Interface synergy between IrOx and H-ZSM-5 in selective C–O hydrogenolysis of glycerol toward 1,3-propanediol

Article abstract of DOI:10.1016/j.jcat.2019.06.025

Site-selective deoxygenation of hydroxyl groups represents essential processes to access valuable functionalized bio-based compounds with industrial potential. One of the challenging tasks in this context is to convert biodiesel-derived glycerol in the presence of abundant water directly to 1,3-propanediol (1,3-PDO), a key component of the emerging polymer industry. Herein, a monometallic iridium supported on H-ZSM-5 in the absence of Re oxophilic metal oxides was prepared via grinding-assisted impregnation procedures and attempted as an effective and recyclable catalyst for the aqueous-phase selective hydrogenolysis of glycerol toward 1,3-PDO in the absence of acid additives. The results revealed the necessity to control the Ir domain dispersions, Ir⁰/Ir³⁺ ratio and the amounts of overall acid/Br?nsted acid sites. Activity depended linearly on the amount of overall and Br?nsted acid sites, and 1,3-PDO selectivity increased in the presence of Ir-induced Br?nsted acid sites, denoted as Ir-O(H)-H-ZSM-5. We speculate that Ir-O(H)-H-ZSM-5 are generated by the interfacial synergistic interaction between IrO_x and H-ZSM-5 through hydrogen spillover and reverse hydrogen spillover according to the reported literatures. The reaction mechanism to elucidate the role of Ir-O(H)-H-ZSM-5 sites in glycerol hydrogenolysis was also postulated based on extensive characterization and catalytic reaction results.

Full text of DOI:10.1016/j.jcat.2019.06.025

Journal of Catalysis 375 (2019) 339–350
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Interface synergy between IrOx and H-ZSM-5 in selective C–O
hydrogenolysis of glycerol toward 1,3-propanediol
Xiaoyue Wan a,b, Qi Zhang b, Mingming Zhu b, Yi Zhao b, Yongmei Liu b, Chunmei Zhou a, Yanhui Yang a,
,
⇑
Yong Cao b,
⇑
a Institute of Advanced Synthesis, School of Chemistry and Molecular Engineering, Jiangsu National Synergetic Innovation Center for Advanced Materials, Nanjing Tech
University, Nanjing 211816, China
b Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Department of Chemistry, Fudan University, Shanghai 200433, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Site-selective deoxygenation of hydroxyl groups represents essential processes to access valuable func-
tionalized bio-based compounds with industrial potential. One of the challenging tasks in this context
is to convert biodiesel-derived glycerol in the presence of abundant water directly to 1,3-propanediol
(1,3-PDO), a key component of the emerging polymer industry. Herein, a monometallic iridium supported
on H-ZSM-5 in the absence of Re oxophilic metal oxides was prepared via grinding-assisted impregnation
procedures and attempted as an effective and recyclable catalyst for the aqueous-phase selective
hydrogenolysis of glycerol toward 1,3-PDO in the absence of acid additives. The results revealed the
necessity to control the Ir domain dispersions, Ir0/Ir3+ ratio and the amounts of overall acid/Brönsted acid
sites. Activity depended linearly on the amount of overall and Brönsted acid sites, and 1,3-PDO selectivity
increased in the presence of Ir-induced Brönsted acid sites, denoted as Ir-O(H)-H-ZSM-5. We speculate
that Ir-O(H)-H-ZSM-5 are generated by the interfacial synergistic interaction between IrOx and H-ZSM-
5 through hydrogen spillover and reverse hydrogen spillover according to the reported literatures. The
reaction mechanism to elucidate the role of Ir-O(H)-H-ZSM-5 sites in glycerol hydrogenolysis was also
postulated based on extensive characterization and catalytic reaction results.
Received 15 March 2019
Revised 13 June 2019
Accepted 13 June 2019
Keywords:
IrOx clusters
Zeolite
Glycerol
C–O hydrogenolysis
1,3-Propanediol
Ó 2019 Elsevier Inc. All rights reserved.
1. Introduction
sought-after target for making value-added chemicals from bio-
renewable glycerol [11].
The development of cost-effective bio-based platforms for the
production of renewable chemicals was critical for sustainable
chemical industries [1,2]. The large amount of glycerol as the
byproduct from biodiesel production has attracted considerable
research interest from the viewpoint of developing feasible pro-
cesses to transform this particular triol molecule into more
value-added compounds [3–8]. In this context, selective C–O
hydrogenolysis of glycerol to 1,2-propanediol (1,2-PDO) and 1,3-
propanediol (1,3-PDO) has been known to be a promising and eco-
nomically viable route [9]. Especially, 1,3-PDO was of great interest
because it was currently produced by a cost-intensive petrochem-
ical multi-step route from ethylene oxide following a sequential
hydroformylation-hydrogenation pathway [10]. Driven by its
emerging application as an essential monomer for making poly-
trimethylene terephthalate (PTT), 1,3-PDO has become a highly
Tremendous research efforts have been dedicated to the glyc-
erol hydrogenolysis, and a number of catalysts exhibited good effi-
ciency with high 1,3-PDO selectivity under specified reaction
conditions [12–29]. Compared with the removal of the C1 hydroxyl
group, that of the C2 position was kinetically and thermodynami-
cally unfavorable [14]. The intensive search for a promising cata-
lyst to convert glycerol into 1,3-PDO has contributed to the
development of a range of bifunctional catalysts based on various
combinations of bimetal-acid catalyst pairs, showing reasonable
selectivity control in glycerol hydrogenolysis [12,13,15–23]. In par-
ticular, interest lied in the discovery of novel bifunctional catalysts
involving the association of a platinum group metal (PGM) and
oxophilic metal oxides, with representative examples including
Pt-WOx and Ir-ReOx [14]. Owing to their favorable cooperative
properties capable of promoting desired dehydration hydrogena-
tion pathways, several recent studies have shown that it was pos-
sible to attain high 1,3-PDO selectivity and formation rate by
carefully controlling the interaction between the noble metal and
oxophilic acid center [12–29]. It was widely accepted that Ir/Pt
⇑
Corresponding authors.
E-mail addresses: yhyang@njtech.edu.cn (Y. Yang), yongcao@fudan.edu.cn
(Y. Cao).
https://doi.org/10.1016/j.jcat.2019.06.025
0021-9517/Ó 2019 Elsevier Inc. All rights reserved.

