Use of short time-on-stream attenuated total internal reflection infrared spectroscopy to probe changes in adsorption geometry for determination of selectivity in the hydrogenation of citral

Article abstract of DOI:10.1021/cs500185n

A new experimental procedure based on attenuated total reflection infrared spectroscopy has been developed to investigate surface species under liquid phase reaction conditions. The technique has been tested by investigating the enhanced selectivity in the hydrogenation of α,β-unsaturated aldehyde citral over a 5% Pt/SiO₂ catalyst toward unsaturated alcohols geraniol/nerol, which occurs when citronellal is added to the reaction. The change in selectivity is proposed to be the result of a change in the citral adsorption mode in the presence of citronellal. Short time on stream attenuated total internal reflection infrared spectroscopy has allowed identification of the adsorption modes of citral. With ssno citronellal, citral adsorbs through both the C = C and C = O groups; however, in the presence of citronellal, citral adsorption occurs through the C = O group only, which is proposed to be the cause of the altered reaction selectivity.

Full text of DOI:10.1021/cs500185n

Research Article
pubs.acs.org/acscatalysis
Use of Short Time-on-Stream Attenuated Total Internal Reflection
Infrared Spectroscopy To Probe Changes in Adsorption Geometry for
Determination of Selectivity in the Hydrogenation of Citral
H. Daly,† H. G. Manyar,† R. Morgan,† J. M. Thompson,† J.-J. Delgado,‡ R. Burch,† and C. Hardacre*,†
†CenTACat, School of Chemistry and Chemical Engineering, Queen’s University Belfast, BT9 5AG, United Kingdom
‡Departamento de Ciencia de los Materiales e Ingeniería Mmetallurgica y Química Inorgan
Cadiz, Spain
́ ́
ica, Universidad de Cadiz, Puerto Real,
́
S
* Supporting Information
ABSTRACT: A new experimental procedure based on attenuated total reflection
infrared spectroscopy has been developed to investigate surface species under liquid
phase reaction conditions. The technique has been tested by investigating the
enhanced selectivity in the hydrogenation of α,β-unsaturated aldehyde citral over a 5%
Pt/SiO2 catalyst toward unsaturated alcohols geraniol/nerol, which occurs when
citronellal is added to the reaction. The change in selectivity is proposed to be the
result of a change in the citral adsorption mode in the presence of citronellal. Short
time on stream attenuated total internal reflection infrared spectroscopy has allowed
identification of the adsorption modes of citral. With no citronellal, citral adsorbs
through both the CC and CO groups; however, in the presence of citronellal,
citral adsorption occurs through the CO group only, which is proposed to be the
cause of the altered reaction selectivity.
KEYWORDS: Pt, ATR, infrared, citral, hydrogenation
The attainment of higher selectivity upon addition of an
organic compound into the liquid phase reaction has also been
reported for the hydrogenation of citral. The addition of
thiophene to Ru/KL catalysts gave a maximum selectivity to
geraniol and nerol of 46% (at 15% citral conversion with 3 ppm
thiophene); in the absence of the thiophene, the Ru catalysts
typically favor CC hydrogenation.5 Therein, it was also
reported that the adsorption mode of the thiophene (relating to
thiophene concentration/surface coverage) altered whether the
citral adsorbed through the CC or the CO bond. Addition
of ethylenediamine during the hydrogenation of citral over a
Ru/AlO(OH) catalyst (in water) resulting in 87% selectivity to
geraniol/nerol (93% conversion), whereas selectivity of 74.3%
(at 60.5% conversion) was obtained without the additive.6
Similar effects have been observed with Pt-based catalysts with
both amines7 and thiols,8 resulting in enhancements to the
unsaturated alcohol selectivity.
INTRODUCTION
■
In chemoselective hydrogenations, such as the hydrogenation of
α,β-unsaturated aldehydes, there is the desire to understand and
control the reaction pathway so as to obtain a specific product
with high selectivity. The hydrogenation of citral is of particular
interest because of the complex reaction network resulting from
the preferential adsorption geometry and strength of interaction
of different functional groups within the molecule, namely the
CO, conjugated CC, and isolated CC bonds the
hydrogenation and further reaction of which can form a wide
range of products illustrated in Scheme 1. The hydrogenation of
citral has been studied extensively with the choice of metal,
support, reaction conditions, and solvent all reported to affect
the selectivity.1 With unsaturated alcohols being important fine
chemical and pharmaceutical intermediates, there is significant
interest in improving the selectivity to these desired products.
Modification of the catalyst through addition of a second metal,
such as Sn, Fe or Ge, has been reported to alter reaction
selectivity toward CO hydrogenation.2 For other chemo-
selective hydrogenations, modification of the catalyst surface via
the adsorption of organic compounds has also been reported,
with n-alkanethiol self-assembled-monolayer-coated Pd catalysts
and thiol-modified Pt/TiO2 catalysts giving excellent selectivity
enhancements in hydrogenation of unsaturated epoxides to
saturated epoxides3 and a complete switch in the products
formed for 4-nitrostyrene hydrogenation following modifica-
tion, respectively.4
In this study, we report the hydrogenation of citral with a
selectivity of 97% to geraniol and nerol at 100% conversion over
a Pt/SiO2 catalyst when citronellal (a minor reaction product) is
added to the reaction; the adsorption of citronellal influences
the adsorption of citral, favoring CO hydrogenation. It is
evident that the adsorption geometry of the reagent is critical in
Received: February 12, 2014
Revised: June 13, 2014
© XXXX American Chemical Society
2470
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
Scheme 1. Reaction Scheme for the Hydrogenation of Citral
determining the reaction selectivity, and in this study,
attenuated total internal reflection infrared (ATR-IR) spectros-
copy is used to probe the mode of citral adsorption and
modification of this adsorption mode in the presence of
citronellal.
