Sorbitol Hydrogenolysis over Hybrid Cu/CaO-Al2O3 Catalysts: Tunable Activity and Selectivity with Solid Base Incorporation

Article abstract of DOI:10.1021/acscatal.5b01324

We report for the first time the performance of hybridized Cu/CaO-Al₂O₃ catalysts for aqueous-phase hydrogenolysis of sorbitol to ethylene glycol (EG), 1,2-propanediol (1,2-PDO), and 1,2-butanediol (1,2-BDO) with linear alcohols as coproducts in a base-free liquid phase. These supported Cu catalysts with solid bases as promoters show significant activity for C-C cleavage and high selectivity (～84%) to glycols and linear alcohols. The effects of Cu loading, catalyst pretreatment conditions, H₂ pressure, and temperature on activity and selectivity of Cu/CaO-Al₂O₃ catalysts were investigated. The strong interaction between Cu and Ca²⁺ cations in the solid support is found to facilitate C-C and C-O cleavage of sorbitol, as evidenced from TEM, SEM, and TPR studies of the catalysts. Surface characterization and activity tests further suggest that Ca_xCu_yAl_zO_p (Phase I) promotes dehydrogenation and isomerization reactions, whereas spinal CuAl₂O₄ (Phase II) species facilitates hydrogenation reactions. In addition, the overall activity and selectivity of the Cu catalysts may be easily tuned by the Cu/Ca²⁺ molar ratio and catalyst preparation conditions. Cu/CaO-Al₂O₃ catalysts also give higher overall yields of value-added glycols (63-82%) for facile conversion of various other sugar polyols such as glycerol (C₃), erythritol (C₄), xylitol (C₅), and mannitol (C₆) under similar reaction conditions. A surface reaction mechanism involving the formation of β-ketoses on multifunctional Cu-Ca²⁺ sites is proposed.

Full text of DOI:10.1021/acscatal.5b01324

Subscriber access provided by UNIV OF ARIZONA
Article
Sorbitol Hydrogenolysis over Hybrid Cu/CaO-Al2O3 Catalysts:
Tunable Activity and Selectivity with Solid Base Incorporation
Xin Jin, Jian Shen, Wenjuan Yan, Meng Zhao, Prem S.
Thapa, Bala Subramaniam, and Raghunath V. Chaudhari
ACS Catal., Just Accepted Manuscript • Publication Date (Web): 24 Sep 2015
Downloaded from http://pubs.acs.org on September 28, 2015
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 free 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 accessible to all
readers and 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.
ACS Catalysis 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 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
Sorbitol Hydrogenolysis over Hybrid Cu/CaO-Al O Catalysts: Tunable
2
3
Activity and Selectivity with Solid Base Incorporation
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
1
3
1,2
1
4
1,2
Xin Jin, Jian Shen, Wenjuan Yan, Meng Zhao, Prem S. Thapa, Bala Subramaniam,
1
,2*
Raghunath V. Chaudhari
1
Center for Environmentally Beneficial Catalysis, University of Kansas, 1501 Wakarusa
Drive, Lawrence, Kansas 66047, USA
2
th
Department of Chemical and Petroleum Engineering, University of Kansas, 1530 W 15
St., Lawrence, Kansas 66045, USA
3
Department of Chemistry, University of Kansas, 1251 Wescoe Hall Drive, Lawrence,
Kansas 66045, USA
4
Microscopy and Analytical Imaging Laboratory, Haworth Hall, 1200 Sunnyside Ave,
University of Kansas, Lawrence, Kansas 66045, USA
*
To whom correspondence should be addressed. Eꢀmail: rvc1948@ku.edu
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
We report for the first time hybridized Cu/CaOꢀAl O catalysts for aqueous phase
2
3
hydrogenolysis of sorbitol to ethylene glycol (EG), 1,2ꢀpropanediol (1,2ꢀPDO) and 1,2ꢀ
butanediol (1,2ꢀBDO) with linear alcohols as coꢀproducts in a base free liquid phase.
These supported Cu catalysts with solid bases as promoters show significant activity for
CꢀC cleavage and high selectivity (~ 84%) to glycols and linear alcohols. The effects of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cu loading, catalyst pretreatment conditions, H pressure and temperature on activity and
2
selectivity of Cu/CaOꢀAl O catalysts were investigated. The strong interaction between
2
3
2
+
Cu and Ca cations in the solid support is found to facilitate CꢀC and CꢀO cleavage of
sorbitol as evidenced from TEM, SEM and TPR studies of the catalysts. Surface
characterization and activity tests further suggest that Ca Cu Al O (Phase I) promotes
x
y
z
p
dehydrogenation and isomerization reactions while spinal CuAl O (Phase II) species
2
4
facilitates hydrogenation reactions. In addition, the overall activity and selectivity of the
2
+
Cu catalysts may be easily tuned by the Cu/Ca molar ratio and catalyst preparation
conditions. Cu/CaOꢀAl O catalysts also give higher overall yields of valueꢀadded glycols
2
3
(63 ꢀ 82%) for facile conversion of various other sugar polyols such as glycerol (C3),
erythritol (C ), xylitol (C ) and mannitol (C ) under similar reaction conditions. A surface
4
5
6
2+
reaction mechanism involving the formation of βꢀketoses on multiꢀfunctional CuꢀCa
sites is proposed.
Key words: sorbitol, hydrogenolysis, glycols, copper catalyst, solid base promoters
2
ACS Paragon Plus Environment

Page 3 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
. Introduction
Hydrogenolysis of biomass derived feedstocks (e.g. sugars and polyols) has been
extensively studied in the past decade for developing sustainable technologies for fuels
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
1
, 2
and chemicals.
In this context, catalytic conversion of cellulosic and hemicellulosic
biomass to fuelꢀgrade compounds and valueꢀadded chemicals has received increasing
attention. In a typical process scheme, cellulose and hemicellulose can be converted to
3, 4
sugarꢀbased polyols via hydrolysisꢀhydrogenation route. The sugar polyols, including
erythritol (C ), xylitol (C ), sorbitol (C ) and mannitol (C ), can be further converted to
4
5
6
6
various useful products that are currently produced from fossil based raw materials. In
particular, sorbitol is commercially produced by catalytic hydrogenation of glucose and
hence is a promising platform chemical. Hydrogenolysis or hydrodeoxygenation (HDO)
of sorbitol can produce valuable chemicals such as ethylene glycol (EG), 1,2ꢀpropanediol
5ꢀ12
(
1,2ꢀPDO), 1,2ꢀbutanediol (1,2ꢀBDO), and lower aliphatic alcohols.
These products
are known commodities with several downstream applications in everyday products such
as renewable plastics, paints, antifreeze and pharmaceuticals.
7
, 13ꢀ16
HDO of sorbitol has been studied previously using supported metal catalysts.
However, one of the longstanding challenges is the development of cost effective
catalysts with high activity and selectivity. Sohounloue and coꢀworkers reported Ru/SiO2
and Raney Ni catalysts for HDO of sorbitol and proposed a retroꢀaldol mechanism (CꢀC
cleavage), based on observation of enhanced catalyst activity and the formation of 1,2ꢀ
1
3
17
PDO and EG in a strong alkaline medium (pH > 12). Montassier et al. proposed that
CꢀC bond cleavage by retroꢀMichael rather than retroꢀaldol mechanism was the main step
on Ru/C catalysts based on the fact that glycerol was the major product during sorbitol
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
18
conversion. But Wang et al. later observed that CꢀC cleavage follows retroꢀaldol
1
9
mechanism by using various diols as model compounds with NaOH as a base promoter.
Clark and coꢀworkers reported that HDO of sorbitol over kieselguhr supported Ni
catalysts with CaO as a base promoter gave EG, 1,2ꢀPDO and glycerol yields of 16%, 17%
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0
and 40%, respectively at 215 °C and 14 MPa hydrogen pressure (P ). Recently, Zhao
H2
and coꢀworkers investigated Ru/carbon nanofiber (CNF) and graphite felt composite
o
catalysts for HDO of sorbitol at 220 C and 7 MPa hydrogen, reporting 68% conversion
5
, 6
and overall glycols (EG and 1,2ꢀPDO) selectivity of approximately 53%. It was found
that the activity of Ru catalysts was enhanced when Ca(OH) , instead of NaOH, was used
2
as an alkali promoter. We recently reported that bimetallic RuRe/C showed improved
performances during sorbitol HDO over monometallic Ru/C catalysts when MgO was
15
added as a promoter, with selectivity to glycols in the range of 45ꢀ53% at 230 °C. In
7
another study, Banu et al. investigated sorbitol conversion over bimetallic NiPt/NaY
catalysts with CaO as a basic promoter at 220 °C and 6 MPa hydrogen, and reported a
11
combined selectivity to EG and 1,2ꢀPDO of 76%. Ye et al. reported that the activity of
Ni catalysts for HDO of sorbitol was enhanced by almost threefold at 230 °C and 7 MPa
4
+
H with trace Ce addition to the Ni/Al O +CaO system. However, side reactions
2
2
3
including methanation and water gas shift reaction (WGS) were also significant (methane
and CO selectivity = 10 ꢀ 22%).
2
It is clear from the previous work that directing product selectivity toward liquid
oxygenates is a major challenge during HDO of sorbitol over Pt, Ru and Niꢀbased
2
1, 22
catalysts.
Cu is an inexpensive candidate that has been investigated for HDO of
2
3
glycerol. However, studies reveal that Cu catalysts are better for CꢀO cleavage with
4
ACS Paragon Plus Environment

