Hydride shift in substituted phenyl glyoxals: Interpretation of experimental rate data using electronic structure and variational transition state theory calculations

Article abstract of DOI:10.1039/b102330f

Experimental rate data for hydride transfer in some p-substituted phenyl glyoxals in 80 : 20 dioxan : water are presented including kinetic hydrogen isotope effects. The roles of different substituents are discussed using electronic structure calculations

Full text of DOI:10.1039/b102330f

View Article Online / Journal Homepage / Table of Contents for this issue
Hydride shift in substituted phenyl glyoxals: Interpretation of
experimental rate data using electronic structure and variational
transition state theory calculations¤
Gary Tresadern, Julie Willis, Ian H. Hillier* and C. Ian F. Watt
Department of Chemistry, University of Manchester, Manchester, UK M13 9PL
Received 12th March 2001, Accepted 20th June 2001
First published as an Advance Article on the web 28th August 2001
Experimental rate data for hydride transfer in some p-substituted phenyl glyoxals in 80 : 20 dioxan : water are
presented including kinetic hydrogen isotope e†ects. The roles of di†erent substituents are discussed using
electronic structure calculations of the potential energy surfaces at the SM5.4/PM3 level together with
calculations employing variational transition state theory and multidimensional tunneling. p-NO substitution
2
has the most pronounced e†ect on the rate parameters, which can be understood in terms of the relative
destabilisation of the reactant and reduced quantum mechanical tunneling through the barrier. The use of
di†erent models for tunneling suggests that when coupled with the SM5.4/PM3 method, the higher level of
theories tend to overestimate the degree of tunneling.
The reaction studied here is the aqueous base induced
Introduction
rearrangement of phenyl glyoxal hydrate to mandelate. The
process has been shown to involve migration of the hydrated-
formyl hydrogen to the adjacent carbonyl carbon19 and is
best formalised as the transfer of a hydride from electron-rich
to electron-deÐcient carbon (Fig. 1). This hydride shift can
occur either in the mono-anion or the di-anion of the hydrate.
The shift in the di-anion is expected to be signiÐcantly faster
than that in the mono-anion and its contribution is thought
to dominate except at low pH.20 At high pH the mono-anion
form is dominant but the reaction occurs via a small concen-
tration of rapid and reversibly formed di-anion. The experi-
mental barrier must therefore be partitioned between the free
energy for the formation of an equilibrium concentration of
the di-anion and the barrier for the subsequent hydride trans-
fer step.20
Hydrogen transfer is a common process in many organic
systems. Both protic and hydridic motions can be important
steps in many biological redox processes.1 It is also observed
in industrially important reactions involving inter- and intra-
molecular transfer in hydrocarbons2 and combustion pro-
cesses.3 The light mass of the transferring particle can
facilitate quantum mechanical tunneling, which is reÑected in
measured kinetic isotope e†ects. Several recent experimental
studies have implicated tunneling in enzymatic systems,4h6
although it is often difficult to quantify this e†ect, particularly
the role of the motion of the enzyme, which has been sug-
gested to enhance the tunneling.4,7
We here report a combined experimental and theoretical
study of the hydride transfer in phenyl glyoxal hydrate, which
occurs in aqueous solution. The aim has been to study the
e†ect of di†erent substituent groups located at the para-
position of the phenyl ring and to see if current computational
methods can predict the subtle changes in the reactivity which
we observe. In particular we shall explore the suitability of
variational transition state theory (VTST) and multidimen-
sional tunneling methods for the study of this system.
Although the phenyl glyoxal hydrate system does not directly
represent any particular biological scheme, it is an aqueous
phase organic hydride transfer similar in many respects to
those observed in enzymatic reactions.
Various theoretical methods have been used to model tun-
neling, and hydrogen transfer in general. In particular classical
molecular dynamics,8,9 and VTST, both with hybrid quantum
mechanical and molecular mechanical (QM/MM) potential
energy calculations10h12 and with wholly quantum mechani-
cal or semi-empirical potential energy surfaces have been
used.13h17 In the work described here canonical variational
transition state theory (CVT)18 has been used to calculate the
overbarrier contribution to the rate. A multiplicative trans-
mission coefficient is then included which accounts for tunnel-
ing through the barrier and is evaluated by di†erent methods.
