Interference of the surface of the solid on the performance of tethered molecular catalysts

Article abstract of DOI:10.1021/ja304181q

The catalytic performance of cinchonidine in the promotion of thiol additions to conjugated ketones was used as a probe to assess the tethering of molecular functionality onto solid surfaces using well-known "click" chemistry involving easy-to-react linke

Full text of DOI:10.1021/ja304181q

Article
pubs.acs.org/JACS
Interference of the Surface of the Solid on the Performance of
Tethered Molecular Catalysts
Junghyun Hong and Francisco Zaera*
Department of Chemistry, University of California, Riverside, California 92521, United States
ABSTRACT: The catalytic performance of cinchonidine in the promotion
of thiol additions to conjugated ketones was used as a probe to assess the
tethering of molecular functionality onto solid surfaces using well-known
“click” chemistry involving easy-to-react linkers. It has been assumed in
many applications that the tethered molecules retain their chemical
properties and dominate the chemistry of the resulting solid systems, but it
is shown here that this is not always the case. Indeed, a loss of
enantioselectivity was observed upon tethering, which could be accounted
for by a combination of at least three effects: (1) the nonselective catalytic
activity of the surface of the solid itself; (2) the activity of the OH species
generated by hydrolysis of some of the Si−alkoxy groups in the trialkoxy moieties used to bind many linkers to oxide surfaces;
and (3) the bonding of the molecule to be tethered directly to the surface. Several ideas were also tested to minimize these
problems, including the silylation of the active OH groups within the surface of the oxide support, the selection of solvents to
optimize silane polymerization and minimize their breaking up via hydrolysis or alcoholysis reactions, and the linking at defined
positions in the molecule to be tethered in order to minimize its ability to interact with the surface.
Here we explore the effect that the surface of the solid exerts
on the catalytic behavior of tethered molecular functionalities.
The test system chosen in this study is a catalyst consisting of
cinchonidine molecules tethered to silica surfaces, used to
promote the addition of aromatic thiols to conjugated
cyclohexanones.16,17 Specifically, we report on results on the
addition of p-tertbutylbenzenethiol to 2-cyclohexene-1-one
according to the reaction in Scheme 1:
1. INTRODUCTION
In general, heterogeneous catalysts are preferred to their
homogeneous counterparts: they are easier to handle and
recycle, and they allow for a more facile separation of the
products of reaction. On the other hand, it is difficult to match
the level of selectivity that can be attained with homogeneous
catalysts using heterogeneous materials. Simply, even with the
new nanotechnologies being developed nowadays, it is not yet
possible to imprint on solid samples the level of molecular
detail required for complex catalytic sites.1,2 A potential
compromise is to “heterogenize” homogeneous catalysts by
anchoring or tethering them onto solid surfaces. Anchoring has
in fact been explored for decades with organometallic
catalysts,3,4 but the results have been mixed, either because
those catalyst slowly leach out of the solid, or because the new
constrained geometry imposed on the organometallic complex
by the surrounding solid changes the catalytic behavior.
Alternatively, molecular functionality may be added to typical
catalyst supports such as silica or alumina to enhance their
catalytic activity.5,6 New “click” chemistry7 has been advanced
to facilitate the attachment of molecular compounds or
structures to surfaces. In one approach that has become quite
popular, silylating reagents are used for this purpose.5,8,9 In fact,
molecular functionalities tethered with such silylating agents
have been extensively used for other applications beyond
catalysis, including for the resolution of racemates in
chromatography and other separation techniques,10,11 as
sensors,12,13 and in biotechnology,13−15 to mention a few.
Surprisingly, though, not much work has been carried out to
characterize the effect of the tethering process on the
performance of the resulting solid, in particular in contrast
with that of the free, nontethered, molecule.
Scheme 1
This reaction is easily promoted by acids or bases, and when
chiral compounds such as cinchonidine are used as catalysts, it
leads to the production of an enantio-enhanced product. The
tethering of the cinchonidine was done by using propyltrie-
thoxysilanes terminated with appropriate linking groups, as
reported before by us18 and others.19−22 Two approaches were
tested in terms of the derivatization of the cinchonidine itself,
one where a carbamate link is made by reacting 3-
isocyanatopropyltriethoxysilane (ICPTEOS) with the OH
moiety of cinchonidine, following the steps in Scheme 2, and
a second where 3-mercaptopropyltriethoxysilane (MerPTEOS)
was used to create a mercapto linkage through the vinyl moiety
of the cinchonidine according to Scheme 3.
Received: April 30, 2012
Published: July 27, 2012
© 2012 American Chemical Society
13056
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
Scheme 2
Scheme 3
Tethering by reversing the steps in Schemes 2 and 3, namely,
by anchoring the ICPTEOS or MerPTEOS linkers to the silica
surface first and adding the cinchonidine afterward, were tried
as well; both sequences worked, but the ones indicated above in
Schemes 2 and 3 led to higher cinchonidine density on the
surface. Finally, tethering was performed on both Aerosil23 and
ordered mesoporous SBA-1524 silica supports. Only typical
examples of our results for selected samples are reported here,
but similar results and conclusions were reached with all these
combinations.
It was found that cinchonidine tethered to silica surfaces can
indeed promote the thiol addition in Scheme 1 at rates
comparable to those seen when using free cinchonidine in
solution. On the other hand, the enantioselectivity obtained
with the tethered catalysts was not always as high as that seen in
homogeneous solutions. Three factors were identified that
explain this behavior: (1) a nonenantioselective promotion of
the reaction by acidic sites within the silica solid; (2) a possible
activity of the free OH groups in the silane end of the linker not
bonded to the silica substrate; and (3) additional direct
bonding of cinchonidine to the silica surface. Ways to minimize
all these effects were tested, and enantioselectivities comparable
to those in solution were eventually accomplished with the
proper tethered catalysts. We believe that the factors identified
in this study may interfere with many other tethered systems,
and the solutions proposed here to be fairly general and
applicable to those as well.
