Investigating the α-effect in gas-phase SN2 reactions of microsolvated anions

Article abstract of DOI:10.1021/ja4066943

The α-effect - enhanced reactivity of nucleophiles with a lone-pair adjacent to the attacking center - was recently demonstrated for gas-phase S_N2 reactions of HOO^-, supporting an intrinsic component of the α-effect. In the present work we explore the gas-phase reactivity of microsolvated nucleophiles in order to investigate in detail how the α-effect is influenced by solvent. We compare the gas-phase reactivity of the microsolvated α-nucleophile HOO^-(H₂O) to that of microsolvated normal alkoxy nucleophiles, RO^-(H₂O), in reaction with CH₃Cl using a flowing afterglow-selected ion flow tube instrument. The results reveal enhanced reactivity of HOO^-(H ₂O) and clearly demonstrate the presence of an α-effect for the microsolvated α-nucleophile. The association of the nucleophile with a single water molecule results in a larger Bronsted β_nuc value than is the case for the unsolvated nucleophiles. Accordingly, the reactions of the microsolvated nucleophiles proceed through later transition states in which bond formation has progressed further. Calculations show a significant difference in solvent interaction for HOO^- relative to the normal nucleophiles at the transition states, indicating that differential solvation may well contribute to the α-effect. The reactions of the microsolvated anions with CH₃Cl can lead to formation of either the bare Cl^- anion or the Cl^-(H₂O) cluster. The product distributions show preferential formation of the Cl^- anion even though the formation of Cl^-(H₂O) would be favored thermodynamically. Although the structure of the HOO^-(H₂O) cluster resembles HO^-(HOOH), we demonstrate that HOO^- is the active nucleophile when the cluster reacts.

Full text of DOI:10.1021/ja4066943

Article
pubs.acs.org/JACS
Investigating the α‑Effect in Gas-Phase SN2 Reactions of
Microsolvated Anions
Ditte L. Thomsen,†,‡ Jennifer N. Reece,‡ Charles M. Nichols,‡ Steen Hammerum,†
and Veronica M. Bierbaum*,‡
†Department of Chemistry, University of Copenhagen, Universitetsparken 5, Copenhagen, DK-2100 Denmark
‡Department of Chemistry and Biochemistry, University of Colorado, 215 UCB, Boulder, Colorado 80309, United States
S
* Supporting Information
ABSTRACT: The α-effectenhanced reactivity of nucleo-
philes with a lone-pair adjacent to the attacking centerwas
recently demonstrated for gas-phase SN2 reactions of HOO−,
supporting an intrinsic component of the α-effect. In the
present work we explore the gas-phase reactivity of micro-
solvated nucleophiles in order to investigate in detail how the
α-effect is influenced by solvent. We compare the gas-phase
reactivity of the microsolvated α-nucleophile HOO−(H2O) to
that of microsolvated normal alkoxy nucleophiles, RO−(H2O),
in reaction with CH3Cl using a flowing afterglow-selected ion
flow tube instrument. The results reveal enhanced reactivity of HOO−(H2O) and clearly demonstrate the presence of an α-effect
for the microsolvated α-nucleophile. The association of the nucleophile with a single water molecule results in a larger Brønsted
βnuc value than is the case for the unsolvated nucleophiles. Accordingly, the reactions of the microsolvated nucleophiles proceed
through later transition states in which bond formation has progressed further. Calculations show a significant difference in
solvent interaction for HOO− relative to the normal nucleophiles at the transition states, indicating that differential solvation may
well contribute to the α-effect. The reactions of the microsolvated anions with CH3Cl can lead to formation of either the bare Cl−
anion or the Cl−(H2O) cluster. The product distributions show preferential formation of the Cl− anion even though the
formation of Cl−(H2O) would be favored thermodynamically. Although the structure of the HOO−(H2O) cluster resembles
HO−(HOOH), we demonstrate that HOO− is the active nucleophile when the cluster reacts.
hydrazine and hydroxylamine. In solution, α-nucleophiles have
been shown to display enhanced nucleophilicity relative to that of
normal nucleophiles of similar basicity, and the term α-effect16
has been used to describe this deviation in reactivity relative to
that expected from Brønsted-type correlations. The α-effect has
been observed in several different types of reactions including
SN2 reactions.17−23 The magnitude of the α-effect for reaction
with specific substrates can be determined as the ratio of rate
constants for the reactions of an α-nucleophile (kα) and a normal
nucleophile (knormal) of similar basicity.24 α-effects (kα/knormal) in
the range of 5−1000 have been reported for numerous reactions
in such diverse solvents as H2O, DMSO, and CH3CN.25 This has
led to an active controversy about whether the α-effect is
controlled by inherent properties of the α-nucleophile or by
external solvent effects.
Gas-phase studies provide a vital link to resolving the intrinsic
nature of the α-effect. The experimental search for a gas-phase α-
effect began with studies of DePuy et al.26 who found that the
hydrogen peroxide anion (HOO−) does not show enhanced
reactivity relative to hydroxide (HO−) in reactions with methyl
formate. The authors concluded that a gas-phase α-effect could
INTRODUCTION
■
Gas-phase ion-molecule studies provide a means to investigate
the intrinsic factors that affect reactions in a solvent free
environment or in an environment with a controlled number of
solvent molecules, also referred to as microsolvation. Drastic
differences are observed between gas-phase and condensed-
phase reactivity, as rate constants of gas-phase ion-molecule
reactions exceed those in solution by as much as 16 orders of
magnitude.1 Gas-phase studies of microsolvated ions are
essential to provide insight into the role of the solvent in a
given reaction and may shed light on the transition between gas-
phase and condensed-phase reactivity.
