Selective oxidation of monoethanolamine to glycine over supported gold catalysts: The influence of support and the promoting effect of polyvinyl alcohol

Article abstract of DOI:10.1016/j.mcat.2019.03.011

Glycine is an important fine chemical used in many fields, and the traditional synthetic methods like Strecker synthesis and ammoniation of chloroacetic acid use highly toxic reagents or produce equal molar byproducts. Herein, selective aerobic oxidation of monoethanolamine (MEA) to glycine using Au catalysts supported on various supports, including Al₂O₃, SiO₂, TiO₂, ZnO, and ZrO₂, was investigated and the correlation between acid-base properties of the catalysts and catalytic performance was established. Catalysts with higher base content exhibited higher initial activity and that with lower acid content gave higher glycine selectivity. The influence of preparation methods was revealed using Au/ZrO₂ with the best catalytic performance and it was demonstrated that the presence of polyvinyl alcohol (PVA) has a significant promoting effect. In-situ FTIR and ¹H NMR analysis revealed that the hydrogen bonds between PVA and MEA can enhance adsorption of MEA on catalysts, resulting in doubled turnover frequency (TOF) and improved MEA conversion; and the preferential hydrogen bonds between the amino group of MEA and the hydroxyl of PVA can prevent the coordination of amino group with Au nanoparticles, favoring reaction of the hydroxyl group of MEA on the active sites, accounting for the enhanced glycine selectivity. The reaction conditions were optimized and the optimal yield of glycine was 95%.

Full text of DOI:10.1016/j.mcat.2019.03.011

MolecularCatalysis469(2019)131–143
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat
Selective oxidation of monoethanolamine to glycine over supported gold
catalysts: The influence of support and the promoting effect of polyvinyl
alcohol
Xiangzhan Menga,b, Yinge Baib, Haiyang Xub,c, Yongqiang Zhangb, Chunshan Lib, Hui Wangb,⁎
,
Zengxi Lia,⁎
a School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100049, PR China
b Beijing Key Laboratory of Ionic Liquids Clean Process, State Key Laboratory of Multiphase Complex Systems, CAS Key Laboratory of Green Process and Engineering,
Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, PR China
c College of Papermaking Science and Technology, Tianjin University of Science and Technology, Tianjin, 300457, PR China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Glycine is an important fine chemical used in many fields, and the traditional synthetic methods like Strecker
synthesis and ammoniation of chloroacetic acid use highly toxic reagents or produce equal molar byproducts.
Herein, selective aerobic oxidation of monoethanolamine (MEA) to glycine using Au catalysts supported on
various supports, including Al2O3, SiO2, TiO2, ZnO, and ZrO2, was investigated and the correlation between acid-
base properties of the catalysts and catalytic performance was established. Catalysts with higher base content
exhibited higher initial activity and that with lower acid content gave higher glycine selectivity. The influence of
preparation methods was revealed using Au/ZrO2 with the best catalytic performance and it was demonstrated
that the presence of polyvinyl alcohol (PVA) has a significant promoting effect. In-situ FTIR and 1H NMR analysis
revealed that the hydrogen bonds between PVA and MEA can enhance adsorption of MEA on catalysts, resulting
in doubled turnover frequency (TOF) and improved MEA conversion; and the preferential hydrogen bonds be-
tween the amino group of MEA and the hydroxyl of PVA can prevent the coordination of amino group with Au
nanoparticles, favoring reaction of the hydroxyl group of MEA on the active sites, accounting for the enhanced
glycine selectivity. The reaction conditions were optimized and the optimal yield of glycine was 95%.
Au nanoparticles
MEA oxidation
Glycine
Polyvinyl alcohol
Promoting effect
Introduction
Ammoniation of chloroacetic acid to produce glycine is widely utilized
in many countries including China, the world’s main supplier for gly-
Amino acids and their derivatives play important roles in both
biology and chemistry fields, and they are mainly obtained from protein
enzymolysis or, as an alternative, by chemical synthesis [1]. Traditional
chemical processes for amino acid synthesis involve amination of halo
acids [2–4], Gabriel synthesis [5], Sorensen method [6], and Strecker
synthesis [7], which are complicated and not environmental benign due
to the use of toxic reagents, catalysts or the formed by-products. For
example, Gabriel synthesis involves protection and deprotection steps
since the unprotected amine groups can undergo side reactions, making
the procedure more complicated with reduced atom efficiency. Strecker
synthesis uses highly toxic cyanide or hydrocyanic acid as the reagent,
limiting its wide application in modern sustainable society.
cine with an annual output of 600,000 tons [9]. However, the hexam-
ethylenetetramine (HMTA) catalyst used in chloroacetic acid ammo-
niation would decompose during the reaction and product purification
processes, resulting in the formation of formaldehyde and increased
production cost. Besides, equal molar amount of ammonium chloride
would form as the byproduct, and lots of volatile methanol is needed in
order to get purified glycine product. These drawbacks restrict the
application of chloroacetic acid ammoniation to produce glycine in the
upcoming future, especially in the context of green chemistry and the
atom economy of modern society [10].
Although hydrolyzing the nitrile group of amino-nitrile under
moderate conditions using microorganisms seems to be a green method
to producing amino acids, this route has disadvantages of low activity
per unit time and difficulty in recovery of the ammonia by-product
Glycine is the simplest amino acid and an important compound in
chemical, food, pharmaceutical, and agricultural industries [8].
⁎ Corresponding authors.
E-mail addresses: huiwang@ipe.ac.cn (H. Wang), lizengxi@ucas.ac.cn (Z. Li).
https://doi.org/10.1016/j.mcat.2019.03.011
Received 10 January 2019; Received in revised form 5 March 2019; Accepted 9 March 2019
2468-8231/©2019ElsevierB.V.Allrightsreserved.

