Acc. Chem. Res.

1985,18, 207-212 207

Thermodynamics of Dissolution of Simple Gases in Water

ROBERTO
FERNANDEZ-~RINI,*
ROSACROVETTO,
MAR~A
L. JAPAS,
and DANIELLARIA
Departamento de Quimica de Reactores, Comisidn Nacional de Energ$ Atdmica, Av. Libertador 8250,1429-Capital Federal, Argentina
Received February 22, 1985

The thermodynamics of dissolution of simple non- in which the thermodynamic quantities of dissolution
polar gases in water has played historically a prominent obtained from the solubility data are reliable enough6
role in signaling out the peculiar characteristics of so- will be reviewed and compared with those predicted by
lutions in aqueous medium. The pioneering work of perturbation theories applied to hard-sphere reference
Eley,' and of Frank and Evans2 laid down the basis for fluids. Arguments will be advanced to explain why
the use by the latter authors of the large negative en- perturbation theories, which do not take account of the
tropy of dissolution of noble gases in water as an in- intermolecular orientational correlation, are applicable
dicator of the distinct type of ion-solvent interactions to the solvent water. The rather controversial view
found in aqueous solutions. The large positive partial advanced in the present work that introduction of small
molar heat capacity of dissolution of the gases stands nonpolar solutes into water does not involve significant
as another characteristic typical of aqueous solutions. reorientation of the solvent molecules has the advantage
These features of solutions in water are considered to of explaining the available experimental results and also
be originated in the strong intermolecular interactions of suggesting some further experiments which would
existing in the liquid due to extensive hydrogen bonding help to support or reject this interpretation of the
amont neighboring water molecules. dissolution of simple gases in water.
Pierotti3 pointed out that the thermodynamics of
dissolution of simple gases in water close to room tem- Calculation of Henry's Constant by
perature could be accounted for in terms of a hard- Perturbation Theories
sphere fluid model with an added Lennard-Jones at- In order to calculate the values of Henry's constants
tractive term. For this purpose he used the scaled for gas-liquid binary systems, a hard-sphere reference
particle theory (SPT). The success of such a simple fluid having a Lennard-Jones perturbation potential
theory when applied to aqueous solutions has been has been employed. As already mentioned, SPT is one
considered fortuitous by Stillinger4 because SPT takes of the simplest theories to be used; however, since it has
no account of the orientational correlation present in been shown to underestimate the magnitude of the
the solvent. work necessary to create the solute cavity in the liquid:
The significance of this point could in principle be we have not used SPT in the present calculations.
assessed by testing the theory over a wide temperature With the assumption that a Lennard-Jones potential
range since it is known that the intermolecular orien- describes adequately the energy of non-hydrogen
tational correlations in water are quite sensitive to a bonding interaction between solvent (1)and solute (2),
change in temperature. IR and dielectric constant u12is given bys
studies5 have shown that the degree of intermolecular = 4C12[x1' - X6]
correlation in liquid water decreases appreciably with
U12 (1)
increasing temperature; however, even at 573 K water where x = ( u 1 2 / r )and the parameters t12 and u12 mea-
is a liquid with extensive intermolecular hydrogen sure the strength and the range of the interaction. For
bonding. the interactions of nonpolar molecules with water in the
As a result of the research program of our laboratory gas phase (low density fluid), eq 1 should include the
related to the study of the thermodynamics of gas-liq- contribution of the dipole-induced dipole interaction.8
uid binary systems, we have determined Henry's con- On the other hand, these interactions in a high density
stant for inert gases and for methane in light and heavy (liquid-like) fluid can only be calculated if the instan-
water6a and for nitrogen, and hydrogen6c in taneous total electric field on the solute particle due to
H20, over a wide temperature range. In this work ex- the solvent dipoles surrounding it is known. On the
perimental results corresponding to a wide range of basis of symmetry considerations, this contribution to
temperature (between room temperature and 473 K) the total intermolecular energy is probably close to zero
Roberto Fernindez-Prini received his Ph.D. from Universidad de Buenos (1)D.D.Eley, Trans. Faraday SOC.,35, 1281,1421 (1939).
Aires (UBA) in 1964. He is at present Head of the Department of Inorganic, (2)H. S. Frank and M. W. Evans, J . Chem. Phys., 13, 507 (1945).
Analytical, and Physical Chemistry at the Faculty of Sciences (UBA) and (3)R. Pierotti, J. Phys. Chem., 67, 1840 (1963);69, 281 (1965).
Head of Division of Coolant Physical Chemistry at Comlsibn Naclonal de (4)F. H.Stillinger, J. S o h t . Chem., 2, 141 (1973).
Energih Atbmica. (5)E.U.Franck and K.Roth, Discuss. Faraday Soc., 43,108 (1967);
Rosa Crovetto has a Ph.D. from Universidad de-Buenos Alres and at K.Heger, M. Uematau, and E. U. Franck, Ber. Bunsenges. Phys. Chem.,
present is working at the Comisibn Naclonal de Energia Atbmlca, as well as 84,758(1980).
being Assistant Professor at the Faculty of Sciences (UBA). Her research (6)(a) Crovetto, R.Femhdez-Prini, and M. L.Japas, J. Chem. Phys.,
has dealt with effects of pressure and temperature on solutlons. 76, 1077 (1982);(b) R.Crovetto, R. Fernindez-Prini, and M. L.Japas,
Ma& L. Japas is an M.Sc. from Universidad de Buenos Alres who Is fin- Ber. Bunsenges. Phys. Chem., 88,484(1984);(c) J. Alvarez, R. Crovetto,
ishing her doctoral thesls on Thermodynamlcs of Dissolution of Gases in Li- and R. Fernindez-Prini, Water Steam: Their Prop. Curr. Ind. Appl.,
quids. She worked for 2 years in the Physical Chemistry of Fluids under Proc. Int. Conf. Prop. Steam 10th (1984);(d) R. Crovetto and R.
Professor E.U. Franck at Karlsruhe University. Fernbdez-Prini, AIChE J., 31, 513 (1985).
Daniel L a h is an M.Sc. from Universidad de Buenos Aires. His research (7) R. Crovetto. R. Fernhdez-Prini, and M. L. Japas, J.Phys. Chem.,
has been devoted to Thermodynamics of Solutions. At present he Is a grad- 86,4094 (1982).
uate student under Professor D. Chandler at the Department of Chemistry, (8)J. Alvarez, R. Crovetto, and R. Fernindez-Prini, 2.Phys. Chem.
University of Pennsylvania. (Munich) 136, 135 (1983).

