C?O Hydrogenolysis of Tetrahydrofurfuryl Alcohol to 1,5-Pentanediol Over Bi-functional Nickel-Tungsten Catalysts

Article abstract of DOI:10.1002/cctc.201800783

In this study, we report a series of bimetallic Ni?WO_x catalyst for the ring-opening of THFA into 15PDO. The structure-performance relationship of the catalysts was discussed based on extensive characterization using techniques such as BET, H₂-TPR, NH₃-TPD, Pyr-IR, IPA-TPD-MS, XRD, XPS and EXAFS/XANES. The acidity measurements show that higher W density leads to the higher amount of acid density, which could be assigned to the creation of Lewis acid sites mainly on the surface of the calcined catalysts. H₂-TPR profiles of Ni?WO_x catalysts show that there is a strong interaction between Ni and W species, enhancing the reducibility of WO_x. XRD measurements of calcined Ni?WO_x catalysts reveal that the dispersion of Ni particles is enhanced after addition of WO_x species. After reduction, different peaks corresponding to metallic Ni and WO_3?x are identified for 10Ni?WO_x catalysts, as well as new peak assigned to Ni?W intermetallic phase on 10Ni?30WO_x catalyst. The formation of Ni?W intermetallic phase was further proved using XPS and EXAFS studies. THFA hydrogenolysis was also conducted under aqueous-phase conditions over Ni?WO_x catalysts, yielding up to 47 % selectivity to 15PDO, along with a highest combined C₅ polyols (i. e., 15PDO and 125PTO) selectivity of approximately 64 %. However, the Ni?WO_x catalytic system suffers from deactivation process due to the hydrothermal dissolution of the active phase. Further investigation reveals the better stability of metallic tungsten and Ni?W intermetallic phase (Ni₄W) against leaching since their corresponding peaks in the XRD patterns of spent catalysts remains nearly unchanged. Finally, 1,4-dioxane as an organic solvent was employed in THFA hydrogenolysis reaction, resulting in different product distribution, with a THP yield of around 54 %. The catalyst crystalline structure is preserved because of very low Ni and W leaching when 1,4-dioxane is used as solvent.

Full text of DOI:10.1002/cctc.201800783

Full Papers
DOI: 10.1002/cctc.201800783
1
2
3
4
5
6
7
8
9
CꢀO Hydrogenolysis of Tetrahydrofurfuryl Alcohol to
1,5-Pentanediol Over Bi-functional Nickel-Tungsten
Catalysts
Elmira Soghrati,[a, b] Chee Kok Poh,[a] Yonghua Du,[a] Feng Gao,[a] Sibudjing Kawi,*[b] and
Armando Borgna*[a]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
In this study, we report a series of bimetallic NiꢀWOx catalyst for
the ring-opening of THFA into 15PDO. The structure-perform-
ance relationship of the catalysts was discussed based on
extensive characterization using techniques such as BET, H2-TPR,
NH3-TPD, Pyr-IR, IPA-TPD-MS, XRD, XPS and EXAFS/XANES. The
acidity measurements show that higher W density leads to the
higher amount of acid density, which could be assigned to the
creation of Lewis acid sites mainly on the surface of the calcined
catalysts. H2-TPR profiles of NiꢀWOx catalysts show that there is
a strong interaction between Ni and W species, enhancing the
reducibility of WOx. XRD measurements of calcined NiꢀWOx
catalysts reveal that the dispersion of Ni particles is enhanced
after addition of WOx species. After reduction, different peaks
corresponding to metallic Ni and WO3ꢀx are identified for
10NiꢀWOx catalysts, as well as new peak assigned to NiꢀW
intermetallic phase on 10Niꢀ30WOx catalyst. The formation of
NiꢀW intermetallic phase was further proved using XPS and
EXAFS studies. THFA hydrogenolysis was also conducted under
aqueous-phase conditions over NiꢀWOx catalysts, yielding up to
47% selectivity to 15PDO, along with a highest combined C5
polyols (i.e., 15PDO and 125PTO) selectivity of approximately
64%. However, the NiꢀWOx catalytic system suffers from
deactivation process due to the hydrothermal dissolution of the
active phase. Further investigation reveals the better stability of
metallic tungsten and NiꢀW intermetallic phase (Ni4W) against
leaching since their corresponding peaks in the XRD patterns of
spent catalysts remains nearly unchanged. Finally, 1,4-dioxane
as an organic solvent was employed in THFA hydrogenolysis
reaction, resulting in different product distribution, with a THP
yield of around 54%. The catalyst crystalline structure is
preserved because of very low Ni and W leaching when 1,4-
dioxane is used as solvent.
31 Introduction
reaction pathway requires typically a bifunctional catalyst,
consisting of a reducible metal and an acidic center to facilitate
hydrogenation and hydrogenolysis processes, respectively.
Recently, the catalytic system comprising of noble metals (Pt,
Rh, and Ir) in combination with oxides of group 6 or 7 metals
such as molybdenum, rhenium, and tungsten has gained much
attention in the catalytic conversion of glycerol,[3] tetrahydro-
furfuryl alcohol (THFA),[4] and tetrahydropyran-2-methanol
(THPM)[5] to their corresponding a,w-diols.[6] In particular,
Tomishige and co-workers have reported a 94% yield of 1,5-
pentanediol (15PDO) from THFA using a RhꢀReOx catalyst.[5a]
32
33 The replacement of petroleum with biomass as a renewable
34 resource for the production of fuel and chemicals has been one
35 of the most important research fields of green chemistry.[1]
36 While fossil fuel feedstocks which mostly lack functionality,
37 biomass feedstocks are highly-functionalized due to the high
38 oxygen content. Therefore, it is evident that deoxygenation
39 processes are needed to convert biomass feedstocks or plat-
40 form chemicals into target molecules to replace petroleum-
41 based fuels and chemicals. Of various deoxygenation processes,
42 CꢀO bond hydrogenolysis is a critical reaction in upgrading
43 biomass-derived oxygenates such as polyols and cyclic ethers
44 into high-value chemicals.[2] One of the major challenges is the
45 selection and design of an efficient catalyst. Hydrogenolysis
46
[7]
[8]
Moreover, PdꢀIrꢀReOx/SiO2 and RhꢀIrꢀReOx/SiO2 catalysts
were employed in one-pot conversion of furfural to 15PDO,
yielding up to 71.4% and 78.2%, respectively.[9] Despite good
selectivity into 15PDO, this catalytic approach suffers from the
high cost of catalyst and relatively low efficiency. Hence, it is
highly desirable to develop an alternative catalyst system based
on non-noble metals.
