An electric field-induced change in the selectivity of a metal oxide-catalyzed epoxide rearrangement

Article abstract of DOI:10.1021/ja210365j

The rearrangement of cis-stilbene oxide catalyzed by Al₂O ₃ was studied in the presence of interfacial electric fields. Thin films of Al₂O₃ deposited on Si electrodes were used as the opposing walls of a reaction vessel. Application of a voltage across the electrodes engendered electrochemical double layer formation at the Al ₂O₃-solution interface. The aldehyde to ketone product ratio of the rearrangement was increased by up to a factor of 63 as the magnitude of the double layer charge density was increased. The results support a field-dipole effect on the selectivity of the catalytic reaction.

Full text of DOI:10.1021/ja210365j

Communication
pubs.acs.org/JACS
An Electric Field−Induced Change in the Selectivity of a Metal
Oxide−Catalyzed Epoxide Rearrangement
Craig F. Gorin, Eugene S. Beh, and Matthew W. Kanan*
Stanford University, Department of Chemistry, 337 Campus Drive, Stanford, California 94305, United States
*
S Supporting Information
nonelectrochemical reactions including gas phase oxidations of
simple hydrocarbons with O , oxidation of H with O , and
ABSTRACT: The rearrangement of cis-stilbene oxide
catalyzed by Al O was studied in the presence of
2
2
2
14−18
2
3
alkene isomerizations.
This electrochemical promotion of
interfacial electric fields. Thin films of Al O deposited
2
3
catalysis (EPOC) results from voltage-induced ion migration
onto the conductive catalyst surface that changes the work
on Si electrodes were used as the opposing walls of a
reaction vessel. Application of a voltage across the
electrodes engendered electrochemical double layer
1
5,19,20
function or available active sites.
In contrast, we
hypothesized that the local electric field of a catalytic reaction
could be controlled without effecting electrochemical changes
to the catalyst by localizing it to a suitably designed electrode−
electrolyte interface. Here we demonstrate that externally
controlled double layer charge density changes the selectivity of
epoxide rearrangement reactions catalyzed by an insulating
Lewis acidic metal oxide. Our results are consistent with a
field−dipole origin of the selectivity changes and provide a
general experimental method to evaluate electric field effects on
catalytic reactions.
formation at the Al O −solution interface. The aldehyde
2
3
to ketone product ratio of the rearrangement was
increased by up to a factor of 63 as the magnitude of
the double layer charge density was increased. The results
support a field−dipole effect on the selectivity of the
catalytic reaction.
ontrolling selectivity is arguably the principal challenge
facing the development of efficient chemical syntheses.
1
,2
C
To study electric field effects on catalytically active metal
oxides, we designed a reaction vessel (the “parallel plate cell”)
that enables the electrostatic environment of a solution-exposed
metal oxide to be controlled with a voltage source (Figures 1
and S3). The parallel plate cell employs two heavily p-doped Si
electrodes. Each electrode is coated on one side with a thin (45
Å) layer of Al O that is deposited by using atomic layer
For an irreversible reaction, selectivity is determined by the
relative magnitudes of competing activation barriers on the
reaction’s potential energy surface. Because this surface is
determined by the chemical structures of the reaction
components, efforts to control selectivity have necessitated
making changes to one or more of these components.
Identifying structural features that induce selectivity can be
extremely challenging because of the complexities of molecular
structure−activity relationships. In principle, however, an
externally applied electric field could also be used to control
selectivity through field−dipole interactions. This concept is
particularly appealing because all unique molecules and
transition states have unique interactions with an electric field
determined by their structure-specific charge distributions.
