Aerobic oxidation of alcohols on Au/TiO2 catalyst: New insights on the role of active sites in the oxidation of primary and secondary alcohols

Article abstract of DOI:10.1007/s00706-015-1616-3

A study of secondary, benzylic, and primary alcohols oxidation on a reference Au/TiO₂ catalyst is reported. Reactions gave rise to ketones, aldehydes, or esters in good yields. The FT-IR results and catalytic tests data provide to elucidate the reaction mechanism and the specific role of the active sites on titania.

Full text of DOI:10.1007/s00706-015-1616-3

Monatsh Chem (2016) 147:391–403
DOI 10.1007/s00706-015-1616-3
ORIGINAL PAPER
Aerobic oxidation of alcohols on Au/TiO2 catalyst: new insights
on the role of active sites in the oxidation of primary
and secondary alcohols
Floriana Vindigni1 Stefano Dughera2 Francesco Armigliato2 Anna Chiorino2
•
•
•
Received: 17 September 2015 / Accepted: 23 November 2015 / Published online: 15 December 2015
Ó Springer-Verlag Wien 2015
Abstract A study of secondary, benzylic, and primary
alcohols oxidation on a reference Au/TiO2 catalyst is
reported. Reactions gave rise to ketones, aldehydes, or
esters in good yields. The FT-IR results and catalytic tests
data provide to elucidate the reaction mechanism and the
specific role of the active sites on titania.
Introduction
Gold catalysis has become a particularly active area of
research within only a few years. In particular, the last
decade has witnessed a remarkable growth in the number
of organic reaction that are catalyzed by gold complexes
such as oxidative processes [1–4], hydrogenations [5–8],
cross-couplings [9–12], and others [13–15].
Graphical abstract
Selective oxidation of alcohols plays key role in organic
synthesis and in industrial practice. The application of
heterogeneous catalysis and molecular oxygen [16] to
oxidation reactions offers a green alternative to traditional,
toxic chemical oxidants. Supported gold nanoparticles (Au
NPs) are a means of carrying out alcohol oxidation reac-
tions under mild and green conditions.
It must be stressed that the catalytic activity of gold is
remarkably sensitive to many factors, such as gold particle
size, preparation method, and support. In particular, the
nature of the support can play a key role in its catalytic
activity; in fact a number of metal oxides such as CeO2
[17], TiO2 [18, 19], mesoporous silica [20], Fe2O3 [21],
NiO [22, 23], MgO [24], metal hydroxides such as
Mg(OH)2 [25] and zeolites [26] were used as a support of
Au NPs. For example, Au/CeO2 is able to catalyze the
selective oxidation of alcohols to aldehydes or ketones and
acids in relatively mild conditions using O2 as oxidizing
agent [17]. On the contrary, the oxidation carried out in the
presence of Au/TiO2 is scarcely studied [18, 19] and often
requires harsh reaction conditions (e.g., under high pres-
sure in an autoclave).
Keywords Alcohol Á Oxidation Á Gold Á
Reaction mechanism Á FT-IR
Electronic supplementary material The online version of this
article (doi:10.1007/s00706-015-1616-3) contains supplementary
material, which is available to authorized users.
& Floriana Vindigni
floriana.vindigni@unito.it
& Stefano Dughera
stefano.dughera@unito.it
In addition, great attention is also being paid to the study
of the specific role that active sites play in the oxidation
reaction. In this regard, recent works have reported
mechanisms based on the activation of oxygen atoms on
the gold surface [27]. Furthermore, it is generally accepted
1
Department of Chemistry, NIS Centre of Excellence,
University of Torino, via P. Giuria 7, Turin, Italy
2
Department of Chemistry, University of Torino, via P. Giuria
7, 10125 Turin, Italy
123

392
F. Vindigni et al.
that the contact boundaries between the gold and the oxide
support play an important role in the catalytic mechanism
[21]. Therefore, since many factors can influence catalytic
activity, i.e., Au dispersion, nature of the support, metal-
support interaction, etc., this research topic is still open to
further study.
2a was obtained in excellent yield. Interestingly, also the
reaction carried out in neat conditions gave 2a in good
yield (91 %, Table 1; entry 9).
The catalyst was easily recovered at the end of the
reactions by simple vacuum filtration and was reused for
other two consecutive catalytic cycles and no significant
decrease in catalytic activity was observed, as reported in
Table A (see Supplementary Material). Alternatively,
instead of recovering the catalyst by filtration at the end of
the reaction, 2a was completely removed by extraction
with diethyl ether; new 1a and toluene were then added to
the same environment and the reaction was repeated sev-
eral times. The catalyst was efficient until the sixth
consecutive run. The results are reported in Table B (see
Supplementary Material).
In the light of this, the purpose of our work is to test a
reference Au/TiO2 (AUROliteTM) catalyst, produced by
STREM Chemicals Inc., in the oxidation of primary,
benzylic, and secondary alcohols. FT-IR and DRUV–Vis
spectroscopy were used to investigate its activity and to
clarify the specific role that its active sites play in the
reaction.
Results and discussion
The high yield and simplicity of the procedure encour-
aged us to further exploit the generality and the scope of
this oxidation, using other secondary alcohols 1. The
results are collected in Table 2. Alcohol 1b was oxidized in
excellent yields with or without toluene (Table 2; entries 1,
2). However, the oxidation of 1c only occurred without
solvent (Table 2; entries 3, 4). No results were obtained
with 1d (Table 2; entry 5).
Oxidation reaction of 1-phenylethanol (1a)
and secondary alcohols 1b–1d
Initially, to optimize the reaction conditions, the oxidation
of 1-phenylethanol (1a) to acetophenone (2a) in the pres-
ence of the Au/TiO2 catalyst was studied under varying
conditions, as reported in Table 1.
Spurred on by these encouraging results, a careful FT-IR
spectroscopy investigation was carried out to obtain new
insight into the interaction between alcohol and the active
sites exposed on the catalyst surface.
Firstly, two tests were carried out with TiO2 (Table 1;
entry 1) and NaOH/TiO2 (Table 1; entry 2) to demonstrate
that gold is the actual catalyst. In both cases, only unre-
acted 1a was detected by GC and GC–MS analyses. On the
other hand, no 2a was obtained when the reaction was
carried out without base (Table 1; entry 3). In the light of
this, it is clear that NaOH, Au/TiO2, and atmospheric
oxygen were necessary to obtain the complete oxidation of
1a–2a. The best conditions are those reported in entry 5
and involve the use of toluene as a solvent and NaOH (3
equivalents) as a base. Note that no 2a was obtained when
carrying out the same reaction under nitrogen flow
(Table 1; entry 6), confirming that oxygen is necessary for
alcohol oxidation.
FTIR characterization
The FT-IR spectra of adsorbed CO at -153 °C, before
1-phenylethanol (1a) interaction, on differently treated
catalysts are reported in Fig. 1a.
According to the attribution and nomenclature proposed
by Hadjiivanov et al. [29], the peaks at 2179 and
2164 cm-1 are associated with CO molecules adsorbed
onto pentacoordinated Ti4? ions exposed on (010) and
(101) faces (c sites) and those on pentacoordinated Ti4?
ions exposed on (001) planes (c sites), respectively.
Moreover, CO adsorption at low temperature also produces
a shoulder at 2155 cm-1, attributed to the interaction
between the probe molecules and OH groups; this feature is
confirmed by the perturbation of the m (OH) pattern in the
3750–3600 cm-1 range (not shown here).
It has been reported [28] that Au may leach into the
solution during the reaction. We, therefore, stopped the
reaction after 12 h and removed the catalyst (Table 1; entry
7). GC analyses showed 1a and 2a in almost equivalent
amounts. The residue was reacted for a further 12 h, but the
reaction still did not complete, demonstrating that this
phenomenon did not affect the catalytic performance of the
system.
It is evident that the intensity of the band associated with
CO-support site interaction is strictly related to the specific
thermal pre-treatment that the catalyst was submitted to. In
particular, in the spectrum of the sample underwent to the
mild treatment (i.e., AuTiDeg) the band intensity is the
lowest, indicating that a smaller amount of Ti4? ions are
accessible to the probe molecules in this case. These fea-
tures have been clarified by a comparison of the FT-IR
spectra of the samples before the CO inlet (see Fig. S1 in
Experiments with lower amounts of NaOH, other bases
(Na2CO3, Table 1; entry 8), and other solvents (Table 1;
entries 12, 13) generally led to unsatisfactory results. It
must be stressed that the oxidation of 1a only partially
occurred when bubbling a continuous oxygen flow without
NaOH (Table 1; entry 10). In the presence of NaOH
(Table 1; entry 11), the reaction was complete after 8 h and
123