Page 5 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
24
poor CꢀC cleavage activity. In general, multiꢀfunctional catalysts that promote
dehydrogenation, isomerization, hydrogenation are needed to favor the formation of
14, 21
carbonyl intermediates and their isomers,
which may undergo CꢀC cleavage via retroꢀ
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
25
aldolization. Due to the monoꢀfunctional nature of conventional Cu catalysts, liquid
alkalis [e.g. aqueous solutions of NaOH, KOH, Ca(OH) ] are often needed to promote the
2
9
, 25ꢀ27
formation of C=O species.
This however results in the formation of excess amounts
of undesired carboxylic salts, which are difficult to further convert to glycols (via CꢀO
2
5, 27ꢀ29
cleavage).
The only previous work on Cu catalyzed HDO of sorbitol reported poor
24
activity and selectivity, even with liquid base promoters.
In this work, we report for the first time a hybrid Cu/CaOꢀAl O material for selective
2
3
HDO of sorbitol to glycols without using a liquid base. The proposed Cu catalysts
immobilized on Al O support containing either CaO, MgO or ZnO as solid base (CaOꢀ
2
3
Al O , MgOꢀAl O and ZnOꢀAl O ) were prepared via a concurrentꢀprecipitation method
2
3
2
3
2
3
3
0, 31
(CP)
and tested for HDO reactions. The Cu catalysts were characterized by N2ꢀ
adsorption, temperature programmed reduction (TPR), Xꢀray diffraction (XRD), UVꢀVis
spectra, scanning electron microscopy (SEM) and transmission electron microscopy
2
+
3+
(
TEM). The interactions of Cu atoms with adjacent Ca and Al ions in the solid
supports, the key to enhancing the activity of CꢀC and CꢀO cleavage, were investigated.
Concentrationꢀtime profiles are presented and discussed using sorbitol as the model
polyol. Based on the reactivity and product distribution observed on the hybrid Cu/CaOꢀ
2
+
Al O catalysts, a general reaction mechanism involved CuꢀCa interaction is proposed.
2
3
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
. Experimental Section
2.1 Catalyst preparation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2.1.1 Supported Cu on metal oxides
Cuꢀbased catalysts on various solid base supports were prepared using CP method
according to the following procedure (Figure S1). Required amounts of
Cu(NO ) .2.5H O (purum, ≥ 98%, Sigma), Ca(NO ) .4H O (≥ 99.0%, Sigma) and
3
2
2
3 2
2
Al(NO ) .9H O (purum, ≥ 98.0%, Sigma) were mixed with deionized water (DI water),
3
3
2
2
+
2+
3+
denoted as solution A. The total concentration of Cu , Ca and Al was about 0.2
3
kmol/m . Predetermined amounts of NaOH (reagent grade, ≥98% pellets, anhydrous,
Sigma) and Na CO .10H O (purum, ≥ 99.0%, Sigma) were then mixed with DI water
2
3
2
3
(solution B). The concentrations of NaOH and Na CO were 0.25 kmol/m and 0.8
2
3
3
kmol/m , respectively. In another 500 mL beaker, 50 mL of deionized water were
introduced under vigorous stirring (>800 rpm) at 60 °C (solution C). In the next step,
solutions A and B were added drop wise simultaneously (concurrently, as shown in
Figure S1) to beaker C (approximately 2 drops every 3 seconds). A blue slurry was then
formed. The pH of the solution in the beaker C was kept >10 throughout the preparation
process to ensure the precipitation of metal cations. The resulting slurry was stirred at
6
0 °C for 16 h. Then, the mixture was filtered and the solids were washed with 2000 mL
+
of DI water at 90 °C to remove surface Na ions. The solid cake obtained was then dried
overnight in a vacuum oven at 100 °C.
The dried solid catalyst sample was charged to a porcelain bowl, which was then
transferred to a calcination furnace (Barnstead/Thermolyne 48000) under air flow (flow
6
ACS Paragon Plus Environment

