up the tube .
On the other hand , if mixing has taken place , as may happen either if the current is too large or if the temperature of the bath is allowed to vary , the bromine is distributed throughout the solution and no clearly defined layer is seen at the anode .
Experimental Results .
The results of the experiments are contained in the folowing table , in which the concentration of the various solutions is expressed under , which gives the number .
of gramme equivalents of dissolved substance per litre of solution , the percentage also being given in the 3rd column .
The 4th column contains the weight of silver on the cathode of the silver voltameter , the 5th column gives the weight of subetanoe transported , the 6th column the cation transport numbers .
VOL. COV .
ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
With the exception of Nos. 1 and 4 , which unfortunately were , all the experiments which have been made are given in the tables .
The transport number of each substance has been measured at two concentrations , the more concentrated solution usually containing about twice as much solute as the other .
It will be seen that the cation transport number is always increased by increase of concentration and the amount of disagreement between parallel experiments , although in some cases approaching 10 per cent. , is never sufficient to leave any doubt as to the influence of change of concentration .
This change from analogy with aqueous solutions indicates an increase in the complexity of the cation as the solution becomes stronger , but the measurements of conductivity and of the molecular weight , which are given in Part II .
, do not appear to confirm this conclusion .
The significance of the change will be discussed in Part IV .
A special significance is to be attached to the results of Series 2 , for the following reasons : the foregoing description of the method of analysis it has been explained that after the solution was transferred to the absorber the hydrogen bromide was allowed to evaporate and to become dissolved in the water .
During the evaporation the temperature of the liquid in A gradually rose until finally it reached that of the atmosphere , when the liquid which remained was a saturated solution of hydrogen bromide in ether .
At this stage little or no decomposition occurred in a solution which had not been electrolysed , or in a dilute solution which had , but in the case of the cathode portion of a concentrated electrolysed solution the decomposition which occurred was sufficient to give results so discordant as those tabulated in Series 2 .
This behaviour seems to indicate the fonnation at the cathode of some extremely unstable substance during the electrolysis of these solutions , and it is possible that a compound is formed by the union of two or more discharged cations by a reaction similar to that by which persulphuric acid results from the electrolysis of sulphuric acid .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. E. H. ARCHIBALD PART IV .
The Abnormal Variation of Conductivity , etc. B. D. STEEL , D. McINTOSH and E. H. ARCmBALD .
IN discussing the nature of those inorganic liquids which are able to act as " " ionising\ldquo ; solvents , WALDEN ( ' Zeit .
fur anorg .
Chemie , ' 1900 , 25 , p. 209 ) states that " " measurable dissociation ( ionisation ) occurs only in combinations of the elements of the 5th and 6th groups of the periodic table and in compounds of these elements with hydrogen and the halogens We have shown in Part I. of this investigation that the hydrides of the halogen elements and of sulphur belong to the class of " " ionising\ldquo ; solvents , so that this class consists of compounds of the elements of the 5th , 6th , and 7th groups amongst themselves and with hydrogen .
Attempts have been frequently made to arrive at some generalisation connecting the so-called ' ' ionising\ldquo ; power with certain physical constants of the pure solvents .
Thus , according to NERNST and THOMSON , a close relationship exists between the dissociating power and the dielectric constant .
These investigators were led to look for this relation by the consideration that the force which two electricaly charged bodies attract or repel each other depends on the magnitude of the dielectric constant of the separating medium , and as the ions are to be regarded as electrically charged bodies , the force attracting two unlike ions will be more weakened , and dissociation aided , in a solvent of high than in one of low dielectric constant .
This expectation is only partially realised in the parallelism which exists for a great number of solvents between the two properties in question ; thus liquefied ammonia which possesses a low dielectric constant is a better dissociating solvent for some substances than water which has a high dielectric constant ; moreover , the majority of electrolytes are far more dissociated in water than in hydrocyanic acid or in hydrogen peroxide , although the dielectric constant of water is less than that of either of these liquids .
