BENTONBROOKS OWEN
24
on the temperature, and the absolute accuracy of the bath depends solely on the purity of the water. By repeating the washing occasionally the constancy can be maintained almost indefinitely. Air-saturation is easily secured and still more easily tested. With a sealed bulb giving the triple point of water :’is a temperature standard there should be no error from conduction down the thermometer. Impurity in the water, however, is exceedingly troublesome, since it is both concentrated and transported by the freezing, and the configuration
[CONTRIBUTION FROM
THE
1701. 56
is continually changing. A particular method of freezing gives a relief which is excellent but only temporary, and then there seems to be no way to avoid beginning preparations all over again. Hence the triple point, while convenient and reliable for a short job, is less so than the above-described “ice-point” for continuous work. If the impurity is small enough in comparison to the requirements, however, the triple-point apparatus is excellent, though it requires more attention and work than the cold cell. RECEIVED AUGUST 2, 1933
WASHINGTON, D. C.
STERLING CHEMISTRY LABORATORY O F YALE UNIVERSITY ]
The Dissociation Constants of Glycine at Various Temperatures BY BENTONBROOKSOWEN R e ~ e n t l y i’t ~was ~ ~shown ~ that the dissociation constants and activity coefficients of weak electrolytes (including ampholytes) can be obtained from measurements of the electromotive forces of cells without liquid junctions. In applying the method to glycine,2 the activity coefficient calculations were complete, but it was pointed out that the dissociation constants were i‘apparent” because insufficient data were a t hand to correct for the medium effect of the glycine ions. The true (thermodynamic) dissociation constants have been evaluated in the present paper.
The acid dissociation constant was determined by the method of Harned and Owen,2 making use of cells of the type (-4)
Neutral glycine is represented by Z*, and glycine hydrochloride by HZC1. Two stock solutions were prepared with slightly different values of the ratio, ,ml/m2, and these, in various dilutions, constitute Series A1 and A2 for which the data are given in I* The ‘lapparent” hydrogen ion concentrations, m&+,were calculated by the eauation E/k
- E O / k + 2 log Y&
+ log m?
=
-log mkt
. -
+
KA =
k.kY;*Yk+/YLB+
and kA
=
ml*mk+/m&+ = (m,
[St]
(2)
+ mg+)m&+/(mz- mk+) (3)
Discussion of the Methods
HdZ* ( m d , HZCl(mdIAgC1, Ag
in which k = RT2.3026/F, and Y& is the activity coefficient of hydrochloric acid in pure water solution a t the concentration W L ~ . Values of ygCl were read from plots of the data of Harned and E h l e r ~ and , ~ E o was calculated from Equation (7) of the same paper. Representing the acid dissociation of glycine by ZH+ = 2” H f , and the true and apparent dissociation constants, respectively, by
i t follows that K A is obtained by extrapolation to p = 0 on a plot of log ka against p. The necessary data are given in Table I. The extrapolation is illustrated in Fig. 1, and the values of log K A are collected in Table 11. The basic dissociation constant was determined by the method of Harned and Ehle1-5,~ making use of cells of the type Hz/Z*
(4, N a G h ) , NaCl(mdlAgC1, Ag
(B)
in which N ~ Grepresents sodium glycinate. The electromotive force of this cell is expressed b y Elk - Ea/k
=
-log
rnH+rnCI-YH-YCI-
(4)
Representing the basic dissociation of glycine by [2O’I4 GHzO = 2‘ OH-, and the true dissocia(1) tion constant by \-,
(I) Harned and Owen, THISJOURKAL, 62, 5079 (1930). (2) Harned and Owen, ibid.,61, 5091 (1930). (3) Harned a n d Ehlers, i b i d . , 64, 1350 (1932). (4) Numbers enclosed in brackets refer t o t h e corresponding Equation numbers i n Ref 2
+
+
- mz~moH-Yz*Yo~-
B -
_____
mc- YG-aw
[7 1
( 5 ) Harned and Ehlers, THISJ O U R N A L , 66,2179 (1933).
