Probing properties of the interfacial perimeter sites in TiOx/Au/SiO2 with 2-propanol decomposition

Article abstract of DOI:10.1016/j.apcata.2017.06.027

The decomposition of 2-propanol was studied over SiO₂, SiO₂ with an overlayer of TiO₂ (Ti/SiO₂), Au/SiO₂, and Au/SiO₂ with an overlayer of TiO₂ (Ti/[Au/SiO₂]) at 170–190 °C. There was no reaction on SiO₂. Propene was the only product on Ti/SiO₂, and its rate of formation increased proportionally with the Ti content. Acetone was the major product (selectivity 65–99%) on all Au-containing catalysts. Its rate of formation also increased with Ti loading. In addition, small amounts of propene were also formed on Ti/[Au/SiO₂] the rate of which increased with Ti loading. Characterization of the catalysts with N₂ adsorption, STEM, DR-UV-vis spectroscopy, XPS, XANES and EXAFS suggested that the Ti formed an amorphous TiO₂ overlayer on the catalyst. At high Ti loadings (4–5 wt.%), there were patches of thick porous TiO₂ layer, and some microdomains of crystalline TiO₂ could be detected. Au was present as 1–3 nm nanoparticles on all catalysts, before and after used in reaction. Only Lewis acid sites were detected based on results from pyridine adsorption, and their quantities increased with Ti loading. Based on the comparison of reaction rates, the dependence of the kinetics on 2-propanol partial pressure, the apparent activation energies, and the effect of co-feeding O₂ among different catalysts, it was concluded that propene was formed on the TiO₂ overlayer, acetone was formed primarily at the Au-TiO₂ interfacial perimeter sites, and α-C-H bond breaking preceding acetone formation was more facile on Au at the interfacial site than other surface Au atoms. Implication of these results to the selective acetone formation in the oxidation of propane in the presence of a O₂/H₂ mixture was discussed.

Full text of DOI:10.1016/j.apcata.2017.06.027

Accepted Manuscript
Title: Probing properties of the interfacial perimeter sites in
TiO /Au/SiO with 2-Propanol decomposition
x
2
Authors: Yi Y. Wu, Harold H. Kung
PII:
S0926-860X(17)30276-4
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.06.027
APCATA 16289
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
17-4-2017
19-6-2017
21-6-2017
Please cite this article as: Yi Y.Wu, Harold H.Kung, Probing properties of the interfacial
perimeter sites in TiOx/Au/SiO2 with 2-Propanol decomposition, Applied Catalysis A,
Generalhttp://dx.doi.org/10.1016/j.apcata.2017.06.027
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Probing Properties of the Interfacial Perimeter Sites in TiO
Propanol Decomposition
x
/Au/SiO with 2-
2
Yi Y. Wu and Harold H. Kung*
Department of Chemical and Biological Engineering, Northwestern University, Evanston IL,
0208, USA
6
Graphical abstract
Highlights:

Interfacial perimeter sites of Au with TiO
acetone selectively.
Interfacial perimeter sites of Au with TiO
2
2
catalyze decomposition of 2-propanol to
are much more active than Au surface.

Abstract
The decomposition of 2-propanol was studied over SiO
Ti/SiO ), Au/SiO , and Au/SiO with an overlayer of TiO (Ti/[Au/SiO
was no reaction on SiO . Propene was the only product on Ti/SiO , and its rate of formation
2
, SiO
2
with an overlayer of TiO
2
(
2
2
2
2
2
]) at 170-190 ºC. There
2
2
increased proportionally with the Ti content. Acetone was the major product (selectivity 65-99%)
on all Au-containing catalysts. Its rate of formation also increased with Ti loading. In addition,
small amounts of propene were also formed on Ti/[Au/SiO
loading. Characterization of the catalysts with N adsorption, STEM, DR-UV-vis spectroscopy,
XPS, XANES and EXAFS suggested that the Ti formed an amorphous TiO overlayer on the
catalyst. At high Ti loadings (4-5 wt.%), there were patches of thick porous TiO layer, and
some microdomains of crystalline TiO could be detected. Au was present as 1 – 3 nm
2
] the rate of which increased with Ti
2
2
2
2
nanoparticles on all catalysts, before and after used in reaction. Only Lewis acid sites were
detected based on results from pyridine adsorption, and their quantities increased with Ti loading.
Based on the comparison of reaction rates, the dependence of the kinetics on 2-propanol partial
1

pressure, the apparent activation energies, and the effect of co-feeding O
catalysts, it was concluded that propene was formed on the TiO overlayer, acetone was formed
primarily at the Au-TiO interfacial perimeter sites, and -C-H bond breaking preceding acetone
2
among different
2
2
formation was more facile on Au at the interfacial site than other surface Au atoms. Implication
of these results to the selective acetone formation in the oxidation of propane in the presence of a
O
2
/H
2
mixture was discussed.
Keywords: catalysis, 2-propanol decomposition, supported Au
1
. Introduction
It has been demonstrated that Au is an effective catalysts for various selective oxidation
reactions, such as benzyl alcohol oxidation [1], selective oxidation of styrene [2] and epoxidation
of propene and butene [3-5]. It is particularly interesting that supported Au catalysts can produce
acetone selectively from propane by oxidation with O
Oyama and coworkers that acetone was formed with 84% selectivity using a Au/TS-1 catalyst,
whereas propene was the selective product (69% selectivity) over Au/TiO [8]. However, the
2
and H
2
[6-8]. This was first reported by
2
very low conversions achieved substantially limited its commercial viability as a route for
acetone production. In order to improve the production, it is useful to identify the active site and
understand the reaction mechanism. Thus, subsequent to Oyama’s report, we demonstrated that
the interfacial perimeter site of Au with (Ti,Si)O
showed that: (i) the reaction was much slower on Au/SiO
sites; (ii) the formation of acetone was faster with a higher density of these sites, and (iii) close
proximity of (Ti,Si)O to Au was necessary for rate enhancement.
2
was essential for acetone formation [9, 10]. We
2
without these interfacial perimeter
2
The fact that Au/SiO
2
was much less active than catalysts with Au-(Ti,Si)O interfacial
2
perimeter sites implies that either activation of oxygen on these sites is much faster than on a Au
surface or the oxygen species that is catalytically active is formed preferentially at these
interfacial perimeter sites. An example of the latter would be an adsorbed hydroperoxy. Other
than these, however, there is little mechanistic information on how these sites participate in the
formation of acetone. In a mechanism proposed by Oyama [6], propane is activated by a surface
hydroperoxy group to form an adsorbed 2-propoxy, which can undergo either -C-H elimination
to form acetone, or dehydration to form propene. Since dehydration of 2-propanol to form
propene occurs readily on TiO
dehydration. On the other hand, only small amounts of acetone have been detected on TiO
2
[11, 12], an interfacial perimeter site is not necessary for
both
2
in steady state [11, 12] and temperature programmed reaction experiments [13-16]. This suggests
participation of Au in its formation. One possibility is that at the interfacial perimeter sites, 2-
propoxy is adsorbed on Ti, and Au is responsible for cleaving the α-C-H bond. If this is correct,
we expect that these interfacial sites will also be very efficient in 2-propanol decomposition to
form acetone. Indeed, there is evidence that Au-oxide interfacial perimeter sites are capable of
acetone formation. In TPD of adsorbed 2-propanol, acetone was detected from Au/TiO
or a Au(111) surface with deposited nano-TiO crystallites, while none was detected from TiO
powder [17] or a single crystal TiO surface [18]. Likewise, adsorbed acetone derived from 2-
propanol was observed by IR on Au/Al but not on Al [19].
2
powder
2
2
2
2
O
3
2
O
3
The objective of this work is to investigate the decomposition of 2-propanol on TiO
x
-
modified Au/SiO , with a focus on generating information on the reaction of 2-propoxy at the
2
2

