Using incoming nucleophile primary hydrogen - Deuterium kinetic isotope effects to model the SN2 transition state

Article abstract of DOI:10.1021/ja000441i

The primary hydrogen - deuterium incoming nucleophile KIEs for the S_N2 reactions between parasubstituted benzyl chlorides and borohydride ion in DMSO at 30.000 ± 0.002 °C are small (≤1.14) and insensitive to a change in substituent at the α-carbon. The small Hammett ρ (0.51) and ρ^r (-0.52) values found when the para substituent on the benzene ring of the substrate is altered indicate there is very little charge on the α-carbon in the transition state. The large, constant secondary α-deuterium KIEs of 1.089 ± 0.002 and the large chlorine leaving group KIEs of 1.0076, 1.0074, and 1.0078 found for the p-methyl-, the p-hydrogen-, and the p-chlorobenzyl chloride reactions suggest that the transition states for these reactions are unsymmetric with short H - C_α and long B - H and C_α - Cl bonds. The decrease in the chlorine leaving group KIE from 1.0076 ± 0.0003 for the p-methylbenzyl chloride reaction to 1.0036 ± 0.0003 for the p-nitrobenzyl chloride reaction indicates the C_α - Cl bond shortens markedly when a strongly electron-withdrawing substituent is on the α-carbon. Unfortunately, the bond strength hypothesis is the only theory that predicts the changes observed in transition-state structure and it only indicates the bond that changes but not how the transition-state structure is altered.

Full text of DOI:10.1021/ja000441i

7342
J. Am. Chem. Soc. 2000, 122, 7342-7350
Using Incoming Nucleophile Primary Hydrogen-Deuterium Kinetic
Isotope Effects To Model the SN2 Transition State
Terry Koerner, Yao-ren Fang, and Kenneth Charles Westaway*
Contribution from the Department of Chemistry and Biochemistry, Laurentian UniVersity,
Sudbury, Ontario, Canada P3E 2C6
ReceiVed February 4, 2000
Abstract: The primary hydrogen-deuterium incoming nucleophile KIEs for the SN2 reactions between para-
substituted benzyl chlorides and borohydride ion in DMSO at 30.000 ( 0.002 °C are small (e1.14) and
insensitive to a change in substituent at the R-carbon. The small Hammett F (0.51) and Fr (-0.52) values
found when the para substituent on the benzene ring of the substrate is altered indicate there is very little
charge on the R-carbon in the transition state. The large, constant secondary R-deuterium KIEs of 1.089 (
0.002 and the large chlorine leaving group KIEs of 1.0076, 1.0074, and 1.0078 found for the p-methyl-, the
p-hydrogen-, and the p-chlorobenzyl chloride reactions suggest that the transition states for these reactions are
unsymmetric with short H-CR and long B-Η and CR-Cl bonds. The decrease in the chlorine leaving group
KIE from 1.0076 ( 0.0003 for the p-methylbenzyl chloride reaction to 1.0036 ( 0.0003 for the p-nitrobenzyl
chloride reaction indicates the CR-Cl bond shortens markedly when a strongly electron-withdrawing substituent
is on the R-carbon. Unfortunately, the bond strength hypothesis is the only theory that predicts the changes
observed in transition-state structure and it only indicates the bond that changes but not how the transition-
state structure is altered.
Introduction
the nitrogen incoming nucleophile KIEs for the SN2 reactions
between anilines and pyridines with methyl and ethyl halides
and arenesulfonates. Unfortunately, these KIEs were too small
to show how changes in the structure of the substrate, the leaving
group, or the nucleophile affected the Nu-CR bond in the SN2
transition state. Matsson et al.11 overcame this problem by
measuring the 11C/14C incoming nucleophile KIEs (the mass
difference of 3 between the isotopes increased the magnitude
of the KIE) for the SN2 reactions between labeled cyanide ion
and para-substituted benzyl chlorides. Their 11C/14C KIEs, which
were large (the smallest KIE was 1.0070 ( 0.0008) compared
to the error in the KIE, showed that the NC-CR transition-
state bond was slightly shorter when a more electron-withdraw-
ing substituent was at the R-carbon.
Because isotopes of hydrogen have the largest percentage
change in mass possible, they produce the largest KIEs. This
led us to measure the primary hydrogen-deuterium incoming
nucleophile KIEs for the SN2 reactions between sodium boro-
hydride and several para-substituted benzyl chlorides in DMSO
at 30.000 °C, eq 2, to determine whether these KIEs would be
Despite a large volume of research, substituent effects on the
structure of the SN2 transition state, eq 1, are not understood.1-4
The problem is that most of the studies5,6 have only demon-
strated how the CR-LG bond changes and since one needs to
know how substituents affect both the reacting bonds (the total
transition-state structure), more methods for estimating the
length of the Nu-CR transition-state bond are needed. The best
method of determining how the Nu-CR transition-state bond
is changed by substituents is by measuring an incoming
nucleophile KIE, and in fact, nitrogen incoming nucleophile
KIEs have been measured for a variety of SN2 reactions.7-10
Ando and co-workers8 and Kurz and co-workers,9,10 measured
* Correspondingauthor: (fax)(705)-675-4844;(e-mail)KWestawa@Nickel.
Laurentian.CA.
(1) Westaway, K. C.: Jiang, W. Can. J. Chem. 1999, 77, 789.
(2) Westaway, K. C.; Fang, Y.-R.; Persson, J.; Matsson. O. J. Am. Chem.
Soc. 1998, 120, 3440.
(3) Westaway, K. C.; Waszczylo, Z. Can. J. Chem. 1982, 60, 2500.
(4) Westaway, K. C.; Ali, S. F. Can. J. Chem. 1979, 57, 1354.
(5) Hargreaves, R. T.; Katz, A. M.; Saunders, W. H., Jr. J. Am. Chem.
Soc. 1976, 98, 2614.
(6) Shiner, V. J., Jr.; Wilgis, F. P. In Isotopes in organic chemistry;
Buncel, E., Saunders, W. H., Jr., Eds.; Elsevier Science Publishing Co.
Inc.: New York, 1992; Vol. 8, pp 289-292, 303-305.
(7) Smith, P. J.: Westaway, K. C. In Supplement F. The chemistry of
amino, nitroso, and nitro compounds and their deriVatiVes; Patai, S., Ed.;
Wiley Interscience; Toronto, 1982; p 1277.
(8) Ando, T.; Yamataka, H.; Wada, E. Isr. J. Chem. 1985, 26, 354.
(9) Kurz, J. L.; Daniels, M. W.; Cook, K. S.; Nasr, M. M. J. Phys. Chem.
1986, 90, 5357
useful for monitoring the changes in the Nu-CR bond in an
SN2 reaction. Finally, a combination of (i) primary hydrogen-
deuterium incoming nucleophile KIEs, (ii) chlorine leaving
group KIEs, (iii) secondary R-deuterium KIEs, and (iv) Hammett
F values have been used to learn how the transition state changes
when the substituent at the R-carbon is changed in these SN2
reactions.
Results and Discussion
The objectives of this study were to determine whether (i)
primary hydrogen-deuterium incoming nucleophile KIEs could
(10) Kurz, J. L.; Pantano, J. E.; Wright, D. R.; Nasr, M. M. J. Phys.
Chem. 1986, 90, 5360
(11) Matsson, O.; Persson, J.; Axelsson, B. S.; Langstrom, B.; Fang,
Y.; Westaway, K. C. J. Am. Chem. Soc. 1996, 118, 6350.
10.1021/ja000441i CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/14/2000

Use of H-D Effects To Model the SN2 Transition State
J. Am. Chem. Soc., Vol. 122, No. 30, 2000 7343
Table 1. Rate Constants and Primary Hydrogen-Deuterium KIEs
for the SN2 Reactions between Borohydride Ion or Borodeuteride
Ion and Para-Substituted Benzyl Chlorides in DMSO at 30.000 °C
be used to monitor the changes in the Nu-CR bond in SN2
reactions and (ii) to learn how a change in substituent at the
R-carbon affected the total structure of the transition state in
the borohydride ion-para-substituted benzyl chloride reactions.
