The Freezing Point Curves of Concentrated Aqueous Sulfuric Acid1

The Freezing Point Curves of Concentrated Aqueous Sulfuric Acid1. J. E. Kunzler, W. F. Giauque. J. Am. Chem. Soc. , 1952, 74 (21), pp 5271–5274...
2 downloads 0 Views 584KB Size
Nov. 5, 1952

5271

FREEZING P O I N T C U R V E S OF CONCENTRATED kIQUEUUS SULPUKIC . i C I D

causes a significant increase in basicity of alkoxide anions but appears to have little effect on the basicity of amines may also suggest that there is an opposing effect of “B-strain” which would be more important in the latter equilibrium between trivalent and tetravalent nitrogen than in the former, between univalent and divalent oxygen. The Relative Effect of Various “R”’s on the Acidity of Different Functional Groups.-This effect was studied by plotting values of log K (in water)zgfor various RCOaH compounds against the values of log K e for the corresponding RCH20H, RCHOHCH3, ROH, RCONH? and RCONHC6Hj compounds. The values of K , have been suitably corrected. For example, they have been divided by two in the case of symmetrical glycols. In Fig. 2, the RCHzOH plot, the data for the hydroxylcontaining R groups have been ignored in drawing the best line, since it seems likely that the anions of the glycols are stabilized by internal hydrogen bonding and hence that the equilibria measured here are really not of the same type as with the monohydroxy alcohols. It is seen that there is a close relation between the effect of an R group on the acidity of -C02H and on -CHZOH, but that it is not a perfect one. With the -CHOHCH3, -CONHz and -CONHCsHb compounds plotted in Fig. 3, the situation appears to be similar, but here the number of compounds studied is more limited. The ROH plot displayed very little correlation and is not shown. (29) Better correlations may have been obtained Iiy using the ionization c o n s l a n t s in isopropyl alcohul solution, but these were n u t available.

[CON~RIBUTIOS FROM

THE

Effect of Solvent Changes on Relative Acidity.All values of K , are, of course, larger in isopropyl alcohol than in ethanol, but it is interesting to relate the size of the change to the area over which the negative charge is spread in the anion derived from the various acids. From the data in Table V TABLEV EFFECTOF SOLVEXT O N RELATIVE ACIDI~Y K e in Compound

EtOH

i-PrOH

Iic(i-l’rOH) Ke(Et0H)

I,l’-l)initrodiphenylamine 21600 43.4 2,1-L)initrodiphenylamine 19600 26.68” 2,1-Uinitroaniline 2700 4,773“ 0.95 0.058 Ethanol 175 5.ja Glycerol 290 1 .I” 1-Sitrobenzamide From the data of Steariis aiid Whclaiid.11

500 730 570 16 32 310

it is seen that this change is smallest with ethanol, where the negative charge is presumably almost entirely on one oxygen atom. It is somewhat larger with glycerol, perhaps indicating that the charge is somewhat dispersed by the hydrogen bonding of an unreacted hydroxyl group to the alkoxide oxygen atom. The largest changes are found with compounds in which the charge may be spread over several atoms by resonance. These data would be expected from the greater ability of ethanol to solvate the anions with more concentrated charges, which are susceptible to stronger solvation. ATLANTA, GEORGIA

CHEMICAL LABORATORY OF THE UNIVERSITY OF CALIFORNIA 1

The Freezing Point Curves of Concentrated Aqueous Sulfuric Acid’ BY J. E. KUNZLER AND W. F. GIAUQUE RECEIVEDMAY26, 1952 The freezing point composition curves for aqueous sulfuric acid have been investigated over the range H?SO4,2H20-

H2S04.H20 eutectic t o somewhat on the SO8 side of HzSOa. The regions near the melting points of H Z S Oand ~ HZSOI.H?O have been considered in detail so t h a t the measurements can be used for calculating the change of free energy with composition. Evidence has been presented t o show that maximum freezing sulfuric acid corresponds to absolute sulfuric acid within a few thousandths of a wt. 70. On this basis we believe t h a t maximum freezing sulfuric acid is a more accurate standard of acidimetry than can be prepared by any other known method. I t has been shown that the relative compositions of maximum freezing anhydrous sulfuric acid and the maximum freezing monohydrate correspond t o the theoretical H2S04:H?SO,.H20 ratio t o within 0.01 wt. % H2SOa. Thus maximum freezing H2SO,.H20 is a pure compound within 0.01 wt. % HzSOa. The melting points of H2S04 and H&04*H20were found t o be 10.371 and 8.489’, respectively.