340
X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
and Brönsted acid sites played vital roles in converting glycerol
toward high selectivity of 1,3-PDO. Ir/Pt decomposed H2 into H
or H+-HÀ; Brönsted acid sites, typically WOx, ReOx anchored the
1° OH group of glycerol, forming a strongly bonded terminal alkox-
ide, and protonating the internal OH group of glycerol (the 2°).
Interestingly, Qiao group also reported that Brönsted acid sites
being generated in-situ on the Pt/W–SBA-15 catalysts gave rise
to the highest 1,3-PDO selectivity of 70.8% at a high glycerol con-
version [25].
to a mortar with 10 mL of ethanol. After grinding into powdery
solid with homogeneous faint yellow color, the resulting powder
was dried in vacuum for 5 h and finally reduced in 5 vol% H2/Ar
at 573 K for 4 h. After cooling to room temperature, the resulting
sample was passivated in the air and then stored in a dryer for
their subsequent manipulation or characterization. For the conven-
tional impregnation method, the H-ZSM-5_100 support material
was placed in an aqueous solution of H2IrCl6. After being stirred
for 24 h, the sample was dried at 343 K, followed by drying at
323 K in vacuum for 12 h, and finally heated under a 5 vol %H2/
Ar gas flow at 573 K for 4 h. The noble metal loadings in all these
samples were ca. 2.0 wt%.
Preparation of metal-oxide supports: ZrO2 support was pre-
pared by a routine precipitation with ammonia solution as precip-
itation agent [39]. The crystal phase of the resultant ZrO2 was
composed of 44% tetragonal phase and 56% monoclinic phase
based on X-ray diffraction (XRD) analysis, with a specific surface
area of 110 m2 gÀ1 [40]. Nb2O5 powders were obtained from an
air calcination of niobium oxalate at 773 K for 4 h [41]. For SiO2–
Al2O3 samples, TEOS was dropped into nitric acid and aluminium
nitrate (Si/Al = 1) solution under vigorous stirring. After aging the
homogenized mixture at 323 K for one week, the xerogel thus
obtained was finally air-calcined at 873 K for 2 h [42]. The partially
Na-exchanged ZSM-5 (Na-ZSM-5_100) was obtained by subjecting
a commercial NH4-ZSM-5_100 sample to a routine ion-exchange
using a NaNO3 (0.05 M) aqueous solution [43].
Ir-ReOx/SiO2 catalyst in the presence of sulfuric acid [12] and Ir-
Re/KIT-6 catalyst (Ir-Re alloy on silica) in the presence of solid acid
amberlyst [15] have also been shown to be effective in the liquid
phase batch conversion of glycerol to 1,3-PDO (Table S1). However,
it still remained a challenge to achieve high selectivity of 1,3-PDO
over monometallic Ir-based (Re-free) catalyst in the absence of
acid additives. In our quest to develop a feasible monometallic
catalyst for glycerol hydrogenolysis, we were interested in the
multivalent iridium-based (IrOx) catalyst that may eventually
bypass the use of oxophilic metal promoters. Two intriguing points
were considered: (1), DFT calculations proved that Y zeolite and
noble-metal (Pd, Pt, Rh, Au, Ir) clusters can interact to form a
‘‘metal-proton adduct” (M-O(H)-Y zeolite) [30–33], and it was
subsequently demonstrated in the conversion of neopentane, it
has been shown that Pd (or Pt)–O(H)–Y zeolite catalysts displayed
a remarkably higher catalytic activity per exposed metal site than
Pd (or Pt)–Y zeolite [30,31]. This finding inspired us to design a
new Ir-based catalyst consisting of zeolite, which could afford
‘the Brönsted acid sites’ Ir-O(H)-zeolite, leading to the highly effi-
cient and selective synthesis of 1,3-PDO. (2), IrOx-based materials
possessed excellent anticorrosive, oxidation-resistant properties
and superior electrochemical performances. These materials have
attracted extensive attention as a promising electrode candidate
for improving energy conversion and storage efficiency [34–38].
Nonetheless, despite these attractive characteristics, the potential
of IrOx entities for the liquid-phase catalytic transformation,
particularly toward the bio-based chemical production, remained
considerably unexplored.
2.2. Characterization
Inductively coupled plasma atomic emission spectroscopy (ICP-
AES) was applied to determine the metal content in the catalysts
(Thermo Electron IRIS Intrepid II XSP spectrometer). X-ray diffrac-
tion (XRD) measurements were operated using a copper X-ray
source with a nickel filter (40 kV, 40 A) and recorded by a Bruker
X-ray diffractometer D8 Advance. Transmission electron micro-
scopy (TEM) images were recorded by a FEI Tecnai G2 F20 S-
Twin micro scope working at 200 kV. The TEM samples were pre-
pared by dropping ethanol dispersed test materials on carbon-
coating copper grid. Argon physisorption was performed on a
Micromeritics ASAP 2010 system at 77 K. The specific surface area
was estimated with the Brunauer-Emmett-Teller (BET) method in
the approximate relative pressure range 0.05–0.3. The total pore
volume was calculated with single point desorption analysis. X-
ray photoelectron spectroscopy (XPS) analysis was performed on
a Quantum 2000 Scanning ESCA Microprobe (Physical Electronics)
In this study, a new bifunctional catalyst comprising multiva-
lent IrOx nanocluster anchored on H-ZSM-5 zeolite was developed
for the selective upgrading of glycerol under aqueous hydrogenol-
ysis conditions in the absence of any acidic additive. High activity
and 1,3-PDO selectivity could be achieved under optimized reac-
tion conditions. The physicochemical state of the IrOx species and
their interfacial synergistic interaction with H-ZSM-5 surface were
speculated to be the key factors for the enhanced catalytic
performance.
with Al-Ka radiation (1846.6 eV). Temperature-programmed
reduction (H2-TPR) was performed by using ASAP 2020
(Micromeritics instrument). After being pretreated at 473 K for
2 h under flowing He, the sample (100 mg) was cooled down to
room temperature in flowing He. The experiment was operated
in a 5 vol% H2/Ar flow (40 mL minÀ1), with raising the temperature
to 1000 K (5 K minÀ1).
2. Experimental
2.1. Materials and synthesis
Commercially available synthetic zeolites including Beta (BEA,
Temperature-programmed desorption of ammonia (NH3-TPD)
was carried out on a chemical adsorption analyzer (AutoChem II
2920, Micromeritics Instrument). The sample was degassed in-
situ in a helium flow (99.99%) and cooled down to 373 K. The
adsorption of NH3 was conducted by exposing the pretreated sam-
ple to a 10 vol% NH3/He stream at 373 K for 1 h. After purging the
weakly adsorbed and gas phase NH3 with He (99.99%) at 373 K for
0.5 h, the temperature was linearly ramped up to 823 K at a rate of
10 K minÀ1 in a high purity He (99.99%) flow. The signal of the des-
orbed NH3 was monitored by an on-line quadrupole mass spec-
trometer (m/e = 16). The area of the desorption peak was
integrated and employed to quantify the amount of the acid sites
Si/Al = 25),
Y (Si/Al = 6.5), Mordenite (Si/Al = 25) and ZSM-5
(Si/Al = 30, 60, 100, 120, 200) were received in their H-forms from
Nan Kai University Catalyst (Tianjin) Co. Ltd. For the sake of brev-
ity, the number appeared at the end of a given zeolite-based mate-
rial denoted the specified Si/Al ratio of this sample. Different
commercially available fumed oxide products including SiO2
(Aerosil-380), Al2O3 (AEROXIDE Alu C), and TiO2 (P25) were
obtained from Evonik. All other chemicals and solvents (reagent
grade) were purchased from Aladdin Reagent (Shanghai) Co., Ltd.
Preparation of IrOx supported catalysts by grinding-assisted
impregnation was carried out as follows: In brief, the support
was placed in 10 mL of an aqueous solution of hexachloroiridic
acid (H2IrCl6). The sonically treated suspension was transferred
by referring to that calibrated by a 100
lL capacity loop using
the same adsorbate. Pyridine-adsorption Fourier transform infra-

X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
341
red (FT-IR) spectra were collected on a Nicolet 6700 instrument
using a mercuric cadmium telluride detector. The sample was pre-
treated in vacuum at 673 K for 30 min. After cooling down to
423 K, the adsorption of pyridine was conducted at 423 K for a suf-
ficient amount of time. After degassing for 1 h to remove weakly
physisorbed and gas phase pyridine, the FT-IR spectra were
recorded.
Table S2, the BET surface areas of zeolite supports vary in the range
of 250–600 m2 gÀ1, and loadings Ir slightly decreases the surface
area of zeolite support. In addition, total pore volumes of Ir cata-
lysts are the same as zeolite supports, implying that Ir may load
on the zeolite surfaces.
XPS spectra in Fig. 2i–l show two sets of Ir 4f lines comprising
four asymmetric peaks, corresponding to Ir 4f7/2 and Ir 4f5/2
,
respectively, confirming that all samples exclusively comprise
metallic Ir0 and oxidized Ir3+. The quantitative deconvolution of
the Ir 4f lines reveals oxidized iridium (Ir3+) as the predominant
phase. Additionally, Cl 2p XPS spectra of the resulting powder after
grinding (unreduced by H2) confirm that it comprises Cl (Figure S1
2.3. Catalytic measurements
The hydrogenolysis was performed in a high pressure Parr reac-
tor (Hastelloy-C type, 50 mL), charged with a mixture of glycerol
(1.13 g), water (10 mL) and the catalyst powder (0.4 mol% vs sub-
strate). After purging with H2 for three times, the reactor was pres-
surized to 8 MPa. If not specified, the reaction temperature was
453 K, and the stirring rate being 800 rpm. The gas products were
analyzed on a Agilent 6820 gas chromatograph fitted with a 2 m-
long TDX-01 packed stainless steel column connected to a thermal
conductivity detector (TCD) and on a ECHROM A90 gas chro-
matograph fitted with a 3 m-long Porapak Q packed-column con-
nected to a flame ionization detector (FID). The liquid products
were analyzed on an Agilent 7820A gas chromatograph using a
(a)), while Cl 2p peak is not perceptible on the final catalyst (Fig-
z+
ure S1(b)). Metallic chloride [MCly] and H-ZSM-5 form [MH(y-z)
]
after removing Cl as HCl during reduction in H2. And the MOx
z +
and [M(OH) (y-z)
]
ÀHz+ can be generated through subsequent oxi-
z+
dation of [MH(y-z)
]
in O2 [44]. Therefore, IrOx forms.
Fig. 2e–h show the TEM characterization results. The narrowly-
distributed Ir domains are evenly dispersed on the external sur-
faces of zeolites (H-Y, H-ZSM-5, H-Beta, and H-Mordenite). The
sizes of the IrOx particles vary in the range of 0.5 and 3.5 nm (mean
size ca. 1.5 nm). Notably, according to TEM images shown in
Fig. 2a–d, conventional impregnation (IM) can only afford the for-
mation of randomly distributed nanostructures comprising of pri-
marily IrOx particle sizes ranging between 3 and 7 nm, some
considerably larger particles sizes up to around 12 nm over the
zeolite surfaces. This observation is consistent with the established
facts that the dispersion of a supported noble metal is markedly
affected by the preparation methods [45–48] and addition of etha-
nol in grinding [49,50] which have a substantial effect on the
phase- and size-dispersion of the metal-oxide nanoparticles.
HP-INNOWax capillary column (30 m  0.32 mm  0.25
lm) and
FID. 1,4-butanediol was added as the internal standard. The carbon
balance (the fraction of all products relative to that of the substrate
on a carbon molar basis) was above 96% in this study.
3. Results and discussion
3.1. Structural uniformity of zeolite-supported IrOx
3.2. Catalytic performance
Fig. 1 shows the wide-angle XRD patterns of various zeolite sup-
ports (H-Y, H-Beta, H-Mordenite and H-ZSM-5) and the corre-
sponding supported Ir catalysts prepared by the grinding-assisted
impregnation method, followed by 5 vol% H2/Ar reduction at
573 K (see the Experimental section for details). All zeolites and
Ir-based catalysts exhibit the characteristic peaks of FAU/BET/
MOR/MFI structures (JCPDS, PDF No. 39-1380, No. 47-0183, No.
43-0171, No. 42-0024), suggesting the single-phase zeolitic assem-
blies. Moreover, loading Ir does not result in the destruction of the
crystal structure for all the supported Ir-based catalysts, and no
peaks corresponding to Ir-containing phases are detected, possibly
due to highly dispersed and small-sized nanoparticles. The ICP
measurements indicate that the Ir species are loaded on the sup-
ports regardless of the zeolite types (Table S2). As seen in
Hydrogenative upgrading of aqueous glycerol is studied over
the different zeolites/metal oxides supported IrOx nanoparticles
under 8 MPa of H2 at 453 K (Table 1). From these preliminary stud-
ies, aluminum-containing H-ZSM-5 possessing a bulk molar Si/Al
ratio of 30 outperforms as the most suitable support material,
affording the preferential formation of 1,3-PDO with a moderate
selectivity of ca. 55.0% and the highest activity of 5.42 hÀ1 (Table 1,
entry 4), along with 1-propanol as the main by-product. Moreover,
besides SiO2–Al2O3, the commonly used oxide-based acidic sup-
port materials, such as Al2O3, SiO2, ZrO2, and Nb2O5 can only render
the predominant formation of 1,2-PDO under otherwise identical
reaction conditions (Table 1, entries 5–10). In addition, activity
for glycerol conversion can hardly be observed over Ir-free
H-ZSM-5 (Table 1, entry 11), revealing that the combination of IrOx
and H-ZSM-5 is essential to promote the selective hydrogenolysis.
Encouraged by these preliminary results, optimizing the cat-
alytic behavior of IrOx/H-ZSM-5 is attempted. Firstly, improving
the activity of this particular catalyst via altering the Si/Al ratio
of H-ZSM-5 is examined. Consistent with those previous studies
on utilizing H-ZSM-5-based materials to facilitate acid-catalyzed
or related reactions [48,51], a prominent Si/Al ratio-dependent
behavior emerges for the conversion of glycerol over these cata-
lysts. Notably, IrOx supported H-ZSM-5 with a Si/Al ratio of 100
(hereafter referred as IrOx/H-ZSM-5_100) performs the best
(Table 2, entries 1–5). Interestingly, over IrOx/H-ZSM-5_100_C,
which formed through calcination of the powder in air after
grinding, its activity is only half as much as the activity over
IrOx/H-ZSM-5_100, which obtained by reducing the powder in H2
after grinding (Table 2, entry 6 and Table 2, entry 3). And i-PO
and n-PO are main products over IrOx/H-ZSM-5_100_C (Table 2,
entry 6). More intriguingly, the IrOx/H-ZSM-5_100, exhibits an
exceptional selectivity of ca. 75.3% toward 1,3-PDO at an initial
Fig. 1. XRD patterns for H-form zeolites and IrOx supported H-form zeolites.