X. Meng, et al.
MolecularCatalysis469(2019)131–143
[11,12]. Amination of biomass-derived α-hydroxyl acids is also a sus-
tainable approach for the direct synthesis of amino acids, however,
severe reaction conditions like temperature as high as 220 °C are re-
quired [13,14]. The oxidation of aminoalcohols is another promising
synthetic pathway for amino acids under moderate conditions without
the usage of fatal toxic reactants or the formation of chloric byproducts
that would contaminate the product [15,16]. In 1943, Billman reported
that glycine could be made from monoethanolamine (MEA) by first
protecting the amino group with acyl radical, then oxidizing the hy-
droxyl group by acidified potassium dichromate, followed by removing
the acyl radical after the formation of the carboxyl group [17]. This
method was then modified by replacing the acyl radical with acid and
converting the amino group into a substituted ammonium ion before
oxidizing the hydroxyl group to the carboxyl group [18]. Among the
tested oxidants, potassium permanganate showed the best performance
in the acid medium [18]. However, both methods mentioned above
involve protection and deprotection steps using stoichiometric strong
and toxic oxidants, which generate a large amount of waste and pol-
lution.
Recently, transition metals catalyzed oxidation of primary or sec-
ondary alcohols to aldehydes, carboxylic acids or ketones has been
intensively investigated [19–22]. A copper-containing system was re-
ported to catalyze the conversion of amino alcohol to amino acid salts
with a large amount of catalyst at higher temperatures (e.g., 160 °C)
[23]. Biella et al. studied the oxidation of various aminoalcohols in the
presence of oxygen using supported Au, Pd and Pt nanoparticles as the
catalysts, among which Au nanoparticles showed the highest activity,
while amino groups could irreversibly adsorb on the surface of Pt and
Pd metals, resulting in deactivation of the Pt and Pd catalysts [15,16]. A
series of Au supported catalysts, including Au/Al2O3, Au/MgAl2O4, Au/
C, etc., were screened as the catalysts for liquid phase oxidation of
aminoalcohols and it was concluded that the substituent groups on the
nitrogen atoms and reaction conditions were of vital importance for the
catalytic performance [24,25]. However, effects of the support nature
and preparation methods on the catalytic performance still need to be
systematically explored and the yields of amino acids should be im-
proved.
NaBH4 (＞96%), and AgNO3 (≥99.8%) were supplied by Sinopharm
Chemicals. Polyvinyl alcohol (PVA; with a molecular weight of 10,000,
80% hydrolyzed) from Sigma-Aldrich (Shanghai, China) was used. MEA
(＞99.0%) and NaOH (＞99.0%) were from Xilong Scientific Co., Ltd.
(Guangdong, China). O2 (99.9%) was supplied by Beijing Beiwen Gas
Factory. Maleic acid (internal standard for quantitative NMR analysis)
and all other standard samples were purchased from Sigma-Aldrich
Trading Co., Ltd. (Shanghai, China).
Catalyst preparation
Sol-immobilization (SI) method
Au catalysts supported on different supports were prepared using
the SI method as described in literature [26,27]. In a typical prepara-
tion process, the required amount of the protecting agent PVA solution
(1 wt%) was added to the aqueous solution of HAuCl4 (0.24 mmol/L),
with the PVA/Au weight ratio of 1.2:1 at room temperature under
vigorous stirring. The obtained solution was then left under stirring for
5 min. A freshly prepared 0.1 M solution of NaBH4 (NaBH4/Au (mol/
mol) = 5) was added to the above solution to form a dark orange-
brown gold sol. After sol generation for 30 min, the colloid gold solu-
tion was immobilized by adding the support under vigorous stirring,
with the amount of supporting material being calculated to give a final
metal loading of 1% by mass. For the loading of gold on SiO2 support,
the pH of colloid gold needed to be adjusted to 3.0 to achieve similar Au
loading content [27]. After 2 h, the slurry was filtered and the obtained
catalyst was thoroughly washed with deionized water to remove all the
water-soluble species. Finally, the catalysts were dried in air at 80 °C
overnight.
SI-calcinated method
In this method, the obtained Au catalysts prepared by SI method
were calcinated at 250 °C for 3 h in static air to get the calcinated
samples. This method for catalyst preparation was named as SI-calci-
nated method.
Deposition–precipitation (DP) method
In this study, gold nanoparticles supported on different kinds of
metal oxides (i.e., Al2O3, SiO2, TiO2, ZnO, and ZrO2) were prepared and
their physicochemical properties were explored using XPS, TEM, TPD,
ICP, etc. Catalytic performance of these catalysts in the selective oxi-
dation of MEA to glycine (Scheme 1) was studied. The influence of
preparation methods on the catalyst property was further revealed by
using the support with the best performance. The role of the protective
agent in improving the catalytic activity and glycine selectivity was
explored with in situ FTIR, and NMR analysis. Optimization of reaction
parameters, including reaction temperature, NaOH/MEA ratio, sub-
strate/metal ratio, and concentration of MEA, was carried out to give
the best conditions.
Catalysts prepared using the deposition-precipitation procedure was
similar to the method described in literature [28,29]. Briefly, a solution
of HAuCl4·3H2O (0.24 mmol/L) was adjusted to pH of 8.0 by adding
NaOH solution (0.1 M). Subsequently, the support with a mass of 1.0 g
was added to the solution under vigorous stirring and the obtained
slurry was readjusted by NaOH solution to make sure the pH was 8.0.
After stirring for 2 h, the sample was filtered and washed with an excess
amount of deionized water until no chloride could be detected by ti-
tration with an aqueous AgNO3 solution. The obtained catalyst was
dried overnight at 80 °C and then calcined for 3 h at 250 °C in air.
Impregnation (IMP) method
The catalyst was also prepared by impregnation method following
procedure described in literature [30,31]. A required amount of the
support was mixed with 100 mL of aqueous solution of HAuCl4·3H2O
(0.24 mmol/L), and all the water in the mixture was evaporated at 80 °C
with vigorous stirring. The solid was dried in an oven at 80 °C for 12 h,
which was then treated with 0.1 M of NaBH4 aqueous solution (the
molar ratio of NaBH4 to metal was 5) at room temperature for 2 h. Then
the catalyst was filtered, washed and dried overnight at 80 °C in air to
give the Au catalysts.
Experimental
Materials
All chemicals (AR grade) were purchased commercially and used as
received unless otherwise noted. The powder supports, including γ-
Al2O3, ZrO2, and ZnO, were purchased from Shanghai Macklin
Biochemical Co. Ltd. (Shanghai, China). TiO2, with an average particle
size of 25 nm, was from Degussa (Frankfurt am Main, Germany). SiO2
was from Sigma-Aldrich (Shanghai, China). HAuCl4·4H2O (99.99%),
Characterization
Powder X-ray diffraction (XRD) of the catalysts was performed on
Rigaku RINT 2500 Diffractometer using monochromated Cu Kα radia-
tion with a scan speed of 15° min−1 and a scan range of 5-90° at 30 kV
and 15 mA. The X-ray photoelectron spectroscopy (XPS) studies were
performed on a PHI-5400 spectrometer with an Al Kα source. The
Scheme 1. Reaction scheme for the oxidation of MEA to glycine.
132

X. Meng, et al.
MolecularCatalysis469(2019)131–143
binding energy scale was calibrated relative to the C1s peak (285.0 eV)
of contaminant carbon [27,32]. The images from transmission electron
microscope (TEM) were collected using a JEOL JEM-2100 F operated at
200 kV. Samples were deposited on the TEM grids after dispersion in
ethanol. At least 200 particles were randomly measured to determine
the average diameter of Au nanoparticles for every sample. The catalyst
surface area, pore volume, and average pore size were measured at
−196 °C from the N2 adsorption-desorption isotherm on a Micro-
meritics ASAP 2460 apparatus. Prior to the measurements, all samples
were degassed overnight at 300 °C. The Au loading of the catalysts was
determined by inductively coupled plasma optical emission spectro-
scopy (ICP-OES) using an IRIS Intrepid II XSP (Thermofisher, USA). 1H
NMR spectra were recorded on a Bruker AVANCE III HD NMR spec-
trometer operating at 600 MHz. High resolution mass spectra (HRMS
(ESI-TOF)) were collected on a Brucker high resolution mass spectro-
meter (micrOTOF-QII).
Temperature-programmed desorption of ammonia and carbon di-
oxide (NH3- and CO2-TPD) experiments were carried out using Cat-Lab
(BEL JAPAN, INC.) equipped with an online QIC-200 quadrupole mass
spectrometer (Inprocess Instruments, GAM 200) as the detector. The
NH3 desorption profiles were measured by fixing the m/z signal at 15,
and those for CO2 desorption were measured by fixing the m/z signal
for molecular CO2 at 44. The loadings of all samples were around 50 mg
and the calculations for the total acidity and basicity concentration of
Au catalysts were based on the weight of the catalysts. In-situ FTIR
spectra were recorded on a Thermo-Nicolet (Nicolet 6700) spectro-
meter equipped with the high temperature reaction chamber and
transmission accessory. Spectra with 4 cm−1 resolution and 64 scans
were recorded at room temperature using ZnSe windows in the range
from 4000 cm−1 to 650 cm−1. The powders (KBr and sample) were
pressed at 10 MPa for 10 min and loaded into a home-made stainless
steel reaction chamber. Then the catalyst was pretreated by heating to
120 °C with a heating rate of 8 °C min-1 in vacuum for 2 h and cooled to
room temperature. Adsorption behavior of MEA on the catalysts were
started by flowing the He/MEA gas mixture at a flow rate of 5 mL·min-1
and FTIR spectra were recorded at 30, 60, 120, 180, 240, 300, 360 min
for analysis.
Oxidation of MEA
The oxidation reaction of MEA was carried out in a 100 mL batch
type Teflon-lined stainless-steel autoclave, which was connected to an
8 L reservoir of O2 with constant pressure. In a typical procedure, 40 mL
aqueous MEA solution (0.3 M), 0.96 g NaOH (molar ratio of NaOH/
MEA was 2), and 0.236 g catalyst (molar ratio of MEA/Au was 1000)
were loaded into the autoclave successively. The autoclave was purged
3 times with O2 and then pressurized at desired oxygen pressure at
room temperature. Once the system reached the desired temperature,
the reaction was initiated by vigorous stirring with a magnetic stirrer
and samples were taken at 15, 30, 60, 120, 180 min for analysis.
The products were identified by HRMS and comparing the 1H NMR
spectra of the reaction solutions with those for standard samples.
Conversion of MEA and selectivity of glycine was quantitatively ana-
lyzed using 1H NMR as described in the literature [24,25]. In the pre-
paration of the samples for analysis, 34.8 mg of maleic acid (standard
for quantitative NMR) was added to each sample solution (2 mL) as the
internal standard to quantify the products. 1H NMR spectra were re-
corded with suppressed water signal using a low power PRESAT pulse
to minimize signal distortions [24,33]. The concentrations of the re-
actant and products were determined by comparing the integration of
the products signals to standard ones.
Fig. 1. TEM micrographs and corresponding Au particle size distributions for
Au catalysts: a -Au/Al2O3, b -Au/TiO2, c -Au/ZnO, d -Au/ZrO2, e -Au/SiO2.
Information (ESI). Au/SiO2 and Au/Al2O3 have higher surface area
compared with other catalysts. All the supports are nanosized particles
with diameters from 10 nm to 50 nm. XRD patterns of Au/TiO2, Au/
ZnO, and Au/ZrO2 in Fig. S1, ESI show well-crystallized diffraction
peaks of the supports, while the diffraction peaks of Au/SiO2 and Au/
Al2O3 are broad, indicating that these supports are not well crystallized.
No diffraction peak of Au was found in any catalyst, indicating that Au
particles are well dispersed.
Support effect
Results and discussion
Nature of the support usually plays an important role in supported
Au catalysts for the oxidation of alcohols, and an in-depth under-
standing of the support effect on the catalyst property and activity
could help to rationally design efficient catalysts. Study of the support
The surface area, pore volume, pore diameter and support particle
size of the five catalysts are shown in Table S1, Electronic Supporting
133