2. EXPERIMENTAL SECTION
The synthetic steps for the preparation of anchored cinchonidine on
silica surfaces and for their subsequent treatments were all done using
commercial chemicals: cinchonidine (Cd, Fluka, 98%), 3-isocyanato-
propyltriethoxysilane (ICPTEOS, Sigma-Aldrich, 95%), 3-mercapto-
propyltriethoxysilane (MerPTEOS, Sigma-Aldrich, 95%), Aerosil A200
silica (Degussa-Huls), SBA-15 (ACS Material), 3-aminopropyltrie-
̈
thoxysilane (APTES, Sigma-Aldrich, 99%), dibutyltindilaurate
((CH3CH2CH2CH2)2Sn(OCO(CH2)10CH3)2, Sigma-Aldrich, 95%),
azoisobutyronitrile (AIBN, Sigma-Aldrich, 98%), hexamethyldisilazane
(HMDS, Sigma-Aldrich, >99%), ethanol (EtOH, Gold Shield, 200
Proof), chloroform (Mallinckrodt chemicals, 99.8%), ethyl ether
(EMD chemicals, 99%), dichloromethane (Fisher chemicals, 99.9%),
pentane (J. T. Baker, 100%), cyclohexane (Fisher chemicals, 99.9%),
tetrahydrofurane (THF, EMD Chemicals, 99.5%), and toluene
(Sigma-Aldrich, 99.8%).
The carbamate-linked catalysts were made by first dissolving the
cinchonidine in THF, adding the ICPTEOS together with a small
amount of dibutyltindilaurate, and refluxing the mixture at 343 K for 3
h. After cooling down and letting the mixture rest for a day, the solvent
(THF) was evaporated and pentane was added to precipitate the new
compound (cinchonidinyl(3-triethoxysilylpropyl)carbamate, Cd−
TEOSPC), which was filtered, washed with more pentane, recrystal-
lized in chloroform and cyclohexane, and dried under nitrogen. The
Cd−TEOSPC was then redissolved in ethanol and/or toluene and
added to the silica, and that mixture was refluxed at 400 K for two days
(or other times, as indicated in Figure 8). The final product, the
cinchonidinyl(3-propylcarbamate)-functionalized silica (Cd−PC−Sili-
13057
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
ca), was let to cool down, washed with ethanol, toluene and/or diethyl
ether, and dried in vacuum before use.
activity and selectivity of the different mixtures tested are
summarized in Figure 1. The left panel, which provides
For the mercapto-linked catalysts, cinchonidine and azoisobutyr-
onitrile (AIBN) were mixed in chloroform, MerPTEOS was added,
and the mixture refluxed at 343 K for 24 h. The mixture was cooled
down, the solvent (chloroform) was evaporated, and pentane was
added to precipitate the new compound (11-(3-
triethoxysilylpropanethio)hydrocinchonidine, Cd−TEOSPMer),
which was filtered, washed with more pentane, recrystallized in
chloroform and cyclohexane, and dried under nitrogen. The Cd−
TEOSPMer was then redissolved in toluene and added to the silica,
and that mixture was refluxed at 400 K for two days. The final product,
the 11-(3-propanethiyl)-hydrocinchonidine-functionalized silica (Cd−
PMer−Silica), was cooled down, washed with ethanol, toluene and/or
diethyl ether, and dried in vacuum before use.
The samples obtained after each step of the synthesis were
characterized by transmission infrared absorption spectroscopy using a
Bruker Tensor 27 Fourier-transform infrared (FTIR) spectrometer. A
small amount of the liquids was placed between two NaCl windows,
the solids were pressed into pellet form, and all samples were then
placed at the focal point of the sample compartment of the FTIR
instrument. The data correspond to averages of 1024 scans, taken at 4
cm−1 resolution and reference to appropriate spectra from an empty
NaCl window setup or from pellets made out of pure silica.
Figure 1. Summary of catalytic performance of samples made out of
mixtures of cinchonidine (Cd) and silica for the promotion of the thiol
addition reaction depicted in Scheme 1. (Left and Center) Yield after
24 h and specific optical rotation of the product, respectively, for the
room-temperature conversion of 4-tert-butylbenzenethiol (0.3 mL, 1.8
mmol) and 2-cyclohexene-1-one (0.17 mL, 1.8 mmol) catalyzed by,
from left to right, Cd alone in solution (4.4 mg, 15 μmol, green), Cd +
25 nm-silica beads (20 mg, blue), Cd + Aerosil silica (20 mg, purple), a
tethered Cd-Carbamate-Aerosil solid (Cd−PC−Aerosil, 20 mg,
magenta), and Aerosil silica alone (20 mg, red). (Right) Specific
optical rotation of the product obtained after 24 h room-temperature
conversion using physical mixtures of Cd (4.4 mg, 15 μmol) and the
indicated amounts of Aerosil. All these data point to significant
catalytic promotion by both Cd and silica but enantioselectivity
associated only with Cd.
Solid-state 29Si CP-MAS NMR spectra were acquired on a Bruker
Avance 600 spectrometer, employing a cross-polarization contact time
of 2 ms, a 1H decoupling bandwidth of 80 kHz, and a recycle time of 3
s. Data were acquired as 12 000 coadded 2048 complex data point
FIDs with a 100 kHz sweep width. Post acquisition processing
consisted of exponential multiplication with 200 Hz of line broadening
and zero filling to 4096 data points. Chemical shifts were referenced to
an external DSS sample.
For the XPS studies reported in Figure 9, a 1 cm × 1 cm square
piece of a Si(100) wafer was exposed to a solution of APTES in
toluene and refluxed at 385 K for a specified amount of time (1, 24, or
48 h, Figure 9). The resulting sample was then cooled down and
washed with acetone and ethanol, and dried under vacuum prior to the
XPS analysis, which was performed using a Kratos AXIS Ultra DLD.
The catalytic performance of our samples was tested by using the
thiol addition reaction shown in Scheme 1.16,17 p-Tertbutylbenzene-
thiol (Sigma-Aldrich, 97%) and 2-cyclohexene-1-one (Sigma-Aldrich,
95%) were mixed in benzene and let to react at room temperature for
a day (with or without catalyst depending of the case being studied).