0001-4842/85/0118-0207$01.50/0 0 1985 American Chemical Society

208 Fernandez-Prini et al. Accounts of Chemical Research
although there may be rather large fluctuation terms. Table I.
This contention is supported by statistical mechanical Molecular Parameters of the Gaseous Solutes Employed in
calculations applied to fluid systems with liquid-like the PY Treatment
densities having a central point dipole. Employing the gas 4, A tplk, K
mean spherical appro~imation,~ or the hypernetted Ne 2.900 36.3
chain quadratic approximation,1° it has been found that Ar 3.412 121.8
for a nonpolar solute particle in this dipolar fluid the Kr 3.709 174.7
Xe 4.054 229.9
total electric field is zero. Consequently we have em- CH, 3.681 151.7
ployed in eq 1 a Lennard-Jones potential without di- CZH, 4.377 236
pole-induced dipole interaction. H2 2.750 29.2
The change of standard chemical potential of the NZ 3.400 95.05
solute, which characterizes the dissolution process (from
ideal gas to infinitely dilute solution taken to a mole For a fluid with the density of water eq 3 and 4 give
fraction of unity), may be expressed by values of Apcav2which are quite close especially at low
temperature^.^
Apo2 = Apid2 + ApcLcBv2 + Apat2 (2) The contribution of the attractive solute-solvent in-
The three terms represent the contributions to Apo2 due teractions is given for both treatments by
respectively to changes in the density of an ideal gas
when going from its standard state (0.1 MPa) to the
density of the reference fluid, to the formation of a
cavity in the reference fluid (solvent) large enough to where u'&) is the perturbation Lennard-Jones po-
accommodate the solute particle, and to the attractive tential. For PY Apat2was calculated from the Laplace
Lennard-Jones solute-solvent intermolecular interac- transform of rglZo(r)for mixtures of hard spheres de-
tion. rived by L e b o ~ i t z . ' ~For PC, g120(r)is obtained from
We have employed two methods to calculate Apo2. a calculation involving the radial distribution function
The first one is based on the Percus-Yevick (PY) of the pure solvent and sphericalized Ornstein-Zernike
hard-sphere equation of state which takes no explicit expressions."
account of the intermolecular correlations existing in The use of perturbation theories based on hard-
the real solvent. The other is the semiempirical per- sphere reference fluids requires values for the di's.
turbation method of Pratt and Chandler1' (PC) which These are determined essentially by the steepness of
uses the experimental distribution function of the the repulsive branch of the intermolecular potentials;
solvent to calculate the various terms in (eq 2). This they may be calculated by either the Barker and Hen-
second method is particularly attractive for the solvent derson14 or the Weeks, Chandler, and Andersen15
water because the correlation among water molecules methods, for aqueous solutions the calculation is ham-
due to hydrogen bonding interactions is explicitly taken pered by the need to decide upon the best water-water
into account. The theory employes spherical correlation intermolecular potential.
functions which are related to the thermodynamic The value of t12 in eq 1can be expressed for nonpolar
properties of the system via the compressibility equa- gas-water interactions by the geometric mean combi-
tion.12 For PC the distribution function of the refer- nation rule, Le., e12 = ( ~ ~ e ~ only ) ' / ~if, el denotes the
ence fluid is that corresponding to a hard-sphere par- water-water intermolecular energy when all hydrogen
ticle introduced in real water. bonding contributions are eliminated. This parameter
According to the PY treatment, pApcav2is given by is not readily available, but we have estimated its value
,8Apcav2 = -In (1- y ) + (3y(R + R2)/(1 - y ) + from gas-phase studiesS8On account of the difficulties
to obtain rigorously the molecular parameters when
(9/2)[yR/(1 - y)l21+ ,8yPHSR3/pl (3) dealing with solvents having strong intermolecular in-
where p1 is the number density of the solvent, PHsis teractions, we have employed a more pragmatic semi-
the pressure of the hard-sphere reference fluid, y is the empirical approach.
packing fraction of the solvent, and R = d2/dl is the
Method of Calculation and Results
ratio of equivalent hard-sphere diameters.
For the PC treatment the cavity term is given by The values employed for the molecular parameters
d2 of the solute gases for the P Y treatment are given in
,8Apcav2= -In (1- y) + 2 7 r p 1 s0 G12"(z)z2dx (4) Table I, the values for PC being only slightly different;
the diameter of water was taken to be 2.70 A at all
Where the second term gives the total work of creation temperatures. With these parameters and with use of
of a cavity of size x which grows up to the diameter dz the experimental values of the solubilities of Ar, Kr, Xe,
of the solute, z = (d, + x)/2 and G120(z) is the contact and CH4 in H 2 0 at 298 K, the best value for cl/k was
value of the radial distribution function of the solvent, determined. This quantity resulted 217 K for the P Y
go12(r),around the solute considered as hard sphere. It treatment and 135 K for the PC treatment. Employing
is gol&) and consequently GoI2(z)which are calculated these values for el, the values of d2 for each gas at each
from the experimental radial distribution function of temperature were adjusted so that the P Y and PC
pure water measured by X-rays or neutron diffradion.'l treatments fitted the corresponding experimental
(9) S. A. Adelman and J. M. Deutch, J. Chem. Phys., 59,3971 (1977).
In (koH).
(10) D. Levesque, J. J. Weiss, and G. N. Patey, Mol. Phys., 38, 535
(1979). (13) J. L. Lebowitz, Phys. Reo. [Sect.] A 133, 895 (1964).
(11) L. R. Pratt and D. Chandler, J. Chem. Phys., 67, 3683 (1977). (14) J. A. Barker and D. Henderson, J.Chem. Phys., 47,2856 (1967).
(12) J. P. Hansen and I. R. McDonald, "Theory of Simple Liquids", (15) J. D. Weeks, D. Chandler, and H. C. Andersen, J.Chem. Phys.,
Academic Press, New York, 1976. 54, 5237 (1971); 55, 5422 (1971).
Vol. 18, 1985 Dissolution of Simple Gases in Water 209