(3
nISSOCIATION CONSTANTS OF
GLYCINE
25
TABLE I ELECTROMOTIVE FORCESOF Pn2
X 10'
5126 6449 6833 8483 10237 14107 14S3 27022 37740 37740
Series
A1
0.52135 ,51376 .51157 .50450 ,49870 .48890 .48725 ,47043 ,46195 .46203
0 51122
49665 ,48728
,46144 ,46151 0.51025 .51014 ,50895 ,50083 .49412 .48764 .47353 .46223 .46235
4659 5477 5661 6998 8627 11892 23135 24530 25440 34488 36960 50907
0.51301 .51275 ,51158 .50315 .49599 .48933 .47456 .46269 ,46274
0 93427
0.93542
,91240 ,89656 .89511
.91315 ,89705 .89566
,88710
,88750
0.93651 .93264 .93161 ,92132 ,91383 ,89747 .89609 ,88779 ,88609 .87830
Series B2
0.52281 ,51502 .51282 .50556 .49967 ,48956 .48797 ,47084 .46210 .46224
0.51434 .51411 .51262 .50404 .49696 .49011 .47510 .46285 .46294
- E"/k
f log K ,
+ log mcl-mZt/mG-
= log K~ log Y ~ , - Y ~ ~ / Y (6) ~ -
Here mcl- == m3, mG- = m2 - moH-, and mzt = ml moK-. The left side of equation (6) is therefore known if mOH- can be estimated. For this purpose it is sufficient to set mOH- = KBmz/ml as a first approximation, and use a rough value of KB obtained from equation (6) by neglecting mOHand the last term on the right. Then by plotting the left side of equation (6)
+
B
350
40'
0.52656 ,51827 .51599 ,50834 .50197 .49146
0.51553
0.51750
.51386 .50501
.51579 .50667
.49076 .47545 .46301
.49190 .47605 .46315
450
.48988
.47157 .4G241 .46257
0.93441 .93335 .92271 .91486 ,89800
,88799 ,88635 .87835
=
0.98487
0.93920 ,93510 ,93408 .92867 .92318 .91527 .89820 .89674 .89588 ,88804 ,88626 ,87823
0.93573 ,93469 ,92372
0.94042 ,93627 ,93528 ,92961 .92407
.89664 ,89571 .E48612 ,87794
.88588 .87754
0.93637 ,92722 ,92649 ,91165 .89866 .88786 ,88009
0.93692 .92763 ,92687 .91186 ,89865 .88765 ,87974
mr/ms = 0.95276
0.93411 ,92547 .92472 ,91056 .89829 .88799 ,88058
equations (4) and (5) may be combined with the dissociation constant for water, K,, to yield Elk
AND
0.52416 .51618 .51389 ,50659 .50038 .49030 ,48860 ,47113 .4D230 .46235
m,/mr
0.93750 ,93356 ,93252 ,92734 ,92208 ,91439 .89778 ,89637 ,89555 .88796 .88627 ,87841
ml/ma = 0.83826
0.93316 .92472 ,92389 .go992 .89794 ,88778 .88048
6231 8902 9166 16218 26604 40535 54955
30"
ml/ms = 1.0187
Series B1
ms X 106
CELLSOF TYPESA
ml/m2 = 1.2981
Series A2
6917 6970 7263 9497 11948 15016 25696 41200 41200
THE
200 250 m / m z = 1.1301
15'
100
0.93498 .92616 ,92541 .91104 .89852 ,88806 ,88052
0.93572 .92675 ,92600 .91140 ,89865 .88798 .88035
against p, log K B is obtained by extrapolation to 0. If the value of K B thus obtained differs6 from the one employed, moH- must be reapproximated using the new value of K B . Finally, by successive approximations and extrapolations, a correct, unvarying value of K , is obtained. The values of K, were taken from the paper of Harned and Hamer.' The necessary data are given in Table I, and the extrapolations illustrated in Fig. 2 . Table I1 contains the values of log K B . u , =
( 6 ) Cf.footnote, Reference 3, p. 1351. (7) Harned and Hamer, TIEIS J O U R N A L , 66. 2194 (1933).
BENTONBROOKS OWEN
26
Materials and Technique Eastman ammonia-free glycine was precipitated once by methanol and recrystallized four times from conductivitv water. I t was kept thereafter in a vacuum desiccator over solid potassium hydroxide, occasionally renewed. I t s moisture content was determined as follows. 0.70‘
0.68 ‘-
0.56
D
1
Fig. t .-The
!
o
0
! 0.01
I
I
I t was standardized gravimetrically, and also against the sodium hydroxide. All standardizations were repeated before the preparation of each of the stock solutions. All concentrations were known to the nearest 0.1 %. For Series AI, B1 and B2 the cells were those previously employed by E h l e r ~ ,and ~ , ~ were filled under vacuum. For Series A2 simpler, smaller cells were , used, and filling was carried out under pressure of pure hydrogen. In this series the electrodes were allowed to 400 stand overnight in small portions of the solutions. Electrodes were of the 300 usual type employed by Harned and his co-workers, and were freshly pre25” pared for each cell. Cells were measured In duplicate, and the mean of their electromotive forces recorded. The average difference between duplicate cells was about 0 05 mv Where this difference exceeded 0 2 mv the 10” measurements were discarded.
