Bimetallic AuPt/TiO2Catalysts for Direct Oxidation of Glucose and Gluconic Acid to Tartaric Acid in the Presence of Molecular O2

Article abstract of DOI:10.1021/acscatal.0c02238

Tartaric acid is an important industrial building block in the food and polymer industry. However, green manufacture of tartaric acid remains a grand challenge in this area. To date, chemical synthesis from nitric acid-facilitated glucose oxidation leads to only <10% yield with significant toxics as byproducts. We reported a one-pot aqueous-phase oxidation of glucose and gluconic acid using bimetallic AuPt/TiO2 catalysts in the presence of molecular O2, with ～50% yield toward tartaric acid at 110 °C and 2 MPa. Structural characterization and density functional theory (DFT) calculation reveal that the lattice mismatch between fcc Pt and bcc Au induces the formation of twinned boundaries in nanoclusters and Jahn-Teller distortion in an electronic field. Such structural and electronic reconfiguration leads to enhanced σ-activation of the C-H bond competing with π-πelectronic sharing of the C═O bond on the catalyst surface. As a result, both C-H (oxidation) and C-C (decarboxylation) bond cleavage reactions synergistically occur on the surface of bimetallic AuPt/TiO2 catalysts. Therefore, glucose and gluconic acid can be efficiently transformed into tartaric acid in a base-free medium. Lattice distortion-enhanced reconfiguration of the electronic field in Pt-based bimetallic nanocatalysts can be utilized in many other energy and environmental fields for catalyzing synergistic oxidation reactions.

Full text of DOI:10.1021/acscatal.0c02238

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
2
Bimetallic AuPt/TiO Catalysts for Direct Oxidation of Glucose and
2
Gluconic Acid to Tartaric Acid in the Presence of Molecular O
Mengyuan Liu, Xin Jin, Guangyu Zhang, Qi Xia, Linyi Lai, Jinyao Wang,
Wenxiang Zhang, Yu Sun, Jie Ding, Hao Yan, and Chaohe Yang
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.0c02238 • Publication Date (Web): 01 Sep 2020
Downloaded from pubs.acs.org on September 1, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
Bimetallic AuPt/TiO2 Catalysts for Direct
Oxidation of Glucose and Gluconic Acid to
Tartaric Acid in the Presence of Molecular O2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Mengyuan Liu, Xin Jin*, Guangyu Zhang, Qi Xia, Linyi Lai, Jinyao Wang, Wenxiang
Zhang, Yu Sun, Jie Ding, Hao Yan, Chaohe Yang*
State Key Laboratory of Heavy Oil Processing, College of Chemical Engineering,
China University of Petroleum, No. 66 Changjiang West Road, Qingdao, Shandong
Province 266580, China
Corresponding Authors: jamesjinxin@upc.edu.cn, yangch@upc.edu.cn
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 40
1
2
3
4
Abstract
5
6
7
8
9
Tartaric acid is an important industrial building block in food and polymer industry.
However, green manufacture of tartaric acid remains a grand challenge in this area. Up
to date, chemical synthesis from nitric acid-facilitated glucose oxidation leads to only
< 10% yield with significant toxics as by-products. We reported a one-pot aqueous
phase oxidation of glucose and gluconic acid using bimetallic AuPt/TiO2 catalysts in
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
the presence of molecular O2, with ~50% yield toward tartaric acid at 110 C and 2
MPa. Structural characterization and DFT calculation reveal that, the lattice mismatch
between fcc Pt and bcc Au induces the formation of twinned boundaries in nanoclusters,
and Jahn-Teller distortion in electronic field. Such structural and electronic
reconfiguration leads to enhanced σ-activation of C-H bond competing with π-π
electronic sharing of C=O bond on catalyst surface. As a result, both C-H (oxidation)
and C-C (decarboxylation) bond cleavage reactions synergistically occur on the surface
of bimetallic AuPt/TiO2 catalysts. Therefore, glucose and gluconic acid can be
efficiently transformed into tartaric acid in base-free medium. Lattice distortion
enhanced reconfiguration of electronic field in Pt-based bimetallic nanocatalysts can be
utilized in many other energy and environmental fields for catalyzing synergistic
oxidation reactions.
Key words: oxidation; bimetallic catalysts; glucose; tartaric acid; lattice distortion;
electronic reconfiguration
2
ACS Paragon Plus Environment

Page 3 of 40
ACS Catalysis
1
2
3
4
5
6
7
1. Introduction
8
9
Replacing petroleum and coal feedstocks with renewable carbon resources is
considered to be critical for existing energy and environmental issues.1, 2 Biomass
feedstocks represent promising alternative sources owing to their great abundance on
earth. Up to date, the catalytic conversion of bio-derived molecules to valuable
chemicals has been the subject of extensive research efforts in both academia and
industry.3 In this context, aqueous phase oxidation (APO) of carbohydrates provides
emerging alternative routes for the sustainable production of renewable carboxylic
acids and derivatives.4, 5 This effort is largely driven by the increasing demands from
bio-based plastic industry. Sugar-derived carboxylic derivatives are major building
blocks for a variety of commodity products such as BPA-free plastics, bio-degradable
polymers and environment friendly detergents for human society.6 Obviously, clean
synthesis of renewable carboxylic acids is essential for reduced reliance on fossil-based
resources, shaping the future sustainability of chemical industry.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Tartaric acid (TA), a C4 hydroxyl dicarboxylic acid, is an important additive and
monomer with broad applications in food and polymer industry.5, 7 For example, TA is
used as an antioxidant and acidification agent. It can also be utilized as a performance
enhancer of cement.8 TA is also an essential component for manufacturing
pharmaceuticals as well as electronic devices.9 To our best knowledge, very limited
work has been reported on catalytic synthesis of TA from either bio- or fossil-derived
feedstocks (see left part of Figure 1). In particular, (a) TA is traditionally obtained via
biological route using enzyme catalysts (Figure 1a, hydrolase). However, it suffers from
poor production rate thus hindered for further industrial applications.10, 11 As an
alternative route, Scheel and colleagues first (in 1769) found that, (b) TA could be
recovered from tartar, a less-valued residue formed during wine production (Figure
1b).12, 13 After century development, TA manufacturing still heavily relies on the wine-
making industry. Poor extraction efficiency, distributed but small scaled batch
production technique, prevents its mass production for chemical industry. (c) In 1951,
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 40
1
2
3
4
5
6
7
8
9
Church and co-workers demonstrated that TA could be synthesized from maleic acid,
in H2O2 solution over tungstic acid as the catalyst (Figure 1c).14 This process generates
significant amounts of corrosives and hazards. (d) Similar issues have also bottlenecked
Mo/W catalyzed oxidation of gluruconic acid (Figure 1d).15 Energy intensive
separation and poor recyclability of homogeneous catalysts have posed a grand
challenge, making the process less sustainable and environmentally benign. Therefore,
it is urgently demanded greener approaches for making TA economically and
efficiently from biomass.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Different conversion routes for synthesis of tartaric acid from fossil- and bio-
feedstocks, (a) biological route, (b) extraction from tartar, (c) H2O2 oxidation of maleic
anhydride, (d) homogeneous gluruconic acid oxidation, (e) proposed direct oxidation of
glucose using O2 as oxidant.
Direct synthesis of TA from glucose provides a promising route for the production of
various downstream polymers and pharmaceutical products (Figure 1e). However,
initial attempts in both academia and industry remains unsuccessful. Numerous reports
have proposed Pt-based mono and bimetallic catalysts for glucose oxidation to glucaric,
tartronic and oxalic acids as valuable hydroxyl dicarboxylic acids.16-18 Overall TA
yields are, however, very low (Y < 10%).18-23 In addition, the detailed mechanism for
glucose oxidation to TA is still under debate, as rational control of C-C and C-H
cleavage of glucose molecule remains to be addressed in this field.
4
ACS Paragon Plus Environment