Page 7 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
3
rate: 5 cm /min/g catalyst). The furnace was then heated at a rate of 1 °C/min to 400 °C,
for 5 h of dwelling (heating) time. Then, the contents in the furnace were cooled naturally
by flowing air (Figure S2). Calcined Cu catalysts as prepared were denoted as CuO/CaOꢀ
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+
Al O ꢀ1 (Cu/Ca : 5.4 molar ratio), CuO/CaOꢀAl O ꢀ2 (Cu/Ca : 3.5), CuO/CaOꢀAl O ꢀ3
2
3
2
3
2
3
2
+
o
o
(Cu/Ca : 5.4, ramping rate used = 5 C/min, instead of 1 C/min). The catalysts were
then activated by a procedure shown in Figure S2. The reduced (i.e. activated) Cu
o
catalysts were denoted as Cu/CaOꢀAl O ꢀ1 (Cu loading = 43w%, 1 C/min), Cu/CaOꢀ
2
3
o
Al O ꢀ2 (Cu loading = 28w%, 1 C/min), Cu/CaOꢀAl O ꢀ3 (Cu loading = 43 w%,
2
3
2
3
ramping rate = 5 °C/min instead of 1 °C/min). The Cu content in the catalysts was
3
2
determined by inductively coupled plasma (ICPꢀAES) analysis.
The same procedure was followed for the preparation of Cu/ZnOꢀAl O , and Cu/MgOꢀ
2
3
Al O catalysts. Zn(NO ) .6H O (reagent grade, 98%, Sigma) and Mg(NO ) .6H O
2
3
3 2
2
3 2
2
(
purum, ≥ 99.0%, Fluka) were used for preparation of solution A to replace
Ca(NO ) .3H O. These catalyst samples were denoted as CuO/MgOꢀAl O , CuO/ZnOꢀ
3
2
2
2
3
Al O for calcined catalysts, while Cu/MgOꢀAl O , Cu/ZnOꢀAl O represent the
2
3
2
3
2
3
corresponding reduced Cu catalysts.
2.1.2 Cu catalyst on H-ZSM5
33
Cu/HꢀZSM5 samples were prepared via impregnation method. A typical synthetic
procedure followed in this approach is as follows: About 5 g of ZSM5 in ammonium
form (NH ꢀZSM5) were pretreated under calcination conditions under flow of air to
4
remove NH , the conditions of which were identical to those in Figure S2. The
3
ammonium ion was removed during calcination such that ZSM5 is now in the form of Hꢀ
ZSM5. Predetermined amounts of Cu(NO ) .2.5H O (purum, ≥ 98%, Sigma) were added
3
2
2
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to HꢀZSM5 in a 50 mL flask (Chemglass) followed by addition of a small amount (5 – 10
mL) of DI water. The resulting slurry was stirred at ~500 rpm for at least 4 h at room
temperature in order to obtain well dispersed solidꢀliquid suspension. The solvent (water)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
in the slurry was then removed using rotary evaporator at 55 C. Next, the sample was
dried at 120 °C in a vacuum oven overnight to further remove the remaining water.
Calcination (ramping rate = 1 °C/min) and activation processes were identical to the
procedures described in Figures S2 and S3. This sample was denoted as Cu/HꢀZSM5.
2.2 Catalyst activity tests
Performance evaluation of Cu catalysts for HDO of sorbitol, mannitol, xylitol, erythritol
and glycerol was carried out in a 300 mL stirred Parr reactor with H addition from
2
external sources to maintain a constant hydrogen partial pressure. Details of the
1
5
experimental setup and procedure are described in a previous publication.
2
.3 Analytical methods
Gas and liquid phase products were analyzed using GC (a RTꢀAPLOT column) and
+
HPLC (a Rezex ROAꢀOrganic Acid H column). An example of a typical HPLC
chromatogram of the product mixture is shown in Figure S4.
2
.4 Catalyst characterization
Brunauer–Emmett–Teller (BET) measurement: N adsorption studies were carried out
2
using a NOVA 2200e Instrument. The measurement procedure is similar to that described
32
previously.
8
ACS Paragon Plus Environment

Page 9 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
Temperature programmed reduction (TPR): TPR study was carried out using an
Autochem 2910 Instrument. Details are presented in supporting information.
UV-Vis spectroscopy: Surface absorbance under UVꢀVis spectroscopy was carried out
using Shimadzu UVꢀ3600 UVꢀVISꢀNIR Spectrophotometer. The samples were dispersed
in hexane solution and the solvent dried on a quartz plate before characterization.
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
Transmission electron microscopy (TEM): Sample preparation and detailed procedures
28
were similar to that described previously. Samples were prepared by suspending the
solid catalyst in ethanol and agitating in an ultrasonic bath. 10 ꢁL of catalyst sample were
placed onto a copper mesh grid with lacey carbon film (from Ted Pella Inc.). The wet
15
grid was allowed to airꢀdry for several minutes prior to being examined under TEM.
Scanning electron microscopy (SEM): A Versa 3D dual beam Scanning Electron
Microscope/ Focused Ion Beam (FEI, Hillsboro, OR, USA) with a silicon drift EDX
detector (Oxford Instruments, XꢀMax, UK) was used to measure the surface morphology,
elemental composition and distribution of metals. The SEM data were obtained at an
acceleration voltage of 15kV, spot size 3.0 and the images were collected with an ET
(Everhart Thornley) detector. The elemental mapping and energy spectra were acquired
with Aztec tools (Oxford Instruments, UK).
Element mapping was carried out via element dispersion xꢀray analysis during SEM
characterization. The xꢀray maps were converted into phase maps. The phase maps show
the constituent elements of the phase, and how the phases are distributed over the sample
(in different colors). The name of each phase is derived from its mapping of main
elements (e.g. CuCaAlO indicates Cu, Ca, Al and O are major elements in CuCaAlO
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
phase). The composition of each element in each phase can also be quantified. Each
phase detected by phase mapping was further confirmed by xꢀray diffraction analysis.
X-ray diffraction (XRD): This measurement was performed on a Bruker D8 powder
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
diffractometer with a copper target (CuK radiation) operating at 40 kV and a current of
α
4
0 mA to analyze the crystal structures of materials.
3. Results and Discussion
Initial experiments on HDO of sorbitol were carried out in a stirred slurry reactor to
compare the activity/selectivity profiles on different Cu catalysts. Selected Cu catalysts
were characterized using XRD, BET, TEM, SEM, TPR and UVꢀVis spectroscopy to gain
insights into structureꢀperformance relationship. Effect of reaction conditions and
concentrationꢀtime profiles were also studied on selected catalysts in a batch slurry
reactor. The results are discussed in terms of TOF, conversion and selectivity. With
selected catalysts, HDO of other substrates such as mannitol, xylitol, erythritol and
glycerol was also studied.
3
.1 Catalyst performance evaluation
The performances of Cu/CaOꢀAl O ꢀ1, Cu/MgOꢀAl O , Cu/ZnOꢀAl O and Cu/HꢀZSM5
2
3
2
3
2
3
catalysts for HDO of sorbitol are compared in Table 1. Cu/CaOꢀAl O ꢀ1 catalyst
2
3
outperforms the other three investigated catalysts in terms of conversion and product
selectivity. Notably, it is found that Cu/CaOꢀAl O ꢀ1 catalyst (Entries#1 and #2) shows
2
3
significantly higher conversion compared with Cu/MgOꢀAl O (Entries#3 and #4) and
2
3
Cu/ZnOꢀAl O (Entries#5 and #6) catalysts under similar reaction conditions. In
2
3
particular, as shown in Entry#1, sorbitol conversion reached 57.1% with Cu/CaOꢀAl O ꢀ
2
3
10
ACS Paragon Plus Environment

