The Effect of Substituents on the Acid Strength of Benzoic Acid. III. In

John H. Elliott, and Martin Kilpatrick. J. Phys. Chem. , 1941, 45 (3), pp 472–485. DOI: 10.1021/j150408a013. Publication Date: March 1941. ACS Legac...
0 downloads 0 Views 629KB Size
4i2

JOHN H. ELLIOTT AND MARTIN KILPhTRICK

T H E EFFECT OF SUBSTITUENTS ON T H E ACID STRENGTH O F BENZOIC ACID. 1111 IN

ETHYLENE GLYCOLAND n-BuTYL ALconoL

JOHS H. ELLIOTT A N D MARTIK KILPATRICK Department of Chemzstry and Chemical Engineering, University of Pennsylvania, Philadelphaa, Pennsylvania Received August 20, 1940

This paper is a report of the potentiometric determination of the relative acid strengths of the substituted benzoic acids in the solvents ethylene glycol and n-butyl alcohol. The paper also includes a discussion of the results presented for the solvents methyl alcohol (10) and ethyl alcohol (11), together with the results available for aqueous solution (5). EXPERIMENTAL PART

The experimental method has already been described (10) and the only difference noted in the application of the method was that longer times were necessary for the establishment of electrode equilibrium in n-butyl alcohol. In addition, an electrode would occasionally become poisoned and have to be replaced. The poisoning of an electrode showed itself by an appreciable potential difference between two electrodes in the same solution, the E.M.F. usually changing with time. It was found that gold-on-glass electrodes prepared by the method of Xewberry (23) gave more reproducible results in n-butyl alcohol, since these electrodes could be cleaned by heating to redness in the flame of an alcohol lamp. All measurements were carried out a t 25°C. Ethylene glycol was purified by the method of ,;kerlof (1). The glycol was allowed to stand over powdered, freshly calcined calcium oxide for several days. The calcium oxide was removed by filtration and the ethylene glycol distilled through a Widmer column. The boiling point for each distillation was very constant. The distillation temperature was 90.0"C. a t 6 mm. of mercury and 95.5OC. a t 11 mm. The index of refraction was nko' = 1.4321. The purity, calculated from a density determination on the assumption that the only impurity was water (20), was 99.93 per cent. n-Butyl alcohol was purified by the method of Mason and Kilpatrick 1 This paper was abstracted from the dissertation presented by John Habersham Elliott to the Faculty of the Graduate School of the University of Pennsylvania in partial fulfilment of the requirements for the degree of Doctor of Philosophy, April, 1940. I t was read a t the Xinety-ninth Meeting of the American Chemical Society, which was held in Cincinnati, Ohio, April, 1940. The authors take this opportunity t o thank Dr. F. H. Westheimer for tables of the various functions employed in the calculation of D E .

EFFECT O F SUBSTITUENTS ON .4CID STRENGTH

4 73

(21). Two liters of alcohol were refluxed with 5 g. of sulfuric acid, distilled, and the middle fraction collected. The product was refluxed with stick sodium hydroxide for several hours, then freshly calcined calcium oxide was added and the mixture refluxed for 8 hr. The alcohol was then distilled in a Snyder column, and the middle fraction with a boiling range of 0.1"C. was collected. The index of refraction was found to be n:" = 1.3976 to 1.3977; the value reported in the literature is 1.3971 (1). S o difficulty was experienced in the preparation of lithium butylate; this is in disagreement with the findings of Wooten and Hammett (28) but in agreement with those of Mason and Kilpatrick (21). Table 1 summarizes the equilibrium constants KA,Bo, obtained in nbutyl alcohol. for the reaction

A,

+ Bo

Ao

+ B,

where the subscript x represents the substituted benzoic acid and the subscript zero represents benzoic acid. For comparison we have included, in column 3, results obtained by measurement of similar concentration cells (28), together with a few colorimetric determinations at low ionic strengths. The agreement between the potentiometric and colorimetric determinations is good. I n the last column are given the negative logarithms of the dissociation constants of the substituted benzoic acids. These were obtained by determining the dissociation constant of benzoic acid by reference t o a suitable solution of hydrogen chloride in n-butyl alcohol and computing the other values from log given in column 2. For benzoic acid pK, was found to be 8.609 at p = 0.05 (mostly lithium chloride). Table 2 gives the determinations in ethylene glycol of KAzBo,the equilibrium constant of equation 1. Table 3 gives the classical dissociation constant of the acids in ethylene glycol containing the solvent salt lithium chloride at a molar concentration of 0.045. There are no results in the literature for comparison. The measure of precision in the determinations in ethylene glycol is estimated as 0.010 unit in log KAzBo,as compared with 0.013 unit in n-butyl alcohol. DISCUSSION OF RESULTS