Table 11.
Experimental Values of AXoHydAs Defined by Eq 6
-Ah0Hyd, AsoHydi -AC,OHyd,
T,K kJ.mol-' J.(mol.K)-' J.(mol-K)-'
298.2 9.08 45.6 205
323.2 10.00 31.8 142
373.2 11.18 17.3 67
423.2 11.87 11.6 28
473.2 12.40 9.7 8
a (5-7) (-14 t o +4)
Figure 1. Change of d, with temperature: open symbols, PC (IThese row quotes typical range for AXoHyd for nonaqueous
treatment; closed symbols, PY treatment; V,Xe; 0,Kr; 0,CH,; solvents.
A, Ar.

Figure 1 illustrates the changes of d2 with tempera- I t -

ture which were determined with each procedure. The
values calculated are quite reasonable when compared
with those predicted by theory.ll
These results indicate that perturbation theories
based on a hard-sphere reference fluid can be success-
fully applied to describe the dissolution of nonpolar 1',
.'\.. t
gases in a solvent like water. The experimental data
were fitted by using d2 as an adjustable parameter due
323 373 L23 L73 T/K
to the difficulties in knowing the intermolecular po-
tential of the pure solvent; this is an apparent limitation Figure 2. A W O H , . ~ and AS OH^^ against temperature: full curves,
of these theoretical treatments. However, on the basis experimental (the vertical bars represent the magnitude of the
of these calculations it is possible to predict several uncertainty); dashed curves, calculated with PY treatment using
temperature-independent d,'s.
features which are usually considered typical of aqueous
solutions and attributed to the peculiar structure of this nonaqueous solvents. The results obtained show that
solvent. in water ApoHyd is more negative as the temperature
Using the available cross-second virial coefficient data increases. This feature, as already noted," casts doubts
for gaseous binary mixtures of water with Ar, CHI, N2, on a straightforward interpretation of this quantity
and C2H6, Alvarez, Crovetto, and Fernandez-Prini8 have because the intermolecular structure of water is weaker
calculated a value of 220 f 10 K for q / k . The closeness at higher temperature and, consequently, any effect
of the value obtained from cross virial coefficients with related to it should decrease with temperature. On the
those calculated with the PY treatment is remarkable other hand, ASoHyd and ACpoHyd become smaller with
and probably partially fortuitous. Nevertheless it increasing temperature and probably closer to the
supports the contention that in the liquid the induced values in other solvents.
dipole-dipole interaction is negligible. On the other It is interesting now to evaluate how the PY pertur-
hand, the value obtained from gas-phase studies is bation theory explains the experimental results shown
larger than that calculated with PC. For this reason in Table 11. A / L O H ~ and
~ ASoHyd have been calculated
and the fact that the PC treatment involves a more by using temperature-independent d2's having the
complicated calculation and requires experimental values required by P Y theory to reproduce the exper-
X-ray data at all the temperatures, we employ only the imental In &OH) at 298 K for each gas. Figure 2 com-
PY method in what follows. pares a plot of the thermodynamic properties calculated
Hydrophobic Effect with a constant d2 to the experimental quantities. The
observed trends in the experimental AXoHydare cor-
From experimental measurements of the solubilities rectly described by the PY theory; this is especially clear
of CHI and C2H6 in water, it is possible to calculate the for the case of the entropy because it has been calcu-