Work in progress in this Laboratory on the thermodynamic properties of sulfuric acid requires the free energy of dilution. This can be determined from the freezing point curves of the several hydrates. Gable, Betz and Maron2 have recently published an excellent phase equilibrium diagram of the sulfur trioxide-water system. However, these authors did not contemplate the use of their data for the determination of free energy from the freezing point lowering. A calculation of this type requires numerous precise values near to the (1) This work wa5 iupported in part by the Office of Kdval Research United States S a v y (2) C \IGablc H r Betr and b €1 X I u o t i , l m b J O U K K A I 72, I l l 2 (1030)

melting points of the several pure phases. I t was found desirable to supplement the data of G., B. and M. by very numerous measurements near the melting points of the anhydrous acid and the monohydrate. At the same time measurements of increased accuracy were obtained over the range from the HzS04.2H20-HzS04.HzO eutectic to somewhat on the SOaside of the anhydrous acid. Apparatus.-The measurements were made in a one-liter dewar which had a long narrow neck. I t was submerged, except for about 2 in. of the neck, in a one-gallon dewar of crushed ice. The sainple dewar was equipped with a motor driven stirrer and the tops of both dewars were protected by Bakelite covers. Several small openings, with removable glass sleeves, were provided for the stirrer shaft, sampling, ;ddiiig iiicremeiits of thc solution coinpoiieiits, and for a tcti’