Page 5 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
In this work, we reported a series of synergistic AuPt/TiO2 catalysts (TOF: 959 h-1 at
110 oC) for direct glucose oxidation to TA using molecular O2. A record high yield of
50% TA was achieved in one-pot glucose conversion. The key finding in this work is
that, the twinned bimetallic AuPt clusters and Jahn-Teller distortion of d orbital at Au-
Pt interface synergistically contribute to enhanced production rate of TA (2.3 mol/mol-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
metal/h) from glucose at 110 C and 2.0 MPa (literature results: ~0.52 mol/mol-
metal/h).19, 20, 22 It is confirmed that, gluconic acid is the key intermediate for the
generation of TA. Experimental studies further demonstrated that, gluconic acid display
significantly high substrate inhibition effect on bimetallic AuPt/TiO2, compared with
glucose in aqueous medium. Based on surface characterization using UV-Vis and
structure-dependency studies obtained from kinetic analysis, it is highly possible that
simultaneous C-H and C=O bond activation competitively proceed on the surface of
AuPt/TiO2 catalysts. Furthermore, the influence of catalyst composition, reaction
temperature, pressure as well as compositional dependent kinetics on oxidation of
glucose and gluconic acid was studied systematically, with the aim to elucidate
plausible reaction mechanism for direct oxidation of glucose to TA. The outcome of
this work offers new insights into atom efficient oxidation of sugars to renewable
chemical products.
2. Experimental
2.1 Materials
TiO2 (rutile & anatase), glucaric acid potassium salts, D-xylonic acid provided by
Sigma-Aldrich. Glyoxylic acid (50% in water), DL-glyceraldehyde (90%) and DL-
tartaric acid (AR) were purchased by Macklin Biochemical Co., Ltd. D-glucose,
H2PtCl6·6H2O, HAuCl4·4H2O, Co(NO3)2·6H2O, Mn(NO3)2, NaOH, Na2CO3, glycolic
acid, NaBH4 were supplied by Sinopharm Chemical Reagent Co., Ltd. Gluconic acid
was supplied by Hubei Wande Chemical Co., Ltd.
2.2 Catalyst preparation
2.2.1 Synthesis of TiO2 supported monometallic Pt and Au catalysts
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 40
1
2
3
4
5
6
7
8
9
We used NaBH4 as the reducing agent to prepare monometallic Pt and Au nanocatalysts,
which is similar to previous reports.24 Specifically, certain amounts of the metal
precursor H2PtCl6 or HAuCl4 were added to a 150 mL aqueous slurry containing
approximately 1.5 g of TiO2 powders. The whole mixture was vigorously stirred at
room temperature for 1 h, after which freshly prepared NaBH4 aqueous solution (0.07
g in 100 mL) was added slowly to the slurry. The catalyst sample was then aged for 10
h under stirring. The solid catalyst samples were filtrated and washed with deionized
(DI) water at least four times to remove the residual species during catalyst synthesis.
The as-prepared catalysts were dried in a vacuum oven at 70 oC for another 10 h before
using in oxidation tests. Metal loading for Pt and Au on TiO2 support is 2 wt%,
respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.2.2 Synthesis of bimetallic catalysts supported on TiO2
Bimetallic AuPt/TiO2 catalysts were also synthesized using NaBH4 as the reducing
agent, the preparation procedure of which is similar with that for monometallic catalysts.
The loading of Pt/Au was varied from 0.5/0.5, 1.0/1.0, 2.0/2.0, 4.0/4.0 wt%/wt%. The
average particle size of these bimetallic AuPt/TiO2 catalysts is 4.2±0.1 nm, 4.8±0.1 nm,
7.5±0.2 nm and 8.4±0.3 nm, respectively. These samples were denoted as AuPt/TiO2-
4.2, AuPt/TiO2-4.8, AuPt/TiO2-7.5 and AuPt/TiO2-8.4, respectively. Similar
procedures were also followed to prepare the AuCo/TiO2, AuMn/TiO2, PtCo/TiO2
PtMn/TiO2 catalysts. The metal loading of these samples is same as 2.0/2.0 wt%/wt%.
2.3 Catalyst characterization
Inductively coupled plasma-optical emission spectrometry (ICP-OES) using an Agilent
730 spectrometer was used to determine Pt and Au content. The N2 adsorption-
desorption isotherms were measured on a Micromeritics ASAP 2020 instrument.
Transmission electron microscope (TEM) images were obtained with a Tecnai G2 F20
S-Twin instrument equipped with a digital imaging system. X-ray photoelectron
spectroscopy (XPS) as performed using a Perkin-Elmer PHI ESCA system. The binding
energy of the C1s peak, i.e., 284.6 eV, was used was the standard for calibration.
6
ACS Paragon Plus Environment