Page 11 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
within only 3 h. The combined selectivity to 1,2ꢀPDO, glycerol and EG was 68%.
Furthermore, as the conversion reached completion (Entry#2), the total selectivity
towards glycerol, glycols and linear alcohols was found to increase from 68% to 84% in
presence of Cu/CaOꢀAl O ꢀ1 catalyst. This is the first time that such high conversion and
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
3
excellent selectivity are reported for both CꢀC and CꢀO bond cleavage during HDO of
sorbitol. Moreover, the pH values of fresh substrate solution and final reaction products
were approximately 7, indicating that our hybridized Cu/CaOꢀAl O materials display
2
3
baseꢀcatalyzed retro aldolization activity even in an alkali free liquid reaction medium.
Compared with Cu/CaOꢀAl O ꢀ1 catalyst, only about 54% sorbitol conversion was
2
3
achieved on Cu/MgOꢀAl O catalyst after 6 h (Entry#3), with lower selectivity towards
2
3
glycols (1,2ꢀPDO and EG). With prolonged reaction time (12 h, Entry#4), combined
selectivity of C products (1,2ꢀPDO, LA and glycerol) was found to decrease, while that
3
of ethanol (EtOH) and methanol (MeOH) increased, suggesting that further conversion of
C products occurred possibly via CꢀC and CꢀO cleavage and HDO reaction to form C
3
1–2
products.
Compared to Cu/MgOꢀAl O catalyst, we observed that Cu/ZnOꢀAl O catalyst displayed
2
3
2
3
slightly higher conversion (56%) after 6 h (Entry#5) under identical reaction conditions.
It is worth noting that after 12 h (Entry#6), 95% conversion was achieved with more C4ꢀ6
polyols (including tetrols, triols, diols, about 29% in selectivity, identified by GCꢀMS) in
comparison with Cu/MgOꢀAl O catalyst (< 6% in selectivity), while selectivity to C
2
3
2ꢀ3
products was lower on Cu/ZnOꢀAl O catalyst. The difference in the product distribution
2
3
suggests that more CꢀO bond breakage of sorbitol molecules occurs over Cu/ZnOꢀAl O
2
3
compared to Cu/MgOꢀAl O catalyst. This product distribution observed with Cu/ZnOꢀ
2
3
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Al O catalyst is consistent with that reported by Blanc and coꢀworkers during HDO of
2
3
24
sorbitol, where CuO/ZnO catalysts were found to favor the formation of C tetrols,
4ꢀ6
triols and diols (about 56% selectivity) rather than CꢀC bond cleavage to smaller
molecules (due to CꢀC bond cleavage) observed when ZnO was used as a support and
promoter.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
+
2+
2+
Given that base promoters (Ca , Mg and Zn ) facilitate CꢀC and CꢀO cleavage, we
also evaluated the performance of acidic promoters during sorbitol conversion. Cu/Hꢀ
ZSM5 catalyst (Entry#7) showed completely different product distribution compared to
the other three catalysts. Under similar conditions, the major products observed were
anhydroglucitol and isosorbide (instead of glycols observed on the base promoted
catalysts) with approximately ~ 30% and 46% selectivity (both products confirmed by
GCꢀMS and HPLC), respectively, at a sorbitol conversion of 85%. These results suggest
that dehydration (DHD), rather than retroꢀaldolization and HDO, is favored on an acidic
support. The results over different Cu catalysts show that the acidity/basicity of solid
supports significantly affects the activity of Cu catalysts and reaction pathways during
34
HDO of sorbitol.
Given our interest in glycols and alcohols as products, we compared the catalytic
performances of the proposed Cu/CaOꢀAl O ꢀ1 catalyst with previously reported metal
2
3
catalysts (see Figure 1). Ru/C and bimetallic RuRe/C catalysts, which were previously
investigated by us and other research groups, show relatively high catalytic activity
compared to Ni and Cu based catalysts. However, the selectivity to glycols was only in
o
15
the range of 45 ꢀ 53% at 200 ꢀ 230 C. In comparison, NiPt/NaY catalyst displays
7, 35
improved selectivity to glycols (S ~ 82%).
NiCe/Al O catalysts exhibit relatively
2 3
12
ACS Paragon Plus Environment

Page 13 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
higher selectivity towards alcoholic products compared to Ru catalysts but tend to also
1
1
form gaseous products (methane and CO ). CuO/ZnO catalyst displays poor activity for
2
sorbitol conversion compared with Ru and Ni catalysts, with low selectivity to glycols
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
24
and other lower products. Our new hybrid catalyst, Cu/CaOꢀAl O ꢀ1, shows improved
2
3
activity compared to CuO/ZnO catalyst. More importantly, the overall yield of glycols
and alcohols is higher compared to other metal catalysts under similar reaction conditions
6
, 9, 27
(
T = 230 °C) although the activity is lower.
Cu/CaOꢀAl O ꢀ1 catalyst thus shows
2 3
good potential for HDO of sorbitol.
2
+
3
.2 Effect Cu/Ca ratio on activity and selectivity of Cu/CaO-Al O catalysts
2 3
2+
Cu/CaOꢀAl O catalysts with various Cu/Ca ratios were prepared and tested in sorbitol
2
3
conversion, the results of which are shown in Figure 2. Interestingly, we observed that
2
+
CaOꢀAl O support alone with no Cu (Cu/Ca = 0) showed very low conversion of
2
3
sorbitol but high selectivity for C products (with LA as the major product, S > 76%) in 6
3
2
+
ꢀ1
h. Catalyst with Cu/Ca ratio of 3.5, showed a TOF value of 0.24 h , and 81%
selectivity to C products (1,2ꢀPDO, glycerol, LA, PrOHs). In comparison, for the
3
2
+
ꢀ1
catalyst with a Cu/Ca ratio of 5.4 (TOF = 1.2 h ), the activity of Cu catalyst was
increased by almost four fold, although the selectivity of C products (1,2ꢀPDO, LA,
3
glycerol, PrOHs) decreased to 64%. Clearly, the Cu content in the support has a strong
influence on the activity and reaction pathways in sorbitol conversion. Therefore, textural
and structural characterization studies, employing BET, TPR, UVꢀVis, XRD, TEM and
SEM tools, were carried out to gain further insights into the surface properties of Cu
catalysts and understand the metalꢀsupport interaction.
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
3
.3 Catalyst characterization
.3.1 BET
BET adsorption results are presented in Figure 3 and Table 2 for Cu/CaOꢀAl O catalysts
2
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2+
with different Cu/Ca ratios. Specifically, adsorption/desorption isotherms, surface area,
2
+
pore size and pore volume of CaOꢀAl O supported Cu catalysts with Cu/Ca atomic
2
3
2
+
ratio of 3.5, 4.0, 5.4 and 6.0 were measured. It is found that as Cu/Ca atomic ratio
increases from 3.5 to 6.0, the total surface area of Cu catalysts is enhanced from 39.4 to
2
8
1.9 m /g. Nitrogen isotherms of these Cu catalysts confirm that these materials have
2
+
macroꢀpores and the mean pore size is also larger as Cu/Ca atomic ratios are higher. As
2+
seen from Table 2, the average pore size of the sample with Cu/Ca ratio of 3.5 is
2
+
approximately 7.69 nm while this value almost doubles for sample with a Cu/Ca value
of 6.0 (13.32 nm). The total pore volume follows the same trend for all four Cu catalyst
samples.
BET data of the four Cu catalysts clearly show that as more Cu was incorporated into the
CaOꢀAl O support, larger pores were formed during the precipitation process. This
2
3
observation implies that Cu species favor the formation of CaOꢀAl O framework in
2
3
aqueous solution and appear to tune pore growth leading to particular pore size and
volume. At this stage we do not have conclusive evidence to show that increasing pore
size and volume could enhance catalytic activity for sorbitol conversion. But later it is
2
+
2+
shown that changing Cu/Ca ratios will tune the interaction between Cu and Ca species
in the support, which might be the key for good HDO activity. This trend is different
from previous experimental findings wherein higher Cu content usually inhibited pore
14
ACS Paragon Plus Environment