Aerobic oxidation of alcohols on Au/TiO2 catalyst…
Table 1 Oxidation of 1-phenylethanol (1a)
393
Entry
Solvent
Base (3 eq.)
T/°C
Time/h
Yield/% of 2a, b, c
d
1
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
–
–
110
110
110
110
110
110
110
80
24
24
24
24
24
24
24
48
72
24
8
–
d
–
2
NaOH
–
3
Traces
12e
100f
4
Na2CO3
NaOH
NaOH
NaOH
Na2CO3
NaOH
–
5
g
–
6
7
55h
8
16
9
–
80
91f, i
40j, k
100j
10
11
12
13
a
Toluene
Toluene
H2O
110
110
110
80
NaOH
NaOH
NaOH
48
48
Traces
Traces
MeCN
All the reactions were carried out with 1 mmol of 1a, 3 mmol of base, and 50 mg of Au/TiO2 (equal to 15 mol% of Au)
b
c
d
e
f
Yields refer to the pure and isolated 2a
Note that with lesser amounts of base, the reaction was not complete
The reaction was carried out without catalyst and in the presence of TiO2 (50 mg)
Unreacted 1a was also isolated (85 %)
The conversion was total
g
h
The reaction was carried out under nitrogen flow
The reaction was stopped after 12 h. GC analyses showed the presence of 1a and 2a in almost equivalent amounts. The catalyst was removed
and the remaining liquid was heated at 110 °C for further 12 h. After the usual work-up, we obtained an oily residue which was chromatographed
on a short silica gel column (eluent: petroleum ether/diethyl ether 9:1). The first eluted product was 2a (66 mg; 55 % yield); the second one was
1a (55 mg; 45 %)
i
The reaction carried out at 110 °C gave lower yield of 2a (60 %)
The reaction was carried out under a continuous oxygen flow
j
k
Unreacted 1a was also isolated (57 %)
Supplementary Material). These spectra also highlighted
the presence of a carbonate-like species and water on the
titania surface.
broad absorption in the 2130–2120 cm-1 region appears.
This latter is attributed to the interaction between CO and
Aud? sites. It is well accepted that the Au sites, particularly
those at the metal-support interface, can be made positively
charged via bonding with residual active oxygen species
which are present after an oxidation pretreatment [28, 30].
Moreover, to further confirm that the intensity decrease of
the band due to the CO–Au0 interaction is not a conse-
quence of an Au particles agglomeration, the same FTIR
measurement was undertaken on the Au/TiO2 submitted to
an oxidative treatment at 200 °C (see Fig. S2 in Supple-
mentary Material).
Coming back to the lower frequency range in Fig. 1a,
the band centered at 2100 cm-1 for AuTiDeg (dotted
line) and AuTiRed (solid line) is the stretching mode of
CO molecules adsorbed on Au0 sites. This band has the
same intensity on both spectra and indicates that metallic
gold sites are available to the reactant after this mild
treatment.
However, oxidation treatment gives rise to significant
changes in the shape and in the position of the peaks.
Particularly, the band due to the CO-Au0 interaction
decreases in intensity and slightly shifts to higher fre-
quency, i.e., 2103 cm-1 and, more interestingly, a new
Now, it is necessary to point out that the reactions
described above were carried out on ‘‘as received’’ Au/
TiO2 under air atmosphere, while its reactivity will be
123