Page 7 of 40
ACS Catalysis
1
2
3
4
5
6
The dispersion of metals was checked by CO adsorption using static volumetric
methods (Micromeritics AutoChem 2920 analyzer). Sample (about 0.05 g) was firstly
7
8
o
pretreated in hydrogen (50 cm3/min) at 150 C for 1 h and purged with helium (50
9
cm3/min) for 1 h at the same temperature. And then, the catalyst was cooled to 50 oC
and CO pulses were injected from a calibrated on-line sampling valve. CO adsorption
equilibrium criterion is 1.2. Stoichiometry of 2 was used for calculations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
UV-Vis absorption spectra and diffuse reflectance (DR) spectra were recorded with a
UV-Vis-NIR Cary 5 (Varian) spectrophotometer. DR spectra were plotted in Kubelka-
Munk (KM) arbitrary unit vs wavelength. UV-vis spectra of liquid phase were recorded
using a spectrophotometer (TU-1901), for which the solutions were left in a cell for the
measurement.
2.4 Oxidation test
The catalytic performances of solid catalysts were evaluated in a 50 mL autoclave. In a
typical experiment, a certain amount of the catalyst and 15 mL of glucose or gluconic
acid aqueous solution (0.3 mol/L) were transferred into the reactor, respectively. The
autoclave was then purged twice with pure O2 at room temperature and then filled with
O2 before heating. The reaction mixture was heated under 200 rpm till target reaction
temperature (e.g., 110 °C), when stirring rate was switched and maintained at 1000 rpm
for pre-determined time (e.g., 4 h). Following each batch reaction, the reactor was
cooled down and the mixture was analyzed by high-performance liquid
chromatography (HPLC) using a Shimadzu LC-20AT equipped with both refractive
index (RID-10A) and ultraviolet-visible (SPD-20A) detectors. A 5 mmol/L aqueous
H2SO4 solution flowing at 0.8 mL/min was used as the mobile phase in conjunction
with an ion-exclusion column (Rezex ROA-Organic Acid H+ (8%), 300 × 7.8 mm) at
60 °C. The definitions of turnover frequency (TOF), conversion, selectivity, carbon
balance (%) and reaction rate used herein are similar to those previously described in
literature.25 In particular, TOF is defined as the amounts of reactant being converted
over certain amounts of surface metals during a period of time. Conversion (X) is
defined as the ratio of amounts of reactant transformed to that initially charged.
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 40
1
2
3
4
5
6
7
8
Selectivity (S) is defined as the ratio of total moles of carbon in particular product to
the carbon mole in the converted reactant. Carbon balance (C%) is defined as the ratio
of total amounts of carbon in the product formed to that existing in converted reactants.
9
△ 푁푟푒푎푐푡푎푛푡
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
TOF = 푁
(calculation of surface metal: refer to Supporting
푠푢푟푓푎푐푒 푚푒푡푎푙 × 푅푒푎푐푡푖표푛 푡푖푚푒
Information)
푖푛푖푡푖푎푙
푓푖푛푎푙
―
Conversion = C
%
C푚표푙푒푠, 푟푒푎푐푡푎푛푡
푚표푙푒푠, 푟푒푎푐푡푎푛푡
푖푛푖푡푖푎푙
× 100
C
푚표푙푒푠, 푟푒푎푐푡푎푛푡
푓푖푛푎푙
C
푚표푙푒푠, 푝푟표푑푢푐푡푠
Selectivity = C
%
× 100
푖푛푖푡푖푎푙
푚표푙푒푠, 푟푒푎푐푡푎푛푡
푓푖푛푎푙
―
C푚표푙푒푠, 푟푒푎푐푡푎푛푡
푓푖푛푎푙
C
푚표푙푒푠, 푝푟표푑푢푐푡푠
푐표푚푣푒푟푡푒푑
C% =
%
× 100
C
푚표푙푒푠, 푟푒푎푐푡푎푛푡
△ 푁푟푒푎푐푡푎푛푡
푀푎푠푠푐푎푡푎푙푦푠푡(푡표푡푎푙) ∙ 푇푖푚푒푟푒푎푐푡푖표푛
Reaction rate =
2.5 Density functional theory (DFT) calculations
DFT calculation was carried out to estimate change of bond length and electronic
reconfiguration in nanoclusters. The procedure has been discussed in our previously
thus stated briefly. Pt (111) surface were modeled using a triple-layer p (4×4) slab
comprising 3 metal layers with 16 atoms per layer and each slab was separated by a
vacuum of 17 Å to ensure the interaction between neighboring cells was negligible,
which is similar to our previously reported method.26, 27 All DFT calculations were
performed using CASTEP code.28,
Using the PBE functional, the exchange-
29
correlation effects were described with the generalized gradient approximation (GGA).
The electron wave functional was expanded by a plane-wave basis set with a cutoff
energy of 400 eV. A 3×3×2 Monkhorst-Pack k-point grid was used for the integrations
of the Brillouin zone. The performance of the three-layer slab model and the calculation
parameters, including the p (3×4) supercell, 3×3×2 k-point, 15 A vacuum zone, and the
energy cutoff of 400 eV, have been tested extensively in previous studies.
3. Result and discussion
3.1 Catalyst screening
8
ACS Paragon Plus Environment