The product, 3-(p-tertbutylbenzenelthio)cyclohexanone, was extracted
using a separation funnel, washed sequentially with benzene, HCl (2.0
N, twice), KOH (2.0 N, twice), and a saturated NaCl solution, and
dried over MgSO4. The resulting benzene solution was filtered, the
solvent evaporated, and the remaining liquid dried in vacuum. The
product was purified using a chromatographic column and a hexane/
ethylacetate = 10:1 mixture as the developing solvent. The identity of
information about the extent of conversion after 24 h of
reaction at room temperature, indicates that all the catalysts
investigated in this set, including the pure Aerosil silica
(without any cinchonidine), are fairly active. On the other
hand, the data in the center panel, which correspond to the
specific optical rotation measured for the product of reaction,
indicate a decrease in enantioselectivity for all catalysts that
include silica solids. The role of the silica surface in diluting the
enantioselective catalysis of cinchonidine is more clearly
indicated by the data in the right panel, where the specific
rotation of the product is plotted as a function of the amount of
silica physically mixed with 4.4 mg (15 μmol) of cinchonidine
in solution: the rotation values decrease monotonically with
increasing amount of added silica.
Also included in the data in Figure 1 are values obtained for a
catalyst prepared by tethering cinchonidine to Aerosil silica via
the hydroxyl moiety using the ICPTEOS linker, to yield a
cinchonidinyl(3-propylcarbamate)-functionalized silica (20 mg
Cd-Carbamate-Aerosil− or Cd−PC−Aerosil−, Scheme 2).18 It
can be seen from the data that the derivatized catalyst also
shows high catalytic activity for the thiol addition reaction in
Scheme 1, even if quantification of this activity in terms of
turnover frequency (TOF) is difficult because of the difficulty
in evaluating the concentration of cinchonidine tethered to the
silica surface. On the basis of previously reported acid−base
titration results using HCl as the titrant and methyl red as the
indicator,18 which estimated the cinchonidine density on
Aerosil to be approximately 50 μmol/g, TOF values of about
1
the product was corroborated by H NMR, and its optical activity
(reported as optical specific rotation) was measured in an ethanol
solution using a MC 241 polarimeter with a Na lamp (λ = 589 nm).
3. RESULTS
Initial testing of the tethering of Cd on silica supports was
provided in a previous report.18 It was found that the tethered
catalyst is still active for the thiol addition reaction depicted in
Scheme 1, but not with the same enantioselectivity as when Cd
is used in free solution. In our quest to understand this partial
loss of enantioselectivity upon tethering of the catalysts, it was
realized that the silica support exhibits some (nonenantiose-
lective) catalytic activity of its own. Evidence for the catalytic
activity of silica supports in this acid/base-catalyzed reaction
and for the dilution of the enantioselectivity provided by the
use of cinchonidine was obtained by testing physical mixtures
of the cinchona alkaloid with different silica samples, including
silica beads (20 mg) and Aerosil silica (20 mg). The data on the
13058
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
1 and 10 s−1 were measured for the free cinchonidine (and the
free Cd-TEOSPC and Cd−TEOSPMer) and the tethered Cd
within the first 5 min of reaction, respectively. Quite possibly,
the derivatized solid catalysts are intrinsically more active
toward the thiol addition reaction than the free cinchonidine
catalysts, an observation that, again, would be consistent with
the silica independently promoting the reaction. From the data
in the right panel of Figure 1, it could be estimated that the
activity of 4.4 mg of Cd can be matched by using 20−30 mg of
pure Aerosil. In terms of enantioselectivity, the Cd-tethered
sample reported in Figure 1 displays approximately one-third of
the enantioselectivity seen with pure Cd in solution (Figure 1,
center panel).
It is clear from the data in Figure 1 that the silica supports do
themselves show catalytic activity toward the thiol addition
reaction in Scheme 1. It could be presumed that it may be the
acidic catalytic sites, most likely the hydroxo groups known to
exist on the surface of silica, that promote the thiol addition
reaction. To minimize this catalytic contribution from the silica,
blocking of the hydroxo sites via a silylation reaction using
hexamethyldisilazane (HMDS) was tested.25 The effectiveness
of blocking OH surface groups using this approach was
evaluated by using infrared absorption spectroscopy, since both
isolated and hydrogen-bonded surface OH groups can be easily
identified by the sharp peak at 3747 cm−1 and the broad band
between approximately 3000 and 3800 cm−1 associated with
those species, respectively. Addition of HMDS to the surface
could also be evaluated by the appearance of its characteristic
IR features, in particular those associated with the stretching
modes of the C−H of the terminal methyl moieties at 2900 and
2965 cm−1, as well as by the disappearance of the N−H
stretching mode seen in the original free HMDS at 3380 cm−1.
This behavior is clearly seen in the spectra shown in Figure 2,
which correspond to the case of HMDS addition to Aerosil
silica. Two silylated samples are reported, after washing with
ethanol and toluene solvents, respectively (together with
reference spectra for the free HMDS and Aerosil).
It was found that HMDS is particularly effective at blocking
the isolated OH groups: the sharp peak for that species at 3747
cm−1 in the IR spectra is completely gone once the surface has
been silylated. There are nevertheless some remaining hydro-
gen-bonded hydroxo groups in these samples. A majority of
those are likely to be internal moieties not accessible by
reactants from solution, but some exposed groups are still
evident by the differences seen between the samples treated
using ethanol versus toluene as the solvent; the latter appear to
be more effective at removing the hydrogen-bonded OH
groups. In addition, it was determined that approximately half
of the activity of Aerosil toward the thiol addition reaction
remains upon silylation even when using toluene as the solvent
(after adding mercapto linkers). This was found to be the case
regardless, whether catalysts were used with the mercapto
surface groups alone or after silylating the remaining sites using
HMDS. Also, it was determined that, in general, the use of
ethanol leads to surfaces with higher concentrations of hydroxo
groups, specifically those associated with the low-frequency
range of the O−H IR band around 3500 cm−1.
Figure 2. Infrared (IR) absorption spectra for, from top to bottom:
free Aerosil silica (blue); Aerosil silylated with hexamethyldisilazane
(HMDS), then washed with ethanol (EtOH, purple); Aerosil silylated
with HMDS, then washed with toluene (magenta); and free HMDS
(green). The addition of the HMDS to the Aerosil is indicated by the
corresponding C−H stretching peaks at 2900 and 2965 cm−1, and the
blocking of the isolated silica OH groups by the disappearance of the
3747 cm−1 sharp feature in the spectra. Some hydrogen-bonded OH
groups, identified by absorption between 3000 and 3800 cm−1, are
blocked as well, more efficiently if toluene (rather than ethanol) is
used as the solvent.
contrasted before versus after silylation in order to evaluate the
effectiveness of that treatment in blocking the silica active sites,
and those samples were also tested after washing with ethanol
versus toluene to further assess the effect of the solvent.