Previous experimental studies have demonstrated that static
electric fields affect the rates of electron transfer reactions in
2
3
21
deposition (ALD) (see Supporting Information). The Al O
2
3
layer of one of the electrodes is used as the catalytically active
oxide in the parallel plate cell experiments (the “catalyst
electrode”); the Al O layer of the other electrode is coated
2
3
with an alkylphosphonic acid monolayer to block reactivity at
this surface (the “counter electrode”). In an assembled parallel
plate cell, the two electrodes are separated by a 500 μm thick
perfluorinated gasket with a rectangular opening in the center;
an electrolyte solution containing the substrate for a reaction
occupies the volume between the two electrodes determined by
the dimensions of the gasket. Application of a voltage between
the two electrodes generates electrochemical double layers at
each interface and places the exposed Al O layer on the
3
−5
enzymatic and synthetic systems,
and a recent study has
provided spectroscopic evidence for an interfacial electric field
effect on the tautomerization equilibrium of an electrode-
2
3
6
attached synthetic molecule. Large electric fields (1−10 V/
catalyst electrode in an interfacial electric field.
nm) within zeolite cavities have been exploited for visible-light
The strength of the interfacial electric field at the surface of
the exposed Al O layer depends on the extent of double layer
7
,8
photooxidations of simple hydrocarbons with O2. Naturally
occurring electric fields in enzyme active sites have also been
2
3
charging, i.e. the charge density on the Si electrode that is
balanced by electrolyte ions at the oxide−electrolyte interface.
To measure this charge density as a function of the applied
9
,10
implicated as major contributors to enzymatic catalysis.
Furthermore, density functional theory studies of reactions in
uniform electric fields have concluded that selectivity is highly
22
voltage (V), we performed double step chronocoulometry
with fully assembled cells. The cells were comprised of 45 Å of
Al O on the catalyst electrode and 45 Å of Al O coated with a
11−13
sensitive to strong fields.
To date, however, a general
method to apply large electric fields to reactions and study their
2
3
2
3
effects on selectivity has not been developed.
Researchers have previously demonstrated that electro-
chemical changes of conductive catalysts can alter the rates of
Received: November 4, 2011
Published: December 16, 2011
©
2011 American Chemical Society
186
dx.doi.org/10.1021/ja210365j | J. Am. Chem.Soc. 2012, 134, 186−189

Journal of the American Chemical Society
Communication
Figure 1. Parallel plate cell. Cross-sectional illustration of the cell (top) and expanded illustration of the metal oxide−electrolyte interfaces (bottom)
at 0 V (left), positive voltage (middle), and negative voltage (right). Significant V-dependent electric fields are present only at the interfaces. The top
Al O layer is blocked with an alkylphosphonic acid monolayer to confine catalysis to the bottom Al O layer.
2
3
2
3
2
3
25
monolayer of octylphosphonic acid on the counter electrode.
In a CH CN or CH Cl solution containing 0.5 mM
of 1 V/nm are present at |V| > 3 V in the parallel plate cell.
Indirect measurements of field strengths at other electrode−
3
2
2
5
,6,26
tetrabutylammonium hexafluorophosphate (TBAPF ), the
electrolyte interfaces have yielded similar values.
Field
6
charge vs time curves exhibit a rapid (< 50 ms) rise that is
characteristic of double layer charging. A similar discharge is
observed upon stepping back to 0 V (Figure S4). The amount
of double layer charging increases approximately linearly with V
strengths of this magnitude are sufficient to significantly affect
the selectivity of a reaction. For example, if two competing
transition states arise from a common intermediate, the field
effect on selectivity is determined solely by the difference
between its effects on the transition states. If the charge
distributions of the transition states in the presence of the
(
Figure 2). For |V| > 2 V, double layer charging in excess of 1
2
μC/cm is observed. These values are comparable to double
layer charge densities measured for metal electrodes in contact
electric field E are approximated as dipole moments μ
the energetic differentiation induced by E is given by ΔU =
− μ . According to this model, a 1 V/nm
1 2
and μ ,
24
with dilute electrolytes in CH CN. The difference in charge
3
densities measured in CH Cl and CH CN likely reflects the
E•Δμ, where Δμ = μ
1
2
2
2
3
difference in polarity of these solvents. The chronocoulometric
measurements are also consistent with the double layer
capacitance values obtained from electrochemical impedance
spectroscopy measurements (Table S1).