thermodynamic quantities of dissolution of these two lated by differentiation of the PY chemical potential
gases in water up to 473 K.6b We have also calculated with respect to temperature without further adjustment
the quantity of d2.
PAPoHyd = PAP02(C2H6)- 2PAP0dCH.i) + In h / P ) Gas Transfer from Light to Heavy Water
(6)
A critical point for any theory which claims to de-
which, as suggested by Ben-Naim,16may be taken as scribe the thermodynamics of simple solutes in water
an indicator of the presence of hydrophobic effects. is its ability to predict the magnitude of the thermo-
[Strictly the definition of Ben-Naim involves another dynamic quantities corresponding to the transfer of the
term in eq 6 which contains the cross second virial solute between light and heavy water. According to the
coefficients of the gases and water. This term has been perturbation theories used to calculate thermodynamic
neglected because it is too small even at 473 K.] Table quantities of dissolution of gases in liquids, the mo-
I1 records values of A ~ O H , . and
~ related thermodynamic lecular properties of the solvent are expressed by the
quantities at various temperatures. The last row in parameters dl and el which should be the same for D 2 0
Table I1 gives the typical values of AX OH^^ found in and for H20.
The difference between H 2 0 and D20 as solvents for
(16) A. Ben-Naim,J. Chem. Phys., 54,1387 (1971); M. Yaacobi and a nonpolar gas are expected to be greater at lower
A. Ben-Naim, J. Phys. Chem., 78,175 (1974); A. Ben-Naim,J. Wilf, and
M. Yaacobi, J. Phys. Chem., 77, 95 (1973). temperatures. For this reason we have employed our
210 Fernandez-Prini et al. Accounts of Chemical Research
change of densities in the two liquids which is repre-
sented in the upper part of Figure 3. The agreement
between predicted and experimental A(pApo2) for
transfer is quite good. Unfortunately this conclusion
cannot be tested more accurately because the experi-
mental error in our high-temperature setup (hatched
1
-
region in Figure 3) is close to the expected value of
A(/3Apo2)(D20 H20) at high temperature. The gen-
eral trend of the experiment is predicted well although
more accurate determinations at higher temperatures
would be of considerable interest.
The successful account of the thermodynamics of
transfer of nonpolar spherical gases from light to heavy
water by the PY treatment is remarkable. According
, I to the theory this is due to the differences in the num-
300 400 500 ber density and their temperature dependences in both
T: K
liquids.
Figure 3. Chemical potential for transfer of gaseous solutes from The arguments which explain the small solubility of
HzO to D20 as a function of temperature: full curves, experi- nonpolar gases in H 2 0 by the extensive hydrogen
mental; heavy dashed curve, calculated with PY treatment for bonded structure prevailing in this liquid would natu-
Ar; the hatched region gives the magnitude of experimental error;
Kr;18 0,Ar;180,CH4;lsV, C2H6;18 0 , CH4;ls V, CzH6." rally lead to expect an even smaller solubility in D20,
sirice this is generally considered as being more struc-
tured than H20. This expectation is contrary to ob-
servation, because the nonpolar gases are more soluble
at low temperature in D20. This apparent paradox and
the relative increase of solubilities in D 2 0 compared to
H20 for temperatures higher than 400 K are correctly
elrplained by PY perturbation theory.
Dissolution of Gases in Nonaqueous Media
i '\\
1

