Control of metal catalyst selectivity through specific noncovalent molecular interactions

Article abstract of DOI:10.1021/ja411973p

The specificity of chemical reactions conducted over solid catalysts can potentially be improved by utilizing noncovalent interactions to direct reactant binding geometry. Here we apply thiolate self-assembled monolayers (SAMs) with an appropriate structure to Pt/Al₂O₃ catalysts to selectively orient the reactant molecule cinnamaldehyde in a configuration associated with hydrogenation to the desired product cinnamyl alcohol. While nonspecific effects on the surface active site were shown to generally enhance selectivity, specific aromatic stacking interactions between the phenyl ring of cinnamaldehyde and phenylated SAMs allowed tuning of reaction selectivity without compromising the rate of desired product formation. Infrared spectroscopy showed that increased selectivity was a result of favorable orientation of the reactant on the catalyst surface. In contrast, hydrogenation of an unsaturated aldehyde without a phenyl ring showed a nontunable improvement in selectivity, indicating that thiol SAMs can improve reaction selectivity through a combination of nonspecific surface effects and ligand-specific near-surface effects.

Full text of DOI:10.1021/ja411973p

Article
pubs.acs.org/JACS
Control of Metal Catalyst Selectivity through Specific Noncovalent
Molecular Interactions
Karl R. Kahsar, Daniel K. Schwartz, and J. Will Medlin*
Department of Chemical and Biological Engineering, University of Colorado Boulder, Boulder, Colorado 80309-0596, United States
*
S Supporting Information
ABSTRACT: The specificity of chemical reactions conducted over solid catalysts can
potentially be improved by utilizing noncovalent interactions to direct reactant binding
geometry. Here we apply thiolate self-assembled monolayers (SAMs) with an
appropriate structure to Pt/Al O catalysts to selectively orient the reactant molecule
2
3
cinnamaldehyde in a configuration associated with hydrogenation to the desired
product cinnamyl alcohol. While nonspecific effects on the surface active site were
shown to generally enhance selectivity, specific aromatic stacking interactions between
the phenyl ring of cinnamaldehyde and phenylated SAMs allowed tuning of reaction
selectivity without compromising the rate of desired product formation. Infrared
spectroscopy showed that increased selectivity was a result of favorable orientation of
the reactant on the catalyst surface. In contrast, hydrogenation of an unsaturated aldehyde without a phenyl ring showed a
nontunable improvement in selectivity, indicating that thiol SAMs can improve reaction selectivity through a combination of
nonspecific surface effects and ligand-specific near-surface effects.
1. INTRODUCTION
Scheme 1. Hydrogenation Pathway of Cinnamaldehyde
Heterogeneous catalysts, such as supported transition metals, are
widely used in industrial applications due to their recoverability,
but they are generally far less selective toward forming desired1
products in comparison to organometallic or enzyme catalysts.
Enzyme catalysts have evolved an advantage for improving
selectivity: the active site is often contained within a binding
pocket, and noncovalent interactions between the reactant and
binding pocket cause reactant molecules to bind in a specific
2
orientation to the active site. Similar interactions have been
successfully exploited to produce branched chain alcohols as
3
biofuels, and alcohol dehydrogenase enzymes achieve selective
hydrogenation of α,β-unsaturated aldehydes through alignment
of the reactant molecules in a hydrophobic barrel within the
catalyst structure. Previous attempts to improve the chemo-
properties of the active catalyst surface nonspecifically through
the interaction of the sulfur headgroup with the surface, but
4
10
selectivity of heterogeneous catalysts have not generally focused
on controlling noncovalent interactions and instead have focused
on changing the size, shape, or composition of the surface
here we show that it is possible to control selectivity specifically
by precise placement of functional groups in the organic tail that
(de)stabilize particular geometries of adsorbed reacting species.
5
−8
6
layer.
While these methods represent powerful tools for
Interestingly, 15 years ago, Gallezot and Richard speculated that
improving selectivity through (for example) control of the
electronic properties of the active site, the engineering of
noncovalent interactions above the active surface may provide an
additional lever for approaching optimal selectivity.
such an approach might be possible, but it has not been
successfully demonstrated previously.
The formation of alkanethiol SAMs on gold surfaces has been
extensively characterized; such systems are deposited on a
surface from a dilute thiol solution through bulk transport
followed by surface adsorption and a final, slow, two-dimensional
organization step to yield a well-ordered monolayer bound to the
In this contribution, we deposit thiolate self-assembled
monolayers (SAMs) on conventional heterogeneous catalysts
to control the selectivity of an important model reaction, the
hydrogenation of cinnamaldehyde (Scheme 1). By changing the
functional groups of the hydrocarbon “tail” of the thiolate
modifier, it is possible to adjust noncovalent interactions in the
near-surface environment in a manner analogous to changing
1
1−13
active metal surface by a metal−sulfur bond.