No measurements of the dielectric constant of the halogen hydrides or of sulphuretted hydrogen have yet been made .
DUTOIT and ASTON have attempted to show that ionic dissociation occurs only in solvents in which the molecules are associated , but , although numerous instances occur in which this parallelism obtains , it is by no means a general rule .
Thus , although both ammonia and sulphur dioxide are unassociated liquids , both are able to form conducting solutions ; and although the vdrides of bromine , iodine , and sulphur are unassociated , and hydrogen chloride is .
ON THE HALOGEN HYDRIDES As CONDUCTING SOLVENTS .
149 ; all four compounds are equally able to act as conducting solvents , whilst ! .
this property is not possessed by hydrogen phosphide , which is associated .
BRUHL ( ' Zeit .
Phys. Chem 1898 , 27 , p. 319 ) has pointed out that unsaturated mpounds , as a rule , are good conducting solvents .
A consideration of the nonoonducting unsaturated solvents phosphorus hydride and trichloride , and of the conducting saturated solvent phosphorus oxychloride , is sufficient to show that this is not a general rule .
The heat of vaporisation ( OBACH , 'Phil .
Mag 1891 , ( 5 ) , 32 , p. 113 ) is a fourth property which has been suggested as being intimately connected with the dissociating power of the solvent .
In this case , as in that of the others considered , the connection is very obscure and many exceptions occur .
The temperature coefficients of conductivity and of viscosity are approximately equal in the case of aqueous solutions .
This is not so in solutions in the solvents examined by us , although in these , also , a rise of temperature conditions an increase of conductivity and a decrease of viscosity .
It is interesting to note that the increase of conductivity is , in nearly all cases , than the decrease of viscosity .
The foregoing summary shows that failure has attended every attempt which has been made to express the power of forming conducting solutions as a function of the solvent only .
As a matter of fact , every solvent exhibits a very marked selective action as regards the nature of the conducting solute .
Thus water dissolves the majority of salts to form solutions which conduct the current ; organic bodies also , other than salts , are in many cases soluble , but the solutions are not conductors .
Hydrocyanic acid behaves similarly to water , but only a few salts are appreciably soluble in this solvent .
Ammonia , sulphur dioxide , and some other solvents form conducting solutions , not only with many salts , but also with a few organic substances not usually classed as electrolytes .
The halogen hydrides , on the other hand , form conducting solutions with non-saline organic substances , as well as with salts of the ammonium bases , but such solutions are not formed with metallic salts .
, is evident , therefore , that the ability to form conducting solution is a function of both the solute and the solvent , and this has been recognised in the various attempts that have been made to connect the ionising power of a solvent with its tendency to form compounds with the solute .
Indeed CADY ( ' Jour .
Phys. Chem 1897 , 1 , p. 707 ) was led to investigate the conductivity of solutions of substances in ammonia from the analogy between the water and the ammonia compounds of copper sulphate .
KAHLENBgRG and SCHLUNDT ( ' Jour .
Phys. Chem 1902 , 6 , p. 447 ) express the opinion that conductivity is due to mutual action between the solute and the solvent ; and an attempt to obtain experimental evidence in support of this view has been made by PATTEN ( ' Jour .
Phys. Chem 1902 , 6 , p. 554 ) .
DR. B. D. STEEL , DR. D. McINTOSH AND DR E. H. ARCHIBALD For many solvents the substances which dissolve to form conducting solutions be broadly designated as those which enter into combination with the solvent .
Thus the metallic salts as a class are characterised by their tendency to form compounds with water , while non-saline organic bodies as a class are not able to form such compounds .
* Many , but not all , salts which form ammonia compounds dissolve in ammonia to form conducting solutions .
Compounds of the solute with the solvent are also clearly indicated in the se- of many conducting solutions in sulphur dioxide , and WALDEN and CKNTNERSZWER ( ' Zeit .