I n this and the other papers of this series the acid strengths relative t o benzoic acid, and t o the solvated proton of the solvent, have been reported. An examination of the results shows that the order of decreasing acid strengths is not independent of the solvent. In the process A

+S

SH'

+B

(2)

Wynne-Jones (29) considers that (1) the electrostatic action between SH' and B depends upon the solvent, ( 2 ) SH' is a different entity in each different basic solvent, and (3) the chemical potentials of the acid

474

JOHN H

. ELLIOTT AND

MARTIN ICILPATRICK

and base may change differently with change in solvent; for example. W e r e n t types of solvation of A and B may occur . Bjerrum and Lanson (2) suggest that. since A and B differ from each other only by a proton. TABLE 1 Relative acid strengths i n n-butyl alcohol at W C.

Ortho-substituted benzoic acids

1.74. I. . . . . . . . . . . . . . . . . . . . . . . . . Br . . . . . . . . . . . . . . . . . . . . . . .

c1........................

1.038 1.090 1.075

F ........................ CHs . . . . . . . . . . . . . . . . . . . . . .

0.926 0.002

OCHs .................... OH .......................

0.281 1.505

1.00' l.llb 0.02' 0.04b 0.31. 1.508

6.829 7.571 7.519 7.534 7.683 8.607 8.328 7.104

Meta-substituted benzoic acids

NO1 .....................

1.100

I. . . . . . . . . . . . . . . . . . . . . . . . Br . . . . . . . . . . . . . . . . . . . . . .

0.566 0.578

c1....................... F. . . . . . . . . . . . . . . . . . . . . . .

0.585

0.60'

0.416 -0.095 -0.159

-0.09.

CHa ..................... OH ......................

1.098 1.22b

7.509 8.043 8.031 8.024 8.193 8.704 8.768

Para-substituted benzoic acids

NO1 . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . .. . . . . . Br . . . . . . . . . . . . . . . . . . . . . . .

c1. . . . . . . . . . . . . . . . . . . . .. .

F. . . . . . . . . . . . . . . . . . . . . . CHa . . . . . . . . . . . . . .. . . . . . . . OCHa . . . . . . . . . . . . . . . . . . . OH . . . . . . . . . . . . . . .. . . . . .

1.141 0.396 0.421 0.394 0.215 -0.193 -0.360

1.12' 0.42. -0.218 -0.40'

7.468 8.213 8.188 8.215 8.394 8.802 8.969 9.175

* Wooten and Hammett (28).p = 0.0525. Mason and Kilpatrick (21).p < 0.001.

their changes in chemical potential upon change of solvent may be set equal . Wynne-J-ones follows this suggestion and then eliminates the solvent and the solvated proton by considering strengths measured relative to some standard acid Ao. as in equation 1. The effect of the dielectric

475

EFFECT OF SUBSTITUENTS ON ACID STRENGTH

constant of the medium may be approximated by the Born equation (18) and if A,, any acid, and Ao, the standard acid, have the same charge we have

TABLE 2 Relative acid strengths i n ethylene glycol at 86'C. u = 0.050

BWBSTITUENT

I Ortho

NO1...................... I. ........................ Br. . . . . . . . . . . . . . . . . . . . . . . Cl . . . . . . . . . . . . . . . . . . . . . . . . F........................ CHs. . . . . . . . . . . . . . . . . . . . . . OCHa. . . . . . . . . . . . . . . . . . . . OH. ......................

1

1.741 1.105 1.198 1.137 0.689 0.046 0.179 1.496

mBta

0.931 0.490 0.544

0.519 0.448 -0.090 -0.034

ParS

0.965 0.357 0.368 0.304 0.169 -0.172 -0.324 -0.448

PKC SWSTITWENT

NOa ...................... I. . . . . . . . . . . . . . . . . . . . . . . . . Br ........................ c1. .......................

F ........................ CHs. ...................... OCHs. . . . . . . . . . . . . . . . . . . . OH. . . . . . . . . . . . . . . . . . . . . . .