Alkanethiol
SAMs exhibit similar well-ordered structures on palladium and
platinum surfaces. For example, the wetting behaviors of
9
functional groups in an enzyme binding pocket. Previous work
Received: November 27, 2013
Published: December 9, 2013
has shown that self-assembled monolayers can change the
©
2013 American Chemical Society
520
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526

Journal of the American Chemical Society
Article
2
8
alkanethiol SAMs on palladium are similar to those observed for
h. To minimize degradation in this study, all reactions were run for 1 h,
minimizing any degradation of the SAM surface in situ. Aging was
achieved by leaving a dry catalyst exposed to ambient laboratory
conditions for the desired time.
14
the same SAMs on gold, silver, and copper substrates. Similarly,
alkanethiols on platinum have been shown using XPS not to
undergo C−S bond scission or to contain weakly adsorbed
species and are oriented much closer to the surface normal than
2
.3. Reaction System. Reactions were run in a 100 mL Parr
semibatch reactor at 50 °C pressurized to 40 bar with hydrogen gas.
Reactor contents were prepared as 48 mL of ethanol solvent, 5 mL of
THF internal standard for the GC analysis, and 1 mL of reactant, giving
the system a reactant concentration of approximately 0.15 M. Solvent
choice has been shown not to significantly influence the selectivity of the
reaction, and light alcohols such as ethanol have been shown previously
15
those on palladium or gold. Because alkanethiol SAMs are
readily fabricated in a controlled, reproducible manner, they offer
an excellent way to control the near-surface environment.
Regioselectivity studies have been pursued previously using
organic monolayers to create favorable geometries for directing
6
16
to give the highest rates. For reactions of uncoated catalysts between 10
organic photoreactions, and recently a wide range of organic
and 50 mg of catalyst was used, and for coated catalysts, up to 300 mg of
the catalyst was used. All reactions were run for 60 min, during which
eight 1.5 mL liquid samples were taken. Hydrogen pressure was used to
push liquid samples out through the sample tube and through a
disposable filter, which was replaced after each sample. By sampling the
liquid with catalyst and filtering out the catalyst, we were able to keep the
ratio of catalyst to reactant constant within the reactor independent of
how much liquid was sampled. Liquid samples were taken in vials and
analyzed in an Agilent 5890A gas chromatograph with a flame ionization
detector. The column was an Agilent HP-5 capillary column with
dimensions of 30 m × 0.32 mm × 0.25 μm.
modifiers have been successfully used to nonspecifically tune the
1
0,17−21
surface of heterogeneous catalysts.
However, the ability
to rationally and precisely tune chemoselectivity through
noncovalent interactions within these monolayers has not been
previously achieved.
The hydrogenation of α,β-unsaturated aldehydes, such as
cinnamaldehyde, to their respective unsaturated alcohols is
among the most widely studied classes of model reactions for
5,6,22−26
understanding methods to promote chemoselectivity.
In
addition, the production of unsaturated alcohols is a less
favorable yet more industrially valuable pathway for the fragrance
2
.4. Selectivity and Rate Calculations. Reaction rates for both the
6
coated and uncoated catalysts were calculated as the moles of reactant
consumed per mole of surface metal per second. The calculations were
made using the dispersion of the uncoated Pt/Al O catalyst and the
and pharmaceutical industries. The selectivity of this reaction
has been shown to be highly dependent on the precious metal
used, and since platinum is among the most selective for
producing the desired cinnamyl alcohol product, it was chosen as
2
3
initial slope of the consumption of reactant per time at short time
intervals where the slope could be approximated as linear. This is the
same approach for calculating rates as in our previous SAM-coated
5
a starting point for this study.
1
0,29,30
catalysis work.
Further characterization is required to determine
which sites alkanethiol SAMs cover and which type of site is active in the
reaction system with or without the SAM coating. In order to avoid
confusion or present a misleading turnover frequency (TOF), we
instead report the rates as moles of reactant consumed per mole of
surface metal per time. Error bars in the reported rates are computed
from replicate measurements and error in the initial slope estimation.
Reaction selectivity was calculated as the conversion to a particular
product divided by the total conversion to all products. Therefore, the
selectivity always totaled 100%. Error bars in selectivity were calculated
from repeat measurements.
2.5. Thin Film PM-RAIRS. Thin films of platinum were prepared by
electron beam evaporation onto soda lime glass slides. The slides were
cleaned with a nanostrip solution (sulfuric acid/peroxide/buffer) for 10
min at 95 °C, rinsed in DI water for 10 min, and then cleaned in a UV
ozone plasma machine at 150 W for 3 min, rinsed in DI water, and blown
dry. An adhesion layer of 30 nm of Ti followed by 150 nm of Pt was
deposited onto the cleaned slides. Thiol coatings were deposited to the
platinum thin films at the same concentration used to prepare thiol-
coated catalysts. When they were removed from solution, platinum thin
films were blown dry with nitrogen rather than under a vacuum
desiccator, as had been done for the catalysts.