Au-TiO interfacial perimeter sites. In particular, reaction kinetics of product formation from 2-
x
propanol decomposition was determined from which contributions from different sites on the
catalysts were derived. Effect of oxygen in the feed was also investigated in order to relate these
data to the selective formation of acetone in the oxidation of propane.
2
2
. Experimental
.1 Catalyst preparation
Au/SiO
2
The gold precursor, Au(en)
2
Cl
3
, was prepared with a procedure modified from [20].
HAuCl
4
·3H O (1 g, > 99.9%, Sigma Aldrich) was dissolved in 4 mL of deionized (DI) water
2
followed by addition of 20 mL ethanol (KOPTEC, 200 proof). After the solution was fully mixed,
ethylenediamine (0.48 mL, ≥ 99%, ReagentPlus®, Sigma Aldrich) was added and a pale yellow
precipitate was formed. The mixture was stirred for 10 minutes and then filtered and washed
with 3 portions of 100 mL ethanol. The resulting pale yellow precipitate with a yield of ~1 g was
stored in a freezer after drying under vacuum for 30 min.
Au/SiO
Cl was dissolved in 40 °C DI water to obtain a 3.3-4.4 mM Au solution. After SiO
nonporous, fumed Cab-O-Sil L90) was added to the gold solution with stirring, the heat was
2
was prepared by a modified cationic deposition method [21]. Typically,
Au(en)
(
2
3
2
turned off. An aqueous 0.75 M ethylenediamine solution was added dropwise to the mixture to
gradually adjust the PH to 9.0. After stirring the mixture at room temperature for 2 h, the
Au/SiO
2
was filtered and washed with 500 mL room temperature DI water. The solid was
) and stirred for 10 min. The mixture was
redispersed in 40 °C DI water (100 mL DI water/g SiO
2
filtered and washed again with 500 mL of DI water at room temperature until the concentration
of Cl¯ ions in the filtrate was below 30 ppm as indicated with Quantab Chloride Titrators. No
detectable Cl was observed by XPS.
TiO
x
/SiO
Ti(OiPr)
used to prepare TiO
Ti/[Au/SiO ]) by impregnation onto SiO
Ti(OiPr) (acac) that would yield approximately 1 wt.% Ti was used to prepare 1-Ti/SiO
Ti/Au/SiO , and 4 to 5 wt.% (in 3 batches of ~1 wt.% each) for 5-Ti/SiO and 4-Ti/Au/SiO
Assuming that the SiOH density was 5 SiOH/nm [22] and that each 2-propoxy on
Ti(OiPr) (acac) would react with one SiOH on the surface, 1.5 wt.% Ti would roughly form one
monolayer equivalent of TiO on the SiO
Au/TiO was prepared by deposition-precipitation. A 1.5 mM DI water solution of
HAuCl ∙3H O was adjusted to pH 8.4 with 1 M NaOH solution. Then 3 g TiO (anatase, 99.8%,
25 nm, 45-55 m /g, Sigma Aldrich) was dispersed in the gold solution. The resulting pH value
was 6.94. The suspension was filtered, washed with 600 mL DI water, and the solid was
calcined in air at 400 °C for 2 h and reduced in 5% H at 150 °C for 1 h.
All the catalysts were treated with 10% ozone in O at 300 mL/min and a thermal ramping
rate of 0.12 °C min to 150 °C to remove organic ligands before storage.
2
, TiO
x
/[Au/SiO
(acac)
/SiO
2
], and Au/TiO
(75 wt.% in 2-propanol, Sigma Aldrich) in tetrahydrofuran (THF) was
(whence labeled Ti/SiO ) and Au/SiO with TiO overcoat (labeled
and Au/SiO , respectively. An amount of
and 1-
2
2
2
x
2
2
2
x
2
2
2
2
2
2
2
2
2
.
2
2
2
x
2
.
2
4
2
2
2
≤
2
2
-1
2
.2 Catalytic reactions
The catalysts were pretreated with 100 mL min 10% H
-1
2
in He at 180 °C for 1 h to
completely reduce Au nanoparticles prior to catalytic testing (Supporting Information Figure S2).
3

All reactions were conducted under atmospheric pressure in a quartz packed bed reactor. Any
empty space in the reactor was filled with 10-16 mesh quartz chips to minimize gas-phase
reactions.
Contact time was defined as the ratio of the catalyst weight to the total flow rate of 2-
-1
propanol (W/F2PrOH), in the unit of (gCat. min mmol) . Similar contact times were achieved by
using different catalyst weights and/or adjusting the total flow rate of 2-propanol accordingly.
Conversions increased linearly with increasing contact time, verifying that the reactions were
carried out without external mass transfer limitation (Figure S3). The Weisz-Prater criteria was
-9
estimated to be 1.2x10 , implying that pore diffusion did not influence the kinetics (See
Supplementary Information and Table S1). The gas phase products were analyzed online by gas
chromatography-mass spectroscopy (Agilent 6890 GC-MSD 5973). The GC was equipped with
a Porapak-Q (80-100 mesh) packed column with a thermal conductivity detector for analysis (O
CO , H O, propane, propene, acetone and 2-propanol), and the analysis was supplemented with a
flame ionization detector and mass spectrometer coupled to a Agilent DB624 capillary column to
2
,
2
2
detect propene, acetone and 2-propanol. No CO
selectivity are defined as follows:
2
formation was observed; conversion and
Conversion
=
moles of (C
Selectivity of a C product
moles of a C product/ total moles of all C
3
products) / moles of (2-PrOH + C
3
products)
3
=
3
3
products
2
-Propanol Decomposition
All catalysts were diluted with 4 times the weight of 40-80 mesh quartz chips, whereas 7
times the weight was used for TiO and Au/TiO . 2-Propanol vapor was introduced via a flow of
2
2
He through a saturator of anhydrous 2-propanol (99.5%, Sigma Aldrich). The saturator delivered
a stream of at least 97% alcohol saturation [23]. The partial pressure of 2-PrOH were adjusted by
changing the saturator temperatures (0 or 20-23 °C) and/or diluting with a stream of He. The
range of 2-PrOH partial pressures tested was 0.6-5.6 kPa. Catalyst weight (25-200 mg) was
adjusted depending on the catalyst activity to keep conversions less than 16% for kinetic studies.
Kinetic data were collected at ~2 h time on stream when 2-propanol decomposition had reached
a steady state (conversions at ~2 h and ~5 h differed by less than 10% of its value). Carbon
balance was within 5% (see Table S1 for 4-Ti/[Au/SiO ] as an example).
2
Propanol Oxidation with O
-propanol vapor (99.5% anhydrous, Sigma Aldrich) was introduced via a flow of He in a
saturator at 20-23 °C. The feed composition was 2-propanol/O /He=1.2/5/93.8 with a total flow
rate of 30 mL min at 180 °C. Catalytic data shown were the average values of data collected at
2
2
2
-1
3
-5 h time-on-stream when the reaction reached a steady state.
2
.3
Characterization
Scintag XDS2000 diffractometer (Cu κα, 1.54 Å) was used to collect the XRD pattern of
°
°
°
°
TS-1 (Figure S4). Scans started from 5 to 50 with a step size of 0.02 at a rate of 0.04 /sec. The
metal loadings were determined by Thermo iCAP 7600 ICP-OES. Samples were digested in
small amounts of a mixture of concentrated hydrofluoric acid, nitric acid and hydrochloric acid
before dilution to a metal concentration of about 4 ppm for analysis. Standards for Ti and Au
4