342
X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
Fig. 2. TEM micrographs, size distributions, and Ir 4f XPS spectra of IrOx nanoparticles loaded on H-form zeolites: (a) IrOx/H-Y_6.5-IM, (b) IrOx/H-Beta_25-IM, (c) IrOx/H-
Mordenite_25-IM, and (d) IrOx/H-ZSM-5_30-IM pretreated by conventional impregnation. (e and i) IrOx/H-Y_6.5, (f and j) IrOx/H-Beta_25, (g and k) IrOx/H-Mordenite_25, and
(h and l) IrOx/H-ZSM-5_30 pretreated by grinding-assisted impregnation.
Table 1
Catalytic behavior of IrOx nanoparticles loaded on various supports for the hydrogenolysis of glycerol to 1,3-propanediol.a
Entry
Catalyst
Activitya (hÀ1
)
Selectivityb (%)
1,3-PDO
1,2-PDO
i-PO
n-PO
1
2
3
4
5
6
7
8
IrOx/H-Y
0.76
0.90
2.19
5.42
0.88
1.33
0.96
1.26
3.53
3.04
0.17
37.5
19.5
16.0
55.0
36.8
15.0
28.0
18.2
7.9
43.2
27.9
50.4
9.8
37.0
72.2
63.0
56.4
73.4
77.4
0
6.2
1.1
4.4
1.0
2.0
3.7
2.2
3.7
4.3
4.6
0
13.1
51.0
29.2
33.1
23.9
9.2
IrOx/H-Beta
IrOx/H-Mordenite
IrOx/H-ZSM-5
IrOx/SiO2-Al2O3
IrOx/Al2O3
IrOx/SiO2
IrOx/ZrO2
IrOx/Nb2O5
IrOx/TiO2
H-ZSM-5
5.8
21.7
13.5
12.0
0
9
10
11
4.9
0
a
Reaction conditions: 10 wt% glycerol, 11.13 g; n(glycerol)/n(metal), 250; H2, 8 MPa; temperature, 453 K, 10–15% glycerol conversion.
1,3-PDO, 1,2-PDO, i-PO, and n-PO denote 1,3-propanediol, 1,2-propanediol, 2-propanol, and 1-propanol, respectively.
b
stage with an activity of 4.50 hÀ1 (Table 2, entry 7). In particular,
the selectivity of 1,2-PDO, the unwanted by-product is effectively
suppressed to a low level of 3.5%. The results in this study demon-
strate that monometallic Ir catalyst indeed can efficiently catalyze
the glycerol hydrogenolysis and it is in contrast to those previous
reports in which the presence of rhenium is crucial to facilitate
the abstraction of the C2-hydroxyl group in glycerol [12,15,27].
The reaction time dependence of the glycerol hydrogenolysis
over IrOx/H-ZSM-5 is shown in Fig. 3a. The selectivity toward
1,3-PDO is high in the first 6 h and it decreases with increasing
reaction time. On the other hand, the selectivity to 1-propanol
gradually increases with increasing the reaction time. Subsequent
investigation focuses on the effect of H2 pressure on glycerol
hydrogenolysis over IrOx/H-ZSM-5_100 at 453 K (Fig. 3b). Glycerol
conversion dramatically boosts from 6.1% to 21.5% with increasing
H2 pressure from 2.0 to 8.0 MPa. Meanwhile, the 1,3-PDO selectiv-
ity increases from 44.6% to 69.6% and concurrently the selectivity
to 1,2-PDO decreases gradually with increasing H2 pressure except
the higher pressure range (ꢀ4 MPa). Thus, increasing hydrogen
pressure is favorable to form 1,3-PDO. Moreover, a separate

X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
343
Table 2
Catalytic behavior of IrOx nanoparticles loaded on H-ZSM-5 zeolites for the hydrogenolysis of glycerol to 1,3-propanediol.a
Entry
Catalyst
Activity (hÀ1
)
Selectivityb (%)
1,3-PDO
n1,3-PDO/n1,2-PDO
1,2-PDO
i-PO
n-PO
1
2
3
4
5
6
7
8
IrOx/H-ZSM-5_30
IrOx/H-ZSM-5_60
5.21
4.88
4.48
2.96
2.10
2.92
4.50
0.52
2.08
2.65
3.04
40.5
60.7
69.6
54.4
50.0
17.3
75.3
18.5
0
10.2
3.2
3.2
11.8
29.4
10.7
3.5
76.2
0
2.0
2.1
1.8
0.2
1.6
16.6
1.3
3.1
38.2
2.3
1.2
47.3
34.0
24.8
32.8
18.9
48.9
18.0
2.1
4.0
19.0
21.8
4.6
1.7
2.2
21.5
0.2
–
2.6
13.3
IrOx/H-ZSM-5_100
IrOx/H-ZSM-5_120
IrOx/H-ZSM-5_200
IrOx/H-ZSM-5_100_Cc
IrOx/H-ZSM-5_100d
IrOx/Na-ZSM-5_100
IrOx/H-ZSM-5_100-IM
IrOx/H-ZSM-5_100e
IrOx/H-ZSM-5_100f
9
10
11
61.7
34.1
38.6
45.8
55.8
17.4
4.2
a
b
c
d
e
f
Reaction conditions: 10 wt% glycerol, 11.13 g; n(glycerol)/n(metal), 250; H2, 8 MPa; temperature, 453 K; time, 12 h.
1,3-PDO, 1,2-PDO, i-PO, and n-PO denote 1,3-propanediol, 1,2-propanediol, 2-propanol, and 1-propanol, respectively.
IrOx/H-ZSM-5_100 after grinding by air-calcined at 823 K.
reaction time, 6 h.
IrOx/H-ZSM-5_100 after grinding reduced at 473 K.
IrOx/H-ZSM-5_100 after grinding reduced at 673 K.
Fig. 3. (a) Time course, (b) effect of H2 pressure and (c) reusability of the of the IrOx/H-ZSM-5_100 catalyst for the hydrogenolysis of glycerol. Reaction conditions: 10 wt%
glycerol, 11.13 g; n(glycerol)/n(metal), 250; H2, 8 MPa (or 2 MPa, 4 MPa); temperature, 453 K, time, 12 h (or 6 h, 24 h, 36 h, 48 h). 1,3-PDO, 1,2-PDO, i-PO and n-PO denote 1,3-
propanediol, 1,2-propanediol, 2-propanol and 1-propanol, respectively. The catalyst lost about 5 wt% mass in each repetitive recovery process. Therefore, in the fifth run,
0.38 g recovered catalyst, 8.9 g 10 wt% glycerol aqueous were used.
À1
reusability test indicates that the activity/selectivity of IrOx /H-
ZSM-5_100 can be maintained for at least five times (Fig. 3c). In
good agreement with this observed durability, the structural char-
acterizations by XRD, TEM, and XPS do not reveal any noticeable
change in both fresh and spent IrOx/H-ZSM-5_100 catalysts
(Fig. 4). Table S11(b) summarizes the specific activities expressed
in terms of the 1,3-PDO formation rate (mmol1,3-PDO mol
metal
timeÀ1). As shown in Table S1, Pt-W bimetallic catalytic systems
have higher conversion of glycerol and selectivity of 1,3-PDO than
Ir-Re bimetallic catalytic systems, while Ir-Re bimetallic catalytic
systems possesses higher 1,3-PDO formation rate than Pt-W
systems. The IrOx/H-ZSM-5_100 catalyst exhibits a glycerol-to-
Fig. 4. Structural properties of IrOx/H-ZSM-5_100 reused four times for the hydrogenolysis of glycerol: (a) XRD pattern, (b) Ir 4f XPS spectra, (c) TEM micrographs and size
distributions.