The SN2 reaction is a fundamental process in organic
chemistry, and numerous experimental1−11 and computa-
tional9−15 studies have examined the reactivity of microsolvated
anions in SN2-type reactions. Nucleophilic substitution of methyl
chloride is a prototypical SN2 reaction, with no competing
reaction pathways (no proton abstraction or adduct formation),
and investigation of the relative nucleophilicity of different
anions is straightforward, making this substrate ideal for the
investigation of an α-effect for microsolvated anions.
α-Nucleophiles possess a lone pair of electrons adjacent to the
attacking center, e.g. ionic nucleophiles such as the hydrogen
peroxide and hypochlorite anions, or neutral nucleophiles such as
Received: July 1, 2013
Published: September 18, 2013
© 2013 American Chemical Society
15508
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514

Journal of the American Chemical Society
Article
afterglow ion source and are mass selected using a quadropole mass filter
prior to injection into the reaction flow tube. HO− was prepared by
electron ionization (70 eV) of methane and nitrous oxide (2:1 ratio) in
helium buffer gas. HOO− was prepared by proton abstraction from
not be observed. This was subsequently questioned, since the
gas-phase basicity of HO− far exceeds that of HOO−, and similar
behavior would imply enhanced reactivity of the α-nucleophile.27
Also, increased single-electron-transfer in the HOO− transition
state led Patterson and Fountain27 to support the existence of an
α-effect. However, an experimental study of the gas-phase
reactivity of the α-nucleophiles HOO−, ClO−, and BrO− with a
comprehensive series of oxygen-centered normal nucleophiles in
SN2 reactions with methyl chloride revealed no gas-phase α-
effect.28 It was recognized that the high exothermicity of these
reactions might obscure the properties of the intrinsic barriers.
Subsequent studies of similar SN2 reactions with reduced
reaction exothermicity definitively showed reactivity enhance-
ments by factors of 2.3−50 when comparing HOO− to the
normal nucleophiles HO−, methoxide (CH3O−), ethoxide
(C2H5O−), and 2-propoxide (i-C3H7O−) in reactions with
methyl fluoride, anisole, and fluoroanisole.29 These results
demonstrate the existence of an intrinsic component of the α-
effect. Recent studies clearly also show the presence of an α-effect
in addition-elimination reactions on carbonyl carbon.30 In
addition, McAnoy et al.31 report a gas-phase α-effect when
comparing the reactions of HOO− and CH3O− with dimethyl
methylphosphonate, based on major differences in the product
branching ratios for the two ions.
Recent theoretical studies support the premise that the α-effect
has a component that can be attributed to intrinsic properties of
the nucleophile.27,32−37 Ren and Yamataka34−37 investigated the
reactivity of the α-nucleophiles, HOO−, ClO−, and BrO−, with a
series of alkyl chlorides by calculating the reaction barrier heights
using the composite G2(+) method and comparing the results to
those for normal nucleophiles. While the normal nucleophiles
display a linear correlation between anion basicity and reaction
barrier height, reactions of the α-nucleophiles appear to have
lower barriers than predicted by this linear basicity-barrier
relation, supporting an intrinsic component of the α-effect.
The α-nucleophile investigated in this study is the micro-
solvated hydrogen peroxide anion, HOO−(H2O). The structure
of this cluster ion has been mentioned briefly by Anick38 in a
comprehensive computational study on the incremental
solvation of the HOO− anion. However, to the best of our
knowledge this is the first experimental study of the gas-phase
reactivity of the α-nucleophile HOO−(H2O). We report overall
rate constants and reaction efficiencies along with product
distributions for the reactions of HOO−, HO−, CH3O−, C2H5O−,
and i-C3H7O− with methyl chloride in the presence of a single
water molecule. We investigate the α-nucleophilicity of
HOO−(H2O) by comparing the reaction efficiency to that of
the microsolvated normal nucleophiles. The absence of an α-
effect for HOO− in previous gas-phase studies of nonsolvated
anions with methyl chloride was ascribed to the large reaction
efficiency. Incremental microsolvation has previously been
shown to significantly decrease the reaction efficiency,1−6 and
we exploit the lower reactivity of the microsolvated anions to
investigate the presence of an α-effect in SN2 reactions with
methyl chloride.
−
H2O2 by NH2 , which was prepared by electron ionization of NH3. All
other anions were generated by proton abstraction of the corresponding
neutrals by HO−. The microsolvated ions were prepared by introducing
a mixture of H2O in tetrahydrofuran slightly downstream of the
production of bare ions. In previous studies, tetrahydrofuran was shown
to enhance the formation of microsolvated ions.41 Upon injection of the
mass-selected ions into the reaction flow tube the ions are entrained in a
flow of helium (∼0.5 Torr), which assures a thermal energy distribution
of the reactant ions prior to reaction with the neutral reagent. A known
flow of neutral reagent is added through a series of equidistant inlets
along the reaction flow tube, and finally ionic reagents and products are
analyzed in the detection region using a quadrupole mass filter coupled
to an electron multiplier.
The decrease in reactant ion signal is monitored as a function of
reaction distance (which is directly correlated to the reaction time), and
a rate constant is derived by assuming pseudo-first-order kinetics
(reactant ion ∼105 ions cm−3, neutral reagent ∼1011 molecules cm−3).