were prepared with 1000 ppm ICP standard solutions (Sigma Aldrich). The relative standard
deviations (RSD) from replicate measurements for a sample were 0.9-2.6% for Au and 0.6-1.6%
for Ti.
The BET surface areas, pore volumes and pore diameters were determined by N
2
adsorption and desorption isotherms using Micrometrics 3Flex. The amount of sample for BET
analysis was ~200 mg. Complete degassing of samples at room temperature was checked by
monitoring the pressure (< 3 μmHg) over 2 min when the vacuum was temporarily turned off.
The result of BET surface area was obtained by analysis of five data points.
High angle annular dark field (HAADF) STEM images and EELS spectra at Ti L-edge
were obtained with a JEOL 2100-F transmission electron microscope (200 kV, probe size 0.7-1.0
nm) at the Electron Probe Instrumentation Center (EPIC) at Northwestern University. STEM
samples were prepared by sonicating ~3 mg of sample in 10 mL ethanol for 10 min. A drop of
the suspension was placed onto a copper grid (Ted Pella, lacey carbon, 300 mesh). The grid was
dried in the ambient environment overnight. Average particle size was calculated by counting
2
00 particles using ImageJ software (Figure S5-S8).
UV-vis spectra were collected with a Shimadzu 3600 UV-visible-NIR spectrophotometer.
Cab-O-Sil L90 was used as the reference. All samples were measured in the ambient
environment. DRIFTS of adsorbed pyridine was collected on a Nicolet Nexus 670 FTIR
spectrometer with a MCT detector and a Harrick reaction cell equipped with KBr windows.
Samples were ground into fine powders without dilution. The temperature ramping rate was 5 °C
-1
min . The background at each temperature, i.e., 180, 120 or 28 °C, was collected under a He
-1
flow rate of 25 mL min after the temperature was stabilized for 30 min. Pyridine (99.8%
-1
anhydrous, Sigma Aldrich) was introduced at 28 °C for 15 min in a He flow of 25 mL min
through a swath of quartz wool pre-wetted with 0.1 mL of pyridine. The adsorbed pyridine
spectra at each temperature were collected 30 min after the temperature was stabilized. Data
-1
were taken with 750 scans and a resolution of 4 cm , and processed by subtracting the
background. The adsorption of 2-propanol followed by pyridine adsorption was only conducted
at 180 °C.
Temperature programmed desorption (TPD) of NH
3
or acetone, and temperature
programmed reaction (TPRxn) of 2-propanol were conducted with an Altamira AMI-200
instrument equipped with a thermal conductivity detector (TCD) and a mass spectrometer
(
Universal Gas Analyzer-100). Samples containing Au nanoparticles were pretreated with 50 mL
-1
-1
min of 10% H
2
2
2
in N
00 °C for 2 h before exposed to 60 sequential pulses of 10 % NH
.43 mol of NH in each pulse. For TPD of NH , the sample temperature was ramped to 500 °C
2
at 180 °C for 1 h. All samples were dehydrated in 30 mL min of He at
3
in He at 50 °C. There were
3
3
-1
-1
at a rate of 30 °C min in 40 mL min of He and held at 500 °C for 10 min. For 2-propanol
TPRxn, a sample was pretreated with 50 mL min of 10% H
down to 40 °C. 2-Propanol was then introduced by bubbling 15 mL min of He into a saturator
at room temperature for 2 h. The sample temperature was ramped up to 500 °C at a rate of 30 °C
min in 15 mL min of N
-1
2
in N at 180 °C for 1 h and cooled
2
-1
-1
-1
2
and held at 500 °C for 6 min. TPD of acetone followed the same
procedure as TPRxn of 2-propanol, while TPD of propene was conducted by the same pulsing
method as for TPD of NH
3
. A calibration curve of known acetone concentrations versus the
5

-1
corresponding pressures of the mass spectrometer was obtained in He at 15 mL min . TPD of
-1
-1
acetone was conducted with 15 mL min of He using a temperature ramp of 30 °C min to 500
°
C, and the amount desorbed was calculated using the calibration curve.
X-ray Photoelectron Spectroscopy (XPS) was performed with a Thermo ESCALAB
2
1
50Xi equipped with an electron flood gun. The binding energies were calibrated using carbon
s (284.8 eV). The number of scans and dwell time for N 1s, Au 4f and Ti 2p were 16 and 100
ms. X-ray absorption spectroscopy (XAS) was performed at beamline 5-BMD of the Dupont-
Northwestern-Dow Collaborative Access Team (DND-CAT) located in the Advanced Photon
-4
Source, Argonne National Laboratory. A Si (1 1 1) monochromator with a resolution of 10 eV
was used. Energies were calibrated with Ti (4966.4 eV) or Au (11918.7 eV) foil. XAS was
measured at the Au L
reaction in the fluorescence mode under ambient conditions. AuCl and Au(OH)
were measured in the transmission mode. For the Ti K-edge measurements, the catalysts and TS-
3
-edge on fresh catalysts in the transmission mode and on catalysts after
3
as references
-1
-1
1
1
were pretreated in a He flow of 300 mL min and a temperature ramping rate of 10 ºC min to
80 ºC for 1 h followed by XAS fluorescence collection in He without heating, except TiO
2
anatase was measured in the ambient. XAS data were processed and analyzed by a package
software, Demeter [24]. Ti pre-edge peaks were fitted with four Gaussian peaks denoted as A
, A and B with an arctan function as a baseline (Figure S9). Theoretical models for Extended
1
,
A
2
3
X-ray Absorption Fine Structure (EXAFS) fitting were from Atoms.inp Archive [25]. Phase
correction was not applied in EXAFS figures and fitting.
3
. Results
.1. Catalyst Structure
3
BET surface areas of the catalysts are shown in Table 1. Addition of ~1 wt.% Ti caused
only small changes in the surface area, suggesting the formation of a well dispersed layer of TiO
After addition of ~5 wt.% Ti, the surface area increased by about 20% based on the weight of
SiO . The adsorption-desorption isotherms showed very small hysteresis loops (Figure 1) as
x
.
2
reported by others previously [26], typical for non-porous fume silica consisting of aggregates
and agglomerates with mesoscale gaps between 2 to 50 nm. For the higher Ti loading samples 5-
Ti/SiO
explained with a model that TiO
expected due to the relatively low temperatures employed in the catalyst preparation.
2
and 4-Ti/[Au/SiO
2
], new small pores of 3-7 nm appeared (Figure 2). These data can be
x
formed a porous overcoat on the non-porous SiO . This was
2
STEM images of Au/SiO
nm diameter, Au particles. Most importantly, addition of TiO
particle size (Table 1). Except for the presence of small Au particles, the morphology of 1-
Ti/SiO and 1-Ti/[Au/SiO ] were similar to that of SiO . However, the morphology of 5-Ti/SiO
and 4-Ti/[Au/SiO ] samples showed higher degrees of nonuniformity. There were regions where
the surface appeared to be smooth like that of SiO , and other regions that were covered with a
poorly organized structure (Figure 3). The latter was probably where a thick overlayer of TiO
was present. Indeed, EELS detected the presence of only a small Ti signal on the smooth region
2
, 1-Ti/[Au/SiO
2
], and 4-Ti/[Au/SiO
2
] showed small, mostly 1-3
x
did not change the average
2
2
2
2
2
2
x
(
(
highlighted with a square in Figure 4) but a strong Ti signal in the poorly organized region
highlighted with a circle).
6

EELS performed in STEM was used to obtain information on the crystallinity of TiO
x
in
these samples. It has been shown that the sub-splitting of the Ti L and L transitions, which
3
2
correspond to transitions from the spin-orbit coupled 2p3/2 and 2p1/2 states, respectively and
separated by ~5 eV, reflects the degree of crystallinity of the Ti environment because of the
crystal-field effect of the coordinating oxygen ions [27]. For Ti atoms in an octahedral
environment of oxygen atoms, such as in TiO
orbitals. A similar splitting can also be observed at O K-edge due to a transition to O 2p-Ti 3d
hybridized states of t2g and e symmetry. Both amorphous titanium oxide microtube and Ti
showed no sub-splitting of Ti L and L peaks or the peak at O K-edge.[28]
2
, the Ti 3d states are split into t2g (3d

) and e
g

(3d )
g
2
O
3
2
3
For sample 5-Ti/SiO
mostly to the amorphous SiO
a strong Ti L-edge peak appeared while the intensity of the O K-edge peak was greatly reduced.
2
, in the region highlighted with a square, the O K-edge peak due
2
did not show sub-splitting. In the region highlighted with a circle,
Neither the Ti nor the O peaks showed sub-splitting, suggesting that the TiO
amorphous.
x
overlayer was
Another technique used to gain information on the coordination symmetry of Ti was X-
ray absorption, using both the pre-edge features and the EXAFS region of the Ti absorption edge.
The pre-edge features of fresh and reaction used catalysts were identical. For TS-1, the pre-edge
showed a single, intense absorption peak, while TiO
). The spectra of 1-Ti/[Au/SiO ] and 4-Ti/[Au/SiO
TS-1 and anatase, with the anatase feature more prominent in 4-Ti/[Au/SiO
Figure 5), which suggested a higher fraction of O Ti in 4-Ti/[Au/SiO ] than 1-Ti/[Au/SiO
and vice versa a higher fraction of T Ti in 1-Ti/[Au/SiO ] than 4-Ti/[Au/SiO ].
2
anatase showed four weak features (Figure
] showed features that were combinations of
5
2
2
2
] than 1-Ti/[Au/SiO
2
]
(
h
2
2
],
d
2
2
It has been suggested that the Ti pre-edge absorption can be fitted with four Gaussian
peaks A , A , A and B, and the ratio of the area of the A + A peaks, which is a measure of the
1
2
3
2
3
availability of unoccupied Ti 3d orbitals, to the total area correlates with the area-weighted
average peak energy [29, 30]. Furthermore, there is a trend that a lower coordination number
showed a higher value of relative area and a lower average pre-edge energy as shown in Figure 6.
After fitting the data in Figure 5 with these four peaks (Figure S9), the results were used to
calculate the points for 1-Ti/[Au/SiO
correlation, the Ti in both 1-Ti/[Au/SiO
to 5. This suggested a mixture of T and O
features and the presence of an amorphous TiO
2
], 4-Ti/[Au/SiO
] and 4-Ti/[Au/SiO
Ti, consistent with conclusions from the preedge
overlayer.
2
], and TiO
2
(anatase) in Figure 6. Using this
2
2
] had an average coordination close
d
h
x
7

