Sept.

, 1937 ACTIVITY COEFFICIENTS OF IONS IN AQvEOtTS SOLUTIONS 1675

Individual Activity Coefficients of Ions in Aqueous Solutions

BY JACOB KIELLAND~

Lewis and Randall2 in 1923 published a table of may be regarded as the effective diameter of the
26 individual ionic activity coefficients, which has hydrated ion. The individual ai-values may now
subsequently been of frequent use to chemists. be approximately calculated by different methods
The authors emphasized, however, that the pre- -for example from the crystal radius and de-
sented values should be regarded as preliminary formability, according to the following equation
values only. In 1927 Redlich3 reprinted the same for cations, given by Bonino”
table, and no corrections were made even by 108aijzi = 0.9 108ri/1024ai 2 + (2)
Jellinek4 in his comprehensive textbook of 1930. or from the ionic mobilities, using the well-known
It has been pointed out5 that the concept of equation
individual activity coefficients cannot be defined lo8& = 182~i/l, (3)
accurately, and such coefficients may not even be
or its empirical modification, given by BriilllO
determined experimentally without some sup-
108ai = 216~‘/2/1, (4)
plementary definition of non-thermodynamic na-
ture. But the concept may not the less be, and One can also determine the chemical hydration
has often been, quite u ~ e f u l when
, ~ ~ ~estimating number by the entropy deficiency method of
mean ionic activity coefficients in cases where Ulich,*2and calculate ai from this and the effec-
great accuracy is not claimed. tive radius of the ion.
I t is the purpose of the present paper to present Results
a revised and extended table of ionic activity
coefficients, which has largely been computed by The diameters of a number of inorganic and
independent means, taking into consideration the organic ions, hydrated to a quite different extent,
diameter of the hydrated ions,* as estimated by have been calculated by these methods, and the
various methods. results are shown in Table I. Values of “effective”
For sufficiently dilute solutions one may use the ionic radii (ri) were taken from Grimm and
well-known Debye-Hiickel formula (aqueous so- Wolff,13 deformability aiof ions from Born and
lution a t 25’) Heisenberg’* or calculated from the ionic refrac-
tion R by the equation
lOZ4cr = 3 X loz4RI4irN = 0.395R (5)
Errera’s values for R being introduced.15 Ionic
mobilities and entropies have been taken from
wherefi denotes the rationalIgand yi the practical Landolt-Bornstein. l 6 Entropy and chemical hy-
activity coefficient of the ith ion with valence q, dration numbers of some gaseous anions were es-
and I’ is the ionic concentration given by timated roughly by the author. The parameters
i=l
with ci in moles per liter. of the other anions, and of the organic ones, have
It has been shown recently by Briill’O that ai been calculated from ionic mobilities only, and are
(1) Research chemist, Norsk Hydro-Elektrisk Kvaelstofaktiesel-
therefore not included in the table (they are,
skab. however, to be found in Table 11).
(2) G. N. Lewis and M. Randall, “Thermodynamics,” McCraw- When computing the activity coefficients from
Hill Book Co., Inc., New York, 1923.
(3) G. N. Lewis and M. Randall, ”Thermodynamik,” translated the Debye-Hiickel formula, it was decided to ar-
and with supplementary notes by 0. Redlich, Verlag von J. Springer, range the various ions in groups according to the
Wien, 1927.
(4) K. Jellinek, “Lehrbuch der physikalischen Chemie,” Vol. 111, (11) G . B. Bonino and G . Centola, Mem. Accad. Italio. 4, 445
Ferd. Enke, Stuttgart, 1930. (1933).
( 5 ) E. A. Guggenheim, J . Phys. Chem., 3.3, 842 (1929). (12) H. Ulich, Z . Elekluochem.. 36, 497 (1930); also J. Sielland,
(6)E. A. Guggenheim and T. D. Schindler, ibid., 38, 533 (1934). J. Chem. Ed., in press.
(7) G . Scatchard, Chem. R r o . , 19, 323 (1936). (13) H. G . Grimm and H. Wolff,Z . physik. Chem., 119, 254 (1926);
(8) The part of the eEect on logfi, which is taken into account by see also K. Fajans and K.F. Herzfeld, Z . Physik, 2, 309 (1920).
the increase nf ai caused by the hydrated water molectlles, is ratha (14) M. Born and W. Heisenberg, ibid., !U, 388 (1924).
small in the case of some anions and a few cattons, but may be of (15) J. Errera, “Polarisation DiClectrique,”Presses Univeisitaires,
m o r e importance in other cases-compare diameters of hydrated and Paris, 1928, p. 144.
not hydrated ions in Table I. (16) From Landolt-Bdrstein Physikalioch.chemiscba Tabellen,
(9) E. A . Guggenheim, P k l . Mag., 19, 588 (1935). 5. Aufl. I11 Erg. bd., Verlag von Springer, Berlin, 1936, and other
(10) L. Briill, Gam. chim. i f a l . , 64, 624 (1931). sources.
1676 JACOB KIELLAND VOl. 59