Mechanism of the Reaction. Although the reaction in eq 2
occurs,12,13 it was necessary to prove these reactions occur by
an SN2 mechanism. Early investigators13-16 found that the
NaBH4 reductions of alkyl halides were second order and
concluded they were SN2 processes. Bell and co-workers, on
the basis of rate-structure profiles, also concluded that these
reactions occurred by an SN2 mechanism.15,17 Although Ashby
and co-workers concluded that the reduction of primary alkyl
bromides and iodides by several metal hydrides occurs by a
single-electron-transfer (SET) mechanism, they were unable to
find any evidence suggesting that alkyl chlorides react by the
SET mechanism and concluded that primary and secondary alkyl
chlorides react by an SN2 pathway.18,19
para
substituent
kH × 103
1.468 ( 0.032
1.360 ( 0.002
kD × 103
(kH/kD)R
1.228 ( 0.033b
1.243 ( 0.009
CH3c
CH3
1.195 ( 0.019a
1.094 ( 0.008
av
1.236 ( 0.01d
H
1.317 ( 0.013
1.239 ( 0.031
1.290 ( 0.017
1.052 ( 0.017
0.988 ( 0.093
1.032 ( 0.032
1.252 ( 0.020
1.254 ( 0.020
1.250 ( 0.017
Hc
Hc
av
1.252 ( 0.002
Cl
1.808 ( 0.012
1.780 ( 0.015
1.889 ( 0.014
1.465 ( 0.015
1.437 ( 0.029
1.517 ( 0.012
1.234 ( 0.015
1.239 ( 0.027
1.245 ( 0.014
Clc
Clc
av
1.239 ( 0.006
NO2c
NO2
3.318 ( 0.010
3.049 ( 0.049
2.644 ( 0.067
2.422 ( 0.042
1.255 ( 0.032
1.259 ( 0.029
The results from this study confirm that all the para-
substituted benzyl chloride-borohydride ion reactions occur by
an SN2 mechanism. All the reactions are second order, first order
in substrate and first order in the nucleophile so both the
nucleophile and the substrate are involved in the rate-determin-
ing step of the reaction. More importantly, the primary incoming
nucleophile hydrogen-deuterium and chlorine leaving group
KIEs found for these reactions (vide infra) indicate both Nu-
CR bond formation and CR-LG bond rupture occur in the
transition state of the rate-determining step of the reaction; i.e.,
that they are all SN2 processes.
Primary Hydrogen-Deuterium KIEs. The primary hydro-
gen-deuterium KIEs for the SN2 reactions between sodium
borohydride and para-substituted benzyl chlorides were deter-
mined by reacting sodium borohydride or sodium borodeuteride
with the para-substituted benzyl chloride in DMSO at 30.000
( 0.002 °C. The average primary hydrogen-deuterium KIEs,
Table 1, are surprising for two reasons. They are (i) all small
and (ii) insensitive to a change in substituent at the R-carbon;
i.e., the average KIE is 1.246 ( 0.010. Small primary
hydrogen-deuterium KIEs have been attributed to a bent20 and/
or an unsymmetric transition state.21,22 To determine which
factor is responsible for the small KIE, the maximum primary
hydrogen-deuterium isotope effect for the SN2 reaction between
borohydride ion and benzyl chloride was estimated using eq 3,
av
1.257 ( 0.003
a Standard deviation for the average rate constant. b The error is
1/kD[(∆kH)2 + (kH/kD)2(∆ kD)2]1/2,where ∆kH and ∆kD are the standard
deviations for the average rate constants for the reactions of the
undeuterated and deuterated nucleophiles, respectively.3 c Measured
in a different batch of DMSO. d Standard deviation of the average KIE.
B-H(D) stretching vibration in the transition state. Substituting
the frequencies of the asymmetric stretching vibrations for the
B-H and B-D bonds in the reactant (νB-H ) 2272 cm-1 and
νB-D ) 1721 cm-1, respectively)23 into eq 3 suggests the
maximum primary hydrogen-deuterium KIE will be ∼3.7 at
30 °C. The observed kH/kD ) 1.25 is only 34% of the maximum
KIE of 3.7. Moreover, the observed hydrogen-deuterium KIE
is the product of a primary hydrogen-deuterium KIE and a
secondary hydrogen-deuterium KIE, eq 4, The (kH/kD)Prim is
(kH/kD)Obs ) (kH/kD)Prim(kH/kD)Sec
(4)
primarily determined by the change in the asymmetric stretching
vibration of the transferring hydron, and the (kH/kD)Sec is
determined by the changes in the B-H(D) vibrations of the
nontransferring hydrons as the reactant is converted into the
transition state. The latter KIE arises because the boron changes
from sp3 to sp2 hybridization as the hydride ion is transferred
to the R-carbon. Because the observed KIE is the product of a
secondary and a primary hydrogen-deuterium KIE, (kH/kD)Prim
must be smaller than the observed KIE of 1.25.
-
νB-D
)
kH/kD ) e-hc/2kT (ν
(3)
B-H
Yamataka and Hanafusa24 measured 12C/14C R-carbon and
hydrogen-deuterium KIEs for the reaction between benzophe-
none and NaBH4, eq 5, and used a BEBOVIB-IV calculation
which assumes the complete disappearance of the asymmetric
(12) Hutchins, R. O.; Hoke, D.; Keogh, J.; Koharski, D. Tetrahedron
Lett. 1969, 40, 3495.
(13) Brown, H. C.; Murray, K. J.; Snover, J. A.; Zweifel, G. J. Am. Chem.
Soc. 1960, 82, 4233.
(14) (a) Brown, H. C.; Tierney, P. A.; J. Am. Chem. Soc. 1958, 80, 1552.
(b) Volpin, M. Dvolaitzky, M.; Levitin, I. Bull. Soc. Chim. Fr. 1970, 4,
1526.
[5]
(15) Bell, H. M.; Vanderslice, C. W.; Spehar, A. J. Org. Chem. 1969,
34, 3923.
(16) Brown, H. C.; Krishnamurthy, S. J. Org. Chem. 1980, 45, 2250.
(17) It is worth noting that the second-order rate constants calculated
from Bell and co-workers’ gas chromatographic study of these reactions at
room temperature15 agree (within 10%) with those found in this study. Given
the difficulty in reproducing rate constants in different batches of DMSO
and the small temperature difference between the two studies, the results
of the two studies are in reasonable agreement.
(18) Ashby, E. C.; DePriest, R. N.; Goel, A. B.; Wenderoth, D.; Pham,
T. N. J. Org. Chem. 1984, 49, 3545.
(19) Ashby, E. C.; Pham, T. N. J. Org. Chem. 1986, 51, 3498.
(20) Hawthorne, M. F.; Lewis, E. S. J. Am. Chem. Soc. 1958, 80, 4296.
(21) Westheimer, F. H. Chem. ReV. 1961, 61, 265.
(22) Melander. L. Isotope effects on reaction rates; Ronald Press: New
York, 1960; p 24.
to match the transition-state structure with the observed KIEs.
Their calculations suggested that the secondary hydrogen-
deuterium KIE increases dramatically from 1.1 to 1.7 as the
CR-H transition-state bond order increases from 0.10 to 0.90.
This presumably occurs because as the CR-H bond order in
the transition state increases, the hybridization at boron changes
from sp3 to sp2, and there is less steric hindrance to the
nontransferring B-H(D) out-of-plane bending vibrations. As a
(23) Shirk, A. E.; Shriver, D. F. J. Am. Chem. Soc. 1973, 95, 5901.
(24) Yamataka, H.; Hanafusa, T. J. Am. Chem. Soc. 1986, 108, 6643.

7344 J. Am. Chem. Soc., Vol. 122, No. 30, 2000
Koerner et al.
Figure 1. A productlike (a) and reactant-like (b) unsymmetric form
for the hydride-transfer portion of the transition state.
result, the zero-point energy difference in the transition state is
small relative to that in the initial state and a large secondary
KIE is observed. The results reported by Yamataka and
Hanafusa can be used to estimate the magnitude of the maximum
primary hydrogen-deuterium KIE in the borohydride ion-
benzyl chloride reactions because (i) the vibrations involved in
the reduction of benzophenone to benzhydrol are similar to those
involved in the reduction of benzyl chloride to toluene. The
only differences will be at the R-carbon because the B-H(D)
vibrations in NaBH4 and NaBD4 are the same in both reactions.