Fitting of the first shell coordination data in EXAFS for both 1-Ti/[Au/SiO
Ti/[Au/SiO ] also showed a coordination number of ~5 (Table 2 and Figure S10). However, the
Ti-O bond distance of 4-Ti/[Au/SiO ] (1.91 Å) was longer than 1-Ti/[Au/SiO ] (1.85 Å). No
obvious Ti-O-Ti shells scattering was observed for these two samples.
2
] and 4-
2
2
2
After reaction, the Au nanoparticles remained metallic on the catalysts (Figure S11) and
EXAFS fitting results of the Au edge showed similar average distances (2.84-2.85 Å) between
Au atoms on all the Au catalysts (Table 2 and Figures S12). Within experimental uncertainties,
the nearest neighbor coordination number was the same on all samples with values consistent
with the STEM results, and in agreement with literature results that Au-Au nearest neighbor
distance of 2.84-2.85 Å would have NAu-Au in the range of 8.5-10 and correspond to particle size
of 1.5-3.2 nm [31].
Finally, DR-UV-vis spectroscopy was also used to investigate the structural properties of
the catalysts, making use of the absorption in the near UV region that offers information on the
coordination environments of TiO
220 nm due to the ligand-to-metal charge transfer (LMCT) transition of isolated tetrahedral
TiO units [32, 33]. Bulk TiO anatase, where Ti is in an octahedral coordination, showed an
absorption peak at ~320 nm. The absorption peaks of 1-Ti/SiO and 1-Ti/[Au/SiO ] were at 260-
80 nm, while those of 5-Ti/SiO and 4-Ti/[Au/SiO ] were at 280-300 nm. They were in the
range for samples containing pentacoordinated Ti species, oligomeric species, and TiO
nanoclusters [32, 33]. The longer wavelengths for 5-Ti/SiO and 4-Ti/[Au/SiO ] suggested that
these samples contained higher fractions of oligomeric Ti species than 1-Ti/SiO and 1-
Ti/[Au/SiO ].
x
species. For TS-1 (Figure 7), an intense peak appeared at
~
4
2
2
2
2
2
2
2
2
2
2
2
The Au-containing samples also showed a broad peak at 520-550 nm due to the size- and
support-dependent surface plasmon vibration of Au nanoparticles [34]. For Au/SiO and 1-
Ti/[Au/SiO ], the peaks centered at ~520 nm. For 4-Ti/[Au/SiO ], the peak was at 550 nm and
broader. Assuming that the STEM and EXAFS data were accurate, that the average Au particle
size were the same for all Au sample, the shift in the plasmon peak for 4-Ti/[Au/SiO ] would be
a results of the more prominent TiO overlayer on Au.
2
2
2
2
x
3
.2.
Ti Oxidation States and Catalyst Acidity
To determine if there was a mixture of oxidation states of Ti species in the amorphous
, XPS spectra were collected on the supported Au catalysts. From the literature, the Ti 2p
TiO
x
XPS spectrum for TiO shows two peaks due to spin-orbit splitting into 2p3/2 and 2p1/2 states at
2
2
+
4
58.8 eV and 464.5 eV, respectively [35]. The binding energies of the Ti 2p3/2 peaks of TiO (Ti )
3
+
and Ti
and 4-Ti/[Au/SiO
(
2
O
3
(Ti ) are at ~456 eV and ~457 eV, respectively [28, 35]. For samples 1-Ti/[Au/SiO
], the Ti 2p3/2 and 2p1/2 peaks were at 459-459.5 and 465 eV, respectively
Figure 8). The slightly higher binding energy for the 2p3/2 peak than in the literature could be
2
]
2
instrument error, or it could be due to contribution from a small quantity of tetrahedral Ti species
8

on silica that has been assigned to a Ti 2p3/2 peak at ~460 eV [36]. Overall, the results indicated
2
+
3+
that there was little TiO (Ti ) or Ti
samples were Ti .
2
O (Ti ) in these supported Au catalysts, and the Ti in these
3
4
+
Two methods were used to probe acidity of these catalysts: adsorption of pyridine
monitored with DRIFTS [37] and adsorption and TPD of NH . No Brønsted acid sites, detected
as pyridinium ions (BPy) at ~1540 (19b) and ~1640 cm (8a) were observed on any of the
3
-1
-1
samples. Pyridine bound to Lewis acid sites (LPy) at ~1445 (19b) and ~1608 cm (8a) were
observed on catalysts containing TiO species (Figures 9, S13, and S14). We also observed a
488 cm band commonly assigned to the vibrations of both BPy and LPy [38-40], and bands
x
-1
1
due to physisorbed (PPy), hydrogen-bonded (HPy) and gas phase pyridine (GPy). Table 3
summarizes the DRIFTS results. Whereas intensities of the peaks were much stronger for the
higher Ti loading samples, the presence of Au did not have much effect. In general, the GPy,
PPy, and HPy peaks gradually decreased or disappeared as the samples were purged with He at
2
1
8 °C or by heating to 120 ºC and higher (Figures 9, S13, S14). The decrease in the HPy peak at
-1
-1
596 cm made the LPy peak at 1608 cm (8a) much more apparent. After heating to 180 °C,
, no LPy peaks were detected after heating to 120 °C or
only LPy peaks remained. On Au/SiO
80 °C.
2
1
To determine whether adsorbed 2-propanol or water could convert a Lewis acid site into
a Brønsted acid site [41], pyridine was introduced after adsorption of 2-propanol at 180 °C. The
sample cell was then purged with He, and DRFIT spectrum was collected after 15 min. Only
-1
peaks attributed to LPy at 1447 and 1606 cm were observed and no BPy peaks (Figure 10).
Thus, there should be no Brønsted acid sites present under our reaction conditions.
The amounts of NH
in Table 6. SiO adsorbed 110 mol g of NH
hydrogen bonding with SiOH groups because TPD-NH
subtracting this value from the total amount of NH
3
adsorbed from a pulse passed over the catalysts at 50 °C are shown
-1
-2
2
3
(~1.2 mol m ), probably resulting from
3
barely showed a desorption peak. By
adsorbed, then the addition of Au
3
-1
nanoparticles resulted in an additional adsorption of 119 mol g on Au/SiO . For the other
2
-1
samples, the additional amounts adsorbed were 217, 703, 288, and 1002 mol g for 1-Ti/SiO
2
,
5
-Ti/SiO
2
, 1-Ti/[Au/SiO
2
], and 4-Ti/[Au/SiO
2
], respectively. By further expressing these values
on the basis of per gram of SiO
2
, and subtracting contributions from both Au and SiO , then the
2
-
1
addition of Ti resulted in roughly 200±20 mol g SiO for adding ~1 wt.% Ti, and roughly
2
-1
8
10±100 mol g for adding 4-5 wt.% Ti.
TPD of these adsorbed NH
maximum varied slightly (Figure 11). For 1-Ti/SiO
Ti/[Au/SiO ], the maximum appeared at 190 °C, 220 °C, 175 °C and 240 °C, respectively. That
3
showed very broad peaks for all catalysts, and the peak
2
, 5-Ti/SiO , 1-Ti[Au/SiO ], and 4-
2
2
2
is, the peak maximum was at a higher temperature for the higher Ti loading samples, which
9