We here describe calculations of energetic and rate data
with kinetic isotope e†ects for the four substituted phenyl
glyoxal hydrates (Fig. 1) and compare these to experiment. In
addition to the unsubstituted species (1) we have studied elec-
tron withdrawing (ÈNO , 4), Hammett p \ 0.778, and elec-
2
tron donating (ÈOMe, 3), p \ [0.286, groups at the para (p-)
position, as well as methyl substitution (2). Our results provide
more insight into the mechanism of the phenyl glyoxal
hydride transfer rearrangement and also identify the role of
tunneling in such processes. Although the highest levels of
theory cannot be used to model this reaction in the aqueous
phase, the work presented here does, however, show that cal-
culations using VTST and multidimensional tunneling
methods can be an extremely useful tool for understanding
organic and biological mechanisms.
¤
Electronic supplementary information available. See http://
Fig. 1 General reaction for the rearrangement involving the 1,2-
www.rsc.org/suppdata/cp/b1/b102330f/
hydride shift, the p-group, X \ H (1), Me (2), OMe (3), NO (4).
2
DOI: 10.1039/b102330f
Phys. Chem. Chem. Phys., 2001, 3, 3967È3972
This journal is ( The Owner Societies 2001
3967

View Article Online
rates were measured using 80 : 20 dioxan : water. At this level,
the acid base properties of the solvent are similar to pure
water and ionisation equilibria are una†ected.22
Experimental details
Phenyl glyoxal is a yellow oil that forms a stable crystalline
hydrate, PGOH. A dissociation constant K \ 3.5 ] 10~4
D
has been quoted.21 The hydrate is a weak acid and Letellier et
al. also report a pK \ 10.85 at 298 K. In solution at
a
pH P 12, it exists mainly as its mono-anion, PGO~.
Reactions for a series of p-substituted phenyl glyoxals were
again bu†ered with phosphate at pH 12.57, and the tem-
perature was held constant at 305 K. The rate data are pre-
sented in Table 2.
Computational details
Electronic structure calculations have been carried out on the
four phenyl glyoxals, which di†er only in the substituent
present at the para position. Our strategy is to evaluate rate
constants using semiclassical variational transition state
theory (SC-VTST), which requires generalised Gibbs energies
Conversion of the mono-anion, PGO~, to mandelate,
MA~, was monitored by the changes in UV absorbance of
solutions at 252 nm, the reaction being found to be Ðrst order
in PGO~.
of activation, * G , along the reaction path. Such calcu-
act T
lations are computationally expensive, since Hessians are
needed to evaluate the required vibrational partition func-
tions. These calculations are generally feasible for quite large
systems with semi-empirical Hamiltonians such as AM1 or
PM3, but may be prohibitive for even quite small systems
when high levels of theory are employed. For this reason the
so-called dual level approach is often used, where information
on full pathways computed at a lower level of theory is cor-
rected by more limited information from higher level calcu-
lations such as that relating to stationary points. We have
previously adopted such an approach in a study of hydride
transfer in polycyclic hydroxy ketones in the condensed
phase.17 This latter study involved a dual level calculation on
a single molecule of a size comparable to those to be studied
here and was at the limit of computational feasibility. We
adopt a di†erent strategy for the four molecules studied here
since initial ab initio calculations on reactant structure (1) at
the B3LYP/6-311&&G**23,24 level could not locate a reactant
minimum, due to cleavage of the C7ÈC8 bond (Fig. 2). Such a
problem may not exist if solvation e†ects were to be included.
However, in view of this problem, as well as the number of
molecules to be studied and the success of semi-empirical
methods that include solvation, we have chosen to use the
latter approach for our study.
The dependence of k on base concentration shows that
the reaction requires a second hydroxide ion. We therefore
believe that there is then a second deprotonation and the
hydride shift occurs in the di-anion. There is no experimental
measurement of the second acidity constant, but many dibasic
obs
acids of the form RX(OH) show Ðrst and second acidity con-
2
stants di†ering by 4 to 5 pK units.
a
Rates for the conversion of PGO~ to MA~ were measured
in aqueous phosphate bu†er (pH 12 at 298 K) at four di†erent
temperatures between 289 and 325 K. The results are sum-
marised in Table 1.
