Reactive Intermediates -55- Prof. Bernhard Jaun
5. Solvation and Solvent Effects
5.1 Solvent polarity scales
Solvents are grouped according to their properties into the following main classes:
1. Protic polar solvents
Examples: Water, alcohols, acids, primary and secondary amides.
Properties: high dielectric constant, hydrogen bond donor and acceptor. Strong solvation of cations
2. Aprotic polar solvents
Examples: HMPT, DMSO, DMF, acetonitrile, nitromethane, tert-amines
Properties: high to medium dielectric constants; good hydrogen bond acceptors; solvation of cations is
good; solvation of anions is very weak.
3. Aprotic nonpolar solvents
Examples: aliphatic alicyclic and aromatic hydrocarbons, halogenated hydrocarbons, ethers.
Properties: low dielectric constant, inefficient solvation of small hard ions (exception: polyethers such
as dimethoxyethane, glymes). Aromatic solvents are known to form donor-acceptor complexes with
suitable solutes. Halogenated solvents are quite good solvents for medium to soft cations such as
In view of the many different types of possible interactions between the solvent and individual solute
molecules/ions, no single polarity scale is suitable for all purposes.
Generic polarity scales
This polarity scale is based on the wavelength of the charge transfer band in the visible spectrum of A
(solvatochromicity; from 453 nm in H2O to 810 nm in diphenyl ether):
Reactive Intermediates -56- Prof. Bernhard Jaun
The ET-parameter is defined as the energy of this transition in kcal/mol. The rates of many types of
reaction, including SN2, obey a linear relation based on ET of the solvent
log k = a•ET + b
The slope of such plots is an indicator for the change in the polarity upon going from the reactant
ground state to the transition state. It will be positive if the TS is more polar than the reactants and
negative in the opposite case.
This scale is based on the solvatochromicity of B and runs practically parallel to the ET scale (formula
for conversion: Z = 1.41 ET + 6.92).
This polarity scale, which was introduced by Winstein, measures the influence of the solvent on the
rate of SN1 type solvolysis reactions.
Therefore, Y is only defined for protic polar solvents. Y is defined as the log of the ratio between the
rate of solvolysis of t-butyl chloride in a given solvent and the corresponding rate in 80% aqueous
Y = log k(solvent) - log k(80% EtOH)
The sensitivity of the rate of a particular reaction to solvent polarity is expressed as the coefficient m
log = mY
By definition, solvolysis of t-butyl chloride has m=1. Typical SN1 reactions have 0.7 < m < 1.2;
SN2 reactions have m < 0.5.
Reactive Intermediates -57- Prof. Bernhard Jaun
4. The Hildebrand solubility parameter δ
The internal pressure c of a pure solvent is defined as the energy (per molar volume) needed to
vaporize the solvent to a vapor pressure of 0 Pascal.
ΔHv - RT ΔHv =enthalpy of vaporisation to 0 pressure
Vm = molar volume
Therefore, c reflects the energy needed to make a “hole” in the solvent that is large enough to ac-
commodate a solute molecule. Hildebrand has defined his solubility parameter as δ = c and has
shown that substances are optimally miscible if they have the same δ value. If the difference between
the δ-values of solute and solvent differ by more than 3, one can no longer expect any solubili-
Solvents with a high internal pressure (high δ) make reactions with positive ΔV more difficult. At least
for nonpolar reactions, where no special solvation effects are to be expected, variation of the solvent
and plotting log k vs. δ allows to crudely estimate ΔV and serves as a “poor man’s” substitute for the
correct measurement of ΔV with high pressure equipment.
5.2 Ion Pairs and Solvolysis
From todayʼs perspective, we need to discriminate between contact ion pairs, solvent separated ion
pairs and free ions. Some authors also include solvent-shared ion pairs into the (actually somewhat
continuous) spectrum of interactions between ions of opposite charge in solution.
contact solvent sharing solvent separated dissociated
Transitions between these species can be observed by physical and spectroscopic methods, such as
13 6 23
conductivity or C-, Li and Na-NMR. In his classical work during the 1960s, Winstein has demon-
strated convincingly that ion pairs are a reality and have to be taken into account in order to explain
the kinetics of solvolysis reactions.
In the following, we give a summary of the developments and follow the arguments of, first, Ingold and
then Winstein (for a contemporary review see: S. Winstein et al. Chem. Soc. Spec. Publ., No. 19,
Reactive Intermediates -58- Prof. Bernhard Jaun
1. The normal salt effect:
Reactions producing ions (or a charge separation in the transition state) show a rate increase if the
ionic strength of the medium in increased. At first approximation, this effect is usually linear:
k = ko + b [salt]
Coefficient b is a measure of the sensitivity of the reaction towards ionic strength.
