]\gt ; INDEX SLIP .
, W. R. , and LOWBY , T. M.\mdash ; Tho Ele ieal Conducbivity and other Properties of Sodium Hydroxide in Aqueous Solution as Elucidating the Mechanism of Conduction .
Phil. Trans. , , vol. 204 , 1905 , pp. LOWRY , T. N. , and BOUSFIELD , W. R.\mdash ; The Electrical Conduotivity and other Properties of Sodium Hydroxide in Aqueous Solution as Elucidating tho of Conduction .
Phil. , vol. 204 , , pp. 26$-322 .
Sodium ation of Concentrated Stan Solutions of , from Ketallic Sodium .
Density , Conductivity , and Viscosity of Solntions of .
BOUBFIELD , W. R. , and LOWRY , T. M. Phil. Trans. , , vol. 204 , 1905 , pp. 253-322 .
nsity of Aqueouo of 8odium Hydroxide .
BOUSrIELD , W. R. , and LOWRY , T. M. Phil. , vol. 204 , 1905 , pp. } Conductivity ( Electricul ) Specific , Molecular , and Intrinsic , of Sodium Hydroxide in Aqueous Solutions .
, W. R. , and LOWRY , T. M. Phil. Trams .
, , vol. 204 , pp. 253-322 .
Viscosity of Aqueoul Solubiona of Sodium Hydroxide Correlated with Electrical Conductivit .
BousrILD , W. R. , and LOWRY , .
M. Phil. Trans. , , vol. 204 , 1905 , pp. 253 ] .
The Electricat Conductivity and othe } Properties of Hydroxide in Aqueous Solution as Elucidating the of onduction .
WILLIAM ROBERT BOUSFIELD , A. , K. C. , M.P. , } THOMAS MARTIN LowRy , D.Sc .
, Lecturer in Physical mistry at the Central College , South Kensington .
by Professor HENRY E. ARMSTRONG , F.R. Received June iead N , 1904 .
CONTENTS .
Page troduction 253 Part I.\mdash ; Preparation of coucentrated standard solutions of sodium oxide from metalIic sodium 357 Part II.\mdash ; Determination of the density of queous solutions of sodiunt hydroxide at C. 265 Part III.\mdash ; Influence of temperature ou the density of aqueous solutions of sodium hydroxide 374 Part IV.\mdash ; The molecular volume of sodium hydroxide in aqueon solution 280 Part V.\mdash ; The conductivity of sodium hydroxide in aqueous solution at C. 383 Part viscosity and intrinsic conductivity of aqueon of sodium hydroxide 289 Part \mdash ; Influence of temperature on the COIlduc.tivity of queous solutions of sodium hydroxide NaOH.\mdash ; Preparation from metallio sodium of aouceutrated standard solutions . .
\mdash ; Densities of aqueous solutions of sodium 7250.\mdash ; Conductivity electric.al ) of solutions of sodium hydroxido ith othor properties .
INTRODUCTION .
IN very dilute solutions the lelationship between solvent and solute is such as to admit of simple and exact mathematical treatment , and it is in this field that the theory of electrolytic dissociation has been applied with the greatest success .
Until recently it was generally believed that the formulae deduced from a consideration of the properties of very dilute solutions could be applied without any serious loss of accuracy to solutions of moderate concentration .
This hope has , however , not been realised .
A careful parative study , especially by WHETHAM and by GRIFFITHS , of electrical and osmotic properties of dilu.te solutions has shown that even at ccnoentratious as low as gram-molecule per litre there a discrepancy between ( 378 .
) 23.3.05 the values for the coefficient of ionisation deduced by the two methods .
It appears , in fsct , that the formulae in which ARRHENIIJS'S hypothesis has usually found expression , cau only be regarded as accurate when applied to soIutions containing more than 100 litres of water for each gram-molecule of dissolved salt .
As a qualitative expression , however , theory is applicable to electrolytes of all kinds , and the meant of exact ical relationships is only hindered by the difficulty of pl.ocul.ing the ecessary experimental data , a difficulty that in many cases can only be overcome by the discovery of new methods of investigation .
The essential postulate of ' theo1y is that a salt may exist both in an active and in an inactive form as egards conductive power , so that when brought into an electrical circuit , only part of the total weight of electrolyte is as a rule directly operative in carrying the .
This conception finds expression in the -efficient of ionisation , , which represents the proportion of the salt which is thus active in electrolytic conduction .
Independently of any particnlar theory as to the exact nature of the difference between the ionised and un-ionised part of the salt , the formula is universally applicable as expressing the fact that the specific conductivity of any ) inary compound is dependent on the mass of the coml)ound contained in unit volume , the proportion that is able to take paxt in the nsport of rent , and the velocities and with which the oppositely charged ions move through the solution under the action of the electrical forces .
In the case of dilute aqueous solutions the numerical values of , and are accurately known .
In the case of less solutions they are approximately known .
In the case of concentrated solutions and in the case of fused salts they are altogether unknown .
The present papel is essentially a contribution to the study of concentrated aqueous solutions .
We hope at some future date to extend the investigation to dilute solutions , but the experiments now described deal almost excIusively with solutions of normal concentration and above .
The electrolyte chosen for investigation was sodium hydroxide , which has the advantage , not possessed by many binary salts , of being exceedingly soluble in water , so that even at C. it is possible to measure of a 50-per cent. solution , though at this temperature such a solution is highly supel.satulated .
In common with others who have been engaged in the study of concentrated aqueous solutions , we ] lave been impressed with the fact that combination between solvent and solute is an essential part of the of solution , and that the of hydrates must be ised not only in the crystals which eparate wheu the solutions are cooled , but also in the solutious themselves .
We not , howevel , beeu able to econcile ovations with the view , iu itself improbable , that this power of combinatiou is no longer operative in dilute solutions , and indeed have obtaiued evidence that if the reduction iu volume which frequently accompanies dissolution may be regarded as an index of the OQiation of solvent and solute , the degree of association reaches a maximum only at dilutions which are not far removed from those required for complete ionisation .
We believe , eoefore , that CONDUCTIVITY OF SODIUM IN SOLUTION .
25.5 combination with the solvent is an essential part of the process of ionisation , and are of opinion that in the case of dilute aqueous solutions ionisation is simply a final stage in the process of hydrate formation , whereby the hydrated molecule is resolved into two or more hydrated ions .
As evidence in support of this view is to be found not only in the observations now recorded , but also in a wide range of current literature , we have reserved the fuller discussion of this theory for a separate communication .
The chief features of the paper may be described under four headings : ( 1 .
) Influence of on \mdash ; In our former paper ( ' Roy .
Soc , Proc 1902 , vol. 71 , pp. 42-54 ) we were able to give a general curve expressing the relation between conductivity and temperature over the whole range of tempera- ture within which electrolysis is possible .
This general curve was arrived at from a consideration of the experimental data with reference to composite electrolytes of all classes , ranging from glass to solutions in liquid ammonia and sulphur dioxide .
These scattered observations we were enabled to piece together by a theoretical consideration of the influences which determine the change of conductivity with temperature .
In the present paper we have been able to give quantitative expression to these different factors , and have obtained results which afford a complete confirmation of our earlier conclusions .
In the formula , which we have used to represent relation between specific conductivity and temperature , allowance has been made for each of the three factors which influence the conductivity .
Changes of concentration are allowed for by introducing as a factor the ratio , which expresses the change of density between and .
Changes of ionic mobility have been allowed for by introducing a factor similar to that used by SLOTTE to represent the variations of with temperature .
Changes in the coefficient of ionisation have been allowed for by introducing a factor similar to that used by ABEGG to express the decay of specific the temperature rises .
Although these factors may not represent perfectly the variations of ionic mobility and of ionisation with temperature , the formula has been found to satisfactory representation of the experimental facts over a wide range of temperature .
It accurately expresses the varied types of conductivity-temperature curves recorded in the present paper , and indicates no only the point of inflexion which is observed in moderately dilute solutions of hydroxide between C. and 10 C. , but also the maximum conductivity and the second point of inflexion shown in our general conductivity-temperature curve .
Even more important is the fact that by means of this formula it is possible , for the first time as we believe , to analyse the conductivity-temperature curves for concentrated as well as for dilute solutions , and to estimate separately the effects produced by the increase of mobility and by the decay of ionisation which accompany a rise of temperature .
MR. .
R. AND DR. T. M. LOWRY ON THE ELECTRICAL ( 2 .
) Influence of Viscosity on Conductivity.\mdash ; In the conventional method of deducing the coefficient of ionisation from the molecular conductivity by means of the equation , it is assumed that the ionic mobility is independent of the concentration .
As , however , there is a marked difference ( amounting to something like 1 per cent. ) between the viscosity of water and that of a decinormal salt solution , it is clear that even at this small concentration some appreciable alteration must be looked for in the iunic mobility , and the method is only strictly accurate at concentrations less than .
In order to correct for the variation of ionic mobility produced by changes of viscosit , , we have assumed as a first approximation that the alterations of mobility are proportional to the alterations in the fluidity of the solution .
We have used the term " " intrinsic conductivity\ldquo ; and the symbol I to represent the ratio obtained by dividing the molecular conductivity by the fluidity of the solution compared with that of water at the same temperature at unity , so that whilst , we have .
The coefficient of ionisation is accurately given by the relation ) ; if were constant it would also be given accurately by the relation ; and with the help of this relation we believe that it will be possible to extend the method of determining the coefficient of ionisation from the conductivity to considerably greater concentrations than those within which the validity of the relation can be recognised .
The possibility that the ionic mobility might be influenced by the dielectric constant of the medium as well as its viscosity has been discussed by SUTHERLAND ( ' Phil. Mag 1902 , ( vi .
) , vol. 3 , p. 161 ) , but cannot be entered into here .
( 3 .
) tolytic Conductirity of ) Salts.\mdash ; All the acids behave as dielectrics when pure , and only become electrolytes when associated with water or some other " " ionising solvent The alkalies , on the other hand , like the majority of salts , are excellent electrolytes when fused , and are believed to conduct the current quite independently of any interaction with other substances .
For the purpose of convenience we propose , at Professor ARMSTRONG'S suggestion , to describe these two classes of electrolytes as heterolytes and autolytes .
It was conceivable that the conductivity of dissolved salts was not wholly dependent on the ionising properties of the solvent , but could be attributed in part to the mere presence of the liquified salt , which could then conduct in the same way as when fused .
An important result of the present investigation has been to show that , when a correction for viscosity has been applied , concentrated aqueous solutions of sodium hydroxide exhibit an " " intrinsic conductivity\ldquo ; which is too great to be accounted for by the direct influence of the solvent , and which actually increases with the concentration of the soda .
This increase we attribute to the autolytic conductivity of the dissolved salt .
A similar effect is manifest in the isothermal conductivity-concentration curves for C. and C. , which appear to indicate that the ffised alkali would be an excellent conductor it over-cooled to these temperatures , and that it contributes largely to the conductivity of the more concentrated solutions .
There can be little doubt that CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
similar influences are at work in the majority of concentrated aqueous salt solutions , but the effects are usually masked by the increased viscosity of these solutions , and very few measurements have been made at concentrations at which the autolytic conductivity of the salt becomes of importance .
* ( 4 .
) Constants of Sodium Hydroxide.\mdash ; The available data with reference to the physical properties of the alkalies are more scanty and less accurate than in the case of the more important acids and salts .
This result is mainly due to the exceptional difficulty of preparing and standardising pure solutions of these substances .
After a long series of experiments we succeeded in devising a method by which it is possible to prepare concentrated solutions of sodium hydroxide directly from weighed quantities of metallic sodium .
We were thus enabled to determine some of the physical properties of the solutions with a closer approach to accuracy than has hitherto been attained .
This was notably the case with the densities at C. Data are also given , though with a somewhat larger error , for every C. from C. to C. The electrical conductivity at C. was also determined with an accuracy of about 1 part in 1000 , and , as the result of experiments on the influence of temperature on conductivity , we are also able to tabulate the conductivities at C. , C. , PART I.\mdash ; PREPARATION 0F RATED STANDARD SOLUTIONS 0F SODIUM HYDROXIDE FR0M METALLIC SODIUX .
Until a few years ago the best material available for experimental purposes was caustic soda purified by dissolution in alcohol .
More recently the commercial preparation of metallic sodium has led to the introduction of caustic soda , prepared by the cautious addition to one another of drops of water and particles of sodium in a silver dish ; this material contains little impurity except carbon dioxide , and this can readily be lemoved by means of baryta .
Solutions prepared from these materials are necessarily of unknown strength , since the sodium hydroxide is never obtained sufficiently free from moisture and other impurities to be directly weighed out .
Standardisation is usually effected by titrating against an acid of known strength , but this process is liable to serious e1rors , and we were soon compelled to abandon the attempt to standardise our solutions in this way .
Ultimately we were led to * SUTKBRLAND loc. observed that in the case of zinc sulphate the ratio , corresponding to our , passed through a minimum value , but attributed this result to altogether different causes .
The measurements of viscosity given in Part VI .
are to be regarded only as approximate determinations .
The errors were probably of the order of 1 per cent. , but the conclusions based upon these measurements would not have been influenced by much larger errors in the viscosity constants of the solutions .
It was considered undesirable to describe in detail a type of viscometer that had not been tested with reference to its tbility to exact measurements .