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Table 1
Selective oxidation of MEA over various Au catalysts.[a].
Catalyst
Au loading (wt%)
Au particle
size (nm)
Initial MEA conversion[b] (%)
Selectivity (%)[c]
GLY
ANHA
FA
Others
No[d]
–
–
4
0
–
–
–
Au/ZnO
Au/Al2O3
Au/ZrO2
Au/TiO2
Au/SiO2
0.86
0.71
0.77
0.76
0.68
2.7
2.6
2.8
2.7
4.6
49
39
34
33
9
46
47
59
55
30
41
34
29
36
22
2
11
16
8
6
33
2.5
3.8
3
15
[a] Catalysts were prepared by SI method; reaction conditions: CMEA = 0.3 M, MEA/Au = 1000/1 (mol/mol), NaOH/Au = 2, T =50 °C, P(O2) = 0.5 Mpa, t =15 min.
[b] Calculated after reaction for 15 min.
[c] GLY: glycine, others include glycolic acid, glycylglycinate, 2-iminoethanol, etc.
[d] Selectivities of some by-products can not be determined accurately due to their very low concentrations.
effects requires that the catalysts are in uniform Au particle size. The
supported Au catalysts were prepared by deposition of Au colloids with
PVA as a protective agent using SI method as this method could produce
Au particles with similar size on different supports [32]. Fig. 1 displays
the representative TEM micrographs and corresponding Au particle size
distributions of the prepared catalysts. An average Au particle size of
2.7 nm with similar particle size distributions (ranging from 1 nm to
5 nm) were obtained on the Al2O3, TiO2, ZnO, and ZrO2 supports, while
the average diameter of gold on SiO2 was 4.6 nm with particle size
distribution from 1 nm to 8 nm. XRD patterns of the catalysts in Fig. S1,
ESI showed no characteristic peaks of metallic gold at 2θ around 44.4°,
indicating that the size of Au particles was less than 5 nm [34], in good
agreement with the TEM observation.
Oxidation of MEA in the presence of these catalysts was carried out
in an autoclave as described in the Experimental section. Products from
the MEA oxidation were qualified by NMR and HRMS (Figs. S2 and S3,
ESI), which were proved to be glycine, 2-amino-N-(2-hydroxyethyl)
acetamide (ANHA), formic acid (FA), etc. The performance of these
catalysts in the selective oxidation of MEA is shown in Table 1. Control
trial was conducted with the absence of the catalyst and the result
showed that under these specific conditions, MEA conversion was only
4% and no glycine was detected. With the addition of Au catalysts, both
MEA conversion and glycine selectivity were improved. Among the
tested catalysts, Au/SiO2 gave much lower activity with only 9% MEA
conversion and 30% selectivity for glycine in 15 min. Au/ZnO showed
the best catalytic activity with 49% MEA conversion, while Au/ZrO2
gave the highest selectivity towards glycine (59%), followed by Au/
TiO2, Au/Al2O3, and Au/ZnO.
The electronic state of the catalysts and the gold content on the
surface were analyzed by XPS, and the results were shown in Fig. 2 and
Table 2. All spectra of the catalysts showed two distinct XPS signals
with binding energies around 83.9 eV and 87.7 eV, corresponding to the
characteristic electrons of 4f7/2 and 4f5/2 in the metallic Au of the
catalysts [35]. The binding energies of Au 4f7/2 were 83.5, 83.4, 83.5
and 83.5 eV for Au/ZrO2, Au/Al2O3, Au/ZnO, and Au/TiO2, respec-
tively. Compared to the binding energy of bulk Au metal at 83.9 eV,
these lower binding energies of Au 4f7/2 in these catalysts indicate
modification of the electronic states of Au due to the interaction be-
tween the gold nanoparticles and nanosized supports [36–40]. For Au/
SiO2, the binding energy of Au 4f7/2 was 83.8 eV, quite close to that of
bulk Au, indicating that the interaction between Au and SiO2 was weak.
The valence state of Au for freshly prepared gold catalysts was also
dependent on the identity of the support even the gold colloid was
prepared by subjecting Au3+ to excessive strong reductive NaBH4 be-
fore deposition. The gold nanoparticles tended to keep in the metallic
state on SiO2, Al2O3, and ZnO, while 17.0% and 5.5% of the gold
species were in the oxidized state for those on supports like TiO2 and
ZrO2, respectively (Table 2).
catalytic activity of alcohol oxidation over supported gold catalysts
[41,42]. Temperature programmed desorption (TPD) of NH3 and CO2
over these Au catalysts was conducted to examine the acid and base
properties of the catalysts and the results are shown in Figs. 3 and 4,
respectively. All the catalysts exhibited abundant acid and base sites
except Au/SiO2, which showed negligible content of acidity or basicity
[32]. From the temperatures and intensities of the NH3 desorption
peaks in Fig. 3, we can see that Au/Al2O3 possessed various sites of
weak, medium and strong acidity, and that both Au/ZnO and Au/ZrO2
had only medium acid sites, with the concentration of acid sites on Au/
ZrO2 being relatively lower as indicated by the intensity. Au/TiO2
possessed some weak acid sites. CO2-TPD profiles in Fig. 4 suggest that
Au/ZnO and Au/Al2O3 possessed a high concentration of basic sites,
while Au/ZrO2 has a similar concentration of basic sites to Au/TiO2.
There was no obvious desorption of CO2 detected on Au/SiO2, in-
dicating that this catalyst had no obvious basicity.
The relationship between the initial MEA conversion and the total
basicity concentration of the catalysts was shown in Fig. 5 (Au/SiO2
was not included in the plots because of the obvious difference in Au
particle size and the solution pH value during preparation). It indicates
that the initial MEA conversion increased almost linearly with the
catalyst basicity except Au/TiO2. For Au/TiO2, it had similar con-
centration of basicity to that of Au/Al2O3 (0.105 vs. 0.109 mmol/g), but
it exhibited relatively lower initial MEA conversion (33% vs. 39%), due
to the less proportion of Au0 on the surface of Au/TiO2 (83% for Au/
TiO2 vs.100% for Au/Al2O3) as reduced gold was the active component
for the reactant activation [19,43]. As shown in Fig. 5, the selectivity of
glycine seems to be negatively correlated with the acidity concentration
of the catalysts. Possessing the highest acidity concentration
(0.159 mmol/g), Au/ZnO showed the lowest selectivity for glycine
(46%) and the highest selectivity for ANHA (41%), which was formed
by the oxidative coupling of hydroxyl group of MEA with amino group
of another MEA molecule (discussed in Section 3.5) as gold nano-
particles can also catalyze amide synthesis from alcohols and amines
[44,45]. Au/ZrO2, with an acidity concentration of 0.125 mmol/g, ex-
hibited the highest selectivity for glycine (59%) and the low selectivity
for ANHA (29%), attributed to the relatively low adsorption and acti-
vation amount of the amino group [46].
To test the durability of Au/Al2O3, Au/TiO2, Au/ZnO, and Au/ZrO2,
which gave good MEA conversion or glycine selectivity in 15 min’ re-
action (Table 1), the oxidation of MEA was carried out for 3 h over these
catalysts and the changes of MEA conversion and selectivities of pro-
ducts during the reaction were shown in Figs. 6 and S4. MEA conver-
sion first increased with reaction time and then reached maximum
values for all the tested catalysts. The same phenomena were also ob-
served by Prati et al. and Lu et al., the invariable conversions of MEA or
benzyl alcohol were attributed to deactivation of the catalysts
[24,25,47]. As shown in Fig. 6, both Au/TiO2 and Au/ZrO2 showed
relatively higher stability, and MEA conversion reached maximum
Acid-base property of the supporting materials usually relates to the
134