junction copper-constautail therinocouple. The acid was The therinocouple was coiiiparcd with a strain-frec platipurified by distillatio~i.~Samples were taken with a pipet num resistancc thcrinotneter calibrated at the U . S. Bureau which withdrew solution by means of an evacuated bulb of Staiidards. which could be opened to the pipet by means of a stopcock. The observations are given in Table I in chroiiological The pipet has been described by Kunzler.3 Its special order. featurc is an extension, in the form of it small chamber endTABLE I itig iii a tip, attached to the pipet by nicans of a t:tpered ground glass joint. A sintered glass filter disk, with an FREEZISC,POIXTLOWERING I N AQUEOUSSULPURIC ACID avenlge porc size of about 10 microns to csclutic the iinc @ = freeziiig poiiit lowering acid slurry, is sealcti across the iiiiddlc of the chuiiber. TVt. '0 I )uring the pcriod bcfore cquilibriuin, a very small stream of HzSO, 0 , oc. iiitrogen bubbles prevented entry of solution so th;tt the !10.288 16.07 4iitered glass filter could not be containinatcd b!. liquitl. 89.!)26 14.07 Son-equilibrium liquid can bc produced bj- the melting of 80.482 11.7!) solid when a pipet a t 25' is introduced, but any such liquid 88.8!)0 9 10 11zSOd.II20 crystals could not enter or be produced in the pipet as used here. S8.42 0 7 . 21 added Method.-The dctermiiiatioii of free energy from freezing point lowering measurcinents near the melting point ol the 76,643 26.51 Serics ,5 acid, or that of onc of its hydrates, rcquirch it very accurate i6.103 L'O.53 S7.718 z.01 knowledge of the conipositioii in this range. \\~ithdra\virig 73.990 30.9!1 8 7 , 3:1X 0 . 80 sanli)lcs corresponding to cach tempcraturc o1,scrv;itioti is 7 5 .786 3:' 8fi 8!1li 2 ,7I _.I,rborious aiitl uiiticccasarily inaccuratc. The iiictliotl tiset1 I J 2.3!! : { 3 87 8ti. 54 2 1 . !IS.$ ih soincwlmt similar to thosc usctl Iiy I ianiinctt a i i r l D C J T U1~ ) 8Ii.Trl1 I . ! !I a i d Cillcspie,6 iri that thc variation of coiiipositioii \ v a s o b 8!i 1 1 3 I. taincd by addition of wcighed amounts of niatcrial. 8;).:!I1 I) 7!l7 A known weight of some 2300 g. of acid, a little on thc ,188 , I!il SO2 side of I I ? S 0 4 , was placed in the dewar. The coinposi. 180 . 23:j tion was kiiown only approximately. About 20'% of the Series >.a acid was in the form of solid at the beginning but t h e :tinoutit was known only w r y roughly. Thus iicithcr the ovcr-all 84,!l09 0.0871 composition iior the compositiou of the liquid portion werc 84.89; ,0821 known a t the start. 8t.880 ,0762 8 4 . 828 ,0362 \Vater was added in the forin of small iiicrements of aque84.775 .01108 ous acid, which was shown b>- analysis to havc a composi84.723 ,0277 tion of 96.20 i 0.01%. This reduced heat introduction 8 4 . 6 7 2 .0l68 and gave increased accuracy with respect to the amount of 84, A 2 0 ,0087 water added. S e a r the melting point the temperature 81,,570 ,0033 differencc was read to 0.0001", and nxas aceuratc to 0.000i'", 8 ) 518 ,000.1 after cach addition. The laboratory i n which thc experi81. '$6') ,0002 inelit was performed was thcrinostated a t 2 3 " . The incre81.I I ' I .(IO20 ments of 96.20% acid were a t 25' when added and thus a t 81,::71 .002!i k n o \ \ ~ iheat content. The heat of stirring plus heat lcak 8% :323 .011: could be evaluated from the rate of change of melting point 81 %i(i 020-I with tinic ovcr long periods. This caii be determined most accurately near the beginning and eiid of such :i series w1iei.c Series 5 fcoiit ) the ratc of change of temperature with the ;tinoutit of solid s I . 00G 0 , 100% melted is grcateit. The dewar vesscl ivaq treated :t< a cal83.132 0.4!)7 orimeter and a n accurate record of it5 change iii heat coritcirt 82.821 1.23Ri was recorded from the beginning to the ciid of it serics of 82.412 1.022 observations. When the cornposition of the acid had 81,;s:; :3,241 reached about 99.79yo, four samples of the cquilibriuiii liquid werc withdrawn and analyzed to within 0 . 0 0 5 ~ o . The data were treated bl- plotting a curve of iriclting point against amount of 96.20% acid adticti. At the nielting point maximum the composition corresponded to purc H&O, and this could be determincd to 0.0002% acid. 80.154 8.31 Since the amount of excess water between the melting point 7!!.446 11.19 iuaximum and the end of the scries was known to about NzSOI.HZOcrystals 0.4%, the amount of solid a t the end of the series could be added calculated. The heat of mixing of 96.209/0 acid in thc variSeries 6 (cont.) ous solutions was available from the paftial specific heat content measurements of Kunzler and G ~ a u q u e although ,~ 78.921 13.63 the calculations required one approximation before the final " I:.]). \lightly high bccausc amount of solici was I I J U result was obtained. Since the heat content \vas kiiorvn for each temperature mcasurenicnt, the amount of solid and S c w inaterial was used a t thc start of cach iiuiiibercti thus the composition of the liquid could bc calculated. The heat of fusion of H2S09,2560 cal. iiiole - I , was takcii from scries ill Table I . Some 30 to 10% of the sample was prcsciit as the equilibrium solid in the form of a slurry of fine crysthe iucasurenients of Rubin aut1 Giauqtrc.' The heat quantities were known with sullicicnt accuracy tals which were thoroughly and continuously stirred. Serics 1 and 2 were measured in the order of descending teinpera80 that the accuracy at each point was essentially not influenccd by the calculation of thc various :tniouiits of solid tures by adding 96.20% sulfuric acid. Series 3 , 4 and thc first part of Series 5 were measured in order of ascending during the experiment. I n regions farther removed from the melting points, the temperatures to take advantage of the heat capacity of the temperatures were observed, and samples removed for system. In Series 3, acid with about 3070 excess SO3 was used to increase the concentration, and finely divided crysanalysis t o 0.005Yo. - ... . . tals of H2S04, cooled by means of solid carbon dioxide, were added where noted. I n Series 4 and 5, except 5a, ice at (3) J. E . Kunzler. work to be published (4) I.. P. H a r u m r t t artri A, J Deyrup. THISJ U I J H K A L , 66, I!lOO 0' was used for dilution, and on the one occasion noted, (I!)?3J crystals of H&04.H20cooled by stoiing the container in solid carbon dioxide were added. In Series 6 and 7 , dilu( 5 ) R.J. Cillespie, J . C h e m .Soc , 21!?3 (1930). ( 6 ) J. E. Kurizlcr atid IV. F. Ci8irilue. TEIWJ o r i R N n l . . 7 4 , 3172 tion was with ice a t 0' or a t liquid nitrogen teinperaturc as (l!)52). requirctl. \Vhrii it wits necessary to add crystals of €I&OI. 1 1 ~ or 0 I I2SOI tlicsc wcrc adjustctl approximately to solutioii $ 7 ) T. IC. Rtibiri atid IV. 1 : . Gi8iit~iiie,i h L ,7 4 , 800 (I!J52J.