394
F. Vindigni et al.
Table 2 Oxidation of secondary alcohols 1b–1d
Entry
R
R0
Solvent
T/°C
Time/h
Yield/% of 2a, b
1
2
3
4
5
a
Ph
Ph
Ph (1b)
Toluene
110
80
15
24
24
52
72
100c (2b)
97c (2b)
– (2c)
Ph (1b)
–
cC6H11 (1c)
cC6H11 (1c)
nC6H13 (1d)
Toluene
110
80
–
–
84c, d (2c)
Me
80
– (2d)
All the reactions were carried out with 1 mmol of 1, 3 mmol of NaOH, and 50 mg of Au/TiO2 (equal to 15 mol% of Au)
Yields refer to the pure and isolated 2
b
c
d
The conversion was total
The reaction carried out at 110 °C gave lower yield of 2c (45 %)
investigated by FTIR measurements on AuTiDeg samples,
The effect of oxygen on pre-adsorbed alcohol is already
evident at room temperature (Fig. 2, dotted and fine lines a
and b). In fact, the oxidation reaction of the alcohol to
ketone is confirmed by the presence of the band at
1670 cm-1, related to the C=O stretching and another at
1281 cm-1 which is typical of C–C phenyl–carbonyl
stretching. Interestingly, the C=O stretching band is split
into two components that are well noticeable in the spec-
trum of alcohol/toluene–O2 interaction (fine line b).
In this case, the comparison of FT-IR spectra of the
acetophenone liquid film and of the acetophenone adsorbed
on Au/TiO2 (Fig. S4, section b) indicated that the ketone
molecule is adsorbed onto the support surface, giving rise
to a complex band in the carbonyl stretching region, while
no interaction between the ketone and gold sites is evident
as confirmed by the spectra of acetophenone adsorbed on
P25 (see Fig. S6 in Supplementary Material).
following the operative setup previously reported. This
mild thermal pretreatment is necessary, in order to have the
catalyst surface adequately clean for FTIR analyses.
In addition, DRUV–Vis spectroscopy and XRD analyses
were used to have information on the effect of the thermal
treatment on the optical and structural properties of the
catalyst (see Fig. S3 and S4 in Supplementary Material).
DRUV–Vis spectra (Fig. S3, section A) showed that
AuTiDeg and the ‘‘as received’’ Au/TiO
are equal.
2
Moreover (Fig. S4), XRD pattern clearly put in evidence
that no structural modifications, both the titania and of gold
nanoparticles, occurred after the different thermal treat-
ment as well as after the reactions.
The spectra collected after 1-phenylethanol-O2(a) and
1-phenylethanol/toluene-O2(b) interactions on the AuTi-
Deg surface are presented in Fig. 2. For the sake of clarity,
we highlight 1750–1000 cm-1 range, where the absorption
of C=O and C–OH groups is located.
The spectra collected at r.t., after heating at 80 °C in the
presence of alcohol–O2 and at 110 °C in alcohol/toluene–
O2 mixtures (Fig. 2, dashed lines, a and b, respectively),
show an increase of the band intensity related to ace-
tophenone and a simultaneous decrease of the band
intensity related to 1-phenylethanol, e.g., d (C–OH), and m
(C–O), at 1206 cm-1 and 1150–1075 cm-1, respectively.
These further confirm the reactivity of the catalyst toward
alcohol oxidation.
The adsorption of 1-phenylethanol onto the catalyst
surface (solid lines, a and b) is confirmed by the presence
of its typical IR bands: m (C=C), i.e., 1600 and 1570 cm-1
(8a and 8b stretching mode), 1490 and 1450 cm-1 (19a and
19b stretching mode); d (CH3), i.e., 1370 cm-1; d (C–OH),
i.e., 1206 cm-1; m (C–O), i.e., 1150–1075 cm-1
.
To study the interaction between the alcohol alone and
the sites exposed at the catalyst surface, FT-IR analyses of
1-phenylethanol adsorbed on AuTiDeg and of the liquid
film were carried out. The results (see Fig. S5, section a in
Supplementary Material) highlight that the alcohol inter-
acts, via H-bonding, only with the OH groups, exposed at
the titania surface.
Starting from these results, with the aim to investigate
the nature of the active sites and the role played by the
oxygen, FT-IR investigations have been performed on Au/
TiO2 submitted both to an oxidizing (in O2 at 400 °C) and
to a reducing (in H2 at 200 °C) treatment. These thermal
treatments produce Au sites with different oxidation states
123

Aerobic oxidation of alcohols on Au/TiO2 catalyst…
395
Fig. 2 FT-IR spectra of 1-phenylethanol–O2 interaction (a) and of
1-phenylethanol/toluene–O2 interaction (b) on AuTiDeg surface.
Solid lines adsorbed alcohol or alcohol/toluene. Dotted lines after O2
inlet. Fine line after 30 min of O2 interaction. Dashed lines mixture
heated at 80 °C (a) and at 110 °C (b)
Fig. 1 a FT-IR spectra of CO adsorbed at -153 °C on AuTiRed
(solid line), AuTiOx (fine line), and AuTiDeg (dotted line). b FT-IR
spectra of CO adsorbed at -153 °C, after 1-phenylethanol-O2
interaction on AuTiRed (bold line) and AuTiOx (dashed line). Fine
line spectrum of CO adsorbed on AuTiRed after the interaction
between alcohol and pre-adsorbed oxygen
present in both samples. This is congruent to the formation
of water, which remains adsorbed on the catalyst surface.
In addition, a new peak, typical of OH bridged groups,
appeared at 3696 cm-1. Water formation in the oxidation
reaction of cyclohexanol and 2-cyclohexen-1-ol on
Au(111) was confirmed by Friend et al. [31], due to the
extraction of the two H atoms of the alcohol by Au–O
species.
(Fig. 1a). We, therefore, wish to stress that, despite the
catalyst is not activated in this way before the catalytic test,
this FT-IR study allows us to explore the active sites
behavior when Au0 or Aud? sites are exposed on the cat-
alyst surface.
The FT-IR experiments were undertaken at r.t. with a
vapor pressure of 0.5 mbar of 1-phenylethanol. In Fig. 3a,
the spectra collected after 10 min of interaction between
the alcohol and the differently treated catalyst are reported.
As reported above, the adsorption of 1-phenylethanol
onto the catalyst surface is confirmed by the presence of
typical IR bands, i.e., m (C=C) (1495 and 1450 cm-1), d
(CH3) (1370 cm-1), d (C–OH) (1205 cm-1) and m (C–O)
If we consider that Au nanoparticles are supported on
TiO2 and that the water formation also occurs on the
reduced catalyst without the presence of O2, it should be
clear that the support has a direct role, as reported by
Gonzalez-Yanez et al. [32] in their study of the surface
reactions of ethanol on an Au/TiO2 catalyst.
However, to have experimental data on the mechanism
of water formation on the surface of the catalyst used in
this study, the same IR experiment was undertaken on the
bare P25 support reduced in H2 at 200 °C, and on reduced
and deuterated P25 (see Fig. S7 in Supplementary
(1120–1075 cm-1),
m
(CH2)
and
m
(CH3;
3100–2860 cm-1). Interestingly, a very large absorption
centered at 3300 cm-1 and a band at 1630 cm-1 are
123