344
X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
1,3-PDO formation rate of 5.4 mol1,3-PDO mol hÀ1 under opti-
ZSM-5 materials. Based on the results obtained from Py-IR (Table 3,
the first ten lines), the Brönsted acid concentrations and amounts
over the IrOx-loaded catalysts, compared with their corresponding
pristine H-ZSM-5 supports, are increased. Such increase in the sur-
face Brönsted acidity in the IrOx-loaded catalysts corresponded to
the generation of new acidic sites that is reasonable to be specu-
lated as a key factor in promoting the site-selective deoxygenation
of glycerol toward 1,3-PDO.
To further validate the role of Brönsted acid sites in facilitating
this reaction, a separate control experiment involving the use of
partially Na-exchanged H-ZSM-5_100 as the support material
under otherwise identical reaction conditions is carried out
(Table 2, entry 8). This material with significantly attenuated
Brönsted acidity (Table 3, the next-to-last entry, Figure S4) exhibits
considerably inferior performance as compared to its Na-free ana-
log (Table 3, the third line from bottom, Figure S5). In other words,
the decline of Brönsted acid amount and lack of Ir-induced acid
sites in close proximity to IrOx domains result in the remarkably
poor catalytic activity (0.52 hÀ1) and 1,3-PDO selectivity (18.5%)
over IrOx/Na-ZSM-5_100 surface, which are far below the perfor-
mances of IrOx/H-ZSM-5_100 (4.48 hÀ1, 69.6%, Table 2, entry 3).
Interestingly, despite bearing similar low level of surface Brönsted
acidity, the performance IrOx/Na-ZSM-5_100 sample is still
remarkably lower than that of IrOx/H-ZSM-5_200.
To understand the benefits of well-dispersed IrOx for glycerol
hydrogenolysis, the IrOx/H-ZSM-5_100-IM sample obtained from
conventional impregnation followed by the same above-
mentioned pre-activation step is evaluated under the optimal reac-
tion conditions. This catalyst, however, appears to be particularly
effective for the exclusive formation of C3 monohydric alcohols
(Table 2, entry 9), albeit with a activity approximately 2.2 times
less than that achieved for its grinding-derived counterpart. These
results, along with the TEM and XPS characterizations shown in
Fig. 5a–d, emphasize the importance of precisely controlling the
dimensions of the IrOx domains to ensure the highly active and
selective conversion of glycerol. Nevertheless, this argument does
not permit a full explanation of the striking product distribution
change between IrOx/H-ZSM-5_100 and IrOx /H-ZSM-5_100-IM.
In this regard, one may notice that, in contrast to its grinding-
À1
Ir
mized conditions. Essentially, this value is comparable to those
obtained in ReOx- or WOx- promoted catalysts (Table S1).
3.3. Origin of the enhanced performance
To rationalize the observed trend in catalytic performances, the
above IrOx/H-ZSM-5-based catalysts, along with their correspond-
ing pristine zeolite supports, are subjected to the combination of
microscopic and spectroscopic characterizations. TEM and XPS
measurements confirm that all IrOx/H-ZSM-5-based samples with
different Si/Al ratios comprise well-dispersed IrOx nanoparticles
with an average size and chemical states are essentially identical
(Fig. 2h, 2 l, Fig. 5a and b, Figure S3). NH3-TPD analyses reveal that
the acid sites in pristine and IrOx-loaded H-ZSM-5 appear in two
temperature regions, i.e., at 373–523 K and around 573–673 K,
assigned to the weak and moderate acid sites, respectively (Fig. 6
(A)). Table S3 summarizes the acid strength distribution quantita-
tively estimated from the amount of ammonia desorbed from dif-
ferent regions. As the Si/Al ratio in H-ZSM-5 increases from 30 to
200, considerable loss of the weak and moderate acid sites occurs.
Notably, loading IrOx induces the genesis of new mild acid sites
without affecting the integrity of moderate acidity over H-ZSM-5
support, thereby increasing the total acidity of IrOx-based catalysts
compared to Ir-free zeolites. These findings imply that an appropri-
ate acid content and the Ir-induced mild acid sites are crucial for
the selective 1,3-PDO production from glycerol over IrOx supported
on zeolite.
Additional insights regarding the essential features of the acid
sites are acquired from IR spectroscopic studies of pyridine absorp-
tion. The IR bands at around 1455 and 1543 cmÀ1 can be ascribed
to the Lewis and Brönsted acid sites, respectively. The IR band at
1490 cmÀ1 results from both Brönsted and Lewis acid sites of these
catalysts [52–55]. The characteristic pyridine-adsorbed IR bands
corresponding to Lewis acid (1446–1452 cmÀ1) and Brönsted acid
sites (1540–1544 cmÀ1) are observed in all samples (Fig. 6(B)). In
line with the information regarding the enhanced total acidity as
inferred from NH3-TPD, the peak areas related to the IrOx-
containing samples are consistently larger than those of the H-
Fig. 5. Structural properties of IrOx/H-ZSM-5_100 pretreated by grinding-assisted impregnation (GM) or conventional impregnation (IM) and reduced at different
temperatures. (a and b), IrOx/H-ZSM-5_100 pretreat by GM and reduced at 573 K. (c and d), IrOx/H-ZSM-5_100-IM pretreat by IM and reduced at 573 K. (e and f) IrOx/H-ZSM-
5_100 pretreat by GM and reduced at 473 K. (g and h) IrOx/H-ZSM-5_100 pretreat by GM and reduced at 673 K.

X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
345
Fig. 6. (A) NH3-TPD profiles and (B) Pyridine-adsorbed FT-IR spectra of H-ZSM-5 (a–e) and IrOx/H-ZSM-5 (f–j) with different Si/Al ratio: (a and f) Si/Al = 30, (b and g)
Si/Al = 60, (c and h) Si/Al = 100, (d and i) Si/Al = 120, (e and j) Si/Al = 200.
Table 3
Acidities of H-ZSM-5 and IrOx/H-ZSM-5 with different Si/Al ratios.
Sample
Acid concentrationa (%)
Brönsteda
Acid amount (mmol gÀ1
Brönstedc
)
Lewisa
Lewisc
Totalb
H-ZSM-5_30
IrOx/H-ZSM-5_30
H-ZSM-5_60
IrOx/H-ZSM-5_60
H-ZSM-5_100
IrOx/H-ZSM-5_100
H-ZSM-5_120
IrOx/H-ZSM-5_120
H-ZSM-5_200
IrOx/H-ZSM-5_200
Na-ZSM-5_100
78
79
73
76
51
60
26
32
17
24
16
15
22
22
21
27
24
49
40
74
68
83
76
84
85
78
0.48
0.50
0.38
0.42
0.22
0.28
0.09
0.12
0.03
0.05
0.06
0.05
0.11
0.13
0.13
0.14
0.13
0.21
0.18
0.26
0.25
0.12
0.12
0.30
0.27
0.30
0.61
0.63
0.52
0.55
0.43
0.46
0.35
0.37
0.15
0.17
0.36
0.32
0.41
IrOx/Na-ZSM-5_100
IrOx/H-ZSM-5_100_Cd
a
Acid concentration (%) of the Brönsted and Lewis acid sites determined by pyridine-IR using extinction coefficients of
= 2.94 cm
molÀ1
e(Brönsted) = 1.67 cm l e(Lewis)
molÀ1 and
l
.
b
Amount of the total acid determined by NH3-TPD and pyridine-IR.
Amount of Brönsted and Lewis acids obtained from NH3-TPD and pyridine-IR.
IrOx/H-ZSM-5_100 after grinding by air-calcined at 823 K.
c
d
derived counterpart, the IrOx/H-ZSM-5_100-IM material prepared
by impregnation exhibits an obviously reduced surface Brönsted
acidity (Table S4), relative to that of the pristine H-ZSM-5_100 sup-
port. Given the indispensable importance of creating additional
surface Brönsted acid sites in close proximity to IrOx domains for
preferential 1,3-PDO formation, the lack of sufficient and desirable