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Fig. 2. Au 4f XPS spectra of the supported Au catalysts.
Table 2
values of 89% and 81% after 2 h, respectively. However, it seems that
Au/Al2O3 and Au/ZnO deactivated in 1 h, when the maximum MEA
conversion reached 63% and 62%, respectively. In fact, the amino
groups can strongly interact with the supported Pt and Pd metals and
decrease the durability of the gold catalysts, as demonstrated by the
contrastive study of the oxidations of ethylene glycol and ethanolamine
[15,25]. Au/Al2O3 and Au/ZnO catalysts with higher acidity con-
centrations (0.21 and 0.29 mmol/g, respectively) compared with Au/
TiO2 and Au/ZrO2 (0.15 and 0.13 mmol/g, respectively) can adsorb
more amino groups on the surface, which could decrease the durability
of the catalysts. As shown in Fig. S4, the selectivity of glycine enhanced
with time going on, while the selectivity of ANHA decreased, which
could be attributed to the hydrolysis of ANHA to glycine in the presence
Surface content and relative surface distribution of gold species in the catalysts.
Catalyst
Ausurf (%)
Au 4f7/2 (eV)
Au 4f5/2(eV)
Au species (%)
Au0
Au+
Au/ZrO2
Au/ZnO
Au/Al2O3
Au/TiO2
Au/SiO2
0.5
0.5
0.5
0.4
0.1
83.5
83.5
83.4
83.5
83.8
87.2
87.2
87.0
87.2
87.5
94.5
100
100
83.0
100
5.5
0
0
17.0
0
135

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Fig. 6. Conversion of MEA using supported gold catalysts as a function of re-
action time (Reaction conditions: CMEA = 0.3 M, MEA/Au = 1000/1 (mol/
mol), NaOH/Au = 2, T =50 °C, P(O2) = 0.5 Mpa).
Fig. 3. NH3-TPD profiles for the supported Au catalysts.
viewed as the best catalyst and chosen as a model catalyst in the fol-
lowing study.
Comparison of catalyst preparation methods
Catalysts prepared using different methods usually have different
properties and thus catalytic activities. Au/ZrO2 prepared using four
different methods including SI, SI-calcinated, DP, and IMP methods
were tested for the selective oxidation of MEA. The results are shown in
Fig. 7. It is obvious that the preparation methods have a significant
influence on the catalytic performance with the catalyst prepared by SI
method showing the best performance, with MEA conversion and gly-
cine selectivity reaching 81% and 67%, respectively in 2 h.
To determine the possible reasons for the difference in the catalytic
performance, the most important properties of electronic state and
particle size that would affect the activity of Au/ZrO2 were character-
ized using XPS and TEM. The Au 4f spectra of all the investigated Au/
ZrO2 samples are shown in Fig. 8. Lower binding energies of Au 4f
around 83.5 eV in all samples were detected due to the modification of
the electronic states of Au by the nanosized support as mentioned be-
fore. In order to determine the possible gold states in Au/ZrO2 prepared
by different methods, the curve fitting of the Au 4f core-level spectra
was performed by using two spin-orbit split Au 4f7/2 and Au 4f5/2
components, separated by 3.7 eV, in a fixed intensity ratio of 1.33 and
reasonable half-maximum values [48]. All the samples showed two Au
4f7/2 components around 83.5 eV and 85.2 eV, which could be attrib-
uted to Au0 and Au+ species, respectively [35,49,50]. It should be
Fig. 4. CO2-TPD profiles for the supported Au catalysts.
Fig. 5. Correlation between initial MEA conversion and the total basicity con-
centration of Au catalysts (■) as well as the correlation between the selectivity
of glycine and the total acidity concentration of Au catalysts (●) (Reaction
conditions: CMEA = 0.3 M, MEA/Au = 1000/1 (mol/mol), NaOH/Au = 2, T
=50 °C, P(O2) = 0.5 Mpa, t =15 min).
of NaOH at 50 °C as verified in Section 3.4.1. The selectivity of formic
acid increased with the reaction time slowly, indicating that cleavage of
the CeC bond was a continuous process. Taking both conversion of
MEA and the selectivity of glycine into consideration, Au/ZrO2 was
Fig. 7. Effect of preparation methods for Au/ZrO2 on the oxidation of MEA
(Reaction conditions: CMEA = 0.3 M, MEA/Au = 1000/1 (mol/mol), NaOH/
Au = 2, T =50 °C, P(O2) = 0.5 Mpa, t =2 h).
136

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Fig. 8. Au 4f XPS spectra of Au/ZrO2 catalysts prepared by a) SI, b) SI-calcinated, c) DP, d) IMP methods.
noted that the relative contents of different gold species for all the in-
vestigated Au/ZrO2 samples were quite similar, for example, the pro-
portions of Au0 were 94.5%, 95.2%, 90.0%, and 95.4% for Au/ZrO2
catalysts prepared by SI, SI-calcinated, DP and IMP, respectively, in-
dicating that the preparation methods didn’t affect the electronic state
too much and that the variation in catalytic performance of the Au/
ZrO2 catalysts was not directly related to the electronic state in this
case.
Fig. 9 shows the representative images and corresponding particle
size distributions for Au/ZrO2 catalysts prepared by SI, SI-calcinated,
DP, and IMP methods. MEA conversion was inversely correlated to the
average particle sizes as shown in Fig. 7. Au/ZrO2 prepared by IMP
method had the largest particle size of 5.3 nm and the lowest activity,
while the catalyst prepared by DP method showed smaller particle and
relatively high MEA conversion. There is no obvious correlation be-
tween glycine selectivity and the average particle size. Interestingly,
Au/ZrO2 prepared by SI method showed similar average particle size to
that prepared by DP method (2.8 vs. 3.1 nm), but much higher con-
version of MEA (81% vs. 55%) and selectivity of glycine (67% vs. 51%)
were observed when using Au/ZrO2 prepared by SI method as the
catalyst. Considering that there was no significant difference of the
surface electronic state of gold as shown in Fig. 8, we attributed the
improvement in both activity and selectivity to the presence of PVA in
Au/ZrO2 obtained by SI method.
accessibility of the reactant to the active site [53]. Thick capping layer
usually leads to shielding effect on the surface of Au active sites for
catalytic reactions, while the reactant would pass through thin layer
and reach the active site [52,54].
To investigate the effect of PVA of the catalyst on the oxidation of
MEA, we used various PVA/Au weight ratios of 0.2, 0.4, 0.8, 1.2, 2.0
and 3.0 to prepare 1% Au/ZrO2 by SI method. The representative TEM
images and Au particle size distributions for Au/ZrO2 prepared with
different PVA amounts are shown in Fig. 10. It can be seen that the
average gold particle size decreased from 3.4 nm to 2.8 nm with the
increase of the PVA/Au ratio from 0.2 to 0.8. With the PVA/Au ratio
further increasing, the average gold particle size remained around
2.8 nm, indicating that excessive amount of PVA cannot reduce the
particle size further as particle growth could be well prevented and the
velocity of nucleation became the main factor to determine the particle
size when the PVA amount reached a certain point [55]. In addition, the
decreasing trend of statistical standard deviation (given in the top right
corner of every graph for Au particle size distributions) of the particle
size with the increase of PVA/Au ratio indicated that aggregation of the
nanoparticles was inhibited at higher PVA/Au ratios.
XPS studies of Au/ZrO2 prepared by SI method using various PVA/
Au ratios were performed to check the possible change of the electronic
state, and the results were shown in Table 3 and Figs. S5 and S6, ESI. As
the ratio of PVA to Au varied, there was no obvious difference in the
binding energy of Au4f, Zr3d, C1s, or O1s signals and the relative
surface distribution of the gold species, which was consistent with the
results that PVA could not lead to the change of binding energy of Au 4f
[54,56]. It should be noted that the atomic ratios of carbon on the
surface increased with the PVA/Au ratio, confirming the increase of
PVA capping layer.
The role of PVA
PVA is a commonly used effective steric stabilizer, which can be
easily adsorbed on the surfaces of the particles to form a capping layer
to protect metal nanoparticles from aggregating through steric hin-
drance of bulky molecules that induce a repulsive force against the van
der Waals attraction between two metal nanoparticles [51–53]. The
ratio of PVA to Au could influence the thickness of the protective layer,
and therefore the stability of the colloidal metal particles and
The effect of the ratio of PVA to Au on the oxidation of MEA is
presented in Table 4. MEA conversion first increased with the increase
of PVA/Au ratio, and it reached the maximum value when the ratio was
1.2. With the ratio of PVA to Au further increasing to 3.0, the MEA
137