396
F. Vindigni et al.
spectroscopy and spin trapping techniques. In addition, the
authors pointed out that alcohol oxidation proceeds via a
radical mechanism.
The absence of a band related to C=O stretching in the
spectra reported in Fig. 3a, clearly shows that the interac-
tion of 1-phenylethanol, at r.t., with the oxidized or reduced
catalyst does not produce any oxidation product, in our
experimental conditions.
Most probably, in our conditions, the oxidation reaction
does not follow the Au–H intermediate formation pathway.
Indeed oxygen is needed for alcohol oxidation, as
demonstrated when we performed the catalytic test under
nitrogen flow (Table 1; entry 6). Therefore, the effect of
oxygen on pre-adsorbed 1-phenylethanol was investigated
at r.t., and the results are shown in Fig. 3b.
The first effect is the intensification of the bands related
to 1-phenylethanol across the whole range. This manifes-
tation can be attributed to the increase of the pressure
inside the IR cell, due to the inlet of O2 on pre-adsorbed
alcohol. The alcohol oxidation reaction on the catalyst
surface is confirmed by the presence of the band at
1657 cm-1, related to C=O stretching and another at
1281 cm-1, typical of phenyl–carbonyl C–C stretching, in
the AuTiRed spectrum (solid line).
Acetophenone formation is the important proof that
oxygen is needed for the alcohol oxidation in our case,
because Au–O species are the active sites promoting the C–
H bond scission, in accordance with the reaction mecha-
nism proposed by Friend and co-workers. Considering both
this latter statement and the results of the FTIR experiment
with CO at low temperature on AuTiOx (Fig. 1a, fine line)
that revealed the presence of Au-O species located at
metal-support interface, we would expect the ketone for-
mation on this sample, even before the O2 inlet.
Surprisingly, however, acetophenone formation is inhibited
on the oxidized catalyst.
Fig. 3 FT-IR spectra of 1-phenylethanol adsorption on AuTiRed
(solid line) and AuTiOx (dashed line). a After the interaction with
0.5 mbar 1-phenylethanol.
1-phenylethanol
b After 30 mbar O2 on adsorbed
Material). Interestingly, the frequency of the adsorbed H2O
bands and of bridged OH groups on Au/TiO2 is shifted
10 cm-1 to higher frequency as compared to those
observed on bare P25. This may be an indication of
adsorbed sites with different chemical surroundings (sup-
port sites in close contact with Au nanoparticles, for
example).
Green and co-workers [35] have recently reported that
the oxidation reactions are catalyzed by the dissociation of
O2 on borderline sites on Au/TiO2 catalyst (Au–O–Ti)
which involve both Au and Ti4? centers, located at the
interface between the Au nanoparticles and the TiO2 sup-
port. The same authors [36] have also highlighted, in the
case of CO oxidation, a significant reaction inhibition when
the Au/TiO2 catalyst is pre-oxidized. In this case, both their
theoretical DFT calculations and experimental results
revealed that chemisorbed oxygen atoms induce a positive
charge in the adjacent Au atoms, doubling the activation
energy for CO oxidation at the borderline sites.
Moreover, considering that this reactivity is totally
absent on AuTiDeg (Fig. 2a, b, solid lines), and that linear
OH groups are involved in H-bond with residual water on
this sample (Fig. S1, section a), it is possible to conclude
that water formation is due to the reaction of alcohol with
these linear hydroxyls.
Actually, the mechanism of gold-catalyzed alcohol
oxidation is not completely clear. Friend and co-workers
have reported the importance of the Au–O species, [27, 31]
also in other reaction types [33]. On the other hand, Corma
et al. [17] have proposed Au–H intermediate formation as
the key step in aerobic alcohol oxidation, limiting the role
of oxygen to the regeneration of the catalysts. This is also
asserted by Conte et al. [34] via the combined use of EPR
Hong and Rahman [37] have recently reported the high
activity of interfacial sites of the Au13/TiO2 (110) model
system in methanol decomposition. Their theoretical
studies have highlighted that this high activity is closely
related to the important charge-transfer-induced Coulomb
123