346
X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
metal-acid bifunctional synergy appears to be a more decisive fac-
tor, leading to the predominant formation of C3 monohydric alco-
hols over the IrOx/H-ZSM-5_100-IM catalyst.
potential to meet the ever-pressing needs for renewable
transportation fuels and polymer processing. Although several
ReOx-containing Ir- or Rh-based catalysts are known to be effective
for this key transformation [57–59], it is of significance to develop
new catalyst that can facilitate the convenient, practical
production of 1,6-HDO from 1,2,6-HTO under Re-free conditions.
By subjecting the IrOx/H-ZSM-5_100 catalyst to a 5 wt% 1,2,6-
HTO-containing aqueous solution under conditions similar to
those applied to glycerol, 12 h of the reaction produces the desired
1,6-HDO with a selectivity of 65% at 55% 1,2,6-HTO conversion,
confirming that the present catalytic protocol may indeed be used
for this type of transformation. In addition, IrOx/H-ZSM-5_100 is
briefly assessed for simultaneous vicinal hydroxyl removal using
1,4-anhydroerythritol (1,4-AHERY) as a representative model com-
pound. In this particular case, the IrOx/H-ZSM-5_100 catalyst exhi-
bits modest activity, and desired tetrahydrofuran (THF) is obtained
in 54% yield after 18 h (Scheme 1b). Further optimization is needed
to attain THF yields greater than those reported previously over the
ReOx-Pd/CeO2- based catalyst [60]. The prominent 1,4-AHERY-to-
THF selectivity achieved with IrOx/ H-ZSM-5_100 once again high-
lights the catalytic potential of multivalent forms of iridium for the
Re-free reductive upgrading of biofeedstocks.
The above-mentioned results highlight the distinct, critical
interaction between highly dispersed IrOx and H-ZSM-5 support
for facilitating 1,3-PDO formation through a desired hydrogenoly-
sis. To further understand the role of the Ir3+ species in preferen-
tially facilitating the hydroxyl abstraction at the C2 position of
the glycerol molecule, an additional set of experiments aiming at
elucidating the effect of reductive pretreatment on IrOx/H-ZSM-
5_100-mediated glycerol hydrogenolysis are carried out. By
subjecting the as-ground Ir-H-ZSM-5_100 precursor to thermal
treatment under H2/Ar over the temperature range from 473 to
673 K, samples comprising essentially identically sized IrOx parti-
cles (ca. 1.5 nm, Fig. 5a, e, g), with the co-existence of the Ir0- Ir3+
components in different portions, are obtained. The reduction
degree of the Ir species calculated from XPS (Fig. 5b, f, h) is 15%
at 473 K, 40% at 573 K, and 60% at 673 K. Notably, prior to reduc-
tion, the unreduced catalyst sample contains only Ir4+ species
(Figure S6). This fact, along with the H2-TPR results shown in
Figure S7, demonstrates the inherent instability of the Ir4+ species
under moderate reductive conditions. The reduction peaks at 400 K
in the H2-TPR profile (Figure S7), probably ascribed to the reduc-
tion of Ir(IV) to Ir(III) and the one at 580 K could be Ir(III) to Ir
and formation of HCl evolution [44,56]. Thus, the best performance
toward activity and 1,3-PDO selectivity can only be achieved in the
presence of a proper Ir0- Ir3+ ratio when pretreatment is carried out
at 573 K (Table 2, entry 3 and entries 10–11). These findings high-
light the need for carefully controlled Ir0-Ir3+ compositions.
Lastly, the substrate extension of this IrOx-mediated method for
the hydrogenolysis of 1,2,6-hexanetriol (1,2,6-HTO) is examined,
which is another representative triol molecule structurally similar
to glycerol, to afford industrially important 1,6-hexanediol (1,6-
HDO). As an integral part of the modern biorefinery schemes, this
transformation attracts particular attention in the context of the
production of a renewable monomer diol (Scheme 1a) sourced
from 5-hydroxymethylfurfural (HMF), which is considered as a
versatile bio-based molecular platform that can offer immense
3.4. Discussion
The primary goal of this work is to develop a monometallic (in
the absence of Re oxophilic metal oxides as the promoters) catalyst
to selectively convert glycerol toward 1,3-PDO without adding acid
additives. In the present study, IrOx nanoclusters with a narrow
size distribution (0.5–3.5 nm) and a small mean size (ca. 1.5 nm)
supported on H-ZSM-5 prepared by a grinding-assisted impregna-
tion method exhibit remarkably high activity and 1,3-PDO selectiv-
ity under optimized reaction conditions. Several parameters are
identified as the key factors affording the satisfactory catalytic
performances such as the Ir domain dispersions, proper Ir0/Ir3+
and Si/Al ratios, Brönsted acid site over ZSM-5 support, etc.
Interestingly, loading Ir nanoclusters afford the formation of
Ir-induced Brönsted acid site in close contact to Ir domains and
Scheme 1. (a) Formation of 1,6-HDO from HMF. [c] Reaction conditions: 5 wt% 1,2,6-HTO; 10 mL H2O; n(1,2,6-HTO)/n(metal), 500; 4 MPa of H2; 423 K; 12 h. (b) Formation of
THF from glucose. [d] Reaction conditions: 10 wt% 1,4-AHERY; 1,4-dioxane, 10 mL; n(1,4-AHERY)/n(metal), 100; 8 MPa of H2; 413 K; 18 h.

X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
347
the catalytic activity and selectivity toward 1,3-PDO is correlated
with this particular Brönsted acid site.
exchange between Ir and OH groups by spillover [33]. Rosch
group’s DFT study indicated that metal clusters adsorbed on a zeo-
lite, forming M6(3H)/zeo as a result of reverse hydrogen spillover
(RHS) from zeolite OH group onto metal nanoparticles, affected
the stability of nearby bridging hydroxyl groups, thereby modify-
ing their deprotonation energies, hence the acidity of these groups
[32]. The study in this work is in accordance with these previously
reported results, showing that the generation of synergetic inter-
face acidic species improves the catalytic activity and be beneficial
for the increase of product selectivity.
As for the IrOx/H-ZSM-5-catalyzed hydrogenolysis of glycerol in
this work, activity tests and characterization results show that
activities are dependent on the amounts of Brönsted acid sites.
Fig. 7 shows the correlation between activity and the amounts of
overall acid/ Brönsted acid sites, as the amount of acid and
Brönsted acid increases, activity also increases linearly. Moreover,
1,3-PDO selectivity increases with the amount of additional
Brönsted acid sites, tentatively denoted as Ir-O(H)-H-ZSM-5, pre-
sumably the Brönsted acid sites generating over the synergetic
interface between IrOx and H-ZSM-5, as shown in Fig. 8. Similar
findings have been reported on the Pd/HY catalyst in hydrogenoly-
tic ring opening of methylcyclopentane (MCP) [30,31]. Sachtler
et al. reported that the ring enlargement of MCP catalyzed by the
‘metal-proton adduct’, formed by the interaction between zeolite
and metal clusters, was more effectively than by separating the
metal and acid sites [30,31]. Besides, Gates et al. showed experi-
mental evidence of hydrides on supported metal (Ir) clusters in
the absence of H2 in the Ir-supported Y zeolite, presumably equili-
brated with OH groups on the zeolite and demonstrated the
The formation of Ir-induced Brönsted acid site relates with the
presence of hydrogen gas in the reduction process based on our
experimental data below and reported Ref. [44]. NH3-TPD analyses
(Table 3, the last line) and pyridine-adsorbed FT-IR spectra of
IrOx/H-ZSM-5_100_C (the powder after grinding were calcined in
air at 823 K) was collected (Table 3, the last line, and Figure S8,
(c)). Contrasting with Brönsted acid site over H-ZSM-5, Brönsted
acid amount IrOx/H-ZSM-5 reduced in 5 vol% H2/Ar at 573 K
increases, while amount over IrOx/H-ZSM-5_100_C was lower than
Brönsted acid amount over H-ZSM-5_100 (Table 3, the last
entry). That is, the Ir-induced Brönsted acid site is generated in
the reduction process with presence of hydrogen gas. In view of
[Ga(OH)x](3Àx)+(x = 1, 2) cations and GaOx formation on the surface
of GaOx and H-MFI zeolite which is confirmed through experiment
and DFT study [44] reported by Alexis T. Bell and co-workers
(Figure S9), the grinding-assisted process results in the exchange
of Brönsted acid O–H groups by [IrCl3] cation. H2 reduction at
573 K of the as-exchanged zeolites results in the stoichiometric
removal of Ir-bound Cl as HCl and the formation of [IrH2] cations.
Upon O2 oxidation, the [IrH3] cations are oxidized to [Ir(OH)
H] ÀH or [Ir(OH)2] ÀH cation pairs and neutral IrOx species. This
process of removing the Cl and of generating Ir-oxide and Ir-
induced Brönsted acid site is similar to the process in Figure S9 [44].
Hydrogenated upgrading of aqueous glycerol is studied over
IrOx/H-ZSM-5_100_C (Table 2, entry 6). The activity for glycerol
conversion over IrOx/H-ZSM-5_100_C which is calcined in air is
lower than that over IrOx/H-ZSM-5_100 reduced in H2. Note that
without hydrogen gas, the reaction over IrOx/H-ZSM-5_100 can
take place, while the activity is only 0.8 hÀ1, and i-PO and n-PO is
the main products. This shows that with the presence of H2, Ir-
induced Brönsted acid site is form in the reduction process and
could be a result of dynamic balance between hydrogen spillover
and RHS in the reaction process.
Fig. 7. Relationship between activities and amount of (Brönsted) acid over IrOx/H-
ZSM-5 with different Si/Al ratios.
In fact, Brönsted acid sites facilitate the generation of 1,3-PDO,
while Lewis acid sites facilitate production of 1,2-PDO [11]. In glyc-
erol hydrogenolysis, Brönsted acid site and Lewis acid site both can
proceed dehydration via elimination of
a hydroxyl group
(Scheme 2). Selective elimination either a primary or a secondary
alcohol results 1,2-propanediol or 1,3-propanediol selectivity in
the transformation of glycerol. And in general, the resulting alde-
hyde is less stable than the ketone formed by elimination of a pri-
mary alcohol. A Brönsted acid will help to eliminate a secondary
alcohol, whereas a Lewis acid can more easily coordinated to a pri-
mary alcohol [61].
As for our catalysts, IrOx/H-ZSM-5_100 is rich in Brönsted acid
(amount of Brönsted acid: Lewis acid = 0.28:0.18, mmol gÀ1
,
Table 3) and Ir-induced Brönsted acid (0.03 mmol gÀ1, Table 3),
so it tends to eliminate the hydroxyl group of C2 position and
forms secondary carbocation, which finally produces mainly 1,3-
PDO (1,3-PDO selectivity, 69%, Table 2). Besides, IrOx/Na-ZSM-
5_100 contains plenty of Lewis acid (amount of Brönsted acid:
Lewis acid = 0.05:0.27, mmol gÀ1, Table 3) and none of Ir-induced
Brönsted acid, results in transformed glycerol towards 1,2-PDO
(1,3-PDO selectivity, 18%, Table 2).
There is an intriguing fact that the TEM results showed that the
sizes of the IrOx particles vary in the range of 0.5 and 3.5 nm (mean
Fig. 8. Relationship between 1,3-PDO selectivities and amount of Ir-O(H)-H-ZSM-5
over IrOx/H(or Na)-ZSM-5 with different Si/Al ratios.