2
. EXPERIMENTAL METHODS
2
.1. Materials. A 5 wt % Pt/Al O catalyst was purchased from
2 3
Sigma-Aldrich. 1-Propanethiol (C SH), 1-hexanethiol (C SH), 1-
dodecanethiol (C SH), 1-octadecanethiol (C SH), thiophenol,
3
6
1
2
18
thiophene, benzyl mercaptan, and 2-phenylethanethiol were all obtained
from Sigma-Aldrich. 3-Phenyl-1-propanethiol (>95%) was obtained
from MolPort, and 4-phenyl-1-butanethiol (>95%) was obtained from
Ukrorgsyntez, Ltd. Ethanol (>99.5%) used as a solvent during SAM
deposition and in the reactor was from Sigma-Aldrich. Tetrahydrofuran
(
>99.5%) internal standard and the reactants cinnamaldehyde and
prenal were also obtained from Sigma-Aldrich. All gases (H , O , and
He) used for catalyst preparation and reaction were Airgas ultrahigh
purity.
2
2
2.2. Surface Area and Preparation of Catalysts. The commercial
uncoated 5 wt % Pt/Al O was characterized by chemisorption of
2
3
hydrogen on a Quantachrome Autosorb-1 to determine the active
surface area. The catalyst was reduced in situ at 473 K for 16 h, after
which chemisorption with CO indicated an active surface area of 2.87
m /g and a dispersion of 26.6%. The particle size was measured by TEM
as 3.9 ± 1.1 nm.
2
SAM-coated catalysts were prepared by immersing the catalyst in an
ethanolic solution of 10 mM for thiols that were liquid at room
temperature and 1 mM for thiols that were solid at room temperature.
After 12−16 h, the thiol solution was poured off and the catalyst was
rinsed for 3 h in ethanol to remove any physisorbed thiols. Finally, this
ethanol supernatant was poured off and the catalyst was dried under
vacuum for 20 min in a vacuum desiccator. All thiol-coated catalysts
were used the same day they were prepared unless specified otherwise.
Phenylated SAM coatings used for the hydrogenation of
cinnamaldehyde were susceptible to subtle changes of the deposition
procedure, including aging in air. Aging of the thiol coating in air
PM-RAIRS (polarization modulation−reflection absorption infrared
spectroscopy) data were taken using a Thermo Scientific Nicolet 6700
spectrometer with a Thermo Scientific TOM (tabletop optical module)
PM-RAIRS attachment. The PEM (photoelastic modulator) was made
by Hinds Instruments and the synchronous sampling demodulator by
GWC Technologies. The signal was received by a Thermo Scientific
MCT-A (mercury cadmium telluride) detector cooled with liquid
nitrogen. In order to hold the cinnamaldehyde or prenal liquid in place
to take the IR scan, an Edmund Optics (1 mm × 50 mm) zinc selenide
coverslip was used. Zinc selenide is IR inactive and therefore served as a
way to maintain a uniform thin film of cinnamaldehyde or prenal while
not disrupting the infrared signal on the surface. Some incident light was
reflected from the surface of the zinc selenide coverslip before reaching
the platinum surface, but this light was not focused at the detector and so
could not interfere with the measured signal. The coverslip had an
antireflective coating rated from 3 to 12 μm to allow as much light to
pass through as possible. The coverslip was clipped in place over top of
the liquid sample so as not to interfere with the signal. Error bars of the
(ambient laboratory conditions) significantly altered the selectivity of
the 3-phenyl-1-propanethiol coating, reducing the selectivity from 90%
to near 15% for times longer than 1 week; this is hypothesized to be due
to degradation of the SAM layer. Numerous effects can contribute to the
degradation of an alkanethiol SAM, including ozone oxidation in air
27
enhanced by time and light. In addition, SAM-coated surfaces degrade
under solvated conditions, though this degradation is minor up to 25
5
21
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526

Journal of the American Chemical Society
Article
η1/η2 ratio were calculated from repeat measurements on freshly
prepared slides.
3
. RESULTS AND DISCUSSION
.1. Noncovalent Control of Selectivity. Over an
uncoated Pt/Al O catalyst (Figure 1a), cinnamyl alcohol was
3
2
3
Figure 2. Selectivity to cinnamyl alcohol for the hydrogenation of
cinnamaldehyde over Pt/Al O catalysts. All reactions were carried out
2
3
at 40 bar of hydrogen pressure and 0.15 M initial cinnamaldehyde
concentration in ethanol solvent. Alkanethiol SAMs nonspecifically
increased the selectivity to cinnamyl alcohol (C SH, C SH, and C SH
3
6
18
shown in Figure S5 (Supporting Information)) over the uncoated case,
but phenylated SAMs showed the ability to dramatically increase or
decrease the selectivity depending on the location of the phenyl ring
with respect to the surface.