Phys. Chem 1903 , 42 , p. 432 ) have isolated and investigated two such compounds containing potassium iodide and sulphur dioxide .
In the halogen hydrides we find that the only substances which conduct are the amines , alcohols , ethers , ketones , \amp ; c. , all of which are able to enter into combination with the solvents .
Many similar cases have been observed amongst organic solvents , and as an example of these reference may be made to solutions in amylamine ( KAHLENBERG and RUHOFF , ' Journ. Phys. Ch 1903 , 7 , p. 254 ) .
The study of the behaviour of aqueous solutions has led to ARRHENIUS ' theory of ionic dissociation and to VAN HOFF'S theory of solutions ; and numerous investigations have been undertaken with the object of testing these theories , when applied to solutions in non-aqueous solvents .
As a result it has been found that , as required by the theory , most substances , when dissolved in ammonia , sulphur dioxide , hydrocyanic acid , and some other solvents , show an increase of the molecular conductivity , , with dilution , but that the opposite change occurs in solutions of a few substances in the same solventa This difference in behaviour cannot therefore be conditioned by the nature of the solvent only , although if we consider the inorganic hydrides as solvents , we find that varies normally , that is to say , increases with dilution , in solutions in water and ammonia , hydrides , namely , of elements in the first series of the periodic table , whereas decreases with dilution in solutions in the remaining hydrides , the variation therefore being abnormal .
The results of the molecular weight determinations in non-aqueous solvents are , as a general rule , not concordant with the conductivity results , many conducting solutions being known in which , contrary to expectations , the dissolved substance appears to be associated .
* The view that compounds of the solute and the solvent exist also in solutions appears to be steadily gaining ground , see MORGAN and KANOLT ( ' Jour .
Amer .
Chem. Soc 1904 , 26 , p. 636 ) and and GETMAN ( ' Zeit .
Phys. Chem 1904 , 49 , p. 390 ) .
Amongst others the following cases have been of solutions in which the molecular conductivity decreases with dilution : Silver nitrate , cadmium iodide , and ferric chloride in amylamine ( KABLENBERG and RUHOFF , 'Jour .
Phys. Chem 1903 , 1 , p. 284 ) ; Antimony bromide and phosphorus pentabromine in bromine ( PLOTNIKOFF , 'Jour .
Russ .
Phys. Chem. Soc 1902 , 34 , p. 466 ; 1903 , 36 , p. 794 ) ; Hydrogen chloride in other and in amyl alcohol ( KABLUKOTF , ' Zoit .
Phys. Chem Y889 , 4 , p. 439 ) ; Hydrogen chloride in cineol SACKUR , , \amp ; c. ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
The consideration of these abnormalities has led KAHLENBERG to conclude that the theory of ionic dissociation is not applica.ble to the majority of conducting solutions .
It is our object to show that the abnormal behaviour of solutions in the solvents examined by us can be simply and consistently explained in terms of the theory of ARRfflNIUS , if the assumption is made that the original dissolved substance , being itself incapable of undergoing ionic dissociation , either polymerises or combines with the solvent to form a compound containing more than one molecule of the solute , and that the polymer , or compound , as the case may be , then acts as the electrolyte .
Those non-saline organic substances which are able to conduct the current when dissolved in certain solvents are considered by WALDEN ( ' Zeit .
Phys. Chem 1903 , 48 , p. 385 ) to be abnormal , in view o their ] -known constitutions and their behaviour in aqueous solutions , but if the foregoing assumption of the formation of compounds be made , these substances are not more abnormal electrolytes than ammonia , which with water forms the compound ammonium hydroxide .
it has been suggested that the existence of compounds of the solute with the solvent is proved by the abnormal variation of the molecular conductivity , to which reference has been made ; but the following considerations will show that an increase of with dilution furnishes no evidence for or against the occurrence of such compounds .