Substituents on the phenyl ring were not expected to alter
the acidity of the phenyl glyoxal greatly. Reasonable extrapo-
lations suggest that the p-NO compound should be more
2
acidic than PGOH by 0.3 pK units at one extreme, and that
a
p-OMe should be less acidic by just 0.1 pK . At pH 12 all
a
should be mostly ionised in solution. Not all compounds were
sufficiently soluble in pure water so, for comparability, all
We employ the SM5.4/PM3 solvation model developed by
Truhlar and Cramer,25 which has been shown to successfully
reproduce vapour pressures as a test of the calculated solva-
tion energies.26 This is a continuum model, employing the
PM3 Hamiltonian for the electronic structure calculation and
a Gibbs energy of solvation that involves two terms. A Born
approximation contribution accounts for the electric polarisa-
tion of the continuum-dielectric solvent. Charge Model 1
(CM1)27 class IV partial charges are used for this and are
Table 1 Temperature dependence of rates for rearrangement of
PGH (1)
T /K
k
/s~1
Activation parameters
obs(H)
289.9
304.8
312.2
324.7
1.94 ] 10~4
1.01 ] 10~3
2.09 ] 10~3
6.75 ] 10~3
A \ 4.27(^1.1) ] 1010 s~1;
E \ 79.5(^0.4) kJ mol~1
a
*Ht \ 77.0(^0.4) kJ mol~1
*St \ [49.8(^21.0) J K~1 mol~1
Table 2 Rate data for the rearrangement of substituted phenyl
glyoxal hydrates. Experimental rates and kinetic isotope e†ects.
Errors ^5% for rates and ^10% for the KIEs apply, all measured at
3
05 K in 80 : 20 dioxan : water, pH 12.57
X
k
/s~1
k
/k
obs(H)
obs(H) obs(D)
H (1)a
1.13 ] 10~3
4.62 ] 10~4
3.86 ] 10~4
1.48 ] 10~2
3.52
3.55
2.56
2.27
p-Me (2)
p-OMe (3)
p-NO (4)
2
a The data in Tables 1 and 2 are from separate studies, so that the
rate constants for 1 di†er slightly.
Fig. 2 Atom labelling of 1.
3968
Phys. Chem. Chem. Phys., 2001, 3, 3967È3972

View Article Online
obtained from the Ðtting of charges from NDDO Mulliken
population analyses such that they more accurately represent
experimental dipole moments. The second term, a solvent
accessible surface area (SASA) term, accounts for the Gibbs
energy of cavity formation, dispersion interactions, and
various e†ects arising from the Ðrst solvation shell. Water
solvent was used for the calculations, which di†ers from the
experimental, non-polar, dioxan : water 80 : 20 mixture.
Using parameters for pure dioxan led to difficulty in Ðnding
stable reactant structures. It is possible that experimentally,
the counter cation has a role to play in stabilising the reac-
tant.28 For simplicity this e†ect was not included in the calcu-
lation and the stabilisation was instead the result of the larger
relative permittivity of water.
describe the use of the associated potential energy surfaces in
VTST calculations of rate constants.
Structures and energetics
As far as the reactants (di-anion) are concerned, a number of
di†erent minimum energy conformers were located and prop-
erly characterised for all four systems studied. These di†ered
mainly in the orientation of the hydrated-formyl group. Two
are of particular importance. These are the lowest energy reac-
tant conformer (R1) and the reactant structure (R2) obtained
by an intrinsic reaction co-ordinate calculation starting from
the transition state. Harmonic vibrational frequencies (cm~1)
for the structures R1, R2 and the four transition states are
shown in Table S1. The latter structure (R2) is approximately
VTST calculations of the rate of hydride transfer in the di-
anion were carried out for all four molecules. This was accom-
plished by using AMSOLRATE,29 interfacing SC-VTST in
the POLYRATE30 code and the SM5.4/PM3 solvation model
from AMSOL.31 Optimisation was carried out using the
eigenvector following method in AMSOL, with semi-
analytical methods to evaluate the gradients and derivatives
and with an SCF energy convergence criterion of 10~9 eV.
Stationary structures were characterised by calculation of har-
monic frequencies. The solvated reaction path was calculated
using the PageÈMcIver32 method in mass scaled co-ordinates
with a reduced mass of 1. Calculation of the Gibbs energy of
activation at all the points along the path requires the gener-
alised normal modes. These were calculated by the harmonic
approximation using curvilinear internal coordinates33 as
opposed to rectilinear Cartesians. This was found to give a
more realistic representation of the frequencies along the path.