Aerobic oxidation of alcohols on Au/TiO2 catalyst…
397
interaction between the gold and the support in the Au–O–
Ti dual perimeter sites. Therefore, the formation of a strong
O–Au bond between gold and methoxy occurs in such
sites, leading to a weakening of the molecular C–O bond
and to a repulsive interaction between the partially positive
gold and the carbon atom bound to the oxygen. These
effects contribute to enhance the reactivity of methoxy via
the extraction of H atom, toward CH2O formation.
With our experimental results and these literature
information in mind, we can assert that these sites, located
at the interface between the Au nanoparticles and the TiO2
support, play a crucial role in acetophenone formation. In
particular, the lack of oxidation product on the AuTiOx
sample, both in the presence and in the absence of oxygen,
can be explained by the existence of an Au-O bond on
borderline sites (Fig. 1a, fine line) that makes them not
available for the activation of the 1-phenylethoxy species.
Therefore, the oxidation reaction is inhibited on the pre-
oxidized samples.
heterogeneous catalysis. We demonstrated (see Tables A
and B in Supplementary Material) that Au/TiO2 can be
reused up to the sixth consecutive run without any loss in
its catalytic effectiveness. However, a study of the surface
site situation after a catalytic reaction can be useful to
make considerations about the deactivation process and, as
a consequence, to set up a cleaning procedure to recycle the
catalysts.
With this in mind, FT-IR experiments of CO adsorbed at
low temperature after the 1-phenylethanol–O2 interaction
were undertaken as a first step, both on AuTiRed and on
AuTiOx that had been simply outgassed for 30 min at r.t. It
is important to stress that outgassing at r.t. is not sufficient
to completely eliminate both ketone and alcohol, because
of their strong interaction with support surface sites.
In Fig. 1b, the FT-IR spectra of AuTiRed (bold line) and
of AuTiOx (dashed line) are reported. Moreover, the
spectrum of adsorbed CO on AuTiRed (fine line) after the
interaction between the alcohol and the pre-adsorbed
oxygen is also shown for comparison.
To further sustain this statement, the oxidation reaction
was repeated on AuTiRed, and the effect of 1-pheny-
lethanol at r.t. on pre-adsorbed O2 was studied. As
previously reported in the literature [38], the effect of
oxygen is to re-oxidize a reduced Au/TiO2 catalyst, even at
low temperature. In our case, the spectrum collected after
the inlet of alcohol on pre-adsorbed oxygen (see Fig. S8 in
Supplementary Material) is quite similar to what was
observed on AuTiOx (Fig. 3b, dashed line), thus confirming
the negative effect of oxygen chemisorbed on Au atoms
before alcohol adsorption.
Firstly, the most interesting feature here is that all the
spectra are quite similar, showing the same surface site
distribution after the reaction, despite having different
reactivity. Moreover, this also occurs when the order of
reactants is reversed, as shown by the spectrum of CO
adsorbed on AuTiRed after the interaction between alcohol
and pre-adsorbed oxygen (fine line).
In addition, a comparison with the spectra of adsorbed
CO, collected before the 1-phenylethanol-O2 interaction on
AuTiRed (Fig. 1a, solid line) and on AuTiOx (Fig. 1a, fine
line) shows that:
At this point, it is important to stress that, if the oxi-
dation of 1-phenylethanol was carried out in conditions
close to those used for FTIR experiments (i.e., in oxygen
flow in absence of the base), the alcohol was not com-
pletely oxidized (see Table 1; entry 9) even when bubble a
continuous oxygen flow through the reaction. Therefore,
the addition of a strong base (NaOH) was necessary to
obtain the complete oxidation of 1-phenylethanol to ace-
tophenone (Table 1; entry 10).
1. a decrease in the intensity of the band related to CO–
Au sites interaction occurs in the AuTiRed spectrum
(Fig. 1b, bold line). As we have FTIR evidence that
both ketone and alcohol are not adsorbed onto Au
sites, we can tentatively explain this feature with the
presence of residual acetophenone and 1-phenyletha-
nol adsorbed onto the support surface near the Au
sites, which makes the diffusion of CO molecules
towards the gold sites difficult.
In the light of this, we can assert that: (1) since ace-
tophenone forms at the beginning of the oxidation reaction,
the sites needed for alcohol activation are available on
catalyst surface; (2) since the complete oxidation of
1-phenylethanol does not occur, many of these sites are
depleted in some way. In fact, only adding NaOH the
oxidation reaction ends.
2. the band due to oxidized gold sites, i.e., in the
2130–2120 cm-1 range, disappears in the AuTiOx
spectrum (Fig. 1b, fine line), indicating that there is no
more oxygen bounded on the Au borderline sites, even
if FT-IR measurements did not show reactivity on this
sample.
We can tentatively explain this as follows: since the OH
groups, present on titania, are depleted after the H alcohol
group extraction (thus producing water in the first step), O–
H bond scission can be promoted by further interaction
between the base and the hydroxyl groups.
A further validation of this latter point can be found in
FTIR spectra of the alcohol oxidation being carried out
again on the ‘‘used’’ AuTiOx (see Fig. S9 in Supplementary
Material). The presence of a band at 1658 cm- 1, due to
ketone C=O stretching after the adsorption of 1-pheny-
lethanol can be considered evidence of the simultaneous
The stability and, if carried out, the recycling of cata-
lysts are two of the main topics in the field of
123

398
F. Vindigni et al.
presence of ‘‘free’’ Au at the Au–TiO2 interface, where the
activation of the alcohol occurs, and residual active oxygen
adsorbed on Au sites, located on non-borderline positions.
Moreover, the effect of a new O2 inlet on the pre-adsorbed
alcohol is to further improve the oxidation reaction, as
shown by the intensity increase in the C=O stretching band.
Therefore, the question is: how has the oxygen been
removed from the Au borderline sites?
In the light of these, we can assert that, under our
experimental conditions, alcohol group H extraction is
carried out by OH groups and O2-basic sites present on
titania and the product water was adsorbed onto the support
surface. The formed alkoxy species is then activated on Au
sites (Au–O–Ti), located at the interface of the Au
nanoparticles and the TiO2, and C–H bond scission is
promoted by the active oxygen species (Au–O), generated
by O2 dissociation on Au nanoparticles. Scheme 1 shows a
withdrawing of the reaction steps.
Coming back to our AuTiOx experimental results, it is
worth noting that FT-IR experiment of adsorbed 1-pheny-
lethanol (Fig. 3a, dashed line) has shown water production,
adsorbed onto support sites in close contact with Au
nanoparticles. The interaction between co-adsorbed H2O
and oxygen has been highlighted by theoretical studies [39,
40], which have revealed the formation of co-adsorbed
complexes which are strongly bound to gold clusters. In
addition, other works have reported an increase in oxygen
mobility on reducible supports in the presence of surface
water [41, 42]. The promotion effect of water on the cat-
alytic activity of Au/TiO2 has also been demonstrated by
Yang et al. [19] in the selective oxidation of benzyl alcohol
to benzaldehyde and by Aghaei and Berger [43] in the
selective oxidation of aqueous ethanol solution. These
results clearly demonstrate the strong interaction between
water and oxygen adsorbed onto the catalyst surface.
Considering our experimental data, along with the relevant
literature information, we tentatively explain the disap-
pearance of the oxygen bound onto Au borderline sites as a
result of an interaction with adsorbed H2O that probably
promotes oxygen spill-over on the support surface.
Oxidation reaction of benzylic (1e-1i) and primary
alcohols (1j-1k)
Benzylic and primary alcohols were also reacted to further
expand the scope of our work. The oxidation of benzyl
alcohol (1e) was initially carried out in the same conditions
used for the oxidation of 1-phenylethanol (1a). As already
observed for 1a, oxidation did not take place without a base
(Table 3; entry 1) and NaOH was, therefore, added to the
reaction mixture.
Unfortunately, we only observed the total decomposi-
tion of 1a (Table 3, entry 2), probably due to the
interaction of the strong base with relatively acidic
methylenic protons. K2CO3 was, therefore, used (Table 3,
entry 3) and, interestingly, benzaldehyde (3a) was obtained
in good yields as well as by-product benzyl benzoate (4a).
Surprisingly, K2CO3 is able to deprotonate the very
weak acid 1e despite it being a much weaker base than
NaOH. Since we also observed that the activation of the
Scheme 1
(a)
(b)
(c)
123