All reported reaction rate constants and product distributions represent
the average of at least three individual measurements. Reaction
efficiencies are the ratio between the experimental rate constant (kexp
)
and the collision rate constant (kcol). Collision rate constants are
calculated from parametrized trajectory collision rate theory.42 Product
distributions are determined by extrapolating the product yields to zero
reaction time in order to exclude any secondary reaction products. The
product distributions were corrected for reactions of the bare ion, which
was always present as a result of collision-induced dissociation upon
injection of the microsolvated ion into the reaction flow tube. Mass
discrimination was minimized, and no corrections were applied to the
data. Absolute uncertainties with respect to the rate constants are 30%,
and with respect to the product distributions they are 50% due to
uncertainties in pressure, flow, and temperature. All reagents were
obtained from commercial sources and are used without further
purification: methyl chloride (CH3Cl, Matheson, 99.5%); hydrogen
peroxide (H2O2, Sigma-Aldrich, 50 wt % solution in water); methanol
(CH3OH, Honeywell B&J, 99.9%); ethanol (C2H5OH, Decon
Laboratories, 200 proof); isopropanol (i-C3H7OH, Fisher Scientific,
99.9%), tetrahydrofuran (THF, Sigma Aldrich, 99.9%), ammonia (NH3,
Airgas, 99.9995%). Helium buffer gas (He, Airgas, 99.995%) was
purified by passage through a molecular sieve trap immersed in liquid
nitrogen.
COMPUTATIONAL METHODS
■
Minimum and transition-state structures were localized as stationary
points on the MP2/6-311++G(d,p) potential energy surface. Harmonic
vibrational frequencies were obtained at the same level and used to verify
the nature of the stationary points as minima and first-order saddle
points, respectively. The structure of the HOO−(H2O) cluster was
further optimized using B3LYP/6-311++G(d,p), MP2/aug-cc-pVTZ,
and CCSD(T)/ aug-cc-pVTZ. Thermodynamic properties were
calculated using a modified version of the composite G3 method in
which all geometry optimizations and harmonic vibrational frequency
calculations were carried out with the 6-31+G(d) basis set instead of the
default 6-31G(d). The diffuse functions were included to allow a better
description of the anions. This treatment is comparable to the G2(+)
method which has successfully been employed in previous studies of SN2
reactions involving anions.36,43 All MP2, B3LYP and composite
calculations were carried out with the Gaussian 09 program.44 CCSD(T)
calculations were carried out with Molpro 2010.1.45
EXPERIMENTAL SECTION
■
The experiments were carried out with a flowing afterglow selected-ion
flow tube (FA-SIFT) mass spectrometer at the University of Colorado,
Boulder. Details of the instrument and methods have been presented
elsewhere.39,40 Briefly, the instrument consists of the following four
sections: an ion-source flow tube, an ion-selection region, a reaction flow
tube, and a detection region. The ions are produced in a flowing
Proton affinities (PA) of the microsolvated nucleophiles correspond
to the calculated reaction enthalpies at 298 K for the reaction:
Nu−(H2O) + H+ → NuH(H2O)
(1)
15509
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514

Journal of the American Chemical Society
Article
Table 1. Thermodynamic data, kinetic data, and product distributions for the SN2 reactions of microsolvated anions with methyl
chloride: Nu−(H2O) + CH3Cl
a
b
c
thermodynamic data
kinetic data
kexp (× 10−11
product distribution
Cl−
PA
ΔHhydr
ΔHr
)
Eff
Cl−(H2O)
d
e
e
HO−(H2O)
1538
1520
1511
1505
1493
−112
−99
−92
−89
−106
−101/−161
−103/−164
−95/−155
−82/−143
−80/−140
36
10
0.15
0.85
0.78
0.47
0.42
0.72
0.15
CH3O−(H2O)
C2H5O−(H2O)
i-C3H7O−(H2O)
HOO−(H2O)
1
0.049
0.016
0.0028
0.044
0.22
0.53
0.58
0.28
3.3 0.1
0.54 0.05
9.3 0.3
a
Thermodynamic data in units of kJ/mol calculated using G3; proton affinities (PA) of the microsolvated anions; hydration enthalpy (ΔHhydr);
b
reaction enthalpies (ΔHr) displayed for Cl− and clustered Cl−(H2O) products respectively (Cl−/Cl−(H2O)). Overall rate constant (kexp) in units of
cm3 molecule−1 s−1. Error bars represent one standard deviation, absolute uncertainty is 30%. Reaction efficiency (Eff) is the ratio of the overall
experimental rate constant (kexp) to the collision rate constant (kcol) calculated using parametrized trajectory collision theory.42 Product distributions
c
d
are determined by extrapolating the product yields to zero reaction time; absolute uncertainty is estimated as 50%. Rate constant for DO−(D2O)
e
+ CH3Cl obtained from ref 46; standard deviation not reported; deuteration was used to distinguish the reactant and product ion masses. Product
distribution is obtained from ref 3.
RESULTS AND DISCUSSION
Reaction Efficiencies. The measured rate constants and
solvated hydrogen peroxide anion definitively shows an α-effect
in the SN2 reaction with methyl chloride.