Time-resolved ATR-IR spectra have been reported in the
study of pure ethanol (and ethanol/2-propanol) oxidation over
Pd/Al2O3.18 On switching from H2- to O2-saturated ethanol,
transient species such as ethyl acetate are identified in the liquid
phase together with the primary oxidation product acetalde-
hyde, which continues to form under an O2 environment.
Adsorption of acetate on the support and CO on the Pd are also
observed, which shows that oxidation (to acetic acid) and
decarbonylation of the acetaldehyde also occur. Ethyl acetate
formation via acetic acid is also proposed from the response of
dissolved ethyl acetate and acetate on the support directly
following the switch to O2-saturated ethanol, showing the need
for rapid monitoring of reactions. This technique has also been
applied to the aqueous phase reforming of glycerol over Pt/
Al2O3.19 By switching between feeds containing water, water +
glycerol, water + O2, and water + H2, the rate of formation of
CO from the reaction of glycerol on various pretreated surfaces
can be examined. In addition, from a comparison of the linear
and bridge-bound CO formed, on switching from the glycerol
feed to pure water, the bridge-bound species was found to react
via the water gas shift reaction, whereas little change with time
was observed for the linear species.
Recently, the application of a short time-on-stream (STOS)
technique coupled with diffuse reflectance infrared fourier
transform (DRIFT) spectroscopy has been shown to be
effective in the gas phase for hydrocarbon selective catalytic
reduction of NO over Ag catalysts.20 Therein, it was possible, by
limiting the exposure time of the catalyst to the reaction feed, to
differentiate the species adsorbed on the metal and the support.
In this case, the observation in the spectroscopy of accumulated
spillover spectator species from the metal onto the support was
decreased by examining the surface species only during the first
few seconds following initiation of the reaction. In this paper,
we have applied the short time on stream (STOS-ATR-IR)
methodology to examine the liquid phase adsorption of an α,β-
unsaturated aldehyde on a Pt/SiO2 catalyst. Herein, by limiting
the exposure of the catalyst to the reagent in solution, it is
possible to detect species adsorbed on the metal before support
Differentiating between potential reaction intermediates and
inactive spectrator species for catalytic reactions in the liquid
phase is very challenging. A technique of choice in the study of
the catalyst−liquid interface over the past decade has been
ATR-IR spectroscopy.9 The use of model metal films10 or
catalyst/reagent systems in which no support adsorption
occurs11 has been reported and enabled spectra showing
interaction with only the metal to be directly obtained.
However, for powdered catalyst films, there are limitations of
the technique in terms of identifying the relevant species
associated with the active metal in spectra that also contain
contributions from species dissolved in the liquid phase (within
the catalyst pores and in the interparticle void volume12) as well
as species adsorbed on the support. Liquid phase contributions
and support adsorption can dominate spectra for supported
metal catalysts because most of the surface area probed by the
IR beam will be associated with the support compared with the
contribution from adsorption on the metal.13
Modulation excitation spectroscopy with phase-sensitive
detection coupled with ATR-IR spectroscopy has been
developed to probe species adsorbed on the catalyst, being
applied to both model metal and catalyst films, for systems such
as alcohol oxidation14 and adsorption of chiral modifiers on Pt
or Pd for chemoselective hydrogenations.15 Other methods
(both experimental and analytical) to overcome the limitation
and distinguish adsorption on the metal in the presence of
contributions from the liquid/support adsorption include
performing reactions with two ATR-IR probes. Therein, one
was coated with catalyst and the other used to monitor the
dissolved species during the reaction.16 Sun and Williams
recorded “blank” experiments for reagents in the liquid over the
support and the catalyst and then deconvoluted the spectra to
distinguish adsorption on the metal.17
2471
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
adsorption, and contributions from the liquid phase species
dominate the spectra.
100 °C and held for 1 h. The reduced Pt/SiO2 and dried SiO2 at
30 °C were used for their respective background spectra. Citral
was delivered to the cell by means of a saturator system with an
Ar flow (20 cm3 min−1, saturator temperature of 25 °C), which
corresponded to 200 ppm of citral. Spectra were recorded with
1024 scans at 4 cm−1 resolution every 5 min over a 30 min
period.
Although the hydrogenation of citral has been widely studied,
there are limited studies probing the adsorption geometry of
citral, in comparison with other α,β-unsaturated aldehydes.
Density functional theory (DFT) studies on the effect of amine
capping agents on reaction selectivity for amine-capped PtCo
nanocrystals7 and the effect of solvent on the hydrogenation of
the isolated versus conjugated CC bonds in citral21 have been
reported. In addition, infrared spectroscopy for the adsorption
of citral on Rh and Pt blacks22 and ATR-IR spectroscopy to
study the deactivation over a Pd/Al2O3 catalyst23 have been
studied. In the ATR-IR study, adsorption of CO on Pd from
decarbonylation of citral was observed during the reaction.
Additional species in the ATR-IR spectra due to citral dissolved
in the solvent and adsorbed on the alumina support, as well as
dissolved products (citronellal/dihydrocitronellal), carboxylate
species, and bands between 1300 and 1450 cm−1 due to
unassigned adsorbed species were observed. The interaction of
citral with Pd was not observed therein.