VOL. CCIV .
determine with special care the density of a series of solutions prepared from weighed quantities of metallic sodium and to construct tables , by reference to which it was possible to determine the percentage of sodium hydroxide in any solution with an error amounting to little more than per cent. The approximate values previously obtained by titration were then foumd to be incorrect to the extent of per cent. to per cent. The quantitative conversion of sodium into sodium hydroxide was first undertaken on a small scale in order to obtain a reliable standard alkali with which to compare our unknown solutions by titration .
But it soon became obvious that a closer approach to purity and accurate standardisation would be attained if we could prepare the whole of the material required for the investigation from weighed quantities of metallic sodium .
Quantities of sodium amounting to 50 grammes or 100 grammes were therefore fused and poured into a weighed silver mug ; the mug ( which we found must be provided with silver baffle-plates to stop the spray ) was placed in a shallow dish of water , covered with a glass shade , and exposed to an atmosphere of steam produced by gently warming the dish .
The action proceeded quietly , and the rate at which the sodium was acted upon was indicated by the escape of bubbles of hydrogen from the glass shade .
This arrangement proved , however , to be unstable , for ultimately the metal , which had become covered with a layer of caustic soda , floated to the top of the solution and caught fire in the steam , in one case after action had proceeded during about 6 hours , and in a second case after about 30 hours .
In order to avoid this result it would have been necessary to prolong the action during many days .
We were therefore led to devise forms of apparatus in which the sodium was suspended in such a way that the caustic soda produced was to flow from the metal , thus maintaining a clean surface and uniform conditions throughout the action .
In this way we succeeded , for the first time , in converting a bar of sodium directly into caustic soda , though the method involved the risk that the sodium might become sufficiently heated to melt and flow , with disastrous results , into the water or soda beneath .
We were also able to carry out the operation quantitatively , though in a somewhat risky manner , by suspending a weighed quantity of metal in a silver cradle at the top of a large platinum test-tube over a quantity of water which was distilled on to the metal , under reduced pressure , by gently warming the tube .
method was completely successful , but was somewhat tedious in the later stages , owing to the great reduction of vapour-pressure caused by the dissolution of caustic soda in the water .
This difficulty was overcome by distilling the water into the platinum tube from a flask , of which the neck was drawn out and bent so as to enter the cover of the tube ; in this form the apparatus was used for preparing some two litres of standard 50-per cent solution .
Some of the details of the method may now be described .
( 1 .
) and Weighing the Sodium.\mdash ; In order to get rid of oxide and other superficial impurities , the sodium was melted under xylene or ffin oil , and was CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
obtained in a clean condition by drawing it up into a short-nosed pipette ( previously warmed over a flame to prevent the sodium from solidifying in the tube ) , and then squirting it into a porcelain basin containing melted paraffin wax .
By allowing the wax to solidify , the metal could be kept in a very convenient form , without undergoing any further oxidation .
The most convenient method of weighing the metal was to draw it up into a weighed pipette , in which the air had been freed from oxygen and moisture by displacing it through a -tube containing chips of metallic sodium ; in this way it was possible to obtain a known weight of untarnished metal , but in a form in which it could not readily be converted into the hydroxide .
We were therefore obliged to modify the method by transferring the melted sodium , before weighing it , from the pipette into a bomb-shaped flask with short narrow neck , in which the air had been freed from moisture and oxygen by suspending in it a fragment of sodium during some hours before the flask was used .
In taking the sample of sodium from under the melted paraffin wax , purified air was blown through the pipette to prevent the entrance of paraffin , and after the metal had been sucked up , the outside of the pipette was wiped before allowing the metal to flow into the bomb .
The flasks , which had a volume of about 200 cub. centims. , were nearly filled with the molten metal , c , losed with a small paraffined cork , allowed to cool , and weighed , after momentarily loosening the cork in order to avoid a partial vacuum in the flask during weighing .
When required for use the metal was again melted , and poured , at a temperature of about 15 C. , into the silver cradle used in converting it into hydroxide .
When cold , the flask was again weighed , after momentarily loosening the cork ; the difference gave the weight of sodium transferred minus the weight of any oxygen absorbed by the metal in the flask during the operation .
In the weight of sodium in the cradle , the metal was protected from the air by a thin cover of silver foil , rapidly cooled by dipping the cradle into cold water , and weighed as quickly as possible .
The increase in weight of the cradle gave the weight of metallic sodium plus that of any oxygen absorbed by the metal in the transfer .
The two weighings of the sodium differed , therefore , by an amount that was dependent on the sum of the oxidations in the flask and in the cradle .
Usually the difference amounted to 10 milligrammes or 15 milligrammes , and only on one occasion rose to 30 milligrammes in consequence of steam produced during the cooling of the cradle obtaining access to the metal under one edge of the silver cover .
As the quantity of metal weighed was about 150 grammes , and the oxidation product was not more than 15 milligrammes , the mean weight of the sodium was probably correct within 1 part in 20,000 .
The error due to oxidation was certainly less than that caused by uncertainty in the value of the atomic weight of the metal .
It may be of interest to mention that in pouring the sodium from the bomb into the cradle , the metal always flowed through a thin tube of oxide , reaching from the mouth of the flask to the surface of the metal , and that even if the bomb had contained a considerable quantity of oxidised metal , this could not have been poured into the cradle , but would have remained adhering to the side of the flask .
It may also be mentioned that , whilst steam has a very penetrating action on the metal , a thin film of oxide is sufficient to protect it from oxidation by moderately dry air .
We relied largely on the use of considerable weights of metal , amounting in each experiment to about 150 grammes , to render inappreciable the effects of superficial oxidation , and the results which we obtained fully justified this course .
( 2 .
) Conversion of the Sodium into Hydroxide.\mdash ; In order to avoid contamination during the conversion of the sodium into hydroxide , the sodium which had been weighed into the silver cradle was transferred to a platinum vessel , and this vessel was also used for many of the subsequent conductivity measurements .
The platinum pot was in the form of a test-tube , as shown in fig. 1 , and held about 500 cub. Fig. 1 .
Fig. 2 .
centims. When used for converting sodium to hydroxide it was provided with a dome-shaped lid , also of platinum , which fitted into the top of the pot .
Two cylindrical apertures were provided for admitting the steam and removing the hydrogen .
The whole apparatus was thickly coated with silver , electrolytically deposited , in order to resist the pressure of the atmosphere , and the dome-shaped lid was further strengthened by means of silver ribs .
An air-tight joint was made CONDUCTIVIK OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
by phcing an indiarubber band between the platinum flanges of the lid and the pot ; these flanges were strenghened by means of heavy silver rings and were held together by means of three clamps .
The silver cradle used to contain the sodium had the form shown in fig. 2 and occupied the whole of the upper half of the pot .
It consisted essentially of a cylindrical box , 2inches in diameter and inches high ; one-third of the cylinder was cut away to allow of ffiling and emptying ( this aperture was covered with silver foil whilst the sodium was weighed ) ; the upper surface was pierced by a circular hole for the admission of , and the projecting part of the lower surface was provided with a series of small holes for the escape of the caustic soda .
A stout silver wire held the cradle in position in the pot , and when the conversion was complete the cradle was lifted out , inverted and used for mixing the solution after the soda adhering to the silver wire had been washed down by means of water . .
The arrangement used in effecting the conversion is shown in fig. 3 .
The pot was heated to about C. and water was distiUed into it under reduced pressure from a pear-shaped flask heated to about C. in a second water-bath .
A manometer was provided to indicate the pressure , and a mercury trap to prevent water from being sucked back into the pot if the pump should cease to act .
The completion of the action was marked by the distillation of water , which was condensed in a vertical tube and allowed to flow back into the pot .
As a result 01 this distillation the upper part of the pot was kept free from soda and no loss was incurred when the flask was removed and the lid of the pot wiped in order to weigh the solution .
The weight of the solution was roughly determined before allowing it to cool , and after adding from a burette the volume of water required to dilute to 50 per cent. a careful weighing was made and the exact percentage calculated .
The water required for dilution was also used to rinse the wire of the cradle and the upper parts of the pot .
The balance used was sensitive to gramme and the weight of the solution was about 400 grammes ; the errors of weighing and the slight errors due to variations in the weight of the apparatus were less than those by using the accepted value for the atomic weight of sodium .
In order to illustrate the method , a typical series of weighings may be given:\mdash ; and sodium after pouring Sodium .
Cradle lid and sodium Sodium Total oxidation . . . .
Mean weight of sodium Air correction True weight Weight of pot , \amp ; c. , and solution 269 805 grammes .
119 .
245 . . . . .
150 560 86 .
402 . . . . .
gramme .
Solution . . . . . . . . .
517 145 Air correction . . . . .
True weight of solution 517 48 grammes .
Weight of sodium hydroxide l50734 grammes .
Concentration of solution . . . . .
per cent. ( 3 .
) Purity of the Material.\mdash ; Although the method used was calculated to give solutions more nearly to accuracy and purity than those prepared by any other of the available methods , both the sodium and the hydroxide were tested for impurities which might have vitiated the subsequent measurements .
By using the metal as raw material it was possible to ensure that the soda contained no acid impurities such as chloride or sulphate , and as the solution was protected from the CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
air and not allowed to come into contact with glass , it was free , at least initially , from carbonate and silicate .
In one experiment the sodium left in the cradle a few milligrammes of a black granular metallic residue ( possibly iron sawdust introduced in opening the case in which the sodium was packed ) ; although the quantity was probably less than per cent. of the weight of sodium , this solution was regarded as of unknown strength and was not used as a standard ; in all the remaining experiments the only residue was occasionally a brown stain on the silver cradle .
The sodium was therefore free from all but the merest trace of heavy metals .
A further test was made for the presence of alkaline earths .
About half a litre of concentrated solution had been allowed to become slightly carbonated , and was decanted from the carbonate that had been precipitated at the bottom of the bottle ; on adding water the precipitate was found to be completely soluble , and there was therefore no appreciable quantity of calcium or magnesium in the metallic sodium or in the hydroxide prepared from it .
The presence of a small quantity of potassium or lithium would have had little influence on the subsequent measurements , but a spectroscopic test did not reveal the presence of these elements , and we were therefore led to conclude that the metal we had used as our standard was free not only from acidic impurities , but also from all but traces of foreign metals .
More difficulty was experienced in ascertaining whether the metal might not be alloyed with carbon or hydrogen in the form of carbide or hydride .
When decomposed by water , these substances would giye rise to volatile products and would not affect the purity of the solutions , but if present in considerable quantities might lead to serious errors in the concentrations .
Sodium carbide does not appear to be soluble in metallic sodium , but under certain conditions large quantities of hydrogen may be absorbed .
The concordance of our measurements for the density of 50-per cent. solutions of sodium hydroxide prepared from different samples of sodium seemed to indicate that there could not be any large quantity of volatile impurity in the .
metal , but on the advice of Dr. SCOTT we tried an additional method oi purification by heating a sample of sodium in an iron ladle in such a way as to expel any carbide or hydride that might initially be present .
The metal was heated until it caught fire and was then poured into melted paraffin wax , but when subsequently converted in the usual way into hydroxide it gave a normal value for the density , and did not appear to have been altered in any way by the roasting .
As a matter of fact , metallic sodium is now made almost exclusively by electrolytic processes and no longer contains the black explosive carbide that was formerly associated with it ; eyen when crude sodium is heated , the black residue that is left in the ladle appears to consist merely of a little lamp-black from the petroleum , and if clean metal is used no black scum is produced .
A test for acetylene in the gases produced by the action of steam on the metal gave a negative result .
An attempt was made to detect the presence of hydride in the metal by determining the total loss in weight of the apparatus during the action .
In this we were unsuccessful ; the quantity of hydrogen to be weighed was about 6 grammes , and under atmospheric pressure would have occupied some 70litres ; working under reduced pressure the volume would be increased to about a thousand litres , and our drying apparatus proved to be insufficient to hold back the moisture contained in this large volume of gas .
As , however , we considered the possible presence of hydrogen in the metal to be the only serious error to which our method of preparing standard solutions was liable , we were led to devise a further method of testing for this impurity .
According to the measurements of TROOST and HAUTEFEUILLE ( ' Comptes Rendus , ' 1874 , vol. 78 , p. 807 ) , sodium hydride begins to dissociate at alittle above 30 C. ; at 33 C. its vapour pressure is 28 millims. and rises to 752 millims. at 42 C. We therefore took a test-tube of hard glass , placed in it 3 grammes of sodium and exhausted by means of a mercury pump .
After any occluded gases had been liberated by fusing the metal , the temperature was raised , by means of a bath of fusible alloy , to , in order to dissociate the hydride , the temperature being recorded by means of a high-pressure mercurial thennometer .
In this experiment no gas was liberated .
In the second experiment , with 5 grammes of metal , a bubble of gas was obtained which occupied about four-tenths of a cubic centimetre , but was probably air rather than hydrogen .
This experiment was conclusive , for the presence of per cent. of hydride-hydrogen would have produced 5 cub. centims. of gas , and at the temperature used the whole of this must have been liberated into the vacuum produced by the pump .
We consider , therefore , that whatever impurity may have been present , none was in sufficient quantity to influence our measurements , and that if the atomic weight of sodium be taken as , the solutions used as standards were accurate within approximately the same limits , that is about 1 part in 5000 .
The solutions required for the measurements of density and conductivity at C. were for the most part stored in platinum vessels until required , but , by taking advantage of the relative inertness of the most concentrated solutions , we were enabled to make use of glass bottles for storing the solutions required for determining the temperature coefficients .