might be due to readsorption of desorbed NH
3
, which would be more severe for higher Ti
loading samples. Thus, overall, the samples showed similar distributions of Lewis acid strengths.
3
.3.
TPRxn of 2-Propanol and TPD of Acetone and Propene
Under the conditions of TPD experiments, there was no evidence of propene adsorption
on these catalysts, suggesting that propene interacted weakly. TPD of acetone showed a peak at
40°C on all the three catalysts containing Au (Figure 12). The activation energy for desorption
1
of acetone could be calculated from the peak temperature assuming uniform sites and a
desorption kinetic that is first-order in the surface density of acetone [42, 43]. Table 4
summarizes the calculated activation energies for acetone desorption for different values of v, the
pre-exponential factor for desorption. The ratios of moles of acetone desorbed to surface Au
were 4-6 for all the three catalysts containing Au (Table 5).
퐸푎,푑푒푠
푅푇푝2
휈
−퐸푎,푑푒푠
푅푇푝
=
exp (
),
훽
where Ea,des is the activation energy for desorption, R is ideal gas constant, T is the desorption
p
peak temperature, v is the pre-exponential factor for desorption and  is the heating rate.
TPRxn of adsorbed 2-PrOH produced different products from different catalysts. On
SiO
around 130 °C (not shown). Propene was also the only product detected on 1-Ti/SiO
Ti/SiO , and the peak maxima were at 270 °C and 230 °C respectively (Figure 13). Small
amounts of propene were observed on Au/SiO , 1-Ti and 4-Ti/[Au/SiO ] at ~140 ºC. For the
latter two samples, there were additional propene peaks at temperatures similar to those from 1-
Ti/SiO and 5-Ti/SiO . The higher temperature peak was about the same area as the lower
temperature peak on 1-Ti/[Au/SiO ], but was significantly larger for 4-Ti/[Au/SiO ].
2
, only propene evolution was observed as a broad, low intensity peak with a maximum at
2
and 5-
2
2
2
2
2
2
2
In addition to propene, acetone was also evolved on the Au-containing samples (Figure
4). On Au/SiO , acetone was evolved at 130-140°C, which was the same as the TPD peak of
acetone. Two acetone peaks were observed for both 1-Ti/[Au/SiO ] and 4-Ti/[Au/SiO ]. There
was a much smaller peak at 130-140ºC, and a much larger peak at 240-260°C.
.4. Catalytic Decomposition of 2-Propanol
1
2
2
2
3
Catalytic 2-propanol decomposition was conducted between 170-190 ºC, where
reasonable conversions and product analyses could be obtained. Table 6 shows the results at 190
ºC for all the catalysts, collected using different catalyst weights in order to maintain <16%
conversions. Results for other reaction conditions are shown in Supplementary Figure S15.
1
0

Without Au, only propene was detected on 1-Ti/SiO
rate of formation, when compared using a common basis of unit weight of SiO
roughly proportionally with Ti loading. Acetone became the predominant product on all Au-
containing catalysts. On Au/SiO , acetone selectivity reached 92%. Addition of TiO to Au/SiO
increased its activity for both acetone and propene formation. The propene formation rate was
roughly proportional to the TiO loading, but the acetone formation rate increased less rapidly.
Consequently, the selectivity for acetone decreased with increasing Ti loading. Since Au/SiO
without Ti had only a small activity for acetone formation, the data indicated that sites composed
of both Au and TiO , such as the Au-TiO interfacial perimeter sites, were responsible for the
majority of acetone formation.
2
and 5-Ti/SiO
2
as expected [44-46], and its
2
, increased
2
x
2
x
2
x
x
The dependence of 2-propanol decomposition rate on the partial pressures of 2-propanol
are shown in Figures 15. On 1-Ti/SiO and 5-Ti/SiO , where only propene was formed, the rate
of 2-propanol decomposition was rather independent of 2-propanol partial pressure. On the other
hand, both Ti/[Au/SiO ] catalysts showed a positive order dependence for propene formation on
2
2
2
propanol partial pressure. The order was positive (0.2-0.4) at lower pressures, and decreased to
near zeroth order at higher pressures (Figure 16). The order for propene formation rates on
Au/SiO was slightly higher (0.5-0.7).
2
Because acetone was the major product on the Au-containing catalysts, the dependence
of the overall 2-propanol decomposition rate and acetone formation rate on partial pressures of 2-
propanol (Figure 15 and 16) were rather similar. Au/SiO showed an order of 0.4 in both rates.
On both Ti/[Au/SiO ] catalysts, the order dependence was 0.7-1.1 at low 2-propanol pressures
2
2
and reached zero and even slightly negative at higher pressures.
Because the reaction rates depended on 2-propanol pressure, the apparent activation
energy E
condition of zeroth order dependence of 2-propanol pressure (P= 5.6 kPa) was chosen to
determine E by varying the temperature and measuring the reaction rates. Although Au/SiO did
not achieve zero order reaction, the reaction rates at this same pressure was used. The obtained
’s and preexponential factors for the overall reaction, acetone and propene formation are
shown in Table 7.
a
also depended on the propanol pressure. For comparison among catalysts, the
a
2
E
a
Because the main reaction for the Au-containing catalysts was acetone formation, values
for ln (A) and E for the overall reaction and acetone formation were similar for each of the
catalysts. However, although the E for acetone formation for 1-Ti/[Au/SiO ] and 4-Ti/[Au/SiO
were similar, they were substantially higher than that for Au/SiO . Since the E for propene
formation were smaller than for acetone formation, these larger E ’s could not be due to
contribution from propene formation. Instead, they must indicate different active sites. Similarly,
the E for propene formation on the two Au-containing catalysts were substantially smaller than
those on Ti/SiO . This implied that the properties of the active sites or the reactive intermediate
involved were different.
a
a
2
2
]
2
a
a
a
2
3
.5.
Catalytic 2-propanol Oxidation with O
2
1
1

Addition of O
for the Ti/SiO catalysts. For the Au-containing catalysts, the 2-propanol conversion increased
with time-on-stream for the first 2-3 h before reaching a steady state. For example, at 30 min,
the conversions for Au/SiO and 1-Ti/[Au/SiO ] were 11% and 14%, respectively. At steady
state, Au/SiO reached a conversion of ~40%, and 1-Ti/[Au/SiO ] reached ~20%.
2
in the feed increased the activities of the Au-containing catalysts but not
2
2
2
2
2
The product selectivities were little affected by the addition of O
, propene was the only product. For the Au-containing catalysts, acetone was the major
product accompanied with significant amounts of water production, indicating oxidative
dehydrogenation. In fact, on both 1-Ti/[Au/SiO ] and 4-Ti/[Au/SiO ] the selectivities for acetone
increased with O present (Table 8).
2
. On 1-Ti/SiO and 5-
2
Ti/SiO
2
2
2
2
4
. Discussion
Results of various characterization suggested that the catalysts contained 1-3 nm Au
nanoparticles on a nonporous SiO support. The samples were covered by an overlayer of
amorphous TiO , which XPS showed to consist of all Ti(IV). At high Ti loadings, the Ti
coverage was nonuniform, there were regions of thick TiO patches independent of whether Au
particles were present or not, and microdomains of crystalline TiO began to appear. The Ti ions
in the overlayer were present in roughly equal amounts of T and O coordination, such that the
2
2
2
2
d
h
average coordination number was about 5. This model was derived from corroborative results
from EELS, EXAFS, XANES, and DR-UV-vis measurements. Horváth et al.[47] prepared a
TiO
2
-decorated Au/SiO using a different Ti precursor than ours and colloidal Au particles but
2
otherwise a similar method of impregnation of Ti onto Au/SiO
dispersed over all surfaces of the catalyst.
2
. They also observed TiO
2
Only Lewis acid and no Brønsted acid sites were detected on all samples using pyridine
as the probe molecule The presence of 2-PrOH did not change the type of sites available, and no
new sites were formed. These Lewis acid sites were associated with the TiO overlayer and were
.
2
the main source of propene formation from 2-PrOH, since they were present in catalysts without
Au, and could account for the formation of propene on all the catalysts. However, it is interesting
to note that the apparent activation energy for propene formation was higher on the two Ti/SiO
catalysts than on the two Ti/[Au/SiO ] catalysts. We attribute this to the fact that on the Au-
2
2
containing catalysts, there are two competing pathways for an adsorbed 2-propoxy, namely to
form acetone versus propene. This will be discussed further later.
Acetone was formed only on Au-containing catalysts. Thus, Au must be responsible for
cleavage of the -C-H bond of the adsorbed 2-propoxy. Since the rates of acetone formation on
the two Ti/[Au/SiO
2
] catalysts were much higher than that on Au/SiO , and the catalyst with a
2
higher Ti loading was the most active, it can be concluded that acetone formation was facilitated
by the simultaneous presence of TiO and Au, such as at the interfacial perimeter sites. Thus, for
these two catalysts, 2-propoxide was formed by dissociative adsorption of 2-propanol on the
TiO at the interfacial perimeter site. This was followed by cleavage of the -C-H bond by the
Au atom to form adsorbed acetone, which desorbed as the product (Figure 17B). Recent studies
2
2
1
2

