124
Chemistry 1983
leagues at the University of Chicago. By the time of the meeting, at the
University of Notre Dame, a meeting I did not participate in, I appreciated in a
general way the special advantages which the Cr
3+/2+
couple offered in the
investigation of the mechanism of electron transfer, and outlined my ideas to N.
Davidson when he visited me en route to the meeting, but the experimental
work which led to the first two papers (16, 17) was not done until 1953. In the
interim, I had failed to interest any of my graduate students in the work,
because, of course, no one foresaw what it might lead to, and because it seemed
less exciting than other work in progress in my laboratories, much of it
concerned with isotopic effects, tracer and kinetic. The bulk of the work
reported in the two papers just cited was done by my own hands; I shall now
outline the background for those early experiments.
My interest in coordination chemistry did not develop until I elected it as a
topic for an advanced course given soon after coming to the University of
Chicago. Instead of using the standard textbook material, I used as major
source the relevant volume of the reference series by Gmelin in which the
chemistry of the cobaltammines is described. At this time I already had a good
background in the literature devoted to substitution at carbon and understood
the issues raised in that context, and I soon became interested in raising the
same issues for substitution at metal centers. Furthermore, it was evident that
the complexes based on metal ions which undergo substitution slowly were
readily amenable to experimental study. I became curious as well about the
reasons underlying the enormous difference in rates of substitution for metal
ions of the same charge and (approximately) the same radii. The ideas that
resulted were presented in my course the next time it was given, but the
extensive literature study that led to the paper published in February, 1952
(18) was not done until 1949 when I was on leave from the University of
Chicago as a Guggenheim Fellow.
In this paper, a correlation with electronic structure was made of observa-
tions, mainly qualitative (“labile” complexes arbitrarily defined as those whose
reactions appear to be complete on mixing and “inert” as those for which
continuing reaction can be observed), for complexes of coordination no. 6. To
make the correlation, it was necessary to break away from the practice which
was common in the USA of classifying complexes as “ionic” and “covalent”
according to electronic structure. Thus, for example, the comparison of the
affinities of Cr(III) and Fe(III), the latter high spin and thus “ionic”, con-
vinced me that in the Fe(III) complexes, the bonds to the ligands might
actually be somewhat more covalent than in the Cr(III) complexes. Further-
more, it appeared to me that in earlier discussions of relative rates of substitu-
tion, where attempts were made to understand the observations in terms of the
existing classification, there was a failure to distinguish between thermodynam-
ic stability, and inertia, the latter being understood as referring solely to rate.
The affinity of Cl
-
for Cr
3
+(aq) is considerably less than it is for Fe
3+
(aq), yet
the aquation rate of CrCl
2+
(aq) is much less than it is for FeCl
2+
(aq). Rates, of
course, cannot be accounted for by considering ground state properties alone,
but the stability of the activated complex relative to the ground state must be
H. Taube
125
taken into account. When the effect of electronic structure on the relative
stabilities is allowed for, a general correlation of rates with electronic structure
emerges. (In the language of ligand field theory, for complexes of coordination
number 6, substitution tends to be slow when the metal ions have each of the
orbitals, but none of the
(anti-bonding) orbitals occupied.
The specific rates of exchange of water between solvent and the hexaaquo ions
o f
a n d
a r e
a n d
1 X 1
respectively ( 19)).
A shortcoming of this early effort is that such rationalizations of the correla-
tion as were offered were given in the language of the valence bond approach to
chemical binding, because, at the time I wrote the paper, I did not understand
the principles of ligand field theory even in a qualitative way. The valence bond
approach provides no simple rationale of the difference in rates of substitution
for labile complexes, and these have been found to cover a very wide range,
thanks to a pioneering study by Bjerrum and Poulsen (20), in which they used
methanol as a solvent to make possible measurements at low temperature, and
those of Eigen (21), in which he introduced relaxation methods to determine
the rates of complex formation for labile systems.
Activated complexes have compositions and structures, and it is necessary to
know what these are if rates are to be understood. It is hardly likely that these
features can be established for the activated complexes if they are not even
known for the reactants, and in 1950 we were not certain of the formula for any
aquo cation in water. It seemed to me important, therefore, to try to determine
the hydration numbers for aquo cations. Hydration number as I use it here
does not mean the average number of water molecules affected by a metal ion
as this is manifested in some property such as mobility, but has a structural
connotation: how many water molecules occupy the first sphere of coordina-
tion? Because the rates of substitution for Cr
3+
(aq) were known to be slow (22),
J. P. Hunt and I undertook to determine the formula for Cr
3+
(aq) in water
(23), with some confidence that we would be successful even with the slow
method we applied, isotopic dilution using
18
O enriched water. That the
formula turned out to be Cr(H
2
O )
6
3+
was no great surprise - although I must
admit that at one point in our studies, before we had taken proper account of
isotopic fractionation effects, Cr(H
2
O )
7
3+
was indicated and, faced with the
apparent necessity, I was quite prepared to give up my preconceived notions. It
was also no great surprise that the exchange is slow (t
1/2
at
ca. 40 hr). Even
so, the experiments were worth doing. They were the first of their kind, and
they attracted the attention also of physical chemists, many of whom were
astonished that aquo complexes could be as kinetically stable as our measure-
ments indicated, and were impressed by the enormous difference in the resi-
dence time of a water molecule in contact with a cation, compared with water
molecules just outside. That we dealt with hydration in terms of detailed
structure rather than in terms of averaged effects may have encouraged the
introduction into the field of other methods, such as nmr, to make the distinc-
tion between cation-bound and free water (19). As I will now detail, it also led
directly to the experiments described in the papers of references 16 and 17.