In consequence , perhaps , of its extreme viscosity and the relatively small amount of water present , a 50-per cent. solution of sodium hydroxide has far less action on glass than a dilute solution .
Its density did not alter by a single unit in the fourth place when stored during ten weeks in a glass bottle .
further advantage in the use of concentrated solutions is found in the fact that the carbonate is almost insoluble , and , if produced by exposure to air , separates as a precipitate at the bottom of the bottle , producing a small change in the concentration of the solution , but hardly any deprec of its purity .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
SODIUM HYDROXIDE AT C. Determinations of the density of aqueous solutions of sodium hydroxide have been made by some six different investigators , but owing to the difficulty of preparing pure solutions of known concentration , the results obtained have not been characterised by any close approach to accuracy , and the published tables are far less trustworthy than those for a substance like common salt , which can be weighed out in an anhydrous condition , or sulphuric acid , which can be purified by distillation and freezing .
In every case the solutions appear to have been standardised by titration , and it is clear that the errors introduced by inaccurate values for the concentration , and by neutral impurities in the solution , must have been many times greater than those due to the mere measurements of density .
In the early tables of RICHTER ( 1795 ) and of DALTON ( 1810 ) the density is given to two places of decimals .
A more ambitious table was published by TuNNERMANN .
in 1827 ( ' Neues Journal der Pharmacie fur Aertzte , ' vol. 18 , p. 2 ) , in which the percentages of were given to three places of decimals ( 1 part in 30,000 ) and the densities to four ; the additional accuracy of this table is for the most part fictitious , errors are to be found even in the second place of decimals .
values , in conjunction with those of DALTON , were , however , made use of 30 years later by SCHIFF ( ' Annalen , ' 1858 , vol. 107 , p. 300 ) , who deduced from them a formula connecting density with concentration , his object being to obtain by extrapolation a value for the density of fused 100-per cent. sodium hydroxide .
SCHIFF'S formula was used by GPRLACH in 1867 ( ' Zeitschrift fur Analytische Chemie , ' vol. 8 , p. 279 ) to calculate the densities for each exact percentage of and NaOH , and GERLACH'S table was reproduced in all the dictionaries and text-books of chemistry until 20 years ago .
It is remarkable that the crude measurements of density made in the early years oflast century should be still in use ; it is even more remarkable that during more than half a century no attempt at revision was made , each successive writer being content to quote from his immediate predecessor the untested figures of a former generation .
A revision of the tables of density of sodium hydroxide was attempted by HAGEN , who published in 1883 ( ' Commentar sir Pharmacopoea Germanica , ' Berlin , 1883 ) a list of densities determined to four places of decimals at no less than 100 concentrations between per cent. and per cent. NaOH .
This list is given in LUNGE'S ' Sulphuric Acid and Alkali , ' and is quoted thence in the current edition of WATT'S Dictionary , but is rather less accurate than the table of GERLACH , which it was intended to replace .
It appears , therefore , that until 10 years ago no table had been published in which the density of aqueous solutions of sodium hydroxide was given with a sufficient to accuracy to render even the second of VOL. CCIV .
decimals altogether reliable .
All the tables referred to above are quoted in full in COMEY'S ' Dictionary of Solubilities .
' The measurements of PICKERING ( ' Phil. Mag vol. 37 , 359 , 1894 ) must be placed in a different category from those described above .
As in previous caaes , the stock solution was standardised by titration , but it had been prepared from metallic sodium and in all probability was substantially pure .
Unusual care appears also to have been taken with the density measurements , which were made at no less than 144 different concentrations between per cent. and 50 per cent. NaOH .
No details are given with regard to the titrations , nor is it stated what acid or alkali was used as standard ; the author considers , however , that the value for the strength of the stock solution was correct within per cent. of the total alkali present .
Had this been the case , the density of the 50-per cent. solution should have been correct within ; actually , however , the value is higher than that which we have found for solutions prepared quantitatively from metallic sodium .
The difference in density corresponds to an error of per cent. in the determination of the total alkali in the solution , and our own experience has shown us that errors of this magnitude may easily be introduced in standardising an alkali by titration against a standard acid .
Although , therefore , PICKERING'S relative values for the density at different concentrations must be regarded as substantially accurate , the absolute values ( which are given to six places of decimals , and are quoted in this form in WATTS'S ' Dictionary of Chemistry ' ) appear to be unreliable to the extent of at 50 per cent. NaOH , and roughly to a proportionate extent at the lower concentrations .
It is to be regretted that PICKERING did not apply in this case the method , which proved so successful in his determinations of the density of sulphuric acid ( ' Trans. Chem. Soc 1890 , vol. 57 , p. 64 ) , of freezing out a hydrate of definite composition and using this to prepare the solutions ; an application of this method might have given results as accurate as any that can be obtained by weighing out metallic sodium , and a comparison of the two sets of data would have been of great value .
In weighing out either a hydrate of sodium hydroxide or metallic sodium , it is necessary to introduce an analytical factor based upon atomic weight determinations .
As the atomic weights are only known to two places of decimals , it would be useless to pretend to any higher approach to accuracy in the concentration of the solutions used in the density determinations .
In the case of sodium hydroxide , this factor may be taken as involving a possible error of about 1 in 5000 in the concentration , corresponding to an error of in the density of a 50-per cent. solution .
At the higher concentrations therefore , the densities can only be relied upon to four places of decimals , and in view of the unstable character of the material it is doubtful whether the purity of the solutions would justify a closer approximation , even if the accuracy of the atomic weights should appear to permit it .
In a 5-per cent. solution an error of 1 part in 5000 in the concentration would produce the CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
same effect as an error of in the density , and it is therefore permissible to give the density to 5 places of decimals .
Our original intention was to measure the densities by suspending a cylindrical sinker in the platinum pot , and determining the loss in weight due to the buoyancy of the soda solution .
As the sinker had a volume of 100 cub. centims. , it was only necessary to measure the loss in weight within a centigramme , in order to determine the density to four places of decimals .
It was thought that by adding successive quantities of water from an accurate burette , and checking at the end the total weight of water added , it would be possible to make a rapid series of accurate observations , and to measure the density at frequent intervals over a range of concentration with little more trouble than that involved in two or three density determinations made by the ordinary method .
In this we were disappointed , at least as regards the more concentrated solutions , for ( 1 .
) The viscosity of the 50-per cent. solution was so great that a sensitive balance required several minutes to show a displacement , even when the weight in the pan was altered by 1 centigramme or 2 centigrammes at a time ; and ( 2 .
) The heat generated by the addition of water was so great , and the temperature adjustment so slow , that at least an hour was required before an observation could safely be taken ; a surprisingly large amount of stirring was also required before the solution was sufficiently uniform to give a constant reading .
For these reasons the method , though excelently adapted for more dilute solutions in which there is no marked heat evolution , proved so tedious in case of the concentrated solutions that it was abandoned in favour of a pyknometer .
Even when using a pyknometer , the densities of the concentrated solutions could not be determined without considerable trouble , for many hours were required before the freshly diluted solutions were free from minute air bubbles , and the temperature adjustment was so slow that a 50-cub .
centim. pyknometer could not safely be adjusted in less than an hour .
Altogether seven standard solutions were independently prepared by the method described in Part I. One of these was used only for conductivity measurements .
Density observations were made on all the other six .
In order to show the extent of the agreement in the values obtained from these different solutions , a list is subjoined of all the density measurements in the immediate neighbourhood of 50 per cent. MR. W. R BOUSFIELD AND DR. T. M. LOWRy ON THE I[LECTRICAL TABLE The extreme variation from the mean density the 50-per cent. solution is only , and the average error , some part at least of which must be attributed to small errors in the measurements of density .
We consider therefore that the mean value , , for the density of a 50-per cent. solution is correct within The method adopted for measuring the density of the less concentrated solutions consisted in weighing the standard solution in the pot , diluting with a calculated volume of water , and again weighing to determine the exact concentration .
After mixing thoroughly by means of a stirrer , or one of the electrodes described in Part V. of the paper , an Ostwald pyknometer of 50 centims. capacity was filled with the solution , and the remainder was again diluted .
Although the weighings were calculated to secure an accuracy of over 1 in 10,000 in the dilution , it was considered desirable after 5 or 6 dilutions had been made to start again ffom one of the standard 50-per cent. solutions , in order to avoid the influence of cumulative errors such , for instance , as contamination by carbon dioxide .
A typical series of hings will illustrate the method used .
Weight of pot per cent. solution .
1218 97 grammes .
weight of solution Air . . . . .
True weight of solution .
Weight of soda . . . . .
Added centims. of water .
* The mixed solutions were stored in a glass bottle ; the remainder had not been allowed to come into contact with glass .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Weight of pot and solution . . .
grammes .
Apparent weight of solution Air . . . . .
True weight of solution .
Strength of solution per cent. From 50 per cent. to 25 per cent. NaOH the density measurements are considered to be accurate within .
Altogether 38 determinations were made within this range , usually at intervals of per cent. These are set out in Table V. In one or two instances individual measurements showed a discrepancy of as much as but no difficulty was experienced in deciding which figure should be adopted , and in every doubtful case fresh dilutions were made direct from the standard 50-per cent. solution .
The measurements made at 25 per cent. and 20 per cent. NaOH are tabulated in order to illustrate the degree of accuracy obtained .
TABLE TABLE NOTE .
sinker ; pyknometer .
At the lower concentrations the density measurements are less tedious and more exact , and the strength of the solution is known with sufficient accuracy to justify the inclusion of the fifth place ol decimals in the table of density .
Within the range from 20 per cent. to per cent. , 28 measurements of density were made .
Usually two determinations were made with each solution , and the densities deduced from the weighings of the two pyknometers did not differ in any case by more than .
Within the whole range , from 20 per cent. to per cent. , the relative values for the densities are probably exact within , but the absolute values are subject to a possible error of about in the range from 25 per cent. to 20 per cent. NaOH , from 20 per cent. to 15 per cent. , and ffom 15 per cent. to 10 per cent. Interpolation Formuloe.\mdash ; The curve connecting density and concentration is approximately linear in character , curvature being most marked from 5 per cent. to per cent. , and less so in the range from 30 per cent. upwards , but apart from this it does not exhibit special peculiarities .
The curve is therefore of little use either for indicating the constitution of the solution or for interpolation .
For the latter purpose we have made use of three kinds of sensitive curves .
I. A very convenient means of interpolation consists in recording the extent to which the ordinates of the density curve differ from those of the diagonal line shown in Diagram I. ( p. 310 ) .
This sensitive curve is obtained by plotting against , and has the form shown in the diagram .
In using it we have set out the percentage as abscissae on such a scale that per cent. was represented by 1 millim. , and the ordinates on such a scale that 1 millim. corresponded to a difference of in the density ; when using this scale , it was easy to read off the density to four places of decimals , and the calculation required was reduced to a minimum .
Thus , in reading the density from the curve we have for per cent. , and II .
The second sensitive curve represents the deviation of the density curve from a straight line by means of the variations in the tangent of the angle between the radius vector and the axis of concentration .
It is obtained by plotting against , and has the shown in Diagram L This curve has the advantage that its sensitiveness increases with the dilution , and it has proved especially useful for interpolation at concentrations at which it was desired to record the fifth place of decimals in the density .
III .
Over the range from 5 per cent. to 18 per cent. NaOH the density-percentage curve is a straight line .
This is shown by the first sensitive curve , but is entirely masked in the second .
The accuracy of the observations may therefore be tested by comparing the values calculated from the linear formula with those which were obtained by interpolation from our erimental values by means of the second sensitive curve .
The result is shown in Table III .
CONDUCTIVITY OF SODIUM HyDROXIDE IN AQUEOUS SOLUTION .
TABLE III .
In no case does the difference exceed , though the absolute values for the density are liable at 18 per cent. NaOH to an error of on account of the fact that the concentration is only known to 1 part in 5000 .
with the Results of other Observers.\mdash ; With the help of the temperature coefficients deduced in Part III .
of the paper , it is possible without any loss of accuracy to calculate the values for the density at C. and so to make a comparison of our results with those of other investigators .
This has been done in Table VI .
of the paper , in which the densities at C. and C. are given for each exact percentage ffom to 50 .
On account both of errors of titration and impurities in the solutions the earlier values are in almost every case higher than those given in Table VI .
, and the errors extend in some cases even into the second place of decimals .
* The deviations from our determination at five typical concentrations are as follows:\mdash ; TABLE IV .
We had expected that , when corrected for the error due to titration , PICKERING'S values would have shown a close agreement with our own , or , in other words , that the differences between the two series of observations would have been proportional *See footnote post , p. 287 .
per cent. value .
Calculated by means of an interpolation formula from the observations of DALTON and TUXNERMANN .
to the concentration .
This , however , is not the case , as is shown by the differences tabulated in the fourth column of Table VI .
Up to 20 per cent. the difference does not exceed , and is for the most part negative in sign ; above this concentration the difference is positive , and increases more or less proportionately ffom at 20 per cent. to at 50 per cent. If the two series of observations had differed only on account of the error introduced by titration , the difference at 25 per cent. should have been half as great as at 50 per cent. , and should have amounted to ; actually it amounts only to , and as the average error of our five observations at this concentration amounted only to , and in the case of PICKERING'S observations to only , it is clear that the differences are not due exclusively to errors either in the density determinations or in the standardisation of the stock solutions , but must be attributed in part to a difference in purity or some similar cause .
TABLE V.\mdash ; Density Observations at C. NOTES .
pyknometer , sinker .
The pyknometer values are the more nearly accurate .