Contrasting data in terms of the enantioselectivity of the thiol
addition reaction, reflected by the specific optical rotation of
the product, are reported in Figure 3. Catalysts made by
tethering cinchonidine via its vinyl group, using (3-
mercaptopropyl)triethoxysilane as the linker, were used in
this case (Scheme 3). It can be seen in Figure 3 that the
enantioselectivity obtained with catalysts made using toluene as
the solvent is much larger than that recorded with catalysts
washed with ethanol, and essentially the same obtained with
free cinchonidine in solution. This was in fact true even if the
silica surface was not treated with the silylation agent,
presumably because the triethoxysilane moiety of the linker
acts as a silylation agent itself (spectroscopic evidence of this is
provided below). When ethanol is used as the solvent, on the
other hand, enantioselectivity does increase somewhat with
silylation using HMDS, but not significantly in comparison with
that seen with the free Cd or with the catalysts treated with
toluene; it never reaches more than 40% of the specific rotation
values obtained with the better samples. Ethanol is a common
choice of solvent in these systems, the reason why we used it in
our original report,18 but it is clear that toluene is a better
choice and should be preferred.
These and other data showed that the choice of solvent for
the treatment of the silica surfaces is critical in defining their
final properties, one of the key observations of this study. That
is further highlighted by the enantioselectivity catalytic data
measured using these materials. Specifically, the catalytic
properties of the cinchona-derivatized Aerosil solids were
The competing chemistries of surface derivatization and
silylation were characterized in more detail next for the case
where toluene was used as the solvent. The stepwise evolution
of the surface species after each of the derivatization and/or
silylation steps is mapped out by the IR absorption spectra
shown in Figure 4. Initial addition of the (3-mercaptopropyl)-
13059
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
the isolated OH groups evidenced by the sharp peak at 3747
cm−1, a result consistent with the conclusion mentioned in the
previous paragraph in connection with the data in Figure 3
regarding the triethoxysilane moiety of the linker acting as
silylation agents itself. Either way, the remaining isolated
hydroxo sites can be blocked easily by silylation using HMDS
(third trace from top). The order of the surface derivatization
can in principle be reversed, by adding the HMDS first, but in
that case the subsequent addition of the linker is limited
because most surface OH sites required for tethering have
already been blocked by the silylation (fourth trace from top).
Additional experiments were performed to probe the nature
of the resulting surfaces from tethering cinchonidine to the
silica surfaces. Specifically, the samples obtained by performing
the cinchonidine tethering step last, after sequentially adding
the mercapto linker and HMDS, were contrasted against a
sample obtained by first linking the MerPTEOS to the
cinchonidine and then tethering the coupled molecule to the
Aerosil support. The relevant IR absorption data, provided in
Figure 5, demonstrate that cinchonidine can certainly be added
Figure 3. Specific optical rotation of the product from the thiol
addition reaction catalyzed by tethered Cd−Mercapto−Aerosil (Cd−
PMer−Aerosil) catalysts, as synthesized (left) and after silylation using
HMDS (center). Data are shown for catalysts washed with both
ethanol (course-hatched red) and toluene (solid blue). The specific
optical rotation value obtained with free Cd is also provided for
reference (right, fine-hatched green). The right-hand scale shows the
appropriate enantiomeric excess (ee), calculated by using literature
information.16 The silylation of the catalyst washed with ethanol leads
to a slight increase in enantioselectivity, but the larger effect is seen
when switching to toluene as the washing agent, with which the
tethered catalyst show a performance comparable to that of free Cd.
Figure 5. IR absorption spectra to illustrate the optimum procedure
for tethering cinchonidine to silylated silica. Shown are, from top to
bottom, traces for: Aerosil first derivatized with MerPTEOS and then
silylated with HMDS (blue, the same as third trace from top in Figure
4); previous sample after Cd addition (red); and Aerosil derivatized
with Cd−TEOSPMer (Cd-Mercapto-Aerosil, green). The latter
synthesis leads to a higher density of Cd on the surface, but shows
some isolated silica OH groups still exposed.
Figure 4. IR absorption spectra to illustrate the evolution of the
surface of Aerosil upon bonding of the mercapto tethering agent
(MerPTEOS) and silylation with HMDS. The traces correspond, from
top to bottom, to: pure Aerosil (green); Aerosil derivatized with
MerPTEOS (ochre); Aerosil first derivatized with MerPTEOS and
then silylated with HMDS (magenta); Aerosil first silylated with
HMDS and then derivatized with MerPTEOS (red); free MerPTEOS
(purple); and free HMDS (blue). The best treatment for molecular
tethering was determined to be a sequential derivatization with the
tethering linker and silylation, in that order.
last: its incorporation into the surface is mainly manifested by
the quinoline deformation modes seen in the low-frequency
end of the appropriate IR trace, in particular the peaks at 1463,
1512, and 1593 cm−1 (middle trace). It should be noted that
the cinchonidine coverage in this case is significantly lower than
when the Cd-TEOSPMer is made first and then tethered into
pure Aerosil (bottom trace). On the other hand, the latter
sample displays much more intensity in the low-frequency end
(∼3400 cm−1) of the hydrogen-bonded OH section of the IR
spectrum, indicating the availability of many more silica sites for
nonselective catalysis.
triethoxysilane linker to Aerosil is facile, as indicated by the
appearance of its characteristic C−H stretching and C−H
deformation IR bands in the 2800−3000 and 1300−1500 cm−1
regions, respectively, in the spectra from the tethered sample
(compare the second versus fifth traces, from the top, in Figure
4). The derivatization consumes most, even if not always all, of
The possibility that the presence of specific surface sites
within the silica support or the mode in which cinchonidine is
tethered to the surface may contribute to the changes in the
13060
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
enantioselectivity reported here was assessed as well. For this,
the specific optical rotation of the product from the thiol
addition reaction was measured for catalysts based on Aerosil
versus SBA-15, an ordered mesoporous silica-based material,
and in samples where the cinchona was derivatized at the
hydroxo (using a carbamate bond) versus at the vinyl (with a
mercapto link) moiety. The data, shown in Figure 6, show
Figure 7. 29Si cross-coupling magic-angle-spinning (CP-MAS) NMR
spectra for Cd-Carbamate-Aerosil catalysts washed with toluene (top,
red) and ethanol (blue, bottom). Two sets of peaks are observed in
these data around −80 to −120 ppm and −40 to −70 ppm,
corresponding to Si atoms from the original silica and the newly
attached silane, respectively. Much more extensive derivatization is
seen with the toluene sample, mostly in the form of polymerized
silanes.