The charge densities measured by chronocoulometry
indicate that voltage-dependent field strengths on the order
electric field would induce an ∼10-fold change in selectivity at
300 K if the projection of Δμ along the field axis were 2.8 D
(E•Δμ = 1.4 kcal/mol).
Lewis acidic metal oxides including Al O are known to
2 3
catalyze numerous organic reactions including epoxide
2
7,28
rearrangements.
To study electric field effects on Al O
2 3
catalysis, we selected the rearrangement of cis-stilbene oxide (1)
to diphenylacetaldehyde (2) and diphenylethanone (3) as a
model reaction (Figure 3a). Parallel plate cell reactions were
performed in cells with 45 Å of Al O on the catalyst electrode
2
3
and 45 Å of Al O coated with an octylphosphonic acid layer on
2
3
the counter electrode. The electrolyte solution consisted of 5
mM 1 and 0.5 mM TBAPF in either CH CN or CH Cl . After
6
3
2
2
assembly of the cell, either the electrodes were connected by a
short circuit (0 V) or a constant voltage was applied across the
two electrodes by using a sourcemeter. Parallel plate cell
reactions were allowed to proceed for 16 h. At the conclusion
of each experiment, the cell was disassembled and the
conversion and product ratio were determined by using high-
pressure liquid chromatography (HPLC). For V ≤ 5 V, a
Figure 2. (a) Schematic expanded view of the Al O −electrolyte
2
3
interface at positive V. Charge density on the underlying Si (σ) is
2
balanced by electrolyte ions. (b) Chronocoulometric measurement of
leakage current of 2−10 nA/cm is observed over the course of
σ at selected voltages in the parallel plate cell with 0.5 mM TBAPF in
6
the experiment (Figures S6 and S7), corresponding to a total
CH CN (black squares) and CH Cl (blue circles). Data are shown
3
2
2
moles of charge passed that is ∼0.5−2.5% of the total moles of
for a cell with 45 Å of Al O on each electrode and an octylphosphonic
2
3
1
in the cell. Voltages > 5 V lead to a dielectric breakdown and
acid monolayer on the counter electrode.
increasing currents over time.
1
87
dx.doi.org/10.1021/ja210365j | J. Am. Chem.Soc. 2012, 134, 186−189

Journal of the American Chemical Society
Communication
quantities of additional unidentified products are observed by
HPLC.
The electric field strength at the surface of the exposed Al O
2
3
layer in the parallel plate cell can also be adjusted by changing
the double layer capacitance of the cell. Accordingly, an electric
field effect on selectivity should depend on capacitance in
addition to V. To test this hypothesis, we prepared a cell in
which the counter electrode was coated with an octadecyl-
phosphonic acid layer instead of an octylphosphonic acid layer.
The extended hydrocarbon chain was expected to lower the
capacitance of the counter electrode interface. In 0.5 mM
TBAPF6 in CH CN, the double layer charge densities
3
determined by chronocoulometry for this cell at +4.5 and
2
2
−
4.5 V are 1.56 and 1.52 μC/cm , respectively, ∼0.5 μC/cm
lower than the corresponding values for a cell with an
octylphosphonic acid coated counter electrode. At 0 V, the
2
:3 ratio is the same for experiments in cells with the two
different counter electrodes; however, at ±4.5 V, the 2:3 ratios
for a cell with an octadecylphosphonic acid−coated counter
electrode are unchanged (Figure 3b, red triangles), in sharp
contrast to the ∼10-fold increases observed for cells with an
octylphosphonic acid layer on the counter electrode. These
results indicate that selectivity changes in the parallel plate cell
depend not on V by itself, but on double layer charge densities,
as required for a field−dipole effect. In this case, the double
layer charge density at the exposed Al O layer and its effect on
2
3
selectivity were lowered by changing the blocking layer on the
counter electrode 500 μm away.