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Fig. 9. Representative TEM micrographs and corresponding Au particle size
distributions for Au/ZrO2 catalysts prepared by a -SI, b -SI-calcinated, c -DP, d
-IMP methods.
conversion decreased. It is worth to mention that when the ratio of PVA
to Au was higher than 0.8, the average gold particle size was 2.8 nm,
implying that the change of the catalytic activity was irrelevant to the
particle size when the PVA/Au ratio varied from 0.8 to 3.0. Interest-
ingly, the MEA turnover frequency (TOF, calculated with conversions
ranging from 11%˜36% at 15 min) was doubled over Au/ZrO2 by SI
method when the PVA/Au ratio changed from 0.4 to 1.2, compared
with that over Au/ZrO2 by DP method (Entry 1, Table 4). Further in-
crease of PVA/Au ratio led to decrease of the TOF to 0.6 s−1. In addi-
tion, the selectivity of glycine continuously increased with the increase
of the PVA/Au ratio (although the increasing extent was retarded at
higher PVA/Au ratios), and the selectivity of FA was inversely pro-
portional to the PVA/Au ratio. These phenomena observed above in-
dicated the promoting effect of suitable amount of PVA in the catalytic
oxidation of MEA over gold catalysts.
In-situ FTIR spectra of MEA adsorbed on ZrO2, Au/ZrO2 prepared by
DP method (no PVA) and Au/ZrO2 prepared by SI method (prepared
using the PVA/Au ratio of 1.2) were collected to reveal the role of PVA.
As shown in Fig. 11, before introduction of MEA into the system
(time = 0), the spectrum of Au/ZrO2 prepared by SI method shows the
characteristic bands at 1731.4 cm−1 (C]O, stretching vibration) and
1384.2 cm−1 (OeH, deformation vibration), which confirmed the
presence of PVA. All samples including ZrO2, Au/ZrO2 prepared by DP
method and Au/ZrO2 prepared by SI method could adsorb MEA and the
Fig. 10. Representative TEM micrographs and corresponding Au particle size
distributions for Au/ZrO2 catalysts prepared by SI method with different PVA/
Au ratios: a -0.2, b -0.4, c -0.8, d -1.2, e -2.0, f -3.0.
amounts of adsorbed MEA increased with time, as indicated by the
intensity increase of the band for MEA C–N stretching vibration at
1076 cm−1 with increasing adsorption time. Particularly, the intensity
of the band at 1076 cm−1 increased significantly in 180 min for MEA
adsorbed on Au/ZrO2 prepared by SI method, while the intensity of the
138

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Table 3
intensity around 1595 cm−1 was observed as the adsorption amount of
MEA increased, indicating the coordination between the amino group
of adsorbed MEA and the surface of Au nanoparticles [34,57,58]. For
the FTIR spectra of MEA adsorbed on Au/ZrO2 with PVA capping layer
around Au particles, the intensity of the characteristic band for MEA at
1595 cm−1 (NeH, deformation vibration) obviously increased with the
adsorption amounts of MEA, indicating that in this case, the NeH bond
was more “free” and coordination of the MEA amino group with Au
nanoparticles was effectively prevented by PVA capping layer. In short,
the presence of PVA can enhance the adsorption of MEA on the catalysts
and prevent the coordination of the amino group with Au nanoparticles.
Therefore, the doubled TOF (from 0.7 s−1 for DP catalyst to 1.4 s−1 for
SI catalyst with the PVA/Au ratio of 1.2), improved MEA conversion
(from 55% for DP catalyst to 81% for SI catalyst with the PVA/Au ratio
of 1.2) and selectivity of glycine (from 51% for DP catalyst to 70% for SI
catalyst with the PVA/Au ratio of 3.0) for Au/ZrO2 prepared by SI
method could be attributed to the ability of PVA to enhance the ad-
sorption of substrate and mediate the interaction between MEA and the
Au nanoparticles, making the hydroxyl group of MEA preferentially
react on the active sites.
Content of carbon on the surface, the binding energy of Au 4f, and the relative
surface distribution of gold species of Au/ZrO2 with different PVA/Au ratios.
PVA/Au
C1s(%)
Au 4f7/2(eV)
Au 4f5/2(eV)
Au species (%)
Au0
Au+
0.2
0.4
0.8
1.2
2.0
3.0
15.4
14.7
15.1
16.3
16.9
19.0
83.6
83.5
83.5
83.5
83.5
83.5
87.3
87.2
87.2
87.2
87.2
87.2
93.5
95.1
93.3
94.5
95.9
91.3
6.5
4.9
6.7
5.5
4.1
8.7
Table 4
MEA oxidation with Au/ZrO2 prepared by DP method and SI method using
different PVA/Au ratios.
PVA/Au Au particle Conversion Selectivity of
Selectivity of
FA (%)[a]
TOF at
15 min
size (nm)
(%)[a]
GLY (%)[a]
(s−1
)
[b]
[c]
-
3.1
3.4
3.3
2.8
2.8
2.8
2.7
55
48
55
70
81
71
60
51
58
62
67
67
68
70
5.8
6.5
6
4.8
4.2
3.6
3.2
0.7
0.5
1.5
1.5
1.4
1.3
0.6
In order to study the reason for the enhanced MEA adsorption on the
catalysts and the prevented coordination of the MEA amino group with
Au nanoparticles by PVA capping layer, the interactions between MEA
and PVA were explored with spectroscopic analysis and simulation. 1H
NMR spectra of aqueous MEA/PVA solutions, with the same con-
centrations of MEA but different PVA/MEA ratios of 0, 1, 3, were col-
lected as shown in Fig. S7, ESI, and the chemical shifts of H-1
(NH2CH2CH2OH) and H-2 (NH2CH2CH2OH) were summarized in
Table 5. It was revealed that with the increase of the ratio of PVA to
MEA, both H-1 and H-2 shifted downfield. These trends indicated that
hydrogen bonds were formed between PVA and MEA, as the oxygen
atom in the hydroxyl group or nitrogen atom in the amino group in
MEA could provide lone pair electrons to form hydrogen bonds with the
hydroxyl hydrogen of PVA, leading to deshielding effect and downfield
shifts of H-1 and H-2 in the neighboring methylene groups. However,
the greater downfield shift of H-1 compared with H-2 (0.5 vs. 0.3 ppm)
indicated a preferential trend for the amino nitrogen atoms to interact
with the hydrogen of −OH in PVA.
0.2[d]
0.4[d]
0.8[d]
1.2[d]
2.0[d]
3.0[d]
[a] Reaction conditions: CMEA = 0.3 M, MEA/Au = 1000/1 (mol/mol), NaOH/
Au = 2, T =50 °C, P(O2) = 0.5 Mpa. Conversion and selectivities were calcu-
lated at reaction time of 2 h.
[b]TOF (s−1) = Number of MEA molecules consumed per second/Number of
exposed Au atoms, calculated after reaction for 15 min.
[c] Au/ZrO2 prepared by DP method.
[d] Au/ZrO2 prepared by SI method using different PVA/Au ratios.
In short, the presence of PVA capping layer on the catalyst was
confirmed by FTIR and XPS analysis. It can not only stabilize Au na-
noparticles but also influence the catalytic performance by the hy-
drogen bonds between PVA and the substrate. As shown in the proposed
collision models of MEA on different catalysts in Fig. 12, the enhanced
adsorption of the substrate on the catalyst significantly promoted the
catalytic activity, and the preferential interactions between the MEA
amino group and PVA favored the reaction of hydroxyl group of MEA
on the active sites. However, when the ratio of PVA to Au was higher
than 1.2, the thicker PVA capping layer could hinder diffusion of the
reactant and decrease the catalytic activity, which was consistent with
the previous report for PVA stabilized Au nanoparticles supported on
activated carbon and TiO2 [26,56].
Fig. 11. FTIR spectra of MEA adsorbed on ZrO2, Au/ZrO2 prepared by DP
method (no PVA) and Au/ZrO2 prepared by SI method (prepared using the
PVA/Au ratio of 1.2) with time (curves from top to bottom: 0, 60, 120, 180,
240, 300, 360 min for MEA adsorbed on ZrO2, Au/ZrO2 prepared by DP
method, and 0, 30, 60, 120, 180 min for MEA adsorbed on Au/ZrO2 prepared by
SI method) .
Influence of the reaction parameters
The effects of reaction parameters, including temperature, NaOH/
Table 5
same band slightly increased in 360 min for MEA adsorbed on ZrO2 and
Au/ZrO2 prepared by DP method, indicating the increased amount and
rate of MEA adsorption on Au/ZrO2 prepared by SI method due to the
presence of PVA compared with the other two. For the FTIR spectra of
MEA adsorbed on the ZrO2 support, the intensity of the band at
1595 cm−1 (NeH of MEA, deformation vibration) increased with the
increasing amounts of MEA, while for the FTIR spectra of MEA ad-
sorbed on Au/ZrO2 prepared by DP method, no obvious change of peak
Chemical shifts of H-1 (NH2CH2CH2OH) and H-2 (NH2CH2CH2OH) in the so-
lutions of MEA with different PVA/MEA ratios.
PVA/MEA
(wt/wt)
δ-CH2OH
δ-CH2NH2
Δ-CH2OH
Δ-CH2NH2
(ppm)
(ppm)
(ppm)
(ppm)
0
1
3
3.60
3.62
3.63
2.73
2.76
2.78
0
0.2
0.3
0
0.3
0.5
139