Page 9 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
We selected several mono and bimetallic catalyst candidates and tested their
performances in direct oxidation of glucose base-free medium. These catalysts include
Au/TiO2, Pt/TiO2, AuPt/TiO2*, PtCo/TiO2, PtMn/TiO2, AuCo/TiO2 and AuMn/TiO2
materials. Conversion and selectivity of glucose over these catalysts are studied at 110
oC and 1 MPa O2, the results of which are presented in Table 1. Before we compared
the conversion and selectivity data, calculation on mass transfer significance was
performed to exclude possible contribution of external and internal mass transfer
limitation (refer to Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Oxidation of glucose over various mono and bimetallic catalysts
Selectivity (%)
Conversion
#
Catalyst
C%
(%)
GA
TA
XA
Others
1
2
3
4
5
6
7
Au/TiO2
Pt/TiO2
48.1
38.2
100.0
38.8
37.2
78.1
72.2
67.6
63.2
57.1
48.9
49.8
62.2
41.6
2.7
12.4
20.0
11.7
12.1
0.0
0.0
0.0
3.8
6.2
5.5
5.6
7.3
11.2
11.7
13.9
13.9
11.5
12.2
13.1
81.5
87.3
94.8
80.8
78.9
80.1
67.2
AuPt/TiO2*
PtCo/TiO2
PtMn/TiO2
AuCo/TiO2
AuMn/TiO2
5.3
Reaction conditions: Cglucose,0: 0.3 M, catalyst amount: 0.05 g, Pt loading: 2wt%, Pt/M (Au,
Co, Mn) mass ratio: 1/1, Au loading: 2 wt%, Au/M (Co, Mn) mass ratio: 1/1, T: 110 oC, 1
MPa O2, 4 h.
AuPt/TiO2* is the AuPt/TiO2-7.5 sample mentioned later.
We found from entries#1 and #2 in Table 1 that, monometallic Au/TiO2 and Pt/TiO2
catalysts exhibit 48.1% and 38.2% conversion of glucose, respectively, after 4 h batch
time. It is found that the major products include gluconic acid (GA), tartaric acid (TA),
xylonic acid (XA), with xylaric acid (XYLA), glucaric acid, tartronic acid and glyceric
acid formed as well. The selectivity towards TA over Pt/TiO2 (12.4%) is slightly higher
than that on Au/TiO2 catalyst (2.7%). We further observed in entry#3 that, bimetallic
AuPt/TiO2* catalyst shows a complete conversion of glucose within 4 h. Selectivity of
TA over bimetallic AuPt/TiO2* catalyst is approximately 20%, which is higher than
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 40
1
2
3
4
5
6
monometallic Au/TiO2 and Pt/TiO2 catalysts. This value is also higher compared with
previously reported data in literature (<10%).18, 20
7
8
9
The conversion and product distribution over other bimetallic catalysts were also
analyzed quantitatively (entries#4-#7). We confirmed that the major oxidation products
over PtCo/TiO2 and PtMn/TiO2 catalysts (entries#4, #5) also include GA, TA and XA,
but the conversion levels are much lower compared to AuPt/TiO2* catalyst. In addition,
we found that the overall conversion of glucose on AuCo/TiO2 and AuMn/TiO2
(entries#6 and #7) is higher than PtMn/TiO2 and PtCo/TiO2 catalysts. However, TA
selectivity is low (< 5%). Based on our preliminary studies, it is possible that, while Au
content shows good activity for glucose oxidation, Pt seems to be more selective
towards TA formation. Therefore, the bimetallic AuPt/TiO2* catalyst show synergism
during glucose conversion.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Since bimetallic AuPt/TiO2* catalyst showed enhanced performances for both activity
and selectivity toward TA, we were also interested in evaluating its performance in
oxidation of gluconic acid, as it is known to be a key primary product from glucose
conversion. Results in Figure 2 show that, for oxidation of gluconic acid, Pt/TiO2
(15.5%) outperforms Au/TiO2 catalyst (8.4%) in term of conversion at 110 oC after 4
h. It is clear that Au/TiO2 gives XA (a C5 product) as the major product, while Pt/TiO2
material leads to the formation of both TA and XA (combined selectivity > 70%).
Therefore, it is plausible that, gluconic acid tends to follow C1-C5 cleavage reaction via
decarboxylation reaction (orange route in Figure 2) over Au/TiO2 catalyst, thus XYLA
is formed as the major product.30 But for Pt/TiO2 catalyst, gluconic acid (GA in Figure
2) seems to undergo subsequent C-C cleavage and oxidation, which eventually gives
XYLA and TA as major products (orange and green routes).22 Again, bimetallic
AuPt/TiO2* catalyst shows improved conversion compared with Au/TiO2 and Pt/TiO2
catalysts, promoting consecutive C-C cleavage of gluconic acid (via orange and green
routes), thus TA is formed as the main product with almost 40% selectivity. Therefore,
it is clear that, selective formation of TA requires two types of active sites for effective
catalytic materials, one catalyzing oxidation reactions while the other facilitating C-C
10
ACS Paragon Plus Environment

Page 11 of 40
ACS Catalysis
1
2
3
4
5
6
bond cleavage. Obviously, AuPt/TiO2* catalyst exhibits strong synergism for such
reactions, thus is chosen for detailed characterization and kinetic studies.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Conversion and selectivity of gluconic acid (GA) oxidation over Au/TiO2, Pt/TiO2
and AuPt/TiO2* catalysts and possible formation route for tartaric acid (TA). Reaction
conditions: CGA0 =0.3 M, catalyst amount=0.05 g, T=110 oC, 1 MPa O2, 4 h.
3.2 Catalyst characterization
3.2.1 Physical properties
Structural properties of Pt/TiO2 and various AuPt/TiO2 catalysts were measured for
surface areas and pore volume. As shown in Table 2, all catalyst samples display very
low surface areas in the range of 8.5 - 9.1 m2/g. The pore volume varies insignificantly
for different catalysts (2.5 - 3.7 10-2m3/g). Metal dispersion for Pt and Au content was
measured by both chemisorption and TEM images. In fact, Au possesses very weak
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 40
1
2
3
4
5
6
7
8
9
chemisorption properties due to the absence of vacancies in its d-bands .31 Since the
AuPt alloy will reduce the amount of CO adsorbed, the measured dispersion data will
be lower than the actual value. Due to unusual adsorption behaviors of CO on Pt and
Au sites, chemisorption data cannot precisely reflect actual metal dispersion state.32, 33
Therefore, we used particle size distribution obtained from TEM characterization to
determine metal dispersion, which is an extensively accepted protocol.34
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Physical properties of solid catalyst samples
Metal content by ICP
Metal dispersion (%)
SBET
(m2/g)
Vpore
(10-2m3/g)
(wt%)
Catalyst
Au
Pt
Chemisorption TEM
Pt/TiO2
8.7
9.1
8.9
9.0
8.5
2.5
3.7
2.6
2.5
2.7
-
1.8
3.2
1.7
0.9
0.6
24.2
0.3
-
AuPt/TiO2-8.4
AuPt/TiO2-7.5
AuPt/TiO2-4.8
AuPt/TiO2-4.2
3.4
1.6
0.9
0.7
13.5
15.1
23.3
26.9
0.2
14.1
12.3
3.2.2 Surface morphologies
TEM images for various AuPt/TiO2 provide important information on particle size
distribution and surface morphology. Figures 3 and S1 present low resolution TEM
images for AuPt/TiO2-8.4 (Figure 3a), AuPt/TiO2-7.5 (Figure 3b), AuPt/TiO2-4.8
(Figure 3c) and AuPt/TiO2-4.2 (Figure 3d), respectively. The particle size distribution
is 4.2±0.1 nm, 4.8±0.1 nm, 7.5±0.2 nm and 8.4±0.3 nm, respectively for the four
samples. The TEM images of monometallic Pt/TiO2 are shown in Figure S3a. It can
clearly see that the average size of Pt is 6.9±0.7 nm in Pt/TiO2. From the deviation,
there is a wide range of particle size distribution. It is confirmed that addition of Au
could promotes uniform particle size. TEM images in Figure 3 and Figure S3a also
confirm the poor porosity of TiO2 support, as majority of metallic particles exist at outer
surface of TiO2 materials. Therefore, there is no doubt that catalytic reactions of glucose
mostly occur at outer surface of TiO2 with negligible contribution from AuPt and Pt
12
ACS Paragon Plus Environment