Because the adsorption geometry is clearly critical in
determining the ultimate selectivity, it is important to be able
to probe this adsorption on the metal, in particular, using a real
catalyst in the presence of the solvent. Therefore, we have
applied STOS-ATR-IR coupled with site-blocking with CO24 to
distinguish signals from citral adsorption of Pt from adsorption
on the silica support as well as citral in the liquid phase. This has
provided information that has been used to understand the
reaction selectivity in the hydrogenation of citral to geraniol and
nerol and changes in adsorption geometry that occur under
different surface modifications.
ATR-IR spectra were recorded using a PIKE ATRMax II
accessory in a Vertex 70 infrared spectrometer (Bruker) over a
ZnSe internal reflection element (IRE) or ZnSe coated with Pt/
SiO2 catalyst. The Pt/SiO2 catalyst layers were prepared by
grinding the catalyst and forming a suspension in ultrapure
water (0.25 g of catalyst/10 g of water). The solution was then
sonicated for 10 min immediately before slurrying 1 cm3 onto
the ZnSe IRE. Thereafter, the film was left in air to dry for 15 h.
This catalyst film was used as background, and all spectra were
recorded with 32 scans at 4 cm−1 resolution. Importantly,
although the depth of penetration of the IR beam into a 2-
propanol-saturated 5% Pt/SiO2 layer is ∼1.5 μm at 1000 cm−1,
the thickness of the deposited layers exceeds this value (∼100
μm), and the particles are also crushed and sieved to <45 μm, so
the catalyst particle size is, on average, much greater than the
depth of penetration of the beam. The catalyst layers were
found to be stable and reproducible with comparable intensities
for bands due to 2-propanol obtained in each experiment for
each new layer of catalyst that was deposited. No catalyst
particles were detected in the outlet of the flow cell, and the
layers were intact following each experiment.
Solutions (0.1 M) of citral and citronellal (96%, Alfa Aesar)
in 2-propanol (>99.8% from Sigma-Aldrich, used as received)
were passed over the catalyst using an in-house flow-through
cell at 0.5 cm3 min−1, and the liquid flow was controlled by
metering valves (Swagelok). To equilibrate the catalyst to the
solvent, 2-propanol was flowed over the catalyst for 15 min (a
steady signal for 2-propanol observed after this time), and then
the feed was switched, using an 8 position electrically actuated
fast switching valve from Vici, to 0.1 M citral or citronellal
solutions with spectra collected every 27 s for the desired time.
To distinguish the adsorption of the compounds on the Pt from
the SiO2 support, CO was preadsorbed onto the catalyst. CO
was bubbled through 2-propanol for 30 min before being flowed
over the catalyst for 90 min, at which point no further changes
to the spectra were observed. The flow was then switched to a
pure 2-propanol feed for 15 min to purge the cell of CO and
remove any weakly bound CO. Thereafter, the feed was
switched to expose the catalyst to 0.1 M citral or citronellal in 2-
propanol.
EXPERIMENTAL SECTION
■
Citral hydrogenation reactions were carried out in the kinetic
regime, free from mass transfer limitations using a 100 cm3
Autoclave Engineer’s high-pressure reactor charged with 0.02
mol of citral (96%, Alfa Aesar) and 0.3 g of 5% Pt/SiO2 catalyst
(Johnson Matthey) suspended in 60 cm3 of 10 mol % 2-
propanol in water. The reactor was purged three times with N2
before heating to 100 °C with the reaction mixture agitated at
1500 rpm. After purging with H2 (high purity from BOC) three
times, the reactor was pressurized to 1 MPa, which corresponds
to t = 0. The reaction was monitored by sampling at regular
intervals with GC analysis of the samples (DB-1 capillary
column and FID detector). The catalyst recycle was performed
by removal of the catalyst from the reaction mixture by
filtration. The catalyst was then washed with IPA and dried at
120 °C before being reused. In the spiking experiments, 0.003
mol of citronellal was added to the citral/solvent reaction
mixture and the reaction was performed as before.
To probe the effect of citronellal on the adsorption of citral,
0.1 M citronellal/2-propanol was flowed over the Pt/SiO2 film
for 25 min before switching the feed to 0.1 M citral/2-propanol
and recording spectra every 27 s following the switch.
Hydrogenation reactions were also performed in pure 2-
propanol, and the same conversion-selectivity profiles were
observed with and without spiking with citronellal. Because of
the strong adsorption bands from water, all the ATR-IR
spectroscopic measurements were performed using 2-propanol
rather than 2-propanol/water mixtures.
DRIFT spectra of citral adsorbed on SiO2 (Davicat) and Pt/
SiO2 were collected using a Collector II accessory with an
environmental chamber (Spectra-Tech) in a Vertex 70 infrared
spectrometer (Bruker). Prior to exposure to citral, the Pt/SiO2
catalyst was reduced at 100 °C in 10% H2 in Ar (50 cm3 min−1)
for 1 h, and the gas was switched to Ar and cooled to 30 °C for
the adsorption study. The SiO2 catalyst was heated under Ar to
In the ATR-IR experiments, the catalyst was reduced by
flowing the 2-propanol over the catalyst film prior to switching
the feed to citral, citronellal, or CO. This reduction procedure
was utilized in the ATR-IR experiments, as opposed to using
hydrogen dissolved in the solvent, to avoid significant amounts
of hydrogen adsorbed on the Pt being present. This hydrogen
would react with the citral on switching to the citral feed and
complicate the spectra as a result of the presence of bands due
to both products and citral adsorption. It should also be noted
that the citral hydrogenation reaction was also performed
without prior reduction of the catalyst, and no effect on the rate
2472
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
or selectivity was observed. A schematic of the ATR-IR reactor
setup is shown in Figure 1.