Let us suppose that a reaction between the solute , AB , and the solvent , , takes place according to the equation AB , CD , and let a be the active masses of the three substances .
Now , provided that moderate dilutions are used , we are justified in regarding as constant , when from the law of mass action constant .
Now if conduction is due to the dissociation of the compound ABCD , the number of dissociated molecules is given by , where is the degree of ionic dissociation ; but ; therefore the concentration of the ions is equal to if is the number of ions formed from one molecule of solute .
But the specific conductivity of the solution is proportional to the ionic concentration , and therefore ( 1 ) , and since the molecular conductivity it must vary with , that is to say , it must increase with dilution even when a compound of the solute with the solvent is formeA DR. B. D. STEEL , DR D. hIcINTOSH AND DR. E. H. ARCHIBALD This is the case , for example , in an aqueous solution of ammonia , to which reference will be made later .
If , however , we assume that two or more molecules of AB unite to form a compound which undergoes ionic dissociation , AB itself being unable to conduct the current , then the molecular conductivity may decrease with dilution whether the solvent enters into the composition of the electrolytic compound or not .
If we consider the two cases:\mdash ; ( 1 ) A compound of molecules of solute with molecules of solvent is formed according to the equation ( CD ) the active masses being and Then , if we again consider sufficiently dilute solutions , may be regarded as constant , and or If ionic dissociation occurs so that of the compound is ionised , then , as before , the ionic concentration The specific conductivity . .
( 2 ) .
The molecular conductivity and since the dilution or In the development of this relation no assumption has been made as to the nature of the ionisation of the electrolyte .
If we consider the second case , for example , there are a number of ways in which the compound can ionise .
Thus ( 1 ) ( 2 ) and generally ( S ) If dissociation takes place according to the first of these equations , 2 ions result from the dissociation of 1 molecule of the electrolyte .
If according to the second equation , the number of ions is , and , generally , the number is Now whatever value may have , the number of ions present is given by and is therefore proportional to ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
( 2 ) Combination between the solute and the soIvent does not occur , but an ionising polymer of the solute is formed .
In this case the equation is the active masses are and and from the law of mass action and which leads to the same expression as before , namely , and Now , since the number of molecules always increases with dissociation , must increase with dilution in whatever manner dissociation takes place ; but unless the increase in is greater than the diminution in which is brought about by dilution , the sum of the effects due to the variation of and of in the equation must produce a diminution in with increasing dilution .
It follows , therefore , that the molecular conductivity may decrease with dilution in the case of any conducting solution in which the electrolyte is a compound of two or more molecules of the dissolved substance , whether it is a simple polymer or a compound containing one or more molecules of the solvent .
It folows , also , from the equation that in the case of a solution in which the ionic dissociation was approaching completion , in which therefore varied but slightly , the specific conductivity of the solution should be very nearly proportional to the power of the concentration of the dissolved substance ; for such solutions we therefore have the relation Although we have shown that it is not necessary that union with the solvent should occur in order to bring about an abnormal variation of , we nevertheless .
consider that the formation of such compounds*affords the best explanation of the behaviour of solutions of organic substances in the halogen hydrides and in sulphuretted hydrogen .
* See also WALKER 'J .
C. 1904 , vol. 85 , p. 1082 , and WALKEn , McINTOSH , and AltCBIBALD , ' J. C. 1904 , vol. 85 , p. VOL , X DR B. D. STEEL , DR. .
McINTOSH AND DR. H. ARCHIBALD We are led to this conclusion by a consideration of the following facts:\mdash ; ( 1 ) Large quantities of heat are evolved when conducting solutes are added to either of these solvents .
This heat evolution we take to indicate chemical union .
( 2 ) Compounds containing a varying number of molecules of solvent have been isolated ARCHIBALD McINTOSH , Jour .
Chem. .
( 3 ) The ionisation of a compound such as is much easier to understand than that of a simple polymer such as In order to apply the foregoing conclusions to a specific case , we will consider a solution of acetone in hydrogen bromide .