5.906 Ortho

::E 6.510 6.958 7.601 7.468 6.151

~

~

1

1

Meta

ParS

6.716 7.157 7.103 7.128 7.199 7.737

6.682 7.290 7.n9 7.343 7.478 7.819 7.971 8.095

7.681

where rA0and are the effective radii of the acids, D is the dielectric constant, and ZA is the charge on the acid. Other developments (13, 14) also yield a relationship where In K A ~ Bis. linear with 1/D. The linear

476

JOHN H. ELLIOTT AND MARTIN KILPATRICK

relationship has been shown to hold for D > 25 by Minnick and Kilpatrick (22) and Kilpatrick and Mears (17). In both of these investigations KAzBawas obtained a t infinite dilution, and the dielectric constant of the solvent was used in equation 3. In this work K A ~was B ~determined in a medium 0.045 molar in lithium chloride and 0.005 molar in the anion of the, benzoic acid. However, since the values of K.~.B~ seem to be independent of the electrolyte concentration, and since the theory predicts that the change of the dielectic constant of a solvent with added electrolyte is small (12), the dielectric constant of the solvent has been em-

FIG. 1. Effect of substituents upon t h e acid strength of benzoic acid. Curve I, o-nitrobenxoic acid; curve 11, m-nitrobenzoic acid; curve 111, p-nitrobenzoic acid; curve IV, o-bromobenzoic acid; curve V, m-bromobenzoic acid; curve VI, p-bromobenzoic acid.

ployed in testing equation 3. The results are presented graphically in figures 1, 2, 3, and 4 from the data summarized in tables 1 and 4. In agreement with our earlier findings, the linear relationship holds from D = 78 to D = 25 and, with the exception of the ortho-substituted acids in ethylene glycol, the agreement between the calculated and observed values is good, the differences being less than 0.02 log unit in most cases. The extrapolation to 1 / D = 0 yields log Kazbo,the intrinsic acid strength, which should be independent of the solvent. The failure of this relationship in media of dielectric constant lower than 25 may be due to the fact that in media of low dielectric constant dipole-dipole interactions between the molecules and the solvent play a more important r6le. The failure

EFFECT OF SUBSTITUENTS O N ACID STRENGTH

477

//D FIG.12. Effect of substituents upon the acid strength of benzoic acid. Curve I, o-iodobenzoic acid; curve 11, m-iodobenzoic acid; curve 111, p-iodobenzoic acid; curve IV, o-methylbenzoic acid; curve V, m-methylbenzoic acid; curve VI, p-methylbenzoic acid.

FIG.3. Effect of substituents upon the acid strength of benzoic acid. Curve I, o-fluorobenzoic acid; curve 11, m-fluorobenzoic acid; curve 111, p-fluorobenzoic acid; curve IV, o-methoxybenzoic acid; curve V, p-methoxybenzoic acid.

478

JOHN H. ELLIOTT AND MARTIN KILPATRICK

of the ortho-substituted benzoic acids in ethylene glycol to fall in line may also be due to difference of solvation in the dihydroxy alcohol, ethylene glycol. In general, these results may be taken as proving that log is linear in (1/D) for pure hydroxylic solvents of dielectric constant greater than 24.2 in the case of meta- and para-substituted benzoic acids, and for monohydroxylic solvents in the case of orthosubstituted benzoic acids. The effect of substituents on acid strength may now be considered in the light of the experimental findings in figures 1 to 4. It is seen that

FIG.4 Effect of substituents upon the acid strength of benzoic acid. Curve I, o-hydroxybenzoic acid; curve 11, m-hydroxybenzoic acid; curve 111, p-hydroxybenzoic acid; curve IV, o-chlorobenzoic acid; curve V, m-chlorobenzoic acid; curve VI, p-chlorobenzoic acid.

ortho-substituted benzoic acids generally have negative slopes, while meta- and para-substituted benzoic acids have positive slopes. An explanation of the positive slopes follows for the electronegative substituents (see equation 5 ) . The possibility that the ortho-substituted benzoic acids can be considered on the basis of an electrostatic effect has been considered by Jenkins (15, 16). Branch and Yabroff (3) offer an ingenious explanation for the positive slope of salicylic acid. They state that salicylic acid may exist in the normal state and the following resonance forms (I and 11). The spatial arrangements of the molecule are such

479

EFFECT OF SUBSTITUENTS ON ACID STRENGTH

$

d

.

.

.