348
X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
Scheme 2. Dehydration of glycerol and hydrogenation of carbonyl group.
size ca. 1.5 nm), which should be on the external surface of the
zeolite. However, small metal clusters less than 0.5 nm were also
reported locating inside the support pores with grinding-assisted
impregnation method [46,61,62]. So in this work, some IrOx clus-
ters less than 0.5 nm would be possible on the internal surfaces,
which should not be ruled out. Although, the Si/Al ratio mainly
affects the internal acid sites rather than the outer surface sites,
some studies confirmed the presence of catalytic active sites on
external surfaces because that reactants (molecular diameter ＞
0.51 nm  0.55 nm) are not likely to diffuse into the microporous
channels of an MFI zeolite [63,64]. And conversion & product dis-
tribution are strongly dependent on the external acidity which in
turn correlates well with the zeolite particle size [65]. Accordingly,
the activity and product selectivity have been influenced by the
Si/Al ratio of zeolites; whether the external or internal acid sites
play a critical role. With respect to our IrOx/HZSM-5 catalysts,
the IrOx particles induced Brönsted acid site are the activity site.
Both of external and internal acid of support H-ZSM-5 may affect
the generation of Ir-induced Brönsted acid site through hydrogen
spillover and reverse hydrogen spillover. As already reported, in
GaOx/ZSM-5, some of the regenerated Brönsted sites are stronger
than those in the original HZSM-5, while in Fe/ZSM-5 and
Zn/ZSM-5, such Brönsted sites are weaker [44,66].
The favorably accepted mechanism for the formation of 1,3-
PDO from glycerol hydrogenolysis is the Adsorption-
Dehydration- Dehydration- Hydrogenation- Hydrolysis route
(ADDHH route). In general, 1° OH group of glycerol is firstly
adsorbed on the surface acid sites of MOx (M = W, Re, Al) cluster,
forming a strongly bonded terminal alkoxide by dehydration. The
2° OH of the alkoxide is protonated by a nearby proton coming
from a Brönsted acid site. After dehydration, an adsorbed 2° carbo-
cation is formed. Subsequently, the carbocation is attacked by a
hydride species coming from the heterolytic activation of H2 on
the metallic active sites. Hydrolysis as the last step yields 1,3-
PDO. Note that the ADDHH route on the Pt/WOx/Al2O3 catalyst
has been experimentally evidenced by in-situ attenuated total
reflection infrared [27]. In the present study, activity tests and
characterization results allow ones to postulate the reaction mech-
anism that is in some ways akin to ADDHH route. As shown in
Scheme 3, 1° and 2° OH groups adsorb on IrOx and nearby Ir-O
Scheme 3. Proposed reaction mechanism for glycerol hydrogenolysis to 1,3-PDO over the IrOx/H-ZSM-5 catalysts.