TABLEI logf, = -0.5 ~ ? p ' / z / ( l (6) +

PARAMETER 10% AS ESTIMATEDBY VARIOUS METHODS which equals -0.354 z y / z / ( i 3.04 X 0.2325 T v 2 ) +
EBective Hydra- (7)
diam. of Bonino Ionic tion no. Rounded
Ion
unhy- formula mobilities and eff. values
dratedion (eq. 2) (eq. 3) (eq. 4) radius (Table 11)
This equation is based on a parameter ai = 3 A.
Lit 0.8 6.2 4.7 5.6 5.3 6 for all ionic species. Of the about 130 ions ex-
Na + 1.0 4.2 3.6 4.3 4.7 4-4.5 amined by the author, 20% have a diameter of 2.5
K' 1.6 2.8 2.5 2.9 3.9
Rb + 1.8 2.45 2.4 2.8 ...
3
2.5 to 3.5 A., 40% 4 to 5 A., 25% 6 to 8 and 15% A.,
CS' 2.1 2.35 2.35 2.8 ... 2.5 9 to 11 A. The inorganic univalent ions are most
NH ... ... 2.45 2.9 I . . 2.5
frequently of the order 3 to 4 A., the divalent ones
T1+ 1.4 2.15 2.4 2.9 ... 2.5
Ag +
1.5 2.35 2.95 3.45 4.7 2.5 4 to 6 8., the trivalent about 9 A. and the tetra-
Bel + 0.6 ... 8.1 6.8 ... 8
valent about 11 A., while the organic ions fall
Mg2 ' .9 10 6.85 5.8 7.0 8
Cas + 1.4 6.25 6.1 5.15 6.3 6 between 4.5 to 7 8. Hence, one may suggest the
Sr2 + 1.7 5.1 6.1 5.15 ... 5 following formulas, which in the case of univalent
Bal+ 2.1 4.9 5.7 4.8 5.9 5
Ra2 ' 2.3 ... 5.45 4.55 ... 5 ions give the same results as Guggenheim's, but
cu2+ ... ... 6.5 5.45 6.8 6 which in the case of organic and polyvalent ions
Znst 1.1 6.05 6.8 5.75 6.8 6
Cd2 1.4 5.0 6.75 5.65 6.4 5 give distinctly higher values for the activity coef-
Hgz 1.5 4.9 ... ... 6.3 5 ficients.
Pb2+ ... 4.5 5.2 4.35 5.9
Mu1 1 5.2 6.8 5.7 ...
4.5
6 Inorganic ions" logfi = -0.5 z;pl/z/(l + z,pl/r) (8)
Fez 1 ... 6.8 5.7 6.8 6 Organic ions logfi = -0.5 z;p%/(l + 2p1/2) (9)
Nil 1 5.6 6.8 5.7 ... 6
Col + 1 ... 6.8 5.7 ... 6 Such formulas as the following of Guggenheim6
SnZ + ... ... ... ...
Ala 0.8 ... 8.65 6.0 > 8
6.2 6
9 - l o g f , = 0.5Z7p1/2/(l f p1lz) + ZKBikCk (10)
Feat ... ... 8.05 5.5 > 8 9
Cra ... ... 8.15 5.6 ... 9 or that of Briilll*
La' 2.0 7.7 7.8 5.4 ... 9
+ +
... ... ... -logfi = 0.358~;r~/z/(l 10*a,0.2325rVz)
Cea
Pra + ... ...
8.15
8.35
5.6
5.75 ...
9
9 A(W + Ph) (11)
Nda+ ... ... 8.5 5.8 ... 9
sea+ 1.4 10.8 8.45 5.8 ... 9 may be more accurate in the more concentrated
Sma + ... ... 8.3 5.7 ... 9 range, but they contain additional constants,
ya + 1.6 8.15 ... ... I . . 9
which must be determined by the activity coef-
Inat 1.4 8.95 ... ... ... 9
Snd + ... 12.5 ... ... ... 11 ficient measurements themselves.