When comparing the reaction efficiencies of the microsolvated
nucleophiles with the reaction efficiencies previously reported for
the corresponding nonsolvated nucleophiles28 (Table 2), it is
■
reaction efficiencies for the SN2 reactions of microsolvated
anions with methyl chloride are displayed in Table 1, together
with calculated reaction thermodynamics and proton affinities of
the microsolvated nucleophiles. Since the reaction with methyl
chloride proceeds exclusively by an SN2 process, the overall
efficiencies can be compared directly; there is no loss of reactant
from competing reaction pathways to take into account. Figure 1
Table 2. Thermodynamic Data, Reaction Efficiencies and
Calculated Bond Lengths for the Transition States of the SN2
Reaction of Unsolvated Anions with Methyl Chloride: Nu− +
CH3Cl
a
a
b
c
c
PA
ΔHr
Eff
RNu−C
RC−Cl
HO−
1635 (1633)
1601 (1598)
1584 (1585)
1574 (1576)
1574 (1575)
−214
−202
−187
−171
−186
0.67
0.60
0.60
0.54
0.52
2.158
2.127
2.109
2.084
2.115
2.090
2.069
2.088
2.109
2.076
CH3O−
C2H5O−
i-C3H7O−
HOO−
a
Thermodynamic data in units of kJ/mol calculated using G3;
calculated proton affinities (PA) of the anions with experimental
values in parentheses;47 calculated reaction enthalpies (ΔHr) displayed
b
for Cl− product formation. Reaction efficiency (Eff) is the ratio of the
overall experimental rate constant (kexp) to the collision rate constant
(kcol) calculated using parametrized trajectory collision theory;42
c
reaction efficiencies are obtained from ref 28. Geometries are
optimized at the MP2/6-311++G(d,p) level of theory; transition state-
bond lengths (RNu−C and RC−Cl) in units of Å.
Figure 1. Reaction efficiency (kexp/kcol) vs anion proton affinity (PA) for
the SN2 reactions of unsolvated and microsolvated anions with methyl
chloride. Normal nucleophiles are displayed in blue, and α-nucleophiles,
in red. The linear trend lines are fits to the normal nucleophile data sets.
The reaction efficiency is the ratio of the overall experimental rate
constant (kexp) to the collision rate constant (kcol), calculated using
parametrized trajectory collision theory.42 Reaction efficiencies for the
unsolvated anions are obtained from ref 28. The anion PAs are
calculated using G3. The vertical error bars represent the absolute
uncertainty of 30%. The horizontal error bars represent the
experimental uncertainty in the PA of the bare ion, against which the
computed PAs are benchmarked.
evident that the association of just a single water molecule has a
drastic effect on the reactivity. The reaction efficiencies of the
nonsolvated nucleophiles are all large and quite similar (between
0.67 and 0.52) despite a proton affinity range of more than 55 kJ/
mol. In contrast, the reaction efficiencies of the microsolvated
nucleophiles vary by a factor of 50 (from 0.15 for HO−(H2O) to
0.003 for i-C3H7O−(H2O)), despite a smaller proton affinity
span.
Incremental microsolvation has previously been shown to
significantly decrease the reaction efficiency,1−6 and indeed,
associating the nucleophile with even a single water molecule
lowers the reaction efficiency and results in large reactivity
differences across the series of nucleophiles; thus, the presence of
an α-effect is revealed, which was not apparent for reactions of
the unsolvated nucleophiles.
shows a Brønsted-type plot of the reaction efficiency versus the
proton affinity of the microsolvated anions. Efficiencies for
reactions of unsolvated nucleophiles are included for compar-
ison. The result is clear: the reaction efficiency of HOO−(H2O)
reacting with methyl chloride is more than an order of magnitude
larger than would be expected from the rate-energy relationship
found for the normal microsolvated nucleophiles. The micro-
Product Distributions. While gas-phase SN2 reactions of
unsolvated nucleophiles with methyl chloride produce only the
bare Cl− ion, reactions of microsolvated nucleophiles have an
15510
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514

Journal of the American Chemical Society
Article
additional product channel in which the solvated Cl−(H2O) ion
is formed:
Nu−(H2O) + CH3Cl → NuCH3 + Cl− + H2O
→ NuCH3 + Cl−(H2O)
(2a)
(2b)
The product distributions for formation of Cl− and Cl−(H2O)
for the series of microsolvated nucleophiles are shown in Table 1.
The microsolvated product ion is formed in the more exothermic
process and this channel could well be expected to be dominant.
Nonetheless, unsolvated Cl− is the main product of the reaction
of the microsolvated HOO−, HO−, and CH3O− ions with methyl
chloride, and reactions of the microsolvated C2H5O− and i-
C3H7O− ions form approximately equal amounts of the solvated
and nonsolvated Cl−.
Previous investigations of SN2 reactions involving micro-
solvated nucleophiles and methyl chloride have shown similar
product distributions in which formation of the nonsolvated
product dominates.3,4 We find that the relative yield of solvated
product, Cl−(H2O), tends to increase with increasing size of the
nucleophile and with decreasing reaction exothermicity. Several
factors may contribute to the preferred formation of the
thermodynamically less favorable nonsolvated product; explan-
ations suggesting steric properties,8 product dynamics in a
postreaction complex,10 as well as redistribution of excess
reaction energy48 have been proposed. In the present study we
obtain a similar product distribution for the microsolvated α-
nucleophile, HOO−(H2O), as that found for the microsolvated
normal nucleophiles, which leads us to conclude that the
enhanced reactivity of HOO−(H2O) is not reflected in
differences in the product distribution.
Figure 3. Lowest-energy structure of the microsolvated anion
HOO−(H2O) optimized at the CCSD(T)/aug-cc-pVTZ level of theory.
completely planar, whereas isolated hydrogen peroxide possesses
a staggered conformation. This structure is even more
remarkable considering that the proton affinity of HO− (1633
kJ/mol)47 is higher than that of HOO− (1575 kJ/mol).47
This peculiar structure of the microsolvated HOO− has been
addressed briefly by Anick38 in a comprehensive computational
study on the incremental solvation of the HOO− anion. In
agreement with his results we find that two minima can be found
when the structure is optimized with B3LYP/6-311++G(d,p),
one corresponding to a HOO−(H2O) structure and one
corresponding to a HO−(HOOH) structure. The two isomers
are separated by only a small barrier of 7 kJ/mol. When MP2/6-
311++G(d,p), MP2/aug-cc-pVTZ or CCSD(T)/aug-cc-pVTZ
is employed only the HO−(HOOH) structure represents a
stationary point.