The symmetric and asymmetric CR-H stretching vibrations
Figure 2. Westheimer-Melander curve for the hydron transfer
showing the maximum primary hydrogen-deuterium KIEs for linear
and bent hydride transfer transition states.
to become broader, Figure 2. This means the position of the
hydride in the transition state can vary markedly without
changing the primary hydrogen-deuterium KIE. Therefore, if
the hydron-transfer portion of the transition state is sufficiently
bent, the primary hydrogen-deuterium KIEs would be insensi-
tive to a change in substituent at the R-carbon, even if the change
in substituent causes a relatively large change in transition-state
structure. Ab initio calculations27 on the methyl chloride-
borohydride ion reaction at the 6-31+G* level indicate that the
transfer of hydride between boron and carbon is nonlinear with
a B-H-CR bond angle of 144°. It is worth noting that
calculations at a higher level are expected to yield a larger
B-H-CR bond angle27 and that a looser transition state is
expected for the benzyl chloride reaction. However, a B-H-
CR bond angle of 144° suggests that the maximum primary
hydrogen-deuterium KIE would be reduced from 3.7 to ∼2.7.
Even if the maximum KIE is 2.7, the small primary hydrogen-
deuterium KIE of e1.14 is still very small with respect to the
maximum value. Therefore, it is clear that the hydron-transfer
portion of the SN2 transition state must be very unsymmetric,
even if it is bent.
formed in these reactions are found at 2930 and 2890 cm-1
,
respectively, for benzhydrol25 while they are at 2960 and 2930
cm-1, respectively, for toluene.25 (ii) Changing the solvent to
DMSO from 2-propanol will not affect this comparison because
Yamataka and Hanafusa concluded that the solvent does not
contribute significantly to the KIE either through hydrogen
bonding to the hydride ion or through nucleophilic assistance
in the hydride transfer. Since the dipolar aprotic solvent, DMSO,
cannot form hydrogen bonds, it should not affect the reaction
or the KIE. (iii) The calculations performed by Yamataka and
Hanafusa and the calculation to determine the maximum KIE
in this study both assumed that the hydride-transfer transition
state is linear.
If the range of secondary deuterium KIEs in the benzophe-
none reduction (between 1.1 and 1.7) is applied to the benzyl
chloride reactions, the observed hydrogen-deuterium KIE of
1.25 corresponds to a primary deuterium KIE of between 1.25/
1.1 ) 1.14 and 1.25/1.7 ) 0.74. The actual value will, of course,
depend on the amount of hydron transfer from B to CR in the
transition state. The Westheimer-Melander theory21,22 states that
the maximum primary hydrogen-deuterium KIE (3.7) is
observed when the hydrogen is symmetrically located (50%
transferred) between the two heavier atoms in the transition state
and that a smaller KIE is found when the hydron-transfer
transition state is unsymmetric. Therefore, if the transition state
is linear, the small primary hydrogen-deuterium KIE of e1.14
suggests the hydride is unsymmetrically located, i.e., either very
near the boron or the R-carbon in the transition state, Figure 1.
Another explanation for the small primary hydrogen-
deuterium KIE is that the hydride-transfer transition state is not
linear. In a linear transition state, the ZPE difference in the
transition state is zero and the KIE is a maximum when the
hydride is symmetrically located.21,22 In a nonlinear, symmetric
transition state, the hydride will not be motionless and the
vibrational energy will be dependent on isotopic substitution.
Hence, the ZPE difference will be larger in a bent transition
state and the KIE will be smaller. Calculations by More
O’Ferrall26 suggest that small deviations of 5° from linearity
have little effect on the KIE. However, deviations greater than
5° cause the magnitude of the KIE to decrease significantly and
the curve relating the position of the hydron in the transition
state to the magnitude of the primary hydrogen-deuterium KIE
The second problem is to explain why the primary hydrogen-
deuterium KIEs (the B-H-CR portion of the transition state)
do not change with the substituent. Westaway28 proposed a bond
strength hypothesis for SN2 reactions which suggests there will
be a significant change in the weaker reacting bond but little or
no change in the stronger reacting bond in an SN2 transition
state when a substituent at the R-carbon is changed. For the
borohydride ion-para-substituted benzyl chloride SN2 reactions,
the CR-H bond stretching vibration for the product (toluene)
is at 2960 cm-1 while the CR-Cl stretching vibration for the
reactant (benzyl chloride) occurs at 650 cm-1 25
Because the
.
stronger CR-H reacting bond in the SN2 transition state should
not change significantly with substituent, the identical primary
hydrogen-deuterium KIEs that are found in these reactions are
expected.
Although it seems certain that the hydron-transfer portion of
the transition state is unsymmetric, it is not known whether the
hydride ion is more or less than 50% transferred to CR in the
transition state. However, a productlike transition state with a
short H-CR bond is favored because in all the SN2 reactions
where the transition state is unsymmetric, the strongest reacting
bond is short and the weaker reacting bond is long.29 Therefore,
one would expect that the stronger H-CR bond would be short
and the weaker CR-Cl bond would be long in the hydride ion
SN2 portion of the transition state. Another reason for faVoring
(27) Saunders, W. H., Jr.; Westaway, K. C. Private communication.
(28) Westaway, K. C. Can. J. Chem. 1993, 71, 2084.
(29) Westaway, K. C.; Pham, T. V.; Fang, Y. J. Am. Chem. Soc. 1997,
119, 3670.
(25) Dyer, J. Application of Absorption Spectroscopy of Organic
Compounds; Prentice Hall: Englewood Cliffs, NJ, 1965; pp 25-36
(26) More O’Ferrall, R. A. J. Chem. Soc. B 1970, 785.

Use of H-D Effects To Model the SN2 Transition State
J. Am. Chem. Soc., Vol. 122, No. 30, 2000 7345
a productlike hydron-transfer portion of the transition state is
that if the bond strength hypothesis can be applied to the hydride
ion-transfer portion of the transition state (a nucleophile is being
transferred between two electrophiles, B and CR, whereas in an
SN2 transition state, an electrophile is being transferred between
two nucleophiles), one would expect the B-H-CR transition
state to be productlike. The stretching vibration for the B-H
bond in sodium borohydride23 and the CR-H bond in toluene25
are at 2294 and 2960 cm-1, respectively. If the bond strength
hypothesis is applied to this hydride-transfer transition state,
one would expect the stronger CR-H reacting bond to be shorter
than the B-H reacting bond (a productlike transition state).
Although this is only speculation, it is worth noting that
Yamataka and Hanafusa proposed a productlike transition state
with B-H and H-CR bond orders of 0.25 and 0.75, respectively,
for the hydride-transfer reaction between borohydride ion and
benzophenone.24
Figure 3. Yukawa-Tsuno plot for the SN2 reactions between
borohydride ion and para-substituted benzyl chlorides in DMSO at
30.000 ( 0.002 °C.