Th4+ 2.2 10.3 ... ... ... 11
zr4 + 1.6 11.5 ... ... ... 11
Comparison with Experimental Values
Ce4 2.0 11 ... ... ... 11
F- 1.5 ... 3.3 3.9 5.3 3.5
CI - 1.9 ... 2.4 2.85 3.9 3 The individual ionic activity coefficients com-
Br- 2.0 ... 2.3 2.75 3.6 3 puted in this paper with independently estimated
1- 2.2 ... 2.4 2.8 2.9 3
ai-values (except in the case of H+), have been
clog- ..* ... 2.85 3.35 3.3 3.5
clod- ... ... 2.7 3.15 3.5 3.5 compared with those obtained experimentally by
NOa - ... ... 2.55 3.0 3.1 3 Hass and Jellinek19 in the case of C1-, Br-, I-,
BrOa- ... ... 3.25 3.85 3.5 3.5
10s- . I . ... 4.45 5.25 4.8 4-4.5 S0-i and (COO)-:, and with those found by
HCOa- ... ... 4.1 4.85 7 4-4.5 Bjerrum and UnmackZ0in the case of H f , PO-:
Ss- 2.2 ... ... ... 5 5
and citratea-. In Fig. I we have plotted these
Coal- ... ... 4.25 5.05 6 4.5
so,*- . . . ... 3.8 4.55 51/2 4 values, which agree fairly well with the calculated
c~04~-. . . ... 4.2 5.0 s1/z 4.5
ones (drawn lines).
parameters found above, and use rounded values (17) Except such complexes as the ferrocyanides, cobaltammines
of ai, as clearly stated in Table 11, where the re- and others which are to be treated as organic ions by the formula 9.
(18) L. Briill, Gazz. chim. ital., 64, 261, 270 (1934).
sulting activity coefficients are given. (The (19) K.Hass and K. Jellinek, Z.physik. Chem., A162, 153 (1932).
parameters of hydrogen and hydroxyl were not The activity coefficients of Soil- were recalculated, using the more
recent value 0.614 for the normal potential of the mercurous sulfate
calculated independently, but taken as 9 and electrode [Shrawder, Cowperthwaite and La Mer, THISJOURNAL,
3.5 Hi.) 66, 2348 (1934)l together with the best value 0.222 for that of the
silver chloride electrode (ref. 16, p. 1855). When computing the
Approximate Formulas solubility product of silver oxalate, we extrapolated against the func-
tion pl/l/(l+ 1.5fi1/2)(H. and J. used p1/2), and found it to be 1.3.
Guggenheim and Schindler6 have proposed the 10-11 at 25'. Hence, all activity coefficients of oxalate anion (as
following formula for ionic activity coefficients up found by H. and J.), were multiplied by 1.2.
(20) N. Bjerrum and A. Unmack, Kgl. Danskc Vidensk. Selskab,
to an ionic concentration of about r = 0.2 MaL-fys. Medd., 9, 1 (1929).
 Sept., 1937 ACTIVITY COEFFICIENTS
OF IONS SOLUTIONS
IN AQTJEOUS 1677