According to our hypothesis , such a solution contains a compound of acetone and hydrogen bromide , the formula of which we will assume to be This compound dissociates simultaneously in two different ways , a certain number of molecules being dissociated into acetone and hydrogen bromide , other molecules being dissociated into ions , and the ratio of the number of molecules undergoing each dissociation will be constant .
Applying the equation , we see that the molecular conductivity will increase with .increasing concentration of the acetone , the increase\mdash ; neglecting variation of \mdash ; being proportional to the square of the concentration .
Similarly , we see that if is nearly constant , the specific conductivity will be proportional to the cube of the acetone concentration .
If , however , is not constant , then will be proportional to the cube of the acetone concentration .
This conception of an intermediate compound which is able to break up in different ways is by no means new to chemists , and the solution of ammonia , which we have already referred to , furnishes an example of such a case , which , in many ways , is analogous to the preceding .
The compound that is formed in this solution is ammonium hydroxide , and the dissociations are ( 1 ) ; ( 2 ) The relation between specific conductity and concentration for such a solution has been already developed in equation ( 1 ) , which a special case of equation ( 2 ) .
Here again so that ; but so that is simply the molecular conductivity at infinite dilution .
ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
In order to test the conclusions that have been arrived at , we require to know the concentration of the unassociated solute for each dilution .
If A gram-molecules of the solute are dissolved in one litre of solvent , and if there are formed gram-molecules of the electrolytic compound , then , if molecules are required to form one molecule of the compound , molecules of solute will have been used , when equilibrium is established , so that the equilibrium equation is or and the general expression for the specific conductivity becomes No solution of any substance in any one of the halogen hydrides has yet been found with high value for the conductivity , a fact which may be assigned to one of three causes , namely , either ( 1 ) the concentration , or ( 2 ) the coefficient of ionisation of the , electrolyte is small , or ( 3 ) the ionic velocities may be very small .
If we assume the first to be the most probable cause , will not differ much from , and we may without sensible error make use of the values for the total concentration in applying the above equation to our results .
This has been done , the equation being used in the form of , and the results of the calculations are given in Tables I. and II .
Table I. contains the values of V and of or for those solutions in which , that is , in which two molecules of solute combine with the solvent to form one molecule of the electrolyte .
Table II .
contains the similar value for those solutions for which It will be noticed that in some cases the figures exhibit considerable irregular variation .
This is to be expected from the fact that the measurements of conductivity and of concentration are subject to considerable experimental error .
These errors were not specially guarded against , as our object was to establish beyond question the nature of the variation of with V rather than to obtain accurate measurements , which , in the present state of our would not possess any special value .
The figures for at very high dilution are , in some cases , quite valueless as a test of our hypothesis , on account of the enormous influence of very slight errors of observation at these dilutions .
The results contained in Tables I. and II .
are shown graphically in figs. 1 and 2 respectively .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. E. H. ARCHIBALD compelled to use , in the calculation of , the total concentration instead of that of the unassociated substance .
It is also possible that more than one type of electrolytic compound is formed in a given solution , as , for example , the compounds ABCD and ; in which case the total conductivity will be the sum of the conductivities due to the ionisation of each of these compounds .
In such a case as this extremely complicated curves might result .
Moreover , we cannot strictly apply the equation to the concentrated solutions , since for these the active mass ( b ) of the solvent is no longer constant .
The fact that the curves , as a whole , are so analogous to those for a simple electrolyte in aqueous solution appears to indicate that , as a general rule , the main effect is due to the ionisation of a single substance .
The equation should also be applicable to abnormal solutions in other solvents .
This is the case for the solutions investigated by PLOTNIKOFF ( ' Zeit .
Phys. Chem 1904 , 48 , p. 224 ) , who found very abnormal variations of for antimony tribromide and phosphorus pentabromide in bromine .