The most ( [ iable solutions are those )ared ( usually by diluting direct from 60 per cent. ) from the mixed solution .
In this series the densities of the solutions were determined by means of two pyknometers .
The mean of the duplicate determinations is given above in some cases as if it were a single observation .
In no case did these determinations differ by more than .
In series V there was some uncertainty in the dilution from 46 per cent. to per cent. , and the concentration of the solution was efore re-calculated from its density at 40 per cent. and 20 per cent. ; the intermediate observations are of interpolation only .
CONDUCTIVITY OF SODIUM HYDROXIDE AQUEOUS SOLUTION .
TABLE VL \mdash ; Density of Aqueous Solutions of Sodium Hydroxide at C. and C. VOL. CCIV .
\mdash ; A. VOL. CCIV .
\mdash ; A. VOL. CCIV .
\mdash ; A. VOL. CCIV .
\mdash ; A. PART III.\mdash ; INFLUENCE 0F TEMPERATURE 0N THE DENSITY 0F AQUEOUS SOLUTIONS 0F SODIUM In order to be able to correct the conductivity measurements for changes of concentration produced by alterations of temperature , it was necessary to know not only the density of the solutions at G. , but also the variation of density over the whole range from C. to 10 C. The only statement that we have been able to find with reference to the density of caustic soda at temperatures other than C. is a note by HAGER to the effect that the correction for temperature amounts to per degree centigrade from 50 per cent. to 40 per cent. , from 29 per cent. to 30 per cent. , from 29 per cent. to 20 per cent. , and from 19 per cent. to 10 per cent. NaOH .
The measurements of the variation of density with temperature , originally undertaken on account of their bearing on the measurements of conductivity , proved to be of some interest on their own account , and were therefore carried considerably further than was originally intended .
No special difficulty was encountered in measuring the density at temperatures up to about C. , but at higher temperatures the operation became somewhat troublesome , and the measurements made at the boiling-point were found to be subject to much larger errors than those made at lower temperatures .
For the mere purpose of correcting the conductivities for the changes of concentration produced by expansion it would have been sufficient to know the density to three places of decimals , but the fourth place was always determined , and , although the values obtained at the boiling-point were occasionally erratic , we believe that the smoothed values are accurate within a few units in the fourth place .
For this series of density measurements we employed a type of density-tube very similar to that described by Dr. PERKIN ( Trans. Chem. Soc 1884 , 45 , p. 444 ) , the actual form being represented in fig. 4 .
The bulb is necessary in order to allow for the expansion of the liquid when determining the density at temperatures below that of the atmosphere , but it is also useful at higher temperatures .
Fig. 4 .
Only in the case of the most dilute solutions at temperatures in the neighbourhood of the -point was difficulty encountered from the distillation of moisture into the bulb .
The tubes used in some of the earlier experiments had a volume of only about 8 cub. centinla , but CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
usually the volume was increased to 15 cub. centims. or 25 cub. centims. in order to reduce the errors of adjustment and weighing ; the accuracy of the observations was not very much increased by using larger tubes , and probably the chief source of error arises from fluctuations of temperature .
In the case of the concentrated solutions the great viscosity called for the exercise of some degree of patience ; to fill a tube , even with the help of an efficient pump , often required an hour or more , and it was noticed that more than an hour elapsed before a tube of 47 per cent. soda ( capacity 25 cub. centims. , diameter less than 2 centims. ) acquired a steady temperature when cooled in powdered ice .
At the lower temperatures the solutions were frequently supersaturated , but we did not experience difficulty on account of crystallisation either in the density tubes or in the conductivity vessels , though occasionally the stock solutions deposited magnificent transparent plates of hydrated soda .
From five to ten measurements were made with each solution , the temperature being steadily raised throughout the series ; the error due to the action of the soda on the glass does not appear to be serious , and no change whatever could be detected in the density at C. of a 47-per cent. solution after heating for an hour each at C. , at C. and at 10 C. Fonn of the Curves .
Formuloe.\mdash ; A consideration of the densitytemperature curves for a series of percentages beginning with water and ending with 50 per cent. soda showed conclusively that all the relationships could be represented with a close approach to accuracy by equations of the form It was further found that at about 12 per cent. NaOH the coefficient of vanished , and that thenceforward up to 50 per cent. the relation could be expressed within the limits of experimental error by an equation of the second degree .
In order to make these results clear , we append in Table VII .
a list of the observed densities for water , decinormal , normal , per cent. , per cent. and per cent. NaOH , together with the densities calculated from the cubic or quadratic formulae .
Only in exceptional cases did the observed and calculated densities differ by more than but it is necessary to add that a close agreement on a single series of observations did not necessarily afford a measure of the accuracy of the series , since a serious error at 10 C. might well be masked by an alteration in the coefficients which would not seriously disturb the values calculated for the density at lower temperatures , and would not appear until different series were compared with one another .
For this reason we deemed it very important to check the density-temperature observations by drawing isothermal density-percentage curves for various tempera- tures and by noting the gradual change in the coefficients of expansion with changing concentration .
The temperatures chosen for the isothermal curves were C. , C. , C. , and 10 C. , and , having already plotted an accurate isothermal curve for C. , it was only necessary now to plot the differences and .
The difference curves had the advantage that they could be plotted on a larger scale than the densities , and they were found to be much simpler in form than the sensitive curves described in Part II .
of the paper .
Where measurements of density had not been made at the exact temperatures chosen for the isothermal curves , values for these temperatures were calculated by means of the cubic or quadratic formulae described in the preceding paragraph ; the interpolated values , being based to some extent on the whole series of measurements , were at least equal , and probably superior , in value to the individual observations of the series .
The difference curves served the double purpose of bringing into relief the experimental errors in measuring the change of density with temperature , and of providing a means of deducing by interpolation the densities for exact percentages at the exact temperatures chosen .
TABLE VII.\mdash ; Soda Densities , Observed and Calc.ulated .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
The observed values for water are taken from WATT'S 'Dictionary of Chemistry , ' where the values deduced by VOLKMANN from the determinations of HAGEN , MATTHIESEN , PIERRE , KOPP and JOLLY are given to five places of decimals .
The differences shown in the table indicate that an additional term must be introduced into the equation in order to express the influence of temperature on the density of water correctly to the fourth decimal place .
The values which we have set out for the densities of caustic soda solutions at different temperatures are all based upon the measurements at C. described in Part Ir .
of the paper .
The values for C. are deduced from those at C. with the help of the difference curye for , and as the maximum deviation of any individual experimental values from this curve , as shown in Diagram II .
, is only , we believe that the values for are accurate within .
This close agreement is of considerable value , as is an important constant in the equation used in deducing the densities at C. , C. , C. , \amp ; c. The values for C. are deduced from those at C. by means of the difference curve for ; in this case the maximum difference between the experimental values and the curve was , and we consider that the values for are probably accurate within The values for 10 C. were deduced from those at C. by means of the difference curve for ; in this series the experimental values differ in two cases from those deduced from the curve by as much as , and the average error amounts to , but we consider that the values for C. are probably accurate within Table VIII .
gives the values of , which were used for the above purpose ; and also the temperature coefficients and A remarkable confirmation of the substantial accuracy of the values for , at which temperature the experimental error reaches a maximum , was afforded by the discovery that the values for could be represented by the simple logarithmic law , where percentage of soda in solution .
This formula was not in any way used to fix the final values for and indeed was not discovered until after these had been tabulated , but the substantial agreement between the two series of values , shown in Table VIII .
, which do not differ in any case by more than , appears to indicate that we have not over-estimated the accuracy of our observations by placing the probable error in ithin TABLE VIII.\mdash ; Expansion of Aqueous Solutions of Sodium Hydroxide between C. and lo C. The coefficients in the equation which are also given on Table VIII .
, were deduced from the final values for and were plotted against the percentage of soda in the solution .
The resulting curves were quite smooth , with the exception of a break at 12 per cent. which was caused by the change from the cube to the square law ; this break might have been avoided by giving to a small negative value above 12 per cent. , but as the effect of this on the densities would have been unappreciable , it was thought better not to reintroduce this coefficient when once it had vanished ffom the equation .
These coefficents were used to calculate the density of the solutions for every 2 per cent. at intervals of C. from C. to C. , and the results are shown in Table IX .
The gradual change in the coefficients is well shown in Diagram IL , which also shows the difference curves for and CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUOIOT .
General Results of the Investigation.\mdash ; Reference may now be made to the more striking results that have followed from the investigation of the influence of temperature on the density of aqueous solutions of caustic soda .
The most marked characteristic , which was noticed very early in the investigation , is the great simplification observed in the behaviour of the solutions as the concentration and the temperature are increased .
This is indicated by the gradual alteration in the coefficients , but is also shown by the density-temperature curves of Diagram IIL The maximum density falls to C. at about per cent. NaOH , and is given by the formula as C. at 2 per cent. NaOH .
The curves become steadily flatter as the concentration increases , and at per cent. the density decreases in a linear manner with the temperature , a behaviour which is in marked contrast to the complex character of the curve that represents the influence of temperature on the density of water .
The gradual reduction in curvature , indicated by the decrease with increasing concentration of the coefficient , was noticed by BREMER ( ' Rec .
Trav .
Chim 1887 , vol. 7 , 268-309 ) in the case of aqueous solutions of calcium chloride , but our observations show further that may actually become zero .
At still higher concen- trations becomes positive in sign and the curvature is reversed .
Interesting results would doubtless accrue from a comparison of the influence of increased concentration and of increased temperature in straightening out the densitytemperature curves .
In general terms it may be stated that the curve connecting density and temperature for water between 10 C. and 2o C. would probably be very similar to that for a solution containing 10 per cent. NaOH between C. and 10 C. , and that if a soIution containing 10 per cent. NaOH were cooled down ffom C. to C. , its behaviour would probably be very similar to that of water between 10 C. and C. The cause of the simplification in behaviour of the concentrated solutions is discussed in connection with the changes of molecular volume described in the following Part IV .
of the paper .
PART MOLECULAR VOLUME 0F SODIUM HYDROXIDE 1N AQUEOUS SOLUTION .
The molecular volumes of dissolved salts are usually much less than those of the solid salts .
This contraction cannot be attributed to the mere liquefaction of the salt , as this would probably give rise to an increase of volume , but it may well be attributed to the combination of solvent and solute .
In the case of a few compoun\amp ; , of which the hydroxides of sodium and lithium are by far the most striking examples , the addition of the substance to water actually causes a decrease in the total volume of liquid , and the molecular volume is negative for dilute solutions .
In the case of sodium hydroxide the effect of this contraction is remarkable ; it is possible to take a litre of water and dissolve in it no less than 100 gramunes of sodium hydroxide without causing any increase in volume , whilst if 50 grammes are added there is actually a contraction of 3 cub. centims. CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
The solution volumes , V , of 1 gramme of sodium hydroxide at C. are given in the last column of Table VI .
, together with the totul dration , , of the solution expressed in molecules of water per molecule of soda .
When these are plotted against one another , they give the hyperbola-like curve of Diagram IV .
The symmetry of this curve , though not exact , is of interest because it appears to indicate that the prooess of hydration , on which the contraction of volume depends , proceeds continuously from the most conoentrated to the most dilute solutions .
From this point of view it was of interest to inquire at what dilution the molecular volume would reach its limiting value .
A more careful study of the curve of Diagram in conjunction with a similar curve in which the molecular volume was plotted against the logarithm of the hydration , revealed the fact that there was an alteration of curvature at about 45 per cent. NaOH and a further change at oout 5 per cent. It was possible , however , to obtain remarkably concordant results for the value of the solution volume by two entirely different methods of extrapolating from the densities between 1 and 5 per cent. The following table shows the agreement between the observed densities and those calculated from the formulae where is the density of water , the density of the solution , and the percentage of sodium hydroxide .
TABLE X. observed)atod ) From formula the solution volume is given by the expression , where ; and from formula ( b ) by the expression where The limiting value deduced from formula is o1423 , and from formula ( b ) is The concordance between these figures renders it probable that they afford an VOL , CCIV .
accurate estimate of the limiting value of the solution volume ; the limiting value previously given was ( ' Dictionary of Chemistry , ' vol. IV .
, p. 491 ) .
By the same process of extrapolation it can be deduced that the molecular volume does not become constant until a dilution of about 100litres is reached .
The molecular conductivity reaches a maximum value at about 1000litres , and it is therefore clear that the decrease of molecular volume proceeds side by side with the process of ionisation , and does not cease until the latter is almost complete .
A study of the influence of temperature on the solution volume of sodium hydroxide gave somewhat remarkable results .
Instead of increasing steadily from C. to 10 C. , as was expected , the solution volume at all percentages was found to reach a maximum value at C. to C. , and then to decrease between C. and lo C. There is thus a close resemblance between the solution volume at C. and that at C. The extent of the variation produced by alterations of temperature decreases with increasing concentration , and at 50 per cent. NaOH amounts only to 7 per cent. The data for the solution volume are given in Table XI .
and are represented graphically in diagrams V. and VI .
In discussing the significance of the variations of solution volume it may be regarded as certain that the great variation in the case of dilute , as compared with concentrated , solutions must be attributed to the complex character of liquid water and not to any complexity introduced by the addition of soda .
It appears , in fact , that the maximum of simplicity is reached in a 50-per cent. solution at lo C. , and that the most complex solution with which we have to deal is water .
The complex character of the water molecule was demonstrated by RAOULT ( ' Ann. Chim .