Scheme 2. Molecules Used in the Hydrogenation System: (a)
Thiol SAMs Used To Coat the Pt/Al O Surface; (b)
2
3
Cinnamaldehyde and Prenal Reactants
Figure 1. Kinetic plots for the hydrogenation of cinnamaldehyde: (a)
hydrogenation over an uncoated 5 wt % Pt/Al O3 catalyst; (b)
2
hydrogenation over a 3-phenyl-1-propanethiol coated 5 wt % Pt/Al O3
2
catalyst.
produced at levels much lower than those of hydrocinnamalde-
hyde at all times and conversions. In fact, the selectivity to
cinnamyl alcohol was constant at 25% (Figure 2) over a broad
range of conversions (20−90%), in agreement with previous
5
,6
reports.
To tune noncovalent interactions in the near-surface environ-
ment, SAMs comprised of a number of organic ligands (Scheme
the potential for additional noncovalent aromatic stacking
interactions. Whereas linear alkyl ligands are not expected to
interact preferentially with a particular region of the
cinnamaldehyde reactant, phenylated ligands can interact with
cinnamaldehyde’s phenyl group through aromatic π−π stack-
ing. For example, cinnamaldehyde exhibits much smaller
contact angles with phenyl-terminated Pt surfaces than with
alkyl-terminated surfaces (Figure S1, Supporting Information).
By changing the vertical position of the modifying phenyl group
within the SAM layer, it is hypothetically possible to control the
2
) were employed. Prenal, a branched nonaromatic α,β-
unsaturated aldehyde, was used as a reactant for control
experiments. Because all α,β-unsaturated aldehydes contain the
same reactive double bond and aldehyde at their terminal end,
they can be expected to respond similarly to changes in the
electronic properties of the catalyst surface; however, cinna-
maldehyde contains a phenyl group, resulting in additional
31
6
spatial constraints within the crowded surface region as well as
5
22
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526

Journal of the American Chemical Society
Article
orientation of the cinnamaldehyde with respect to the active
surface via these π−π stacking interactions. Prenal, which lacks an
aromatic moiety, exhibits no such specific interaction with
phenylated SAMs.
(Supporting Information), the selectivity was improved for thiol
-coated catalysts in comparison to the uncoated case, but the
increase in selectivity did not depend on the functionality of the
coating. This loss of ligand-specific selectivity control suggests
that, for small particles with high curvature, the ligand-specific
control is difficult to maintain.
Using a catalyst coated with a 3-phenyl-1-propanethiol SAM
(Figures 1b and 2) increased the hydrogenation selectivity of
cinnamaldehyde to greater than 90%, indicating a highly
favorable specific interaction between the properly spaced
phenyl ring from the SAM and the phenyl ring of the
cinnamaldehyde. 2-Phenylethanethiol and benzomercaptan
modifiers improved selectivity to a lesser extent, while catalyst
modification with thiophene and thiophenol decreased selectiv-
ity. In other words, the highest selectivity was associated with a
three-methylene spacer between the S atom and the phenyl ring,
a structure that approximately matches that of cinnamaldehyde.
Over long time intervals, at 100% conversion, the series reaction
to produce 3-phenyl-1-propanol occurred; however, high
selectivity was observed even at conversions of 90%.
Alkanethiol SAMs formed from propanethiol, hexanethiol,
dodecanethiol, and octadecanethiol were all shown to increase
the selectivity of the reaction (Figure 2 and Figure S5
(Supporting Information)) to up to 60% selectivity. These
selectivity values were improved over those of the uncoated case,
but they were not specific to the length of the thiol tail, similar to
the nonspecific effect typically seen from sulfur poisoning. This
result suggests that a nonspecific selectivity improvement may be
realized through modification by inorganic sulfur sources, as has
been observed previously in the hydrogenation of 1-epoxy-3-
10
butene on alkanethiol- and H S-modified Pd catalysts. Studies
2
of H S-modified Pt (reported in the Supporting Information) for
2
A longer 4-phenylbutanethiol coating showed an increase in
selectivity similar to that of the 2-phenylethanethiol coating, but
not as good as the 3-phenylpropanethiol coating, indicating a
peak in selectivity at the proper spacing length of the
cinnamaldehyde molecule. The longer spacer also decreased
the rate of reaction by a factor of 2 compared to the shorter
phenylated SAM cases. An even larger rate decrease was seen for
the hydrogenation of benzaldehyde (C H CHO) on uncoated
and 3-phenyl-1-propanethiol-coated catalysts (Table S1, Sup-
porting Information). That is, in cases where the spacer length in
the modifier is greater than the distance between the phenyl ring
and the carbonyl function in the reactant, the reaction rate is
suppressed. This result suggests that if aromatic stacking
interactions occur too far above the surface, the carbonyl
function is hindered from reaching the surface.
cinnamaldehyde hydrogenation showed a decrease in rate with
no improvement in selectivity. However, we cannot rule out the
possibility that a different form of exposure to H S or another
2
inorganic sulfur source can improve selectivity; as reported
elsewhere, sulfur deposition for selectivity modification is
2
6,33−36
difficult to control.