It is pertinent to consider whether the reactive nucleophile
should be described as HOO− solvated by a water molecule or as
HO− solvated by hydrogen peroxide. In the experiment the
reactant ion is selected solely on the basis of mass, which does not
distinguish between the two structures. Upon injection of the
cluster into the reaction flow tube collision-induced dissociation
results only in the formation of the bare HOO−; no HO− is
observed (Figure 4). Accordingly, dissociation of the adduct into
Water Cluster Structures. Representative calculated
structures of the microsolvated anions are shown in Figure 2.
Figure 2. Lowest-energy structures of the microsolvated anions
HO−(H2O), CH3O−(H2O), and CH3CH2O−(H2O) optimized at the
MP2/6-311++G(d,p) level of theory.
All normal-nucleophiles form clusters in which the water
molecule has one hydrogen oriented toward the anionic center.
The computed hydration enthalpy (ΔHhydr) is obtained as the
enthalpy difference between the water cluster and the isolated
species (Table 1). The magnitude of ΔHhydr varies in the series of
normal nucleophiles from −112 kJ/mol for HO−(H2O) to −89
kJ/mol for i-C3H7O−(H2O) tracking the decreasing proton
affinity for the isolated anions.
Figure 4. The mass selected spectrum of the microsolvated hydrogen
peroxide anion HOO−(H2O) (m/z 51). The bare HOO− ion (m/z 33)
was always present as a result of collision-induced dissociation upon
injection of the microsolvated ion into the reaction flow tube.
Compared to the normal nucleophiles, the HOO−(H2O)
cluster has a high ΔHhydr of −106 kJ/mol, even though the
proton affinity of the HOO− anion is similar to that of the i-
C3H7O− anion. The optimized structure of microsolvated
HOO− is displayed in Figure 3 along with geometrical
parameters. The HOO−(H2O) cluster is distinctly different
from the water cluster of the normal anions in that it appears to
resemble HO−(HOOH), in which the HOOH behaves as a
bidentate hydrogen donor toward HO−. The structure is almost
symmetrical, with nearly identical hydrogen bond lengths and
hydrogen bond angles. The most conspicuous geometrical
feature is the structure of the HOOH moiety, which is
HOO− and H2O is 60 kJ/mol more favorable than dissociation
into HO− and H2O2 based on experimental heats of formation
for the separated species.47 The calculated CCSD(T)/aug-cc-
pVTZ energy difference between the two dissociation channels is
similar (58 kJ/mol). However, this does not prove the identity of
the reactive nucleophile in the cluster.
The experimental product distributions provide us with direct
evidence to support the contention that HOO− is the reactive
15511
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514

Journal of the American Chemical Society
Article
nucleophile. It is not possible to directly demonstrate whether
methyl hydrogen peroxide or methanol is formed since only the
ionic products are monitored. Fortunately, the clustered ionic
products differ depending on whether HOO− or HO− is the
active nucleophile:
HOO−(H2O) + CH3Cl → HOOCH3 + Cl−(H2O)
(3)
HO−(HOOH) + CH3Cl → HOCH3 + Cl−(HOOH)
(4)
The reaction of microsolvated HOO− with methyl chloride
produces 28% Cl−(H2O); formation of a Cl−(HOOH) product
is not observed (Figure 5). This result indicates that the attacking
Figure 6. IRC surfaces (solid lines) for reaction of the microsolvated
hydrogen peroxide anion with methyl chloride calculated using MP2/6-
311++G(d,p). The black surface corresponds to HOO− attack, the
purple surface corresponds to HO− attack. The dotted line only serves
to illustrate a connection to the reactants. The structures represent
MP2/6-311++G(d,p) optimized transition states (TS), prereaction
complexes (RC) and reactants. In parentheses are the transition state
enthalpies calculated using G3.
Figure 5. The product mass spectrum for the reaction between
HOO−(H2O) and CH3Cl. Only two ionic products are detected: Cl− at
m/z 35/37 and Cl−(H2O) at m/z 53/55. Residual reactant ion,
HOO−(H2O) is seen at m/z 51.
nucleophile is in fact HOO−. Nevertheless, as noted by a
reviewer, we cannot rigorously exclude the possibility that some
of the unsolvated Cl− is formed through a reaction in which HO−
is the reactive nucleophile. However, when the HOO−(H2O)
cluster reacts with methyl formate the products formed are solely
the peroxyformate anion and clusters thereof, conclusively
demonstrating that HOO− is the only reacting nucleophile in the
cluster ion.49 Furthermore, reaction paths for reactions of
HOO−(H2O) and HO−(HOOH) with methyl chloride were
examined computationally (Figure 6). The enthalpy of the
transition state corresponding to HOO−(H2O) attack is 17 kJ/
mol below that of the transition state for HO−(HOOH) attack.
This is consistent with the experimental results, supporting that
HOO− is the reactive nucleophile in the HOO−(H2O) cluster.
This in turn supports that the enhanced reactivity of
HOO−(H2O) can indeed be ascribed to the α-nucleophilic
character of the HOO− anion.
An intrinsic reaction coordinate (IRC) calculation was
performed to investigate how the transition state corresponding
to HOO− attack connects to the HO−(HOOH) structure in
Figure 3. As shown in Figure 6 (black line) the HOO−(H2O)
transition state connects to a prereaction complex in which the
structure of the attacking cluster is closely related to that found in
the transition state. However, in the prereaction complex we see
that the hydrogen bonding OH of the water molecule is
elongated, suggesting incipient proton transfer. Upon further
increase of the distance between HOO−(H2O) and CH3Cl the
proton is completely transferred from the water to the HOO−
and the attacking cluster converts into the HO−(HOOH)
cluster.