Figure 1. Schematic of ATR-IR reactor setup.
RESULTS AND DISCUSSION
■
The selective hydrogenation of citral using a 5% Pt/SiO2
catalyst has been studied, and the time profile is shown in
Figure 2a. A 97% selectivity to gernaiol/nerol at 100%
conversion for the reaction performed in water/2-propanol
(9:1 mol ratio) was observed (Figure 2). It should be noted that
with the exception of the formation of citronellal and geraniol/
nerol, all other products made up <1% of the composition in
the liquid phase. The higher apparent citral conversion observed
in the initial 60 min of reaction is thought to be due to
adsorption of citral on the catalyst. For example, after 10 min of
reaction, 1.74 × 10−3 mol of citral is found to be converted into
products, indicating that 2.48 × 10−3 mol of citral is not
observed because of adsorption; that is, the mass balance was
∼86.9% at the early stages of reaction. This is consistent with
adsorption isotherms measured using the same mass of catalyst
that showed 2.6 × 10−3 mol of citral was adsorbed on the metal
and support. In this adsorption-dominated period of the
reaction, up to 60 min of the reaction, the selectivity gradually
increased from 70% to geraniol/nerol and 30% to citronellal at
10 min and, as the reaction proceeded to 100% conversion, the
selectivity to geraniol/nerol increased to 97%. Thereafter, the
mass balance increased, and by 180 min of reaction, 97% of the
molecules were accounted for. At this point, some of the
citronellal formed at the start of the reaction was not observed
in the liquid phase and is thought to be adsorbed on the catalyst.
Interestingly, on recycling the catalyst, enhanced selectivity to
geraniol/nerol was observed from the start of the reaction
(Figure 3), indicating that the adsorbed species at the end of the
first reaction modifies the surface of the catalyst.
Figure 2. (a) Percent disappearance of citral from the liquid phase
⧫
during the hydrogenation, in mol % ( ), and selectivity to geraniol +
□
nerol ( ) as a function of reaction time. (b) Comparison of percent
selectivity of geraniol + nerol during citral hydrogenation with a fresh
◇
catalyst ( ) and after the reaction is spiked with 15 mol % citronellal
■
( ) as a function of percent conversion to products calculated from
product concentrations in the liquid phase. Reaction conditions: citral,
0.02 mol; solvent, 60 cm3 (10 mol % 2-propanol in water); 5% Pt/SiO2
catalyst, 0.3 g (particle size <45 μm); 100 °C; H2 pressure, 1 MPa;
stirrer speed, 1500 rpm.
has been probed using DRIFT and ATR-IR spectroscopy
coupled with the liquid phase STOS technique.
Figure 4 shows that when using conventional DRIFT
spectroscopy, no significant differences after adsorption of gas
phase citral on the Pt/SiO2 catalyst compared with the blank
SiO2 are observed. In addition, after blocking the Pt by
adsorption of CO, similar spectra are obtained for the catalyst
and the blank support. This indicates that the conventional
DRIFT spectra are dominated by the interaction between the
citral and silica and, therefore, cannot be used to probe the
surface interaction with the active site.
To investigate the adsorption in the presence of the 2-
propanol solvent, ATR-IR spectroscopy was used. In these
experiments, the solvent was equilibrated with the catalyst layer
on the ATR crystal using a flow cell, after which the feed was
switched to a solution of 0.1 M citral in 2-propanol. The
subsequent ATR-IR spectra of citral on the Pt/SiO2 catalyst are
shown in Figures 5 and S1 (Supporting Information; SI) for
different exposure times to the feed.
To examine whether this could be due to citronellal
adsorption, a fresh catalyst−solvent mixture was spiked with
citronellal. After spiking the reaction mixture with citronellal,
the minor reaction product, selectivities of >97% to gernaiol/
nerol from the start of the reaction were observed (Figure 2b).
Therefore, it is postulated that the citronellal product adsorbs
onto the catalyst, altering the mode of citral adsorption and thus
increasing the selectivity for CO rather than CC
hydrogenation either through templating the surface or
adsorbing on unselective sites. In the present paper, this effect
At long exposure times (SI Figure S1), the bands observed on
the catalyst film were comparable to those found in the liquid
phase citral spectrum over the blank ZnSe crystal, both in
position and relative intensity (Figure 5). In this case, it is
2473
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
thought that the spectra obtained using the catalyst film are
dominated by liquid phase citral in the catalyst pores or the
interparticle void volume of the catalyst powder film. In
contrast, when the catalyst was exposed to the citral/2-propanol
feed for only a short time (Figure 5), changes in the relative
intensities of the bands were detected when compared with the
pure liquid phase spectra. For example, the spectra observed 81
s following the switch in feed showed that the ν(CO) bands
at 1667/1681 cm−1 (trans/cis CO citral, respectively; the cis
and trans ν(CC) bands occur at 1610/1634 cm−1) were of
comparable intensity to the bands found at 1379 and 1396 cm−1
due to δ(CH3)sym and aldehyde C−H rock, respectively.25,26
A
band at 1027 cm−1 was also observed 81 s after switching the
feed, which over 4 min of exposure became more intense than
the C−H deformation bands (SI Figure S1). In the citral
spectrum in the absence of the catalyst, that is, the uncoated
ZnSe IRE crystal, the bands in this region are of much lower
intensity than the alkane deformation bands, which shows that
significant intensity enhancement occurs in these spectra.