The maximum in the Gibbs energy was then found and the
rate constant calculated by CVT for that point. The rate was
then combined with a multiplicative transmission coefficient
that accounts for tunneling by either the Wigner (W),34 zero
curvature tunneling (ZCT)18,35 or small curvature tunneling
2
2 kJ mol~1 higher in energy than the most stable reactant
structure. These two di†erent structures may be characterised
by the di†erent values of the O10ÈC7ÈC8ÈH12 dihedral angle
(
Fig. 2). In structures R1 and R2 these values are [170¡ and
[
142¡ respectively, showing that in the reactant structure R2,
H12 is more favourably positioned for attack at C7.
The origin of the di†erence in energies between the struc-
tures R1 and R2 can be seen from Table 3. Whereas the inter-
nal energy favours R2, which is the structure that connects
directly to the transition state, the greater solvation energy of
R1 results in this structure being favoured overall, by D22 kJ
mol~1. This increase in solvation arises from the di†ering
orientation of the hydrated-formyl group in the two struc-
tures. In R2, in contrast to R1, the close proximity of H13 to
O11 restricts the solvation around the negatively charged O11
and this leads to a reduced solvation energy for R2. This
feature is conÐrmed as the origin of the energy di†erence by
consideration of O9, which is fully solvated in both structures.
Thus, the reaction proceeds Ðrst by deprotonation of the
mono-anion, to give structure R1, followed by a conforma-
tional change to give R2, from which hydride transfer occurs.
These steps are shown schematically in Fig. 3, the energy dif-
ference between the mono-anion and di-anion being estimated
(
SCT)14,36 methods. The Wigner method depends only on the
curvature at the top of the barrier. The ZCT method calcu-
lates the e†ect of tunneling through the minimum energy path
and the SCT method allows for tunneling through the adia-
batic curve. The ZCT and SCT models di†er by the use of an
e†ective reduced mass in the SCT case as opposed to a more
simple reduced mass for ZCT. The e†ective reduced mass
includes the e†ects of turning points and reaction path curva-
ture thereby including the e†ects of corner cutting.
using a *pK of 4.5.
a
All calculations employed the re-orientation of the dividing
surface (RODS)37 algorithm which maximises the Gibbs
energy of the generalised transition state at every point. This
was found to eliminate dips which arose when the minimum
energy path was followed by conventional methods, an e†ect
similar to that seen in previous work.38 The resultant
minimum energy pathway (MEP) and adiabatic curves were
much smoother and allowed for more accurate rate constants
and transmission coefficients to be calculated.
Computational results
We Ðrst describe the structural and energetic features of the
stationary structures that we have located. We will then
Fig. 3 Energy proÐle (kJ mol~1) for the reaction involving 1.
Table 3 Contribution to the energies (kJ mol~1) of reactant structures (R1, R2) of 1
Solvation Gibbs energy
Heat of
formation
of solute
Gibbs
energy
Zero point
energy
Gibbs
Energya
O9
H13
O11
Total
R1
R2
[656.9
[653.5
38.6
69.5
[659.4
[570.7
[1264.0
[1184.9
235.6
182.4
[1028.4
[1002.5
321.3
317.6
[707.1
[684.9
a Gibbs energy including zero point energy.
Phys. Chem. Chem. Phys., 2001, 3, 3967È3972
3969

View Article Online
Table 4 Selected bond lengths (A
Ž
) and energies (kJ mol~1) of reactant (R), transition state (TS) and product (P). Selected bond lengths (A
Ž
) for
di†erent species studied are shown in Table S2¤
p-H (1)
p-Me (2)
p-OMe (3)
p-NO (4)
2
R
TS
P
R
TS
P
R
TS
P
R
TS
P
C ÈO
8
1.39
1.39
1.57
1.14
2.14
1.23
1.49
1.10
([1002.5)
0
1.32
1.33
1.52
1.42
1.69
1.26
1.50
1.10
1.27 1.39
1.28 1.39
1.56 1.57
2.16 1.14
1.12 2.14
1.38 1.23
1.52 1.49
1.10 1.49
1.32
1.33
1.52
1.43
1.68
1.26
1.50
1.49
1.27 1.39
1.28 1.39
1.56 1.57
2.16 1.14
1.13 2.15
1.38 1.23
1.52 1.49
1.49 1.40
1.32
1.33
1.52
1.43
1.68
1.26
1.50
1.40
1.27 1.39
1.28 1.39
1.56 1.57
2.16 1.13
1.12 2.15
1.38 1.23
1.52 1.50
1.40 1.48
1.33
1.33
1.52
1.41
1.69
1.26
1.50
1.48
1.27
1.28
1.56
2.19
1.12
1.38
1.52
1.47
9
C ÈO
8
11
C ÈC
8
7
C ÈH
8
12
12
10
C ÈH
7
C ÈO
7
C ÈC
C ÈX
6
7
3
Relative
energya
Relative
energyb
48.1
[160.9
([1040.7)
47.7
[160.7
([1162.7)
48.3
[159.3
([1043.4)
45.8
[169.2
0
0
0
0
0
0
0
36.8
[160.1
36.9
[198.7
35.9
[160.6
34.9
[171.1
Energies are given relative to the di-anion reactant, a excluding and b including zero point energy. c Absolute energies for reactants in
parentheses.