TABLEI1
ACTIVITYCOEFFICIENTS
INDIVIDUAL IN WATER
OF IONS

Parameter
1oSai 0,001
Total ionic concentration
0.002 0.005 0.01
r
0.02
=
cs i=l
0.05
cis,
0.1 0.2
Inorganic ions:
H+ 9 0.975 0.967 0.950 0.933 0.914 0.88 0.86 0.83
Li + 6 .975 ,965 .948 .929 ,907 .87 ,835 .80
Rb+,CS+, NH4+, T1+, Ag+ 2.5 .975 ,964 ,945 .924 ,898 .85 .80 .75
K+, C1-, Br-, I-, CN-, NOz-, NOS- 3 .975 ,964 .945 ,925 ,899 .85 ,805 .755
.76

.775

,355

.37

868 ,805 . 744 .67 . 555 . 465 .38

CaZ+, Cu2+, Zn2+, SnZ+, Mn2+, Fez+, Ni2+,
,905 ,870 ,809 ,749 ,675 .57 ,485 ,405
co2 +

Mg*+, Be2+ 8 ,906 ,872 ,813 ,755 .69 ,595 .52 .45
.796 ,725 .612 .505 ,395 .25 .16 ,095
.798 ,731 ,620 .52 ,415 .28 ,195 .13
,802 .738 ,632 .54 ,445 ,325 ,245 .18
.668 .57 ,425 .31 .20 .10 ,048 ,021
.670 ,575 .43 ,315 .21 .105 ,055 ,027
.678 .588 ,455 .35 ,255 ,155 .10 ,065
.a2 .43 .28 .18 ,105 .045 ,020 ,009

,975 ,964 ,946 ,926 ,900 ,855 .81 .76

.975 ,964 ,947 ,927 ,901 ,855 ,815 .77
.975 ,964 .947 .928 ,902 .86 .82 ,775

.975 .964 ,947 .928 .904 .865 .83 .79

.975 .965 ,948 .929 .go7 .87 ,835 .80

.975 ,965 ,948 ,930 ,909 ,875 .845 .81

.975 ,966 .949 ,931 ,912 ,880 .85 .82
.903 ,867 .804 ,741 ,662 .55 .45 .36
.903 .868 .805 ,744 .67 ,555 ,465 .38
.905 ,870 ,809 .749 .675 .57 ,485 ,405

[OOC(CHz)sCOO ]'-, [OOC(CH2)eCOO1'- .906 .872 ,812 ,755 ,685 .58 .50 .425
Congo red anion2-
Citratea- 5 .796 .728 .616 .51 .405 .27 .18 .115

As another test we have in Table 111 compared umns represent: (1) author's Table I1 (formula 1,
the experimentall6 mean activity coefficients for with calculated ai-values) ; (2) author's approxi-
some strong electrolytes with those calculated mate formulas ; (3) Guggenheim's formula ; and
from individual coefficients. The last four col- (4)Lewis and Randall's old tables.
1678 AND F. B. LAFORGE
H. L. HALLER Vol. 59
.2 ,726 .75 .76 .76 ...
.4 .658 .69 .70 .70 I..