2. The kinetics of SN1-reactions according to Ingold:
R–X R+ + X– RO–S + H+ + X–
If the first step, the dissociation into ions, is reversible we can apply the steady state approximation for
the concentration of the carbenium ion and the expression for the overall rate is:
d[RX] k2 k1
– = with k2 = k2' [SOH]
dt k2 + k-1 [X-]
Because, during the reaction, more and more X is generated, the rate decreases with increasing
turnover number. If, on the other hand, we deliberately add an excess of a salt M X , the concentration
of X will be independent of turnover and one observes pseudo first order kinetics (with a reduced
apparent rate constant). This phenomenon is called “common ion rate depression” (CIRD).
At the same time, however, addition of an excess of salt M X will also cause a normal salt effect
which accelerates the reaction:
– = with k2 = k2' [SOH] k1 = k0(1+b[X-])
dt k2 + k-1 [X-]
Whether the normal salt effect or CIRD dominates depends on the reactivity of the carbenium ion: for
highly reactive R , k–1 is comparable to k2ʼ. Because [SOH] >> [X ], the normal salt effect will dominate
and an increase of the overall rate is observed.
+ – –
For more stable R , k–1 >> k2ʼ (since X is charged while SOH is neutral, X is expected to be the better
nucleophile than SOH) and the overall rate decreases because the CIRD effect dominates.
3. Ion pairs. So far (up to Ingold), it was assumed that solvolysis generates free ions directly. The
results obtained with the following solvolysis reaction prompted Winstein to postulate the formation of
ion pairs as intermediates:
Reactive Intermediates -59- Prof. Bernhard Jaun
H3C H3C H3C
CH3 Cl CH2 CH3
CH3 OAc CH3
The overall rate of this solvolysis in aqueous acetic acid/acetate buffer was determined by acid/base
titration (consumption of acetate). For the solvolysis of A, pseudo-1 order kinetics was observed and
addition of Cl did not lead to a CIRD.
The rate of solvolysis of B was, at first, higher than that for A but then gradually decreased and
eventually approached the rate found for A. If B was isolated back at incomplete turnover, it was found
that, in parallel to solvolysis, B had also rearranged to A (which explains the change in rate with time).
The crucial argument of Winstein was that, since external Cl did not show a CIRD, the rearrangement
of B into A could not have gone via the free dimethylallyl cation but that there must be an additional
intermediate which is not accessible for external chloride ions but allows 1,3 migration of chloride to
give A. This led to the postulate of a contact ion pair.
CH3 Cl CH3 CH3
A contact ion pair B
CH3 OAc CH3
Reactive Intermediates -60- Prof. Bernhard Jaun
4. Solvent separated ion pairs. In order to explain the phenomena observed in the solvolysis of
substrates that racemize in parallel to solvolysis, Winstein was eventually forced to introduce an
additional intermediate, termed Solvent separated ion pair, into the kinetic scheme:
contact ion pair solvent separated free ions
k1 k2 k3
RX R+X– R+ || X– R+ + X–
k-1 k-2 k-3
k4 SOH k5 SOH
Bs = Br S brosylate
OBs contact ion pair
H CH k1
H k-1/2 H
OBs Me H
CH3 k4 solvent seperated ion pair
H Me H
H k3 k-3
H CH +
H3C H free ions
products of acetolysis Me
Of course, Winstein and coworkers first tried to analyze the kinetics of this system based on their
original scheme with only one intermediate, the contact ion pair, between reactant and free ions. This
amounts to leave out rate constants k2, k–2 and k4 from scheme I above.
Reactive Intermediates -61- Prof. Bernhard Jaun
Experimentally, two rates could be followed: the rate of racemization of the starting material kα , and the
overall rate (acetate consumption) kt.