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Fig. 12. Proposed collision models of MEA on (a) naked gold nanoparticles and (b) PVA capped gold nanoparticles (N -blue; H -white; C -gray; O -red) (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).
MEA ratio, substrate/metal ratio, and MEA concentration, on the con-
version of MEA and selectivities of the products were investigated over
1% Au/ZrO2 prepared by SI method using PVA/Au ratio of 1.2.
more base to yield both glycine and MEA. The selectivity of formic acid
increased with the increase of temperature, suggesting cleavage of the
CeC bond was boosted, which was consistent with the phenomena
observed in gold catalyzed glycerol oxidation that the amount of pro-
ducts from the CeC cleavage increased by increasing the reaction
temperature [59]. Considering the activity and stability of the catalyst,
50 °C was chosen as the reaction temperature in the following study.
Effect of reaction temperature
Reaction temperature is an important parameter to influence a
chemical process. The temperature effect on MEA oxidation was studied
in the range from 10 °C to 90 °C and the results are shown in Fig. 13. As
expected, the conversion of MEA increased with the temperature in-
creasing from 10 °C to 50 °C, and reached the maximum value at 50 °C.
With the reaction temperature further increasing to 90 °C, the MEA
conversion decreased apparently from 81% at 50 °C to 59% at 90 °C,
which could be attributed to the quick removal of PVA capping layer in
water at higher temperatures, similar to the fact that Au/ZrO2 prepared
by DP method with no PVA exhibited lower MEA conversion of 55%
[26]. However, the selectivity of glycine increased with temperature in
the tested range, and the selectivity of ANHA decreased to nearly zero
at 90 °C. The reaction solutions obtained at 50 °C were further treated at
90 °C or at 50 °C with 4 more eqv. NaOH in the absence of the catalyst
for 2 h, respectively, and the results are shown in Table S2, ESI. The
content of ANHA decreased to zero and the concentrations of glycine
and MEA improved accordingly (corresponding to increased glycine
selectivity and decreased MEA conversion), indicating that ANHA could
be easily hydrolyzed under higher temperature or with the presence of
Effect of the molar ratio of NaOH to MEA
The presence of strong base has important effects on both conver-
sion and product distribution in the oxidation of alcohols [60–62]. In
this study, oxidation of MEA was carried out over Au/ZrO2 catalysts
keeping all parameters constant except the initial molar ratio of NaOH
to MEA, and the results were shown in Fig. 14. As expected, the con-
version of MEA per pass increased with increasing ratio of NaOH to
MEA. Despite the amino group can provide a basic environment, the
conversion of MEA and the selectivity of glycine were very low in the
absence of NaOH (11% and 23%, respectively). When the ratio of NaOH
to MEA increased to 1.0, the conversion of MEA dramatically increased
to a maximum value of 80% in 2 h. The need for the addition of strong
base is consistent with the situations of the glycerol oxidation, where
strong base helped the initial deprotonation of a primary alcohol group,
followed by being rapidly dehydrogenated to glyceraldehyde
[60,61,63]. For the product distribution, the selectivity of glycine
Fig. 13. Effect of reaction temperature on the oxidation of MEA (Catalyst: Au/
ZrO2 prepared by SI method using PVA/Au ratio of 1.2; reaction conditions:
Fig. 14. Effect of NaOH/MEA ratio on the oxidation of MEA (Catalyst: Au/ZrO2
prepared by SI method using PVA/Au ratio of 1.2; reaction conditions:
CMEA = 0.3 M, MEA/Au = 1000/1 (mol/mol), T =50 °C, P(O2) = 0.5 Mpa, t
=2 h).
CMEA = 0.3 M,
MEA/Au = 1000/1
(mol/mol),
NaOH/Au = 2,
P
(O2) = 0.5 Mpa, t =2 h).
140