Table 2. Frand F Values Found for SN2 Reactions between
Various Anionic Nucleophiles and Para-Substituted Benzyl
Chlorides
Hammett G Value. Because the rate constants for all the para-
substituted benzyl chlorides are greater than the rate constant
for benzyl chloride (see the kH rate constants in Table 1), a
U-shaped Hammett plot is observed for the SN2 reactions
between sodium borohydride and para-substituted benzyl chlo-
rides. Three explanations have been suggested for the curved
Hammett F plots found for the SN2 reactions between benzylic
substrates and anionic nucleophiles. One is a change from an
SN2 to a carbocation mechanism when a more electron-donating
substituent is on the substrate.30-32 The second is a change in
the relative importance of bond breaking and bond making in
the transition state,33-35 and the third is that the curvature is
due to a greater resonance effect of the +M (methyl and chloro)
substituents.3,36 The curvature in the Hammett plot is not due
to a change in mechanism because all of the substrates react by
an SN2 mechanism (vide supra). Nor is it due to a change from
a productlike to a reactant-like transition state because the
primary hydrogen-deuterium KIEs are identical for all the para-
substituted benzyl chloride reactions. Therefore, it must be due
to a greater resonance effect by the +M substitutents.
nucleophile
solvent
DMSO
temp (°C)
Fr
F
ref
borohydride ion
thiophenoxide ion LiOCH3/CH3OH
thiosulfate ion
thiosulfate ion
cyanide ion
30
20
30
30
30
20
-0.55 0.51
-1.67 0.94
-3.99 0.52 37
-0.96 0.67 37
-6.39 1.44 38
-4.08 0.89 39
a
3
60% aq acetone
60% aq acetoneb
80% aq dioxane
n-butyl sulfide ion LiOCH3/CH3OH
a This study. b The substrates in these reactions were para-substituted
m-chlorobenzyl chlorides.
p-hydrogen- and p-nitro-substituted benzyl chlorides (the p-
hydrogen and p-nitro substituents have little or no resonance
interaction with the benzene ring and their (σ+ - σn) values
are 0.000 and 0.012, respectively) gives a line with slope F of
0.51. When the other substituents were placed on this Hammett
plot, the points corresponding to the p-methyl and p-chloro
substituents show positive deviations from the F ) 0.51 line.
Plotting these deviations (∆log kX/kH) vs (σ+ - σn) for the
p-methyl and p-chloro substituents gives a line with a slope Fr
) - 0.55. When the values of F and Fr are used, the Yukawa-
Tsuno plot, Figure 3, has a high correlation coefficient of 0.994.
This confirms that the curvature in the Hammett plot is caused
by a variable resonance effect.
The variable resonance effect can be demonstrated by
applying the modified Yukawa-Tsuno expression,36 eq 6, to
log k/ko ) Fσn + Fr(σ+ - σn)
(6)
Although the F and Fr values are only approximate, a
comparison of the F and Fr values found for these reactions
with those found for other SN2 reactions between para-
substituted benzyl chlorides and anionic nucleophiles, Table 2,
allows the structure of the transition state for the borohydride
ion reactions to be elucidated in more detail. The small Fr )
-0.55 is between 0.5 (Fr ) -0.96 for the reactions with
thiosulfate ion) and 0.08 times (Fr ) -6.39 for the cyanide ion
reactions) the Fr values found for the other SN2 reactions. This
suggests there is little conjugation between the benzene ring
and the R-carbon in the transition state and that there is very
little charge on the R-carbon in the transition state. If this is
correct, the F value, which is primarily due to charge develop-
ment on the R-carbon in going to the transition state, should
also be small. In fact, the F value of 0.51 found in this study is
the smallest value; i.e., it is equal to the F value of 0.52 found
for the reactions with thiosulfate ion and is only 0.33 of the
largest F value (F ) 1.44 for the cyanide ion reaction) found
for SN2 reactions between benzyl chlorides and anionic nucleo-
philes, Table 2. This also suggests there is only a small charge
on the R-carbon in the transition states of these reactions and
supports the conclusion based on the Fr value.
the data. The F and Fr terms represent the sensitivity of the
reaction to the polar and resonance effects of the substituent,
respectively. The σ+ terms are substituent constants representing
the interaction of a substituent with a carbocation while the σn
terms represent the substituent’s polar (no resonance) effect on
the reaction site. If resonance between the R-carbon and the
benzene ring (Fr(σ+ - σn)) is important in a reaction series
where the substituent on the substrate is varied, Fr is large. Also,
if conjugation is important in the transition state, Fr will be larger
than F.3 A plot of log(kX/kH) vs σn for the reactions of the
(30) Stein, A. R. Tetrahedron Lett. 1974, 4145.
(31) Johnson, C. D. In The Hammett Equation; Cambridge University
Press: London. 1973; pp 1-50.
(32) Aronovitch, H.; Pross, A. Tetrahedron Lett. 1977, 2729.
(33) Hudson, R. F.; Klopman, G. J. Chem. Soc. 1962, 1062.
(34) Bowden, K.; Cook, R. S. J. Chem. Soc. B 1968, 1529.
(35) Swain, C. G.; Langsdorf, W. P., Jr. J. Am. Chem. Soc. 1951, 73,
2813.
(36) Deleted in proof.
(37) Fuchs, R.; Carlton, D. M. J. Org. Chem. 1962, 27, 1520.
(38) Hill, J. W.; Fry, A. J. Am. Chem. Soc. 1962, 84, 2763.
(39) Grimsrud, E. P. Ph.D. Dissertation, University of Wisconsin,
Madison, WI, 1971.
(40) Deleted in proof.
(41) Johnson, C. D. Chem. ReV. 1975, 75, 755.
Although the use of Hammett F values to suggest transition-

7346 J. Am. Chem. Soc., Vol. 122, No. 30, 2000
Koerner et al.
Table 3. Average Rate Constants and Secondary R-Deuterium
KIEs for the Sodium Borohydride SN2 Reactions with
Para-Substituted Benzyl Chlorides in DMSO at 30.000 ( 0.002 °C
para
(kH/kD)/
R-D
substituent
kH × 103
kD × 103
(kH/kD)R
CH3c
CH3
1.735 ( 0.015a 1.595 ( 0.004a 1.088 ( 0.010b 1.043
1.698 ( 0.003 1.553 ( 0.017 1.093 ( 0.012
1.045
av
1.091 ( 0.004d 1.044
H
1.102 ( 0.005 1.005 ( 0.018 1.096 ( 0.020
1.047
1.040
Hc
1.121 ( 0.011 1.035 ( 0.003 1.082 ( 0.01
1.089 ( 0.010
av
1.044
Figure 4. Possible unsymmetric SN2 transition states for the reaction
between borohydride ion and para-substituted benzyl chlorides.
Cl
1.806 ( 0.012 1.655 ( 0.009 1.091 ( 0.009
1.640 ( 0.009 1.502 ( 0.002 1.092 ( 0.006
1.045
1.045
Clc
av
1.092 ( 0.0007 1.045
state structure has been criticized,41,42 the Hammett F value
appears to be a valuable tool for determining the relative charge
density on atoms in the transition state in a series of similar
reactions because the substituent effects suggested by KIE data
and the Hammett F values are identical.3,4,43,44
NO2c
NO2
3.420 ( 0.093 3.160 ( 0.004 1.082 ( 0.029
1.040
1.043
3.471 ( 0.044 3.161 ( 0.049 1.088 ( 0.022
1.085 ( 0.004
av
1.042
a Standard deviation. b The error is 1/kD[(∆kH)2 + (kH/kD)2(∆
kD)2]1/2,where ∆kH and ∆kD are the standard deviations for the average
rate constants for the reactions of the undeuterated and deuterated
nucleophiles, respectively.3 c Measured in a different batch of DMSO.
d Standard deviation of the average KIE.
The primary hydrogen-deuterium KIEs indicated that the
hydride is unsymmetrically placed between the boron and the
R-carbon in the transition state. There are four possible linear,
unsymmetric, hydride ion-transfer transition states, Figure 4.
Transition state a with bond rupture ahead of bond formation
would have significant positive charge on the R-carbon and a
large negative Hammett F value would be found. Transition state
b has short and strong H-CR and CR-Cl bonds, there would
be significant negative charge on the R-carbon, and the F value
would be large and positive. The small F and Fr values found
in this study are obviously not consistent with either of these
transition states. However, small F and Fr values are consistent
with transition states c and d, where the R-carbon is unsym-
metrically located between the hydride and the chloride ions
and there is only a small charge on the R-carbon. Although the
small Hammett F and Fr values cannot distinguish between
unsymmetric reactant-like c and productlike d SN2 transition
states, a productlike transition state is more likely for two
reasons. First, the primary hydrogen-deuterium KIEs suggest
that the B-H-C portion of the SN2 transition states for these
reactions are productlike. Second, the small positive F value
suggests there is more bond formation than bond rupture in the
transition states.45
rate-determining, the (kH/kD)R would be ∼1.15 per R-deuteri-
um.48 Recent work has shown that a (kH/kD)R e 1.06 per
R-deuterium is indicative of an SN2 mechanism3,49-52 so the
secondary R-deuterium KIEs of between 1.042 and 1.045/R-D
found in this study confirm that these reactions are SN2
processes.