Figure 6. Specific optical rotation of the product from the thiol
addition reaction catalyzed by three solid catalysts consisting of Cd
tethered to silica, namely (from left to right): Cd-Carbamate-Aerosil;
Cd-Mercapto-Aerosil; and Cd-Mercapto-SBA-15. Data are shown for
catalysts washed with both ethanol (hatched red) and toluene (solid
blue). The right-hand scale shows the appropriate enantiomeric excess
(ee), calculated by using literature information.16 Minor differences in
catalytic performance are observed depending on the nature of the
silica or the bonding mode of the Cd to the linker, but the main factor
determining enantioselectivity remains the choice of solvent used
during synthesis.
on the Aerosil sample washed with toluene. Furthermore, that
increase is almost exclusively associated with silane silicon
atoms without any free hydroxo groups.
Quantitation of the extent of derivatization seen in this
samples indicates a total number of new silicon atoms,
originating from the added silane, exceeding the total number
of OH groups originally available on the silica surface: the new-
silane/old-SiOH-in-silica silicon atom ratios amount to about
1.2 and 3.1 for the ethanol- and toluene-washed samples,
respectively. This is consistent with the silylation effect seen in
the catalytic performance of the tethered catalysts as discussed
above in connection with Figure 3: most of the initial surface
OH groups are derivatized and made catalytically nonactive for
our acid−base catalyzed reaction. It also means that some
surface derivatization must involve OH groups from silanes
partially bonded to the silica, which means that the anchored
silanes in fact polymerize on the surface. The extent of this
polymerization is significantly higher in the case of the toluene-
washed sample, suggesting that ethanol may reverse that
process, possibly by realcoholizing some of the silane bonds.
Also from the data in Figure 7 it can be estimated that the bond
of the silane to the surface involves on average approximately
two Si−O−Si links in all cases.
The silane polymerization process on the silica surface is
more clearly seen in Figure 8, where the evolution of the 29Si
CP-MAS NMR spectra is shown as a function of derivatization
time for the case of the synthesis of Cd-carbamate-Aerosil
samples using ethanol as the washing solvent. In general, it is
clear that the amount of silane increases with derivatization
time: the features in the −40 to −70 ppm region, which
correspond to the silicon atoms from the Cd-carbamate-linker
added to the surface, grow in intensity. Direct bonding to the
original silica surface occurs quite rapidly, within the first hour
some small differences in performance across the three catalysts
referenced there. For instance, the carbamate-tethered sample
washed with toluene exhibits a slightly lower enantioselectivity
than the other toluene-washed catalysts, perhaps because
derivatization via the hydroxo moiety may disturb the chiral
pocket during catalysis.26−28 Also, the sample made using SBA-
15 and ethanol as the solvent shows a marginally better
performance compared to the other ethanol-washed samples,
possibly because the bonds made during tethering are sturdier
in that case. However, none of these effects are significant in
comparison with the main factor defining the enatioselectivity
of these cinchonidine catalysts, which remains the solvent used
for washing the tethered samples.
To better understand the difference induced by washing the
catalysts with ethanol versus toluene during the tethering
process, the nature of the bonding of the linker and
cinchonidine to the silica surface was explored further by
using 29Si CP-MAS NMR. Figure 7 displays typical data,
obtained for the Cd-Carbamate-Aerosil case after washing with
ethanol versus toluene. Analysis of these spectra can be made
based on well-established peak assignments:18,29,30 the group of
features around −80 to −120 ppm are associated with Si atoms
from the original silica and the peaks in the −40 to −70 ppm
range correspond to Si from the newly attached silane, and
within each group, three peaks are identified with chemical
shifts that decrease with increasing number of OH
substitutions. On that basis, it is clear from Figure 7 that
there is a significantly higher density of anchored silane groups
13061
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
Figure 9. Si 2p (left), C 1s (center), and N 1s (right) XPS data for a
Si(100) wafer treated with 3-aminopropyltriethoxysilane (APTES) as a
function of derivatization time. The thin film of silicon dioxide that
exists on untreated Si(100) surfaces, indicated by the Si2+ peak seen in
the Si 2p XPS of the original sample, was used here to emulate the
surface of silica. The initial reaction of the triethoxysilane moieties in
APTES with the surface is manifested by the increase in the Si2+ signal
and the decrease in the Si(0) peak intensity seen in the Si 2p XPS data
after only 1 h, and the addition of the APTES by the growth of the
main peaks observed in the C 1s and N 1s XPS. Afterward, a new N 1s
XPS feature, associated with direct bonding of the amine to the silica,
grows more slowly, and the silicon surface also continues to oxidize,
until no reduced Si atoms are seen after 48 h.
Figure 8. 29Si CP-MAS NMR spectra for Cd-Carbamate-Aerosil
catalysts as a function of derivatization time in a synthesis where
ethanol was used as the washing solvent. The initial fast derivatization
of the OH groups in silica is followed by a slower further growth
associated with silane polymerization.
of derivatization, at which point a significant number of single
(Si−OH, −100 ppm) and geminal (Si(OH)2, −91 ppm)
hydroxo groups from the original surface are converted to form
fully oxygen-bonded (SiO4, −108.5 ppm) silicon atoms. This
means that new Si−O−Si bonds are formed and that the
hydroxo-terminated silicon surface atoms in the original
substrate are incorporated into the silicon dioxide network
just underneath the surface. No significant additional changes
are observed in the silicon atoms associated with the original
silica surface after the first hour of derivatization, but further
silane anchoring is still observed. This anchoring is first
dominated by species bonded via two Si−O−Si bonds, but it is
overtaken by the formation of three Si−O−Si links at later
times, especially after 48 h of reaction. This is consistent with
the silane polymerization process discussed above.