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Fig. 15. Effect of substrate/metal molar ratio on the oxidation of MEA
(Catalyst: Au/ZrO2 prepared by SI method using PVA/Au ratio of 1.2; reaction
conditions: CMEA = 0.3 M, NaOH/Au = 4, T =50 °C, P(O2) = 0.5 Mpa, t =2 h).
Fig. 16. Effect of initial substrate concentration on the oxidation of MEA
(Catalyst: Au/ZrO2 prepared by SI method using PVA/Au ratio of 1.2; reaction
conditions: MEA/Au = 500/1 (mol/mol), NaOH/Au = 4,
T =50 °C, P
(O2) = 0.5 Mpa, t =2 h).
enhanced while the selectivity of ANHA decreased with the NaOH/MEA
molar ratio, which could be attributed to the hydrolysis of ANHA to
glycine in the presence of NaOH. Taking the MEA conversion and the
glycine selectivity into consideration, the NaOH/MEA ratio of 4.0 was
chosen for further experiments.
this work. Results (Fig. S9, ESI) indicate that the recycled catalyst ex-
hibited much lower activity than the fresh one. In addition, it was found
that the Au nanoparticles seriously aggregated after being used for 3
cycles (Fig. S10, ESI). Based on the above discussion, Au particle ag-
gregation, PVA leaching from the catalyst into the solution, and pos-
sible adsorption of the reactant on the catalyst surface led to the de-
creased activity of the recycled catalyst. Work to solve the problem of
catalyst deactivation is currently undergoing in our laboratory.
Effect of substrate/metal molar ratio
Catalyst loading is another parameter that could influence the
processing of a chemical reaction. The MEA/Au molar ratio was varied
in the range from 300 to 1500 while the other reaction parameters were
kept constant to study the effect of substrate/metal ratio on MEA
conversion and product distribution. Fig. 15 shows that the conversion
of MEA and selectivities of the products were dependent on the catalyst
amount. The decrease of MEA/Au ratio (i.e., increase catalyst amount)
helped to improve MEA conversion from 65% to 100% when the ratio
varied from 1500 to 500 because of the increase of the concentration of
active sites. Glycine selectivity increase from 80% to 85% was ob-
served. Meanwhile, the selectivities of FA and ANHA decreased from
4.5% and 9% to 1.3% and 0.4%, respectively, when the MEA/Au ratio
decreased from 1500 to 300. Lower MEA/Au ratio meant increased
catalyst amount and number of active sites, which would lead to in-
creased initial reaction rate and MEA conversion in a short time before
removal of PVA (PVA could be removed from the gold surfaces by water
reflux [26] and this phenomenon had been confirmed by FTIR as shown
in Fig. S8, ESI). Therefore, the total selectivity of glycine increased and
the selectivity of formic acid was reduced at lower MEA/Au ratios.
Reaction pathway for MEA oxidation
Possible reaction pathways of MEA oxidation on Au catalysts were
proposed based on the detected intermediates and products, as well as
the widely accepted mechanisms of gold-catalyzed alcohol oxidation
and oxidative dehydrogenation of primary and secondary amines
[15,19,44,64–67]. As shown in Scheme 2, there exist at least two
pathways as both the hydroxyl and amino groups of the MEA molecule
are active on Au catalysts. In pathway A, the oxidation of the hydroxyl
group of MEA follows a similar process to the common alcohol oxida-
tion on Au to produce 2-aminoacetaldehyde, which could be subse-
quently oxidized to the target product glycine (pathway A1) or could
couple with the primary amine to produce amide or imine by-products
[15,19,44,45]. Amide (mainly ANHA, m/z 117.07, Fig. S2 and Table S3,
ESI) was the Au catalyzed oxidative coupling products of the aldehyde
intermediate
with
MEA.
Imine
(mainly
H2N−CH2-C =
N−CH2−CH2−OH, m/z 101.07, Fig. S2 and Table S3, ESI), was the
coupling products of the aldehyde intermediate with MEA [44,45]. In
pathway B, the MEA molecules proceeded through the oxidative de-
hydrogenation of the amine to the imine intermediate RCH = NH (m/z
58.03), as detected by MS (Fig. S2, ESI), which subsequently reacted
with water to form glycolaldehyde [64–68]. According to the literature,
glycolaldehyde could be subsequently oxidized into formic acid and
glycolic acid, which were detected by NMR and HRMS (Fig. S2 and S3,
ESI) [69,70].
Effect of the concentration of MEA
Various concentrations of MEA aqueous solutions show different
base strengths, which would influence the proceeding of the reaction.
The effects of initial MEA concentration on its selective oxidation were
illustrated in Fig. 16. Strong dependence of catalytic activity and pro-
duct selectivity on the substrate concentration was observed. At lower
substrate concentrations (e.g., 0.1-0.5 M), conversions higher than 96%
could be achieved. Further increasing the concentration led to de-
creased MEA conversion. For instance, when the MEA concentration
was 1.0 M, the conversion decreased to 79%, indicating that too high
MEA concentration accelerated deactivation of Au/ZrO2 catalyst. The
selectivity of glycine enhanced with the concentration of MEA until
0.5 M. At MEA concentrations higher than 0.5 M, apparent decrease in
selectivity of glycine and increase in formic acid selectivity were ob-
served, indicating that the side reactions were boosted under higher
MEA concentrations. The selectivity of ANHA decreased with the in-
creasing concentration of MEA due to hydrolysis of ANHA at higher
base concentration.
Conclusions
In this study, a series of Au catalysts supported on different metal
oxides were synthesized and characterized by TEM, XRD, ICP, XPS, and
TPD. These catalysts were used to catalyze MEA oxidation to produce
glycine. It was shown that the catalytic activity, product selectivity, and
stability of the catalysts were strongly dependent on the acid-base
properties of the supporting metal oxides. The catalytic activity was
positively correlated with the total basicity concentration of catalysts,
Recycle and reuse of Au/ZrO2 prepared by SI method was studied in
141