Figure 7. Transition-state structures for the SN2 reaction of the
microsolvated anions HOO−(H2O) and HO−(H2O) with methyl
chloride optimized at the MP2/6-311++G(d,p) level of theory. The
bond lengths RNu−C, RC−Cl, and RNu−O, and the vibrational frequencies of
the hydrogen-bonded OH stretch, νOHb, are found in Table 3 for all
anions.
bipyramidal transition state. Throughout the reaction the water
molecule retains one hydrogen pointing toward the nucleophilic
center. The distances between the nucleophile and the carbon
center (RNu−C) and between the leaving group Cl− and the
carbon center (RC−Cl) are displayed in Table 3. The transition-
state bond lengths vary only little between the different
Table 3. Calculated Bond Lengths, Vibrational OHb
Stretching Frequencies and OHb Red Shifts for the Transition
State of the SN2 Reaction of Microsolvated Anions with
Methyl Chloride: Nu−(H2O) + CH3Cl
a
a
a
b
b
RNu−C
RC−Cl
RNu−O
νOHb
ΔνOHb
HO−(H2O)
2.062
2.044
2.030
2.026
2.023
2.157
2.133
2.151
2.157
2.141
2.602
2.593
2.604
2.617
2.637
2961
2964
2999
3043
3127
991
986
947
904
820
CH3O−(H2O)
C2H5O−(H2O)
i-C3H7O−(H2O)
HOO−(H2O)
a
Geometries are optimized and vibrational frequencies calculated at
the MP2/6-311++G(d,p) level of theory; bond lengths are in units of
Å, and frequencies and red shifts are in units of cm−1; RNu−C, RC−Cl
,
b
RNu−O, and νOHb are indicated in Figure 7. The red shifts of the
bonded OH stretching frequency (ΔνOHb) are calculated relative to
the frequency of the free OH stretch (νOHf) in the attached water
molecule (ΔνOHb = νOHf − νOHb). Deuterated alkoxy groups were used
in the frequency calculation to avoid coupling between the OHb- and
CH-stretches.
Transition-State Structures. Representative calculated
transition-state structures are displayed in Figure 7. In the SN2
reactions examined the nucleophile attacks the methyl chloride
opposite to the leaving chloride group, forming a trigonal
15512
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514

Journal of the American Chemical Society
Article
nucleophiles, although the HOO−(H2O) cluster has a slightly
tighter transition state in which bond formation has progressed
most (shortest RNu−C). The corresponding geometrical param-
eters for the transition states of unsolvated nucleophiles are
displayed in Table 2. It appears that the presence of a single water
molecule results in a later transition state with more advanced
Nu−C bond formation. A direct relationship between the
magnitude of the α-effect and the extent of bond formation in the
transition state has previously been established for condensed-
phase reactions of other systems.21 The Brønsted βnuc parameter,
which is the slope of a Brønsted plot, is also known to be directly
related to the extent of bond formation in the transition state.
This in turn implies that the α-effect is directly related to the
magnitude of the βnuc value.21 Furthermore, it has been argued
that the βnuc value reflects the amount of single-electron-transfer
(SET) in the transition state,50−52 and that enhanced SET
character would be directly related to a larger α-effect.27,53 The
slopes of the Brønsted plot shown in Figure 1 provide qualitative
βnuc values for reactions of the unsolvated and microsolvated
nucleophiles. Although the data for the microsolvated
nucleophiles exhibit some deviation from a straight line, a
qualitative comparison demonstrates that the βnuc value for the
microsolvated nucleophiles is significantly larger than the βnuc
value for the unsolvated nucleophiles. The drastic change of the
βnuc value upon microsolvation is in complete agreement with the
experimental finding that there is no apparent α-effect for the
unsolvated nucleophiles, whereas the effect strongly influences
the reactions when even a single water molecule is present.
To investigate the interaction with the solvent water molecule
in the transition state, the distance between the nucleophilic
center and the oxygen of the water molecule (RNu−O) as well as
the frequency (νOHb) and red shift (ΔνOHb) of the hydrogen-
bonded OH stretch in the water molecule are displayed in Table
3. It is well established that the IR red shift can be used to assess
the strength of the hydrogen bond, as a larger red shift
corresponds to a stronger hydrogen bond.54 For the normal
nucleophiles the red shifts are between 1000 and 900 cm−1,
decreasing from HO−(H2O) to i-C3H7O−(H2O). The red shift
calculated for the HOO−(H2O) transition state (ΔvOHb = 820
cm−1) is lower, corresponding to a weaker hydrogen bond to the
water molecule in the transition state. This is in accordance with
the transition state of HOO−(H2O) having a slightly longer
distance to the water molecule than do the normal nucleophiles.
Although the HOO−(H2O) cluster has the second-highest
hydration enthalpy (Table 1), surpassed only by that of the
HO−(H2O) cluster, it exhibits the weakest hydrogen bond to the
water molecule in the transition state. This implies that the
solvent interactions for the α-nucleophile differ from those of the
normal nucleophiles, and indicates that, in addition to inherent
properties of the α-nucleophile, differential solvation may well
contribute to the α-effect.