The experimental figures for antimony tribromide lead to the following values for V and The molecular conductivity of phosphorus pentabromide in bromine varies so irregularly as to suggest that some disturbing effect is at work rendering the figures valueless .
Another solvent in which increases with concentration is amylamine ( KAHLENBERG and RUHOFF , ' Jour .
Phys. Chem 1903 , 7 , p. 254 ) , and the equation has been applied to the measurements of conductivity for cadmium iodide , ferric chloride and silver nitrate dissolved in this solvent .
The results of the calculations when , 3 and 4 are given in Table III .
It will be noticed that maxima are shown in each case .
Passing on now to the consideration of the molecular weight determinations which are recorded in Part II .
, we find that some of these afford confirmation of our hypothesis , inasmuch as ether and acetone in dilute solutions possess a greater molecular weight than the theoretical .
DR. B. D. STEEL , DR. D. McINTOSH AND DR. H. ARCHIBALD Thus WALDEN and CENTNERSZWER ( Zeit .
Phys. Chem 1902 , 39 , p. 513 ) found that the molecular weight of potassium iodide dissolved in hydrocyanic acid is twice as large as the normaL ABEGG ( ' Die Theory der electrolytischen Dissociation , ' p. 103 ) has pointed out that this can be explained by the assumption that the undissociated substance is polymerised ; in which case a high average molecular weight might occur even with considerable ionisation .
A compound such as , for example , if it were completely dissociated into two ions , would have an average molecular weight of 332 .
We find that in most cases the molecular weight increases with increasing concentration , and although the opposite change occurs in dilute aqueous solution , this variation is the same as that which takes place in more concentrated aqueous solutions .
This will be seen from the following comparison of the figures for acetone dissolved in hydrogen bromide with those for lithium bromide dissolved in water , the latter figures being taken from a recent paper by JONES and GETMAN ( ' Zeit .
Phys. Chem 1904 , 49 , p. 390 ) .
Acetone in hydrogen bromide\mdash ; concentration in gram-molecules per litre ; molecular depression ( b ) Lithium bromide in water\mdash ; ; JONES and GETMAN attribute the apparent increase in the number of m olecules- in more concentrated solution to the formation of hydrates in solution .
The low molecular weight which we have found for triethylammoniam chloride in sulphuretted hydrogen , although at first sight difficult to reconcile with the hypothesis of association , is not inconsistent with it .
Thus if the compound formation and subsequent dissociation takes place according to the general scheme and and if dissociation were nearly complete , it is evident that if is equal to or greater than , a larger number of molecules than would be formed , and therefore the average would be less than the theoretical molecular weight .
We can offer no suggestion as to why toluene , when dissolved in hydrogen chloride , although it absolutely fails to conduct the current , possesses such an extremely low molecular weight .
Similar cases have been observed by KAHLENBERG , but no explanation has been suggested .
ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS .
165 A explanation of the abnormal variation of molecular conductivity might be fotlnd in the hypothesis that when acetone or ether is added to hydrogen bromide oetone other aonising solventWhen loint o ' variation owhich actuaHy occursand tydrogen.bromide ionised .
appears as a normal one .
This explanation is , however , shown to be incorrect when we come to consider the transport number experiments .
Thus , during the electrolysis of ether in hydrogen bromide , the deposition of 1 gram-molecule of silver by the current is accompanied by a transport of 8 grammolecule of ether to the cathode .
But , if the ether did not take part in the electrolysis , the same result would be obtained by the transport of a sufficient quantity of bromine as anion from the cathode to the anode .
A simple calculation , however , shows that in order to bring this about no less than 23 gram-molecules of bromine must be transported for every gram-molecule of silver .
Now we have shown that FARADAy'S law is valid for solutions in hydrogen bromide , and accordingly we conclude that ether takes part in the carriage of the current , and that conduction is not due to ionisation of the hydrogen bromide .
Information regarding the constitution of the electrolyte is also afforded by the port number .