47
48
49
50
51
52
53
54
55
56
57
[a] E. Soghrati, Dr. C. Kok Poh, Dr. Y. Du, Dr. F. Gao, Dr. A. Borgna
Institute of Chemical and Engineering Sciences (ICES)
Agency for Science, Technology and Research (A*STAR)
1 Pesek Road
Jurong Island 627833 (Singapore)
E-mail: Armando_ABorgna@ices.a-star.edu.sg
[b] E. Soghrati, Prof. S. Kawi
A pioneer work developed a multiple approach to synthe-
size 15PDO from THFA via a dehydration-hydration-hydro-
genation pathway, with an overall yield of ca. 70%.[10] Very
recently, this process has been modified by Dumesic et al. for
the production of 15PDO from biomass-derived THFA and
furfural over inexpensive catalysts, reaching a significantly
higher overall yield of ca. 90%.[11] Among non-noble metals, Ni-
based catalysts are widely used in deoxygenation of various
bio-derived oxygenates. Resasco et al. reported the CꢀO bond
Department of Chemical and Biomolecular Engineering
National University of Singapore
4 Engineering Drive 4
Singapore 117585 (Singapore)
E-mail: chekawis@nus.edu.sg
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/cctc.201800783
ChemCatChem 2018, 10, 1–14
1
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
cleavage of tetrahydrofuran ring over Ni catalysts.[12] Later,
Huber et al. measured the initial activity of alumina-supported
monometallic catalysts (Pd, Pt, Ru, Rh, Ni, and Co), showing that
only the Ni catalyst can produce 15PDO in aqueous-phase
hydrogenolysis of THFA at temperatures above 2508C.[13]
Subsequently, we have reported that Ni-based catalysts in THFA
hydrogenolysis under aqueous-phase conditions were capable
of CꢀO bond cleavage of THFA at the more sterically hindered
secondary CꢀO bond, yielding mostly 15PDO.[14] On the
metal catalysts, tungsten is earth-abundant and relatively
inexpensive, being an attractive candidate in view of industrial
catalytic applications. It was widely reported that various
tungsten compounds, such as Tungsten carbide (WCx),[17]
tungstic acid (H2WO4),[18] W,[19] WO3,[20] and WOx[3c,d,21] by
themselves or pairing with hydrogenation metals such as Ni, Pt,
and Ru, are effective in promoting the catalytic activity and
selectivity in various hydrogenolysis reactions, particularly in
cellulose conversion to ethylene glycol (EG).[22] To understand
the performance of these tungsten species, Zhang et al.[18a]
conducted extensive characterization of the catalyst before and
immediately after the reaction, showing that the dissolved
tungsten bronze (HxWO3) was the catalytically active phase in
CꢀC bond cleavage of cellulose, through a homogeneous
reaction pathway. Upon cooling down to room temperature
and exposure to air, the dissolved tungsten bronze was
precipitated and transformed into tungstic acid and finally WO3,
which are not soluble in water. Therefore, this catalyst
combination is particularly attractive due to its advantages of
both heterogeneous and homogeneous catalysts through a
temperature-controlled phase transfer of the catalyst.[23]
10 contrary, the CꢀO bond cleavage of the less sterically hindered
11 side of THFA molecule into 1,2-pentanediol (12PDO) was not
12 favored over Ni-based catalysts. A similar result was obtained
13 over NiꢀY2O3 catalyst in the hydrogenolysis of
14
furfuryl alcohol, selectively producing 15PDO with THFA as
15 reaction intermediate.[15] We have also shown that THFA hydro-
16 genolysis proceeded in a parallel pathway through protonation
17 and SN2-type reaction with water (i.e., solvent), yielding 1,2,5-
18 pentanetriol (125PTO) as a new reaction intermediate of THFA
19 conversion under aqueous-phase condition.[14a] However, the
20 initial activity of THFA hydrogenolysis over Ni/SiO2 was
21 considerably lower as compared to the initial hydrogenolysis
22 rates over the bifunctional Ni/HZSM-5 catalysts. It is important
23 to mention that in the presence of acid sites on Ni/HZSM-5
24 catalysts, the C5 polyols underwent ring closure pathways
25 towards the formation of tetrahydropyran (THP), as depicted in
In this contribution, we first examined the promoting effect
of WOx, using a series of NiꢀWOx catalysts, on the catalytic
performances of the ring-opening of THFA into 15PDO. The
structure-performance relationship of the prepared catalysts is
discussed in detail. Subsequently, we explored the recyclability
of the catalysts as well as different approaches to deal with
catalyst deactivation in hydrogenolysis reaction over NiꢀWOx
catalysts.
26 Scheme 1.
27
28
29
30
31
32
33
34
35
36
Results and Discussion
Characterization Results
The main physicochemical properties of the NiꢀWOx/SiO2
catalysts along with their corresponding support and mono-
metallic Ni and W catalysts are summarized in Table 1. The BET
surface area of SiO2 is notably high (i.e., 792.9 m2gꢀ1). Upon
addition of Ni with increasing tungsten oxide loading, the BET
surface areas continuously decrease up to around 300 m2 gꢀ1
for 10Niꢀ30WOx catalyst. N2 adsorption-desorption isotherms at
ꢀ1968C for SiO2 supported NiꢀWOx catalysts are shown in
Figure S1 (Supplementary Information). All samples show Type-
37
38
Scheme 1. The reaction pathway for CꢀO hydrogenolysis of THFA over Ni-
based catalysts under aqueous-phase conditions at 2508C.