Phys 1884 , ( 6 ) , vol. 2 , p. 66 ) , and was made use of by RoNTGEN ( ' Wied .
Ann 1891 , vol. 45 , p. 91 ) , who attributed the increase of density between C. and C. to the dissociation of " " ice molecules\ldquo ; having a greater molecular volume than the water molecules .
This view was elaborated by SUTHERLAND ( ' Phil. Mag 1900 , vol. 50 , p. 460 ) , who suggested that RoNTGEN'S ice molecules had the formula and might be trihydrol , whilst the water molecules consisted of dihydrol , .
A further discussion is given by STRADDLING ( ' Jour .
Frank .
Inst 1901 , vol. 152 , pp. 257-268 ) with reference especially to the compressibi]ity of water .
It is probable that the simplification in the density-temperature curves of the preceding Part III .
of the paper , which results from the addition of soda or ffom an increase of temperature , is due primarily to the destruction of the ice molecules to which water owes most of its abnormal properties .
A similar explanation may be given of the large negative solution volume of sodium hydroxide , which is most pronounced at those low temperatures at which the water contains a large proportion of the bulky ice molecules .
This explanation does not , however , in any way account for the fact that the molecular volume of sodium hydroxide decreases ffom C. to lo C. The only explanation that we can offer to account for this unexpected behaviour is that over this range of temperature the water contains alarge proportion of steam CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEO(IS SOLUTION .
molecules , having the simple formula , that these , like the ice molecules , have a larger molecular volume than the water molecules , and are destroyed by adding soda to the solution .
The addition of soda may then be regarded as simplifying the solution mainly by lessening the proportion of associated and dissociated water molecules , the water molecules themselves being converted into hydrates of soda having somewhat similar properties to dihydrol , but an even smaller molecular volume .
This view , which is an extension of the theories of RoNTGEN and of SUTHERLAND , appears to be fully justified by the experimental facts , and indeed it is difficult to avoid the conclusion that in the neighbourhood of the boiling-point the simple steam molecules form , if not the largest , at least a very important proportion of the total number .
TABLE XI.\mdash ; Solution Volume of 1 gramme of Sodium Hydroxide between C. and C. * C. - .
07 - .
06 - .
04 - .
03 - .
012 -.003 PART V.\mdash ; THE CONDUCTIVITY 0F SODIUN HYDROXIDEi AQUEOUS SOLUTION AT C. The only measurements that we have been able to find of the electrical conductivity of sodium hydroxide in aqueous solution , apart from a single measurement by LOOMIS Phys. Chem. of the specific conductivity and density of a normal solution , are those of KOHLRAUSCH ( ' Wied .
Ann 1897 , vol. , pp. 1 and 145 ; l885 , vol. 26 , p. 161 ) .
As , however , the values given by KOHLRAUSCH for the densities are considerably higher than those which we have deduced in Part .
of this paper , it .
would seem that the accuracy of the electrical measurements was very largely neutralised by the presence of neutral impurities in the * Our thanks are due to Mr. R. .
QUERTIER for his help in calculating Tables IX .
and XI .
[ December , 1904.\mdash ; The first four rows of figures have been amended .
] MR. W. R. BOUSFIELD AND DR. T. M. LOWRY ON THE ELfflCTBICAL solution*or by errors introduced in standardising the solutions by titration .
We have therefore redetermined the conductivities at C. , and although our electrioal measurements are probably not so accurate as those of KOBLRAUSCH , we consider that the purity and accuracy of our solutions and the increased number of observations combine to render our measurements at C. more reliable than those previously recorded .
( 1 .
) The SOlutions used for the electrical measurements were similar to those employed in measuring the density , and indeed in one series of dilutions the same solutions were used for measuring the two properties .
As it was likely that the conductivity would be influenced more than the density if the alkali should become partially neutralised by carbonic or silicic acid , special care was taken to maintain as far as possible the original purity of the solutions .
The solutions used in measuring the conductivity from 50 per cent. to 20 per cent. were kept in platinum vessels and only allowed to come into contact with the glass of the electrode that was introduced in measuring the conductivity .
The solutions below 20 per cent. were made up from standard solutions which had been kept in platinum or stored in a glass bottle as 50-per cent. solutions during periods ranging ffom a few days up to two months , but as the density remained unchanged at the end of ten weeks , and the maximum conductivity measured at the end of five months was within per cent. of the value obtained before transferring to the glass bottle , it is believeA that no appreciable fouling had resulted from the contact of the viscous 50-per cent. solution with the glass .
( 2 .
) The Conductivity Vessel.\mdash ; No difficulty is experienced in designing a conductivity apparatus of low resistance capacity in which the liquid is allowed to come into contact only with platinum surfaces , but it is very difficult , perhaps Fig. 5 .
even impossible , to design a simple apparatus of this kind in which the resistance capacity is sufficiently large to give satisfactory measurements with solutions of high conductivity .
We were able , however , to effect a very satisfactory compromise by making use of the platinum vessel already described and introducing into it an electrode shielded within a glass tube of suitable shape .
The total area of glass in the electrode was comparatively snBU , and 300 cub. centims. to 400 cub. centims. of solution used for each observation .
Under these conditions the fouling of the solution by contact with the glass of the See footnote , p. 287 .
QONDUCTIVITY OF HYDROXIDE IN AQUROIJS OLUTION .
electrode must have been very slight The surface of the platinum pot was used as the second electrode .
* ( 3 .
) libration of the Electrode.\mdash ; The majority of the measurements at C. were made with an electrode having the form shown in fig. 5 .
The required resistance capacity was obtained by suitably fixing the length and cross section of the tubes leading outwards from the electrode .
The electrode most used was one which had a resistance capacity of .
For calibrating it a normal solution of potassium chloride was used and two series of measurements gave the capacity as 1 .
) with 200 .
centims. of solution in the pot , and ( 2 .
) with 300 cub. centims. of solution .
A very valuable check on the accuracy of the calibration was obtained by comparing the values for the specific conductivity of the solutions as deduced from measurements with different electrodes .
Using a carefuly calibrated electrode of the type described in Part VI .
of the paper as a standard , the capacity deduced from measurements of the maximum conductivity of caustic soda was per cent. higher , of the Solutiom .
The possibility has been suggested that the electrical measurements might have been vitiated by impurities introduced by the contact of the soda with the glass of the electrodes .
Such fouling might have occurred if the electrode figured above had been heated with the solution during several hours .
This , however , was not the case , as this electrode was used exclusively for low-temperature measurements , and was removed from the solution when not in use .
At atmospheric temperatures the foulin is exceedingly slow , and the experiments quoted on p. 284 may be regarded as affording experimental justification of the view that no appreciable fouling could have taken place in the short period of time during which the electrode was in contact with the solution .
( 2 .
) The advantages which we believe to have resulted from the use of a platinum vessel with glasscased electrode over a glass cell of the ordinary kind may be briefly indicated:\mdash ; ( a ) It is reasonable to suppose that under comparable conditions the total fouling is proportional to the surface of glass exposed to the solvent action of the soda , but is not directly dependent on the volume of the solution .
If this be so , the percentage error due to fouling would be inversely proportional to the volume of the solution , and by using 300 cub. centims. or 400 cub. centims. in place of 10 cub. centims. or 20 cub. centims. the risk of fouling would be proportionately reduced .
The electrode was used as a stirrer and was frequently drawn above the surface of the solution so as to be completely emptied ; there was therefore no local fouling .
( b ) Although the fouling at atmospheric temperatures is too slow to be easily detected , the advantages of the platinum vessels were clearly demonstrated in the experiments at higher temperatures , described in Part .
of the paper .
Solutions heated in glass vessels showed a change of conductivity amounting in some instances to as much as 1 6 per cent. ( p. 300 ) , but with platinum vessels this source of error was altogether eliminated ( p. 300 ) .
Thus , in an actual experiment with the electrode figured opposite , the initial resistance at C. of the solution was ohms ; after making eight observations between C. and C. the final resistance was still 169 2 ohms .
This experiment not only shows the superiority of the platinum but also justifies the conclusion that at low temperatures the fouling of the solution by contact with the glass of the electrode must have been very slight .
The final experiments quoted in Part VII .
were made with electrode figured on p. 298 , fig. 9 , which had an even smaller surface of glass .
and from measurements of a 45-per cent. solution per cent. higher than the value given above .
( 4 .
) The Measurements of Conductivity.\mdash ; The resistances of the solutions were measured by using an alternating current and a telephone .
The wheel-bridge ecommended by KOHLRAUSCH ( KOHLRAUSCH and HOLBORN , 'Leitvermogen der ektrolyte , ' p. 42 was used and found to be both convenient and accurate .
The main disadvantage of the bridge is that the sensitiveness decreases considerably when used in comparing unequal resistances , and that the bridge-wire must be carefully calibrated at intervals of a few months if the instrument is in constant use .
These difficulties were overcome by using the comparison coils ( 1 , 10 , 100 , 1000 and 10,000 ohms ) supplied with the bridge only in determining the approximate resistance of the electrolyte , whilst the final measurements were made by comparison with an equal resistance .
The resistance-box used for this purpose was specialy wound so as to be almost free from electrostatic capacity .
The exact position of silence on the bridge-wire was determined with three resistances differing slightly from one another , and in this way three readings were obtained , from each of which it was possible to calculate the resistance .
As the readings were all close to the centre of the wire , a maximum of sensitiveness was obtained , and it was only necessary to check from time to time the exact position of the electrical centre of the bridge-wire .
The three readings were usually concordant within one or two parts in 10,000 , and if necessary were repeated until concordant results were obtained .
The method may be illustrated by an example taken from a recent series of measurements .
Resistance Reading of Resistance of coils .
bridge-wire .
Mean . . .
Leads . . .
\mdash ; \mdash ; ffisistance capacity 13 738 cm .
; specific conductivity cm .
Concordant readings of the instrument do not necessarily indicate a corresponding approach to accuracy in the values deduced for the conductivity of the electrolyte .
The errors due to heating , polarisation , self-induction , .
and capacity are fully discussed by KOHLRAUSCH and HOLBORN in the volume already referred to ( pp. 54-59 ) .
Of these the most serious are those due to polarisation when the resistance is low , and to electrostatic capacity when the resistance is high .
In our experience the most favourable resistance to measure is about 50 ohms to 100 ohms , but in determining the conductivity of a solution over a range of temperature involving large changes in conductivity we have occasionally been obliged to measure resistances as high as CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
1000 ohms and as low as 5ohms .
The resistance of the leads was reduced as far as possible by employing stout copper rods to connect the bridge with the vessel , and by using copper enclosed in glass , rather than platinum , for the leads inside the conductivity vessel ; there was , however , some loss of accuracy when the resistance to be measured was much below10ohms .
To measure resistances above ohms it was necessary to introduce a correction for the slight electrostatic capacity of the coils of the resistance box ; the correction was determined by stretching a thin wire resistance of 2000 ohms or 3000 ohms over the ceiling in such a way as to be free from capacity and then measuring its apparent resistance ( 1 ) by a direct current and galvanometer , and ( 2 ) by an alternating current and telephone ; the difference gave the correction to be applied to the readings of the telephone , and this correction was found to be propo tional to the resistance measured .
Taking into account the various sources of error to which the measurement of the electrolytic conductivity of solutions of varying concentrations is liable , we consider the general accuracy of our tabulated results for C. may be taken to be within 1 part in 1000 , although an inspection of Diagram VII .
, in which all the observations are set out , shows that there are some individual readings which are below this standard .
The results obtained in measuring the conductivity of aqueous solutions of sodium hydroxide at C. are shown in Tables and XIIT .
, and are plotted out in Diagram , together with the values given by KOHLRAUSCH .
The maximum conductivity , , is at about 15 per cent. NaOH , but there is no substantial alteration of conductivity between per cent. and per cent. value for 15 per cent. is .
For a normal solution LOOMIS gives , whilst our values are The values given by are shown on Diagram by open circles , plotted out in reference to the percentages given by him .
If the position of these circles be shifted to the percentages deduced from the densities given by KOHLRAUSCH , most of the observations come very nearly on to our own curve , but the uncorrected values differ by as much as on at 35 per cent. KOHLRAUSCH'S value for the maximum , which should not be affected by errors of concentration , differs from ours by nearly 1 per cent. , and it is therefore probable that his solutions contained appreciable quantities of neutral impurity Prior to the introduction of metallic sodium as a source of purffied caustic soda this alkali was made by the action of lime upon sodium carbonate .
In concentrated solutions the interaction NaOH is probably reversible and the solution would therefore contain a definite proportion of sodium carbonate , in addition to traces of calcium salts .
In this way it is possible to account for the wide difference that exists between the densities tabulated by SCHIFF , KOHLRAUSCH and HAGEN , and the later observations , tabulated by , and those in the present paper , in which the alkali was prepared from metaIlic sodium .
TABLE XII.\mdash ; Conductivity Observations at C. TABLE XIII.\mdash ; Specific , Molecular and Intrinsic Conductivities of Sodium Hydroxide at Exact Percentages .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
2@9 PART \mdash ; THE VISCOSITY AND INTRINSIC CONDUCTIVITY 0F AQUEOUS SOLUTIONS 0F SODIUM HYDROXIDE .
Reference has been made in the earlier parts of the paper to the great viscosity of concentrated solutions of sodium hydroxide , which indeed manifests itself even in so simple an operation as filling and emptying a pipette .