It seems important to test the applicability of the PY

theory in describing the behavior of solutions of gases
45 ~
in nonpolar solvents where no "structural effects"
should be expected. Thus we have examined the solu-
bility of methane in n-heptane, which is known over a
very wide temperature range (between 183 and 511 K).
The In (koH)data reported by Gunn, Yamada, and
Whitmanlg have been compared to those calculated
Figure 4. In (koH) against (1/n for CH, dissolved in n-C7HlG: with the PY perturbation theory over the entire ex-
full curve and points, experimental data;lgdashed curve, calculated perimental temperature range. We have employed the
with PY treatment using temperature-independent dz.
value of In (koH)at 310.94 K to determine (tl/k) = 525
data combined with those reported by Battino et al.17 K with the PY treatment using dl = 6.25 A, d2 = 3.70
for Ar and CH4from 278 to 520 K in order to calculate A, and ( t 2 / k )= 157 K. We have performed the calcu-
the change of chemical potential of these gases upon lations assuming that d2 was temperature independent.
transfer to heavy water. These data are given as full The effect of such an assumption is illustrated in the
curves in Figure 3 together with those for Xe and Kr curves given in Figure 4. The calculated values of
obtained for the high temperature range in our labo- Henry's constants agree quite well with the experi-
ratory.& Figure 4 also contains the data for Ar, Kr, and mental data between 230 and 400 K, and the maximum
CHI reported by Cosgrove and Walkley16for low tem- of the curve is correctly reproduced. Above and below
peratures as well as data for C2H6.16J8 The experi- that temperature range, however, the calculated solu-
mental values are compared in Figure 3 with those bility is larger than that observed. We have not tried
calculated with the PY perturbation theory for Ar to fit the results more closely by adjusting tl and d2
(dashed curve) using the same values of the molecular because the most important cohclusion of this com-
parameters for both solvents and the value of d 2 ( T ) parison is attained with this siinple calculation: the PY
found for Ar in H20. As the solute size increases, the perturbation theory is very successful in explaining the
calculated curves are displaced upwards up to 0.05 unit solubility of CH4in n-heptane o b r all the experimental
for a change in size equivalent to that between Ar and temperature range. The values of In (KOH) obtained with
CzH6* PY result from a rather delicate cancelation of the
The calculated change i,n Apo2 for Ar when it is contribution of the cavity and thd attractive terms in
transferred from D20 to H 2 0 passes through a shallow eq 2 which are very close in magnitude.
minimum; its value increases again at higher tempera- Figure 5 shows the difference in experimental ASo2
tures. This unexpected behavior is due to the relative and ACpo2 for methane dissolution in H 2 0 and in n-
heptane as function of the reduced temperature. The
(17)E. Wilhelm, R. Battino, and R. J. Wilcock, Chem. Reu., 77, 219
(1977);T.R. Rettich, Y. P. Handa, R. Battino, and E. Wilhelm, J. Phys.
Chem., 85, 3230 (1981). (19)R. D. Gunn, T. Yamada, and D. Whitman, MChE J., 20, 906
(18) B. A. Cosgrove and J. Walkley, J.Chromatogr., 216, 161 (1981). (1974).
Vol. 18, 1985 Dissolution of Simple Gases in Water 211