39
40
41
To enhance catalytic activity and maintain the selectivity
42 towards 15PDO, Ni-based bimetallic catalysts could be consid-
43 ered since often show different chemical and electronic proper-
44 ties than their parent metals.[16] Among the reported non-noble
45
46
47
48
49
50
51
52
53
54
55
56
57
Table 1. Physicochemical properties of NiꢀWOx/SiO2 Catalysts.
[a]
Catalyst
S BET
Loading[b] [wt%]
Ni
WO3 Surface density
[Wnm2]
Total Acidity[c]
Brønsted/Lewis acid
site ratio [d]
[m2 gꢀ1
]
W
[mmolgꢀ1
]
SiO2
10Ni
792.9
623.6
556.9
487.8
353.8
303
–
–
–
–
–
ND
134
–
–
–
–
0.11
0.12
ND
ND
9.9
9.7
9.6
9.8
9.7
–
10Niꢀ2.5WOx
10Niꢀ7.5WOx
10Niꢀ15WOx
10Niꢀ30WOx
15WOx
2.4
7.4
14.6
29.4
14.8
14.7
0.11
0.33
0.66
1.32
0.66
0.66
171.4
317.5
872.6
1038.5
451.9
581.7
572.7
449.1
5Niꢀ15WOx
4.8
[a] Determined by BET method; [b] Determined by XRF; [c] Determined by NH3-TPD; [d] Determined by IPA-TPD. ND: not determined.
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
2
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
IV isotherm according to the IUPAC classification, with an H1
hysteresis loop, which is characteristic of mesoporous materi-
als.[24] The shape of the isotherm of SiO2 support is preserved
after loading of Ni and W, indicating that loading of nickel and
tungsten precursors and their subsequent calcination do not
change significantly the mesoporous nature of silica support.
Besides, the hysteresis loops are shifted to lower relative
pressures with the loading of metals, as compared to the bare
silica support. Also, it should be noted that the hysteresis loop
to the previous studies[3f,25–26], below monolayer coverage
(tungsten surface density !4–5 Wnmꢀ2), only Lewis acidity
attributed to monotungstate WOx species are present on the
surface the catalyst. At higher tungsten surface density (i.e.,
<4–5 Wnmꢀ2), polytungstate surface WOx species start to be
formed, yielding Brønsted acid sites. For monolayer coverages
higher than 5 Wnmꢀ2, crystalline WO3 nanoparticles are also
formed on the surface. Finally, for the coverage higher than
10 Wnmꢀ2, bulk-like WO3 crystallites are the dominant species
on the surface of the catalyst. Therefore, it can be concluded
that tungsten surface density plays a central role in the
formation of the Brønsted acid sites on the surface of the W-
containing catalyst. To further explore the presence of Brønsted
acid sites, 10Niꢀ15WOx and 10Niꢀ30WOx catalysts were ana-
lyzed by IPA-TPD-MS method and the results are reported in
Table 1. Gorte et al.[27] reported that Brønsted acid sites
protonate amines to form alkylammonium ions which sub-
sequently decompose to produce ammonia and propylene
above 3008C. Moreover, unreacted IPA is desorbed from both
Lewis and Brønsted acid sites, resulting in a broad peak starting
around 908C. Around 756 mmolgꢀ1 (Lewis)+85 mmolgꢀ1
(Brønsted) and 850 mmolgꢀ1 (Lewis)+98 mmolgꢀ1 (Brønsted) of
desorbed IPA and propylene are recorded for 10Niꢀ15WOx and
10Niꢀ30WOx, respectively. Interestingly, very small amount of
Brønsted acid sites (B:L molar ratio of 0.11–0.12, Table 1) is
detected for both 10NiꢀWOx catalysts with the highest W
surface densities, which could not be observed by Pyr-IR
measurements.
10 shows a gradual decrease as the W loading is increased.
11
Acid characteristics of NiꢀWOx/SiO2 catalysts were measured
12 by TPD of NH3 and IPA, and FTIR of adsorbed pyridine. The total
13 acidity and the ratio of Brønsted to Lewis acid sites were
14 quantified and are listed in Table 1. Several studies have
15 investigated the effect of tungsten oxide by determining the
16 types and strength of acid sites as a function of tungsten oxide
17 surface density.[25] The tungsten oxide surface density (Wnmꢀ2)
18 can be derived from Equation (1):
19
20
21
22
23
24
WO3 loading ꢁ 6:023 ꢁ 1023
ð1Þ
WO3 surface density ¼
231:84 ꢁ BET surface area ꢁ 1020
where WO3 loading is in wt%, and BET surface area is in m2 gꢀ1.
The tungsten oxide surface density was calculated by using
25 the initial BET surface area of the silica support since it can be
26 preserved after relatively mild calcination temperature. Besides,
27 it is known from previous reports that surface nickel and
28 tungsten species do not induce any sintering of the oxide
29 supports.[25] Thus, the calculated tungsten oxide surface density
30 are also listed in Table 1. As shown in this table, 10Ni catalyst
Additionally, Diffuse reflectance Fourier transform infrared
spectroscopy (DRIFTS) was employed to investigate the effect
of H2 reduction on the acid properties of 10Niꢀ30WOx catalyst
and the result is shown in Figure S3 (Supplementary Informa-
tion). The formation of Brønsted acid sites during H2 treatment
has been directly observed by pyridine adsorbed DRIFTS
measurements. It is reported in the literature that the density of
Brønsted acid sites increases during contact with H2, while the
density of Lewis acid sites concurrently decreases, suggesting
that Lewis acid sites are converted to Brønsted acid sites during
H2 reduction. Specifically, WOx containing samples can be
31 shows
a
small amount of adsorbed ammonia (i.e.,
32 134 mmolgꢀ1), corresponding to the weak acid sites. Compared
33 to the 10Ni catalyst, the amount of acid sites continuously
34 increased up to 1038.5 mmolgꢀ1 as the W loading increases,
35 which could be ascribed to the generation of a wide range of
36 medium acid sites on the surface of the catalysts.