X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
349
(H)-H-ZSM-5, respectively. Alkoxide and the secondary carboca-
Appendix A. Supplementary data
tion are formed successively through dehydration. Then, hydride
species activated on the IrOx attack the secondary carbocation,
and 1,3-PDO is obtained after hydrolysis. However, the distinctive
characteristics of the IrOx/H-ZSM-5-catalyzed hydrogenolysis of
glycerol is that the use of oxophilic metal oxides, primarily respon-
sible for the acid site, is replaced by the Brönsted acid site over zeo-
litic support, in particular, Ir-O(H)-H-ZSM-5 acid sites at interfaces.
The formation of this Ir-O(H)-H-ZSM-5 acid site at the inter-
faces can be rationalized by the balance between the hydride
spices formed from H2 activation over Ir and the RHS from the
hydrogen-rich zeolite surface to the anchored Ir nanoparticles.
We expect the protons around the Ir species via RHS have multiple
efficacies. The RHS is associated with the partial oxidation of IrOx
species by charge transfer. It can also stabilize IrOx sites against
being entirely reduced. Moreover, RHS ensures that the Brönsted
acid site is in close contact to the hydride site on Ir metallic sur-
faces, which facilitates the formation of 2° carbocation and the
attack of the hydride species for the fast hydrogenation of the 2°
carbocation.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcat.2019.06.025.
References
[1] J.N. Chheda, G.W. Huber, J.A. Dumesic, Liquid-phase catalytic processing of
biomass-derived oxygenated hydrocarbons to fuels and chemicals, Angew.
Chem. Int. Ed. 46 (2007) 7164–7183.
[2] A. Corma, S. Iborra, A. Velty, Chemical routes for the transformation of biomass
into chemicals, Chem. Rev. 107 (2007) 2411–2502.
[3] W.D. Bossaert, D.E. De Vos, W.M. Van Rhijn, J. Bullen, P.J. Grobet, P.A. Jacobs,
Mesoporous sulfonic acids as selective heterogeneous catalysts for the
synthesis of monoglycerides, J. Catal. 182 (1999) 156–164.
[4] P. Lauriol-Garbay, J. Millet, S. Loridant, V. Bellière-Baca, P. Rey, New efficient
and long-life catalyst for gas-phase glycerol dehydration to acrolein, J. Catal.
280 (2011) 68–76.
[5] D. Roy, B. Subramaniam, R.V. Chaudhari, Cu-based catalysts show low
temperature activity for glycerol conversion to lactic acid, ACS Catal.
(2011) 548–551.
1
[6] D.D. Falcone, J.H. Hack, A.Y. Klyushin, A. Knop-Gericke, R. Schlögl, R.J. Davis,
Evidence for the bifunctional nature of Pt–Re catalysts for selective glycerol
hydrogenolysis, ACS Catal. 5 (2015) 5679–5695.
[7] W. Oberhauser, C. Evangelisti, C. Tiozzo, F. Vizza, R. Psaro, Lactic acid from
glycerol by ethylene-stabilized platinum-nanoparticles, ACS Catal. 6 (2016)
1671–1674.
4. Conclusions
[8] J. Chaminand, L. aurent Djakovitch, P. Gallezot, P. Marion, C. Pinel, C. Rosier,
Glycerol hydrogenolysis on heterogeneous catalysts, Green Chem. 6 (2004)
359–361.
[9] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C. Della Pina, From glycerol to
value-added products, Angew. Chem. Int. Ed. 46 (2007) 4434–4440.
[10] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Improved utilisation of
renewable resources: new important derivatives of glycerol, Green Chem. 10
(2008) 13–30.
[11] J. Ten Dam, U. Hanefeld, Renewable chemicals: dehydroxylation of glycerol
and polyols, ChemSusChem 4 (2011) 1017–1034.
[12] Y. Nakagawa, Y. Shinmi, S. Koso, K. Tomishige, Direct hydrogenolysis of
glycerol into 1,3-propanediol over rhenium-modified iridium catalyst, J. Catal.
272 (2010) 191–194.
[13] Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Modification of Rh/
SiO2 catalyst for the hydrogenolysis of glycerol in water, Appl. Catal. B Environ.
94 (2010) 318–326.
[14] Y. Nakagawa, M. Tamura, K. Tomishige, Catalytic materials for the
hydrogenolysis of glycerol to 1,3-propanediol, J. Mater. Chem. A 2 (2014)
6688–6702.
[15] C. Deng, X. Duan, J. Zhou, X. Zhou, W. Yuan, S.L. Scott, Ir–Re alloy as a highly
active catalyst for the hydrogenolysis of glycerol to 1,3-propanediol, Catal. Sci.
Technol. 5 (2015) 1540–1547.
[16] L. Gong, Y. Lu, Y. Ding, R. Lin, J. Li, W. Dong, T. Wang, W. Chen, Selective
hydrogenolysis of glycerol to 1,3-propanediol over a Pt/WO3/TiO2/SiO2 catalyst
in aqueous media, Appl. Catal. A Gen. 390 (2010) 119–126.
[17] L.Z. Qin, M.J. Song, C.L. Chen, Aqueous-phase deoxygenation of glycerol to 1,3-
propanediol over Pt/WO3/ZrO2 catalysts in a fixed-bed reactor, Green Chem. 12
(2010) 1466–1472.
In this work, it has been demonstrated that the well dispersed
monometallic iridium nanoclusters supported H-ZSM-5 was effec-
tive and recyclable catalyst for the aqueous-phase selective
hydrogenolysis of glycerol toward 1,3-PDO in the absence of acid
additives. IrOx/H-ZSM-5_100 exhibited an excellent 1,3-PDO selec-
tivity of ca. 75% with an activity of 4.50 hÀ1, and could be reused at
least five times without the change of the activity and 1,3-PDO
selectivity. Characterization and catalytic testing proved that activ-
ity and 1,3-PDO selectivity depended on the physicochemical
states of iridium domains including Ir dispersions and proper Ir0/
Ir3+ ratios and more importantly the amounts of overall acid/
Brönsted acid sites. TEM and XPS results of IrOx clusters and cat-
alytic behaviors indicate that the smaller the IrOx particle size
(ꢁ1.7 nm), the higher activity and 1,3-PDO selectivity. With
increasing the Ir0-Ir3+ ratio, the activity and 1,3-PDO selectivity
increase firstly and then decrease. When Ir0-Ir3+ ratio was 40:60,
the activity and 1,3-PDO selectivity were the highest. The activity
and selectivity toward 1,3-PDO strongly correlated with overall
acid/Brönsted acid sites and the Ir-induced additional Brönsted
acid site Ir-O(H)-H-ZSM-5 at Ir-zeolite interfaces, respectively.
The Ir-induced acid sites were speculated as coming from the
interfacial synergistic interaction between IrOx and H-ZSM-5
through hydrogen spillover and RHS. The reaction mechanism
was proposed to elucidate the role of Ir-O(H)-H-ZSM-5 sites in
glycerol secondary C–O cleavage. This particular IrOx/H-ZSM-5 also
exhibited remarkable selective hydrogenolysis of structurally sim-
ilar triol molecules. This study highlighted new opportunities in
the use of multivalent forms of the IrOx nanoclusters for polyol
hydrogenolysis. Bifunctional IrOx-based catalysts with tunable
interface synergy showed immense potential in other liquid-
phase catalytic transformations, which are currently the focus of
ongoing studies in our lab.
[18] S. Zhu, Y. Zhu, S. Hao, L. Chen, B. Zhang, Y. Li, Aqueous-phase hydrogenolysis of
glycerol to 1,3-propanediol over Pt-H4SiW12O40/SiO2, Catal. Lett. 142 (2011)
267–274.
[19] R. Arundhathi, T. Mizugaki, T. Mitsudome, K. Jitsukawa, K. Kaneda, Highly
selective hydrogenolysis of glycerol to 1,3-propanediol over
a boehmite-
supported platinum/tungsten catalyst, ChemSusChem 6 (2013) 1345–1347.
[20] S. Zhu, Y. Qiu, Y. Zhu, S. Hao, H. Zheng, Y. Li, Hydrogenolysis of glycerol to 1,3-
propanediol over bifunctional catalysts containing Pt and heteropolyacids,
Catal. Today 212 (2013) 120–126.
[21] S.H. Zhu, X.Q. Gao, Y.L. Zhu, Y.F. Zhu, X.M. Xiang, C.X. Hu, Y.W. Li, Alkaline
metals modified Pt-H4SiW12O40/ZrO2 catalysts for the selective hydrogenolysis
of glycerol to 1,3-propanediol, Appl. Catal. B Environ. 140 (2013) 60–67.
[22] C. Deng, X. Duan, J. Zhou, D. Chen, X. Zhou, W. Yuan, Size effects of Pt-Re
bimetallic catalysts for glycerol hydrogenolysis, Catal. Today 234 (2014) 208–
214.
[23] S. Zhu, X. Gao, Y. Zhu, J. Cui, H. Zheng, Y. Li, SiO2 promoted Pt/WOx/ZrO2
catalysts for the selective hydrogenolysis of glycerol to 1,3-propanediol, Appl.
Catal. B Environ. 158–159 (2014) 391–399.
Acknowledgments
[24] J. Wang, X. Zhao, N. Lei, L. Li, L. Zhang, S. Xu, S. Miao, X. Pan, A. Wang, T. Zhang,
Hydrogenolysis of glycerol to 1,3-propanediol under low hydrogen pressure
over WOx-supported single/pseudo-single atom Pt catalyst, ChemSusChem 9
(2016) 784–790.
[25] Y. Fan, S. Cheng, H. Wang, J. Tian, S. Xie, Y. Pei, M. Qiao, B. Zong, Pt–WOx on
monoclinic or tetrahedral ZrO2: crystal phase effect of zirconia on glycerol
hydrogenolysis to 1,3-propanediol, Appl. Catal. B Environ. 217 (2017) 331–
341.
We are grateful to the support from the National Natural
Science Foundation of China (91545108, 21473035, 21773033,
21503187, 91645201), SINOPEC (X514005), Science and Technol-
ogy Commission of Shanghai Municipality (16ZR1440400), China
Postdoctoral Science Foundation (2015M580288), Nanjing Tech
University and SICAM Fellowship from Jiangsu National Synergetic
Innovation Center for Advanced Materials.
[26] Y. Fan, S. Cheng, H. Wang, D. Ye, S. Xie, Y. Pei, H. Hu, Z. Li, W. Hua, M. Qiao, B.
Zong, Nanoparticulate Pt on mesoporous SBA-15 doped with extremely low