12.3
Figure 5. ASo2 and ACpoz for CHI dissolved in water and in
n-heptane as function of the reciprocal reduced temperature
(Tc/TL 2 4 6 8
r/A"
figure shows clearly the different behavior encountered Figure 6. Radial distribution functions: 0,experimental for
in both solvents which is usually attributed to the water,22- - -, hard-sphere fluid with d = 2.70 A and pd3 = 0.70;
structural effects present in the aqueous media. As the -+, solute-water radial distribution function according to PC
temperature increases ASo2 and ACPo2for water and treatment" at 298 K (d, = 2.70 A). The number of f i t neighbors
n-heptane solvents approach each other. However, even calculated for each distribution function are shown under each
curve.
at 500 K the difference between water and heptane
stands out clearly. These differences are correctly
predicted by the PY perturbation treatment. water close to room temperature using the self-con-
The origin of the bell-shaped curve of In &OH) against sistent method.21 In Figure 6 we compare the radial
1/ T for CHI in n-C7HI6is quite different from that for distribution functions of pure water (298 K)22with that
the aqueous system. For water the change of sing in of its hard-sphere analogue. There we also exhibit re-
AHo2 is due to the rapidly increasing contribution of sults based on the PC?l method which gives the radial
the cavity formation term which is zero at the tem- distribution function of water molecules surrounding
perature where the expansivity of water is zero. For a hard-sphere solute having its same diameter. It must
n-heptane the change in sign of the enthalpy of disso- be borne in mind that the PC method starts from the
lution is due to an appreciable decrease in the attractive experimental value of the radial distribution function
contribution with increasing temperature which origi- of the solvent. As a consequence it would be expected
nates in the reduction of the number density of solvent to reflect the structural properties of water. The great
molecules with temperature. It is usually considered similarity of the curves obtained for a hard-sphere fluid
that the dissolution of solutes in water is governed by and for a hard-sphere solute immersed in actual water
the large negative values of ASo2 which characterizes is remarkable. The similarity probably holds the an-
the dissolution process. According to PY theory the swer to the question posed at the beginning of this
magnitude of ASo2 in water results from the balance section. Furthermore, the number of first neighbors
between the large work necessary to create a cavity in calculated with both curves are very similar and, ob-
liquid water (manifested in a large bo2 due to the large viously, very different from that corresponding to a
size ratio R in eq 3) and the very small enthalpy of central water molecule. That is, a hard-sphere solute
cavity formation in the vicinity of room temperature immersed in water has practically the same number of
as a consequence of the s m d expansivity of the solvent. first neighbor H 2 0 molecules as a particle of a pure
Thus ASo2 is large and negative due to a large Apo2 and hard-sphere liquid having the same number density and
a small AHcav2,rather than due to a structural rear- size as water. Moreover the calculated number of first
rangement of the solvent itself. This may be inter- neighbors compares very well with those determined by
preted as implying that the introduction of small non- Monte and molecular dynamic21 simulation
polar spherical solutes in water does not entail a sub- studies of water with nonpolar solutes (usually noble
stantial reorientation of the surrounding water mole- gases or methane).
cules. With use of the Pratt and Chandler method, we have