: ' :" ": '., 8 + .: .: .. : .. . J , , .E y

3 i

.

. . . . . . . . . . . . . . . . . . .. .. .. .. . . . . . . . . . .. . ... ... ... ... ... ... ... .. ., : : 42 E . . . . . . . a : . :. :. : : 1 : I : I I ! E ..: ..: ..: ..: :.. :.. :.. :.. . . . : . . . . . * . . . . . . . .

. ;. ;. ;. ;.

,

:.z. ~

.[ ;. ,$ ;

. :. ;. . ;. ;. .i.5 n :: :: 1: : ;: :: .h: 1:

4 ." { & , $ & , . . . :WE : 36 do : i" D V Z ~ z ~ m o* ~ A 0o o o ~ z - m ~ ~ v o ~ z ~ 0 E & ,

'

:L;

'

480

JOHN H. ELLIOTT AND MARTIN KILPATRICK

that the transfer of the phenolic hydrogen t o the carboxyl group involves only an electron shift.

I

I1

These resonance forms, I and 11, are considered to show an enhanced acidity, due to the presence of a positive charge on the carboxyl group. Now, as the number of hydroxyl groups in the solvent available for solvation of the carboxyl group decreases (which occurs in alcohols as 1/D increases), there will be a greater tendency for salicylic acid to be in the forms I and I1 and thus to become relatively stronger. A somewhat similar explanation may be offered for the negative slopes observed for m- and p-hydroxybenzoic acids. If the strength of the bond between the hydroxyl substituent and the oxygen of an alcohol molecule becomes stronger as 1/D increases, the inductive effect of the phenolic hydroxyl group will become less, owing to a shifting of the binding electrons in the phenolic hydroxyl group toward oxygen, thus decreasing its acid-strengthening properties. The list of intrinsic acid strengths, Ka,bo,given in column 3 of table 4, shows that the ortho-substituted benzoic acids are invariably stronger than their meta- and para-isomers. The order of decreasing acid strength for the substituted benzoic acids is not that expected on the basis of inductive effects ( 5 ) , and it is necessary to consider an acid-weakening mesomeric effect to account for the order in the case of the halogen acids.

o\

C

/OH

Inductive effect

Mesomeric effect

In the case of the meta-substituted benzoic acids the departure from the inductive order is explained by Dippy and Lewis on the basis of an electromeric effect (6). I n the case of the o-and p-nitrobenzoic acids, the inductive effect and the mesomeric effect are both acid-strengthening, and the sum of these two effects may be greater for p-nitrobenzoic acid than the inductive effect alone for m-nitrobenzoic acid. This may account for the fact that p-nitrobenzoic acid is stronger than m-nitrobenzoic acid. 3 n the quantitative side attempts have been made to compute the

481

EFFECT OF SUBSTITUENTS ON ACID STRENGTH

electrostatic effect of a substituent on acid strength by an equation of the form (25)

where e is the electronic charge, p the dipole moment of the substituent, 8 the angle between the dipole and the line joining the ionizable proton with the center of the dipole, and r the distance between the center of the dipole and the proton. D E is the effective dielectric constant of the medium through which the radius vector r operates. Computations by this equation can be made by setting p equal to the dipole moment of the corresponding monosubstituted benzene and assuming reasonable values is very sensitive to the for r and 0. It should be emphasized that KAsBo TABLE 5 Log KA.B, calculated from the Kirkwood-Westheimer model and compared with the exuerimental values D,POLE

SOLVENT: H

a

$g$gi

'

D = 37.8

D

D E = 8.1

i

Calcd. Obsd. Calcd. Obsd.

~ _ -___ __ CH,.. , . . 0.40 -0.06 -0.17 -0.07 F . -1.45 0.22 0.06 0.26IC1. . . . . . ,-1.52 0.23 0.22 0.27 B r . . . . . . ,-1.51 0 . 2 3 0.23 0.27 I . .. . , -1.30 0.201 0.23 KO2. . . . . ~ - 3 . 9 3 0.611 0.78 0.71 ~

,

a

~ = 31.6

DE

7.7

~

8oLvENT' CiHaOH

-

D = 24.2

DE

7.1

~

~

SOLVENT:

C4HoOH

'

D = 17.4. Ds 8.4

Calcd. Obsd. Calcd. Obsd. Calcd. Obsd.