Furthermore, other bands of citral (for example, the band at
1195 cm−1) were not detected at this exposure time. The
enhanced intensity of these bands relative to the CO band
over the catalyst film (Figures 5 and SI S1), which is normally
the most intense band in the citral spectrum, suggests that these
bands are associated with adsorbed citral or that the reduction
in intensity of the CO band is associated with binding of the
aldehydic part of the molecule. The changes in the STOS
spectra with exposure time are striking and show that it is
possible to differentiate between citral adsorbed on a real
catalyst and the dissolved citral in the liquid phase within the
catalyst pores/interparticle void volume.
Of course, although the short time-on-stream ATR-IR
spectroscopy has clearly identified bands due to adsorbed citral,
as opposed to liquid phase citral, these adsorbed species may
not be on the metal sites. In fact, the DRIFT spectra indicated
that significant adsorption on the support was possible from the
gas phase. To elucidate where the citral was adsorbing and, thus,
to be able to comment on the mode of the adsorption, CO(g)-
saturated 2-propanol was passed over the catalyst to block the
Pt sites.24 A Pt−CO band at 2075 cm−1 due to linearly adsorbed
Figure 3. (a) Percent conversion to products for the hydrogenation of
⧫
■
citral: , fresh catalyst; , after first recycle; and Δ, after a second
recycle as a function of reaction time; and (b) percent selectivity to
⧫
■
geraniol/nerol for , fresh catalyst; , after first recycle; and Δ, after a
second recycle as a function of conversion to products. Reaction
conditions: citral, 0.02 mol; solvent, 60 cm3 (10 mol % 2-propanol in
water); 5% Pt/SiO2 catalyst, 0.3 g (particle size <45 μm); 100 °C; H2
pressure, 1 MPa; stirrer speed, 1500 rpm.
Figure 4. DRIFT spectra of citral adsorbed on (a) SiO2, (b) Pt/SiO2, and (c) Pt/SiO2 after adsorption of CO.
2474
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
Figure 5. ATR-IR spectra normalized to the ν(CO) band at 1667 cm−1 of (a) citral over a blank ZnSe, (b) 0.1 M citral in 2-propanol over a ZnSe
IRE 81 s exposure to the feed (short time-on-stream ATR-IR spectrum) and (c) over Pt/SiO2, 81 s exposure to the feed (short time-on-stream ATR-
IR spectrum). Bands due to 2-propanol have been subtracted from all spectra.
Figure 6. Short time-on-stream ATR-IR spectra of 0.1 M citral in 2-propanol over Pt/SiO2 (gray spectra) and CO-Pt/SiO2 (black spectra) after (a)
81 s for Pt/SiO2 and 108 s of exposure for CO-Pt/SiO2 and (b) 4 min of exposure for both Pt/SiO2 and CO-Pt/SiO2. Bands due to 2-propanol have
◇
been subtracted. , bands due to citral adsorbed on support; *, bands due to citral adsorbed on Pt.
CO was observed, and this band was retained following a purge
with 2-propanol, showing strong adsorption of CO on the Pt.
The feed was then switched to 0.1 M citral/2-propanol, and SI
Figure S2 shows the change with time over CO-Pt/SiO2
following the switch. Clear differences are observable, even
after 4 min of exposure to the citral-containing feed, as
compared with those obtained for citral adsorbed on Pt/SiO2
without pre-exposure to CO (Figure 5 and SI S1).
A comparison of spectra following citral adsorption 81 s and
4 min after the switch in the feed on Pt/SiO2 and CO−Pt/SiO2
is shown in Figure 6. After 81 s, only bands at 1668 and1632/
1607 cm−1 were observed in the CO−Pt/SiO2 spectra, whereas
in the Pt/SiO2 spectra, bands at 1461, 1450, 1396, 1379, and
1027 cm−1 were also observed (Figure 6a). It is evident that
bands that are present on Pt/SiO2 are lost when CO is
preadsorbed on the catalyst, indicating that these bands are
related to adsorbed citral species on Pt and not on the support.
After 4 min of exposure to the feed, the 1668 cm−1 band in the
spectra over the CO−Pt/SiO2 catalyst was found to increase in
intensity, with additional bands at 1450, 1401, and 1195 cm−1
also being observed to grow (Figure 6b). These bands are also
present after 4 min of exposure over Pt/SiO2 and are proposed
to be due to adsorption of citral on the silica support. These
spectral features are not likely to be due to liquid phase citral
because if this were the case, all the bands would be expected to
occur with the same relative intensities as in the reference liquid
phase spectrum, and this is not found. For example, the peak at
1379 cm−1 is relatively much weaker in the spectrum after 4 min
of exposure compared with that after 15 min of exposure to the
feed, that is, when the spectrum has more of a contribution
2475
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
Figure 7. ATR-IR spectra of 0.1 M citral in 2-propanol on Pt/SiO2 after 25 min of exposure to 0.1 M citronellal in 2-propanol. (a) Black spectrum:
citral adsorption 81 s after switching feed (no preadsorption of citronellal). Gray spectrum: citral adsorption 81 s after switching feed for Pt/SiO2 with
adsorbed citronellal. (b) Citral adsorbed 25 min after switching the feed for Pt/SiO2 with adsorbed citronellal. Bands due to adsorbed citronellal/2-
propanol have been subtracted.
from the liquid in the pore (SI Figure S2b). In addition, changes
in the region 1000−1200 cm−1 are observed. However, these
are relatively small changes, indicating a weak interaction with
the support, as compared with the Pt; thus, it is difficult to
determine the mode of adsorption.