also determined the number of first neighbors in water
Concluding Remarks as a function of solute size. The results show that for
We have performed additional calculations in order a solute having zero diameter the number of first
to understand more clearly how is it that a perturbation
theory based on a hard-sphere reference fluid can ac- (21)T. M. Reed and K. E. Gubbins, "Applied Statistical Mechanics",
count for the results of nonpolar gas dissolution in McGraw-Hill-Kogakusha, Tokyo, 1973.
water. In particular we have calculatedm the radial (22)A. H. Narten and H. D. Levy, J. Chen. Phys., 55,2263(1971);A.
H. Narten, M. D. Danford, and H. A. Levy, Discuss. Faraday SOC.,43,
distribution function of a hard-sphere fluid having both 97 (1967).
the diameter (2.70 A) and number density of liquid (23)G.Bolis and E. Clementi, Chem. Phys. Lett., 82, 147 (1981);G.
Alagona and A. Tani, J. Chem. Phys., 72,580(1980);S . Swaminathan,
S.W. Harrison, and D. L. Beveridge, J. Am. Chem. SOC.,100,5705(1978).
(20)R. Fernindez-Prini, Monogr. Acad. Nac. Cs. Ex. Ha. Nat. (Bue- (24)A. Geiger, A. Rahman, and F. H. Stillinger, J . Chem. Phys., 70,
nos Aires), 2, 19 (1982). 263 (1979).
212 Acc. Chem. Res. 1985, 18, 212-219
neighbors H20molecules is 7.2; this quantity increases very different from that of the solvent molecules sur-
with d2 leading to the numbers reported in Figure 6 and rounding a lattice point where a central H20 molecule
in simulation studies for simple gaseous solutes. is sited. It seems that in many instances reorientation
The dissolution of nonpolar gases in water is suc- of water molecules has been invoked because this point
cessfully reproduced by the hard-sphere perturbation has been overlooked.
treatments for the simple reason that dissolution entails In order to verify our description of the dissolution
the creation of a cavity which starts growing from a of nonpolar solutes in water, the following points could
natural void in the liquid. Viewed from the positions be studied either by simulation or experiment: (1) the
occupied by the nonpolar solutes, water does not differ average orientation of the first H20neighbors to empty
too much from a hard-sphere fluid having the same points in the pure liquid, compared to the distribution
molecular size and number density. Furthermore, the of first neighbor water molecules surrounding spherical
temperature dependence of the dissolution properties nonpolar solutes of variable size (including d2 = 0); (2)
is correctly given if the experimental change of density accurate determination of the thermodynamics of
with temperature for the pure solvent is employed. transfer of an inert gas from D20 to H20 over a wide
In other words, the dissolution of small nonpolar temperature range; (3) the average total electric field
solutes in water does not involve significant reorienta- on a nonpolar spherical molecule.
tion of the neighboring water molecules. If this were
the case, a hard-sphere perturbation theory would not Note Added in Proof. Recently Wood and co-
be capable of describing the thermodynamics of the workers (J. Phys. Chem., in press) have determined
process over a wide temperature range. This is in experimentally Cpo2for Ar in H20up to 578 K. These
agreement with the view expressed above that the large important experimental data can be successfully re-
negative ASo2 is not the origin of the low solubility in produced by the perturbation treatment presented here
aqueous medium, but rather the consequence. The by using the same molecular parameters (work to be
orientation of H20 molecules surrounding an empty submitted for publication). This is strong evidence in
point or a nonpolar solute in liquid water is obviously favor of the views expressed in this article.