Page 13 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
nanoparticles embedded in pore channels. There are no diffraction peaks of metal
components except for the peaks of TiO2 in XRD patterns of AuPt/TiO2 catalysts
(Figure S4). It also indicates that the bimetallic AuPt nanoparticles have a small size,
which cannot be detected using XRD technique (see supporting information for more
discussion on XRD results).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. TEM images of various AuPt/TiO2 catalysts
Figure 4. HR-TEM images for (a, b) AuPt/TiO2-8.4 and (c, d) AuPt/TiO2-7.5 catalysts. (e) is
elemental mappings of AuPt/TiO2-8.4.
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 40
1
2
3
4
5
6
7
8
9
HR-TEM images in Figures 4 (a-d) and S2 further show that, there exist large quantities
twinned grains with clear boundaries at Au-Pt interfaces. We selected AuPt/TiO2-8.4
and AuPt/TiO2-7.5 samples for detailed characterization of their lattice structures. As
shown in Figure 4a, we observed variously oriented lattice planes within selected
nanoparticles for AuPt/TiO2-8.4 sample. The orientation of facets highlighted by
yellow dash lines suggest that, the lattice mismatch between Au (100) and Pt (111)
induces formation of new interfaces, thus twinned grains are formed in this system.35,36-
38 While similar observation was found in Figure 4b for the same sample, measurement
of lattice spacing of AuPt/TiO2-7.5 sample in Figures 4c and 4d also implies typical Pt
(111) and Au (100) facet co-existing in bimetallic AuPt nanoparticles. However, HR-
TEM images in Figure S3b shows that, there are the regular lattice planes in Pt/TiO2,
which is consist with Pt (111) plane. Since Pt and Au show fcc and bcc crystallinity in
monometallic clusters, respectively, incorporation of Au in Pt facets undoubtedly
creates large surface strain during the growth of clusters. As a result, the tensile strain
at Pt and Au interface often leads to anisotropic growth of bimetallic nanoclusters.39
EDS-mapping was conducted to examine the elemental distribution in AuPt/TiO2-8.4
(Figure 4e). Figure 4e shows the representative STEM images and corresponding color-
coded EDS spectral maps of catalysts AuPt/TiO2-8.4. Obviously, Au (green) and Pt
(red) are uniformly distributed on the same particle in the AuPt/TiO2-8.4 sample
without obvious segregation, which confirming the formation of AuPt alloy.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In this work, at the very early stage, the precursors of Pt and Au were quickly reduced
to form metallic atoms upon the addition of NaBH4, owing to the strong reducing
property of NaBH4. Instantaneous formation of large amounts of Pt and Au nuclei
contributes as perfect template to homogeneous crystal growth for both monometallic
clusters. From the view of the kinetics, TiO2 provides the electron transfer channel from
Au3+ to Pt4+ which will help in the co-reduction of Au and Pt ions. It should be pointed
out that the Au-Pt interaction exists in the solution, which provides the possible
nucleation of AuPt alloys when the core is induced to mix at the atomic level.40 Thus
the newly formed Au and Pt atoms quickly fuse together to form small AuPt nuclei.41
14
ACS Paragon Plus Environment

Page 15 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
However, co-existence of structurally distinguished clusters and strong tendency for
agglomeration under polymer-free condition, leads to interfacial tension-determined
anisotropic growth, as we already described in previous cases with PtFe and PtMn
clusters.25, 42 In general, such strong interface strain without surface mediating agents
could possibly induce formulation of random Janus structures. As seen from Figure 4,
it is noted that the low energy AuPt crystal (100) and (111) planes as the directing agent.
Therefore, different metal loadings could determine the final alloy particle size. Such
growth underwent approximately 10 h aging in aqueous phase. Therefore, the
anisotropic growth of small AuPt clusters overcomes strong interfacial force and
eventually formulated twinned boundaries within bimetallic particles. The strong
metal-metal interaction will lead to changes in the structure of the catalyst interface,
and the stabilization of transition states on alloy catalysts is an additional benefit.
Prolonged aging time results in structural reconstruction with several defected sites on
catalyst surface. In addition, the slight difference in electronegativity between Pt and
Au could also intrinsically affect the final morphology. Slightly higher electron affinity
of Au element might lead to faster reduction rates in aqueous solution. As a result, bond
length and electronic configuration of Pt-Pt sites might be altered on Au-Au template
(confirmed in later sections), which contributes to enhanced catalytic performances
during oxidation of glucose and gluconic acid.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3.2.3 Electronic reconfiguration
(1) Electronic binding properties. XPS analysis was conducted to evaluate the
variation in the chemical states of Au and Pt elements and possible changes of binding
energies (BE). Two chemical states of Ti were separated in the Ti 2p spectrum at the
BE of Ti 2p3/2 peak close to 458.6 eV and 458.1 eV, which can be identified as Ti4+ and
Ti3+, respectively (Figure 5c). The Au 4f spectra were fitted successfully with Auδ+ and
Au0 components (Figure 5a), which were separated at the BE of Au 4f7/2 peak at around
83.5 eV and 82.8eV. The Au0 species are oxidized to Auδ+ on the surface of TiO2 facets,
which could be ascribed to the strong interaction between metallic Au clusters and the
surface Ti3+ centers at defects in the TiO2 [43, 44]. Pt XPS spectra (Figure. 5b) can be
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 40
1
2
3
4
5
6
7
8
9
divided into two pairs of peaks. The binding energy bands centered at 70.4 eV can be
assigned to the Pt metal state (Pt 4f7/2 energy levels). The other peaks are in good
agreement with Pt2+, indicating the presence of oxidized Pt species [45, 46]. In the spectra
of Au 4f and Pt 4f(as shown in Figure 5a and 5b), the variation of BE for Auδ+ and Pt2+
contents show quite a distinct tendency in mono and bimetallic samples. Therefore, it
is indicate the formation of Au/Pt alloy compounds.[47] Compared with monometallic
samples, the Pt2+ content ascends, while the Auδ+ content slightly reduces in AuPt/TiO2
samples. It is clear that the strain at Au-Pt interface also induces electronic
redistribution. The positively charged Pt sites are more conducive to the formation of
π-π bonds with -C=O groups. As a result, the bimetallic AuPt/TiO2 catalyst promotes
the C-C cracking (decarboxylation reaction) caused by this interaction, which
ultimately leads to the formation of smaller products. The varied size for AuPt
nanoparticles shows the trend on BE values that the positively charged Pt increases as
the size of the AuPt alloy particles increases. In the following two sections, therefore,
we conducted detailed ex-situ and in-situ characterization of AuPt/TiO2 catalysts to
further reveal the plausible mechanism of C-H and C=O groups on the modified catalyst
surface. The surface Au-Pt ratio (Table S2) measured by XPS is consist with the bulk
Au-Pt ratio, which confirms that the AuPt alloy is uniformly distributed on the surface.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
16
ACS Paragon Plus Environment