A contribution to the CO region of the spectra with and
without adsorbed CO is also observed. Citral adsorbed on Pt/
SiO2 has a contribution at 1649 cm−1, a shoulder between the
1667 cm−1 ν(CO) band and the 1634/1610 cm−1 ν(CC)
bands of citral that suggests a weakly perturbed CO band in
the adsorbed citral molecule. When CO was preadsorbed, the
1649 cm−1 contribution was lost, together with the bands at
1461, 1396, 1379, and 1027 cm−1. All of these bands are related
to citral adsorption on Pt.
It should be noted that in the period immediately following a
switch from 2-propanol to 0.1 M citral in 2-propanol, the initial
citral concentration in the flow cell will be significantly less than
0.1 M. As the citral solution flows, the concentration of citral
molecules in the catalyst layer will increase until the layer (Pt +
support) is saturated. Spectra recorded after switching to the
citral feed are, therefore, effectively taken over an increasing
concentration range. In the first few spectra after the switch,
that is, as the concentration in the feed on the surface of the
catalyst increases, there is no significant change in the citral
bands observed. The bands change only where it is thought that
the support adsorption/pore filling dominates, which indicates
that there is no significant effect on the spectra as a function of
the surface citral concentration on the metal.
The hydrogenation results showed that adsorbed citronellal
alters the interaction of citral with the catalyst, thus enhancing
CO hydrogenation. STOS-ATR-IR spectra probing the
perturbation in citral adsorption following preadsorption of
citronellal were studied. The catalyst was pre-exposed to 0.1 M
citronellal/2-propanol for 25 min (SI Figure S3) before
switching the feed to 0.1 M citral in 2-propanol. When citral
is adsorbed on the as-received catalyst, bands at 1634, 1461,
1396, 1379, and 1027 cm−1 are observed in the spectra as a
result of citral adsorbed on Pt (Figures 5 and 7a). However,
when citral is adsorbed on the catalyst pre-exposed to
citronellal, there is loss of the band at 1394 cm−1 (aldehyde
C−H rock) and some loss of intensity at 1462 and 1379 cm−1
(CH3 asym/sym deformation) (Figure 7a). The absence of the
band associated with the aldehydic C−H suggests that the
CO group is adsorbing. This is consistent with the higher
selectivity to geraniol/nerol observed in the reaction spiked
with citronellal. In the spectrum observed 81 s after the switch,
the band associated with support adsorption at ∼1450 cm−1 is
of comparable intensity in the two spectra, yet the CO band
is of significantly weaker intensity for citral after citronellal
adsorption (Figure 7a). The spectra due to citral adsorption at
long exposure times after switching the feed show this band at
1396 cm−1, which is due to either citral in the liquid phase or
citral adsorption because citronellal desorbs under flow (Figure
7b).
Spiking the reaction with citronellal gave 97% selectivity to
geraniol/nerol from the start of the reaction, and in the STOS-
ATR-IR spectra, citral, following preadsorption of citronellal, is
2476
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
Figure 8. Short time-on-stream ATR-IR spectra of 0.1 M citronellal in 2-propanol on Pt/SiO2: after 108 s (black spectrum) and 4 min (gray
spectrum) of exposure to the citronellal solution. Bands due to 2-propanol have been subtracted.
found to preferentially adsorb through the CO group. For the
reaction without spiking, the initial selectivity in the reaction is
70% to CO and 30% to CC hydrogenation. STOS-ATR-
IR spectra of citral without citronellal present show the band at
1394 cm−1, which suggests that citral adsorbs through both the
CO and CC groups either separately as η2 modes (di-σ(CC)
higher selectivity for CO hydrogenation is observed (Scheme
2).