Page 15 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
formation and reduced the total pore volume in samples prepared from MgOꢀAl O ,
2
3
3
0, 31, 33, 36
ZnOꢀAl O and ZnO.
2
3
3.3.2 TPR
The Cu/MgOꢀAl O , Cu/ZnOꢀAl O , Cu/CaOꢀAl O ꢀ1 (Cu/Ca = 5.4) and Cu/CaOꢀAlꢀ
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
3
2
3
2
3
2
+
O ꢀ2 (Cu/Ca = 3.5) catalysts were further characterized by TPR. The TPR temperature
2
3
ramping program was set at conditions similar to the reductive atmosphere employed
during catalyst activation in the tube furnace (Figure S3). Thus, the TPR profiles should
reflect the actual metalꢀsupport interaction during catalyst activation. As shown in Figure
4
, the hydrogen reduction peak of Cu/MgOꢀAl O catalyst occurs at 180 °C, which is the
2 3
II
0
37
normal reduction temperature for Cu to Cu species. Cu/ZnOꢀAl O sample exhibits
2
3
reduction behavior similar to Cu/MgOꢀAl O catalyst. TPR profile for Cu/CaOꢀAl O ꢀ1
2
3
2
3
II
sample shows a peak at a different temperature. The reduction peak of Cu species is
shifted from 180 °C to 225 ꢀ 230 °C. Further, the comparison between CuO/CaOꢀAl O ꢀ1
2
3
2
+
2+
(Cu/Ca = 5.4) and CuO/CaOꢀAl O ꢀ2 (Cu/Ca = 3.5) catalysts suggests that the
2
3
II
reduction peak of Cu species is progressively shifted to higher temperature (~ 250 °C)
2
+
2+
when Cu/Ca ratio is lower, implying that Cu and Ca interaction is stronger with an
increase in basic sites.
The shifts in reduction peak positions indicate that the interaction between Cu species
and CaOꢀAl O support is stronger compared with the other two metal oxides.
2
3
Considering the fact that Cu/CaOꢀAl O ꢀ1 and Cu/CaOꢀAl O ꢀ2 catalysts display higher
2
3
2
3
propensity for CꢀC cleavage than Cu/MgOꢀAl O , Cu/ZnOꢀAl O catalysts, metalꢀ
2
3
2
3
support interaction might be the key to activate metal sites during sorbitol conversion.
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
.3.3 UV-Vis Spectra
In order to further understand the metalꢀsupport interaction, three selected Cu/CaOꢀAlꢀ
O ꢀ1, Cu/CaOꢀAl O ꢀ2 and Cu/CaOꢀAl O ꢀ3 catalysts were also characterized using
2
3
2
3
2
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
UVꢀVis as shown in Figure 5. It is found that both Cu/CaOꢀAl O ꢀ1 (solid line) and
2
3
Cu/CaOꢀAl O ꢀ2 (dash line) catalysts exhibit strong characteristic adsorption at 290 nm,
2
3
although the absorbance magnitude is slightly different for the two catalysts. This peak is
3
8
slightly blue shifted compared to that obtained with 300 nm on CuO nanoparticles. In
contrast, the absorption associated with Cu/CaOꢀAl O ꢀ3 catalyst (dotꢀdash line) is
2
3
significantly milder. Besides, the characteristic absorption peak for Cu particles at 570
39
nm was not observed on the three samples, indicating negligible presence of Cu
particles.
3
.3.4 XRD
Xꢀray diffraction patterns for Cu/CaOꢀAl O ꢀ1, Cu/CaOꢀAl O ꢀ2 and Cu/CaOꢀAl O ꢀ3
2
3
2
3
2
3
catalysts are shown in Figure 6. All catalyst samples show characteristic peaks of
CuAl O , CaO and Al O species although their positions and intensities are different.
2
4
2
3
4
0ꢀ43
Specifically, Cu/CaOꢀAl O ꢀ1 displays strong peaks for CaO at 30°, 43°, 56° and 68°,
2
3
while both Cu/CaOꢀAl O ꢀ2 and Cu/CaOꢀAl O ꢀ3 exhibit lower intensity at these
2
3
2
3
positions. Interestingly, we also found that diffraction peaks of CuAl O species in
2
4
4
4ꢀ46
Cu/CaOꢀAl O ꢀ1 are shifted to lower diffraction angles, from 23° and 30°
to 20° and
2
3
2
7°, respectively. This indicates possible interaction of spinel CuAl O with other species
2 4
within the catalyst structure. For Al O diffraction, we found that the two characteristic
2
3
4
7ꢀ49
peaks are slightly moved to higher angles (from 36° and 39°
to 37° and 40°,
respectively) for Cu/CaOꢀAl O ꢀ1 catalyst compared with the other two samples.
2
3
16
ACS Paragon Plus Environment

Page 17 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
3
.3.5 TEM
The surface morphology of Cu/CaOꢀAl O ꢀ1, Cu/CaOꢀAl O ꢀ2 and Cu/CaOꢀAl O ꢀ3 was
2
3
2
3
2
3
investigated using TEM characterization, as shown in Figures 7, S5 and S6. Unlike the
conventional Cu/MgOꢀAl O and Cu/ZnOꢀAl O catalysts, which exhibit Cu particles of
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
3
2
3
3
0, 31, 36, 50
50ꢀ200 nm in size,
Cu/CaOꢀAl O ꢀ1 sample displays a highly crystallized
2 3
structure, as shown in Figures 7 (a) and (b). The Cu/CaOꢀAl O ꢀ1 sample does not show
2
3
any Cu particles with detectable sizes. In other words, Cu element is well dispersed and
mixed with CaOꢀAl O support. This observation is consistent with UVꢀVis spectra
2
3
shown in Figure 5.
EDX analysis of bulk composition of CuO/CaOꢀAl O ꢀ1 [Figure 7 (c)] also confirms that
2
3
Cu, Ca and Al elements are well distributed in the catalyst sample. Similarly, highly
crystalline structures were observed on Cu/CaOꢀAl O ꢀ2 (Figure S5) sample. Besides,
2
3
Cu/CaOꢀAl O ꢀ3 (Figure S6) exhibits ordered structured lattices, but several bulky
2
3
o
phases appear to exist, indicating that fast ramping rate (5 C/min) during synthesis may
result in a phase separation.
3
.3.6 SEM
SEM characterization of Cu/CaOꢀAl O ꢀ1, Cu/CaOꢀAl O ꢀ2 and Cu/CaOꢀAl O ꢀ3
2
3
2
3
2
3
catalysts confirms the phase distribution pattern observed earlier by XRD. In particular,
Figure S7 shows SEM images of the distribution of Cu, Ca, Al and O elements on
Cu/CaOꢀAl O ꢀ1 catalyst surface. Elemental analysis of the selected catalyst particle
2
3
shows that all four elements are well dispersed.
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
As already shown in TEM images (Figure 7), Cu, Ca, Al and O in Cu/CaOꢀAl O ꢀ1 form
2
3
hybrid structures and hence the size of Cu particles is not detectable. The hybrid
composition is however not known from the elemental analysis. Phase analysis was
therefore conducted in order to reveal the chemical composition of Cu catalysts. Figure 8
shows that CuAlO is the major phase, while Ca is concentrated as CaCuAlO being
dominant on the surface of Cu/CaOꢀAl O ꢀ1. The molar ratio of the two phases is about
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
1
8/1 as evaluated using a data processing software counted for 693 thousand points in the
region shown Figure S12. Therefore, it is clear that the highly crystalline structure mainly
represents CuAlO and CaCuAlO phases. The existence of CuAlO and CaCuAlO phases
also confirm a strong interaction between Cu and metal oxide supports, consistent with
TPR results and TEM images.
5
1ꢀ
It is known that CuAl O spinel structure is the major active phase for HDO reactions.
2
4
5
3
Based on the EDX analysis for CuAlO phase, the molar ratio of Cu, Al and O is
roughly 0.6/1.1/2.5 (or 1.8/3.3/7.5 for another spot on the same sample), which is
consistent with the composition of spinel CuAl O species or its dimer (dicopper
2
4
aluminum oxide, Cu Al O ) structure. The EDX analysis of another phase CaCuAlO
2
4
7
phase is consistent with a structure such as Ca Cu Al O .
5
3
6
17
SEM images of Cu/CaOꢀAl O ꢀ2 sample are shown in Figures S8 and S9. It is found that
2
3
Ca, Al and O are well distributed throughout the sample, while Cu species displays only a
slight phase separation. Compared with Cu/CaOꢀAl O ꢀ1 sample (Figures 8 and S7, Cu
2
3
loading = 43 w%), Cu content in this sample is much lower (Cu loading = 28 w%). The
difference in SEM images clearly indicates that the active phases in the two samples are
distinct from each other. Phase diagram of Cu/CaOꢀAl O ꢀ2 sample was thus carried out.
2
3
18
ACS Paragon Plus Environment