The Single Electron Shift as a Fundamental Process in

Organic Chemistry: The Relationship between Polar and
Electron-Transfer Pathways
ADDYPROM
Department of Chemistry, Ben Gurion University of the Negev, Beer Sheva, Israel
Received January 14, 1985 (Revised Manuscript Received May 14, 1985)

The concept of electronic reorganization is a funda- N:- + R-X -* N. + (R-X)- (2)

mental tenet of chemistry. Chemical reactions come
about through the rearrangement of valence electrons. nucleophilic substitution reactions into either polar or
It is these electrons that hold the atoms together in the SET pathways is equally applicable to all the other
reacting molecules and which, by redistribution, govern fundamental organic mechanisms. Thus hydride re-
the structure of the product molecules. duction, electrophilic, and nucleophilic aromatic sub-
Yet a strange dichotomy concerning the movement stitution, and nucleophilic addition, to quote a few ex-
of electrons during chemical reactions has emerged over amples, may also be formulated in terms of either SET
the years. In electron-transfer theory a fundamental or polar mechanisms.
postulate is that electrons may only be transferred one Recognition of the SET pathway in organic reactions
at a time.lV2 But, within organic chemistry, for polar has grown enormously over the last 20 years, primarily
reactions at least, electrons are considered to move following the pioneering work of Kornblum, Russell,
about in pairs, as illustrated by the sN2 process (eq 1). - ~ impact of SET pathways on or-
and B ~ n n e t t . ~The
ganometallic chemistry has been extensive, largely due
N:--RL\ - N-R + X- (1) to the contributions of Kochi and co-worker~.~,~ More
It is only for radical producing reactions that the idea
of a single electron transfer (SET) is widely entertained,
(1) Semenov, N. N. "Some Problems in Chemical Kinetics and
as illustrated by the first step of the S R N l pathway for Reactivity";Princeton University Press: Princeton, 1958.
nucleophilic substitution (eq 2).3 This subdivision of (2) For a recent excellent review on organic electrontransfer reactions,
see: Eberson, L. Adu. Phys. Org. Chem. 1982, 18, 79.
Addy Pross received his B.Sc. and Ph.D. degrees tom Sydney University. (3) Bunnett, J. F. Acc. Chem. Res. 1978, 11, 413.
After postdoctoral posltlons at Kings College, London, and the Hebrew Unlv- (4) Kornblum, N. Angew. Chem., Znt. Ed. Engl. 1976,14, 734.
erslty, Jerusalem, He was appolnted to the Department of Chemlstry, Ben (5) Russell, G. A.; Danen, W. C. J. Am. Chem. So;. 1966, 88, 5663;
Qurion University, where he holds the poshlon of Associate Professor. He 1968,90, 347.
has held visiting appointments in a number of lnstltutbns including the Aus- (6) Kim, J. K.;Bunnett, J. F. J. Am. Chem. Soc. 1970,92,7463,7464.
tralian National University, University of Auckiand, and Stanford Unlversity. (7) Application of electron transfer concepts in organic chemistry by
His current research interests Include apppiicatiins of qualitative VB and MO Soviet workers has been popu1p.r for many years. For reviews, see: (a)
theory to mechanistic organic chemistry and the elucidation of transitlon-state Todres, 2.V. Russ. Chem. Reu. (Engl. Traml.) 1978,47,148; (b) Bilevich,
structure and its application to general problems of organic reactivity. K. A.; Okhlobystin, 0. Y. Russ. Chem. Reu. (Engl. Transl.) 1968, 37, 1.

0001-4842/85/0118-0212$01.50/0 0 1985 American Chemical Society