Larger V-dependent selectivity changes are observed for
parallel plate cell reactions performed in 0.5 mM TBAPF in
6
CH Cl with octylphosphonic acid on the counter electrode
2
2
than in CH CN (Figure 3c). At 0 V, the reaction proceeds in
3
4
% conversion to form a 1:3.7 ratio of 2:3. For |V| > 3 V, the
conversions are > 10-fold higher and the 2:3 ratios increase
exponentially with voltage to reach 16.9:1 at +5 V and 11.3:1 at
−
4.5 V, representing ∼63- and ∼42-fold enhancements,
respectively. The larger selectivity changes in CH Cl may
2
2
Figure 3. (a) Reaction scheme for the rearrangement of cis-stilbene
oxide in the parallel plate cell. Product ratios as a function of voltage
are shown for experiments in CH CN (b) and CH Cl (c). Black
squares correspond to experiments with octylphosphonic acid on the
counter electrode. Red triangles correspond to experiments with
octadecylphosphonic acid on the counter electrode.
22
reflect a more compact electrochemical double layer, which
engenders larger fields at the Al O surface. This effect could
2
3
3
2
2
outweigh the somewhat smaller double layer charge density in
CH Cl compared to CH CN, as measured by chronocoulom-
etry (see Figure 2b).
2
2
3
In control experiments, very low (< 1%) conversions of 1 are
observed for reactions in parallel plate cells with octylphos-
phonic acid monolayers on both Al O layers at 0 V or ±4.5 V,
The product ratio for the rearrangement of 1 catalyzed by an
Al O layer in the parallel plate cell exhibits strong dependence
2
3
2
3
on V (Figure 3b). In all cases, the reaction proceeds cleanly
such that the products 2 and 3 and remaining starting material
account for nearly all of the material detected by HPLC. At 0
indicating that the reaction takes place at the Al O surface, as
2 3
expected. To determine whether voltage permanently alters the
Al O layers, catalyst electrodes that had previously been used
1
2
3
V in CH CN, 5% of 1 is converted to 2 and 3 in a 1:2.2 ratio,
similar to the ratio that is observed for the reaction catalyzed by
at +4.5 V in a parallel plate cell reaction were subsequently
reused in a parallel plate cell at 0 V with a fresh solution of 1.
The conversion and selectivity with the reused electrode at 0 V
matched that of a fresh catalyst electrode at 0 V, indicating that
an applied voltage does not lead to an irreversible change in the
Al O layer that alters its catalytic activity. Additional control
experiments ruled out the contribution of electrolytically
generated H+ to the observed selectivity changes (see
Supporting Information).
3
Al O3 powder in the same solution (see Supporting
2
Information for V-dependent conversion data and for experi-
ments with Al O powders). For |V| ≤ 3 V, negligible changes
2
3
in the selectivity are observed. However, for V > 3 V, the 2:3
ratio increases exponentially as V is increased and the
conversions are 2−4-fold higher than at 0 V. At +5 V, the
2
3
2:3 ratio is 10.2:1, representing an ∼22-fold enhancement over
the ratio at 0 V. Strikingly, a similar trend is observed at
negative voltages. For V < −3 V, the 2:3 ratio increases
exponentially as the voltage is decreased and the conversions
are 2−5-fold higher than at 0 V; at −4.5 V the ratio is 4.5:1,
representing an ∼10-fold enhancement over the 0 V ratio. At
voltages where dielectric breakdown is observed, large
Together, the chronocoulometry, voltage-dependent selec-
tivity data and associated control experiments are consistent
with an electric field effect on the rearrangement of cis-stilbene
oxide catalyzed by Al O . The double layer charge density
increases approximately linearly with V, generating V-depend-
ent interfacial electric fields. While the strength of the field at
2
3
1
88
dx.doi.org/10.1021/ja210365j | J. Am. Chem.Soc. 2012, 134, 186−189

Journal of the American Chemical Society
Communication
the molecular level cannot be determined by the measurements
described here, field strengths on the order of 1 V/nm are
reasonable given the measured charge densities. Significantly
higher transient electric fields may also be accessible in the
vicinity of the electrolyte ions that accumulate at the interface.