temperature before addition. The portion of Series 5 TABLE I1 designated 5a and Series 8 covering the regions near the compounds H&O4.H20 and H ~ S O I ,respectively, utilized MELTINGP O I N T DATA ON H,SOd CONTAININGA SMALL the calorimetric accounting method of determining concenAMOUNTO F HzS207 trations as described above. The dilution in Series 5a was -A, by means of water at 25" and in Series 8, 96.20% acid at 25' was used. Equilibrium-One of the sources of error in freezing point ~~. . measurements depends on the extent to which equilibrium couple therm. Av. liquid melted ih-f, is attained between the solid phase and the bulk liquid which 0.29 0.29 0.29 0.0052 0.15 0,0008 is analyzed. Even when heat leak is reduced to a minimum there will be a small amount of heat developed by the stir,145 .15 ,155 .0032 .22 .0007 ring which is essential to the attainment of uniform concen,128 ,128 ,127 ,0029 .30 .0009 tration. Thus prior t o and during the period of observa,079 ,080 ,081 ,0021 .37 .0008 tion there will be a small amount of crystal melting. In ,058 .om . o m .ooi7 .44 . o m 7 these circumstances one cannot escape the conclusion t h a t the concentration in the immediate vicinity of the melting ,000 .053 ,046 .0016 .51 .0008 crystals will be rich with respect to the liquid resulting from .050 ,050 .050 .0016 .58 .0009 the solid phase. I n a n y two-component system this must raise the temperature above the value which would result Average 0.0008 =k 0,0001 from true equilibrium between the solid and the bulk liquid. The present measurements were made with a full appreciaportant to realize t h a t the temperature composition data tion of this difficulty and the conditions were such as to reused as a basis for treating the data of R. and G. were obduce it to a minimum. The finely ground crystals presented tained by a method in which the amount of the solid changed it very large surface. After additions of liquid, equilibrium by only 1.6% during the investigation which has been menwas established within two minutes a t the higher temperatures and within about 10 minutes at the lower tempera- tioned above. I t is also of interest to note that the actures. When cold solids, such as ice at liquid nitrogen tem- tivity of H~S201in the solution with 15% melted is about peratures, were added, a long period was necessary to bring twice t h a t when 58% is melted. This ratio is readily calthem into equilibrium. Aside from this reason for slow- culable from the freezing point curve. We believe that the ness, equilibrium was reached rapidly even in the more con- above facts support the conclusion t h a t H&07 is very incentrated regions and at lower temperatures where increased soluble in solid absolute sulfuric acid. There is no trend in the amount of H2S207 shown in the last column of Table viscosity would reduce the rate of diffusion. Comparison of Maximum Freezing Sulfuric Acid and 11. Rubin and Giauque7 made measurements of heat capacity Sulfuric Acid Monohydrate.-During the course of the present work the sulfuric acid content of maximum freezing below the melting point and evaluated the premelting heat sulfuric acid and its monohydrate were compared. The calorimetrically. They give the result A = -0.0000 as freezing point curve of the monohydrate is much flatter than compared to -0.0008 in Table 11. They expressed the t h a t of the anhydrous acid but the maximum was located a t opinion that their results were low because insufficient 8.489" within a composition accuracy of 0.003 wt. 70. time was allowed for equilibrium. Such a situation will was determined to a n result in a low concentration of impurity a t the melting The maximum at 10.371" for accuracy of 0.0002 wt. %. The comparison showed that boundary and require a long time for equilibrium by diffuthe two maxima were in agreement within 0.005% and i t is sion and heat conduction The results in Table I1 were based on the two-hour equilibrium period mentioned above conservative to assume that the total possible error is withiti and if a substantial part of this time had been used for 0.01%. Evidence that H2S04 Crystallizes as Absolute Acid to a equilibrium in the premelting heat measurements they High Degree of Accuracy.-LichtyS mixed equimolal would probably have given agreement. Rubin and Giauque? also estimated the amount of H & 0 7 amounts of sulfur trioxide and water and proved t h a t the impurity present in the liquid remaining a t the H&O4-H2maximum melting point corresponded to absolute acid S207 eutectic point as 0.0010 mole of H&O7 per mole of within 0.01 wt. %. Rubin and Giauque' made a considerable number of un- HaSO4 in the calorimeter. This estimate was based on a published determinations of the melting point of sulfuric small amount of eutectic melting and is somewhat uncertain acid as a function of the fraction melted. There was a due to the fact t h a t the heat of fusion of the H2S207 phase is small excess of H~S207in their sample. These measurements known only roughly. The value 0.0010 may be taken as were made after waiting long periods, of about two hours, comparable with the more reliable value of 0.0008 given in Table 11. It may be shown that the activity of H&O, for equilibrium. They were not used by R . and G . to inincreases by a factor of about 150 between the melting point vestigate the amount of impurity because i t was not possible to estimate the freezing point lowering with sufficient of HzSO, and the eutectic and despite this large increase the accuracy, and also because related detailed measurements results indicate that the H2S04eutectic solid phase does not contain any appreciable amount of H&07. I n reaching as a function of composition were unavailable. The needed this conclusion it is not assumed that the entire solid woultl observations were obtained during the present research. In order to ensure uniformity in the temperature scales of continuously reconstitute itself along the freezing point the two researches, which is particularly important when the curve. Even the assumption that a solid built up of layers freezing point lowering, e, is the objective, the melting point deposited in equilibrium with the liquid during cooling of sulfuric acid monohydrate, 281.649'K., was used as an should deplete the total available amount of 0.0008 mole H ~ S Z Oby ? a n appreciable amount if much solubility existed. intermediate temperature reference. This melting point is not very sensitive to composition and thus could be de- We believe t h a t the data as a whole indicate that maximum termined rather accurately. We have applied corrections melting sulfuric acid corresponds to absolute H?SOI within t o the melting point data of R . and G. to make them as a few thousandths of a per cent. and that this procedure for preparing absolute acid is more accurate than any other accurate as possible for the present purpose. The results are given in Table 11. The quantity - A , representing the known method. This is important in the present series of moles deficiency of HnO or the excess SO3, per mole of HzSO4 researches on sulfuric acid in this Laboratory since analytical procedures have been based on this standard. The tleis calculated provisionally on the assumption t h a t the maximum melting point corresponds t o pure sulfuric acid. It tails of these procedures will be published elsewhere. 3 The fact that maximum melting sulfuric acid nionohyshould perhaps be emphasized that ideal solution laws were drate and maximum melting sulfuric acid have relative not assumed in a n y of these calculations. It will be seen that the total amount of H&07 present in compositions which agree with the ideal composition ratio the liquid does not change when the amou'nt of the solid for these compounds, taken with the fact that the latter is melted changes from 15 t o 58%. It is evident t h a t the pure, also demonstrates that maximum melting H?S04.Hr0 solid does not have a n appreciable amount of HzS201 to con- corresponds to the pure compound with high accuracy. tribute to the liquid produced by the melting. It is imComparison with Other Data.--Gillespieb cov~