More details on the nature of the system that results from
bonding linking agents to silicon oxide surfaces were extracted
from experiments using X-ray photoelectron spectroscopy
(XPS). In this case, an untreated Si(100) wafer, which is
covered by a ∼1−2 nm-thick native silicon oxide film,31 was
derivatized with APTES, another common propyltriethoxysi-
lane linker. Figure 9 displays the Si 2p (left), C 1s (center), and
N 1s (right) XPS traces obtained after derivatization for various
amounts of time (1, 24, and 48 h). Reference traces are also
provided for the wafer before treatment for comparison. The Si
2p spectrum of the initial surface displays the features for Si(0)
and silicon dioxide typically seen in these samples, at 99.1 and
103.0 eV, respectively, as well as some intensity in the C 1s data
indicative of carbon contamination. Surface derivatization
becomes clear both by the increase in signal seen in the
285.3 eV peak in the C 1s XPS traces and by the development
of a new peak around 399.9 eV in the N 1s XPS data. In fact,
that derivatization is about 70−80% complete already after only
1 h of reaction of the APTES with the surface. Given that the C
1s and N 1s XPS values are typical for those atoms in free
amines,32 it is straightforward to conclude that the initial
anchoring occurs via reaction of the surface with the
triethoxysilane moiety, and that the amine end of the APTES
remains free, not directly interacting with the silicon substrate.
However, following that initial deposition stage, an additional
slower growth is seen in the N 1s signal in a second feature
centered around 401.9 eV, a value characteristic of amine
salts.32 This strongly suggests that a direct interaction of the
nitrogen atoms in the terminal amine moiety with silicon atoms
on the surface occurs at this point involving the newly added
APTES. It is also interesting to note that, upon treatment with
the APTES linker, the signal from the reduced silicon in the Si
2p XPS data slowly disappears at the expense of the formation
of an oxidized film, until only the latter is seen after 48 h. It is
possible that the surface is slowly oxidized either by the solvent
or, more likely, by the ethanol byproduct from the silane
addition to the Si(100) surface. Alternatively, this behavior
could be a reflection of the silane polymerization reaction
mentioned before.
The XPS data, in particular the evolution of the N 1s XPS
spectra with derivatization time, identified a new potential
complication in the chemistry of tethering to silica surfaces,
namely, the possibility of additional direct interactions between
the surface and the functional group to be tethered. This may
be particularly likely when that functionality is an acidic or basic
moiety, like the case of the terminal amine group in APTES. A
similar problem may occur in the case of the tethering of
cinchonidine being reported here, since the nitrogen atom in
the tertiary amine of the quinuclidine is somewhat basic, and
since there is also an OH group present in the main chiral
center of the molecule. Indeed, a new type of surface silicon
atoms was identified by 29Si CP-MAS NMR for the case of the
Cd-Mercapto moiety tethered to either Aerosil or SBA-15
13062
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
(Figure 10), a new sharp peak at −48 ppm, consistent with the
direct interaction of surface silicon atoms with either the
conjugated ketones indicated in Scheme 1. The activity of the
silica support for that reaction was shown to be comparable to
that of the cinchonidine with which the surface was to be
modified. It could be thought that the presence of additional
acidic sites on the silica only add to the basic centers provided
by the cinchonidine, leading to even more active catalysts.
However, that potential additional activity comes at the expense
of selectivity, the reason for the interest in using cinchonidine
as the promoter in this reaction. The loss of enantioselectivity
in the production of the thioether reported in Figure 1 was
used here mainly as a probe to separate the contributions from
the silica and the cinchonidine to the promotion of the
reaction, but could also constitute a major drawback if such
type of selectivity is the motivation behind the choice of the
particular functionality to be added to the surface. Molecular
selectivity may be induced by the steric or electronic properties
of the tethered modifier, which can be used to control regio-,
stereo-, or enantioselectivity in addition, elimination, or
isomerization reactions, for instance. That selectivity may be
lost if a significant fraction of the conversion is promoted by the
nonselective surface sites of the solid support.
We have also provided a potential solution for this problem,
namely, the blocking of the silica acidic sites with inert moieties
via silylation treatments. Silylation is a well-known and
common treatment, used, for instance, for the passivation of
glass surfaces,25 and was shown here to be effective in blocking
OH groups on high-surface-area silica supports (Figure 2).
Infrared absorption spectroscopy data indicated that the use of
HMDS in particular leads to the complete blocking of all
isolated OH groups within silica surfaces, and also the capping
of most, if not all, of the hydrogen-bonded OH sites available
for reaction. Interestingly, though, it was found that, at least for
the systems studied here, this treatment does not lead to
significant changes in the activity or selectivity of the catalysts
(Figure 3). Presumably, the linkers used for tethering can
themselves be viewed as silylation agents, consuming most, if
not all, of the surface OH groups available on silica (Figure 5).
Nevertheless, the addition of HMDS makes surface passivation
more extensive (Figure 5). To make the tethering of the
molecular functionality on the surface the most efficient, it was
shown that it is important to carefully consider the sequence of
steps used for tethering and silylation: it is more effective to
perform the silylation step last in order to avoid the premature
blocking of the OH surface groups needed for tethering (Figure
4). A recent report provides an alternative sequence where a
protecting group is used to preserve the tethered molecular
functionality during the silylation process.33
Even if all initial OH groups within the silica surface are
blocked, or even if they are not critical to the performance of
the molecularly modified surfaces, there is a second issue to be
considered in connection to the tethering of molecular
functionality to these solids, that bonding of the linker to the
surface may not be complete and may leave unreacted
functionalities at the interface with the surface. In the case of
the trialkoxysilanes so often used in these applications, the
potential for the formation of up to three Si−O−Si bonds per
linker is seldom realized, and an average of approximately two
bonds is seen in most cases instead (Figure 7). That leaves, on
average, a free Si−OH moiety per linker available for reactivity
(after hydrolysis of the original Si−OR). Those species may be
catalytically active, and may, again, interfere with the selectivity
of acid−base catalysis. Additionally, it was found that those Si−
OH groups are amenable to polymerization reactions with
Figure 10. 29Si CP-MAS NMR spectra from two Cd-Mercapto-Silica
catalysts, based on SBA-15 (top, red) and Aerosil (bottom, blue) as
substrates. Toluene was used as the washing solvent. Less
derivatization is seen on SBA-15, but in both cases, significant direct
bonding of the Cd to the silica is manifested by a new sharp peak
around −48 ppm. No equivalent direct bonding was seen with the Cd-
Carbamates or when ethanol was used for washing.
quinuclidine nitrogen or the alcoholic oxygen of cinchonidine.