One explanation for the large secondary R-deuterium KIEs
observed for these hydride-transfer reactions is tunneling.
Klinman and co-workers53 measured primary and secondary
hydrogen-tritium (kH/kT) and deuterium-tritium (kD/kT) KIEs
for the oxidation of benzyl alcohol to benzaldehyde by yeast
alcohol dehydrogenase. Significant deviations from Saunders54
semiclassical relationship, eq 7, which neglects tunneling, were
kH/kT ) (kD/kT)3.26
(7)
observed for both the primary and the secondary KIEs.
Substituting the observed secondary R (kD/kT) of 1.03 ( 0.01
into eq 7 gives an expected semiclassical secondary R (kH/kT)
) 1.11 ( 0.02. However, the observed secondary R (kH/kT) )
1.35 ( 0.04 was significantly larger than the calculated value,
and Klinman’s group concluded it was the result of tunneling
by the transferring hydride. This tunneling affected the second-
ary hydrogen KIE because the motion of the transferring and
nonreacting hydrogens were coupled in the reaction coordinate.
Although the large secondary R-deuterium and small primary
hydrogen-deuterium KIEs observed in the borohydride ion-
para-substituted benzyl chloride reactions could arise because
the motion of the transferring hydride is coupled to the vibrations
of the benzylic hydrogens and tunneling is present, this is
unlikely for a several reasons. First, it is difficult to explain the
constant secondary R-deuterium KIEs in the presence of
tunneling. If the H-CR bond is identical in all the transition
states in the reaction series, as the primary hydrogen-deuterium
KIEs suggest, the contribution of tunneling must also be constant
throughout the series. This is unlikely because the probability
Secondary r-Deuterium KIEs
The secondary R-deuterium KIEs for the reactions between
sodium borohydride and the para-substituted benzyl chlorides
were determined by reacting sodium borohydride with the
unlabeled or the labeled para-substituted benzyl-1,1-d2 chloride
in DMSO at 30.000 ( 0.002 °C. All the average secondary
R-deuterium KIEs, Table 3, are large, i.e., ∼1.044 per R-D,
and do not change with substituent. The (kH/kD)R per R-deute-
rium for a carbenium ion SN reaction of an alkyl chloride, where
the first step is fully rate-determining, is 1.11.46-48 If dissociation
of the intimate ion pair or formation of the free carbocation is
(42) Hoz, S.; Basch, H.; Goldberg, M. J. Am. Chem. Soc. 1992, 114,
4364.
(43) Lee, I.; Lee, W. H.; Sohn, S. C.; Kim, C. S. Tetrahedron Lett. 1985,
41, 2653.
(44) Dubois, J. E.; Ruasse, M. F.; Argile, A. J. Am. Chem. Soc. 1984,
106, 4840.
(45) Miller, D. J.; Subramanian, Rm; Saunders, W. H., Jr. J. Am. Chem.
Soc. 1981, 103, 3519.
(46) Shiner, V. J., Jr. In Isotope Effects in Chemical Reactions; Collins,
C. J., Bowman, N. S., Eds.; ACS Publication 167; Van Nostrand Rhein-
hold: New York, 1970; pp 90-159.
(47) Shiner, V. J., Jr.; Rapp, M. W.; Halevi, E. A.; Wolfsberg, M. J.
Am. Chem. Soc. 1968, 90, 7171.
(48) Westaway, K. C. In Isotopes in organic chemistry; Buncel, E., Lee,
C. C., Eds.; Elsevier Science Publishing Co. Inc.: New York, 1987; p 312.
(49) Koshy, K. M.; Robertson, R. E. J. Am. Chem. Soc. 1974, 96, 914.
(50) Strecker, H.; Elias, H. Chem. Ber. 1969, 102, 1270.
(51) Westaway, K. C.; Lai., Z. Can. J. Chem. 1988, 66, 1263.
(52) Westaway, K. C. In Isotopes in organic chemistry; Buncel, E., Lee,
C. C., Eds.; Elsevier Science Publishing Co. Inc.: New York, 1987; pp
313-319.
(53) Cha, Y.; Murray, C. J.; Klinman, J. P. Science 1989, 243, 1325.
(54) Saunders, W. H., Jr. J. Am. Chem. Soc. 1985, 107, 164.

Use of H-D Effects To Model the SN2 Transition State
J. Am. Chem. Soc., Vol. 122, No. 30, 2000 7347
of tunneling should be smaller for the faster reactions because
more of the molecules will have the energy to cross over the
barrier rather than tunnel through it. Second, eliminating (kH/
kT)R from the semiclassical relationship proposed by Saunders,54
and the semiclassical relationship proposed by Swain and co-
workers,55
and a short CR-Cl bond, Figure 4c, is unlikely for several
reasons. First, the bond strength hypothesis28 suggests that the
stronger H-CR bond will not change significantly when a
substituent is changed in the substrate. This suggestion is
supported by the constant primary hydrogen-deuterium KIEs.
If the structure of the transition state changes with substituent,
as has been observed in every system that has been examined,29
a change in substituent must cause a change in the CR-Cl bond.
If the CR-Cl bond were short and the H-CR bond long in the
transition state, the change in the CR-Cl bond with substituent
would cause a significant change in the secondary R-deuterium
KIE. Since the secondary R-deuterium KIEs do not change with
substituent, the transition state cannot have a short CR-Cl bond.
The small positive Hammett F value of 0.51 is also not
consistent with transition state c. When bond formation lags
behind bond rupture, there would be a positive charge on the
R-carbon in the transition state and a small negative F value
would be found.
kH/kT ) (kH/kD)1.44
leaves a semiclassical expression,
kH/kT ) (kD/kT)3.26 ) (kH/kD)1.44
relating (kH/kD) with (kD/kT)
kH/kD ) (kD/kT)2.26
(8)
(9)
(10)
Because of the larger mass of deuterium and tritium, the
tunneling contribution to (kD/kT) is small and hence this KIE is
primarily due to the semiclassical ZPE differences between the
reactant and transition state. Because the semiclassical KIEs
are related by reduced-mass considerations and a carbon-
hydrogen bond is being formed in both systems, to a first
approximation, the systems can be compared. Substituting the
average secondary kD/kT ) 1.03/R-deuterium found by Klin-
man’s group in eq 10 gives a kH/kD ) 1.069/R-D. Since a kH/kD
> 1.069 suggests tunneling is important, the smaller (kH/kD)/
R-D of 1.044 found in this study suggests there is no tunneling
in the borohydride ion reactions. Two other factors suggest that
tunneling is not the cause of the large secondary R-deuterium
KIEs observed in this system. The probability of tunneling
increases as the energy of reactants and products approach one
another.56 Because the energy of the reactants in the sodium
borohydride-para-substituted benzyl chloride reactions is sig-
nificantly greater than the energy of the products, the probability
of tunneling is small. Tunneling is also unlikely because carbon
motion probably contributes to the reaction coordinate motion
in these unsymmetric transition states.57
A possible explanation for the large, and identical, secondary
R-deuterium KIEs found in these reactions is provided by the
work of Westaway and co-workers29 who suggested that, in
some transition states, the magnitude of the secondary R-deu-
terium KIE is not determined by the Nu‚‚‚LG distance58,59 but
rather by only the shortest (strongest) reacting bond in an
unsymmetric SN2 transition state. This occurs because the
weakest bond to the nucleophiles in the transition state is long
and any changes in this bond occur too far away from the
R-carbon to affect the CR-H(D) out-of-plane bending vibrations
and, hence, the KIE. In these cases, the secondary R-deuterium
KIE does not change with substituent. Thus, the primary
hydrogen-deuterium KIEs, the Hammett F and Fr values, and
the secondary R-deuterium KIEs all indicate the transition states
for these SN2 reactions are unsymmetric.