Page 17 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. XPS spectra of mono and bimetallic catalysts
(2) Electronic coupling at Au-Pt and AuPt-TiO2 interfaces. UV-Vis DR spectra for
monometallic Pt/TiO2, Au/TiO2, bimetallic AuPt/TiO2-7.5 and blank TiO2 samples
confirm the Au-Pt electronic coupled effect and strong AuPt-TiO2 interaction (Figure
6). We also showed that activation of gluconic acid follows alternative mechanism on
the surface of mono and bimetallic catalysts (Figure 7). Specifically, as shown in Figure
6a, the blank TiO2 sample (black curve) displays no signal in visible area (> 400 nm).
Signals for monometallic Pt/TiO2, Au/TiO2 and bimetallic AuPt/TiO2 catalysts are
different from the blank TiO2 support. In particular, Au/TiO2 sample (yellow curve in
Figure 6a) displays a typical peak around 554 nm, which is a characteristic response for
surface plasmonic resonance of Au element,35 while no such peak is not found for
AuPt/TiO2-7.5 catalyst. This observation suggests that electronic transfer at Au-Pt
interface alters the configuration of d orbital at Au sites in bimetallic AuPt/TiO2-7.5
catalyst.
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. (a) UV-Vis absorption spectra and (b) band gap energy estimation of TiO2
supported mono Pt, Au and bimetallic AuPt catalyst samples
Band gap energy estimation for TiO2 support is shown in Figure 6b. As it is observed,
the value of band gap energy for blank TiO2 sample is approximately 3.2 eV, while
AuPt immobilized TiO2 sample has a higher band gap. Increased band gap induces
oxidation Ti3+ to Ti4+ species in TiO2 framework, and subsequent electronic transfer
from oxide structure to metallic sites.48, 49 The result is also consistent with our
benchmark data over anatase and rutile TiO2 supports (Table S1), where the both
supports show similar promotional effect for oxidation of glucose to TA. Therefore, it
is clear that Au-Pt and AuPt-TiO2 interaction leads to reconfiguration of electronic
orbitals at metallic sites.
(3) Activation of -C-OH and -C=O groups on catalyst surface. According to the
possible reaction pathways shown in Figure 2, conversion of glucose to TA is achieved
by forming gluconic acid as the primary product, followed by decarboxylation (C-C
cleavage) and deep oxidation reactions. Hence, we conducted UV-Vis spectra
measurement for solid catalysts in aqueous gluconic acid solution, to reveal possible
mechanism for C-OH and C=O bond activation on catalyst surface. Three key findings
are presented and discussed as below.
18
ACS Paragon Plus Environment

Page 19 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. UV-Vis spectra for gluconic acid solution in the presence/absence of solid catalysts
(i) n to σ* Excitation. As shown in Figure 7a, the blank gluconic acid solution sample
displays typical peaks for n to σ* excitation of hydroxyl (-C-OH) and aldehyde (-C=O)
groups. However, the original conformation is interrupted by adding TiO2 support in
the aqueous gluconic acid solution (Figure 7b). It displays detectable signals for n to σ*
excitation of -C=O and -C-OH groups.50 This observation is well agreed with previous
findings that, surface -OH groups of TiO2 could interact with -C-OH groups and form
H bond.51, 52
Further control measurements were conducted in the presence of Pt/TiO2 (Figure 7c),
Au/TiO2 (Figure 7d) and AuPt/TiO2-7.5 (Figure 7e) catalysts. Distinguished from other
four samples, UV-Vis spectra for Pt/TiO2 catalyst in gluconic acid solution (Figure 7c)
display a stronger peak for n to σ* excitation of -C-OH group than that of -C=O group,
suggesting that Pt sites play a determining role in activating gluconic acid molecules to
aqueous oxidation.53 Compared with Pt/TiO2 catalyst, n to σ* excitation of -C-OH
group is relatively weaker in the presence of Au/TiO2 catalyst (see Figure 7d). However,
the peak for n to σ* excitation of -C=O group has been intensified to a higher level.
Taking into account the poor catalytic activity of Au/TiO2 catalyst (Figure 2), weakened
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 40
1
2
3
4
5
6
7
8
activation of -C-OH group leads to poor conversion of gluconic acid to dicarboxylic
acids, while strong -C=O interaction favors formation of active per-acid species with
strong oxidative properties.54, 55
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(ii) n to π* Excitation. For bimetallic AuPt/TiO2-7.5 catalyst, the presence of Au
species induces the generation an adjacent active sites closing to Pt, leading a more
strong signal of n to π* excitation (see Figures 7c, 7e).56 It tends to interact strongly
with unpaired electrons in gluconic acid molecule. Therefore, activation of gluconic
acid on bimetallic AuPt/TiO2 catalysts is highly possibly following n to π* mechanism.
(iii) π to π* Excitation. We also detected signals for π to π* excitation in Pt/TiO2
(Figure 7c), Au/TiO2 (Figure 7d) and bimetallic AuPt/TiO2-7.5 (Figure 7e) catalyst
samples. Such signal is absent for blank gluconic acid solution. Compared with TiO2
and Au/TiO2 samples, π to π* excitation for -C=O group on Pt/TiO2 catalyst display a
red shift, suggesting facile π-π interaction on the surface of Pt/TiO2 catalyst.
Based on the above-mentioned findings, it is possible that, the existing two competing
activation mechanism for gluconic acid, (a) one predominately initiated by the
activation of terminal -C-OH group, (b) the other involving π-π excitation of -C=O
group. The former one results in the formulation of dicarboxylic acids by oxidative C-
H bond cleavage (σ bond activation), while the latter case facilitates C-C bond cleavage,
due to the fact that decarboxylation reactions starts with strong π-π excitation of
carboxylic groups.57, 58 Simultaneous occurrence of the two competing reactions
eventually contribute to the formation of C5, C4 and C3 dicarboxylic acids, such as
XLYA (C5), TA (C4) and tartronic acid (C3) acids.
3.3 Structure-sensitivity
(1) Size-dependency of AuPt particle on oxidation reactions. It is demonstrated in
this part that, both glucose and gluconic acid conversion over bimetallic AuPt/TiO2
catalysts display a typical dual-site mechanism involving π-π interaction on catalyst
surface. As shown in Figure 8a, as cluster size of AuPt increases from 4.2 nm to 8.4
nm, TOF values measured on glucose conversion show significant enhancement from
20
ACS Paragon Plus Environment