Scheme 2. Schematic of Citral Adsorption Modes and the
Influence of Citronellal on the Reaction Selectivity
or η2 di-σ(CO)) or through both groups in an η3 (di-σ(CC)−σ(O)
)
or η4 (di-σ(CC) + di-σ(CO)) mode.24,25 We postulate that at the
start of the reaction, citral adsorption on Pt can occur through
both η3 (di-σ(CC)−σ(O)) and η2 (di-σ(CO)) adsorption modes. If
interaction of the CC bond of citral with Pt was through an
η4 (di-σ(CC) + di-σ(CO)) mode, bands due to aldehydic species
would not be expected in the spectra of citral on Pt, which is not
the case. The weakly perturbed CO band at 1649 cm−1
observed for citral adsorbed on Pt is associated with the η3 (di-
σ(CC)−σ(O)) mode; the band at 1027 cm−1, with the C−O
(single bond character) of an η2 di-σ(CO) mode hydrogen-
bonded with the 2-propanol. The C−O stretch of an alcohol
would be found in this region.27
Citral adsorption at the start of the reaction, through
individual CC and CO adsorption modes results in the
formation of citronellal and geraniol/nerol. As the reaction
proceeds, the citronellal formed initially adsorbs on Pt. STOS-
ATR-IR spectra of 0.1 M citronellal/2-propanol adsorption
show a band at 1016 cm−1 that is of greater intensity than the
ν(CO) bands at 1731/1718 cm−1 (Figure 8). This 1016 cm−1
band is proposed to be due to citronellal adsorbed in an η2 (di-
σ(CO)) mode hydrogen bonded with 2-propanol, analogous to
the η2 (di-σ(CO)) citral band at 1027 cm−1. The most stable
adsorption mode of C1−C4 aldehydes has been calculated from
DFT to be the η2 (di-σ(CO)) mode.28
The adsorption of citronellal alters the subsequent adsorption
of citral in the reaction, which causes the change in selectivity as
the reaction proceeds, providing enhanced CO hydro-
genation. When citronellal is present in the system from the
start of the reaction (recycling the catalyst as well as the spiking
experiments), citral adsorption on the Pt through the CC
bond (η3 (di-σ(CC)−σ(O)) mode) does not occur, and hence,
These results can be compared with the changes in citral
hydrogenation on other Pt-based catalysts following adsorption
of organic modifiers. Using Pt3Co nanocrystals capped with
aliphatic amines, with increasing chain length of the amine from
butyl amine to octadecylamine, the rate decreased significantly,
but importantly, the selectivity to the CO hydrogenation
versus the CC reduction was enhanced.7 This was attributed
2477
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478

ACS Catalysis
Research Article
171. (e) Siani, A.; Alexeev, O. S.; Lafaye, G.; Amiridis, M. D. J. Catal.
2009, 266, 26−38. (f) Stassi, J. P.; Zgolica, P. D.; de Miguel, S. R.;
Scelza, O. A. J. Catal. 2013, 306, 11−29.
to the long alkyl chain amines stopping the CC bond from
interacting with the surface as a consequence of steric
hindrance, forcing the citral to adsorb in a tilted geometry
through the CO bond only. Kahsar et al. demonstrated that
thiols interacting with Pt/Al2O3 could also lead to enhanced
selectivity for the unsaturated alcohol in the case of
cinnamaldehyde and prenal.8 In both cases, this effect was
attributed to noncovalent interactions changing the adsorption
geometry compared with the uncoated catalyst. Interestingly,
although little effect was found on comparing aromatic and
aliphatic thiols in the case of prenal hydrogenation, further
increases in selectivity were observed for the case of
cinnamaldehyde, which is reported to be due to π−π
interactions between the substrate and the coating. Although
similar effects are found between the results reported herein and
these studies, it is important to note the large differences in the
adsorption strengths in each case, from the very strong thiol
binding to the weakly bound citronellal in the present study.
This illustrates that such effects are a general phenomenon and
are likely to occur in many selective reactions of multifunctional
molecules.
(3) Marshall, S. T.; O’Brien, M.; Oetter, B.; Corpuz, A.; Richards, R.
M.; Schwartz, D. K.; Medlin, J. W. Nat. Mater. 2010, 9, 853−858.
(4) Makosch, M.; Lin, W.-I.; Bumbalek, V.; Sa, J.; Medlin, J. W.;
Hungerbuhler, K.; van Bokhoven, J. A. ACS Catal. 2012, 2, 2079−
2081.
́
(5) Alvarez-Rodríguez, J.; Guerrero-Ruiz, A.; Arcoya, A.; Rodríguez-
Ramos, I. Catal. Lett. 2009, 129, 376−382.
(6) Jiang, H.-J.; Jiang, H.-B.; Zhu, D.-M.; Zheng, X.-L.; Fu, H.-Y.;
Chen, H.; Li, R.-X. Appl. Catal., A 2012, 445−446, 351−358.
(7) Wu, B.; Huang, H.; Yang, J.; Zheng, N.; Fu, G. Angew. Chem., Int.
Ed. 2012, 51, 3440−3443.
(8) Kahsar, K. R.; Schwartz, D. K.; Medlin, J. W. J. Am. Chem. Soc.
2014, 136, 520−526.
(9) (a) Burgi, T.; Baiker, A. Adv. Catal. 2006, 50, 227−283. (b) Burgi,
T. J. Catal. 2005, 229, 55−63. (c) Ebbesen, S. D.; Mojet, B. L.; Lefferts,
L. J. Catal. 2008, 256, 15−23. (d) Keresszegi, C.; Ferri, D.; Mallat, T.;
Baiker, A. J. Phys. Chem. B 2005, 109, 958−967. (e) Burgener, M.;
Wirz, R.; Mallat, T.; Baiker, A. J. Catal. 2004, 228, 152−161.
(f) Andason, J.-M.; Baiker, A. Chem. Soc. Rev. 2010, 39, 4571−4584.
(10) (a) Meier, D. M.; Mallat, T.; Ferri, D.; Baiker, A. J. Catal. 2006,
244, 260−263. (b) Ferri, D.; Burgi, T.; Baiker, A. J. Phys. Chem. B 2001,
̈
105, 3187−3195.
CONCLUSIONS
■
(11) Ebbesen, S. D.; Mojet, B. L.; Lefferts, L. Langmuir 2008, 24,
STOS-ATR-IR spectroscopy has allowed us to investigate the
adsorption mode of citral on Pt, even when adsorption on the
silica support is possible and eventually dominates the spectra.
Using this technique, we have identified a possible explanation
for why the selectivity in the citral hydrogenation reaction
changes as the reaction proceeds. It appears that the formation
of a product (such as citronellal) can modify the adsorption of
the citral and force the reaction down one particular pathway,
leading to very high selectivity to geraniol and nerol.