The productlike SN2 transition state d is favored for several
reasons. First, the small, positive Hammett F value of 0.51 is
consistent with the productlike SN2 transition state d where bond
making exceeds bond breaking.45 Second, in all the reactions
where it has been possible to determine the relative lengths of
the two reacting bonds in an unsymmetric SN2 transition state,
the shorter bond has been the stronger of the two reacting bonds
whether the substituent change is in the nucleophile, the
substrate, or the leaving group in an SN2 reaction.2,29 Because
the H-CR bond is much stronger than the CR-Cl bond, one
would predict that the H-CR bond will be short and the CR-
Cl bond will be long in these transition states. Also, the bond
strength hypothesis28 predicts there will be little or no change
in the stronger reacting bond but a significant change in the
weaker reacting bond in an SN2 transition state with a change
in substituent at the R-carbon. Since there will be little change
in the shorter, stronger H-CR reacting bond in the SN2 transition
state, the secondary R-deuterium KIE, which is determined by
the short, strong (CR-H) reacting bond in the transition state,
will not change with substituent as is observed. Transition state
d is also preferred because the secondary R-deuterium KIEs
are all very large for SN2 reactions with unsymmetric transition
states. The secondary R-deuterium KIEs for all the other SN2
reactions with unsymmetric transition states2,29 have small
secondary R-deuterium KIEs, i.e., close to unity. This presum-
ably occurs because the bond to one nucleophile is short and
the energy of the CR-H(D) out-of-plane bending vibrations is
not significantly different from that in the tetrahedral reactant
or product. As a result, the difference in the zero-point energy
between the reactant and the transition state is small and the
KIE is small. This makes the large secondary R-deuterium KIEs
observed in this system difficult to understand for an unsym-
metric transition state where either the H-CR or the CR-Cl
bond is short. However, it seems that large secondary R-deu-
terium KIEs could be found for these unsymmetric SN2
transition states because the hydride ion is very small. Even if
the H-CR transition state bond is short, the small incoming
hydride ion will be too small to affect the CR-H(D) out-of-
plane bending vibrations in the transition state significantly,
Figure 5. If this is the case, the energy of the CR-H(D) out-of
plane bending vibrations in the transition state will be small,
the ZPE difference between the reactant and the transition state
will be large and a large KIE will be found even though the
bond to the nucleophile is short in the transition state. Finally,
it is important to note that if the chloride ion were close to the
R-carbon in the transition state, the large chlorine would affect
Of the two possible unsymmetric SN2 transition states, Figure
4c and d, the unsymmetric transition state with a long H-CR
(55) Swain C. G.; Stivers, E. C.; Reuwer, J. F., Jr.; Schaad, L. J. J. Am.
Chem. Soc. 1958, 80, 5885.
(56) De La Vega, J. R. Acc. Chem. Res. 1982, 15, 185.
(57) Theoretical calculations by Saunders and co-workers (Table 2, ref
45) have shown that the contribution of heavy atom motion to the reaction
coordinate motion is more pronounced in unsymmetric transition states.
Heavy atom motion increases the effective mass of the transferring hydride
ion and reduces the probability of tunnelling.
(58) Poirier, R. A.; Wang, Y.; Westaway, K. C. J. Am. Chem. Soc. 1994,
116, 2528.
(59) Barnes, J. A.; Williams, I. H. J. Chem. Soc., Chem. Commun. 1993,
1286.

7348 J. Am. Chem. Soc., Vol. 122, No. 30, 2000
Koerner et al.
Table 4. Comparison of the Bending and Stretching Contributions
to the Secondary R-Deuterium KIE for Hydride Ion and Other
Nucleophiles with a Similar Nucleophile Leaving Group Distance in
the Transition State
methyl chloride
methyl fluoride
OH-
H-
% diff
F-
H-
% diff
(kH/kD)total
Nu‚‚‚LG distance 4.37
(kH/kD)bend
0.907 0.983
8.4
1.1
15.9
1.0
0.846 1.013
20.0
1.1
42.0
3.5
4.32
1.112 1.288
0.714 0.707
3.69
3.65
0.978 1.389
0.677 0.653
Figure 5. Transition state for the hydride ion SN2 reactions showing
that the hydride ion is too small to affect the CR-H(D) out-of-plane
bending vibrations significantly.
(kH/kD)stretch
Table 5. Chlorine Leaving Group KIEs for the Sodium
Borohydride SN2 Reactions with Para-Substituted Benzyl Chlorides
in DMSO at 30.000 ( 0.002 °C
k35/k37
trial 2a,b
para
substituent
trial 1a
average
CH3
H
Cl
1.0078 ( 0.00006c 1.0073 ( 0.0001c 1.0076 ( 0.0004d
1.0076 ( 0.0002
1.0082 ( 0.0014
1.0039 ( 0.0005
1.0072 ( 0.0003 1.0074 ( 0.0003
1.0074( 0.0006 1.0078 ( 0.0006
1.0033 ( 0.0009 1.0036 ( 0.0004
NO2
a The KIE was calculated according to the formula k35/k37 ) ln(1 -
f)/ln(1 - (Ro/Rf)f), where f is the average fraction of reaction, Ro is the
average 35Cl/37Cl ratio in five samples taken after the reaction reached
100% of completion, and Rf is the 35Cl/37Cl ratio in the chloride ion
collected after different fractions of reaction ranging from 9 to 25% of
completion.75 b The isotope effects were measured by two different
workers several months apart. c Standard deviation for the individual
measurements of the KIE from five different reactions. d Standard
deviation for the average KIE.
of the KIEs only differ by 3.5%). Because the small hydride
ion does not affect the CR-H(D) out-of-plane bending vibrations
in the transition state as much as the larger fluoride ion, Figure
5, one finds a larger secondary R-deuterium KIE in hydride
ion reaction even when the H-CR transition-state bond is short.
This means that transition state d, Figure 4, is possible whereas
transition state c with a short CR-Cl bond cannot display the
large secondary R-deuterium KIEs that are observed.
Chlorine Leaving Group KIEs. The best explanation of the
small and constant primary hydrogen-deuterium KIEs, the
small Hammett F and Fr values, and the large and constant
secondary R-deuterium KIEs is that the transition states for the
borohydride ion-para-substituted benzyl chloride reactions are
unsymmetric with short H-CR and long CR-Cl bonds. In fact,
the CR-Cl bond must be long enough so that the changes that
occur in the CR-Cl bond when the substituent on the substrate
is altered do not affect the CR-H(D) out-of-plane bending
vibrations (the magnitude of the secondary R hydrogen-
deuterium KIE) significantly. Also, since the H-CR bond
changes very little with a change in substituent on the R-carbon,
the changes in transition-state structure that occur when the
substituent is changed on the R-carbon must be in the CR-Cl
bond. The chlorine leaving group KIEs were determined for
these reactions in order to learn how the CR-Cl bond changes
with substituent.
Figure 6. Secondary R-deuterium KIE as a function of nucleophile
leaving group distance in the transition state for the SN2 reactions
between methyl chloride and fluoride and various nucleophiles.
the out-of-plane bending vibrations markedly and small second-
ary R-deuterium KIEs would be observed.
A theoretical study in which Poirier, Wang, and Westaway58,60
calculated the transition-state structures and secondary R-deu-
terium KIEs for a series of SN2 reactions between methyl
chloride and fluoride and various nucleophiles at the ab initio
6-31+G* level showed that the hydride ion reactions have much
larger than expected secondary R-deuterium KIEs given their
Nu‚‚‚LG distance in the transition state, Figure 6. For instance,
the hydroxide ion and hydride ion reactions with methyl chloride
have similar Nu‚‚‚LG distances in the transition states, but the
hydride ion reaction has a significantly larger KIE. This is
because the bending contribution to the KIE (the portion of the
KIE due to transition-state structure) is 15.9% larger for the
hydride ion reaction than for the hydroxide ion reaction, Table
4. This presumably occurs because the hydride ion does not
increase the energy of the transition-state CR-H(D) out-of-plane
bending vibrations as much as the larger hydroxide ion, Figure
5.60 As a result, the total secondary R-deuterium KIE for the
hydride ion reaction is much larger than expected given the Nu‚
‚‚LG distance in the transition state. This effect is even more
dramatic in the methyl fluoride reactions. Here, the hydride ion
and fluoride ion reactions have similar Nu‚‚‚LG distances in
the transition state. However, the bending contribution to the
KIE in the hydride ion reaction is 42.0% larger than that
observed for hydroxide ion reaction (the stretching components
With the exception of the p-nitrobenzyl chloride reaction, all
the average chlorine KIEs found in Table 5 are large; i.e., they
are ∼50% of the theoretical maximum chlorine isotope effect.61
Since the magnitude of the chlorine KIE increases with the
amount of CR-Cl bond rupture in the transition state,62,63 the
CR-Cl bonds are long in the transition states of the p-methyl-,
(61) Kirsch, J. F. In Isotope effects on enzyme catalyzed reactions;
Cleland, W. W., O’Leary, M. H., Northrop, D. B., Eds.; University Park
Press: Baltimore, MD. 1977; p 18.