Whereas the phenylated SAMs exerted specific control on the
basis of their tail lengths, the alkanethiol modifiers nonspecifi-
cally enhanced the reaction selectivity. Previous studies have
shown that unsaturated alcohol selectivity is increased by
weakening the binding strength of the desired product to the
6
5
3
7,38
catalyst surface,
and the same adsorption weakening effect is
hypothesized to be responsible for the nonspecific increase in
selectivity over alkanethiol-coated catalysts.
3.2. Hydrogenation of Prenal. To test this hypothesis, we
also investigated the hydrogenation of a nonaromatic α,β-
unsaturated aldehyde, prenal, which contains the same reactive
groups as cinnamaldehyde without a phenyl moiety. Whereas
cinnamaldehyde showed specific control of selectivity via
phenylated thiols, prenal hydrogenation selectivity was insensi-
tive to the presence of a phenyl group in the SAM. Prenal
hydrogenation selectivity (Figure 3) increased similarly for each
of the alkanethiol-coated catalysts as well as the 2-phenyl-
ethanethiol- and the 3-phenyl-1-propanethiol-coated platinum
catalysts. This suggested a lack of specific interaction effects
between the thiol coating and the reactant; the general increase
in selectivity was consistent with previous hydrogenation studies
of small molecules over thiol-coated catalysts due to changes in
The molecular order of the phenylated SAMs for these
hydrogenation experiments was critical to their efficacy. Well-
ordered SAMs contain few surface vacancies and exhibit a
standing-up adsorption geometry, while poorly ordered SAMs
adopt a lying-down structure and are prone to C−S bond scission
2
9,32
due to greater tail disorder on the catalytic surface.
For the
hydrogenation of cinnamaldehyde, the quality of the SAM was
shown to have a dramatic impact on the selectivity of the
reaction. A 3-phenyl-1-propanethiol SAM that had been aged for
1−3 weeks in air decreased the selectivity to cinnamyl alcohol to
lower than that for the uncoated case (Figure 2), similar to the
effect caused by shorter phenylated SAMs, consistent with a
strong dependence on the position of the SAM phenyl ring
relative to the catalyst surface. The 2-phenylethanethiol SAM was
even more sensitive to aging, showing a decrease in selectivity
after less than 1 day. Degradation of SAMs within the reaction
solution was also considered, as shown in Figures S2 and S3
10
the electronic properties of the catalyst surface. These results
indicate that, although nonspecific electronic effects can increase
the selectivity of the reaction, the position of the phenyl ring from
the surface provides the extra functional handle with which to
direct selectivity still higher or lower from there.
(Supporting Information). The results show that there is some
selectivity loss when the catalyst is recycled without regeneration,
but this can be avoided by maintaining a small concentration of
thiols in solution. A detailed investigation of the effects of various
recycling procedures is in progress.
The low selectivity observed for catalysts modified with
thiophene and thiophenol was likely due to the close proximity of
the aromatic functional group to the surface. For example,
thiophene is known to undergo extensive C−S bond-breaking
39
Metal particle size has been shown previously to significantly
affect the selectivity of cinnamaldehyde hydrogenation, where
large particles, with larger flat facets, favor selective hydro-
genation and small particles with large curvature have low
reactions on noble metal surfaces, consistent with a greater
deposition of surface sulfur observed in this study (Figure S6
(Supporting Information)).
3.3. PM-RAIRS Study of Reactant Binding Geometry.
We hypothesize that the direction of selectivity change (i.e.,
increase or decrease) for the hydrogenation of cinnamaldehyde is
controlled in part by the orientation of adsorbed cinnamaldehyde
in relation to the surface (Scheme 3) via the interaction of its
6
selectivity to cinnamyl alcohol. The 5 wt % Pt/Al O catalyst
2
3
used here had a measured average particle size of 3.9 ± 1.1 nm,
but we also studied a 0.5 wt % Pt/Al O catalyst particle size of
2
3
0.8 ± 0.1 nm. For these smaller particles, as shown in Figure S4
5
23
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526

Journal of the American Chemical Society
Article
catalysts. Each spectrum was normalized to its background and
−1
adjusted to a zero baseline between 2500 and 2600 cm .
There are distinct differences between the spectra collected
after adsorption of cinnamaldehyde on each of the four surfaces
−1
(
Figure 4a). Peaks in the frequency range 1500−1750 cm
correspond to the vibrational modes of CC double bonds as
well as CO aldehyde stretches. The large qualitative
differences in peak structure in this region suggested that the
various coatings had a strong effect on the cinnamaldehyde
adsorption geometry. In contrast, the general structure of the
prenal spectra was insensitive to the various coatings (Figure 4b).