I
I
0.02
0.03
!J. acid dissociation of glycine:
Vol. 56
0.04
0, series A1 ; 0 , series A2.
Discussion of Results The observed dissociation constants contained in Table I1 can be accurately expressed by the empirical equations
A 20-g. sample, maintained a t 103 t o 107” in air, lost log K . ~= -2,4514 + 0,0052t - 0,000047t2 (7) 0.19% in weight in two days, and a further 0 . 0 2 ~ in o ten and days. A fifteen gram sample, maintained at 100” in the vacuum produced by a Cenco-megavac pump, operating log KB = -4.4200 0.0093t - 0.000050t2 (8) continuously, was weighed hourly for thirteen hours. By substituting the derivatives in the van’t Hoff The loss in weight approached 0.16% asymptotically. The purity of the glycine was therefore taken as 99.8%. equation, the corresponding heats of dissociation A series of small samples and stock 9.95 solution B1 were analyzed by a microKjeldahl method,8 which confirmed this figure. b 0.90 Baker “Analyzed” sodium chloride was precipitated by hydrogen chloride, + 0.85 ignited and recrystallized three times from conductivity water. I t was dried 2 overnight at 110” before being used. M 2 0.80 The sodium hydroxide was Baker I “Analyzed.” The carbonate was allowed to settle from a 50% solution of k 0.75 the hydroxide. Standardization, storage, dilution and all subsequent manip+ 0.70 ulation of this, as well as all other s o h - Ic3 tions, were carried out in an atmosphere 0.65 of tank hydrogen purified by passage over red hot copper turnings, and 0.02 0.04 0.06 0.08 0.10 0.12 P. through a soda-lime tower. Full details of this technique are given elseFig. 2.-The basic dissociation of glycine: 0, series B1; 0, series B2. where.* Benzoic acid and acid potassium phthalate from the Bureau of Standards were used as were calculated a t various temperatures and inalkalimetric standards. cluded in Table 11. These values can be reThe hydrochloric acid was Baker “Analyzed,” redistilled. produced exactly by ___(8) The author takes pleasure in acknowledging his indebtedness A H ~= 1773 - 18.6t - 0.24t2 calories (9)
+
< -
to Professor David I . Hitchcock, of the Department of Physiology, for these arialyses [Hitchcock and Helden. I n d . Eng. Chem., Anal. Ed., 6, 102 (19?3)1.
and AHB = 3170
- 10.2t - 0.24t2calories
(10)
DISSOCIATION CONSTANTS OF GLYCINE
Jan., 1934
27
TABLE I1 THEDISSOCIATION CONSTANTS AND HEATS OF DISSOCIATION OF GLYCINE f,
"C
+ log K A 3 + log K B
100
15'
0 616" 595 670 ,709 A H 4 (cal ) 15632 1439 AHg (cal ) 3044 2962 a Calculated by Equation (7). I)
200
25'
300
350
0.634
0 650
0 602
0 673"
.746
781
1305 2869
11.79 2765
814 1000
In view of the wide disagreement among the values of K 4 and KB reported in the literature, and the fact that their evaluation involved the estimation or "elimination" of liquid junction potentials, a critical comparison with earlier values will not be attempted.$ Heats of dissociation, depending upon the temperature coefficient of the electromotive force rather than upon its absolute magnitude, are in more satisfactory agreement. Thus, Bjerrum and Unmacklo found AHA = 1,510cal. between 0 and 37", and AHB = 2710 cal. a t 25'. Although Harned and OwenZ expressed the opinion that the apparent constants, ki and ki, determined by them were probably within 1% of the true constants, KA and K B , the observed differences are, respectively, 8 and 6% of the former, recalculated to conform to the same values of E ' and R T / F . If this difference is not entirely due to the medium effect of glycine ions, but is due in part t o analytical errors, i t is obvious that the latter are much less probable in the present research. Moreover, the efects of some of the uncertainties, discussed below, are much greater in the former, unbuffered measurements. The estimated numerical uncertainty of the present results is of the order of *0.002 in the (9) Readers desiring to make this comparison should consult the compilation by Branch and Miyarnoto [THIS JOURNAL, 6% 863 (1930)], and also Reference 10. The equations li,, = K , / K 5 , Kh, c K,-/K*, I