Page 21 of 40
ACS Catalysis
1
2
3
4
o
434 h-1 to approximately 982 h-1 at 110 C and 1 MPa O2 (purple line in Figure 8a).
5
6
7
8
Although the overall reaction rates for gluconic acid (blue line) over bimetallic
AuPt/TiO2 catalysts are much lower than those for glucose oxidation, it is clear that a
similar size-dependent trend was also observed. Although σ-σ interaction for C-H bond
on AuPt sites is critical for oxidation reactions, the overall conversion rate of glucose
and gluconic acid is determined by π-π bond activation of C=O group (a dual-site
mechanism). Clearly, the size-dependency found on these AuPt/TiO2 catalysts with
different sizes in Figure 8a is agreed very well with the results obtained from UV-Vis
spectra (Figure 7).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(2) Substrate inhibition. We conducted additional sets of control experiments to
understand the significance of σ-σ and π-π bond interaction under different
concentration of glucose and gluconic acid. As seen in Figures 8b and 8c, control
experiments were carried out over two selected catalysts, Pt/TiO2 and AuPt/TiO2-7.5.
It is found that, as glucose concentration increases, initial reaction rates measured on
both Pt/TiO2 and AuPt/TiO2-7.5 catalysts are enhanced (blue zone in Figure 8b),
although AuPt/TiO2-7.5 catalyst displays much higher reaction rates.
However, when glucose concentration is higher than 1.5 mol/L, we observed a typical
Langmuir-type adsorption behavior, with almost unchanged initial reaction rates (see
2.4 for definition) over both catalysts (orange zone in Figure 8b). This observation
implies that for glucose conversion, σ-activation for C-H bond is important for
oxidation step to form gluconic acid. The π-π interaction is also important, but the fact
that decarbonylation of glucose is almost negligible in the presence of Pt/TiO2 or
AuPt/TiO2-7.5 catalysts, indicates C=O induced C-C bond cleavage is negligible under
the reaction condition. Therefore, initial reaction rates of glucose display such profiles
over both mono and bimetallic catalysts.
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Conversion profiles on mono and bimetallic catalysts for glucose and gluconic acid
conversion at 110 oC and 1 MPa O2. (a) TOF vs size of AuPt clusters for both glucose and
gluconic acid conversion (Reaction conditions: CGA,0 (Cglucose,0): 0.3 M, catalyst amount:
AuPt/TiO2-4.2(0.2g), AuPt/TiO2-4.8(0.1g), AuPt/TiO2-7.5(0.05 g), AuPt/TiO2-8.4 (0.025g)) , (b)
influence of glucose and (c) gluconic acid concentration on initial reaction rate over
monometallic Pt/TiO2 and bimetallic AuPt/TiO2-7.5 catalysts (refer Figure 2 for other
experimental details; refer 2.4 for definition of reaction rate)
Interesting results were observed for oxidation of gluconic acid (Figure 8c). We were
surprised to find that concentration-dependence on initial reaction rate of gluconic acid
is completely different from that of glucose. It is clear that overall conversion rate of
gluconic acid is much lower than that of glucose over both catalysts. More critically,
the apparent conversion rates for gluconic acid decrease dramatically when its
concentration is higher than 0.3 mol/L. Therefore, the substrate inhibition effect of
gluconic acid is significantly stronger compared with glucose molecule. The big orange
22
ACS Paragon Plus Environment

Page 23 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
area in Figure 8c obviously illustrates that this is a typical substrate inhibitory effect
caused by a strong π-π interaction. This observation implies that for gluconic acid
conversion, π-activation for C=O bond is important for oxidation step.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In particularly, initial reaction rate measured on Pt/TiO2 catalyst reaches maximum
value at 340 mmol/g/h, which is surprisingly higher than AuPt/TiO2-7.5 catalyst. But
the curve drops much more rapidly when concentration of gluconic acid is higher than
0.3 mol/L. Therefore, the substrate inhibition effect of gluconic acid on Pt/TiO2 is
stronger than on AuPt/TiO2-7.5 surface.
3.4 Plausible reaction mechanism
(1) Role of Au on σ-σ and π-π bond interaction. As already confirmed (Table 1,
Figures 2, 4 and 7), the presence of Au in Pt clusters creates more active sites for both
σ-σ and π-π bond activation. HR-TEM images in Figures 4 (a-d) and S2 show the lattice
distortion at Au-Pt interfaces. In addition, the devolution of the XPS spectra of the
mono and bimetallic samples in Figure 5 have confirmed that, there exists the electronic
reconfiguration at AuPt/TiO2 samples. However, the detailed picture for them are still
not clear at this stage. Therefore, we performed DFT calculation on metal-metal
bonding and Mulliken charge distribution to understand plausible electronic and
structural reconfiguration in bimetallic AuPt clusters. Therefore, reaction mechanism
on catalyst surface can be established accordingly. As shown in Figure 9a, it is found
that Pt-Pt bond displays a dramatic distortion when Au is introduced to Pt clusters.
Specifically, the average Pt-Pt bond length increases from 0.27 nm to approximately
0.28 nm, with a wider distribution from < 0.27 nm to 0.3 nm. The widened distribution
of Pt-Pt bond confirms the lattice distortion induced by Pt-Au interfacial strain. Such
distortion is found in both random alloy (AuPt-1 and AuPt-2 in Figure 9) and ordered
alloyed structures (AuPt-3 and AuPt-4 in Figure 9). The main difference in random and
ordered alloy is that, Pt-Pt lattice tends to undergo both lattice compression and
expansion due to the introduction of Au in the former case (phase segregation), while
Pt-Pt preferably tends to expand in ordered alloy case. Such lattice distortion
contributes to the formation of twinned structures (TEM images in Figure 4) and more
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 40
1
2
3
4
5
6
active sites for activation of glucose and gluconic acid molecules. Thus catalytic
activity for oxidation reactions can be enhanced by almost two-fold (Table 1, Figure 2).
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Metal-metal bond and surface charge distribution of Pt and AuPt clusters. (a) Pt-Pt
and Pt-Au length, (b) surface electronic charge for Pt and Au sites
The strain at Au-Pt interface also induces electronic redistribution. It is not surprising
to find that Au atoms display more negative charges while overall charge for Pt sites
are more positive (Figures 9b). Detailed inspection on charge distribution on Pt sites
further show that both positively and negatively charged Pt sites co-exist in bimetallic
AuPt clusters, while Au sites are shifting towards more negatively charged
characteristics. Therefore, it can be seen there is electron transfer from Pt to Au in the
AuPt alloy system. Such electronic redistribution contributes that the positively charged
Pt2+ content increases, while the Auδ+ content slightly reduces in AuPt/TiO2 samples
24
ACS Paragon Plus Environment