CONCLUSIONS
■
Gas-phase SN2 reactions of microsolvated oxygen-centered
anions with methyl chloride demonstrate that an α-effect
influences the reactions of microsolvated HOO− anions. To
associate the anions with just a single water molecule lowers the
reaction efficiency and allows observation of an α-effect, which
was not apparent in the studies of the reactions of unsolvated
anions. Reactions of the microsolvated anions can lead to
formation of either the bare Cl− anion or the solvated product,
Cl−(H2O). The experimental product distributions show
preferential formation of the Cl− anion despite the thermody-
namic preference for formation of Cl−(H2O). Reaction of the
microsolvated α-nucleophile, HOO−(H2O), results in a product
distribution (Cl−: Cl−(H2O) of 72%: 28%) similar to that of the
microsolvated normal nucleophiles, showing that the enhanced
reactivity of HOO−(H2O) is not reflected in differences in the
product distribution.
The structure of the HOO−(H2O) cluster was investigated
computationally. This structure can adequately be described as
HO−(HOOH), in which the HOOH behaves as a bidentate
hydrogen donor bonded to a HO− moiety. Despite this peculiar
structure, the experimental product distribution supports that
HOO− is the reacting nucleophile within the cluster, and in turn,
that the enhanced reactivity of HOO−(H2O) is ascribed to the α-
nucleophilic character of the HOO− anion.
In comparison to the unsolvated anions, the association of the
nucleophile with a single water molecule results in later transition
states in which the bond formation is more progressed. This is
directly related to larger Brønsted βnuc values obtained from the
slopes of the Brønsted plots. Accordingly, we find that the
microsolvated anions display an α-effect with methyl chloride
even though the unsolvated anions do not. Hydrogen-bonding
interactions with the water molecule in the transition state
suggest that the α-nucleophile has a much weaker interaction
with the solvating water molecule in the transition state than do
microsolvated normal nucleophiles.
Microsolvation with even a single water molecule provides
insight into the role of solvent on the α-effect; further studies of
the behavior of the α-effect with incremental solvation will
contribute materially to our understanding of the transition from
gas-phase to condensed-phase reactivity.
ASSOCIATED CONTENT
■
S
* Supporting Information
Energies and optimized geometries of all compounds. This
material is available free of charge via the Internet at http://pubs.
acs.org.
AUTHOR INFORMATION
■
Corresponding Author
veronica.bierbaum@colorado.edu
Formation of the clustered product Cl−(H2O) could also
proceed through a transition state in which the water molecule is
transferred to the leaving group, Cl−, prior to nucleophilic
displacement. However, computational studies on the SN2
reaction of HO−(H2O) with CH3Cl have shown that this entails
a higher reaction barrier,12 as would be expected from the better
gas-phase leaving-group ability of the Cl− anion than of the
oxygen-centered anions. Direct dynamic simulations would be
valuable in providing a better understanding of the water
molecule transfer during the reaction.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We dedicate this paper to the memory of Prof. Charles H.
DePuy; we are grateful for his inspiration and friendship. We
thank Nicholas Demarais for helpful discussions and we
gratefully acknowledge NSF (CHE-1012321 and CHE-
1300886) for their financial support. D.L.T. thanks the Danish
Chemical Society and the Augustinus Foundation, C.M.N.
15513
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514

Journal of the American Chemical Society
Article
(40) Bierbaum, V. M. In Encyclopedia of Mass Spectrometry; Gross, M.
L., Caprioli, R., Eds.; Elsevier: Amsterdam, 2003; Vol. 1, pp 98−109.
(41) Bickelhaupt, F. M.; de Koning, L. J.; Nibbering, N. M. M.
Tetrahedron 1993, 49, 2077−2092.
thanks the AFOSR, and J.N.R. thanks the Colorado Diversity
Initiative for financial support.
REFERENCES
(42) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183−5185.
(43) Glukhovtsev, M. N.; Pross, A.; Radom, L. J. Am. Chem. Soc. 1995,
117, 2024−2032.
■
(1) Bohme, D. K.; Mackay, G. I. J. Am. Chem. Soc. 1981, 103, 978−979.
(2) Bohme, D. K.; Raksit, A. B. J. Am. Chem. Soc. 1984, 106, 3447−
3452.
(44) Frisch, M. J., et al. Gaussian 09, Revision B.01; Gaussian, Inc.:
Wallingford CT, 2010.
(45) Werner, H.-J. et al. MOLPRO, version 2010.1. A package of ab
initio programs; see www.molpro.net.
(46) DePuy, C. H.; Gronert, S.; Mullin, A.; Bierbaum, V. M. J. Am.
Chem. Soc. 1990, 112, 8650−8655.
(47) Bartmess, J. E. Negative Ion Energetics Data in NIST Chemistry
WebBook, NIST Standard Reference Database, Number 69; Lindstrom, P.
J., Mallard, W. G., Eds.; National Institute of Standards and Technology:
Gaithersburg, MD, 20899, http://webbook.nist.gov (retrieved Nov 16,
2012).
(3) Hierl, P. M.; Ahrens, A. F.; Henchman, M.; Viggiano, A. A.;
Paulson, J. F.; Clary, D. C. J. Am. Chem. Soc. 1986, 108, 3142−3143.
(4) Hierl, P. M.; Paulson, J. F.; Henchman, M. J. J. Phys. Chem. 1995,
99, 15655−15661.
(5) Seeley, J. V.; Morris, R. A.; Viggiano, A. A. J. Phys. Chem. A 1997,
101, 4598−4601.
(6) Viggiano, A. A.; Arnold, S. T.; Morris, R. A.; Ahrens, A. F.; Hierl, P.
M. J. Phys. Chem. 1996, 100, 14397−14402.
(7) Takashima, K.; Riveros, J. M. Mass Spectrom. Rev. 1998, 17, 409−
430.
(48) Baer, T.; Hase, W. L. In Unimolecular Reaction Dynamics: Theory
and Experiment; Oxford University Press: New York: 1996, pp 1−448.