also reported the formation of acetone after the introduction of Au nanoparticles onto TiO
Al [17, 19].
2
and -
2
O
3
Adsorbed acetone desorbed at the same temperature of ~140 °C in TPD on all Au-
containing catalysts (Figure 12), which was lower than the reaction temperatures of 170-190 °C.
Assuming a first order desorption kinetics for adsorbed acetone and a pre-exponential factor of
1
1
12
-1
-1
1
0 - 10 min , the desorption activation energy was 85 - 93 kJ mol (Table 4). This is lower
-1
than the apparent acetone formation activation energy of 94-105 kJ mol for 1-Ti/[Au/SiO
-Ti/[Au/SiO ] (Table 7). These suggest -C-H bond cleavage is the rate-limiting step in 2-
PrOH decomposition. The addition of O to the feed increased the acetone formation rate
substantially. Since desorption of acetone is not expected to be affected by O , but -C-H bond
2
] and
4
2
2
2
cleavage could be affected by the presence of adsorbed O on Au, either by enabling the
formation of Au-OH or changing the electron density at the Au atom, the observation of
substantial rate enhancement by O
2
supports -C-H bond cleavage to be the rate limiting step.
In the absence of Ti, on the other hand, the activation energy for acetone formation over
-
1
-1
Au/SiO
similar pre-exponential factor. The very different apparent E
suggested that the active sites were different on these catalysts, further supporting the model that
on Ti/[Au/SiO ] catalysts, acetone formation was at the interfacial perimeter sites. It also
suggested a different reaction pathway on Au/SiO . One possible mechanism is shown in Figure
7A. 2-Propanol is adsorbed on Au followed by -C-H bond cleavage, and cleavage of O-H
bond is the rate limiting step. Isotope labeling experiments could be used to test this proposal.
2
was only 66 kJ mol , higher than its desorption activation energy (55 kJ mol ) with a
a
than those for Ti/[Au/SiO
2
]
2
2
1
Whereas TPRxn of 2-PrOH on Au/SiO
second, higher temperature peak was observed for 1-Ti/[Au/SiO
believe that the higher temperature peak is from the Au-TiO
intensity increased with Ti content.
2
showed only one acetone formation peak, a
] and 4-Ti/[Au/SiO ]. We
interfacial perimeter sites, as its
2
2
x
Propene formation took place readily on Lewis acid sites on the TiO
rate increased with increasing Ti content. The activity on the TiO overlayer was sufficient to
account for the propene formation on both 1-Ti/[Au/SiO ] and 4-Ti/[Au/SiO ]. Thus, there is no
2
overlayer, and the
2
2
2
evidence that propene was formed on the Au surface or the interfacial perimeter sites. It is
interesting that the 2-propanol partial pressure dependence of propene formation was different on
samples with and without Au. Without Au, the dependence was nearly zero order, and with Au,
there was a small positive order. It is possible that the apparent positive order on Au-containing
catalysts was due to the order dependence of faster acetone formation, which competes for
adsorbed 2-propoxide. This would be possible if the adsorbed 2-propoxide is mobile on the
surface and diffuses among different sites. It is also interesting to note that under our conditions,
TiO
2
was nearly inactive even taking into account the surface area. This suggests that O Ti is
h
much less active than Ti of a lower coordination.
The results of this study provide further support of the previous conclusion that the
interfacial perimeter sites are essential for selective acetone formation in the oxidation of
propane in the presence of a H
2
/O
2
mixture.[9, 10] They offer additional insight on the possible
reaction mechanism for that reaction. It is likely that adsorbed O or hydroperoxide (-OOH) on
1
3

Au is essential in the first step of propane activation by cleaving the O-H bond to form Au-OH
and adsorbed propoxide on TiO
as in this reaction.
x
at the interface. The propoxide then reacts in the same manner
5
. Conclusions
We have demonstrated the formation of acetone on Au-containing catalysts from
adsorbed 2-propoxide, the rate of which is facilitated by the presence of a porous, amorphous
TiO overlayer on Au. The high activity in the simultaneous presence of Au and amorphous TiO
substantiates the importance of Au-TiO interfacial perimeter sites for this reaction. The data
imply that Au at the interfacial perimeter site is responsible for -C-H bond scission, which is
the rate limiting step. This step is facilitated by the presence of O , suggesting that it is assisted
by adsorbed O on Au via enabling the formation of Au-OH. The fact that Au/SiO is capable of
forming acetone also suggests that TiO is not critical, and other oxides may also be effective.
Perhaps TiO is an attractive support because it easily forms an overlayer on Au, which results in
2
2
2
2
2
2
2
2
a high density of interfacial perimeter sites and thus a highly active catalyst. The results
gathered and the understanding derived could help future development of Au catalyst for
hydrocarbon conversions.
Acknowledgement
The authors acknowledge financial support from Institute for Catalysis in Energy Processes
supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences
under Award Number DOE DE-FG02-03-ER15457. The following facilities at Northwestern
University were used in this work: the EPIC, Keck-II, and/or SPID facility(ies) of the NUANCE
Center, which received support from the Soft and Hybrid Nanotechnology Experimental
(SHyNE) Resource (NSF ECCS-1542205), the MRSEC program (NSF DMR-1121262) at the
Materials Research Center, the International Institute for Nanotechnology (IIN), the Keck
Foundation, and the State of Illinois, through the IIN; the J.B. Cohen X-Ray Diffraction Facility
supported by the MRSEC program of the National Science Foundation (DMR- 1121262); DND-
CAT located at Sector 5 of the Advanced Photon Source (APS), whichis supported by
Northwestern University, E.I. DuPont de Nemours & Co., and The Dow Chemical Company; the
Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science user facility
operated for the DOE Office of Science by the Argonne National Laboratory under Contract No.
DE-AC02-06CH11357; the CleanCat Core facility funded by the U.S. Department of Energy
(DE-FG02-03ER15457); and the Quantitative Bio-element Imaging Center. The authors also
would like to thank Dr. Neil M. Schweitzer for discussion and experimental assistance, Professor
Justin Notestein for use of the DR-UV-vis spectrometer, and Dr. Qing Ma at the DuPont-
Northwestern-Dow Collaborative Access Team (DND-CAT) for discussion and experimental
assistance.
1
4