Previous studies have used a combination of experimental
vibrational spectroscopy and density functional theory (DFT) to
identify the stretching modes in this region associated with
binding configurations of prenal and crotonaldehyde (2-butenal)
2
4,37
on platinum and tin-doped platinum surfaces.
The DFT
studies identified numerous binding configurations, two of which
are especially prominent in high-coverage conditions on Pt: the
1
η configuration (associated with CO hydrogenation) where
the molecule is bound through the carbonyl oxygen lone pair
electrons in an upright geometry similar to that in Scheme 3a and
Figure 3. Selectivity of cinnamaldehyde and prenal hydrogenation to its
2
the η configuration (associated with CC hydrogenation)
respective unsaturated alcohol. The alkanethiols C SH, C SH, C SH,
3
6
12
where the molecule is bound in a di-σ configuration through its
CC
and C SH were averaged for one data point. Reactions were run at 40
18
CC double bond parallel to the surface similar to the position
bar of H and 50 °C and 0.15 M reactant solvated in ethanol. The
2
24,41
shown in Scheme 3b.
Assignments from the prior studies
selectivity was reported at 50% conversion.
(
Table S2 (Supporting Information)) were used together with
spectra of related molecules such as hydrocinnamaldehyde
Figure S7 (Supporting Information)) to assign peaks for the
adsorption of cinnamaldehyde and prenal on platinum thin films.
Scheme 3. SAM-Enhanced Orientation of Cinnamaldehyde
with the Catalyst Surface: (a) Cinnamaldehyde Interacting
with an Uncoated Platinum Surface through the CC Double
Bond; (b) 3-Phenyl-1-propanethiol SAMs Favoring Aldehyde
Hydrogenation by Creating an Upright Molecular
Orientation
(
−1
The peaks at 1683 and 1687 cm corresponded to the aldehyde
stretches of an η binding configuration for cinnamaldehyde and
prenal, respectively, and the peaks at 1575 and 1545 cm
corresponded to the aldehyde stretches of the η binding
configuration, respectively. The spectra in Figure 4a show a
relative increase in the prominence of the η binding
configuration in comparison to the η binding configuration for
2
−1
1
2
1
thiophenol SAMs, which exhibit low selectivity to the
unsaturated alcohol. Conversely, the highly selective 3-phenyl-
1
1
-propanethiol SAM shows a stronger signal from the η binding
1
configuration. Since the η binding configuration is associated
2
with CO hydrogenation while the η binding configuration is
associated with CC hydrogenation, the effect observed for
cinnamaldehyde orientation on these surfaces is consistent with
the selectivity data for the hydrogenation of cinnamaldehyde.
This effect was quantified in Figure 4c, where peaks identified in
Figure 4a,b were integrated (Figures S8 and S9 (Supporting
phenyl moiety with the phenyl ring of the SAM. For the 3-
phenyl-1-propanethiol SAM, the distance between the phenyl
ring and surface is such that cinnamaldehyde is directed to a
standing-up orientation in which only the aldehyde group
interacts with the surface.
1
2
Information)) and a ratio of the η /η peak area for
cinnamaldehyde and for prenal was each compared to its
reaction selectivity for each SAM coating. Cinnamaldehyde
reaction selectivity was shown to trend directly with this ratio,
while the adsorbed states of prenal were uncorrelated with the
ratio of peak intensities. Consistent with previous studies, this
suggests a different mode of selectivity enhancement for prenal
which is nonspecific to the organic function of the SAM and has
been previously attributed to a weakened adsorption state of
It has been shown previously that binding in a horizontal
configuration favors CC hydrogenation, while binding in a
40
vertical orientation favors CO hydrogenation. In order to
probe binding geometries, polarization modulation reflection−
absorption infrared spectroscopy (PM-RAIRS) was used to
examine how the adsorption of cinnamaldehyde varies in the
presence of different thiol coatings. The advantage of this
technique is the ability to isolate the spectra of adsorbed
molecules in the presence of an isotropic solution phase. Because
PM-RAIRS requires an optically reflective surface, the study was
performed on a thin film of platinum (150 nm) deposited on a
silica glass wafer. Thiol SAM-coated surfaces were prepared in
ethanolic solutions identical with the solutions used to prepare
24
prenal on the catalyst surface.
As the electronic properties of a catalyst are altered, the
selectivity typically improves at the expense of reactivity,
consistent with the weakening of reactant adsorption to a
2
4,26
catalyst surface.