Given that this NMR feature was not observed in the Cd-
Carbamate cases, the latter seems the most likely.
4. DISCUSSION
As mentioned in the introduction, the tethering of molecular
functionality to the surfaces of oxides is becoming a popular
way to modify the chemical properties of solids. The use of
“click” chemistry, where a linker is employed, has developed
into a particularly easy and widespread approach for this
purpose. Numerous examples exist in the literature already for
the specific case of the use of propyltriethoxysilanes, which are
available with many linking groups (such as amine, thiol,
isocyanate, azido, or organic acid moieties), and silica surfaces
as the linker and substrate, respectively. Applications for the
derivatized surfaces have ranged from catalysis to chromatog-
raphy, sensing, microelectronic devices, and biotechnology. In
most of the cases reported in the literature, it has been assumed
that the tethered molecules retains their chemical behavior
upon anchoring on the surface, and that such chemistry
dominates the reactivity of the newly prepared solid.
We have shown here that this hypothesis may not be
necessarily valid. Instead, additional active groups on the
surface may interfere with the desired chemistry of the added
molecular modifier. Three specific potential interfering factors
were identified. The first is the chemical activity of given surface
sites within the support. Even on silica surfaces, which are often
assumed to be inert, the terminal hydroxo moieties can act as
acidic centers. In the test case used in our study, those acidic
sites were shown to be capable, by themselves, of promoting
acid−base catalyzed reactions such as the thiol addition to
13063
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
other silanes (Figures 7 and 8). That leads to extensive surface
derivatization, beyond what is possible based on the initial
availability of OH groups on the silica surface, but also to the
clustering of the molecular functionality added to the solid.
The exact effect that this clustering may cause on the
performance of the derivatized surface is not clear. Never-
theless, it became evident from the results of our experiments
that the extent of this polymerization may be affected by the
solvent used in the tethering process. In particular, ethanol may
unravel some of the silane polymers that may build up on the
surface and/or slow down the polymerization process, perhaps
because the ethanol molecules can attack the newly formed Si−
O−Si bonds and realcoholize them. The data in Figures 7 and 8
clearly show that more extensive derivatization and polymer-
ization is obtained if the use of alcohols is avoided and toluene
(or another inert solvent) is used instead. Interestingly, the
catalysts prepared using toluene, where silane polymerization is
more extensive, are more enantioselective for the thiol-addition
reaction (Figures 3 and 6). One possible explanation for this
difference is that the silane polymerization consumes the
dangling OH groups left unreacted during the initial tethering
reaction, and that exposure of the catalyst to ethanol alcoholizes
some of the new Si−O−Si bonds and creates more catalytically
active but nonselective OH sites. This is undesirable, so the use
of nonreactive solvents (and the avoidance of exposure to
water, which could hydrolyze the Si−O−Si bonds) during
tethering is recommended.
Finally, a third complication comes from the potential direct
interaction of the molecular modifier with the silica surface. It
was shown above that even with the simple amine group at the
end of the APTES linker such interactions are possible: XPS
evidence was obtained for the direct bonding of the nitrogen
atom to silicon atoms within the silica surface (Figure 9). For
the case of cinchonidine tethered via a mercapto linker attached
to the peripheral vinyl moiety in the quinuclidine ring (Scheme
3), 29Si CP-MAS NMR data strongly suggest that a direct bond
is formed between the surface and the OH group of the
cinchona alkaloid (Figure 10). The latter problem could be
avoided by tethering the cinchonidine using a carbamate linker
instead, since that attaches directly to the alcohol moiety
(Scheme 2), but, either way, the chiral pocket in cinchonidine,
believed to be responsible for much of its enantioselective
catalysis,26,34,35 may be compromised. Alternatively, the
cinchonidine−surface interaction may be avoided or broken
by changing the solvent: it does not appear to survive after
washing the catalyst with ethanol. The controlled use of ethanol
may be advisable to prevent the direct bonding of the modifier
while avoiding the alcoholysis discussed above.
between cinchonidine and the acidic sites of the silica substrate;
(2) additional promotion by the OH groups created by
incomplete conversion of the triethoxysilane fragments in the
linker; and (3) a direct interaction between cinchonidine and
the silica surface.
Several possibilities were explored to eliminate, or at least
mitigate, these interferences. First, it was shown that the
hydroxo groups that act as acidic sites in silica may be blocked
using a silylating agent such as hexamethyldisilazane. By
following a proper sequence of derivatization steps involving
the propyltriethoxysilane linker and the silylation agent, all of
the isolated OH groups and most of the hydrogen-bonded OH
moieties of the silica surfaces may be passivated. In terms of the
new OH groups created by hydrolysis of the Si−ethoxy bonds
from the linker that do not participate in the attachment to the
silica surface, which were estimated to initially amount to
approximately one of the three in the triethoxysilane, those
were found to be able to participate in subsequent silane
polymerization reactions as long as no alcohols were used in
the surface derivatization process. Indeed, silica-tethered
cinchonidine catalysts prepared using only toluene as the
solvent displayed enantioselectivities comparable to those seen
with free cinchonidine. Finally, it was found that the OH group
in cinchonidine may bind directly to the silica surface,
presumably affecting its catalytic performance. This problem
could be minimized either by tethering the cinchonidine using a
carbamate linkage at that OH position, or by washing the final
catalyst briefly with ethanol. A balance needs to be struck to
avoid alcoholysis of the polymerized silanes, however, since that
leads to a decrease in enantioselectivity.
The studies reported here focused on the performance of a
specific set of tethered catalysts, but the conclusions reached in
terms of interference of the solid upon heterogenization of
homogeneous catalysts may be general. The effects cited in
terms of the potential chemical activity of the solid itself or of
the OH groups generated upon reaction of silane-based linkers
with surfaces and of the possible bonding of the tethered
molecule directly to the surface may be easily encountered in
many of the other systems that rely on similar tethering
approaches, not only in catalysis, but also in chromatography,
racemic separation applications, sensors, and biotechnologies.