6
. References
[
1] X. Yang, X. Wang, C. Liang, W. Su, C. Wang, Z. Feng, C. Li, J. Qiu, Aerobic oxidation of alcohols
over Au/TiO2: An insight on the promotion effect of water on the catalytic activity of Au/TiO2,
Catal. Commun., 9 (2008) 2278-2281.
[
2] M. Turner, V.B. Golovko, O.P.H. Vaughan, P. Abdulkin, A. Berenguer-Murcia, M.S. Tikhov,
B.F.G. Johnson, R.M. Lambert, Selective oxidation with dioxygen by gold nanoparticle catalysts
derived from 55-atom clusters, Nature, 454 (2008) 981-983.
[
3] B. Taylor, J. Lauterbach, G.E. Blau, W.N. Delgass, Reaction kinetic analysis of the gas-phase
epoxidation of propylene over Au/TS-1, J. Catal., 242 (2006) 142-152.
4] M. Haruta, B. Uphade, S. Tsubota, A. Miyamoto, Selective oxidation of propylene over gold
deposited on titanium-based oxides, Res. Chem. Intermed., 24 (1998) 329-336.
5] J. Jiang, H. Kung, M. Kung, J. Ma, Aqueous phase epoxidation of 1-butene catalyzed by
suspension of Au/TiO2 +TS-1, Gold Bull., 42 (2009) 280-287.
6] K.K.B. J. J. Bravo-Suarez, T. Akita, T. Fujitani, T. J. Fuhrer and S. T. Oyama, <Propane reacts
[
[
[
with O2 and H2 on gold supported TS-1 to form oxygenates with high selectivity.pdf>, Chem.
Commun., (2008) 3272-3274.
[
7] J.J. Bravo-Suarez, K.K. Bando, T. Fujitani, S.T. Oyama, Mechanistic study of propane selective
oxidation with H-2 and O-2 on Au/TS-1, J. Catal., 257 (2008) 32-42.
8] J.J. Bravo-Suárez, K.K. Bando, J. Lu, T. Fujitani, S.T. Oyama, Oxidation of propane to
propylene oxide on gold catalysts, J. Catal., 255 (2008) 114-126.
9] Y.Y. Wu, N.A. Mashayekhi, H.H. Kung, Au-metal oxide support interface as catalytic active
sites, Catal. Sci. Technol., 3 (2013) 2881-2891.
10] N.A. Mashayekhi, Y.Y. Wu, M.C. Kung, H.H. Kung, Metal nanoparticle catalysts decorated
with metal oxide clusters, Chem. Commun., 48 (2012) 10096-10098.
11] D. Courcot, B. Grzybowska, Y. Barbaux, M. Rigole, A. Ponchel, M. Guelton, Effect of
[
[
[
[
potassium addition to the TiO2 support on the structure of V2O5/TiO2 and its catalytic
properties in the oxidative dehydrogenation of propane, Journal of the Chemical Society,
Faraday Transactions, 92 (1996) 1609-1617.
[
12] F.M. Bautista, J.M. Campelo, D. Luna, J. Luque, J.M. Marinas, Gas-phase selective oxidation
of toluene on TiO2-sepiolite supported vanadium oxides, Catalysis Today, 128 (2007) 183-190.
13] K.S. Kim, M.A. Barteau, Reactions of aliphatic alcohols on the {011}-facetted titanium
dioxide (001) surface, J. Mol. Catal., 63 (1990) 103-117.
14] K.S. Kim, M.A. Barteau, W.E. Farneth, Adsorption and decomposition of aliphatic alcohols
on titania, Langmuir, 4 (1988) 533-543.
15] O. Bondarchuk, Y.K. Kim, J.M. White, J. Kim, B.D. Kay, Z. Dohnalek, Surface Chemistry of 2-
[
[
[
Propanol on TiO2(110): Low- and High-Temperature Dehydration, Isotope Effects, and Influence
of Local Surface Structure, J. Phys. Chem. C, 111 (2007) 11059-11067.
[
16] E. Farfan-Arribas, R.J. Madix, Role of defects in the adsorption of aliphatic alcohols on the
TiO2(110) surface, Journal of Physical Chemistry B, 106 (2002) 10680-10692.
17] M.C. Holz, K. Kahler, K. Tolle, A.C. van Veen, M. Muhler, Gas-phase oxidation of 2-propanol
over Au/TiO2 catalysts to probe metal-support interactions, Phys. Status Solidi B, 250 (2013)
094-1106.
[
1
1
5

[
18] D.V. Potapenko, Z. Li, Y. Lou, Y. Guo, R.M. Osgood Jr, 2-Propanol reactivity on in situ
prepared Au(1 1 1)-supported TiO2 nanocrystals, J. Catal., 297 (2013) 281-288.
19] Z. Martinez-Ramirez, J.A. Gonzalez-Calderon, A. Almendarez-Camarillo, J.C. Fierro-Gonzalez,
[
Adsorption and dehydrogenation of 2-propanol on the surface of γ-Al2O3-supported gold, Surf.
Sci., 606 (2012) 1167-1172.
[
20] B.P. Block, J.C. Bailar, The Reaction of Gold(III) with Some Bidentate Coördinating Groups1,
J. Am. Chem. Soc., 73 (1951) 4722-4725.
21] R. Zanella, A. Sandoval, P. Santiago, V.A. Basiuk, J.M. Saniger, New preparation method of
gold nanoparticles on SiO2, J. Phys. Chem. B, 110 (2006) 8559-8565.
22] L.T. Zhuravlev, Concentration of hydroxyl groups on the surface of amorphous silicas,
Langmuir, 3 (1987) 316-318.
23] G.S. Parks, B. Barton, VAPOR PRESSURE DATA FOR ISOPROPYL ALCOHOL AND TERTIARY
BUTYL ALCOHOL, J. Am. Chem. Soc., 50 (1928) 24-26.
24] B. Ravel, M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption
spectroscopy using IFEFFIT, Journal of Synchrotron Radiation, 12 (2005) 537-541.
[
[
[
[
[
[
25] M. Newville, Atoms.inp Archive: Crystallographic Data from GSECARS.
26] V.M. Gun'ko, I.F. Mironyuk, V.I. Zarko, E.F. Voronin, V.V. Turov, E.M. Pakhlov, E.V.
Goncharuk, Y.M. Nychiporuk, N.N. Vlasova, P.P. Gorbik, O.A. Mishchuk, A.A. Chuiko, T.V. Kulik,
B.B. Palyanytsya, S.V. Pakhovchishin, J. Skubiszewska-Zięba, W. Janusz, A.V. Turov, R. Leboda,
Morphology and surface properties of fumed silicas, J. Colloid Interface Sci., 289 (2005) 427-445.
[
27] G. Bertoni, E. Beyers, J. Verbeeck, M. Mertens, P. Cool, E.F. Vansant, G. Van Tendeloo,
Quantification of crystalline and amorphous content in porous TiO2 samples from electron
energy loss spectroscopy, Ultramicroscopy, 106 (2006) 630-635.
[
28] C.-N. Huang, J.-S. Bow, Y. Zheng, S.-Y. Chen, N. Ho, P. Shen, Nonstoichiometric Titanium
Oxides via Pulsed Laser Ablation in Water, Nanoscale Research Letters, 5 (2010) 972-985.
29] T.R. Eaton, M.P. Campos, K.A. Gray, J.M. Notestein, Quantifying accessible sites and
reactivity on titania–silica (photo)catalysts: Refining TOF calculations, J. Catal., 309 (2014) 156-
65.
30] J.M. Notestein, L.R. Andrini, V.I. Kalchenko, F.G. Requejo, A. Katz, E. Iglesia, Structural
[
1
[
Assessment and Catalytic Consequences of the Oxygen Coordination Environment in Grafted
Ti−Calixarenes, J. Am. Chem. Soc., 129 (2007) 1122-1131.
[
31] J.T. Miller, A.J. Kropf, Y. Zha, J.R. Regalbuto, L. Delannoy, C. Louis, E. Bus, J.A. van Bokhoven,
The effect of gold particle size on AuAu bond length and reactivity toward oxygen in supported
catalysts, J. Catal., 240 (2006) 222-234.
[
32] F. Berube, B. Nohair, F. Kleitz, S. Kaliaguine, Controlled Postgrafting of Titanium Chelates
for Improved Synthesis of Ti-SBA-15 Epoxidation Catalysts, Chem. Mater., 22 (2010) 1988-2000.
33] L. Raboin, J. Yano, T.D. Tilley, Epoxidation catalysts derived from introduction of titanium
[
centers onto the surface of mesoporous aluminophosphate: Comparisons with analogous
catalysts based on mesoporous silica, J. Catal., 285 (2012) 168-176.
[
34] E. Sacaliuc, A.M. Beale, B.M. Weckhuysen, T.A. Nijhuis, Propene epoxidation over Au/Ti-
SBA-15 catalysts, J. Catal., 248 (2007) 235-248.
35] N. Kruse, S. Chenakin, XPS characterization of Au/TiO2 catalysts: Binding energy
assessment and irradiation effects, Applied Catalysis A: General, 391 (2011) 367-376.
[
1
6