X. Meng, et al.
MolecularCatalysis469(2019)131–143
Scheme 2. Proposed reaction pathways for the gold-catalyzed aerobic oxidation of MEA.
and the selectivity of glycine and the stability were negatively corre-
lated with the acidity concentration. The highest catalytic performance
in the oxidation of MEA among the tested catalysts was observed on
Au/ZrO2.
[12] T. Aoki, K. Kawakami, K. Otsubo, Google Patents, (2005).
[13] W. Deng, Y. Wang, S. Zhang, K.M. Gupta, M.J. Hülsey, H. Asakura, L. Liu, Y. Han,
E.M. Karp, G.T. Beckham, Proc. Natl. Acad. Sci. 115 (2018) 5093–5098.
[14] G. Liang, A. Wang, L. Li, G. Xu, N. Yan, T. Zhang, Angew. Chem. Int. Ed. 129 (2017)
3096–3100.
[15] C. Della Pina, E. Falletta, L. Prati, M. Rossi, Chem. Soc. Rev. 37 (2008) 2077–2095.
[16] S. Biella, G. Castiglioni, C. Fumagalli, L. Prati, M. Rossi, Catal. Today 72 (2002)
43–49.
[17] J.H. Billman, E.E. Parker, J. Am. Chem. Soc. 65 (1943) 761–762.
[18] J.H. Billman, E.E. Parker, W.T. Smith, J. Biol. Chem. 180 (1949) 29–31.
[19] T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037–3058.
[20] A. Villa, N. Dimitratos, C.E. Chan-Thaw, C. Hammond, L. Prati, G.J. Hutchings, Acc.
Chem. Res. 48 (2015) 1403–1412.
[21] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.F. Carley, A.A. Herzing,
M. Watanabe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Science 311 (2006) 362–365.
[22] K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 41 (2002) 4538–4542.
[23] S.F.I. Thaddeus, W.L. Moench Jr, Google Patents, (2003).
[24] A. Gaiassi, L. Prati, Catal. Today 141 (2009) 378–384.
[25] A. Villa, S. Campisi, M. Schiavoni, L. Prati, Materials 6 (2013) 2777–2788.
[26] J.A. Lopez-Sanchez, N. Dimitratos, C. Hammond, G.L. Brett, L. Kesavan, S. White,
P. Miedziak, R. Tiruvalam, R.L. Jenkins, A.F. Carley, Nat. Chem. 3 (2011) 551.
[27] Y. Li, Y. Zheng, L. Wang, Z. Fu, ChemCatChem 9 (2017) 1960–1968.
[28] Y. Kotolevich, E. Kolobova, E. Khramov, M.H. Farías, Y. Zubavichus, H. Tiznado,
S. Martínez-González, V. Cortés Corberán, J.D. Mota-Morales, A. Pestryakov,
N. Bogdanchikova, Mol. Catal. 427 (2017) 1–10.
Preparation methods have a significant influence on the catalytic
activity of the Au nanoparticles. Au/ZrO2 catalyst prepared by SI
method exhibited good activity and glycine selectivity because of the
presence of PVA capping layer and the appropriate Au particle size. In-
situ FTIR and 1H NMR analysis demonstrated that the PVA capping
layer can enhance MEA adsorption on the catalysts and MEA molecules
were preferentially hydrogen bonded with PVA through interactions
between amino group of MEA and hydroxyl group of PVA, favoring the
reaction of the hydroxyl group of MEA on the active sites, which led to
doubled TOF, improved MEA conversion and selectivity of glycine
when the catalyst was capped with PVA layer. The reaction conditions
for the oxidation of MEA were also optimized and under these optimal
conditions, the MEA conversion and the selectivity of glycine could
reach 97% and 98%, respectively. This study provides a systematic
study on an alternative route for glycine synthesis and a deeper un-
derstanding of the role of the PVA layer, which is helpful for designing
more efficient catalysts to improve the catalytic performance.
[29] C.C. Torres, V.A. Jiménez, C.H. Campos, J.B. Alderete, R. Dinamarca,
T.M. Bustamente, B. Pawelec, Mol. Catal. 447 (2018) 21–27.
[30] C. Xu, Z. Wang, X. Huangfu, H. Wang, RSC Adv. 4 (2014) 27337–27345.
[31] S. Li, D. Gong, H. Tang, Z. Ma, Z.-T. Liu, Y. Liu, Chem. Eng. J. 334 (2018)
2167–2178.
[32] Y.-H. Ke, X.-X. Qin, C.-L. Liu, R.-Z. Yang, W.-S. Dong, Catal. Sci. Technol. 4 (2014)
3141–3150.
[33] T. Rundlöf, M. Mathiasson, S. Bekiroglu, B. Hakkarainen, T. Bowden, T. Arvidsson,
J. Pharm. Biomed. Anal. 52 (2010) 645–651.
Acknowledgments
This work was financially supported by General Program Youth of
National Natural Science Foundation of China (21603232), National
Natural Science Foundation of China (21676270, 21676277), Project of
CAS-Hengshui Science & Technology Transfer Center, CAS Pioneer
Hundred Talents Program, K.C.Wong Education Foundation (No. GJTD-
2018-04), and National Key R&D Program of China (No.
2017YFB0603301).
[34] Y. Yang, J. Li, J. Mu, H. Rong, L. Jiang, Nanotechnology 17 (2005) 461.
[35] E.D. Park, J.S. Lee, J. Catal. 186 (1999) 1–11.
[36] X. Zhang, H. Wang, B.-Q. Xu, J. Phys. Chem. B 109 (2005) 9678–9683.
[37] M. Comotti, C. Weidenthaler, W.-C. Li, F. Schüth, Top. Catal. 44 (2007) 275–284.
[38] K. Yang, J. Liu, R. Si, X. Chen, W. Dai, X. Fu, J. Catal. 317 (2014) 229–239.
[39] Y.-W. Chen, D.-S. Lee, H.-J. Chen, Int. J. Hydrogen Energy 37 (2012) 15140–15155.
[40] W. Wang, Y. Xie, S. Zhang, X. Liu, M. Haruta, J. Huang, Catalysts 8 (2018) 60.
[41] Z. Yuan, Z. Gao, B.-Q. Xu, Chin. J. Catal. 36 (2015) 1543–1551.
[42] A. Villa, S. Campisi, K.M. Mohammed, N. Dimitratos, F. Vindigni, M. Manzoli,
W. Jones, M. Bowker, G.J. Hutchings, L. Prati, Catal. Sci. Technol. 5 (2015)
1126–1132.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.mcat.2019.03.011.
[43] N. Weiher, A.M. Beesley, N. Tsapatsaris, L. Delannoy, C. Louis, J.A. van Bokhoven,
S.L. Schroeder, J. Am. Chem. Soc. 129 (2007) 2240–2241.
[44] J.-Fo. Soulé, H. Miyamura, S. Kobayashi, J. Am. Chem. Soc. 133 (2011)
18550–18553.
[45] L. Zhang, W. Wang, A. Wang, Y. Cui, X. Yang, Y. Huang, X. Liu, W. Liu, J.-Y. Son,
H. Oji, Green Chem. 15 (2013) 2680–2684.
References
[46] T. Ishida, R. Takamura, T. Takei, T. Akita, M. Haruta, Appl. Catal. A Gen. 413
(2012) 261–266.
[47] X. Huang, X. Wang, X. Wang, X. Wang, M. Tan, W. Ding, X. Lu, J. Catal. 301 (2013)
[1] R.J. Block, Chem. Rev. 38 (1946) 501–571.
[2] J.M. Orten, R.M. Hill, J. Am. Chem. Soc. 53 (1931) 2797–2799.
[3] W.C. Tobie, G.B. Ayres, J. Am. Chem. Soc. 64 (1942) 725.
[4] Y. Ogata, A. Kawasaki, T. Goto, J. Org. Chem. 33 (1968) 1107–1111.
[5] M. Gibson, R. Bradshaw, Angew. Chem. Int. Ed. 7 (1968) 919–930.
[6] E.J. Corey, D.E. Cane, J. Org. Chem. 35 (1970) 3405–3409.
[7] R. Williams, Oxford, (1989).
[8] A. Huges, Amino Acids, Peptides and proteins in organic chemistry, Origins and
Synthesis of Amino Acids Vol. 1 Wiley-VCH Verlag GmbH & Co., KGaA, Weinheim,
2009, p. 610.
217–226.
[48] M. Casaletto, A. Longo, A. Martorana, A. Prestianni, A. Venezia, Surf. Interface
Anal. 38 (2006) 215–218.
[49] A.M. Venezia, G. Pantaleo, A. Longo, G. Di Carlo, M.P. Casaletto, F.L. Liotta,
G. Deganello, J. Phys. Chem. B 109 (2005) 2821–2827.
[50] J. Chastain, R.C. King, J. Moulder, Handbook of X-ray Photoelectron Spectroscopy:
a Reference Book of Standard Spectra for Identification and Interpretation of XPS
Data, Physical Electronics Division, Perkin-elmer Corporation Eden Prairie,
Minnesota, (1992).
[9] Y. Zeng, Z. Li, G.P. Demopoulos, Ind. Eng. Chem. Res. 53 (2014) 16864–16872.
[10] Y. Zeng, Z. Li, G.P. Demopoulos, Ind. Eng. Chem. Res. 55 (2016) 2426–2437.
[11] H. Shimizu, C. Fujita, T. Endo, I. Watanabe, Google Patents, (1993).
[51] D.F. Evans, H. Wennerström, (1999).
142

X. Meng, et al.
MolecularCatalysis469(2019)131–143
[52] A. Villa, D. Wang, D. Su, G.M. Veith, L. Prati, Phys. Chem. Chem. Phys. 12 (2010)
2183–2189.
[53] A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 102 (2002) 3757–3778.
[54] Y. Cao, X. Liu, S. Iqbal, P.J. Miedziak, J.K. Edwards, R.D. Armstrong, D.J. Morgan,
J. Wang, G.J. Hutchings, Catal. Sci. Technol. 6 (2016) 107–117.
[55] J. Polte, CrystEngComm 17 (2015) 6809–6830.
[61] F. Porta, L. Prati, J. Catal. 224 (2004) 397–403.
[62] W.C. Ketchie, M. Murayama, R.J. Davis, Top. Catal. 44 (2007) 307–317.
[63] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, C.J. Kiely, G.A. Attard,
G.J. Hutchings, Top. Catal. 27 (2004) 131–136.
[64] S. Chi, G.T. Rochelle, Ind. Eng. Chem. Res. 41 (2002) 4178–4186.
[65] A.J. Sexton, G.T. Rochelle, Ind. Eng. Chem. Res. 50 (2010) 667–673.
[66] M. Largeron, Eur. J. Org. Chem. 2013 (2013) 5225–5235.
[67] B. Zhu, M. Lazar, B.G. Trewyn, R.J. Angelici, J. Catal. 260 (2008) 1–6.
[68] K. Matsuoka, Y. Iriyama, T. Abe, M. Matsuoka, Z. Ogumi, Electrochim. Acta 51
(2005) 1085–1090.
[56] A. Villa, D. Wang, G.M. Veith, F. Vindigni, L. Prati, Catal. Sci. Technol. 3 (2013)
3036–3041.
[57] D.V. Leff, L. Brandt, J.R. Heath, Langmuir 12 (1996) 4723–4730.
[58] A. Kumar, S. Mandal, P. Selvakannan, R. Pasricha, A. Mandale, M. Sastry, Langmuir
19 (2003) 6277–6282.
[59] L. Prati, P. Spontoni, A. Gaiassi, Top. Catal. 52 (2009) 288.
[60] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, G.J. Hutchings, Chem. Commun.
(Camb.) (2002) 696–697.
[69] Z. Wang, C. Xu, H. Wang, Catal. Lett. 144 (2014) 1919–1929.
[70] F. Wang, Y. Wang, F. Jin, G. Yao, Z. Huo, X. Zeng, Z. Jing, Ind. Eng. Chem. Res. 53
(2014) 7939–7946.
143