Aerobic oxidation of alcohols on Au/TiO2 catalyst…
Table 3 Oxidation of benzylic alcohol 1e
399
Entry
Solvent
Base
T/°C
Time/h
Yield/% of 3a
Yield/% of 4aa, b
c
1
2
3
4
a
Toluene
Toluene
Toluene
–
–
110
110
110
80
24
24
15
20
–
–
d
–
NaOH (3 eq.)
K2CO3 (6 eq.)e
K2CO3 (6 eq.)g
–
83f
17
97f, h
–
All the reactions were carried out with 1 mmol of 1e, 3 or 6 mmol of base, and 50 mg of Au/TiO2 (equal to 15 mol % of Au)
Yields refer to the pure and isolated 3a or 4a
b
c
d
e
Unreacted 1e was only recovered
Total decomposition of 1e was observed
Using Na2CO3, the yield of 3a was lower (60 %). Also using 3 mmol of K2CO3, the yield was considerably lower (18 %); f the conversion was
total
g
With lesser amounts of base, the reaction was not complete
h
The reaction carried out at 110 °C gave lower yield of 4a (52 %)
alcohol on Au sites (Au–O–Ti) located at the Au–TiO2
interface is essential, we can assume that this interaction
facilitates proton extraction by K2CO3 by weakening the
OH bond of the alcohol group. Moreover, proton extraction
from alcohol groups was also performed by OH groups and
O2-basic sites present on titania, producing water adsorbed
on the support surface, as we have demonstrated.
According to these observations, we can consider that the
combined action of TiO2 and K2CO3 was essential to ini-
tializing the oxidation reaction of 1e.
These results suggest that benzyl benzoate formation
follows the pathway reported in the Scheme 3; the alco-
holate is adsorbed onto Au borderline sites, while the
benzaldehyde interacts through the carbonyl group with the
Ti4? sites neighboring the gold nanoparticles. This inter-
action makes the carbonyl group more available to the
nucleophilic attack of the alcoholate.
After optimizing the reaction conditions, other benzylic
alcohols and primary alcohols were reacted. The results are
shown in Table 4. First of all, aldehydes 3b–3f (Table 4;
entries 1, 3, 5, 7) or esters 4b–4d, 4f (Table 4; entries 2, 4,
6) were selectively obtained from benzylic alcohols 1f–1k,
depended on the presence or the absence of the solvent,
respectively. Moreover, it is interesting to note that the
different electronic effects of the substituent bound to the
aromatic ring did not influence the progress of the reaction.
Corma et al. [17] have proposed a reaction mechanism in
which the breaking of benzylic CH bond was faster than
the formation of the CO p bond, so that a partial positive
transition state was assumed.
Taking into account these observations, we propose the
benzaldehyde formation reaction pathway as shown in
Scheme 2 (most probably the steps showed in the
scheme are simultaneous). Interestingly, only 4a was
obtained in excellent yield (98 %) when the reaction was
performed without toluene (Table 3, entry 4). It is rea-
sonable to assume the formation of 4a due to the
alcoholate’s nucleophilic attack on the aldehyde. This
reaction may be promoted by the activation of the ben-
zaldehyde carbonyl group on the catalyst surface.
To validate this hypothesis, we have studied the
nature of the interaction between the benzaldehyde and
the catalyst by FTIR spectroscopy. The FTIR spectra
of the aldehyde adsorbed on AuTiDeg and on P25 (see
Fig S10 in Supplementary Material) indicate that the
molecule is adsorbed onto the support sites, without
any evidence of an interaction between the benzalde-
hyde and gold sites.
Their hypothesis was confirmed by the fact that benzylic
alcohols bearing electron donating groups were oxidized
faster than those bearing electron-withdrawing groups.
However, the results reported in Table 4 certainly confirm
that no cationic or partially cationic intermediate is feasible
in our case, since the formation of the CO group p bond
and the breakage of CH bond are almost simultaneous
(Scheme 1b).
123

400
F. Vindigni et al.
Table 4 Oxidation of benzyl 1f–1i and primary alcohols 1j, 1k
Entry
R
Solvent
T/°C
Time/h
Yield/% of 3
Yield/% of 4a, b
1
4-MeC6H4 (1f)
4-MeC6H4 (1f)
4-MeOC6H4 (1g)
4-MeOC6H4 (1g)
4-ClC6H4 (1h)
4-ClC6H4 (1h)
4-NO2C6H4 (1i)
4-NO2C6H4 (1i)
C6H5CH=CH (1j)
nC5H11 (1k)
Toluene
–
110
80
48
24
24
24
24
24
24
24
24
72
48
92c (3b)
Traces (4b)
87c, d (4b)
Traces (4c)
89c, d (4b)
Traces (4d)
88c, d (4b)
2
–
3
Toluene
_
110
80
95c (3c)
4
–
5
Toluene
110
80
89c (3d)
6
–
–
7
Toluene
–
110
80
60c (3e)
–
c, e
8
–
–
–
–
(4e)
9
Toluene
Toluene
–
110
110
80
93c (3e)
10
11
a
–
–
nC5H11 (1k)
90c, d (4f)
All the reactions were carried out with 1 mmol of 1, 6 mmol of K2CO3, and 50 mg of Au/TiO2 (equal to 15 mol % of Au)
b
c
d
e
Yields refer to the pure and isolated 3 or 4
The conversion was total
The reactions carried out at 110 °C gave lower yield of 4
GC and GC–MS analyses of the crude residue showed the presence of 4e (m/z = 302, M?). However, it was not possible to isolate it with
adequate purity
Scheme 2
FT-IR characterization
Moreover, the alcohol oxidation reaction was chemos-
elective, and the cinnamyl alcohol (1j) double bond
remained unchanged at the end of the reaction (Table 4;
entry 9). In addition, the oxidation of hexan-1-ol (1k)
allowed us to obtain 4f only (Table 4; entry 11). In fact, the
reaction that was carried out in toluene failed (Table 4;
entry 10). It is likely that hexanal (3g) spontaneously
decomposed in these conditions. We excluded its further
oxidation because we did not recover the corresponding
carboxylic acid; its thermal decomposition seemed more
probable [44].
To understand why the presence/absence of the solvent
influences the selectivity of the oxidation reaction, FT-IR
experiments of toluene adsorbed were undertaken at r.t. on
AuTiDeg. The C–H stretching mode region of the spectra
is shown in Fig. 4a.
The spectrum of toluene adsorbed onto the catalyst at r.t.
(bold curve) shows multiple bands in the 3100–3000 cm-1
region attributable to aromatic C–H stretching vibrations,
while the methyl group has typical bands near 2925 and
123