As this was likely to have a great eflect upon the ionic mobility , a series of measurements of viscosity was with the help of which it was hoped that it might be possible to correct the observed conductivities for the influence of viscosity .
Relation between Conductivity , Ionisation and Ionic Mobitity.\mdash ; The usual mode of expressing the relationship between conductivity and ionic velocities is derived from the consideration of the flow of electricity from side to side of a unit cube of the electrolyte under a potential gradient of 1 volt per centim. Thus if and V are the ionic velocities under such a potential gradient , the concentration of the solute in gramme equivalents per litre ( and therefore the number of gramme equivalents in the unit cube ) , the fraction of the solute that is in the ionised or current-carrying state , and the quantity of electricity of either sign liberated by 1 gramme equivalent of any monad ( coulombs ) , then the current .
Since the potential difference between the sides of the cube is unity , the current ( by OHM'S law ) is equal to the conductivity .
Hence , or since , we have , iting u the equation .
Whilst and V are the absolute velocities of the ions expressed in centims. per second , and are what are known as the .
" " ionic mobilities\ldquo ; and are of the same dimensions as The ionic velocities , and therefore also ths ionic mobilities , are determined , not alone by the nature of the ion , but also by the solvent , the temperature , and the concentration of the solution .
In considering the theory of ionisation it is customary to deal exclusively with the case of ' ' dilute solutions\ldquo ; for the values of the ionic mobilities do not differ appreciably from the values at infinite dilution , so that and approximately .
This is the relation from:which the coefficient of ionisation is usually deduced , but recent observations have shown that whilst the relation is universally true , the relation is only valid for solutions of concentrations below about .
At concentrations up to about decinormal it is probably legitimate to deduce .
the coefficient of ionisation from the osmotic constants of the solution or from the electromotive force of concentration cells ( JAHN ) , but it is clear that , except in dilute solutions , the coefficients of ionisation and the ionic mobilities are altogether unknown quantities , and much confusion has arisen from the general tendency to VOL. CCIV .
regard the ionic mobilities as constants for concentrated as well as for dilute solutions .
Correlation between Ionic Mobility and Fluidity.\mdash ; The exact way in which the ionic mobility varies with the temperature and with the concentration of the solution is not known .
Without doubt the mobility depends chiefly upon the form and dimensions of the ions and the fluidity of the medium through which they travel When the temperature or concentration of the solution is altered there is usually a marked alteration in its fluidity , the experimental data point to the conclusion that this produces a change in the ionic mobilities which may , as a first approximation , be justly regarded as proportional to the change of fluidity , so that and , where is the fluidity .
In certain cases this simple linear relationship undoubtedly exists .
Thus , in the case of the excessively dilute solution which constitutes purified water , and in which the process of ionisation must be substantially complete , the conductivity is actually in direct proportion to the fluidity over the range of temperature from C. to C. ( BOUSFIELD and LOWRY , loc. cit. , p. 48 ) .
KORLRAUSCH has shown that the same relationship holds good in the case of dilute solutions of salts such as sodium valerate , in which the mobility of one of the ions is very small ( ' Roy .
Soc. Proo 1903 , pp. 338-356 ) .
In other oases the linear relationship is less accurate , and although the ionic mobilities are closely related to the fluidity , they no longer obey a direct linear law .
Thus KOHLRAUSCH Sitz .
Akad .
Wiss .
Berlin , has shown that the temperature coefficients of ionio mobility in dilute solution are usually less than the fluidity coefficient , the greatest difference being found in the oase of the mobile hydrogen and hydroxyl ion , for which the coefficients are and per degree Centigrade , as compared with , for the conductivity and for the fluidity of water at C. ; for the remaining ions the ooefficients range ffom for the nitrate ion to for lithium .
Furthermore , 'WOLP ( ' Zeit .
Electrochem 1902 , pp. 117-119 ) and RUDORF ( ' Zeit .
Phys. Chem , vol. 43 , pp. 257-304 ) have sho.wn that when the viscosity of a salt solution is increased at constant temperature by the addition of acetic acid or a non-electrolyte , the decrease of conductivity is proportionately less than the decrease of fluidity .
It is , therefore , clear that the relation between mobility and fluidity is in general only approximately expressed by a linear law , but at present the necessary data are not available for any closer approximation to the form of the function .
The linear law , which is an accurate expression in certain instances , we to use generally as a first approximation to the relationship between mobility and fluidity .
Intrinsic \mdash ; The relations , and being *In certain instances , and especially in the ease of complex or hydrated ions , the dimensions of the ions may alter with the temperature and concentration of the solution , but ugually thig effect will be of eecondary importance .
ONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
aocurate , it follows that the true value of for a given concentration and temperature is not , as is usually assumed , but That is to say , we must correct for the change of mobility in order to arrive at the true value of .
The theoretical corrective factor for is , and we propose to call the product of by this corrective factor the " " intrinsic conductivity It is the value of reduced to the conditions of mobility which obtain at infinite dilution .
For practical purposes , and as a first approximation , we assume that Hence , taking the fluidity of the solvent ( water ) as unity and measuring the actual fluidity , , at the given concentration in these units we have Denoting the intrinsic conductivity by I , we have , therefore , , and hence .
Whilst does not represent the intrinsic conductivity as defined above with complete accuracy , it enables us to make a convenient practical approximation to the true value of , which is much nearer than the current approximation The whole correction for changes of mobility is not large except in solutions of considerable concentration , so that any error in the correction applied will produce a much smaller error proportionately in the values deduced for the intrinsic conductivity and for the coefficient of ionisation .
We believe , therefore , that by using the ratio , in place of the ratio , it will be possible to extend the method of determining from conductivity measurements to solutions of considerably greater concentration than those for which the validity of the equation can be recognised ; and even in concentrated solutions the intrinsic conductivity deduced from the formula will enable a closer approximation to be made to the actual amount of ionisation than has been possible when this could only be deduced from the molecular conductivity of the solution .
Hitherto the majority of the conductivity measurements that have been made have been concerned with dilute solutions , in which the correction for viscosity is relatively small , but there is little doubt that when the necessary data of conductivity and viscosity are available for concentrated solutions , the study of the intrinsic conductivity of these solutions will lead to results at least as important as those derived from the study of dilute solutions .
Unfortunately , the measurements that have been made of the viscosity of aqueous solutions , like those of the conductivity , have in almost every case been restricted to dilute solutions , the data for the alkalies even more scanty than those for acids and salts .
* The only figures that we have been abie to find are those of KANITZ ( ' Zeit .
Phys. Chem 1897 , vol. 22 , pp. 336-357 ) , who gives the folowing table .
No measurements have been made in more concentrated solutions:\mdash ; The results of our measurements , which were made by means of a new type of viscometer , which need not now be described , are set out in the following Table XIV .
, and are represented graphically in Diagram VIIL The relationship between the viscosity and the concentration can be expressed by the formula where is the concentration in gram-molecules per litre , and the values derived from this formula are shown in column 5 of the table .
XIV.\mdash ; Viscosity and Fluidity of Aqueous Solutions of Sodium Hydroxide at C. Concentration in gram-molecules per litre .
Hydration in molecules of water per molecule of sodium hydroxide .
* Data for the conductivity , density , and fluidity of solutions of sodium chloride Srom to between C. and C. are given by LYLB and HOSKING ( Phil. Mag 1902 , vol. 3 , 487-498 ) .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
On correcting the conductivities shown in Table XIII .
for the changes of viscosity shown above , a series of values was obtained for the intrinsic conductivity , and these are shown in contrast with the values for the molecular conductivity in the last column of the same table .
A graphical comparison of the two series of constants is given in Diagram IX .
The values for the intrinsic conductivity show a remarkable increase of conductivity in concentrated solutions which is entirely masked in the specific and molecular conductivities .
The minimum value occurs at about 8 per cent. NaOH , and in a 50-per cent. solution the intrinsic conductivity is nearly half as great again as in the most dilute solutions .
This increase of conductivity we attribute to the fact that sodium hydroxide is an autolyte , that is to say , an electrolyte when fused , and may therefore be able in concentrated solutions to conduct the current quite independently of the ionising properties of the solvent .
Although it was obvious that the conductivity of concentrated aqueous salt solutions might be in part autolytic , we had hitherto been able to obtain no direct experimental evidence in support of this view .
The case of sodium hydroxide is , however , especially favourable for the observations of this property , since its high solubility and low melting-point combine to produce in the concentrated solution a condition approximating unusually closely to that of the fused salt .
Thus the minimum value in the intrinsic conductivity of zinc sulphate , to which attention has been called by SUTHERLAND ( ' Phil. Mag 1902 , VI .
, vol. 3 , p. 171 ) , is reached in a normal solution for which , but the measurements of conductivity and viscosity have only been carried as far as , when the value of rises to This small increase in the intrinsic conductivity might conceivably have been due , not to the autolytic conductivity of the salt , but to some secondary cause , such for instance as over-correcting for the influence of viscosity .
Such an explanation can scarcely be maintained in the case of sodium hydroxide , in which the value of after fallin to at 8 per cent. NaOH , rises again to in a 50-per cent. solution .
These values apply to solutions at C. , a temperature at which the uncorrected specific and molecular conductivities give no indication of autolytic conductivity .
As the temperature rises , however , the extreme viscosity of the concentrated solutions disappears , and the condition of the solutions approximates more and more to that of the fused salt , so that at C. and C. evidence of autolytic conductivity may be seen in the simple isothermal conductivity-concentration curves of Diagram XIII .
These are described in Part VII .
of the papel .
Clear indications having thus been obtained of autolytic conductivity in aqueous solutions of sodium hydroxide , it was not difficult to find other cases in which the abnormal properties of a solution could be referred to the same cause .
This is notably the case with the solutions of zinc sulphate , ammonium sulphate and calcium chloride , for which SUTHERLAND found a minimum value of These minima were not sufficiently pronounced to justify the unsupported assertion that their behaviour was due to the phenomenon now under discussion , but , when placed sida by side with the curve showing the intrinsic conductivity of sodium hydroxide , the behaviour of these salts affords a striking confirmation of the view that the behaviour 01 sodium hydroxide is not unique , but is characteristic of the whole group of autolytes .
Owing to the great ionising power of water , the conductivity of dilute aqueous solutions is usually almost entirely heterolytic .
With non-aqueous solvents it is often quite otherwise , and it is even possible that in certain cases the conductivity of the solution may be wholly autolytic .
Thus it is stated that acetone and formic acid give conducting solutions of the autolytes , KBr and , but not of the heterolytes HC1 , and ( see WALDEN and CENTNERSZWER , 'Zeit .
Phys. Chem , ' 1902 , vol. 39 , p. 564 ) , a result that is most easily explained by supposing that the acetone acts merely as a solvent , but is not able to bring about heterolytic conduction .
The task of distinguishing between autolytic and heterolytic conduction is greatly facilitated by the fact , which is brought out very clearly in the curve showing the intrinsic conductivity of sodium hydroxide at different concentrations , that whilst heterolytic conductivity rises to a maximum in the most dilute solutions , autolytic conductivity is greatest in the most concentrated solutions .
The existence of autolytic conductivity affords , therefore , a simple explanation of the increase of molecular conductivity with concentration that has been not infrequently observed .
Two examples of this kind were recorded by EULER ( ' Zeit .
Phys. Chem 1899 , vol. , p. 622 , who attributed the increase with concentration of the molecular conductivity of sodium bromide and iodide to the influence of the salt in increasing the ionising power of the solvent , that is , to an indirect , rather than to a autolytic influence .
In two of the cases recorded by KAHLENBERG ( ' Journ. Phys. Chem 1903 , vol. 7 , pp. 254-258 ) the molecular conductivity ( of silver nitrate and of ferric chloride in amylamine ) did not increase continuously with the concentration , but reached a maximum and then decreased ; it is clear , however , that in these solutions the final decrease of molecular conductivity may be attributed t , o the great viscosity of the concentrated solutions , to which reference is made in the paper .
It is of interest to note that KABLUKOFF ( ' Phys. Chem 1889 , vol. 4 , p. 429 ) noticed an increase with concentration in the ular conductivity of hydrogen chloride dissolved in ether ; this may perhaps be attributed to the autolytic conductivity of the oxonium salt .
Similar observations have been made more recently by McINTOSH and ARCHIBALD ( ' Roy .
Soc. Proc 1904 , vol. 73 , pp. 454-455 ) and by J. W. WALKER ( ' Trans. Chem. Soc 1904 , vol. 85 , pp. 1082-1098 ) .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
PART VII.\mdash ; INFLUENCE or TEMPERATURE 0N CONDUCTIVITY 0F AQUEOUS SOLUTIONS 0F SODIUM HYDROXIDE .
Amongst the earliest investigations of the influence of temperature on conductivity is that of KOHLRAUSCH ( ' Wied .
Ann 1879 , vol. 6 , pp. 1 and 145 ; 1885 , vol. 26 , p. 161 ) , who measured the linear temperature coefficients of most of the solutions examined , and in several cases determined also the coefficient of in the equation .
He found that the salts gave positive values for , and the acids negative values .
A dilute solution of sodium hydroxide gave a very small positive coefficient , the relationship between temperature and conductivity being very nearly linear over the range of temperature studied .
Subsequent investigations have been made by ARRHENIUS , who discovered that the coefficients for phosphoric and hypophosphous acids become negative above C. and C. respectively , so that the conductivity actually reaches a maximum value , whilst SACH , HOLLAND , and ( ' Ann. Phys. Chem. , ' 1890 , III .