The structure and energetics of the reactants, transition
state and the products for the four systems studied are shown
in Table 4. The expected geometric changes for this type of
hydride transfer are found for all four molecules. During
hydride transfer there is a compression of the C7ÈC8 bond in
the transition state with the transferring hydrogen being
approximately midway between the two carbons (C7,C8).
There is also the associated shortening and lengthening of C8È
O9,ÈO11 and C7ÈO10 respectively as the reaction proceeds.
The e†ect of the di†erent substituents at the para position of
the phenyl ring is rather more subtle but can deÐnitely be
identiÐed. The role of such substituents is to modify the conju-
gation between the phenyl ring and the carbonyl group at
which hydride attack occurs. However, such conjugation is
absent in the product, so that any e†ect of the substituent will
be manifest in the reactant region of the potential energy
surface rather than in the product region. The e†ect of the
varying by only 2.1 kJ mol~1. When zero point energy e†ects
are included the barrier for the p-NO substituent is still the
2
lowest, but now the p-OMe substituent has a slightly lower
barrier than the unsubstituted molecule. The lower activation
barrier for the p-NO species parallels an earlier transition
2
state with a shorter C8ÈH12 bond length (by D0.02 A
Ž
), and a
smaller imaginary frequency (by 70È90 cm~1, Table 5) than in
the three other systems.
These calculations of the stationary structures on the reac-
tion surfaces allow us to rationalise the reaction rates and the
KIEs, which we have reported here. Thus the lower barrier
which we Ðnd for the p-NO species correlates with the obser-
2
vation that the rate for hydride transfer is greater for this
system and the somewhat Ñatter barrier would explain the
smaller measured KIE if tunneling were important.
We now describe the results of the VTST calculations of the
hydride transfer reaction to evaluate the contribution that
such calculations can make to understanding the e†ect of the
substituent on the reaction rates that we have measured.
p-NO group is to reduce the degree of ring-carbonyl conju-
2
gation, so that in the reactant the C3ÈC7 bond is longer than
in the unsubstituted molecule, and the C6ÈN(O ) bond short-
2
ens as the reaction proceeds and the ring-carbonyl (C3ÈC7)
Calculation of reaction rates
conjugation disappears. Thus, considering only the ring-
carbonyl conjugation, the role of the p-NO group is to desta-
bilise the reactant, and render it somewhat more transition
The experimentally determined rate constant is for the conver-
sion of the phenyl glyoxal mono-anion to the mandelate ion.
This proceeds via deprotonation to form the di-anion, which,
from our computational results undergoes a conformational
change, followed by hydride transfer (Fig. 3). The large di†er-
ence in the magnitude of the experimental and theoretical
rates (Table 5) can be understood when one considers that the
2
state-like, leading to a lower barrier compared to the
unsubstituted species (1). The role of the p-OMe group is the
opposite, stabilising the reactant, and leading to a slightly
greater barrier than for (1). These e†ects are however, small,
with the barriers (in the absence of zero point energy e†ects)
Table 5 Kinetic parameters for 1, 2, 3 and 4 at 305 K
TST
CVT
CVT/Wa
CVT/ZCT
CVT/SCT
Experimentalb
Rate constants/s~1
p-H (1)
3.54 ] 106
2.61 ] 106
1.40 ] 106
3.10 ] 106
15.7 ] 106
5.39 ] 106
9.46 ] 106
6.44 ] 106
9.93 ] 106
54.0 ] 106
13.3 ] 106
7.40 ] 106
11.4 ] 106
75.9 ] 106
1.13 ] 10~3
4.62 ] 10~4
3.86 ] 10~4
1.48 ] 10~2
(
(
(
1071i)
p-Me (2)
2.47 ] 106
3.24 ] 106
18.7 ] 106
2.96 ] 106
1099i)
p-OMe (3)
6.43 ] 106
1077i)
p-NO (4)
30.4 ] 106
2
(
1005i)
KIE
p-H (1)
p-Me (2)
p-OMe (3)
2.62
2.64
2.63
2.10
2.46
2.30
2.74
2.08
2.92
2.74
3.27
2.40
3.71
4.03
3.96
3.72
4.21
4.20
4.07
4.15
3.52
3.55
2.56
2.27
p-NO (4)
2
a The imaginary frequencies (cm~1) are given in parentheses. b The large di†erence between the magnitudes of the calculated and experimental
rate constants arises from the reactive structure (R2) being 43.5 kJ mol~1 above the mono-anion (Fig. 3).