( 8 ) D. M Lichty, THISJOURNAL, SO, 1845 (1908).

ered the range 99.7% to 100.4% H$04 and Ham-

rnett and Deyrup' covered approximately the same rmgc. Both of these researches were put ( 1 1 1 '111 ,ib~olutchas15 h v rcfrrcncc t o thr i r i r i \ i i i i i i i i i t rcc/irig Iioiiit. (hllcsl)~c c,tlciil,itccl Ihc rLit't ob I ) o t l i researches to this b,isis for cotii~~arisoii. -1 wide investigation of the freezing point coiiiposition diagram has been made by several observers but, except near the melting point of mhydrous acid, their work IS less accurate than that oi Chble, Retz and hlaron? mho g i w references to most of the previous work. I n the immediate vicinity [99.% to 100.0.5wt. of absolute acid the present research is more accurate than those of 1%. and D. or C,. Between 00.7 and 99.9 wt. c( H2S04the present values of 0 lie about midway between the two previous researches in this region. 8 is smaller in the work of Gillespie and larger in the work of Hammett and Deyrup. Curves through the results of either observer lie just outside the estimated limit of error of the present results. There is essentially no error in the teniperature measurement but the possible analytical error of +0.005 wt. 5; a t 99.8L?, IS equivalent to 0.0:3@. The niethod used reduces error proportionally between the point of analysis and the inaxiniutn melting point so that a t 1)9.9("b the possible error WAS .t0.002.? wt. corresponding to about 0.01 2'". The accuracy of the present work could have been iniproved by performirig the :rnalysis on a sample taken a t a greater 8. Gillespic gives 1 0 :W ;is the melting point of