Using the steady state approximation for both, contact ion pair and free carbenium ion gives:
d[AcO- ] k1 k3 k5
– = [ reactant]
dt k5 (k-1 + k3 ) + k-1k-3 [BsO– ]
k1 k3 k5
i.e. kt =
k5 (k-1 + k3 ) + k-1k-3 [BsO– ]
Looking at this expression, one would expect that addition of BsO should lead to a CIRD effect. This
was, however, not observed. Therefore, we have to assume that
k5(k-1 + k3) >> k-1k3[BsO ]
k1 k3 kα k-1
i.e. kt = k1 = kα and = +1
k-1 + k3 kt k3
In other words: the ratio between kα and kt should vary if the ratio between k–1 and k3 changes. Addi-
tion of an “innocent” salt such as LiClO4 should decrease k-1 and increase k3 through a normal salt
effect (higher ionic strength favors the direction of dissociation). In fact (see curve C in the plot below)
the ratio between kα and kt did decrease when LiClO4 was added. But a plot of [LiClO4] vs. rate
revealed that this increase was far from linear (as would be expected for a normal salt effect) and
leveled off at quite low [LiClO4], with kα/kt becoming constant again at higher [LiClO4].
Salt effects in SN1 solvolysis
LiClO4 A= normal salt effect
special salt effect
kt B= CIRD
C= special salt effect
normal salt effect
and induced CIRD
0 [salt] 0.05 M
Winstein interpreted this puzzling result as follows: there must be an additional intermediate between
the contact ion pair and the free carbenium ion. In contrast to the contact ion pair, this second inter-
mediate must be accessible for counter ion exchange: brosylate can be exchanged by external
Reactive Intermediates -62- Prof. Bernhard Jaun
perchlorate which makes the backwards reactions k–2 and k–1 impossible and increases the overall
rate kt. Winstein also realized that this hypothesis can be checked: if ClO4 can replace BsO in this
new intermediate, addition of BsO should reverse this effect. Indeed, addition of BsO after the
special salt effect had first been induced with LiClO4 cancelled the latter (downwards curve in the plot).
This CIRD, which is only seen in the presence of LiClO4, was termed the induced CIRD.
The complete scheme with contact ion pair and solvent separated ion pair as intermediates gives the
following analytical expression for the overall rate kt:
k1k2 k4 +
k5 + k-3[BsO–]
(k-1 + k2 ) k4 + + k-1k-2
k5 + k-3[BsO–]
Analyses of this equation for conditions compatible with the absence of a normal CIRD leads to four
Case I : k4 >> k3 product formation from the solvent seperated ion pair
is much faster than dissociation into free ions.
No free carbenium ions are formed at all.
(k-1 + k2) k4 + k-1k-2
Case II : k5 >> k-3 [BrO- ] product formation from free ions is so fast that the
solvent seperated ion pair is never reformed
from the free ions
k1k2(k4 + k3 )
(k-1 + k2)( k4 + k3)+ k-1k-2
Case III : k-3 [BrO- ] >> k5 and k3 ≈ k4 products are exclusively formed from
the solvent separated ion pair, never from
the the free carbenium ion.
kt = (as in case I)
(k-1 + k2) k4 + k-1k-2
Case IV : k-1 ≈ 0 and k-2 ≈ 0 No re-association to contact ion pair or
starting material from the solvent seperated ion pair
k-1 + k2
(this is equivalent to the case without the solvent
seperated ion pair as an additional intermediate)
Reactive Intermediates -63- Prof. Bernhard Jaun
Only cases I and III make it necessary to include the solvent separated ion pair into the kinetic
scheme. Case IV can be checked for by back-isolation of starting material after having added either
labeled brosylate or e.g. tosylate instead of brosylate. Case II requires that the free carbenium ion is
very reactive. Winstein excluded it on the basis of the reactivity-selectivity principle by showing that
acetate and azide were reacting not according to their concentrations (statistically) but according to
their nucleophilicities. See however the comments on the RSP earlier in chapter 3).
For the reactant discussed here, Winstein concluded that only cases I or III could apply making the
solvent separated ion pair a necessary component in the kinetic scheme. Winstein and others investi-
gated many other substrates. Sometimes case IV applies, whereas other cases showed the normal
CIRD as described by Ingold. In essence, the more stable the carbenium ion, the more likely it is that
solvent separated ion pairs and free ions are formed before the carbenium ion is trapped by the
5. The rates of dissociation and racemization are not necessarily equal. In the example shown below
it was demonstrated that kα< k1:
H 18O O O
contact ion pair inversion
still chiral inversion
H 18 18O-exchange
C CH CH
O retention 18 O 18
The rearrangement in the contact ion pair to give inversion requires an “anion above the cation→anion
below the cation” movement, whereas the oxygen exchange requires only a small translation of the
carboxylate group in the plane. Since O exchange requires only a minimal movement but can occur
only when the bond to carbon is broken, it is probably the experimentally accessible rate that comes
closest to the true dissociation rate k1.
Reactive Intermediates -64- Prof. Bernhard Jaun