Generally the catalysts investigated here
showed a tradeoff between activity and selectivity, as would be
expected for a modified catalytic system, but the 3-phenyl-1-
propanethiol SAM did not compromise activity (Figure 5) due to
5
24
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526

Journal of the American Chemical Society
Article
Figure 5. Rate versus selectivity to unsaturated alcohol for both
uncoated and SAM-coated Pt/Al O catalysts. Rates are shown as the
2
3
moles of reactant consumed per moles of surface platinum per second.
Alkanethiols C SH, C SH, C SH, and C SH were averaged for one
3
6
12
18
data point.
production was indistinguishable, within experimental error, on
the uncoated and 3-phenyl-1-propanethiol-coated catalysts
(
Table S1 (Supporting Information)). As also shown in Table
S1, the effect of a phenylated SAM with a longer alkyl spacer is
hypothesized to further decrease the rate of reaction.
4
. CONCLUSIONS
Here we demonstrate that thiol modifiers can improve the
selectivity in hydrogenation of α,β-unsaturated aldehydes by a
combination of electronic effects and specific noncovalent
interactions of the reactant in the near-surface environment.
The selective hydrogenation of both prenal and cinnamaldehyde
to the desired unsaturated alcohols was enhanced with a thiol-
coated catalyst surface, but the selective hydrogenation of
cinnamaldehyde was further controlled through interactions of
its phenyl ring with aromatic ligands within the SAM-coating
layer. The selective hydrogenation of α,β-unsaturated aldehydes
6
,22
is commonly achieved through active site modification,
but
here we demonstrate that ligand-specific control exhibited by
phenylated SAMs can create a reaction environment that
functions in analogy to biological catalysts. This ability to
exercise control over the selective hydrogenation of cinnamalde-
hyde by ligand-specific interactions provides a promising new
method for controlling a reactive system beyond modifying the
active site of heterogeneous catalysts.
Figure 4. PM-RAIRS analysis of reactant adsorption geometry. (a)
Cinnamaldehyde and (b) prenal adsorbed on thiophenol-coated
platinum, uncoated platinum, octadecanethiol-coated platinum, and 3-
phenyl-1-propanethiol-coated platinum thin film surfaces. (c) Decon-
volution of the spectra to obtain the ratio of η to η CO peaks shown
with selectivity data for cinnamaldehyde and prenal hydrogenation.
Error bars were calculated from replicate measurements.
ASSOCIATED CONTENT
Supporting Information
■
*
S
1
2
Text, tables, and figures giving contact angle measurements, ICP-
AES analysis, peak assignments and IR analysis, and recycle
studies. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
Will.medlin@colorado.edu
its different mode of selectivity improvement with cinnamalde-
hyde. Although the rate of cinnamaldehyde consumption
decreased by a factor of 3, the rate of cinnamyl alcohol
■
5
25
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526

Journal of the American Chemical Society
Article
Notes
(32) Floridia Addato, M. a. A.; Rubert, A. A.; Benítez, G. A.; Fonticelli,
M. H.; Carrasco, J.; Carro, P.; Salvarezza, R. C. J. Phys. Chem. C 2011,
The authors declare no competing financial interest.
1
(
15, 17788.
33) Baldyga, L. M.; Blavo, S. O.; Kuo, C.-H.; Tsung, C.-K.; Kuhn, J. N.
ACS Catal. 2012, 2, 2626.
34) Hutchings, G. J.; King, F.; Okoye, I. P.; Rochester, C. H. Appl.
Catal., A 1992, 83, L7.
ACKNOWLEDGMENTS
■
(
This research was supported by the Department of Energy,
Office of Science, Basic Energy Sciences Program, Chemical
Sciences, Geosciences, and Biosciences Division, under Grant
No. DE-FG02-10ER16206. We also thank S. Johnson and C.
Hartman for their work on the liquid reactor.
(
(
(
35) Oudar, J. Catal. Rev. 1980, 22, 171.
36) Marinelli, T. B. L. W.; Ponec, V. J. Catal. 1995, 156, 51.
37) Haubrich, J.; Loffreda, D.; Delbecq, F.; Sautet, P.; Jugnet, Y.;
Krupski, A.; Becker, C.; Wandelt, K. J. Phys. Chem. C 2008, 112, 3701.
(
(
(
(
38) Delbecq, F.; Sautet, P. J. Catal. 2002, 211, 398.
39) Lang, J. F.; Masel, R. I. Surf. Sci. 1987, 183, 44.
40) Medlin, J. W. ACS Catal. 2011, 1, 1284.
REFERENCES
■
(
1) de Jong, K. P. Synthesis of Solid Catalysts; Wiley-VCH: Darmstadt,
Germany, 2009.
2) Siegel, J. B.; Zanghellini, A.; Lovick, H. M.; Kiss, G.; Lambert, A. R.;
41) Neurock, M.; Pallassana, V.; van Santen, R. A. J. Am. Chem. Soc.
2
000, 122, 1150.
(
St. Clair, J. L.; Gallaher, J. L.; Hilvert, D.; Gelb, M. H.; Stoddard, B. L.;
Houk, K. N.; Michael, F. E.; Baker, D. Science 2010, 329, 309.