The largest selectivity change observed here, a factor of 63,
corresponds to a 2.5 kcal/mol perturbation of the difference
between activation barriers leading to the two products. To
account for this perturbation with a field dipole model in which
the field is 1 V/nm, the difference in dipole moments between
the competing activation barriers (Δμ) must be 5.1 D along the
direction of the field. We hypothesize that the transition state
leading irreversibly to 2 has a dipole moment that is several
debye larger than the transition state leading to 3 and that the
reactants/transition states are readily aligned in the local field at
the interface. Thus, field−dipole interactions lower the barrier
to 2 relative to 3 to an extent that depends on the magnitude of
the local field but not its direction, accounting for the symmetry
of the selectivity changes at ±V. Interactions between the local
field and induced dipole moments as well as ordering of the
solvent molecules at the interface due to field−solvent dipole
interactions may also contribute to the selectivity changes.
research. The Eli Lilly New Faculty Award (M.K.), the
Benchmark Stanford Graduate Fellowship (C.G.), and the
Althouse Family Stanford Graduate Fellowship (E.B.) are also
gratefully acknowledged. We thank Profs. C. E. D. Chidsey and
J. Du Bois for helpful discussions and Prof. J. Brauman for a
thoughtful review of the manuscript.
REFERENCES
■
(1) Yudin, A. K.; Afagh, N. A. Angew. Chem., Int. Ed. 2010, 49, 262.
(2) Young, I. S.; Baran, P. S. Nat. Chem. 2009, 1, 193.
(3) Boxer, S. G. J. Phys. Chem. B 2009, 113, 2972.
(4) Ohta, N. Bull. Chem. Soc. Jpn. 2002, 75, 1637.
(5) Murgida, D. H.; Hildebrandt, P. J. Phys. Chem. B 2002, 106,
1
(
2
2814.
6) Harris, J. M.; Oklejas, V.; Uibel, R. H.; Horton, R. Anal. Chem.
008, 80, 1891.
(7) Blatter, F.; Frei, H. J. Am. Chem. Soc. 1994, 116, 1812.
(8) Blatter, F.; Sun, H.; Vasenkov, S.; Frei, H. Catal. Today 1998, 41,
297.
(9) Warshel, A.; Sharma, P. K.; Kato, M.; Parson, W. W. Biochim.
Biophys. Acta 2006, 1764, 1647.
10) Warshel, A.; Sharma, P. K.; Kato, M.; Xiang, Y.; Liu, H. B.;
Olsson, M. H. M. Chem. Rev. 2006, 106, 3210.
11) Shaik, S.; de Visser, S. P.; Kumar, D. J. Am. Chem. Soc. 2004,
26, 11746.
12) Shaik, S.; Hirao, H.; Chen, H.; Carvajal, M. A.; Wang, Y. J. Am.
Chem. Soc. 2008, 130, 3319.
13) Shaik, S.; Lai, W. Z.; Chen, H.; Cho, K. B. J. Phys. Chem. Lett.
010, 1, 2082.
(
(
1
(
The results obtained for the reaction of 1 catalyzed by Al O3
2
in the parallel plate cell are not unique to this combination of
substrate and catalyst. In a preliminary examination of an
alternative substrate, a 30−40-fold change in the ratio of
aldehyde to ketone products is observed at ±4.5 V relative to 0
V if 2-(4-chlorophenyl)-3-phenyloxirane is used as the
substrate. In addition, a 2−7-fold change in the ratio of 2:3 is
(
2
(14) Vayenas, C. G.; Bebelis, S.; Pliangos, C.; Brosda, S.; Tsiplakides,
D. Electrochemical Activation of Catalysis: Promotion, Electrochemical
Promotion, and Metal-Support Interactions; Kluwer Academic: New
York, 2001.
observed at ±4.5 V relative to 0 V if an HfO layer is used as
2
the catalyst instead of Al O (see Supporting Information).