37
The acid properties of calcined NiꢀWOx catalysts were also
38 investigated by FTIR of adsorbed pyridine. As shown in
39 Figure S2 (Supplementary Information), the IR bands at around
40 1445, 1595 and 1612 cmꢀ1 are assigned to the adsorption of
41 pyridine on Lewis acid sites. In the case of 10Ni catalyst, it only
42 shows two Lewis acid-related peaks at around 1445 and
43 1595 cmꢀ1 after desorption at 1008C (Figure S2-a). However,
44 after increasing the desorption temperature to 2008C, these
45 peaks corresponding to Lewis acid sites almost disappear.
[26b,c]
transformed into Hd+(WOx)ndꢀ species in the presence of H2
H2-TPR measurements were performed to explore the
reducibility of the catalyst, to quantify the H2 consumption, and
to determine the NiꢀW interaction. The H2-TPR profiles of the
fresh monometallic 10Ni, 15WOx catalysts, along with the
bimetallic 10NiꢀWOx catalysts are shown in Figure 1. According
to the previous studies[20b,28], pure WO3 normally exhibits three
major reduction peaks, corresponding to the reduction of
WO3ꢀWO2.72 (~7508C), WO2.72ꢀWO2 (~8008C), and WO2ꢀW
(higher than 10008C). As shown in Figure 1(a), the above–
mentioned reduction peaks for the 15WOx catalyst is shifted
slightly to lower temperatures due to the dispersion effect of
the support (i.e. silica).[28a] 10Ni catalyst (Figure 1(b)) shows a
hydrogen consumption profile, starting at approximately 3008C
with the main peak at around 370–3808C and a shoulder peak
at a higher temperature (around 5008C). The first peak
corresponds to the reduction of free NiO species weakly
interacting with the silica support, while the latter shoulder is
46
The concentration of Lewis acid sites gradually increases on
47 the surface of the 10NiꢀWOx catalysts by increasing the W
48 loading. The bands at around 1545 and 1637 cmꢀ1 are
49 attributed to Brønsted acid sites, while the band near
50 1490 cmꢀ1 is due to the presence of both Brønsted and Lewis
51 acid sites. Interestingly, that is no evidence of the characteristic
52 IR bands attributed to Brønsted acid sites for calcined 10Ni and
53 10NiꢀWOx catalysts. Nevertheless, the intensity of peak around
54 1490 cmꢀ1 increased with increasing tungsten density on the
55 surface of the catalyst. These results indicate that the acidity
56 can mainly be attributed to Lewis acid sites on the surface of
57 catalysts at tungsten surface densities lower than 1.3. According
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
3
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Figure 1. H2ꢀTPR profiles of NiꢀWOx/SiO2 catalysts: (a) 15WOx, (b) 10Ni, (c)
10Niꢀ2.5WOx, (d) 10Niꢀ7.5WOx, (e) 10Niꢀ15WOx, and (f) 10Niꢀ30WOx.
17 therefore assigned to NiO species with stronger metal-support
18 interaction. After addition of WOx to 10Ni, a few broad peaks
19 are observed for the bimetallic catalysts, indicating the co-
20 reduction of both metal oxide species. Obviously, for all
21 10NiꢀWOx catalysts shown in Figure 1(c)–(f), the reduction
22 peaks of WO3 species significantly shifted to lower temper-
23 atures as compared to 15WOx catalyst, roughly starting at
24 6008C. This indicates the existence of a strong electronic
25 interaction between NiO and WOx, enhancing the reducibility of
26 WOx species.[28a,29] In contrast, the opposite trend is observed
27 after addition of WOx up to 7.5 wt% into 10Ni catalyst, shifting
28 the NiO reduction peaks to slightly higher temperatures as
29 compared to the monometallic 10Ni catalyst. When higher WOx
30 loadings are used (i.e., 15 and 30 wt%), the reduction is
31 initiated at even lower temperature (around 2708C). In the case
32 of 10Niꢀ30WOx catalyst, the reduction peaks attributed to WOx
33 species are shifted again to higher temperatures, as compared
34 to the catalysts with lower W loadings.
Figure 2. a) (top) XRD patterns of calcined NiꢀWOx/SiO2 catalysts: (a) 10Ni,
(b) 10Niꢀ2.5WOx, (c) 10Niꢀ7.5WOx, (d) 10Niꢀ15WOx, and (e) 10Niꢀ30WOx. b)
(bottom) XRD patterns of reduced NiꢀWOx/SiO2 catalysts: (a) 10Ni, (b)
10Niꢀ2.5WOx, (c) 10Niꢀ7.5WOx, (d) 10Niꢀ15WOx, and (e) 10Niꢀ30WOx.
resulting in variation of crystalline orientation of WꢀO in WO3
structure.[28a] This can also facilitate the reduction of WO3
species due to the strong electronic interaction between NiO
and WOx, which is consistent with H2-TPR results.