AUTHOR INFORMATION
Corresponding Author
zaera@ucr.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
5. CONCLUSIONS
Financial support for this research was provided mainly by a
grant from the U.S. Department of Energy, Office of Science,
Basic Energy Sciences Program. Additional funding was
provided by the U.S. National Science Foundation. The XPS
data was acquired with an instrument purchased with funds
from the U.S. National Science Foundation, Grant DMR-
0958796.
In this study, the enantioselectivity of the addition of a thiol to
a conjugated cyclic ketone was used as a probe to evaluate the
performance of catalysts made by tethering cinchonidine to
silica surfaces. The tethering was performed by using
established “click” chemistry in which a C3 hydrocarbon linker
is bonded to both the surface and the cinchona alkaloid using
triethoxysilane and either isocyanato or mercapto links,
respectively.
The resulting heterogeneous catalysts were found to show
activity comparable to that of free cinchonidine in solution, but,
in many cases, diminished enantioselectivity. At least three
factors were identified as responsible for this deterioration in
catalytic performance: (1) a competition in catalytic promotion
REFERENCES
(1) Zaera, F. J. Phys. Chem. Lett. 2010, 1, 621.
(2) Zaera, F. Catal. Lett. 2012, 5, 501.
■
(3) Notestein, J. M.; Katz, A. Chem.Eur. J. 2006, 12, 3954.
(4) Coperet, C.; Basset, J.-M. Adv. Synth. Catal. 2007, 349, 78.
́
(5) Corma, A.; Garcia, H. Adv. Synth. Catal. 2006, 348, 1391.
13064
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065

Journal of the American Chemical Society
Article
(6) Margelefsky, E. L.; Zeidan, R. K.; Davis, M. E. Chem. Soc. Rev.
2008, 37, 1118.
(7) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed.
2001, 40, 2004.
(8) Kim, G.-J.; Shin, J.-H. Tetrahedron Lett. 1999, 40, 6827.
(9) Choong, E. S. Annu. Rep. Prog. Chem., Sect. C: Phys. Chem. 2005,
101, 143.
(10) Hellriegel, C.; Skogsberg, U.; Albert, K.; Lammerhofer, M.;
̈
Maier, N. M.; Lindner, W. J. Am. Chem. Soc. 2004, 126, 3809.
(11) Kacprzak, K.; Gawronski, J., Resolution of Racemates and
Enantioselective Analytics by Cinchona Alkaloids and Their
Derivatives. In Cinchona Alkaloids in Synthesis and Catalysis: Ligands,
Immobilization and Organocatalysis; Song, C. E., Ed.; Wiley-VCH:
Weinheim, Germany, 2009.
(12) Kacprzak, K.; Grajewski, J.; Gawronski, J. Tetrahedron:
Asymmetry 2006, 17, 1332.
(13) Zhao, Y.; Liu, Y.; Lee, I.; Song, Y.; Qin, X.; Zaera, F.; Liao, J. J.
Biomed. Mater. Res., Part A 2012, 100 A, 103.
(14) Jo, S.; Park, K. Biomaterials 2000, 21, 605.
́
(15) Mansur, H. S.; Orefice, R. L.; Vasconcelos, W. L.; Lobato, Z. P.;
Machado, L. J. C. J. Mater. Sci.: Mater. Med. 2005, 16, 333.
(16) Hiemstra, H.; Wynberg, H. J. Am. Chem. Soc. 1981, 103, 417.
(17) Tian, S.-K.; Chen, Y.; Hang, J.; Tang, L.; McDaid, P.; Deng, L.
Acc. Chem. Res. 2004, 37, 621.
(18) Hong, J.; Lee, I.; Zaera, F. Top. Catal. 2011, 54, 1340.
(19) Lindholm, A.; Maki-Arvela, P.; Toukoniitty, E.; Pakkanen, T. A.;
Hirvi, J. T.; Salmi, T.; Murzin, D. Y.; Sjoholm, R.; Leino, R. J. Chem.
Soc., Perkin Trans. 1 2002, 2605.
(20) Kacprzak, K. M.; Maier, N. M.; Lindner, W. Tetrahedron Lett.
2006, 47, 8721.
(21) Azmat, M. U.; Guo, Y.; Guo, Y.; Wang, Y.; Lu, G. J. Mol. Catal.
A: Chem. 2011, 336, 42.
(22) Huang, Y.; Xu, S.; Lin, V. S. Y. ChemCatChem 2011, 3, 690.
(23) Degussa. Product Information: Aerosil 200 Hydrophilic Fumed
Silica; 2004.
(24) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.;
Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548.
(25) Deyhimi, F.; Coles, J. A. Helv. Chim. Acta 1982, 65, 1752.
(26) Mallat, T.; Orglmeister, E.; Baiker, A. Chem. Rev. 2007, 107,
4863.
(27) Zaera, F. J. Phys. Chem. C 2008, 112, 16196.
(28) Lai, J.; Ma, Z.; Mink, L.; Mueller, L. J.; Zaera, F. J. Phys. Chem. B
2009, 113, 11696.
(29) Caravajal, G. S.; Leyden, D. E.; Quinting, G. R.; Maciel, G. E.
Anal. Chem. 1988, 60, 1776.
́
(30) Krupczynska, K.; Buszewski, B.; Jandera, P. Anal. Chem. 2004,
76, 226 A.
(31) Raider, S. I.; Flitsch, R.; Palmer, M. J. J. Electrochem. Soc. 1975,
122, 413.
(32) Wagner, C. D.; Riggs, W. M.; Davis, L. E.; Moulder, J. F.;
Muilenberg, G. E. Handbook of X-Ray Photoelectron Spectroscopy;
Perkin-Elmer Corporation: Eden Prairie, MN, 1978.
(33) Bass, J. D.; Anderson, S. L.; Katz, A. Angew. Chem., Int. Ed. 2003,
42, 5219.
(34) Mink, L.; Ma, Z.; Olsen, R. A.; James, J. N.; Sholl, D. S.; Mueller,
L. J.; Zaera, F. Top. Catal. 2008, 48, 120.
(35) Ma, Z.; Zaera, F. Chiral modification of catalytic surfaces. In
Design of Heterogeneous Catalysis: New Approaches Based on Synthesis,
Characterization, and Modelling; Ozkan, U. S., Ed.; Wiley-VCH:
Weinheim, 2009; pp 113−140.
13065
dx.doi.org/10.1021/ja304181q | J. Am. Chem. Soc. 2012, 134, 13056−13065