(
3) Atsumi, S.; Hanai, T.; Liao, J. C. Nature 2008, 451, 86.
(
4) Fersht, A. Structure and mechanism in protein sciences; W. H.
Freeman: New York, 1999.
5) da Silva, A. B.; Jordao, E.; Mendes, M. J.; Fouilloux, P. Appl. Catal.,
A 1997, 148, 253.
(
̃
(
(
(
6) Gallezot, P.; Richard, D. Catal. Rev. 1998, 40, 81.
7) Marimuthu, A.; Zhang, J.; Linic, S. Science 2013, 339, 1590.
8) Lee, I.; Delbecq, F.; Morales, R.; Albiter, M. A.; Zaera, F. Nat.
Mater. 2009, 8, 132.
9) Pecsi, I.; Leveles, I.; Harmat, V.; Vertessy, B. G.; Toth, J. Nucleic
Acids Res. 2010, 38, 7179.
10) Marshall, S. T.; O’Brien, M.; Oetter, B.; Corpuz, A.; Richards, R.
M.; Schwartz, D. K.; Medlin, J. W. Nat. Mater. 2010, 9, 853.
11) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides,
G. M. Chem. Rev. 2005, 105, 1103.
(
(
(
(
(
(
12) Ulman, A. Chem. Rev. 1996, 96, 1533.
13) Schwartz, D. K. Annu. Rev. Phys. Chem. 2001, 52, 107.
14) Love, J. C.; Wolfe, D. B.; Haasch, R.; Chabinyc, M. L.; Paul, K. E.;
Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 2003, 125, 2597.
15) Petrovykh, D. Y.; Kimura-Suda, H.; Opdahl, A.; Richter, L. J.;
Tarlov, M. J.; Whitman, L. J. Langmuir 2006, 22, 2578.
16) Kim, M.; Hohman, J. N.; Cao, Y.; Houk, K. N.; Ma, H.; Jen, A. K.-
Y.; Weiss, P. S. Science 2011, 331, 1312.
17) Wu, B.; Huang, H.; Yang, J.; Zheng, N.; Fu, G. Angew. Chem., Int.
Ed. 2012, 51, 3440.
18) Kwon, S. G.; Krylova, G.; Sumer, A.; Schwartz, M. M.; Bunel, E.
(
(
(
(
E.; Marshall, C. L.; Chattopadhyay, S.; Lee, B.; Jellinek, J.; Shevchenko,
E. V. Nano Lett. 2012, 12, 5382.
(
19) Taguchi, T.; Isozaki, K.; Miki, K. Adv. Mater. 2012, 24, 6462.
(
20) Snelders, D. J. M.; Yan, N.; Gan, W.; Laurenczy, G.; Dyson, P. J.
ACS Catal. 2011, 2, 201.
21) Makosch, M.; Lin, W.-I.; Bumbal
Hungerbuhler, K.; van Bokhoven, J. A. ACS Catal. 2012, 2, 2079.
22) Delbecq, F.; Sautet, P. J. Catal. 2003, 220, 115.
23) Pradier, C. M.; Birchem, T.; Berthier, Y.; Cordier, G. Catal. Lett.
(
́ ́
ek, V.; Sa, J.; Medlin, J. W.;
̈
(
(
1
(
994, 29, 371.
24) Haubrich, J.; Loffreda, D.; Delbecq, F.; Sautet, P.; Jugnet, Y.;
Becker, C.; Wandelt, K. J. Phys. Chem. C 2009, 114, 1073.
(
(
25) de Jesus
́
, J. C.; Zaera, F. Surf. Sci. 1999, 430, 99.
26) Hutchings, G. J.; King, F.; Okoye, I. P.; Padley, M. B.; Rochester,
C. H. J. Catal. 1994, 148, 453.
27) Schoenfisch, M. H.; Pemberton, J. E. J. Am. Chem. Soc. 1998, 120,
502.
(
4
(
(
28) Schlenoff, J. B.; Li, M.; Ly, H. J. Am. Chem. Soc. 1995, 117, 12528.
29) Kahsar, K. R.; Schwartz, D. K.; Medlin, J. W. Appl. Catal., A 2012,
4
45−446, 102.
30) Schoenbaum, C. A.; Schwartz, D. K.; Medlin, J. W. J. Catal. 2013,
03, 92.
31) Dou, R.-F.; Ma, X.-C.; Xi, L.; Yip, H. L.; Wong, K. Y.; Lau, W. M.;
Jia, J.-F.; Xue, Q.-K.; Yang, W.-S.; Ma, H.; Jen, A. K. Y. Langmuir 2006,
2, 3049.
(
3
(
2
5
26
dx.doi.org/10.1021/ja411973p | J. Am. Chem. Soc. 2014, 136, 520−526