(62) Saunders, W. H., Jr. Chem. Scr. 1975, 8, 27
(60) Koerner, T.; Westaway, K. C.; Poirier, R. A.; Wang, Z. Can. J.
Chem., in press.
(63) Sims, L. B.; Fry, A.; Netherton, L. T.; Wilson, J. C.; Reppond, K.
D.; Crook, S. W. J. Am. Chem. Soc. 1962, 94, 1364

Use of H-D Effects To Model the SN2 Transition State
J. Am. Chem. Soc., Vol. 122, No. 30, 2000 7349
Table 6. Predicted Changes in Transition-State Structure When a
More Electron-Donating Substituent Is Added to the R-Carbon in
the SN2 Reactions between Borohydride Ion and Para-Substituted
Benzyl Chlorides
the p-hydrogen-, and the p-chlorobenzyl chlorides. Also, the
CR-Cl bond in the p-methyl transition state is slightly longer
than the CR-Cl bond in the benzyl chloride transition state.
The slightly larger chlorine KIE in the p-chlorobenzyl chloride
reaction suggests the CR-Cl bond is slightly longer in this
transition state although this conclusion may be the result of
the larger error in this KIE. Surprisingly, the slightly larger KIE
in the p-chlorobenzyl chloride reaction has been observed in
another study39 but no explanation has been suggested and it is
not apparent why it is found.
The chlorine KIEs found by Hill and Fry for the SN2 reactions
between cyanide ion and para-substituted benzyl chlorides38
decrease from 1.0078 to 1.0060 while those found for the same
substrates in the borohydride ion reaction with the p-methyl-,
the p-hydrogen-, and the p-chlorobenzyl chlorides range from
1.0078 to 1.0074. The transition states in the cyanide ion-para-
substituted benzyl chloride system are thought to be unsym-
metric with a short NC-CR and a long CR-Cl bond.11 In fact,
the CR-Cl bonds in the cyanide ion reactions are so long that
the changes in the CR-Cl bond with substituent do not affect
the CR-H(D) out-of-plane bending vibrations in the transition
state (the magnitude of the secondary R-deuterium KIEs).11 The
equally large chlorine KIEs found in the borohydride ion
reactions must mean the CR-Cl bond in the p-methyl-,
p-hydrogen-, and p-chlorobenzyl chloride transition states is also
too long to affect the secondary R-deuterium KIEs.64 Thus, the
chlorine KIEs for these reactions are in agreement with the
interpretation based on the other KIEs.
change in transition-state bond on
adding a more electron-donating
substituent to CR
theory
H-CR
CR-Cl
longer
shorter
longer
shorter
longer
longer
longer
experimental results
unchanged
longer
longer
BEP principle (Hammond)
Thornton reacting bond rulea
Harris and Kurz ruleb
longer
More O’Ferrall-Jencks methoda longer
More O’Ferrall-Jencks methodc shorter
Pross-Shaik method
longer
bond strength hypothesisd
little or no change significant change
a The transition state is symmetrically placed in the energy surface.
b H is lighter than Cl and in different rows of the periodic table.
c Productlike transition state. d The H-CR bond strength is greater than
the CR-Cl bond strength.
p-methyl to p-chloro and that there is a big change in the CR-
Cl bond when the substituent changes from p-chloro to p-nitro.
We are unable to provide an explanation for this observation.
It is worth noting that the general change in transition-state
structure with substituent is consistent with the bond strength
hypothesis,28 which predicts there will be little change in the
stronger H-CR reacting bond and a large change in the weaker
CR-Cl reacting bond in the transition state when the structure
of the substrate is altered.
The chlorine KIE for the p-nitrobenzyl chloride is only 24%
of the maximum KIE so there is less CR-Cl bond rupture in
this transition state. It is also interesting that the secondary
R-deuterium KIE found for the p-nitrobenzyl chloride reaction
(1.085) is slightly smaller than those for the other para-
substituted benzyl chloride reactions. This smaller isotope effect
in the p-nitrobenzyl chloride reaction may have been found
because the CR-Cl the transition-state bond is short enough to
affect the CR-H(D) out-of-plane bending vibrations and reduce
the magnitude of the KIE.
The large decrease in the CR-Cl bond when the para
substituent is changed from methyl to nitro is consistent with
what had been suggested by the primary and secondary KIEs
and the Hammett F and Fr values, i.e., that the transition state
is unsymmetric with a short H-CR bond, which remains
constant for the series of para substituents, and a long CR-Cl
bond, which varies with a change in substituent at the R-carbon.
What is surprising is that the change in the CR-Cl bond with
substituent is very small when the substituent changes from
Theoretical Models. The experimental results indicate the
transition states for the borohydride ion-para-substituted benzyl
chloride reactions are productlike and that adding a more
electron-donating substituent to the benzene ring of the substrate
leads to a transition state with a longer CR-Cl bond and little
or no change in the H-CR bond. These results can be compared,
Table 6, to the predictions based on the theoretical models that
indicate how substituents change transition-state structure.
Since the rate of the reaction increases when both electron-
donating and electron-withdrawing substituents are added to the
R-carbon, it is not possible to apply Hammond’s thermal
postulate (the Bell-Evans-Polyani principle) to this system.
Thornton’s reacting bond rule,69 the More O’Ferrall-Jencks
energy surface method,70,71 and the Pross and Shaik model72
all predict that adding a more electron-donating substituent to
the R-carbon will lead to a transition state with longer H-CR
and CR-Cl bonds. Therefore, none of these models predicts
the experimental results. The More O’Ferrall-Jencks energy
surface method reduces to a parallel effect if the transition state
is productlike. Because a more electron-donating group on the
R-carbon stabilizes the higher energy reactant more than the
product, the parallel effect would lead to a transition state with
a shorter H-CR and a longer CR-Cl bond. This also is not
consistent with the experimental results. Harris and Kurz73
extended Thornton’s rule to include nucleophiles in different
rows of the periodic table. Because hydride and chloride are in
different rows, a more electron-donating substituent on the
R-carbon should lengthen the bond to the lighter nucleophile
and shorten the bond to the heavier nucleophile; i.e., the
transition state should have a longer H-CR bond and shorter
CR-Cl bond when a more electron-donating substituent is
present. This prediction is also not consistent with the experi-
(64) Hill and Fry’s chlorine KIEs might be smaller than expected for
the amount of CR-Cl bond rupture in the transition state because they were
measured in 20% aqueous dioxane rather than in DMSO and hydrogen
bonding to the developing chloride ion in the transition state could add
vibrational energy to the chlorine and reduce the KIE.65 If hydrogen bonding
has a significant effect on the KIE, the KIE versus percent CR-Cl bond
rupture in the transition-state curve would have a lower slope in 20%
aqueous dioxane than in DMSO. Although theoretical calculations by
Brubaker66 showed that hydrogen bonding to a developing chloride ion
decreased the magnitude of a chlorine KIE, several experimental studies
on SN2 reactions suggest the chlorine KIE changes very little with even
drastic changes in solvent.67,68 Although the effect of hydrogen bonding to
the developing chloride ion on the chlorine KIE is not known, it seems
likely that the CR-Cl bonds are long in all the transition states, except
perhaps the reaction with p-nitrobenzyl chloride (vide infra).
(65) Thornton, E. R. In SolVolysis mechanisms; Ronald Press: New York,
1964; pp 199-206.