~ - _ _ _ _ - _ _ 0 17 -0.07 -0.18 -0.08 -0.18 -0.09 -0.19 ' 0.17 0.27 0.18 0.29 0.23 0.32 0.22 0.30 0.28 0.34 0.30 0.42 0.34 0.40 0.371 0.28 0.42 0.30 0.47 0.34 0.42 0.361 0.24 0.39 0.26 0.45 0.29 0.40 0.961 0.74, 1.02 0.81 1.171 0.90 1.14

Wooten and Hammett (28).

* Trans. Faraday SOC.30, .4ppendix (1934). value chosen for r. Kirkwood and Westheimer' (19, 27) have evaluated D e by considering the molecules and ions as cavities of low dielectric constant. Westheimer (26) has recently evaluated log K A Z ~ for@ the para-substituted benzoic acids, using an ellipsoidal model. This method of calculation has been applied to the results in the various alcohols, and the observed and calculated values are given in table 5. The value of r is 5.9 i., in agreement with the value chosen by Westheimer. The agreement between the observed and calculated values is poorer in the alcohols than in water. Figure 5 and table 6 show that the observed relative acid strengths are more sensitive to the macrodielectric constant than the predictions of the theory. To emphasize further the importance of the value chosen for T , a calculation of r using the value of DE for aqueous solutio? and the experimental values of K.A.B~yields 5.5, 6.0, 5.9, 7.4, and 4.9 A., respectively, for the substituents

482

JOHN H, ELLIOTT AND MARTIN KILPATRICK

Eblvent .............. E U B B T F C ~ N T(PABA)

HIO

CELOH

CtHiOH

C&OH

---_______---

. 1.00

. 1.00 ,

(CHX)H):

Calod. Obsd. Cdcd. Obsd. Calod. Obsd. Calod. Obsd. Calcd. Obsd.

1.00

. 1.00

NO,, . . . . . . . . . . . 1.00

1.00 1.00 1.00 1.00 1.00

1.16 1.16 1.16 1.16 1.16

1.01 2.78 1.39 1.58 1.24

'-

1.22 1.22 1.22 1.22 1.22

1.05 3.02 1.57 1.82 1.32

1.32 1.32 1.32 1.32 1.32

1.03 3.77 1.90 2.01 1.61

1.47 1.47 1.47 1.47 1.47

1.14 3.52 1.80 1.81 1.47

EFFECT OF SUBSTITUENTS ON ACID STRENGTH

483

8 3

I ?3

I

i

m w t - m m i o m 0 0 0 0 0 0 0 0 0 0

.. ... . ... .. ..

. . . . . ... ... ... ... ... ... ... ... ... ... ... ... ..

..

... ...

... ...

... ...

... . ..

.. .. ... ... . ..

.. .. ... ... . ..

.

...

... ... . ..

. . .. .. .. ... ... .. .. .. .. .. .. ..

.

484

JOHN H. ELLIOTT AND MARTIN XILPATRICK

From the data now available, a further test (see table 7) of the equation

given by Hammett may be carried out. The substituent constant u has been evaluated from the data for the substituted benzoic acids in water, setting p = 1. With this nucleus of u values, p and u have been evaluated for many reactions. The constant p depends upon the reaction, the medium, and the temperature. If u depends only on the substituent, p should be constant for a given solvent and temperature for all the metaand para-substituted acids. If we omit the values for m-toluic acid, ptoluic acid, and p-fluorobenzoic acid, p is found to be fairly constant. ~ for these acids are The omissions seem justifiable, since the K A = Bvalues so small that the errors in p are large. The change in p with the dielectric constant can be expressed by the equation p =

0.72

23 +D

(9)

The equation for p permits a calculation of K A z B o for any medium of known dielectric constant down to D = 25, provided equation 3 is obeyed. SUMMARY

1. Relative acid strengths of the substituted benzoic acids have been determined in the solvents ethylene glycol and n-butyl alcohol containing lithium chloride. 2. Intrinsic acid strengths have been determined by extrapolation from the data for the solvents water, ethylene glycol, methyl alcohol, and ethyl alcohol. 3. The order of relative acid strengths has been shown to change when the acids are compared in different solvents. 4. The experimental results have been compared with the calculated computed using the Kirkwood-Westheimer model. values of log KA=B~, These calculated values are in poor agreement with the experimental results, indicating that there are factors affecting relative acid strength which this model does not take into consideration. 5. The results indicate the validity of Hammett’s assumption that the substituent constant is independent of the medium. REFERESCES (1) AKERLOF,G . : J. Am. Chem. SOC.64, 4125 (1932). (2) BJERRUM,X., A N D L.4RSSOX, E . : Z . physik. Chem. 12’7, 358 (1927). (3) BHAKCH, G. E. K . , A N D YABROFF, D. L.: J. Am. Chem. SOC.66, 2668 (1934). (4) BRUNEI.L, R. F., CRESSHAW,J. L., A X D TOBIK, E.: J. Am. Chem. SOC. 43,

561 (1921). (5) DIPPY, J. F. J.: Chem. Rev. 26. 151 (1939).