Page 19 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
As shown in Figure S9, we counted for more than 708 thousand points for the region
shown and found that major phases for Cu/CaOꢀAl O ꢀ2 sample are CuO/Cu and
2
3
CaCuAlO. In sharp contrast, CaCuAlO is the dominant phase (98.1%) in this sample
compared with Cu/CaOꢀAl O ꢀ1 (CaCuAlO ≈ 6%). CuO/Cu is < 2% in this sample.
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
3
Figures S10 and S11 presents the distribution of Cu, Ca, Al and O elements on Cu/CaOꢀ
Al O ꢀ3 catalyst surface. Elemental analysis on a selected catalyst particle shows that all
2
3
four elements are not well dispersed. It shows a trend of elemental segregation on the
catalyst surface. It is important to point out that the bulk composition of Cu/CaOꢀAl O ꢀ3
2
3
catalyst is almost identical with Cu/CaOꢀAl O ꢀ1 catalyst. The pretreatment conditions
2
3
are however different from each other. As mentioned in the experimental section,
o
Cu/CaOꢀAl O ꢀ1 sample underwent a 1 C/min ramping rate during calcination (under
2
3
o
flowing air), while Cu/CaOꢀAl O ꢀ3 experienced a 5 C/min ramping rate. The
2
3
uniformity and crystallinity of Cu/CaOꢀAl O ꢀ3 are not as good as the Cu/CaOꢀAl O ꢀ1
2
3
2
3
sample. Figure S11 shows that Al O is a dominant phase, which means that phase
2
3
segregation led to the formation of mixed metal oxides rather than hybridized ones.
Therefore, it is highly possible that fast heating rate results in phase separation.
SEM images of used Cu/CaOꢀAl O ꢀ1 catalysts after three recycles are shown in Figure 9.
2
3
The surface morphologies have changed from layered structures (Figure S7) to layered
and spherical mixtures (Figure 9). We find that both Ca and Al are still well distributed
after reactions, while Cu species tend to agglomerated during recycles. Particularly, EDX
mapping in Figure 9 shows that Cu tend to phase out from Cu/CaOꢀAl O ꢀ1 materials and
2
3
form spherical particles. This observation is also supported by STEM images shown in
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 10. Clearly, Cu species tend to form bulky phases after reactions [see overlayered
element mapping in Figure 10 (b)].
3.4 Structure-activity correlation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The conversion of sorbitol on the three Cu catalysts, Cu/CaOꢀAl O ꢀ1, Cu/CaOꢀAl O ꢀ2
2
3
2
3
and Cu/CaOꢀAl O ꢀ3 was investigated in a batch reactor. As shown in Table 3, Cu/CaOꢀ
2
3
Al O ꢀ1 catalyst displayed almost 100% conversion of sorbitol in 6 h, while Cu/CaOꢀAlꢀ
2
3
O ꢀ2 and Cu/CaOꢀAl O ꢀ3 only showed < 20% conversion under the same reaction
2
3
2
3
conditions. The selectivities towards 1,2ꢀPDO, LA+PAD and EG are significantly
different on the three catalysts. It is found that the combined selectivity to 1,2ꢀPDO,
glycerol and EG was about 73% on Cu/CaOꢀAl O ꢀ1 catalyst, but this value was only 20
2
3
ꢀ
32% on other two catalysts. Further, whereas the selectivity of LA+PAD was < 2% on
the active Cu/CaOꢀAl O ꢀ1 catalyst, the selectivity values on the other two catalysts were
2
3
as high as 64 ꢀ 75%. It is believed that C ꢀC cleavage occurs after dehydrogenation (DH)
3
3
of sorbitol leading to the formation glyceraldehyde (GLA) and dihydroxyacetone (DHA)
as C intermediates, which undergo rearrangement to form LA (in alkali medium or
3
5
4
35
supercritical water ) or HDO reactions to generate 1,2ꢀPDO, as shown in Scheme 1. It
can be inferred that whereas Cu/CaOꢀAl O ꢀ1 is active for both DH and HDO reactions,
2
3
only DH occurs on Cu/CaOꢀAl O ꢀ2 and Cu/CaOꢀAl O ꢀ3 catalysts.
2
3
2
3
Furthermore, the selectivity towards C products (1,2ꢀPDO, LA and glycerol) is 61 ꢀ 98%,
3
on all the three catalysts implying that C ꢀC cleavage is dominant. Further, as shown
3
3
earlier, XRD and SEM characterizations reveal that Cu/CaOꢀAl O ꢀ2 sample has only
2
3
one phase (CaCuAlO), while Cu/CaOꢀAl O ꢀ1 displays both CaCuAlO and CuAlO
2
3
20
ACS Paragon Plus Environment

Page 21 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
phases. It is therefore hypothesized that CaCuAlO is the active site for the DH and C ꢀC
3
3
cleavage reactions of sorbitol and CuAlO (spinel CuAl O ) species are active for
2
4
hydrogenation reactions.
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
Hydrogenation reactions consume C intermediates species formed from DH and retroꢀ
3
aldolization of sorbitol on Cu catalysts, which in turn helps overcome the equilibrium
limitation associated with DH reactions due to product accumulation. Therefore, the
Cu/CaOꢀAl O ꢀ1 catalyst shows much higher sorbitol conversion. In contrast, Cu/CaOꢀ
2
3
Al O ꢀ2 and Cu/CaOꢀAl O ꢀ3 catalysts have almost no active sites for hydrogenation
2
3
2
3
reactions, causing the sorbitol conversion on these catalysts to be hindered by equilibrium
limitations.
The higher activities for CꢀC and CꢀO cleavage reactions observed with Cu/CaOꢀAl O ꢀ1
2
3
compared to other catalysts suggests that the activity and selectivity of Cu/CaOꢀAl O
2
3
catalyst in general are dictated by the relative distributions of DH (CaCuAlO phase) and
hydrogenation (spinel CuAl O phase) sites.
2
4
3
.5 Reaction profiles
Since Cu/CaOꢀAl O ꢀ1 catalyst shows superior performances compared to other catalysts,
2
3
detailed concentrationꢀtime profiles were obtained to better discern possible reaction
pathways. Figures 11 (a and b) show typical temporal product distribution [including
1
,2ꢀPDO, EG, glycerol, butylene glycol (1,2ꢀbutanediol, 1,2ꢀBDO)] profiles during
sorbitol conversion. The concentrations of 1,2ꢀPDO, EG and glycerol increase almost
linearly with reaction time, while that of LA+PAD (pyruvaldehyde) first increase and
then decrease at longer reaction times, implying that LA and PAD under further
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
hydrogenated to linear alcohols and glycols as products. The concentration of linear
alcohols (Figure 11) increased sharply with reaction time, as a result of consecutive HDO
of 1,2ꢀPDO and EG to propanols (PrOH), EtOH and MeOH. A linear dependence of
sorbitol conversion with time is observed until almost 100% conversion. As inferred from
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figures 11 (a) and (b), the HDO rates show a weak dependence on sorbitol concentration
o
(
230 C). While lower sorbitol conversion and product formation rates are observed at
o
lower reaction temperature (210 C), the product distribution is similar at the two
temperatures.
Pressure effect on the product distribution during sorbitol conversion is unclear from
13, 35
previous literature reports.
As shown in Figure 12, more LA is formed, with
relatively low selectivity compared to 1,2ꢀPDO at low H pressure (2.8 MPa). With
2
increasing hydrogen pressures, the combined selectivity of glycerol and LA decreased by
about 22% while the initial selectivity of 1,2ꢀPDO as well as other alcohols increased by
20%. The difference in product selectivity further confirms the possible reaction network
shown in Scheme 1.
3
.6 Plausible reaction pathways and mechanism
The product distribution pattern observed in sorbitol conversion with Cu/CaOꢀAl O
2
3
catalysts shows that a series of consecutive and parallel reactions occur. In particular, Cꢀ
C and CꢀO cleavage, HDO, DH and DHD reactions are expected to occur on the catalyst
surface and a possible reaction pathway is proposed in Scheme 2. Generally, sorbitol is
believed to undergo DH to form different unsaturated aldoses and ketoses, such as
1
6
aldohexose, βꢀketohexose and γꢀketohexose. The formation of aldohexose (primary C
6
22
ACS Paragon Plus Environment