(66) Brubaker, D. M. Ph.D. Dissertation, University of Arkansas,
Fayetteville, AR, 1978.
(67) Graczyk, D. G.; Taylor, J. W.; Turnquist, C. R. J. Am. Chem. Soc.
1978, 100, 7333.
(69) Thornton, E. R. J. Am. Chem. Soc. 1967, 89, 2915.
(70) More O’Ferrall, R. A. J. Chem. Soc., B 1970, 274.
(71) Jencks, W. P. Chem. ReV. 1972, 72, 705.
(72) Pross, A.; Shaik, S. J. Am. Chem. Soc. 1981, 103, 3702.
(73) Harris, J. C.; Kurz, J. L. J. Am. Chem. Soc. 1970, 92, 349.
(68) Westaway, K. C. Can. J. Chem. 1978, 56, 2691.

7350 J. Am. Chem. Soc., Vol. 122, No. 30, 2000
Koerner et al.
mental results. Finally, the bond strength hypothesis28 predicts
that changing the substituent on the R-carbon in this reaction
would not change the stronger H-CR bond but would lead to a
significant change in the weaker CR-Cl bond. Thus, the place
where the changes in transition-state structure occur is predicted
by the bond atrength hypothesis. However, this theory does not
predict how the transition-state bonds will change when an
electron-donating substituent is added to the R-carbon. Although
the bond strength hypothesis was able to predict which
transition-state bonds changed when a substituent was altered
at the R-carbon, in fact, none of the theoretical models were
able to predict the changes that occurred.
withdrawing substituent to the benzene ring on the R-carbon
leads to a transition state with a shorter CR-Cl bond and
unchanged H-CR and B-H bonds.
Finally, neither the Thornton’s reacting bond rule, the More
O’Ferrall-Jencks energy surface method, the Pross-Shaik
method, or the Harris-Kurz extension of Thornton’s reacting
bond rule predicts the observed change in transition-state
structure. The bond strength hypothesis predicts the bond that
changes with substituent but does not indicate how the transition-
state structure is altered.
Experimental Section
Conclusion
Preparation of Reagents. The DMSO used in this study was a
distilled in glass grade purchased from Caledon Laboratories Ltd.
However, each isotope effect measurement was done with the same
batch of solvent in order to reduce the effect small differences in the
water content of the solvent had on the rate constant. The sodium
borohydride and sodium borodeuteride were purchased from Aldrich
Chemical Co. and used with no additional purification. However, they
were stored in a vacuum desiccator once they had been opened. All
the para-substituted benzyl chlorides were purchased from Aldrich
Chemical Co. The benzyl chloride, p-methylbenzyl chloride, and
p-chlorobenzyl chloride were purified by vacuum distillation. The
p-nitrobenzyl chloride was recrystallized once from methanol and once
from n-hexane. The 1H NMR (200 MHz, CDCl3) spectra of the purified
substrates were as expected. A subsequent NMR analysis of the distilled
and recrystallized substrates indicated they were stable throughout the
study.
This study has shown that primary incoming nucleophile
hydrogen-deuterium KIEs can be measured, that they are large
with respect to the experimental error, and that they can be used
to determine the changes that occur in the H-CR transition-
state bond when a substituent is changed on the R-carbon.
The small, and identical, incoming nucleophile hydrogen-
deuterium KIEs of 1.25 found for the borohydride ion-para-
substituted benzyl chloride SN2 reactions in DMSO at 30.000
°C indicate that the hydride is unsymmetrically located between
the boron and the R-carbon in the transition state and that the
H-CR transition-state bond does not change with substituent
on the benzene ring.74 The small Hammett Fr value of -0.55
indicates there is little resonance between the R-carbon and the
benzene ring in the transition state while the small Hammett F
value of 0.51 suggests there is only a small negative charge on
the R-carbon in the transition state. These data suggest that the
transition states for these reactions are unsymmetric. Further-
more, the small positive Hammett F value indicates H-CR bond
formation is ahead of CR-Cl bond rupture, i.e., that the
unsymmetric SN2 transition states have short H-CR bonds that
do not change with substituent, and long CR-Cl bonds. The
large, and identical, (kH/kD)R ) 1.09 found for these reactions
are consistent with transition states with short H-CR and long
CR-Cl bonds. The change in the long CR-Cl bonds caused by
a change in substituent does not affect the CR-H(D) out-of-
plane bending vibrations so the magnitude of the secondary
R-deuterium KIE is only determined by the length of the short
H-CR bond rather than the H‚‚‚Cl transition-state distance.
Finally, these larger than expected secondary R-deuterium KIEs
suggest that the small hydride ion does not significantly affect
the CR-H(D) out-of-plane bending vibrations in the transition
state even though the H-CR bond is short. The reasonably large
(∼50% of the theoretical maximum KIE) chlorine leaving group
KIEs for the p-methyl-, p-hydrogen-, and p-chlorobenzyl
chlorides indicate the CR-Cl transition-state bond is long in
these transition states. The chlorine KIEs decrease when a more
electron-withdrawing substituent is on the R-carbon so the CR-
Cl bond changes significantly when the substituent on the
benzene ring is changed. Therefore, adding a more electron-
The preparation of the para-substituted benzyl-1,1-d2 chlorides has
been reported previously.3 The 1H NMR (200 MHz, CDCl3) spectrum
of benzyl-1,1-d2 chloride had a single absorption at a δ of 7.42 ppm.
Expansion of the region between 4.00 and 5.00 ppm, had a very small
peak at a δ 4.64 ppm corresponding to the benzylic protons. Analysis
of the peak areas suggested that the product was 98.5% deuterated at
the benzylic position. A similar analysis of the purified p-nitrobenzyl-
1,1-d2 chloride, the p-chlorobenzyl-1,1-d2 chloride, and the p-methyl-
benzyl-1,1-d2 chloride suggested that these substrates were 99.4, 99.6,
and 99.2%, respectively, deuterated at the benzylic carbon.
Kinetic Measurements. The detailed procedure used to measure
the rate constants for the primary and secondary hydrogen-deuterium
KIEs and the chlorine KIEs has been reported.75
The primary deuterium KIEs were measured as follows. Three
borohydride ion and three borodeuteride ion rate constants were
measured simultaneously, averaged, and divided to obtain the hydrogen-
deuterium KIE. Each KIE was determined in at least two different
batches of solvent. All the second-order plots were first order in both
the substrate and the nucleophile and linear up to 90% reaction with
correlation coefficients of g0.997. An NMR analysis of the products
showed that no side reactions occurred.
The secondary R-deuterium KIEs were determined from three
reactions of the labeled and three reactions of the unlabeled substrate
that were done simultaneously. The three rate constants for the unlabeled
(kH) and labeled (kD) substrates were averaged and divided to obtain
the secondary R-deuterium KIE. Each KIE was determined twice and
all the kinetic plots were linear to at least 90% of completion and had
correlation coefficients g0.998.
(74) A reviewer suggested the identical obserVed KIEs (eq 4) found when
borodeuteride ion was the nucleophile in these reactions did not necessarily
mean that the B-H-CR portion of the transition states were identical.
Although the exact magnitude of the primary hydride ion KIE in each
reaction is not known, the only way the identical obserVed KIEs found in
this system could not be related to transition-state structure is if a change
in transition-state structure led to an equal and opposite change in the
primary and secondary deuterium KIEs, i.e., if the secondary deuterium
KIE increased by the same amount as the primary deuterium KIE decreased
when the structure of the transition state changed. This is unlikely because
the secondary deuterium KIEs are significantly larger than the primary
deuterium KIEs in these reactions. In any case, the identical secondary
R-deuterium KIEs found for these reactions indicate that the H-CR bonds,
and probably the B-H-CR portion of all these transition states, are identical.
Acknowledgment. The authors gratefully acknowledge the
financial support provided by the Natural Sciences and Engi-
neering Research council of Canada (NSERC).
JA000441I
(75) Westaway, K. C.; Koerner, T.; Fang, Y.-R.; Rudzinski, J.; Paneth,
P. Anal.Chem. 1998, 70, 3548.