Aerobic oxidation of alcohols on Au/TiO2 catalyst…
401
Fig. 4 a FT-IR spectra of toluene adsorbed on AuTiDeg at r.t. (bold
curve), heated at 110 °C (dotted curve) and outgassed 30 min at r.t.
(fine curve). b FT-IR spectra of CO adsorbed at -153 °C on AuTiDeg
before the interaction with toluene (fine curve), after the interaction
with the solvent at r.t. (dotted curve) and after heating in toluene at
110 °C (bold curve)
2870 cm-1 both attributable to the symmetric CH3
stretching mode.
Conclusions
After heating in toluene at 110 °C (dotted curve), no
difference was observed in the aromatic stretching region
while the bands in the aliphatic stretching one show an
intensity increase and, more interestingly, they are irre-
versible to outgassing (fine curve).
From the synthetic point of view, we proposed here an
efficient and easy method to oxidize primary, benzylic, or
secondary alcohols on a commercial Au/TiO2 catalyst in
the presence of atmospheric oxygen under relatively mild
and neat conditions.
SincethesameexperimentcarriedoutonP25didnotreveal
any change in the spectrum (data not shown for the sake of
brevity), these featuresprovide evidenceofgoldnanoparticles
having an effect on toluene molecules. In particular, the peak
at 2855 cm-1 can be attributed to the symmetric stretching
vibration of CH2 groups. Even if it is not easy to associate
these multiple bands to a specific compound, their relative
intensity is comparable to that observed in the spectrum of
methylcyclohexane[45].The effectoftheinteractionbetween
toluene and the surface sites was investigated using CO as a
probe molecule, adsorbed at low temperature. A comparison
of the spectra collected before and after the interaction with
the solvent (Fig. 4b) has revealed a decrease in intensity of the
band related to the CO–Au (2100 cm-1) and CO–Ti4?
interaction (2180–2150 cm-1).
FT-IR investigations highlighted both the role of bor-
derline sites (Au–O–Ti), located at the Au nanoparticles-
TiO2 support interface, in the activation of the alcohoxy
species, and the role of the active oxygen species (Au–O)
which is generated by O2 dissociation on Au nanoparticles,
in C–H bond scission and demonstrated that, in our experi-
mental conditions, ‘‘free’’ Au sites, located at Au–TiO2
interface, and Au–O species are necessary to perform the
selective oxidation of the alcohols on commercial Au/TiO2.
Experimental section
The catalyst used in this study was a commercial AUR-
OliteTM sample (Au/TiO2) purchased from STREM
Chemicals (please see: http://www.strem.com/catalog/v/79-
0165/25/gold_7440-57-5). The sample was prepared to
support gold nanoparticles, on TiO2 (P25 Degussa) extru-
dates, with a specific surface area (SSA) of 40–50 m2/g. The
gold loading is 1 % wt and the catalyst was pretreated by the
manufacturer to ensure that Au particle average size was
2–3 nm. Furthermore, this catalyst was characterized and
tested in other catalytic reactions, such as CO oxidation [46].
The catalyst was used ‘‘as received’’, without any pre-
treatment. Analytical grade reagents and solvents were
used and reactions were monitored by GC and GC–MS.
Petroleum ether refers to the fraction boiling in the range
In the light of this, we assume that toluene and aliphatic
compounds (derived from it) are adsorbed onto the support
surface near the Au sites which makes the diffusion of the
CO molecules towards the gold sites difficult.
These findings can explain the different selectivity that
occurs in the presence or absence of toluene. Indeed, the
presence of compounds adsorbed on Ti4? sites neighboring
to the gold nanoparticles makes these sites not available for
the activation of the carbonyl group of the aldehyde (as
shown in the Scheme 3). This means why benzaldehyde is
the main product of the benzyl alcohol oxidation in the
presence of the solvent.
123