Page 23 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
aldehyde) initially will generate C and C (Route I) products, while βꢀketohexose will
2
4
lead to C ꢀC cleavage (Route III). γꢀKetohexose formed initially results in C ꢀC
5
3
3
1
breakage (Route II) and C may further convert to βꢀketopentose (which can be
5
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
isomerized to aldopentose) to generate C and C products (Route IV). The likelihood of
2
3
the formation of βꢀketohexose is high under low hydrogen pressure on Cu catalysts,
because combined selectivity of C products is much higher than other products, as
3
shown in Figure 12. However, it should also be noted that selectivity of C products
2
(mainly EG) is higher than C (mainly 1,2ꢀBDO). This finding suggests that C ꢀC
5
4
1
cleavage also occurs over Cu catalysts, although C is formed in minor quantities
1
compared to C ꢀC cleavage products. This also implies that DH over Cu catalysts tends
3
3
to generate βꢀketohexose rather than γꢀketohexose and aldohexose. In other words, at
higher H pressures, aldehydeꢀlike intermediate compounds are likely to be formed (still
2
minor compared with βꢀketohexose) while at lower hydrogen pressures βꢀketone like
intermediate is favored which leads to 1,2ꢀPDO and EG formation.
Sun & Liu’s work implies that in xylitol conversion, the formation of aldopentose
(
primary C aldehyde) was favored rather than ketopentose because C and C products
5
1
4
9
, 27
were found to be in trace amounts.
In fact, only γꢀketopentose leads to the formation
of C and C products. From the mechanism shown (Scheme 2), it is obvious that βꢀ
1
4
ketopentose will also facilitate C ꢀC cleavage. In order to further understand the possible
2
3
position where the C=O bond is formed after DH reactions, xylitol (C polyol), erythritol
5
(
C polyol) and glycerol (C polyol) were also studied as substrates.
4 3
Results on catalytic HDO of mannitol, xylitol, erythritol and glycerol on Cu/CaOꢀAl O ꢀ
2
3
1
catalyst are summarized in Table 4. We observed that mannitol conversion led to > 65%
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
C selectivity (Entry#1), while the combined selectivity to EG and EtOH (C products)
3
2
was only 15%. The conversion of xylitol showed mainly C and C products (Entry#2),
2
3
1
8
possibly following retroꢀaldolization mechanism. It is obvious that the theoretical
selectivities of C and C should be 60% and 40%, respectively. Here, it is found that the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
2
total selectivity of 1,2ꢀPDO, LA, PAD and glycerol is approximately 56.7% over
Cu/CaOꢀAl O ꢀ1 catalyst and combined selectivity toward EG and EtOH is about 31%.
2
3
The ratio of C /C is thus almost equal to the theoretical value (60/40). The limited
3
2
selectivity of MeOH and 1,2ꢀBDO implies that xylitol conversion follows primarily C2ꢀ
9
C cleavage. These results agree well with Sun & Liu’s observations on Ru/C catalyst.
3
HDO of mesoꢀerythritol showed that the dominant product is 1,2ꢀBDO (S ≈ 39%). This
suggests that, as the carbon number of the polyol decreases from 6 (sorbitol, mannitol) to
5
(xylitol) to 4 (erythritol), the tendency of CꢀC cleavage on the Cu catalysts becomes
less significant, while the propensity toward CꢀO cleavage is enhanced. Therefore, it is
proposed that erythritol follows sequential dehydration (DHD) and HDO reactions to
form 1,2ꢀBDO (Route V in Scheme 3). C1ꢀ3 products were also detected in HPLC,
including glycerol, LA, 1,2ꢀPDO, EG and MeOH. The ratio of C /C was found to be
3
1
almost 3/1. Besides, as shown in Table 4, the combined selectivity of C +C products is
1
3
much higher than C . This observation clearly indicates that C ꢀC bond breakage (Route
2
1
3
VI in Scheme 2) rather than C ꢀC cleavage (Route VII) prevails over Cu/CaOꢀAl O
3
2
2
2
catalyst. The observed facts support our hypothesis that the formation of βꢀketose (e.g. βꢀ
ketohexose, ketopentose and ketoerythrose) is favored on Cu/CaOꢀAl O catalysts.
2
3
The argument on whether 1,2,3ꢀbutanetriol (1,2,3ꢀBTO, Route V) or 1,2,4ꢀbutanetriol
1,2,4ꢀBTO, Route VIII) is the precursor for 1,2ꢀBDO can be answered by the detected
(
24
ACS Paragon Plus Environment

Page 25 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
species in GCꢀMS. Trace 1,2,4ꢀBTO (S < 1.8%) was found at small sorbitol conversion
values (X < 10%), which indicates that the generation of 1,2ꢀBDO is instantaneous on
Cu/CaOꢀAl O ꢀ1 catalyst. Moreover, this observation confirms that active basic species
2
3
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
tend to attack terminal carbon and promote DHD at primary and secondary carbons.
However, the reason that sequential DHD and hydrogenation occur to form 1,2ꢀBDO is
still not clear. Here a representative mechanism is proposed for the formation of 1,2ꢀBDO.
Unlike sorbitol and xylitol, the βꢀketose intermediate from erythritol might undergo either
retroꢀaldolization (Route VI) or secondary DHD (Route V in Scheme 3) because the
stable conjugated structure might also be favored. A detailed description of the reaction
mechanism over Cu/CaOꢀAl O ꢀ1 catalyst is presented in the next section.
2
3
9
, 15
In sharp contrast to the results reported on Ru catalysts,
the observed product
distribution during conversions of sorbitol, mannitol, xylitol and erythritol over Cu/CaOꢀ
Al O ꢀ1 catalyst implies that the potential of CꢀC breakage of C molecules is restrained
2
3
3
on this catalyst. These observations are further confirmed by our study of glycerol HDO
over the Cu/CaOꢀAl O ꢀ1 catalyst that shows a conversion of 67.5% with 1,2ꢀPDO (S =
2
3
8
1%) as a major product in 10 h under 2.8 MPa H pressure. Interestingly, compared with
2
the HDO of sorbitol (C ), xylitol (C ) and erythritol (C ), the CꢀC cleavage of glycerol
6
5
4
(
C ) molecules is significantly restrained as inferred from the low yields of C products.
3
1
ꢀ
2
Previous studies found that the addition of liquid base (NaOH and CaO) to the reaction
mixture increased glycerol conversion to both EG (S ≈ 20%) and 1,2ꢀPDO (S ≈ 27%)
because the retro aldolization and HDO reactions are enhanced by adsorbed hydroxyls
5
even under much milder conditions. Again, such a difference can be explained by
alternate representative pathways such as the formation of DHA, which is a type of βꢀ
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ketose in Scheme 4 (Route IX) over Cu/CaOꢀAl O catalyst rather than GLA, which can
2
3
lead to CꢀC cleavage in glycerol conversion. Once DHA is formed, it undergoes further
DHD to generate a conjugated unsaturated alcohol, which hydrogenates to form 1,2ꢀPDO.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
As indicated above, product distribution from HDO of sorbitol, xylitol and erythritol
suggests that CꢀC cleavage is likely to be initiated by βꢀketone and thus displays a C ꢀC
x
3
(x = 1, 2, 3 for erythritol, xylitol and sorbitol, respectively) like cleavage in the presence
2+
of divalent Ca species. Therefore, a general CꢀC cleavage mechanism over the Cu/CaOꢀ
Al O catalysts is proposed here for the first time. Because DH of αꢀalcohols occurs
2
3
easily on Cu catalysts whereas DH of secondary and tertiary alcohols demands extremely
55
harsh conditions, sorbitol most likely undergoes DH reaction to form aldehexose
catalyzed by Cu sites, as shown in Scheme 5. As supported by TPR (reduction peak
2
+
position) profiles and SEM (phase diagram) characterization, Ca cations neighboring
Cu species have a strong influence on Cu sites and tend to isomerize the aldohexose
intermediates forming βꢀketohexose (or γꢀketohexose). This is on the basis of retroꢀaldol
mechanism by which the formation of βꢀketohexose results in C ꢀC cleavage. In the case
3
3
of xylitol, this mechanism also applies and similar equilibrium structures (aldopentose, βꢀ
ketopentose and γꢀketopentose) are formed. It is also clear that aldopentose and βꢀ
ketopentose both result in C ꢀC scission, although in different ways.
2
3
There have been several debates on the role of cations during CꢀC and CꢀO cleavage. It
3
+
was found that Cr could isomerize aldohexose (glucose) molecules via bonding with α
and β carbon chains to form glucoseꢀfructose equilibrium structures, which easily
undergo DHD in the presence of an acidic promoter (pH < 5 in the solution containing
3
+ 56
Cr ). Later reports also found similar effects over Lewis acid, Snꢀcontaining zeolites
26
ACS Paragon Plus Environment