, vol. 41 , pp. 259-287 ) showed that the curves for certain sulphates and for the chlorides of copper and cobalt were inflected at the higher temperatures , a supposed abnormality which they attributed to loss of water of crystallisation by the dissolved salts .
A detailed bibliography of temperature-conductivity measurements is given by KOHLRAUSCH and HOLBORN ( ' Leitvermogen , ' pp. 145-158 , 195-199 ) .
The first indication of any general law governing the temperature coefficients of electrolytic conductivity was obtained by AUSCkI ( ' Sitz .
Preuss .
Akad .
Wiss 1901 , vol. 42 , p. 1026 ) , who found that in the case of the most dilute solutions , and within the somewhat narrow range of temperatures investigated , the two coefficients in the formula were related to one another by the simple equation .
In this way it was shown that if were known , the value of could be deduced without any further experimental observations , so that between C. and C. the whole influence of temperature on conductivity could be expressed by means of a single arbitrary constant .
The relationship between the coefficients may be expressed almost equally well by the equation , and when this value for is substituted in the equation for , it is seen that becomes zero when whatever the value of , in other words , the whole of the parabolic conductivitytemperature curves must , if produced ( without taking account of change of the constants ) , converge to a common point on the axis of temperature , situated at C. below the freezing-point of water .
This point has been named the " " conductivity-zero\ldquo ; for water , and the most remarkable observation recorded in KOHLRAUSCH'S paper is to the effect that the viscosity of water also extrapolates to a limiting value at this point .
It was thus evident that the common factor determining the convergence of the conductivity-temperature curves was simply the viscosity of the solvent .
We are indebted to Dr. LARMOR for the suggestion that this result may be indicated by factorising the equation and writing it in the form , where is a coefficient depending on the ( viscosity of the ) solvent , whilst is a specific constant of the electrolyte under consideration , both and , however , being constant only over a narrow range of temperature .
A fuller study of the influence of temperature on the conductivity of composite electrolytes led to the conclusion that it was possible to represent all the experimental data by means of a generalised conductivity-temperature curve having the form shown in fig. 6 BOUSFIELD LoWRY , .
The most important Fig. 6 .
features of this curve are ( 1 ) a more or less tangential approach A to the axis of temperature , ( 2 ) a point of inflexion , ( 3 ) a maximum conductivity at , and ( 4 ) rapid fall almost to zero in the neighbourhood of the critical temperature of the solution , with possibly a second point of inflexion T. In addition to the evidence brought forward at the time , confirmation of this view has been afforded by the measurements by KUNZ ( ' Comptes Rendus , ' 1902 , vol. 105 , p. 788 ; 'Zeit .
Phys. Chem 1903 , vol. 42 , pp. 591-596 ) of over-cooled aqueous solutions , which show that the conductivity-temperature curves , instead of cutting the axis of temperature as is indicated by the parabolic formula , approach it nearly tangentially ; in the other direction , NoYES and COOLIDGE ( ' Zeit .
Phys. Chem 1903 , Jubelband , p. 372 ) , working with aqueous solutions under high pressures , have shown that even powerful electrolytes like sodium and potassium chlorides reach a maximum conductivity at high temperatures ; in decinormal solutions this occurs at about 28 C. , and above that temperature these salts exhibit negative coefficients .
As regards non-aqueous solutions , the measurements by EVERSHEIM ( ' Ann. Physik , ' 1904 , iii .
, vol. 13 , pp. 492-511 ) of the conductivity of liquid ammonia , of ammonia containing 1 per cent. of copper nitrate , and of sulphur dioxide , have yielded additional evidence of the correctness of the curve at upper temperature of conductivity .
The original object of the investigation described below was to obtain.evidence of the correctness of our view that the point of inflexion was a normal feature of the CONDUCTIVITy OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
conductiyity-temperature curves , and not the product of some abnormal change in the character of the solution .
A series of measurements , made some years ago , had shown that dilute solutions of the alkalies gave inflected curves .
As there was no reason to suspect the existence of any special complications in the case of such simple substances , the conditions appeared to be particularly favourable for the investigation of the nature of the point of inflexion and its dependence on the character of the alkali and the concentration of the solution .
The experiments now described deal exclusively with sodium hydroxide , but the few measurements that we have made with other alkalies have shown that they all behave in a similar manner , as indeed had been indicated by our earlier observations .
Turning now to the specific case of sodium hydroxide , it may be noted that although the linear temperature coefficient of conductivity has frequently been noted , very few measurements have been made of its variation with temperature .
KOHLRAUSCH ( loc. cib , 1885 ) investigated solutions containing per cent. and per cent. NaOH and gave the formulae NaOH 2 per cent. .
He also noticed the great increase of conductivity with temperature in the case of the more concentrated solutions , and recorded an increase of 100-fold in the conductivity of a 40-per cent. solution between C. and C. His figures are given below .
The calculated values in this table are derived from the parabolic formula .
In neither case did the observations come within the range within which a point of inflexion may occur .
DfiGUISNE ( ' Dissertation , Strassburg , ' 1895 ) , working between C. and C. , was also able to represent his measurements of the alkalies by means of a flat parabola , and the existence of a point of inflexion in these curves appears hitherto to have been entirely overlooked .
This result is largely to be attributed to the fact that the point of inflexion for solutions of caustic soda never falls much below C. and recedes above 10 C. both in dilute and in concentrated solutions , and also to the fact that a parabolic curve has been assumed to express universally the temperature-conductivity relation .
VOL. CCIV.\mdash ; A. 2 ExPERIMBlVTAL M ( 1 .
) The Glass Vessels.\mdash ; In measuring the temperature coefficients of conductivity the two chief difficulties were ( 1 ) the exact regulation of the temperature in the higher ranges , and ( 2 ) the fouling of the solutions by the action of the glass of the conductivity vessels .
The earlier observations were made in glass vessels of the form shown in fig. 7 ; the resistance capacity , which was dependent upon the length and diameter of the -tube between the electrodes , was measured with the help of the Fig. 7 . ! .
Fig. 9 .
standard solutions recommended by KOHLRAUSCH , HOLBORN and DIESSELHORST ( ' Wied .
Ann vol. 64 , pp. 440 and 451 , 1898 ) ; the capacities of the vessek used were as follows : cm .
cm .
cm .
cm .
; cm . .
For the lower capacities the -vessel was modified , as shown in fig. 8 ; a vessel of this type was used for the more dilute solutions and had the capacity cm .
( 2 .
) Adjustment of Temperature.\mdash ; These glass vessels were placed in a large water-bath , holding about 100 .
of water , and the conductivity was measured at CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
intervals of C. on the ascending scale and C. on the descending scale of temperatures .
It was found that under these conditions the heating and cooling occupied almost exactly the same period of time , and we were therefore able to apply a simple correction for the loss*of conductivity due to the neutralisation of the alkali by the silica of the glass , by assuming that the fouling took place to an equal extent during the periods 01 heating and cooling .
The temperatures were measured by means of a pair of thermometers graduated in tenths on an open scale and reading from C. to C. and from C. to C. ; these thermometers had been recently calibrated at the Reichsanstalt , Charlottenburg , and the errors were known within C. In the lower ranges an exact adjustment of temperature was possible , but in the higher ranges it was not found possible to avoid fluctuations of several hundredths on either side of the steady temperature at which it was desired to determine the conductivity .
( 3 .
) Determination of Temperature Coefficients with the Help of Platinum \mdash ; During the later part of the investigation we had at our disposal a silver-encased platinum bottle of about 500 cub. centims. capacity , in which we were able to carry out conductivity measurements between C. and 10 C. without appreciable fouling of the solutions .
The bottle and the electrodes used with it had the form shown in fig. 9 .
The resistance capacity was dependent upon the form of the glass tube surrounding the inner electrode ; the electrodes principally used for this series of observations had the capacities and cm . .
The area of glass was much less than in the electrode designed for measurements at and the electrode was lifted out of the solution and allowed to drain when not actually in use .
The solution within the tubes was displaced from time to time by forcing air free from carbon dioxide through the side tube of the electrode .
The smaller neck of the bottle was used to admit a thermometer , by which the actual temperature of the solution could be detelmined ; the four thermometers used for this purpose were gladuated in tenths from C. to C. , C. to C. , C. to C. and C. to C. , and had recently been calibrated at the Reichsallstalt , the errors being given to C. at the lower and C. at the higher temperatures .
The platinum pot , described in Part .
I. of the paper , was also used in a modified form for some of the latest temperature-conductivity measurements .
In these measurements the solution was initially heated to C. , and the conductivity determined at a series of temperatures from C. to C. ; the density of the solution was then determined and its exact concentration deduced ; any change of concentrations that might have been produced by the distillation of water from the solution during the initial heating was thus allowed for .
( 4 .
) Accuracy of the Observations.\mdash ; As has already been pointed out , the chief errors in determining the temperature coefficients of conductivity are ( 1 ) those due to temperature adjustment , and ( 2 ) those due to fouling of the solutions , but in * Tho most dilute solutions gained in conductivity by the dissolution of the glass .
addition to these definite sources of error the whole operation of measuring the conductivity becomes troublesome when the temperature is raised , and it is therefore not possible to work with the same exactitude as at lower temperatures , at which the distraction and haste arising from temperature adjustment is absent .
In dilute solutions , for which the temperature coefficient is about 2 per cent. , an error of C. in the temperature produces an error of per cent. in the conductivity , but in the concentrated solutions the temperature coefficients are enormously greater .
The conductivity of a 50-per cent. solution instead of increasing in the latio of 7 : 2 between C. and C. increases 5-fold between C. and C. , and 80-fold between C. and 10 C. For this reason we are unable to claim for our individual measurements at the higher temperatures an accuracy greater than about 1 part in 100 , though the error on a series of measurements may be considerably less .
The total change of conductivity due to the fouling of the solutions heated and cooled in glass vessels amounted in two instances to as much as per cent. , but the greater part of this is eliminated by our method of correction , and the residual error can scarcely amount to more than per cent. , since the total change in either the ascending or the descending series cannot have differed ffom the half-value , per cent. , that was taken as the basis of our calculations .
When using platinum vessels this source of error was eliminated , but the results obtained were so similar in their general character to those obtained when using glass vessels , that we did not consider it necessary to repeat all our observations in order to secure the increase of accuracy that might have resulted from the removal of only one of the sources of error .
TABLE XV.\mdash ; Influence of Temperature on Conductivity .
NOTE.\mdash ; The values for , and which are enclosed in brackete are derived from conductivity-temperature observations in vessek that had not been independently calibrated ; values of were taken from the isothermal observations recorded in Table XII .
, the concentration of the solutions being deduced from their densities at C. ; the values of , and were then calculated from the ratios tabulated above .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
The general nature of the experimental results may be indicated by giving conductivity-temperature curves for two different concentrations .
In Diagram X. are given ( 1 ) a curve showing the influence of temperature on the molecular conductivity of a normal ( 4-per cent. ) solution ; this curve shows a well-marked point of inflexion at is typical of the curves for moderate dilution .
( 2 ) similar curve for .
a 40-per cent. solution , which is not inflected between C. and 10 C. , and may be regarded as typical of the behaviour of concentrated solutions .
The data on which these curves are based are shown in Table XIX .
, p. 305 .
variation of the temperature of inflexion with the concentration of the solution is shown in Table XV .
and Diagram XI .
, and presents some interesting features .
The minimum value for the temperature of inflexion is at C. in a normal solution , but a rapid rise occurs as the concentration is increased , and at 25 per cent. NaOH the inflexion has risen to C. ; above this concentration the inflexion is no longer observed between C. and 10 C. , but the formulae described below indicate that even at the highest concentrations the temperature of inflexion is only a few degrees above 10 C. The most dilute solutions also appear to give curves which are not inflected , but we are unable at present to supply definite data with reference to the variations in the tempelature of inflexion at concentrations below normal .
In explanation of the great variations in the character of the conductivitytemperature curves , it should be recalled that the point of inflexion is in the neighbourhood where the increase of conductivity due to increasing fluidity begins to be overbalanced by the diminution of conductivity due to decay of ionisation with rising temperature .
Now in the most dilute solutions " " ionisation\ldquo ; is nearly complete , and , as the coefficient of ionisation is thus not greatly influenced by temperature , the form of the curve is determined mainly by changes in the viscosity of the solution .
So also in the most concentrated solutions the alterations of viscosity are so great that any changes in the coefficient of ionisation are only of secondary importance .
It is , therefore , only at intermediate dilutions that the point of inflexion can be observed at temperatures between .
and 10 C. Another way of showing the great differeIlce in the influence of temperature on the conductivity of concentrated and dilute solutions is by calculating the ratios and of the conductivities at C. , C. and 1 C. to those at ; in Table the ratio given is that of the molecular conductivities , since a correction has been introduced for the alteration of volume-concentration produced by the contraction or expansion of the solution .
The figures given in the table are plotted in Diagram XIV .
, and indicate very clearly the more important characteristics of the different solutions .
Having determined the temperature coefficients for a considerable range of dilutions , * In the curve showing the influence of temperature on the specific conductivity ( compare Table XVII .
, p. 303 ) the inflexion is at C. it was of interest to determine the character of the change produced by alterations of temperature in the isothermal conductivity-concentration curve of Diagram VII .
The curves for C. , C. , C. , and C. are shown in Diagram XIII .
, and the data are tabulated in Table XV .
The most striking features are : ( 1 .