Page 25 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
(XPS spectra in Figure 5). This observation indicates that, there actually exists distortion
induced phase segregation for Pt phase in bimetallic AuPt clusters. In other words, two
categories of catalytically active Pt sites are available. The more positively charged Pt
sites is more favorable for π-π bond formation with -C=O group. As a result, C-C
cleavage (decarboxylation reaction) induced by such interaction is facilitated on
bimetallic AuPt/TiO2 catalysts, which eventually leads to the formation of smaller
products.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In contrast, the relatively negatively charged Pt site are capable for nucleophilic attack
on C-H bond (σ-σ activation), therefore oxidation of aldehyde (-CHO) and terminal
hydroxyl (-CH2-OH) groups to carboxylic groups can be achieved under mild
conditions. The synergism of the above-mentioned features on bimetallic AuPt/TiO2
catalysts leads to σ-σ and π-π bond activation, selectivity for TA can thus be enhanced
eventually. Results from DFT calculation is consistent with the observation from UV-
Vis spectra shown in Figure 7 and size-dependency trend presented in Figure 8.
Figure 10. Orbital hybridization for Pt atom induced lattice and electronic reconfiguration in
bimetallic AuPt cluster (orbital hybridization refer to ref 59)
(2) Reaction mechanism. We also analyzed orbital hybridization of Pt atom in
bimetallic AuPt clusters, to further understand the underlying mechanism for the
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 40
1
2
3
4
5
6
7
8
9
synergism of bimetallic AuPt/TiO2 catalysts during glucose conversion. It is known that
Pt active center tends to formulate tetrahedral coordination with dxy, dxz and dyz orbitals
elevated to relatively higher energy levels.59 This is the reason that Pt catalysts easily
interact with -C=O groups via π-π bonding. The presence of Au atom leads to strong
lattice strain at Pt-Au interface. Such structural uniqueness lead to altered binding
energies (BE, see Figure 5). As DFT results shown in Figure 9, lattice mismatch
between Pt and Au results in structural distortion with elongated Pt-Pt bond at
equatorial orientation at Pt-Au interface. Such structural evolution induces distortion of
electronic orbitals.59 Based on the schematic description of ball-model in Figure 10, the
elongated Pt-Pt bond at near equatorial of tetrahedral coordination induces further
orbital split of dx2-y2 and dz2 for σ-bonding. Such orbital reconfiguration leads to
following two key outcomes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(i) Enhanced C-H bond activation. Such upshift (orbital split of dx2-y2 and dz2)
minimizes the energetic difference for crystal field splitting (△’ in Figure 10).
Considering the fact that Au atoms have relatively higher electronegativity than Pt
elements, the unpaired electron at dxy, dxz and dyz levels has a strong tendency for
electron transfer (coupling) to adjacent Au site. As a result, the paired electrons at dz2
orbital can be relatively easier to excite to dxy, dxz and dyz levels. Such facile electron
excitation undoubtedly induces Jahn-Teller distortion in Pt orbitals.60-63 The lone
electron at dz2 orbital is very active for σ-bonding formation via perpendicular
orientation to the surface of Pt facet. This is the underlying mechanism for enhanced
catalytic activity for oxidation of C-H bond (Figure 11).
Figure 11. Schematic description of glucose oxidation to gluconic acid over monometallic Pt
26
ACS Paragon Plus Environment

Page 27 of 40
ACS Catalysis
1
2
3
4
and bimetallic AuPt catalysts
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 12. Schematic description for C-C cleavage of gluconic acid over monometallic Pt and
bimetallic AuPt catalysts
(ii) Facile C-C cleavage. Strong π-π electron sharing with C=O bonds on the surface
of noble metals, leads to either inhibited surface for consecutive reactions,22, 64 or C-C
cleavage at terminal position of carbon chain to CO2.65, 66 But π-π interaction is
weakened on AuPt/TiO2 catalysts to some extent. This is because that the unpaired
electrons are excited from dz2 prior to π-π bond formation (Figures 10 and 12), such
electronic reconfiguration induced π-π electron sharing results in C-C cleavage, rather
than substrate inhibition effect. This is the underlying reason that, conversion of both
glucose and gluconic acid over bimetallic AuPt/TiO2-7.5 catalyst displays a less π-π
bonding-determined trend compared with monometallic Pt/TiO2 catalyst (Figures 8b
and 8c).
3.5 Kinetic modeling on Pt/TiO2 and AuPt/TiO2-7.5 catalysts
Since the electronic reconfiguration effect on C-H and C-C bond cleavage has been
clarified by both experimental and computational studies, we were interested in
obtaining the activation energies for conversion of glucose and gluconic acid on both
Pt/TiO2 and AuPt/TiO2-7.5 catalysts. Kinetic data on the initial reaction rates of glucose
and gluconic acid is presented in Figures 13a and 13b. Data fitting and regression
suggest that conversion of both substrates over Pt/TiO2 and AuPt/TiO2-7.5 catalysts
follows a 1nd order kinetics (Figures 13a and 13b), which agrees well with previous
findings.16, 25 It is found from inset table of Figure 13 that, activation of glucose
molecule over AuPt/TiO2-7.5 catalyst needs to overcome much higher energy barrier
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 40
1
2
3
4
5
6
7
8
9
(90.6 kJ/mol) than Pt/TiO2 catalyst (36.3 kJ/mol). This result appears to be inconsistent
with reaction results shown in Table 1 and Figure 2. Detailed inspection on kinetic data
(Figure 13a) actually reveals that, the pre-exponential factor (A) of rate constant of
glucose oxidation on AuPt/TiO2-7.5 catalyst is 108 times that of Pt/TiO2, suggesting
that the presence of Au on Pt surface creates additional active sites for effective glucose
conversion to primary and secondary products. This result is consistent with Figures 4
and 8. As already confirmed (Table 1, Figures 2, 4 and 7), the presence of Au in Pt
clusters creates more active sites for both σ-σ and π-π bond activation. On the one hand,
Au is the active sites for σ-σ activation. On the other hand, the addition of Au causes
lattice distortion at Au-Pt interfaces, leading to the electronic reconfiguration at
AuPt/TiO2 samples. Just as the XPS and DFT discussions, such electronic redistribution
contributes that the positively charged Pt2+ content increases. The positively charged
Pt sites are more conducive to the formation of π-π bonds with -C=O groups. For
oxidation of glucose, σ-activation for C-H bond is important for oxidation step to form
gluconic acid (as seen in Figure 8b). However, π-activation for C=O bond is important
for oxidation of gluconic acid (as seen in Figure 8c). Obviously, the addition of Au
provides more σ-σ active sites than π-π. So the number of pre-exponential factor for
oxidation of glucose is higher than gluconic acid oxidation. Similar observation has
also been made for conversion of gluconic acid over the two catalysts (inset table of
Figure 13).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28
ACS Paragon Plus Environment

Page 29 of 40
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 13. Kinetic data on the initial reaction rates of (a) glucose and (b) gluconic acid
conversion on Pt/TiO2 and AuPt/TiO2-7.5 catalysts at three temperatures (100, 110, 120 oC),
A: pre-exponential factor.
Clearly, the addition of Au changes the reaction behaviors. We further studied the
kinetics for C-C & C-H bond cleavage reactions based on the reaction rates and product
distribution of gluconic acid conversion (Figure 14 and Table S3, the calculation
method described in supporting information). It is found in Figure 14a that, activation
of C-C cleavage over AuPt/TiO2-7.5 catalyst needs to overcome a relatively higher
energy barrier (64.2 kJ/mol) than Pt/TiO2 catalyst (35.9 kJ/mol). This is because the
interaction of -C=O with Pt is weakened due to adjacent Au sites.67 This observation
reveals that the introduction of Au to Pt surface can efficiently prevent irreversible
29
ACS Paragon Plus Environment