3970
Phys. Chem. Chem. Phys., 2001, 3, 3967È3972

View Article Online
although the barriers for the p-H, p-Me and p-OMe species
are quite close, as indeed are the corresponding experimental
rates. In line with their similar rates, the p-H and p-Me species
have very similar KIEs, which are substantially larger than
that for the p-NO species. In the absence of tunneling this
2
qualitative trend is reproduced by the calculations but the
spread of the KIEs is smaller than experiment, suggesting tun-
neling is more important for the p-H and p-Me species than
for the p-NO species. We note that the measured KIE of the
2
p-OMe species is deÐnitely smaller than those of the other
three species which is not shown by the calculations that do
not include tunneling. However, the rate data from Table 2 do
give a good Hammett plot (r2 \ 0.995) but there is no obvious
relationship between the rate and the KIE. When tunneling is
included, all three treatments (W, ZCT, SCT) predict
increased KIEs, the e†ect being largest for the SCT model, but
no model predicts the actual spread of the experimental KIEs.
In general the KIEs predicted using the Wigner correction are
too small and those for the ZCT and SCT methods are too
large.
The origin of the discrepancy between theory and experi-
ment in the case of the KIEs is unclear at present. It may
result from the low level of electronic structure theory used, or
may reÑect the continuum treatment of the solvent, and shows
the need for further studies on well deÐned systems.
Fig. 4 Eyring plot for the hydride transfer in 1.
VTST calculations were conducted on simply the Ðnal step of
Fig. 3.
The results of the VTST calculations for hydride transfer in
reactant R2 are summarised in Table 5. The rate constant is
smaller for the CVT than for the TST model, due to the maxi-
misation of the Gibbs energy that is implicit in the VTST pro-
cedure. The Eyring plot (Fig. 4) shows little deviation from
linearity over the temperature range (273È373 K). The
increase in the rate due to tunneling is clearly evident, where
the increased rate with SCT compared to that with ZCT
shows the importance of corner cutting. At all levels the rate
Acknowledgements
We thank BBSRC and EPSRC for the support of this
research, and Professor D. G. Truhlar for use of the POLY-
RATE codes.
of hydride transfer for the p-NO species is greater, by a factor
2
of about 5, compared to that of the three other species, in
qualitative agreement with experiment. For the other three
species the rates are quite close and the relative order is
dependent upon the computational method employed.
The prediction of the KIEs presents a greater challenge.
Experimentally the values vary by only 1.3 for all four species,
References
1
2
3
C. Walsh, Enzymatic Reaction Mechanisms, Freeman, San Fran-
cisco, 1979; C. I. F. Watt, Comprehensive Biological Catalysis,
Academic Press, New York, 1998.
G. M. Kramer, in Industrial and L aboratory Alkylation, ed. L. F.
Albright and A. R. Gold, American Chemical Society, Washing-
ton, DC, 1977, ch. 1.
and are in the range 2È4. The value for the p-NO species is
2
the smallest, a result predicted using both TST and CVT, and
with tunneling described by the Wigner approximation. These
methods also predict values for the p-NO species essentially
2
in agreement with experiment. All three methods predict that
W.-P. Hu, I. Rossi, J. C. Corchado and D. G. Truhlar, J. Phys.
Chem. A, 1997, 101, 6911.
the KIEs for the other three species (p-H, p-Me, p-OMe) are
quite close and larger than that for the p-NO species. This is
in accord with experiment except for the value for the p-OMe
4
5
6
7
A. Kohen and J. P. Klinman, Acc. Chem. Res., 1998, 31, 397.
A. Kohen and J. P. Klinman, Chem. Biol., 1999, 6, R191.
Y.-K. Cho and D. B. Northrop, Biochemistry, 1999, 38, 7470.