Page 27 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
and depicted the same mechanism for the formation of hydroxymethylfurfural during
5
7, 58
glucose and fructose conversion.
In contrast, Yan and coworkers found that the LA
(
(
C ) yield from glucose (C ) was facilitated by Ca(OH) to a greater extent than NaOH
3
6
2
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
54
at the same initial OH concentration) during hydrothermal conversion (270ꢀ400 C).
2
+
These observations confirm that Ca is the key to C ꢀC scission over Cu catalysts
3
3
during HDO of sorbitol and xylitol (Scheme 5).
For xylitol, dominant C ꢀC cleavage occurs because of the formation of the βꢀketose.
2
3
Analogously, formation of C and C via asymmetrical C ꢀC cleavage can also occur via
1
3
1
3
forming a βꢀketose from a C polyol and is confirmed by our experimental results from
4
the conversion of erythritol, although the potential of CꢀC scission is much lower than
sorbitol and xylitol. The asymmetrical cleavage seems to be not favored for sorbitol (C1ꢀ
C ) and xylitol (C ꢀC ) conversions on Cu/CaOꢀAl O catalyst. These unsaturated
5
1
4
2
3
intermediates generated from DH and CꢀC cleavage are then hydrogenated to alcoholic
products.
3
.7 Recycle studies
Recycle studies on sorbitol conversion using Cu/CaOꢀAl O ꢀ1 catalyst (refer Table 1 for
2
3
reaction conditions) were further carried out. As shown in Table 5, an observable
deactivation of Cu/CaOꢀAl O ꢀ1 catalyst was found. In particular, conversion of sorbitol
2
3
o
st
at 230 C decreases from 98% (Table 1) to 88% after 1 test, and then further to < 70%
after three recycles (see Table 5). The selectivities towards 1,2ꢀPDO, EtOH and MeOH
remain almost unchanged during recycles. But selectivity to glycerol and EG decreases
with increasing LA+PAD selectivity after three recycle tests. It is highly plausible that
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the decreased activity of Cu/CaOꢀAl O ꢀ1 catalyst is due to the phase separation of Cu
2
3
species from Cu/CaOꢀAl O ꢀ1 materials (Figures 9 and 10), which results in relatively
2
3
2
+
weaker interaction between Cu and Ca cations in the support.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
. Conclusion
Hydrogenolysis of sugar polyols such as sorbitol, mannitol, xylitol, erythritol and
glycerol using the novel Cuꢀbased triꢀfunctional catalysts has been investigated. It is
2+
2+
2+
found that, in comparison with Mg and Zn , Ca displays significant promotional
effects on the activity as well as the selectivity of Cu catalysts. Detailed surface
characterization confirms the existence of binary phases, CaCuAlO and CuAlO in Cu
catalysts, which are responsible for DH and HDO, respectively, during the conversion of
2
+
polyols. Results with different polyols further show that Ca also promotes the
isomerization of aldoses, and that C ꢀC (x = 1, 2, 3) cleavage is significant. A possible
3
x
reaction mechanism involving DH and isomerization is proposed for the first time for Cuꢀ
based catalysts. The proposed Cuꢀbased catalysts display good performances for catalytic
upgrading of sugar polyols to valueꢀadded glycols and alcohols.
Associated content
Supporting Information. This material is available free of charge on the ACS Publications
website at http://pubs.acs.org. Additional information on catalyst preparation, analytical
conditions, TEM and SEM characterization.
28
ACS Paragon Plus Environment

Page 29 of 55
ACS Catalysis
1
2
3
4
5
6
7
8
9
Acknowledgement
Partial support from United States Department of Agriculture (USDA/NIFA Award
011ꢀ10006ꢀ30362) and National Science Foundation (CHEꢀ1543673) is gratefully
2
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
acknowledged. We want to thank Dr. Victor Day for XRD analysis.
References
1
2
3
4
. Huber, G. W.; Cortright, R. D.; Dumesic, J. A. Angew. Chem., Int. Ed. 2004, 43, 1549−1551.
. Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044−4098.
. Ohara, M.; Takagaki, A.; Nishimura, S.; Ebitani, K. Appl. Catal., A 2010, 383, 149−155.
. Tong, X. L.; Li, M. R.; Yan, N.; Ma, Y.; Dyson, P. J.; Li, Y. D. Catal. Today 2011, 175,
5
24−527.
5
. Zhao, L.; Zhou, J. H.; Chen, H.; Zhang, M. G.; Sui, Z. J.; Zhou, X. G. Korean J. Chem. Eng.
2010, 27, 1412−1418.
6
. Zhao, L.; Zhou, J. H.; Sui, Z. J.; Zhou, X. G. Chem. Eng. Sci. 2010, 65, 30−35.
7
. Banu, M.; Sivasanker, S.; Sankaranarayanan, T. M.; Venuvanalingam, P. Catal. Commun.
2011, 12, 673−677.
8. Jin, X.; Subramaniam, B.; Chaudhari, R. V. Novel Materials for Catalysis and Fuels
Processing; BravoꢀSuarez, J. J., Kidder, K. K., Schwartz, S., Eds.; ACS Symposium Series
1
132; American Chemical Society: Washington, DC, 2013; pp 273−285.
9
. Sun, J. Y.; Liu, H. C. Green Chem. 2011, 13, 135−142.
1
0. Jin, X.; Subramaniam, B.; Chaudhari, R. V. Proceedings of the International Congress of
Catalysis Meeting, Munich, Germany, July 1−6, 2012.
1
1. Ye, L. M.; Duan, X. P.; Lin, H. Q.; Yuan, Y. Z. Catal. Today 2012, 183, 65−71.
1
2. Jin, X.; Subramaniam, B.; Chaudhari, R. V. Presented at the Center for Environmentally
Beneficial Catalysis Industry Advisory Board Meeting, Lawrence, KS, Oct. 3−4, 2011.
1
3. Sohounloue, D. K.; Montassier, C.; Barbier, J. React. Kinet. Catal. Lett. 1983, 22, 391−397.
29
ACS Paragon Plus Environment