[
36] S. Sakugawa, K. Wada, M. Inoue, Ti-bridged silsesquioxanes as precursors of silica-
supported titanium oxide catalysts for the epoxidation of cyclooctene, J. Catal., 275 (2010) 280-
87.
37] M.I. Zaki, M.A. Hasan, F.A. Al-Sagheer, L. Pasupulety, In situ FTIR spectra of pyridine
2
[
adsorbed on SiO2–Al2O3, TiO2, ZrO2 and CeO2: general considerations for the identification of
acid sites on surfaces of finely divided metal oxides, Colloids and Surfaces A: Physicochemical
and Engineering Aspects, 190 (2001) 261-274.
[
38] M. Tamura, K.-i. Shimizu, A. Satsuma, Comprehensive IR study on acid/base properties of
metal oxides, Appl. Catal., A, 433 (2012) 135-145.
39] S.M. Jung, P. Grange, TiO2–SiO2 mixed oxide modified with H2SO4: II. Acid properties and
their SCR reactivity, Applied Catalysis A: General, 228 (2002) 65-73.
40] B. Notari, R.J. Willey, M. Panizza, G. Busca, Which sites are the active sites in TiO2–SiO2
mixed oxides?, Catal. Today, 116 (2006) 99-110.
41] A.P. Kulkarni, D.S. Muggli, The effect of water on the acidity of TiO2 and sulfated titania,
Applied Catalysis A: General, 302 (2006) 274-282.
42] X. Xia, J. Strunk, S. Litvinov, M. Muhler, Influence of Re-adsorption and Surface
[
[
[
[
Heterogeneity on the Microkinetic Analysis of Temperature-Programmed Desorption
Experiments, J. Phys. Chem. C, 111 (2007) 6000-6008.
[
[
43] P.A. Redhead, Thermal desorption of gases, Vacuum, 12 (1962) 203-211.
44] Z.F. Liu, J. Tabora, R.J. Davis, RELATIONSHIPS BETWEEN MICROSTRUCTURE AND SURFACE-
ACIDITY OF TI-SI MIXED-OXIDE CATALYSTS, Journal of Catalysis, 149 (1994) 117-126.
45] A. Gervasini, J. Fenyvesi, A. Auroux, Study of the acidic character of modified metal oxide
surfaces using the test of isopropanol decomposition, Catal. Lett., 43 (1997) 219-228.
46] D. Haffad, A. Chambellan, J.C. Lavalley, Propan-2-ol transformation on simple metal oxides
TiO2, ZrO2 and CeO2, J. Mol. Catal. A, 168 (2001) 153-164.
47] A. Horváth, A. Beck, A. Sárkány, G. Stefler, Z. Varga, O. Geszti, L. Tóth, L. Guczi, Silica-
[
[
[
Supported Au Nanoparticles Decorated by TiO2:ꢀ Formation, Morphology, and CO Oxidation
Activity, J. Phys. Chem. B, 110 (2006) 15417-15425.
1
7

Figure captions:
Figure 1. N adsorption isotherms of catalysts. The isotherms are vertically displaced for clarity.
2
Figure 2. Barrett-Joyner-Halenda (BJH) pore size distributions. The bottom panel emphasizes the
smaller diameter region.
Figure 3. STEM images of (a) Au/SiO
2
, (b) 1-Ti/[Au/SiO
2
], (c) 4-Ti/[Au/SiO
2
] and (d) 5-Ti/SiO
2
The STEM images were collected on catalysts after 2-propanol decomposition.
Figure 4. 5-Ti/SiO after ozone treatment at 150 °C. (a) STEM image and (b) EELS spectra with
2
the corresponding areas in the circle and square.
Figure 5. Pre-edge features of the Ti K-edge of the catalysts.
Figure 6. The relative area of peaks (A
determined by the positions of A and A
solid circle) and anatase (solid triangle) in this study are plotted. Ba
fresnoite (empty square) and anatase (cross) are taken from Todd et al.[29] A
pre-edge peaks (A =A + A + A + A ).
2
+ A
. TS-1 (empty triangle), 1-Ti/[Au/SiO
TiO (empty circle),
is the total area of
3
)/A
T
versus area-weighted average pre-edge energy,
2
3
2
], 4-Ti/[Au/SiO ]
2
(
2
4
T
T
1
2
3
B
Figure 7. DR-UV-vis spectra of catalysts after reaction.
Figure 8. XPS of Ti 2p spectra for 1-Ti/[Au/SiO
2
] and 4-Ti/[Au/SiO
at 150 °C as part of the preparation procedure, and 50
/He at 180 °C for 1 h pretreatment before catalytic tests. They were exposed
2
]. The samples had been
3
-1
treated with 300 cm min of 10% O
3
/O
2
3
-1
cm min of 10% H
2
to air for XPS measurement.
Figure 9. DRIFT spectra with pyridine adsorption on 1-Ti/[Au/SiO
2
] and 4-Ti/[Au/SiO ].
2
Figure 10. DRIFT spectra of 4-Ti/[Au/SiO ] after adsorption of 2PrOH at 180 °C followed by
2
exposure to pyridine at the same temperature. The spectrum was collected after He purging at 25
mL/min for 15 min.
Figure 11. TPD-NH
3
spectra monitored with a mass spectrometer on (a) 1-Ti/SiO
] and 4-Ti/[Au/SiO ].
2
and 5-Ti/SiO
2
(
b) SiO , Au/SiO , 1-Ti[Au/SiO
2
2
2
2
Figure 12. TPD of acetone on Au/SiO
2
, 1-Ti/[Au/SiO
2
] and 4-Ti/[Au/SiO ].
2
Figure 13. Propene evolution during TPRxn of 2-PrOH.
Figure 14. Acetone evolution during TPRxn of 2-propanol.
Figure 15. Dependence of 2-propanol decomposition rates on the partial pressure of 2-propanol
on (a) 1-Ti/SiO , (b) 5-Ti/SiO , (c) Au/SiO , (d) 1-Ti/[Au/SiO ], and (e) 4-Ti/[Au/SiO ]. For
2
2
2
2
2
1
8

both d and e, the rates shown are the observed rate minus the rates on Au/SiO
order obtained from the slopes.
2
. n is the reaction
Figure 16. Dependence of propene (a, c, and e) and acetone (b, d, and f) production rates on the
partial pressure of 2-propanol on (a, b) Au/SiO , (c, d) 1-Ti/[Au/SiO ], and (e, f) 4-Ti/[Au/SiO ].
2
2
2
For c, d, e, and f, the rates shown are the observed rate minus the rates on Au/SiO
reaction order obtained from the slopes.
2
. n is the
Figure 17. 2-propanol decomposition on (A) Au/SiO
2
and (B) Ti/[Au/SiO ].
2
1
9

1
2
3
: SiO
: Au/SiO
: 1-Ti/[Au/SiO
2
4: 4-Ti/[Au/SiO
5: 1-Ti/SiO
] 6: 5-Ti/SiO
2
]
2
2
2
2
6
5
4
3
2
1
Figure 1. N adsorption isotherms of catalysts. The isotherms are vertically displaced for clarity.
2
2
0

Figure 2. Barrett-Joyner-Halenda (BJH) pore size distributions. The bottom panel emphasizes the
smaller diameter region.
2
1

Au/SiO
2
(a)
1-Ti/[Au/SiO
2
]
(b)
4
-Ti/[Au/SiO
2
]
(c)
5-Ti/SiO
2
(d)
Figure 3. STEM images of (a) Au/SiO
2
, (b) 1-Ti/[Au/SiO
2
], (c) 4-Ti/[Au/SiO
2
] and (d) 5-Ti/SiO
2
The STEM images were collected on catalysts after 2-propanol decomposition.
2
2

(a)
(
b)
Figure 4. 5-Ti/SiO
2
after ozone treatment at 150 °C. (a) STEM image and (b) EELS spectra with
the corresponding areas in the circle and square.
2
3

Figure 5. Pre-edge features of the Ti K-edge of the catalysts.
2
4

Figure 6. The relative area of peaks (A
determined by the positions of A and A
solid circle) and anatase (solid triangle) in this study are plotted. Ba
fresnoite (empty square) and anatase (cross) are taken from Todd et al.[29] A
pre-edge peaks (A =A + A + A + A ).
2
+ A
. TS-1 (empty triangle), 1-Ti/[Au/SiO
TiO
is the total area of
3
)/A
T
versus area-weighted average pre-edge energy,
], 4-Ti/[Au/SiO
(empty circle),
2
3
2
2
]
(
2
4
T
T
1
2
3
B
2
5

Figure 7. DR-UV-vis spectra of catalysts after reaction.
2
6

Figure 8. XPS of Ti 2p spectra for 1-Ti/[Au/SiO
2
] and 4-Ti/[Au/SiO
at 150 °C as part of the preparation procedure, and 50
/He at 180 °C for 1 h pretreatment before catalytic tests. They were exposed
to air for XPS measurement.
2
]. The samples had been
3
-1
treated with 300 cm min of 10% O
3
/O
2
3
-1
cm min of 10% H
2
2
7

Figure 9. DRIFT spectra with pyridine adsorption on 1-Ti/[Au/SiO
2
] and 4-Ti/[Au/SiO ].
2
2
8

Figure 10. DRIFT spectra of 4-Ti/[Au/SiO ] after adsorption of 2PrOH at 180 °C followed by
2
exposure to pyridine at the same temperature. The spectrum was collected after He purging at 25
mL/min for 15 min.
2
9