35
XRD measurements were carried out to understand the
After being reduced with H2 at 5008C for 2 h, three other
peaks around 448, 528, and 768 for Ni (111), Ni (200), and Ni
(220) planes are detected (Figure 2-b). It should be noted that
the profiles of reduced 10Ni, 10Niꢀ2.5WOx and 10Niꢀ7.5WOx
catalysts still show the peaks assigned to NiO, indicating the
incomplete transformation of NiO into metallic Ni after
reduction in H2. This might be due to the passivation step with
1%O2 and exposure to air before running the XRD measure-
ments. Furthermore, broader Ni peaks are observed as the
amount of W is increased, indicating that tungsten oxide might
enhance the dispersion of Ni species on the catalyst surface.[28a]
This result is further supported by TEM image of 10Niꢀ30WOx
catalyst as reported in Figure S4. A mean Ni particle size of
around 5.5 nm is observed for 10Niꢀ30WOx catalyst, whereas
the mean particle size derived from XRD is around 13 nm for
the unprompted 10Ni catalyst. More importantly, when W
loading increases to 30 wt%, the main diffraction peaks of Ni
are slightly shifted to lower angles, which might be due to the
formation of a NiꢀW intermetallic phase (Ni4W) with a broad
peak at around 43.68. According to previous reports, Ni4W and
Ni17W3 intermetallic phases were reported in NiꢀW/SBA-15[30]
and NiꢀW/MCM-41[31] catalysts, respectively. However, it is
challenging to distinguish between Ni and NiꢀW intermetallic
36 crystalline structure of 10NiꢀWOx catalysts, and the results for
37 calcined and reduced catalysts are displayed in Figure 2-a and
38 2-b, respectively. The XRD pattern of calcined Ni/SiO2 catalyst
39 shows a broad hump tailing at around 238, which is attributed
40 to amorphous silica. The strong reflections of the (111), (200),
41 (220), (222) and (311) planes at 2q=378, 438, 62.48, 74.98 and
42 78.88, attributed to the NiO crystallite phase, are clearly
43 observed for the calcined 10Ni catalyst. For the 10NiꢀWOx
44 catalysts, these peaks are slightly shifted to higher angles with
45 lower intensitites as the amount of W is increased, as shown in
46 Figure 2-a. Interestingly, diffraction peaks of NiO almost
47 disappeared for 10Niꢀ30WOx catalyst, excepted the most
48 intense one at around 43.98. The characteristic peaks of WO3
49 are not observed for bimetallic catalysts, excepting 10Niꢀ30WOx
50 catalyst, implying that tungsten oxide is highly-dispersed on
51 the support surface. It is interesting to note that weak and
52 broad diffraction peaks at 2q=24.38, 28.28, 32.68, 38.48, and
53 47.58 can correspond to both WO3 and oxygen-deficient
54 tungsten oxide (WO3ꢀx), which are only observed for the
55 calcined 10Niꢀ30WOx catalyst. Previous studies also reported
56 the peak shift as compared to the pure WO3. This result
57 suggests that NiO can be incorporated into WO3 lattice,
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
4
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
phases by only using XRD. Besides, peaks corresponding to the
crystallite WO2 appear at 25.28, 36.78, 52,78, and 59.68 for
reduced 10Niꢀ30WOx catalyst.
W6+4f7/2 and W5+4f7/2 species, respectively.[28a,32] Besides, the
broad peak around 40 eV corresponds to W6+3p3/2. In contrast
to the Ni2p trend, with increasing W loading, both W4f
doublets are shifted to lower binding energies (i.e. 34.8 and
33.3, respectively). A further shift is observed for 10Niꢀ30WOx
catalyst. It is important to stress that the doublets observed in
the spectrum of 10Niꢀ30WOx catalyst are very close to the
typical binding energies reported in the literature for oxygen-
deficient tungsten oxide (WO3ꢀx) species or W5+ species.[32]
These results suggest that nickel oxide donated electrons to
the tungsten oxide species, making the nickel electron-deficient
and the tungsten electron-enriched. This result is consistent
with the H2ꢀTPR and XRD findings, suggesting that a possible
electronic interaction exists between Ni and W.
The surface chemical states of nickel and tungsten on the
calcined catalysts were analyzed by XPS measurements and are
shown in Figure 3-a and 3-b, respectively. The XPS parameters
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
The Ni2p and W4f XPS spectra of in-situ reduced
10Niꢀ30WOx catalysts are presented in Figure 4-a and 4-b,
Figure 3. a) (top) Ni2p XPS spectra of calcined NiꢀWOx/SiO2 catalysts: (a)
10Ni, (b) 10Niꢀ2.5WOx, (c) 10Niꢀ7.5WOx, (d) 10Niꢀ15WOx, and (e)
10Niꢀ30WOx b) (bottom) W4f XPS spectra of calcined NiꢀWOx/SiO2 catalysts:
(a) 10Niꢀ2.5WOx, (b) 10Niꢀ7.5WOx, (c) 10Niꢀ15WOx, and (d) 10Niꢀ30WOx.
44 derived from the Ni2p, and W4f peaks in the calcined 10NiꢀWOx
45 catalysts are also listed in Table S1 (see Supplementary
46 Information). Obviously, the incorporation of tungsten to nickel
47 catalyst has a significant impact on the surface chemical
48 environment of the nickel catalyst. As shown in the Ni2p
49 spectra of calcined catalysts, tungsten incorporation causes a
50 slight increase in binding energy to 855.8 eV as compared to
51 the NiO binding energy of Ni2p3/2 at 855.5 eV. With further
52 addition of W to 30 wt%, the Ni2p3/2 signal is shifted from
53 855.8 eV to 856.1 eV.[28a,31] This could be as a result of electron
54 transfer from Ni2+ to W6+, facilitating the formation of Ni3+ and
Figure 4. a) (top) Ni2p XPS spectra of in-situ reduced catalysts: (a) 10Ni, and
(b) 10Niꢀ30WOx. b) (bottom) W4f XPS spectrum of in-situ reduced
10Niꢀ30WOx catalyst.
respectively. For 10Ni catalyst after in-situ reduction step
(Figure 4-a), a new doublet peak, which is attributed to metallic
Ni appears at a binding energy of around 853.7 eV (Ni0 2p3/2).
Similarly, the same signal is identified for the reduced
10Niꢀ30WOx catalyst at a slightly higher binding energy (i.e.,
854.2 eV) as compared to the monometallic 10Ni catalyst. This
might be due to the formation of NiꢀW intermetallic phase,
55 W5+
.
56
Figure 3-b presents the W4f spectra of 10NiꢀWOx catalysts.
57 Two W4f doublets are detected at 35.4 and 34 eV, attributed to
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
5
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
which is suggested by the XRD profiles of the reduced
10Niꢀ30WOx catalyst. It is suggested that the electron transfer
from Ni to W can enhance the NiꢀW interaction, stabilizing Ni
in the form of NiꢀW intermetallic phase.[31] As shown in W4f
XPS spectra of 10Niꢀ30WOx catalyst after reduction (Figure 4-b),
two new overlapped W4f doublets with the binding energy of
32.5 eV and 30.9 eV are identified, assigned to tungsten in the
form of WO2 (W4+) and the metallic state (W0), respectively.