2
3
(
(
15) Vayenas, C. G.; Koutsodontis, C. G. J. Chem. Phys. 2008, 128.
16) Neophytides, S. G.; Tsiplakides, D.; Stonehart, P.; Jaksic, M. M.;
The breakdown voltage and interfacial capacitance of the
Al O -coated Si electrodes limit the maximum attainable
2
3
Vayenas, C. G. Nature 1994, 370, 45.
(17) Ploense, L.; Salazar, M.; Gurau, B.; Smotkin, E. S. J. Am. Chem.
Soc. 1997, 119, 11550.
interfacial field strength in the parallel plate cell. The use of
29
alternative insulating layers with higher dielectric constants or
30,31
alternative electrolytes such as ionic liquids
may enable the
(
(
(
18) Salazar, M.; Smotkin, E. S. J. Appl. Electrochem. 2006, 36, 1237.
19) Imbihl, R. Prog. Surf. Sci. 2010, 85, 241.
20) Toghan, A.; Rosken, L. M.; Imbihl, R. Phys. Chem. Chem. Phys.
application of substantially larger fields. For reactions in which
‡
E•Δμ contributes to ΔΔG , such increases would dramatically
improve the control over selectivity afforded by this approach.
Localizing other thin film, nanoparticle, or molecular
catalysts in interfacial fields will enable the study of field effects
across a wide spectrum of reactions. Furthermore, the design
principles of the parallel plate cell may in theory be extended to
flow reactors that enable field effects to be exploited on a
preparative scale. We anticipate that field effects will provide a
general approach toward controlling the selectivity of catalytic
reactions.
2010, 12, 9811.
(21) George, S. M. Chem. Rev. 2010, 110, 111.
(22) Bard, A. J.; Faulkner, L. R. Electrochemical Methods:
Fundamentals and Applications; John Wiley & Sons: New York, 2001.
(
23) Kahn, A.; Vaynzof, Y.; Dennes, T. J.; Schwartz, J. Appl. Phys.
Lett. 2008, 93, 103305.
24) Petrii, O. A.; Khomchenko, I. G. J. Electroanal. Chem. 1980, 106,
77.
25) In a simplified model, a charge density of 1 μC/cm on the
(
2
(
2
underlying Si electrode balanced by an oppositely charged plane of
ions in the double layer generates a field of 1.1 V/nm at the oxide
2
ASSOCIATED CONTENT
surface. The actual field at the oxide surface with 1 μC/cm on the
■
underlying Si deviates from this value depending on the diffusivity of
the double layer, the proportion of ions that are specifically adsorbed,
and the polarization of the reaction medium in the vicinity of the
surface. The field also fluctuates due to the mobility of ions in the
double layer.
*
S
Supporting Information
Experimental procedures, characterization data for the oxide-
coated wafers, and additional control experiments. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(26) Becucci, L.; Guryanov, I.; Maran, F.; Guidelli, R. J. Am. Chem.
Soc. 2010, 132, 6194.
(
(
(
(
27) Kabalka, G. W.; Pagni, R. M. Tetrahedron 1997, 53, 7999.
28) Ruizhitzky, E.; Casal, B. J. Catal. 1985, 92, 291.
29) Leskela, M.; Ritala, M. Thin Solid Films 2002, 409, 138.
30) Li, J. H.; Liu, H. T.; Liu, Y. Phys. Chem. Chem. Phys. 2010, 12,
AUTHOR INFORMATION
■
Corresponding Author
mkanan@stanford.edu
1
685.
ACKNOWLEDGMENTS
(31) Simon, P.; Gogotsi, Y. Nat. Mater. 2008, 7, 845.
■
We thank Stanford University and the Air Force Office of
Scientific Research (FA9550-11-1-0293) for support for this
1
89
dx.doi.org/10.1021/ja210365j | J. Am. Chem.Soc. 2012, 134, 186−189