0.2
1.0 .59 ,605 .62 -62 ...
LiCiHiSOa 0.04 .877 .88 .875 .87 ...
.25 .782 .785 .77 .74 ...
k0.4 1.0 .703 .71 ,675 .62 ...
M HCOONa 0.004 ,954 ,955 .953 .953 ...
.Ol .Q31 .927 ,927 .927 ...
I 0.6 .02 .908 ,901 .90 .QO ...
.04 .879 .87 ,865 .86,5 ...
LinSO4 .012 .862 .85 .855 .85 .82
0.8 .03 .SO8 .78 .79 .78 .75
.06 .754 ,725 .73 .71 .70
.3 ,801 .57 .58 ,525 ...
ZnClr .01 ,857 .861 ,865 .86 ,845
0.2 0.4 0.6 0.8 .03 .778 .78 .7Q .78 .77
Square root of ionic strength, 4;. .1 .660 .68 .685 .66 .67
.17 .609 .63 .63 .5Q .62
Fig. 1.-Individual activity coefficients of ions: drawn .5 .51 .53 * 53 .465 ...
lines represent calculated values, with rounded ai - fig- La(N0a)a .0145 .796 .785 ,785 ,765 .79
ures given in Table I. .034 .720 .71 .71 .675 .72
.066 .645 .645 .645 .59 .64
TABLE
111 ,118 .570 .5Q .59 .51 .57
Mean ionic activitv coefficient ,221 ,505 .525 .525 .42 .so
Ionic Gusgen- Lewis--
concn. Table Approx. heim Randall KaFe(CN)s ,012 ,785 .79 .79 .78 .82
Electrolyte r Exptl. I1 formula formula tables ,024 .717 .72 .73 .71 .76
HI 0.01 0.927 0.928 0,927 0.927 0.95 .06 .618 .61 .63 .60 .69
.02 ,902 ,906 .90 .QO -92 .12 .547 .52 .55 .51 .62
.04 ,870 ,875 ,865 .865 ,895
.I .822 .83 .81 .81 .86 Summary
.2 .787 .79 .76 .76 .815
HCI .01 .Q28 .Q28 ,927 ,927 .95 Individual activity coefficients of 130 inorganic
.02 .QO5 .QO6 .QO 90 .92
.04 .878 ,875 ,865
I

,865 .895
and organic ions in water a t concentrations up to
.1 .831 .83 .81 .81 ,813 I' = 0.2 have been computed and tabulated;
.2 ,799 .79 .76 .76 .815
parameters ai were calculated by various methods.
HCI in LaCla .125 ,818 .82 ,795 ,795 .85
.15 ,811 .81 .78 .78 .84 For approximative work, these individual figures
,175 .794 .80 .77 .77 ,825 have been shown to give mean coefficients in suf-
NH4N03 .02 ,882 ,898 .90 .90 ... ficient accordance with experimental values.
.04 ,840 .86: .87 .87 ...
.1 ,783 .80 .81 .81 ... NORWAY RECEIVED
PORSGRUNN, NOVEMBER
23, 1936

[CONTRIBUTION
FROM THE BUREAUOF ENTOMOLOGY
AND PLANT UNITEDSTATES
QUARANTINE, DEPARTMENT
OF AGRI-
CULTURE ]

Constituents of Pyrethrum Flowers. IX. The Optical Rotation of Pyrethrolone and

the Partial Synthesis of Pyrethrins'
BY H. L. HALLERAND F. B. LAFORGE

Among the observations on the chemical and
physical properties of tetrahydropyrethrolone,
we have previously reported its specific rotation
as +11.9°.2 This value agrees in magnitude but
differs in sign from the one ( - 11.3') reported by
Staudinger and Ruzicka3 for the same compound.
The material on which their rotation is reported
was prepared by hydrogenation of a pyrethro-
lone preparation obtained from a mixture of the
semicarbazones of pyrethrins I and I1 in which
the relative proportions of each were unknown.
(1) For Article VI11 of this series, see J. Org. Chew , 2, 56 (1937).
(2) LaForge and Haller, THISJOURNAL, 68, 1777 (1936).
(3) S t n u d i n g e r a n d Ruzicka, Rdo. Chim. Acln, 7, 212 (1024).
Since the rotation reported by us was also ob-
served on material originating from a mixture
of both pyrethrins, the discrepancy between our
value and that reported by Staudinger and
Ruzicka might be explained with the assump-
tion that the pyrethrolone present in pyrethrin
I differed optically from the one in pyrethrin 11.
We have previously described a method4 by
which pyrethrin concentrates may be separated
into fractions in each of which one of the pyre-
thrins predominates. From a fraction in which
pyrethrin I predominates its semicarbazone may
(4) 1,aForRe and Haller, THIS
JOURNAL, 67, 1893 (1936).