More importantly, the doublets corresponding to W6+4f7/2 and
tively. The detailed structural parameters obtained from fittings
are also listed in Table 2. Besides, the WꢀO distance for 30WOx
Table 2. Structural parameters obtained from EXAFS analysis.
2
2
˚
˚
Catalyst
10Ni
Edge Pair
CN
R [A]
s [A ]
Ni K
NiꢀNi 11.2
(ꢂ1.2)
2.48
0.0064
(ꢂ0.02)
(ꢂ0.0006)
10Niꢀ30WOx Ni K
NiꢀNi 7.4 (ꢂ0.8) 2.49
0.0089
(fixed)
NiꢀW 2.3 (ꢂ0.2) 2.55
(fixed)
W L3 WꢀO 2.8 (ꢂ0.3) 1.84
(ꢂ0.02)
2.52
(ꢂ0.03)
WꢀW 2.0 (ꢂ0.2) 2.65
(ꢂ0.03)
W L3 WꢀO 4.0 (ꢂ0.4) 1.84
(ꢂ0.02)
(ꢂ0.0009)
0.0071 (ꢂ0.007)
10 W5+4f7/2 species are shifted to higher binding energies, 35.4 eV
11 and 34.5 eV, respectively, as compared to the calcined cata-
12 lyst.[32] More detailed information regarding the effect of in-situ
13 reduction as compared to ex-situ reduction, along with XPS
14 parameters derived from Ni2p, and W4f peaks are also
15 presented in Figure S5 and Table S2 (Supplementary Informa-
16 tion), respectively. Besides, four peaks corresponding to W3p3/2
17 with different valence states are identified.
0.0114
(ꢂ0.0011)
WꢀNi 1.3
(ꢂ0.1)
0.0040
(ꢂ0.0004)
0.0040
(ꢂ0.0004)
30WOx
0.0048
(ꢂ0.0005)
18
The Fourier transform of the Extended X-ray Absorption
19 Fine Structure (EXAFS) spectrum taken at the WꢀL3 edge of the
20 in-situ reduced 10Niꢀ30WOx catalyst is shown in Figure 5-a. It is
(Figure S6 of Supplementary Information) and 10Niꢀ30WOx
˚
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
sample at 1.84 A probably corresponds to the typical WꢀO
distance in WO3ꢀx crystals, which is substantially larger than W6+
ꢀO distance.[32]
Fourier transform of the EXAFS spectra at the NiꢀK edge for
10Ni and 10Niꢀ30WOx catalysts, along with the fitting data of
the first coordination shell are shown in Figure 5-b. The first
˚
peak around 2.10 A corresponds to the first nearest neighbor
around the Ni atoms in the 10Ni catalyst. It can be seen that
the position and shape of the peak, corresponding to the first
shell surrounding Ni atoms in 10Niꢀ30WOx catalyst is very
similar to that of 10Ni catalyst, suggesting that it is difficult to
distinguish between NiꢀNi and NiꢀW first shell. Besides, a
reduction in the EXAFS signal, which might be due to a lower
Ni coordination is also observed for 10Niꢀ30WOx catalyst. This
behavior might be attributed to the NiꢀW intermetallic
formation, which was observed in the previous study on Ni4W
and Ni samples.[33] A previous study on NiꢀW alloy phase also
˚
˚
reported similar distances of 2.49 A and 2.55 A for NiꢀNi and
NiꢀW, respectively.[34] This could be due to the fact that the
Ni4W crystal structure is a combination of the body-centered-
˚
˚
tetrahedron tungsten (a: 5.73 A, c: 3.55 A) and face-centered-
[35]
˚
˚
tetrahedral nickel (a: 3.62 A, c: 3.55 A). It is also reported in
the literature that the coordination numbers of A and B atoms
in bimetallic alloys are in the same proportion as the bulk
concentrations of these elements in the sample. For example,
the coordination numbers of NiꢀW and WꢀNi should follow
CN(NiꢀW):CN(WꢀNi)=XW :XNi, in which X is the bulk concen-
trations. However, it should be noted that this relation is not
applicable for 10Niꢀ30WOx catalyst due to the presence of a
large amount of tungsten in the oxidized forms.
47
48
49
50
51
52
Figure 5. a) (top) Fourier Transform of the EXAFS spectrum of 10Niꢀ30WOx/
SiO2 catalyst at the W L3-edge (solid line) along with the fitting of the
experimental data (dashed line). b) (bottom) Fourier Transform of the EXAFS
spectra of 10Ni and 10Niꢀ30WOx catalysts at the Ni K-edge (solid lines) along
with the fitting of the experimental data (dashed lines).
To better understand the electronic structure of Ni and W
atoms in 10Niꢀ30WOx catalyst, the X-ray absorption near edge
structure (XANES) spectra at Ni K-edge are shown in Figure 6.
For 10Niꢀ30WOx catalyst, different intensity of the white line
peak is observed. Consequently, from the higher intensity of
the white line peak in 10Niꢀ30WOx catalyst cannot conclude
that 10Ni possesses higher reduction degree as compared to
53 evident that a considerable deviation in the environment
54 around the W atom in the 10Niꢀ30WOx catalyst is observed,
55 most likely due to the presence of NiꢀW intermetallic species.
56 Three peaks corresponding to WꢀO, WꢀNi and WꢀW are
˚
˚
˚
57 identified, with distances of 1.84 A, 2.52 A and 2.65 A, respec-
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
6
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
conditions tested, with ca. 80% combined selectivity of 15PDO,
125PTO and THP.