A. Kohen, R. Cannio, S. Bartolucci and J. P. Klinman, Nature,
2
species which, experimentally, is signiÐcantly lower than that
for the p-H and p-Me species. We Ðnd that all KIEs are
increased at the ZCT and SCT levels, compared to the values
using the Wigner approximation, reÑecting the importance of
multidimensional tunneling (ZCT) and corner cutting (SCT).
The importance of corner cutting is shown by the increase in
the KIEs (by 4È16%) at the SCT compared to the ZCT level.
However, at both these levels, the values for all four species
are larger than the experimental ones, and the relative values
are not as good as that found at the simpler level of theory
1
999, 399, 496; N. S. Scrutton, J. Basran and M. J. Sutcli†e, Eur.
J. Biochem., 1999, 264, 666.
C. L. Brooks, D. L. Miller, M. Karplus and B. M. Pettitt, Adv.
Chem. Phys., 1988, 71, 1.
R. V. Stanton, D. L. Miller and P. A. Kollman, in Modern
Methods for Multidimensional Dynamics Computations in Chem-
istry, ed. D. L. Thompson, World ScientiÐc, Singapore, 1998,
p. 355.
8
9
1
0
J. A. Barnes and I. H. Williams, Chem. Commun., 1996, 2, 193.
11 C. Alhambra, J. C. Corchado, M. L. Sanchez, J. Gao and D. G.

(
TST, CVT, CVT/W), in particular the value for the p-NO
Truhlar, J. Am. Chem. Soc., 2000, 122, 8197.
R. M. Nicoll, S. A. Hindle, G. MacKenzie, I. H. Hillier and N. A.
Burton, T heor. Chem. Acc., 2001, 106, 105.
Y.-P. Liu, G. C. Lynch, T. N. Truong, D.-H. Lu, D. G. Truhlar
and B. C. Garrett, J. Am. Chem. Soc., 1993, 115, 2408.
Y.-P. Liu, D.-H. Lu, A. Gonzalez-Lafont, D. G. Truhlar and B. C.
Garrett, J. Am. Chem. Soc., 1993, 115, 7806.
Y. Kim and H. J. Hwang, J. Am. Chem. Soc., 1999, 121, 4669.
D. K. Maity, R. L. Bell and T. N. Truong, J. Am. Chem. Soc.,
2
12
13
14
species is not always the smallest.
Discussion
We have attempted the prediction of quite subtle variations in
the reactivity of substituted glyoxals towards hydride transfer,
brought about by substituent e†ects well removed from the
site of transfer itself. Such a substituent will stabilise or desta-
bilise both the mono- and di-anion depending on its p value.
The relative e†ect of the substituent on the mono- and di-
1
1
5
6
2
000, 122, 897.
1
7
R. M. Nicoll, I. H. Hillier and D. G. Truhlar, J. Am. Chem. Soc.,
2001, 123, 1459.
18 B. C. Garrett and D. G. Truhlar, J. Chem. Phys., 1979, 70, 1593;
D. G. Truhlar and B. C. Garrett, Acc. Chem. Res., 1980, 13, 440;
D. G. Truhlar and B. C. Garrett, Annu. Rev. Phys. Chem., 1984,
anion, needed to estimate the pK of the mono-anion, is diffi-
a
cult to quantify accurately, preventing a calculation of the
3
5, 159; D. G. Truhlar, B. C. Garrett and S. J. Klippenstein, J.
overall reaction rate. Nevertheless, the calculated energy bar-
riers for the hydride transfer process itself do broadly reÑect
Phys. Chem., 1996, 100, 12771.
19 W. v. E. Doering, T. I. Taylor and E. F. Schoenwaldt, J. Am.
Chem. Soc., 1948, 70, 455.
the experimental rates, p-NO [ p-H [ p-Me [ p-OMe,
2
Phys. Chem. Chem. Phys., 2001, 3, 3967È3972
3971

View Article Online
2
0
1
I. H. Hillier, S. Smith, S. C. Mason, S. N. Whittleton, C. I. F.
Watt and J. Willis, J. Chem. Soc., Perkin T rans 2, 1988, 1345.
S. Letellier, J. C. Dufresne and M. B. Fleury, Electrochim. Acta,
30 J. C. Corchado, Y.-Y. Chuang, P. L. Fast, J. Villa
Y.-P. Liu, G. C. Lynch, K. A. Nguyen, C. F. Jackels, V. S. Melis-
sas, B. J. Lynch, I. Rossi, E. L. Coitino, A. Fernandez-Ramos, R.