) The great increase in the maximum specific conductivity and the gradual shifting of the maximum in the direction of the more concentrated solutions ; this is shown in Table TABLE XVI .
maxi ( 2 .
) A remarkable holding up of the conductivity of the most concentrated solutions .
Thus the conductivity of a 50-per cent. solution at C. is 5 per cent. of the maximum , at C. 24 per cent. , at C. 47 per cent. , and at 10 C. 84 per cent. ( 3 .
) It is noteworthy that the isothermal curves for C. and 10 C. give clear indications that at these temperatures liquid ( over-cooled ) sodium hydroxide would possess a very considerable conductivity .
The power to conduct in the ffised state is necessary , in order to explain the form of the two isothermal curves referred to , and the fact that this effect can already be detected in solutions contaimng less than 50 per cent. NaOH affords very strong confirmatory evidence of the view arrived at from the study of the intrinsic conductivity in Part VI .
of the paper , that in concentrated aqueous solutions the conductivity of sodium hydroxide is to a large extent independent of the ionising power of the solvent , and depends on the self-ionising properties of the liquid alkali .
If reference be made to the curves in Diagram XIII .
, it will be seen that whilst the curve for C. leads down almost to zero value for the conductivity at concentrations above 50 per cent. NaOH , the curves for C. and lo C. cannot be brought down to the axis of concentration , except by introducing an additional point of inflexion in the curve , which there is no reason to suppose exists .
Form of the Curves .
Cubic Formula.\mdash ; Hitherto the conductivity-temperature relations have been expressed exclusively by linear or parabolic formulae , but these are obviously incapable of reresentin a curve containing a of inflexion .
The normal solution of sodium hydroxide , which shows the inflexion very clearly , formed the starting-point of the investigation of the form of the curves which is now described .
The series of values given in Table XVIL having been obtained , a curve was plotted of the approximate values of at different temperatures obtained by CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
gubtracting consecutive values of and dividing by the temperature differences .
The curve so plotted was obviously a parabola with axis at right angles to the axis of temperature and vertex at the value of approximately .
The equation of this parabola was of the which , on integration , gave the conductivity-temperature relation in the form of a simple cubic equation The values of the constants worked out at ; using these constants , the calculated values shown in the table were obtained .
As be seen , they agree with the observed values within the limits of experimental error .
TABLE XVII .
Significant the Conductivity-Temperature Cwrves.\mdash ; The cubic formula gave a satisfactory agreement in the case of those curves which exhibit a point of inflexion between C. and C. , but was not so satisfactory when applied to curves which are not inflected between these limits .
We were , therefore , led to devise a formula which should give expression to the variation with temperature of the three factors of concentration , ionic mobility and ionisation , on which the specific conductivity depends .
Of these only one can be directly determined , na mely , the concentration , which decreases with rising temperature in the same ratio of the density .
In arriving at the significant formula we have assumed , in accordance with the considerations advanced in Part VI .
of the paper , that the changes of ionic mobility may be expressed by a formula similar to that which represents the influence of temperature on fluidity , an assumption which does not depend on the existence of any direct proportionality between mobility and fluidity .
In a similar manner we have assumed that the decay of ionisation may be expressed by a formula similar to that which represents the decrease of specific inductive capacity with rising temperature .
This assumption , also , does not postulate any direct proportionality between the two quantities , but is sufficiently justified by the general relationships to which J. J. and RNST have called attention .
In order to express the variations of ionic mobility we have made use of SLOITE'S fluidity formula : .
where is viscosity , is temperature , are constants , the general applicability of which has been fully demonstrated by THORPE and RODGPR in their Bakerian Lecture Phil. Trans .
To express the variations of ionisation we have adopted the formula which is used by ABEGG and SEITZ to express the relation between the dielectric constant and the temperature for a series of alcohols ( ' Zeit .
Phys. Chem 1899 , vol. 29 , p. 242 ) , and which we find to be applicable also to EVERSHEIM'S values for the dielectric constant of sulphur dioxide almost up to the critical temperature .
Combining these formulae with that given in Part VI .
for the specific conductivity , we get the relation for the variations of specific conductivity with temperature , and for the variations of molecular conductivity It is to be noted that the formulae thus deduced have an empirical value which is independent of the theoretical consideration that led to their construction .
For the purpose of expressing and tabulating the experimental results it was convenient to correct at once for the measured variations of concentration and to work with the molecular rather than with the specific conductivity .
It was further convenient for the purpose of comparing the influence of temperature on the conductivity of solutions of different concentrations to refer the molecular conductivities to those at C. as unity and to tabulate only the ratios , \amp ; .
The exponential factors in the equations given above have been taken to base 10 , and the form generally employed for arithmetical computation was CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Verification of the Significant Formula.\mdash ; In order to test the validity of the formula , two solutions of widely differing concentrations were taken : a normal solution and solution of cent. strength .
The approximate values of the expression were plotted out against temperature and gave the form of a rectangular hyperbola .
This is the relation which we get at once by differentiating the above equation , and it may be written in the form In the two sets of results given below the values of the constants were obtained from this , but subsequently a 11wre convenient method calculating them was devised .
The constants so obtained were : For normal solution For -per cent. solution Using these constants to calculate the values of , we obtain the results which are shown in the subjoined Table XVIII .
and in Diagram XII .
in comparison with the observed results .
TABLE XVIII .
Having regard to the method adopted for calculating the and to the experimental errors introduced by the use of glass vessels , the agreement appears to to be sufficiently close to demonstrate the general validity of the relation .
V0L .
We were thus enabled , for the first time as we believe , to express the relation between the temperature and conductivity of an electrolyte in terms which exhibit the separate effect of the two main factors involved , namely , change of viscosity and change of ionisation .
It now remained to determine the variation of the coefficients over the whole range of concentrations .
In the present communication we have been obliged to limit our observations to the range of concentrations from 50 per cent. down to 4 per cent. ( normal ) , but we hope at some future time to resume the investigation of the more dilute solutions , which promise to yield results at least equal in importance to those derived from the study of the more concentrated solutions .
The values of the coefficients are somewhat egular if derived either by the method described above or by calculation from four particular observations of any series , since small experimental errors might produce much larger errors in the values of the coefficients .
The points chosen for the initial computation of the coefficients were the experimental values for and for each of the concentrations studied .
On plotting out the values so obtained it appeared that the values for and could be very simply expressed in terms of the concentration .
The relation of to the concentration of the solution was , and we were able to adjust the constants in this expression so that when , the index given by THORPE and RODGER for the index in SLOTTE'S formula for the fluidity of water .
This relation , we have reason to believe , holds good right through the range of dilute solutions down to water .
From 50 per cent. nearly down to normal the coefficient is related in alinear manner to the concentration , a result that may be expressed by means of the equation .
In the case of\ldquo ; water\ldquo ; we have reason to believe that there is no appreciable " " decay of ionisation\ldquo ; between C. and 10 C. , so that when we should have .
The rapid drop in the value of , which actually occurs below nol.mal concentration , is not expressed by the above linear relationship , and we have proyisionally added a small term to the linear equation , which is then made to read .
The effect of this added term is to make when .
We believe that further investigation will justify some such addition , which hardly affects the figures for concentrations above normal , and which corresponds roughly to the facts indicated by our incomplete experiments on dilute solutions .
Having thus determined the relations of and to the concentration from a study of all the individual values for , we now took the smooth values for and given by the above expressions , and recalculated the values of from the equation - at , by substituting the experimental values of and , and the smooth values for and .
The same process was repeated using the experimental values of in place of those of .
In this way all the errors of observation , whether from inaccurate reading or fouling of solutions or determinaCONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
tion of density and concentration , together with any errors in the adjustment of the constants for and , are thrown upon the coefficient .
The accuracy of the experimental work , and the validity of the formula for the influence of temperature on conductivity , can then be tested ( 1 ) by comparing the two values of deduced from respectively ( Table XIX ( 2 ) by the smoothness or otherwise of the curve representing the variation of with the concentration of the solution ( Diagram XIV In Table XIX .
are given the values of the coefficients for all the solutions referred to in Table XV .
The values of and are the values derived from the expressions formulated above .
The values for are those derived from and .
It will be observed that in the column of differences for the values , there is a ressive series of positive differences followed by a series of negative differences .
is mainly due to the fact that the slope of the line re .
the variation with concentration of the constant was chosen so as to give the values for at 40 per cent. and at 4 per cent as close as possible to these individual values .
A closer agreement would have resulted if the linear relationship had been adjusted for 40 per cent. and 10 per cent. , but the differences were so small that it was not deemed necessary to repeat the somewhat tedious calculations involved .
The values of the constants are plotted out in Diagram XIV .
, and the smoothness of the individual values for is a satisfactory indication of the validity of the formula employed .
TABLE XIX .
ents in the Equation Physical Meaning of the Coeficients.\mdash ; The formula used to represent the influence of temperature on conductivity is of value not only because it gives us an exact representation of the different types of conductivity-temperature curve , but ako because a definite physical meaning can be assigned to each of the coefficients .
Moreover , the formula has a special value because it renders it possible more or less accurately to resolve the two opposing influences to which the conductivitytemperature curves owe their peculiar form .
Although it is not possible to test , by direct experiment , the accuracy or otherwise of the constants which we have deduced for the temperature variations of the coefficient of ionisation and the ionic mobilities , the constants are of the right order of magnitude for the physical properties which they are intended to represent .
Thus the constant , which we may term the index of mobility , varies from for water to for a 50-per cent. solution of sodium hydroxide .
Although these values were deduced entirely from electrical measurements , it was expected that they would be comparable with the indices in SLOTTE'S viscosity formula , and the values of , given in Table XIX .
, cover almost the whole range of variation of this constant in the viscosity measurements of THORPE and RODGER .
A high constant represents a rapid decrease of viscosity , and may usually be attributed to rapid depolymerisation of the liquid .
It is noteworthy that values of , comparable with those given in our table for concentrated solutions of sodium hydroxide , were observed only in the case of the alcohols : ethyl alcohol , ; isopropyl alcohol , to ; amyl alcohol , ; all of which are known to exist largely in the form of complex molecules .
The constant is related by the expression to the " " conductivity zero\ldquo ; of the solution BOUSFIELD LOWRY , , and although this zero does not represent an actual physical fact , it gives a valuable mental picture of the way in which the increase of viscosity at low temperatures tends to destroy the conductivity of the solution .
The values of are given in the eighth column of Table XIX .
, and correspond closely with the conductivity zeros deduced from a simple parabolic formula for the influence of temperature on conductivity .
The constant , tabulated in the third column of Table XIX .
, measures the rate of decay of ionisation with rising temperature , and is greatest in the most concentrated solutions .
Points of Inflexion.\mdash ; The above expression for the influence of temperature on conductivity enables us to calculate the positions of both the points of inflexion and the maximum of conductivity of our general curve .
The results are given in Table XIX .
, which shows under and the temperatures corresponding to the lower and upper points of inflexion in the curve of molecular conductivity , whilst the temperature of maximum conductivity is given by .
The values of occurring as they do within or close to the experimental range of temperatures , represent actual physical facts , but the values of and can only be regarded as an approximation to the probable behaviour of the solutions at higher temperatures .
CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
The , values of and are deduced as follows .
By differentiating the logarithmic equation we get . . . . . . . .
( 1 ) , where is the modulus A second differentiation gives ( 2 ) .
Putting substituting from ( 1 ) we get , whence ; , and and the points of inflexion are given by For the temperature of maximum molecular conductivity we put in ( 1 ) , and obtain for the value ; so that the temperature of maximum conductivity is the mean between the two tropical temperatures\ldquo ; and Diagram I. of aqueous solutio1ls of sodium hydroxide , with sensitive curves .
( See p. 270 .
) CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Diagram II .
Expansion of aqueous solutions of sodium hydroxide .
Difference curves for and , and coefficients in the equation ( See p. 277 .
) MR. W. R. BOUSFIELD AND DR. T. M. LOWRY ON THE ELECTRICAL Diagram III .
Density-temperature curves for aqueous solutions of sodium hydroxide .
( See p. 280 .
) CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Diagram .
Solution volume of 1 gramme of soda at C. , plotted against the hydration of the solution expressed in molecules of water per molecule of soda .
( See p. 281 .
) VOL. CCIV . .
2 V. Variation of solution volume with concentration at different temperatures .
( See p. 282 .
) CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Diagram iation of solution volumc tcmperature .
( Sce p. 28 MR. W. R. BOUSFIELD AND DR. T. M. LOWRy ON THE ELECTRICAL CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
.317 Diagram VIII .
Fluidity and viscosity of aqueous tions of sodium hydroxide at C. ( See p. 292 .
) Diagram IX .
Molecular and intrinsic conductivity of sodium hydroxide in aqueous solutions .
( Sec p. 293 .
) CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Diagram X. Typical conductivity-temperature curves for concentrated and dilute solutions .
In each case the specific conductivity at C. is taken as unity .
( See p. 301 .
) Diagram XI .
Variation of temperature of inflexion , , and tivity zero , , with concentration .
( See p. 301 .
) Diagram XII .
Influence of temperature on molecular conductivity .
In each case the moIecular conductivity at C. , , is taken as umity .
( See p. 305 .
) CONDUCTIVITY OF SODIUM HYDROXIDE IN AQUEOUS SOLUTION .
Diagram XIII .
Isothermal conductivity-concentration curves for C. , C. , C. and C. ( See p. 302 .
) VOL CCIV.\mdash ; A. 2 XIV .
Coefficients in the equation ( See ) and 302 .
) . .