3
4
5
6
7
8
9
10
11
12
13
14
15
To get more insights into the product distribution at
comparable conversion level (~14%), we have also carried out
THFA hydrogenolysis using different amounts of 10NiꢀWOx
catalysts (Figure S7, see the Supplementary Information for
more details). For all 10NiꢀWOx catalyst, slightly higher
selectivities towards desired products are obtained as com-
pared to the 10Ni catalyst, suggesting the key role of Ni in
determining the product selectivity. However, much higher
reaction rates are achieved over the catalysts with higher
tungsten surface density (i.e., a higher amount of Lewis acidity).
Meanwhile, we have investigated the impact of Ni loading
on a 15WOx catalyst to have a better understanding of the
synergistic effect of Ni and WOx species in the ring-opening of
THFA (Table 3, entries 6,7). Approximately similar conversion
level was obtained for both 15WOx and 5Niꢀ15WOx (ca. 32%).
28% selectivity toward 15PDO is obtained over 15WOx catalyst,
indicating the catalytic activity of WOx species in the ring-
opening reaction. It was also reported in the previous studies
that oxygen-deficient WOx species can easily activate and
dissociate molecular hydrogen, suggesting great potential as a
hydrogenation catalyst.[36] An enhancement in the selectivity
towards 15PDO and THP is observed for 5Niꢀ15WOx catalyst as
compared to 15WOx catalyst. This shows that the presence of
nickel can facilitate the formation of dissociated hydrogen,
which is necessary to form WꢀOxH acid sites. More importantly,
the better sum of the selectivity, around 80%, is obtained over
5Niꢀ15WOx as compared to the one obtained with 15WOx
catalyst (about 55%). This suggests that higher availability of
dissociated hydrogen at the surface of bimetallic NiꢀWOx
catalyst can enhance elimination of ionic intermediate from the
catalyst surface before further participating in cracking and
polymerization reactions.[37] These results along with the acid
properties of monometallic and bimetallic catalysts indicate
that (metal+Brønsted acid sites) and (metal+Lewis acid sites)
are responsible for the generation of pentane polyols and THP,
respectively.
Figure 6. XANES spectra at the Ni K-edge of 10Ni and 10Niꢀ30WOx catalysts.
16 10Niꢀ30WOx catalyst. The increase in the intensity of white line
17 peak can be attributed to electron transfer from Ni to W due to
18 the formation NiꢀW intermetallic phase. This result can further
19 supports the formation of NiꢀW intermetallic phase in
20 10Niꢀ30WOx catalyst.
21
22
23 Catalytic Evaluation
24
25 The catalytic activity during THFA hydrogenolysis over 10Ni,
26 NiꢀWOx and WOx catalysts is reported in Table 3. The results
27 over 10Ni catalyst show that Ni itself is able to undergo ring-
28 opening reaction at the more sterically hindered secondary
29 CꢀO bond, producing mostly 15PDO (33%), followed by
30 125PTO (11%), 12PDO (7.5%) and THP (7%).[14a] Various minor
31 products, including 3-hydroxytetrahydropyran (3-HTHP), 2-
32 methyltetrahydrofuran (2-MTHF), tetrahydrofuran (THF), and 1-
33 butanol, are labelled as “others” in Table 3. Comparing to the
34 unpromoted 10Ni catalyst, THFA conversion was progressively
35 enhanced from around 8% to 39% as tungsten surface density
36 is increased, suggesting that the catalytic activity is strongly
37 dependent on tungsten content. Likewise, the selectivity to
38 15PDO is also increased up to approximately 43% over
39 10Niꢀ15WOx catalyst. Considering product yields and mass
40 balance, both promoted 10Ni catalysts with 15 and 30 wt% W
41 result in the best catalytic performance under the reaction
42
Moreover, a synergistic effect exists between nickel and
tungsten since higher THFA conversion and higher selectivity
towards 15PDO and THP are achieved using physical mixture of
43
44
Table 3. CꢀO hydrogenolysis of THFA over NiꢀWOx catalysts.[a]
Carbon balance [c] Rate
45
46
47
48
49
50
51
52
53
54
55
56
57
Entry Catalyst
THFA conversion Product selectivity [%]
[%]
15PDO 125PTO 12PDO THP 1-pentanol others [b] [%]
[mmolgcatꢀ1 minꢀ1
]
1
2
3
4
5
6
7
8
9
10
10Ni
7.9
33.1
35.3
37.6
42.9
39.8
27.8
39.7
29.3
33.8
47.3
11.1
18.1
12.7
5.8
5.2
7.3
7.5
4
1.5
0.3
0.1
0
0.1
0
0
6.7
9.2
19.3 4.6
30.5 5.1
2.3
2.6
18
10.1
9.5
14.4
7.3
97.9
96.3
95.8
99.0
95.6
85.4
93.4
89.8
90.6
96.7
33.5
69.2
10Niꢀ2.5WOx
10Niꢀ7.5WOx
10Niꢀ15WOx
10Niꢀ30WOx
15WOx
5Niꢀ15WOx
10Ni+15WOx
10Ni+H2WO4
16.6
25.1
35.6
39.2
31.7
31.9
35.9
34.3
104.8
130.3
155.7
131.3
128.1
143.5[g]
145.5[g]
117.3
33
4.3
12.2 1.8
23.2 3.1
23.9 3.3
13.7 1.5
21.1 2.7
5.8
6.7
6.8
[d]
[e]
11.8
11.6
16
3.2
12.1[f]
2.1
10Niꢀ30WOx-500-800 28.7
0
[a] Reaction conditions: 2508C, 34 bar H2, 4 h, 5wt % THFA in water (60 g), 0.3 g catalyst. [b] Others include: 3-HTHP, 2-MTHF, THF and 1-butanol; [c] Moles of
liquid carbon measured by TOC analyzer; [d] Physically mixed, 10Ni (0.3 g) and 15WOx (0.3 g); [e] Physically mixed, 10Ni (0.3 g) and H2WO4 (0.3 g); [f] Others
plus 2-HTHP; [g] Rate calculated based on the amount of 10Ni catalyst.