If we again consider the case of ether dissolved in hydrogen bromide , there is in solution an electrolyte of the formula , which can ionise either ( 1 ) into ions and a complex anion or ( 2 ) , , , , , , cation If the former , the ether will be transported to the anode as a component of a complex anion ; if the latter , it will be carried to the cathode as a component of a complex cation .
Experiment has proved that the latter is the case not only , for ether but also for the other substances which have been examined .
It has been found that the cation transport number increases considerably with concentration .
This increase can be easily explained if we assume , with JONES and GETMAN ( loc. that the number of molecules of solvent in combination with one molecule of solute is greater in the more dilute solution .
According to the theory of ABEGG and BODLANDER ( ' Zeit .
fur Anorg .
Chem 1899 , 20 , p. 453 ) , the resulting change of constitution of the electrolyte would be conditioned as follows : Any salt in whioh one ion is much weaker than the other manifests a tendency to ' form complex ions by the addition of a neutral molecule to the weaker ion .
In the solutions under discussion the weaker ion would undoubtedly be the complex cation , which , when the active mass of the solvent ( neutral molecules ) was increased by :dlution , would tend to become still more complex by the addition of more solvent moleculea DR. B. D. STEEL , DR. D. McINTOSH AND DR H. ARCHIBALD The effect of this increased complexity would be that the velocity of the ion would be diminished without altering the ionic change , and also that the concentration change at the cathode would be lessened , owing to the carriage of extra solvent molecules to the cathode .
Both of these effects would cause a diminution of the cation transport number as the solution was diluted .
It will be noticed that this explanation involves a change in the active mass of the solvent , and , as a matter of fact , it was not possible to measure the transport number except in solutions which were so concentrated , that the assumption of a constant active mass for the solvent was no longer justified .
We have not been able to calculate , even approximately , the velocity of the various ions , as we had no means of determining the actual nature , concentration , or degree of dissociation of the corresponding electrolytes .
The foregoing pages contain an account of measurements of the vapour pressures , densities , surface energies , and viscosities of the liquefied hydrides of chlorine , bromine , iodine , sulphur , and phosphorus .
solvent action of these substances has also been investigated , and we have shown that , with the exception of phosphuretted hydrogen , they are all able to act as ionising solvents , and the conductivity , molecular weight , and transport number of certain dissolved substances have been measured .
The results of the measurements , although abnormal , are not inconsistent with the ionic theory ; since we have shown that\mdash ; ( 1 ) If in a given solution the electrolyte is a compound containing molecules 01 the dissolved substance , the concentration of this compound will be proportional to the er of the concentration of the dissolved substance , and therefore the expression for the molecular conductivity of the electrolyte becomes instead of We have also shown that , and therefore the molecular conductivity of the electrolyte increases with dilution in these solutions in the same manner as in aqueous solutions .
The variation of the molecular conductivity of the electrolyte with dilution is probably complicated by the occurrence of compounds which contain a different number of solvent molecules at different dilutions .
( 2 ) The want of agreement between conductivity and cryoscopic measurements is a necessary consequence of the occurrence of polymers or compounds in solution , and may be taken as evidence of the existence of such compounds .
( 3 ) The conduction of organic substances when dissolved in the halogen hydrides is best explained by the occurrence of electrolytic compounds of the organic substance with the solvent .
Transport number measurements have shown that organic substance is carried to the cathode as a component of the complex cation In xpress ohanks trofessor BHYDRIDES AONDUCTING SOLVENTS .
eriot Wollege , lacing facilities aisposal aDioector ocGill Universi Chemical Laborat a interest taken in the work .
Our thanks are also due to Professor JOHN Cox , Director of the McGill Physical Laboratory , for the use of apparatus , and for kindly supplying us with large quantities of liquid air .
We also wish to express our thanks to the Research Grant Committee of the Chemical Society for a grant made to one of us , by means of which a large portion of the expense of the work of Part III .
has been met .