402
F. Vindigni et al.
Scheme 3
H
C
2
O
O
H
C
O2
H
H
C
O
C
O
O
O
H
O
O2
H
O
O
O
O
O
O
u
A
u
A
u
A
O2
u
u
O Ti O A
u
A
Ti
A
O
O
Ti
O
Ti
O2
Ti
O2
Ti
O2
8. Claus P, Bruckner A, Mohr C, Hofmeister H (2000) J Am Chem
Soc 122:11430
9. Gonzalez-Arellano C, Corma A, Iglesias M, Sanchez F (2006) J
Catal 238:497
40–70 °C. All reagents were purchased from Sigma-
Aldrich. Unless otherwise stated, all the reactions were
performed in opened air flasks.
The structures and purity of the products of the oxida-
tion of secondary alcohols, namely acetophenone (2a),
benzophenone (2b), and cyclohexanone (2c), were con-
firmed by comparison of their physical and spectral data
with those of a commercial sample (Sigma-Aldrich). Also
the structures and purity of the products of the oxidation of
benzylic and primary alcohols, namely benzaldehyde (3a),
4-methylbenzaldehyde (3b), 4-methoxybenzaldehyde (3c),
4-chlorobenzaldehyde (3d), 4-nitrobenzaldehyde (3e),
trans-cinnamaldehyde (3f), benzyl benzoate (4a), and
hexyl hexanoate (4f), were confirmed by comparison of
their physical and spectral data with those of a commercial
sample (Sigma-Aldrich); 4-methylbenzyl 40-methylben-
zoate (4b) [47], 4-methoxybenzyl 40-methoxybenzoate (4c)
[47], 4-chlorobenzyl 40-chlorobenzoate (4d) [48] with
those reported in the literature.
1H NMR spectra were recorded on a Bruker Avance 200
spectrometer at 200 MHz. Mass spectra were recorded on
an HP 5989B mass selective detector connected to an HP
5890 GC with cross-linked methyl silicone capillary col-
umn. High purity grade gases (H2, CO, O2) were used for
sample pretreatments and FTIR experiments. Sample
characterization, FTIR experiments details and synthetic
procedures of the catalytic tests are reported in Supple-
mentary Material.
10. Gonzalez-Arellano C, Abad A, Corma A, Garcia H, Iglesias M,
Sanchez F (2007) Angew Chem Int Ed 46:1536
11. Carrettin S, Concepcion P, Corma A, Nieto JML, Puntes VF
(2004) Angew Chem Int Ed 43:2538
12. Han J, Liu Y, Guo R (2009) J Am Chem Soc 131:2060
13. Hashmi ASK, Hutchings GJ (2006) Angew Chem Int Ed 45:7896
14. Corma A, Garcia H (2008) Chem Soc Rev 37:2096
15. Zhang Y, Cui X, Shi F, Deng Y (2012) Chem Rev 112:2467
16. Dimitratos N, Lopez-Sanchez JA, Hutchings GJ (2012) Chem Sci
3:20
17. Abad A, Corma A, Garcia H (2008) Chem Eur J 14:212
18. Bianchi C, Porta F, Prati L, Rossi M (2000) Top Catal 13:231
19. Yang X, Wang X, Liang C, Su W, Wang C, Feng Z, Li C, Qiu J
(2008) Catal Comm 9:2278
20. Martinez-Gonzalez S, Gomez-Aviles A, Martynyuk O, Pestrya-
kov A, Bogdanchikova N, Cortes Corberan V (2014) Catal Today
227:65
21. Zhao J, Liu H, Ye S, Cui Y, Xue N, Peng L, Guo X, Ding V
(2013) Nanoscale 5:9546
22. Villa A, Chan-Thaw CE, Veith GM, More KL, Ferri D, Prati L
(2011) ChemCatChem 3:1612
23. Villa A, Wang D, Veith GM, Vindigni F, Prati L (2013) Catal Sci
Tech 3:3036
24. Costa VV, Estrada M, Demidova Y, Prosvirin I, Kriventsov V,
Cotta RF, Fuentes S, Simakov A, Gusevskaya EV (2012) J Catal
292:148
25. Estrada M, Costa VV, Beloshapkin S, Fuentes S, Stoyanov E,
Gusevskaya EV, Simakov A (2014) Appl Catal A-Gen 473:96
26. Simakova OA, Smolentseva E, Estrada M, Murzina EV,
¨
Beloshapkin S, Willfor SM, Simakov AV, Murzin DY (2012) J
Catal 291:95
27. Baker TA, Liu X, Friend CM (2011) Phys Chem Chem Phys
13:34
28. Panayotov DA, Burrows SP, Yates JT Jr, Morris JR (2011) J Phys
Chem C 115:22400
Acknowledgments This work was supported by the University of
`
Torino and by Ministero dell’Universita e della Ricerca.
29. Hadjiivanov K, Lamotte J, Lavalley JC (1997) Langmuir 13:3374
30. Boccuzzi F, Chiorino A, Tsubota S, Haruta M (1996) J Phys
Chem 100:3625
References
31. Liu X, Friend CM (2010) Langmuir 26:16552
32. Gonzalez-Yanez EO, Fuentes GA, Hernandez-Teran ME, Fierro-
Gonzalez JC (2013) Appl Catal A-Gen 464:374
33. Xu B, Haubrich J, Freyschlag CG, Madix RJ, Friend CM (2010)
Chem Sci 1:310
1. Haruta M (2005) Nature 437:1098
2. Hayashi T, Tanaka K, Haruta M (1998) J Catal 78:566
3. Haruta M, Date M (2001) Appl Catal A-Gen 222:427
4. Prati L, Rossi M (1998) J Catal 176:552
5. Okumura M, Akita T, Haruta M (2002) Catal Today 74:265
6. Piccolo L, Piednoir A, Bertolini JC (2005) Surf Sci 592:169
7. Bailie JE, Hutchings GJ (1999) Chem Comm 2151
34. Conte M, Miyamura H, Kobayashi S, Chechik V (2009) J Am
Chem Soc 131:7189
123

Aerobic oxidation of alcohols on Au/TiO2 catalyst…
403
35. Green IX, Tang WJ, Neurock M, Yates JT Jr (2011) Science
333:736
43. Aghaei P, Berger RJ (2013) Appl Catal B-Env 132:195
44. Ho SK (1963) Proc R Soc London Ser A 276:278
36. Green IX, Tang WJ, McEntee M, Neurock M, Yates JT Jr (2012)
J Am Chem Soc 134:12717
37. Hong S, Rahman TS (2013) J Am Chem Soc 135:7629
38. Boccuzzi F, Chiorino A, Manzoli M, Lu P, Akita T, Ichikawa S,
Haruta M (2001) J Catal 202:256
45. Colthup NB, Daly LH, Wiberly SE (1975) Introduction to
infrared and Raman spectroscopy, 2nd edn. Academic Press, New
York, p 220
46. Walther G, Mowbray DJ, Jiang T, Jones G, Jensen S, Quaade UJ,
Horch S (2008) J Catal 260:86
39. Bongiorno A, Landman U (2005) Phys Rev Lett 95:106102
40. Liu LM, McAllister B, Ye HQ, Hu P (2006) J Am Chem Soc
128:4017
47. Simon M-O, Darses S (2010) Adv Synth Catal 352:305
48. Cronin L, Manoni F, O’Connor CJ, Connon SJ (2010) Angew
Chem Int Ed 49:3045
41. Pozdnyakova-Tellinger O, Teschner D, Kroehnert J, Jentoft FC,
Knop-Gericke A, Schlogl R, Wootsch A (2007) J Phys Chem C
111:5426
49. Saavedra J, Powell C, Panthi B, Pursell CJ, Chandler BD (2013) J
Catal 307:37
42. Teschner D, Wootsch A, Pozdnyakova-Tellinger O, Krohnert J,
Vass EM, Havecker M, Zafeiratos S, Schnorch P, Jentoft PC,
Knop-Gericke A, Schlogl R (2007) J Catal 249:318
123