350
X. Wan et al. / Journal of Catalysis 375 (2019) 339–350
amount of W as highly selective catalyst for glycerol hydrogenolysis to 1,3-
propanediol, Green Chem. 19 (2017) 2174–2183.
[27] S. García-Fernández, I. Gandarias, J. Requies, F. Soulimani, P.L. Arias, B.M.
Weckhuysen, The role of tungsten oxide in the selective hydrogenolysis of
glycerol to 1,3-propanediol over Pt/WOx/Al2O3, Appl. Catal. B Environ. 204
(2017) 260–272.
[47] H. Zhu, C. Liang, W. Yan, S.H. Overbury, S. Dai, Preparation of highly active
silica-supported Au catalysts for CO oxidation by a solution-based technique, J.
Phys. Chem. B 110 (2006) 10842–10848.
[48] Y. Lou, P. He, L. Zhao, W. Cheng, H. Song, Olefin upgrading over Ir/ZSM-5
catalysts under methane environment, Appl. Catal. B Environ. 201 (2017) 278–
289.
[28] X. Zhao, J. Wang, M. Yang, N. Lei, L. Li, B. Hou, S. Miao, X. Pan, A. Wang, T.
Zhang, Selective hydrogenolysis of glycerol to 1,3-propanediol: manipulating
the frustrated Lewis pairs by Introducing gold to Pt/WOx, ChemSusChem 10
(2017) 819–824.
[49] M. Niederberger, M.H. Bartl, G.D. Stucky, Benzyl alcohol and titanium
tetrachloride:
a versatile reaction system for the nonaqueous and low-
temperature preparation of crystalline and luminescent titania
nanoparticles, Chem. Mater. 14 (2002) 4364–4370.
[29] W. Zhou, J. Luo, Y. Wang, J.F. Liu, Y.J. Zhao, S.P. Wang, X.B. Ma, WOx domain
size, acid properties and mechanistic aspects of glycerol hydrogenolysis over
Pt/WOx/ZrO2, Appl. Catal. B Environ. 242 (2019) 410–421.
[30] X. Bai, W.M.H. Sachtler, Methylcyclopentane conversion catalysis by zeolite
encaged palladium clusters and palladium-proton adducts, J. Catal. 129 (1991)
121–129.
[50] M. Niederberger, M.H. Bartl, G.D. Stucky, Benzyl alcohol and transition metal
chlorides as
a versatile reaction system for the nonaqueous and low-
temperature synthesis of crystalline nano-objects with controlled
dimensionality, J. Am. Chem. Soc. 124 (2002) 13642–13643.
[51] K. Chen, M. Tamura, Z. Yuan, Y. Nakagawa, K. Tomishige, One-pot conversion of
sugar and sugar polyols to n-alkanes without C-C dissociation over the Ir-
ReOx/SiO2 catalyst combined with H-ZSM-5, ChemSusChem 6 (2013) 613–621.
[31] T.J. McCarthy, G.D. Lei, W.M.H. Sachtler, Methylcyclopentane conversion
ˇ
catalysis over zeolite-Y encaged Rhodium:
adduct model, J. Catal. 159 (1996) 90–98.
a
test for the metal–proton
[52] J. Koo, N. Jiang, S. Saravanamurugan, M. Bejblováb, Z. Musilováb, J. Cejka, S.-E.
Park, Direct synthesis of carbon-templating mesoporous ZSM-5 using
microwave heating, J. Catal. 276 (2010) 327–334.
[32] E.A.I. Shor, V.A. Nasluzov, A.M. Shor, G.N. Vayssilov, N.R. sch, Reverse hydrogen
spillover onto zeolite-supported metal clusters: an embedded cluster density
functional study of models M6 (M) Rh, Ir, or Au), J. Phys. Chem. C 111 (2007)
12340–12351.
[33] F. Li, P. Yu, M. Hartl, L.L. Daemen, J. Eckert, B.C. Gates, Zeolite-encaged iridium
clusters with hydride ligands: characterization by extended X-ray absorption
fine structure, NMR, and inelastic neutron scattering vibrational
spectroscopies, Z. Phys. Chem. 220 (2006) 1553–1568.
[53] G.L. Woolery, G.H. Kuehl, H.C. Timken, A.W. Chester, J.C. Vartuli, On the nature
of framework Brönsted and Lewis acid sites in ZSM-5, Zeolites 19 (1997) 288–
296.
ˇ
[54] D.P. Serrano, R.A. Garca, G. Vicente, M. Linares, D. Prochzkov, J. Cejka, Acidic
and catalytic properties of hierarchical zeolites and hybrid ordered
mesoporous materials assembled from MFI protozeolitic units, J. Catal. 279
(2011) 366–380.
[34] T. Pauporté, F. Andolfatto, R. Durand, Some electrocatalytic properties of
anodic iridium oxide nanoparticles in acidic solution, Electrochim. Acta 45
(1999) 431–439.
[35] G. Dima, A. De Vooys, M. Koper, Electrocatalytic reduction of nitrate at low
concentration on coinage and transition-metal electrodes in acid solutions, J.
Electroanal. Chem. 554 (2003) 15–23.
[36] W.L. Kwong, C.C. Lee, A. Shchukarev, E. Björn, J. Messinger, High-performance
iron (III) oxide electrocatalyst for water oxidation in strongly acidic media, J.
Catal. 365 (2018) 29–35.
[37] G. Nahor, P. Hapiot, P. Neta, A. Harriman, Changes in the redox state of iridium
oxide clusters and their relation to catalytic water oxidation: radiolytic and
electrochemical studies, J. Phys. Chem. 95 (1991) 616–621.
[38] L.C. Seitz, C.F. Dickens, K. Nishio, Y. Hikita, J. Montoya, A. Doyle, C. Kirk, A.
Vojvodic, H.Y. Hwang, J.K. Norskov, A highly active and stable IrOx/SrIrO3
catalyst for the oxygen evolution reaction, Science 353 (2016) 1011–1014.
[39] A. Zecchina, S. Bordiga, C. Lamberti, G. Spoto, L. Carnelli, Low-temperature
Fourier transform infrared study of the interaction of CO with cations in alkali-
metal exchanged ZSM-5 zeolites, J. Phys. Chem. 98 (1994) 9577–9582.
[40] M. Yabuki, R. Takahashi, S. Sato, T. Sodesawa, K. Ogura, Silica–alumina
catalysts prepared in sol–gel process of TEOS with organic additives, Phys.
Chem. Chem. Phys. 4 (2002) 4830–4837.
[55] K. Cheng, L. Zhang, J. Kang, X. Peng, Q. Zhang, Y. Wang, Selective
transformation of syngas into gasoline-range hydrocarbons over mesoporous
H-ZSM-5-supported cobalt nanoparticles, Chem. Eur. J. 21 (2015) 1928–1937.
[56] P. Reyes, J. Fernández, G. Pecchi, Resistance to sulphur poisoning of alumina-
supported iridium catalysts in toluene hydrogenation and methylcyclohexane
dehydrogenation, J. Chem. Technol. Biotechnol. 73 (1998) 1–6.
[57] M. Chia, Y.J. Pagán-Torres, D. Hibbitts, Q. Tan, H.N. Pham, A.K. Datye, M.
Neurock, R.J. Davis, J.A. Dumesic, Selective hydrogenolysis of polyols and cyclic
ethers over bifunctional surface sites on rhodium–rhenium catalysts, J. Am.
Chem. Soc. 133 (2011) 12675–12689.
[58] T. Buntara, S. Noel, P.H. Phua, I. Melián-Cabrera, J.G. de Vries, H.J. Heeres, From
5-hydroxymethylfurfural (HMF) to polymer precursors: catalyst screening
studies on the conversion of 1, 2, 6-hexanetriol to 1, 6-hexanediol, Top. Catal.
55 (2012) 612–619.
[59] N. Ota, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige,
Hydrodeoxygenation of vicinal OH groups over heterogeneous rhenium
catalyst promoted by palladium and ceria support, Angew. Chem. Int. Ed. 54
(2015) 1897–1900.
[60] B. Xiao, M. Zheng, X. Li, J. Pang, R. Sun, H. Wang, X. Pang, A. Wang, X. Wang, T.
Zhang, Synthesis of 1, 6-hexanediol from HMF over double-layered catalysts of
Pd/SiO2+ Ir–ReOx/SiO2 in a fixed-bed reactor, Green Chem. 18 (2016) 2175–
2184.
[41] J. Ohyama, R. Kanao, A. Esaki, A. Satsuma, Conversion of 5-
hydroxymethylfurfural
rearrangement over supported Au nanoparticles, Chem. Commun. 50 (2014)
5633–5636.
to
a
cyclopentanone derivative
by ring
[61] T.M. Eggenhuisen, J.P. den Breejen, D. Verdoes, Fundamentals of melt
infiltration for the preparation of supported metal catalysts. The case of Co/
SiO2 for Fischer-Tropsch synthesis, J. Am. Chem. Soc. 132 (2010) 18318–18325.
[62] I. Gandarias, P.J. Miedziak, E. Nowicka, Selective oxidation of n-butanol using
gold-palladium supported nanoparticles under base-free conditions,
ChemSusChem 8 (2015) 473–480.
[63] Y. Seo, K. Cho, Y. Jung, Characterization of the surface acidity of MFI zeolite
nanosheets by 31P NMR of adsorbed phosphine oxides and catalytic cracking
of decalin, ACS Catal. 3 (2014) 713–720.
[64] L. Hu, Y. Song, A. Riisager, Zeolite and zeotype-catalysed transformations of
biofuranic compounds, Green Chem. 18 (2016) 5701–5735.
[65] S. Melson, F. Schüth, The influence of the external acidity of H-ZSM-5 on its
shape selective properties in the disproportionation of ethylbenzene, J. Catal.
170 (1997) 46–53.
[42] X.L. Du, L. He, S. Zhao, Y.M. Liu, Y. Cao, H.Y. He, K.N. Fan, Hydrogen-
Independent reductive transformation of carbohydrate biomass into
valerolactone and pyrrolidone derivatives with supported gold catalysts,
Angew. Chem. Int. Ed. 123 (2011) 7961–7965.
c-
[43] L. Yu, J. Yuan, Q. Zhang, Y.M. Liu, H.Y. He, K.N. Fan, Y. Cao, Propylene from
renewable resources: catalytic conversion of glycerol into propylene,
ChemSusChem 7 (2014) 743–747.
[44] N.M. Phadke, J. Van der Mynsbrugge, E. Mansoor, A.B. Getsoian, M. Head-
Gordon, A.T. Bell, Characterization of isolated Ga3+ cations in Ga/H-MFI
prepared by vapor-phase exchange of H-MFI zeolite with GaCl3, ACS Catal. 8
(2018) 6106–6126.
[45] T. Ishida, N. Kinoshita, H. Okatsu, T. Akita, T. Takei, M. Haruta, Influence of the
support and the size of gold clusters on catalytic activity for glucose oxidation,
Angew. Chem. Int. Ed. 47 (2008) 9265–9268.
[66] E. El-Malki, R.A. van Santen, W.M.H. Sachtler, Introduction of Zn, Ga, and Fe
into HZSM-5 cavities by sublimation: identification of acid sites, J. Phys. Chem.
B 103 (1999) 4611–4622.
[46] T. Ishida, M. Nagaoka, T. Akita, M. Haruta, Deposition of gold clusters on porous
coordination polymers by solid grinding and their catalytic activity in aerobic
oxidation of alcohols, Chem. Eur. J. 14 (2008) 8456–8460.