869−879.
(12) Ortiz-Hernandez, I.; Williams, C. T. Langmuir 2003, 19, 2956−
2962.
(13) Meier, D. M.; Ferri, D.; Mallat, T.; Baiker, A. J. Catal. 2007, 248,
68−76.
(14) Burgi, T.; Bieri, M. J. Phys. Chem. B 2004, 108, 13364−13369.
(15) (a) Maeda, N.; Hungerbuhler, K.; Baiker, A. J. Am. Chem. Soc.
̈
2011, 133, 19567−19659. (b) Meemken, F.; Maeda, N.;
Hungerbuhler, K.; Baiker, A. Angew. Chem., Int. Ed. 2012, 51, 8212−
̈
8216.
(16) Richner, G.; van Bokhoven, J. A.; Neuhold, Y.-M.; Makosch, M.;
ASSOCIATED CONTENT
* Supporting Information
■
Hungerbuhler, K. Phys. Chem. Chem. Phys. 2011, 13, 12463−12471.
̈
S
(17) Sun, X.; Williams, C. T. Catal. Commun. 2012, 17, 13−17.
(18) Burgi, T. J. Catal. 2005, 229, 55−63.
Overlaid ATR-IR spectra following switching in a 0.1 M citral
solution over a untreated and CO treated Pt/SiO2 catalyst.
Overlaid ATR-IR spectra following switching in a 0.1 M
citronellal solution over a untreated. Pt/SiO2 catalyst character-
ization: BET, XRD, and TEM. This material is available free of
charge via the Internet at http://pubs.acs.org.
(19) Copeland, J. R.; Foo, G. S.; Harrison, L. A.; Sievers, C. Catal.
Today 2013, 205, 49−59.
(20) Chansai, S.; Burch, R.; Hardacre, C.; Breen, J.; Meunier, F. C. J.
Catal. 2011, 281, 98−105. (b) Chansai, S.; Burch, R.; Hardacre, C. J.
Catal. 2012, 295, 223−231.
(21) Chatterjee, A.; Chatterjee, M.; Ikushima, Y.; Mizukami, F. Chem.
Phys. Lett. 2004, 395, 143−149.
AUTHOR INFORMATION
Corresponding Author
*E-mail: c.hardacre@qub.ac.uk.
Notes
■
(22) Pak, A. M.; Vozdvizhenskii, V. F.; Turganbaeva, S. M.; Levintova,
T. D.; Konuspaev, S. R. React. Kinet. Catal. Lett. 1983, 23, 1−5.
(23) Burgener, M.; Wirz, R.; Mallat, T.; Baiker, A. J. Catal. 2004, 228,
152−161.
(24) Ferri, D.; Mondelli, C.; Krumeich, F.; Baiker, A. J. Phys. Chem. B
2006, 110, 22982−22986.
The authors declare no competing financial interest.
(25) (a) Colthup, N. B.; Daly, L. H.; Wiberley, S. E. Introduction to
Infrared and Raman Spectroscopy; Academic Press: New York, 1975.
(b) Haubrich, J.; Loffreda, D.; Delbecq, F.; Sautet, P.; Jugnet, Y.;
Krupski, A.; Becker, C.; Wandelt, K. J. Phys. Chem. C 2008, 112, 3701−
3718. (c) Dandekar, A.; Vannice, M. A. J. Catal. 1999, 183, 344−354.
ACKNOWLEDGMENTS
We acknowledge EPSRC for funding as part of the CASTech
Grant (EP/G011397/1).
■
(d) Silvestre-Albero, J.; Sepulveda-Escribano, A.; Rodríguez-Reinoso,
́
REFERENCES
■
F.; Anderson, J. A. Phys. Chem. Chem. Phys. 2003, 5, 208−216.
(26) (a) Delbecq, F.; Sautet, P. J. Catal. 2002, 211, 398−406.
(b) Kliewer, C. J.; Bieri, M.; Somorjai, G. A. J. Am. Chem. Soc. 2009,
131, 9958−9966.
(27) Merlo, A. B.; Santori, G. F.; Sambeth, J.; Siri, G. J.; Casella, M. L.;
Ferretti, O. A. Catal. Commun. 2006, 7, 204−208.
(28) Sinha, N. K.; Neurock, M. J. Catal. 2012, 295, 31−44.
(1) (a) Stolle, A.; Gallert, T.; Schmoger, C.; Ondruschka, B. RSC Adv.
̈
2013, 3, 2112−2153. (b) Singh, U. K.; Vannice, M. A. J. Catal. 2000,
190, 165−180. (c) Singh, U. K.; Sysak, M. N.; Vannice, M. A. J. Catal.
2000, 191, 181−191. (d) Mukherjee, S.; Vannice, M. A. J. Catal. 2006,
243, 108−130. (e) Mukherjee, S.; Vannice, M. A. J. Catal. 2006, 243,
131−148.
(2) (a) Ponec, V. Appl. Catal., A 1997, 149, 27−48. (b) Hirschl, R.;
Delbecq, F.; Sautet, P.; Hafner, J. J. Catal. 2003, 217, 354−366.
(c) Delbecq, F.; Sautet, P. J. Catal. 2003, 220, 115−126. (d) Vilella, I.
M. J.; de Miguel, S. R.; Scelza, O. A. J. Mol. Catal. A 2008, 284, 161−
2478
dx.doi.org/10.1021/cs500185n | ACS Catal. 2014, 4, 2470−2478