(49) Thomsen, D. L.; Nichols, C. M.; Reece, J. N.; Hammerum, S.;
Bierbaum, V. M. J. Am. Soc. Mass. Spectrom. 2013, Submitted.
(50) Pross, A.; Shaik, S. S. J. Am. Chem. Soc. 1981, 103, 3702−3709.
(51) Bordwell, F. G.; Clemens, A. H. J. Org. Chem. 1981, 46, 1035−
1037.
(8) Otto, R.; Brox, J.; Trippel, S.; Stei, M.; Best, T.; Wester, R. Nat.
Chem. 2012, 4, 534−538.
(9) Otto, R.; Xie, J.; Brox, J.; Trippel, S.; Stei, M.; Best, T.; Siebert, M.
R.; Hase, W. L.; Wester, R. Faraday Discuss. 2012, 157, 41−57.
(10) Otto, R.; Brox, J.; Trippel, S.; Stei, M.; Best, T.; Wester, R. J. Phys.
Chem. A 2013, 117, 8139−8144.
(11) Chen, X.; Regan, C. K.; Craig, S. L.; Krenske, E. H.; Houk, K. N.;
Jorgensen, W. L.; Brauman, J. I. J. Am. Chem. Soc. 2009, 131, 16162−
16170.
(12) Ohta, K.; Morokuma, K. J. Phys. Chem. 1985, 89, 5845−5849.
(13) Hu, W. P.; Truhlar, D. G. J. Am. Chem. Soc. 1994, 116, 7797−
7800.
(52) Terrier, F.; Mokhtari, M.; Goumont, W.; Halle, J. C.; Buncel, E.
Org. Biomol. Chem. 2003, 1, 1757−1763.
(53) Hoz, S. J. Org. Chem. 1982, 47, 3545−3547.
(54) Jeffry, G. A. An Introduction to Hydrogen Bonding; Oxford
University Press: New York, 1997.
(14) Tucker, S. C.; Truhlar, D. G. J. Am. Chem. Soc. 1990, 112, 3347−
3361.
(15) Vayner, G.; Houk, K. N.; Jorgensen, W. L.; Brauman, J. I. J. Am.
Chem. Soc. 2004, 126, 9054−9058.
(16) Edwards, J. O.; Pearson, R. G. J. Am. Chem. Soc. 1962, 84, 16−24.
(17) Fountain, K. R.; Patel, K. D. J. Org. Chem. 1997, 62, 4795−4797.
(18) Fountain, K. R.; Felkerson, C. J.; Driskell, J. D.; Lamp, B. D. J. Org.
Chem. 2003, 68, 1810−1814.
(19) McIsaac, J. E.; Mulhause, H. A.; Behrman, E. J.; Subbaram, J.;
Subbaram, L. R. J. Org. Chem. 1972, 37, 1037−1041.
(20) Dixon, J. E.; Bruice, T. C. J. Am. Chem. Soc. 1971, 93, 6592−6597.
(21) Dixon, J. E.; Bruice, T. C. J. Am. Chem. Soc. 1972, 94, 2052−2056.
(22) Gregory, M. J.; Bruice, T. C. J. Am. Chem. Soc. 1967, 89, 4400−
4402.
(23) Buncel, E.; Wilson, H.; Chuaqui, C. Bull. Soc. Chim. Belg. 1982, 91,
420−420.
(24) Hoz, S.; Buncel, E. Isr. J. Chem. 1985, 26, 313−319.
(25) Buncel, E.; Um, I. H. Tetrahedron 2004, 60, 7801−7825.
(26) DePuy, C. H.; Della, E. W.; Filley, J.; Grabowski, J. J.; Bierbaum, V.
M. J. Am. Chem. Soc. 1983, 105, 2481−2482.
(27) Patterson, E. V.; Fountain, K. R. J. Org. Chem. 2006, 71, 8121−
8125.
(28) Villano, S. M.; Eyet, N.; Lineberger, W. C.; Bierbaum, V. M. J. Am.
Chem. Soc. 2009, 131, 8227−8233.
(29) Garver, J. M.; Gronert, S.; Bierbaum, V. M. J. Am. Chem. Soc. 2011,
133, 13894−13897.
(30) Garver, J. M.; Yang, Z.; Wehres, N.; Nichols, C. M.; Worker, B. B.;
Bierbaum, V. M. Int. J. Mass Spectrom. 2012, 330, 182−190.
(31) McAnoy, A. M.; Paine, M. R. L.; Blanksby, S. J. Org. Biomol. Chem.
2008, 6, 2316−2326.
(32) Fountain, K. R. J. Phys. Org. Chem. 2005, 18, 481−485.
(33) Afzal, D.; Fountain, K. R. Can. J. Chem. 2011, 89, 1343−1354.
(34) Ren, Y.; Yamataka, H. Org. Lett. 2006, 8, 119−121.
(35) Ren, Y.; Yamataka, H. J. Org. Chem. 2007, 72, 5660−5667.
(36) Ren, Y.; Yamataka, H. Chem.Eur. J. 2007, 13, 677−682.
(37) Ren, Y.; Yamataka, H. J. Comput. Chem. 2009, 30, 358−365.
(38) Anick, D. J. J. Phys. Chem. A 2011, 115, 6327−6338.
(39) Van Doren, J. M.; Barlow, S. E.; DePuy, C. H.; Bierbaum, V. M.
Int. J. Mass Spectrom. Ion Processes 1987, 81, 85−100.
15514
dx.doi.org/10.1021/ja4066943 | J. Am. Chem. Soc. 2013, 135, 15508−15514