[COSTRIDTJTIOS FROM

absolute acid. The value obtained in the present rcsenrch is 10.371" (283.531 "K.j. I'hr Ijreseii 1 rcwlts diffrr c.oxisiderably Ircun thosf of (hhle, Bet/ antl M'iroii, cspeci'tlly a t tlic lowcr teiiiper,iturcs betweeii H2SO4 and IiaS04.Hs0 'l'hcir results are high, soinetinies by as much as 2" There is abundant evidence froiii melting and other fixed points i n other concentrations, even a t lower temperaturcs, to show that their temperature measurements are substantially correct. Some of the difference could be attributed to errors in sampling or analysis but it is not possible to explain the difference in that way-. The results are high 011 both sides of the eutectic, whereas systematic analytical errors would be expected to raise the results on one side and lower them on the other side of the H?SOd-HzSOI HzO eutectic. The only self consistent way of explaining the difference is in terms of unattaiiied equilibrium in the boundary region about the melting crystals. The viscosity of sulfuric acid solutions is very high in this concentrated region which requires minimum heat leak, large crystal surface and thorough stirring to approach equilibrium. The results of G., B. and M. on the dihydrate side of the monohydrate agree much better and in the more dilute regions, beyond the rnono-dieutectic, it is expected that the results of G., B. and M. will suffice for the free energy calculations which will he presented in a later paper. RERKEI E \ ,

HARPURCOI.I.EGl?,9 r A T E

c.il I T O R \ I 9

t!SIVERSITY

0 1 7 XBTY

Y(IRK 1

The Solubilities of Naphthalene and Biphenyl in Aqueous Solutions of Electrolytes Bs MARTINA. PAUL RECEIVED M . ~ Y13, 19.2 The solubilities of naphthalene and Iiiphenyl have been determined in aqueous solutions of several strong electrolytes a t 25' Iiy means of ultraviolet absorption. The data fit the Setchenow equation, in which the logarithm of the non-electrolyte's activity coefficient is expressed as a linear function of the concentration of the particular electrolyte. The effects of different electrolytes are generally similar t o those reported for other non-polar solutes. Comparison with published data for benzene shows a progressive increase in the magnittide of the salting-out parameter for a given electrolyte from benzene to naphthalene to biphenyl, apparently accompanying the increase in the molal liquid volume of the non-polar solute. The results nrr discussed in terms of a theorv Dronosed hv 1IcT)evit antl Long in connection with their data for benzene. Data are reimrtrrl ?Iw for ~inphthaleneat 0'

The influence of various electrolytes on the solubility of benzene in aqueous solutions has recently been investigated by SlcDevit and Long1 and by Saylor, IYhitten, Claiborne and Gross. 1lcDevit arid Long have proposed a theory according to which the non-polar solute has the effect of modifying the interaction between the solvent water molecules and the ions of the electrolyte to a first order of approximation merely by occupying \ olume. The salting-out parameter for a given electrolyte acting on a non-polar non-electrolyte should then be proportional in a first approxiination t o the non-electrolyte's own molal volume (in the liquid state), and for different electrolytes should f

(1) n' 1 I [IX)

\rcIl?iit

Jnd I

4 Ion:

'IIII>

J o i ~ i 4 1

74, 1773

(2) J 11 Saylor, 4 I Wliittrn T ClAiImrne w d P I\.I Gro59 rbid , 74, 1778 (10521

depend on the relative extents of interaction between solvent and ions, as measured for example by the contraction in volume accompanying dissolution of the electrolyte. This theory has the nierit of not only reproducing correctly the order of effectiveness of various electrolytes in salting out a particular non-polar solute but also accounting for the salting in observed with perchloric acid and certain other electrolytes having large anions or cations. It seemed desirable to obtain inforination for the larger non-polar molecules o€ naphthalene and biphenyl. For polar non-electrolytes, of course, interactions of a more specific nature niay be expected between the neutral solute niolecules and the ions on the one hand and the solvent iiiolecules on the other, and this situation is indeed indicated by the nature of the experimental results for such siibstances when contrasted with certain