EFFECT OF SUBSTITUENTS ON ACID STRENGTH

485

DIPPY,J. F. J., AND LEWIS,R. 11.: J. Chem. Soc. 1936, 644. DIPPY,J. F. J . , ASD LEWIS,R. H . : J. Chem. Soc. 1937, 1426. DIPPY,J. F. J., A N D WILLIAMS, F. R . : J. Chem. SOC. 1934, 1888. DIPPY,J. F. J., WILLIAMS, F. R., A Z D LEWIS,R. H.: J. Chem. SOC.1936, 343. ELLIOTT, J. H., ASD KILPATRICK, M.: J. Phys. Chem. 4G, 454 (1941). ELLIOTT, J. H., A X D KILPATRICK, 11.:J . Phys. Chem. 46, 466 (1941). FALKENHAGEX, H.: Electrolytes. Translated from the German by R. P. Bell. Oxford University Press, Oxford (1934). (13) HAMMETT, L. P.: J. Chem. Phys. 4, 613 (1936). (14) HAhlMETT, L. P . : J. .4m. Chem. SOC. 69, 96 (1937). (15) JENKINS,H. 0.: J. Chem. SOC.1939, 640. H. 0 . : J. Chem. SOC.1939, 1137. (16) JEXKINS, M., AND XIEARS,W.H . : J. Am. Chem. SOC.62, 3051 (1940). (17) KILPATRICK, (18) KILPATRICK, M.:Trans Electrochem. SOC. 72, 95 (1937). (19) KIRKWOOD, J. G., A N D WESTHEIhlER, F. H. : J. Chem. Phys. 6, 506 (1938). (20) LAWRIE,J. W.: Glycerol and the Glycols. The Chemical Catalog Company, Inc., New Tork (1928). (21) MASON,R. B., AND KILPATRICK, M.: J. Am. Chem. SOC. 69, 572 (1937). (22) MIINNICK, L. J . , AND KILPATRICK, M.J . : Phys. Chem. 43, 259 (1939). E.: Trans. Electrochem. SOC.64, 209 (1933). (23) NEWBERRY, B., AND MEIER,H. F.: J. 4 m . Chem. SOC.56, 1918 (1934). (24) SAXTON, (25) SCHWARZENBhCH, G., AND EGLI,H.: Helv. Chim. Acta 17, 1183 (1934). F. H . : J. Am. Chem. SOC.61, 1977 (1939). (26) WESTHEIMER, F. H., AND KIRKWOOD, J. G . : J. Chem. Phys. 6, 513 (1938). (27) WESTHEIMER, L. A , , AND HAMMETT, L. P.: J. Am. Chem. SOC.57, 2289 (193.5). (28) WOOTEN, W. F. K.: Proc. Roy. SOC. (London) Ala, 440 (1933). (29) WYNNE-JONES,

(6) (7) (8) (9) (10) (11) (12)

T H E EFFECT OF SUBSTITUEXTS OK THE ACID STREKGTH OF BENZOIC ACID. IV1

IN DIOXANE-WATER JOHK H. ELLIOTT

AND

MARTIS KILPATRICK

Department o j Chemistry and Chemical Engineering, University of Pennsylvania, Philadelphia, Pennsylvania Received August 20, 1940

This paper extends to mixed solvents the study of the effect of solvent on relative acid strengths. Dioxane-water mixtures permit a study over a wide range of dielectric constants (1). Minnick and Kilpatrick (10) 1 This paper was abstracted from the dissertation presented by John Habersham Elliott t o the Faculty of the Graduate School of the University of Pennsylvania in partial fulfilment of the requirements for the degree of Doctor of Philosophy, April, 1940. It was read a t the Sinety-ninth Meeting of the American Chemical Society, which was held in Cincinnati, Ohio, April, 1940.