`
, W.-P. Hu,
2
8
1
980, 28, 1043.
Steckler, B. C. Garrett, A. D. Isaacson and D. G. Truhlar,
POL Y RAT E-version 8.5, University of Minnesota, Minneapolis,
2000.
2
2
2
3
H. P. Marshall and E. Grunwald, J. Chem. Phys., 1953, 21, 2143.
P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch,
J. Phys. Chem., 1994, 98, 11623; W. Kohn, A. D. Becke and R. G.
Parr, J. Phys. Chem., 1996, 100, 12974.
31 AMSOL -version 6.6 by G. D. Hawkins, D. J. Giesen, G. C.
Lynch, C. C. Chambers, I. Rossi, J. W. Storer, J. Li, P. Winget, D.
Rinaldi, D. A. Liotard, C. J. Cramer and D. G. Truhlar, Uni-
versity of Minnesota, Minneapolis, 1997, based in part on
AMPAC-version 2.1 by D. A. Liotard, E. F. Healy, J. M. Ruiz
and M. J. S. Dewar and on the EF routines by Frank Jensen.
32 M. Page and J. W. McIver, J. Chem. Phys., 1987, 88, 922.
33 K. A. Nguyen, C. F. Jackels and D. G. Truhlar, J. Chem. Phys.,
1996, 104, 6491.
34 E. P. Z. Wigner, Z. Phys. Chem. B, 1932, 19, 203; R. P. Bell, T he
T unnel E†ect In Chemistry, Chapman & Hall, London, 1980.
35 D. G. Truhlar and A. Kuppermann, J. Am. Chem. Soc., 1971, 93,
1840; B. C. Garrett and D. G. Truhlar, J. Phys. Chem., 1979, 83,
2921.
2
2
2
2
4
5
6
7
W. J. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab
Initio Molecular Orbital T heory, Wiley, New York, 1986.
C. C. Chambers, C. J. Cramer and D. G. Truhlar, J. Phys Chem.,
1
996, 100, 16385.
P. Winget, G. D. Hawkins, C. J. Cramer and D. G. Truhlar, J.
Phys. Chem. B, 2000, 104, 4726.
M. H. Abraham, P. L. Grellier, D. V. Prior, P. P. Duce, J. J.
Morris and P. J. Taylor, J. Chem. Soc., Perkin T rans. 2, 1989,
6
99.
2
8
9
T. Okuyama, K. Kimura and T. Fueno, Bull. Chem. Soc. Jpn.,
1
AMSOL RAT E-version 8.6/P8.5.1-A6.6 by Y.-Y. Chuang, Y.-P.
Liu and D. G. Truhlar, University of Minnesota, Minneapolis,
982, 55, 1493.
2
36 T. C. Allison and D. G. Truhlar, in Modern Methods for Multidi-
mensional Dynamics Computations in Chemistry, ed. D. L. Thomp-
son, World ScientiÐc, Singapore, 1998, p. 618.
1
999 based on POL Y RAT E-version 8.5.1 by J. C. Corchado,
Y.-Y. Chuang, P. L. Fast, J. Villa, W.-P. Hu, Y.-P. Liu, G. C.
`
Lynch, K. A. Nguyen, F. Jackels, V. S. Melissas, B. J. Lynch, I.
Rossi, E. L. Coitino, A. Fernandez-Ramos, R. Steckler, B. C.
Garrett, A. D. Isaacson and D. G. Truhlar, University of Minne-
sota, Minneapolis, 2000 and AMSOL -version 6.6 by G. D.
Hawkins, D. J. Giesen, G. C. Lynch, C. C. Chambers, I. Rossi,
J. W. Storer, J. Li, T. Zhu, D. Rinaldi, D. A. Liotard, C. J.
Cramer and D. G. Truhlar, University of Minnesota, Minnea-
polis, 2000.
37 J. Villa
38 A. Gonzalez-Lafont, J. Villa
and D. G. Truhlar, J. Phys. Chem. A, 1998, 102, 3420.
`
and D. G. Truhlar, T heor. Chem. Acc., 1997, 97, 317.
`
, J. M. Lluch, J. Bertran, R. Steckler
3972
Phys. Chem. Chem. Phys., 2001, 3, 3967È3972