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
7
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
10Ni and 15WOx catalysts (Table 3, entry 8) as compared to the
values obtained over both of them separately. Although lower
selectivity towards 15PDO and THP over physical mixed
catalysts is obtained as compared to the bimetallic catalyst, it
seems that the synergistic effect might not necessarily require
an intimate contact between nickel and tungsten components,
at least in the solid phase, although WOx species could be
converted into soluble HxWO3, which can closely interact with
Ni species in the aqueous phase. However, an initial intimate
Recyclability of Spent Catalysts
Catalyst recyclability is one of the most critical challenges in
heterogeneous catalysis. Hence, it is of great importance to
further investigate the recyclability of these bimetallic catalysts
and to identify the deactivation mechanisms, particularly in
liquid–phase reactions. It is important to note that the
recyclability experiments and deactivation process in batch
reactions can be misinterpreted if the repeated runs are
conducted at high conversion levels.[38] Herein, we have
examined the performance of the recycled 10Niꢀ30WOx catalyst
in THFA hydrogenolysis at shorter reaction time (i.e., conversion
level of below 20%) and the results are shown in Figure S8 (see
the Supplementary Information for more details). It is evident
that the THFA conversion level decreased from around 19.6%
to 10.2% in the second recycle run and dropped to 5.2% in the
third run. Likewise, slightly lower 15PDO selectivity, 33% as
compared to 38% in the first run, is achieved in the second run,
followed by further drop to around 23% in the third run. In
contrast, more 125PTO is produced in the second and third
runs, with a selectivity of 45% and 40%, respectively.
10 interaction between Ni and WOx enhances the selectivity
11 towards 15PDO. This might be because of the nature of
12 different reaction steps, proceeding independently on the
13 different types of catalytic sites. Similar trend is observed over
14 the physical mixture between 10Ni catalyst and tungstic acid
15 (H2WO4) promoter (Table 3, entry 9). Small amount of 2-
16 hydroxytetrahydropyran (2-HTHP) is observed as reaction by-
17 product which was also reported in our previous study.[14a]Other
18 parameters, including amount of catalyst and reaction time,
19 were subsequently investigated by comparing the performance
20 of 10Niꢀ30WOx catalyst in the THFA conversion. Figure 7 (a)
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
To determine the cause of the catalyst activity loss, the
presence of soluble metal species in reaction solutions was
investigated, as well as the changes of the spent catalysts due
to hydrothermal degradation. Ni and W losses by leaching were
evaluated by ICP analysis of reaction solution after the first run
over different 10NiꢀWOx catalysts at nearly comparable con-
version level (~14%), and the results are reported in Table S3.
Compared to the initial loading, 2.6% weight loss is observed
for 10Ni catalyst, indicating a weak interaction between Ni and
silica support. As W loading increases, both Ni and W weight
losses are increased up to around 16% for 10Niꢀ15WOx
(Table S3, entry 4). Surprisingly, in the case of 10Niꢀ30WOx
catalyst, Ni and W concentrations considerably decrease to
around 5% mass loss for both metals (Table S3, entry 5), despite
the fact that higher surface tungsten density is present to leach
into the reaction solution. This indicates that there might be a
possible correlation between the intermetallic phase formation
on 10Niꢀ30WOx and its better stability against leaching.
36
Figure 7. CꢀO bond hydrogenolysis of THFA over 10Niꢀ30WOx/SiO2 catalysts
at 2508C, 34 bar H2, 5wt % THFA in water/1,4-Dioxane (60 g): (a) Effect of
catalyst loading in water, (b) Effect of reaction time in water, and (c) Effect of
reaction time in 1,4-dioxane. Others include: 1-pentanol (for (a) and (b) only),
12PDO, 3-HTHP, 2-MTHF, THF and 1-butanol.
37
38
39
40
41
Additionally, we have explored the catalyst degradation
using characterization techniques (mostly XRD) for the spent
catalysts (Figure 8). As shown in the XRD patterns of spent
catalysts, a significant growth of Ni crystallites size after
hydrogenolysis reaction is observed for 10Ni and 10Niꢀ2.5WOx
catalysts. Moreover, new reflection peaks corresponding to
crystalline NiWO4 also appear for 10Niꢀ2.5WOx catalyst, most
likely due to the leaching of the Ni and W species into hot
water and their re–deposition on the catalyst surface.[31] These
peaks are significantly strengthened with increasing tungsten
loading, whereas the crystalline Ni and NiꢀW intermetallic
phase disappear. However, it is important to note that the
NiꢀW intermetallic phase is still present on the XRD profiles of
spent catalysts (see Figure S9 of the Supplementary Information
for more details) during different recycling experiments con-
ducted at shorter reaction time. This may further support that
the intermetallic NiꢀW phase is more stable under hydro-
thermal condition. Previous studies by Zhang et. al.[22] revealed
42 and 7(b) show the distribution of products formed during THFA
43 hydrogenolysis over 10Niꢀ30WOx catalyst under aqueous-phase
44 conditions as a function of conversion level. These results show
45 that 15PDO and 125PTO are initially formed over 10Niꢀ30WOx
46 via hydrogenolysis and hydrolysis pathways, respectively, with a
47 highest combined selectivity of approximately 64% (at 15.5%
48 conversion). Subsequently, at higher conversion level, which
49 can be achieved by either employing a higher amount of
50 catalyst or longer reaction times, a higher THP selectivity of
51 nearly 50% (at around 76% conversion) is observed. In contrast,
52 selectivity to 15PDO and 125PTO decreased to less than 20%,
53 indicating that these polyols are participating in the cyclo-
54 dehydration pathways in the presence of acid sites (Scheme 1).
55
56
57
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
8
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
catalysts were denoted as Ni4WꢀXꢀY and 10Niꢀ30WOxꢀXꢀY,
where X and Y indicate the calcination and reduction temper-
atures, respectively. The H2-TPR profiles of the catalysts are
presented in Figure S11 (see the Supplementary Information),
respectively. The TPR result of calcined Ni4W catalyst reveals
broad peaks from around 4508C to 8008C, which correspond to
the reduction of NiO and NiWO4 particles, having strong
interaction with the surface of the catalyst.[39] XRD profiles of
calcined 10Niꢀ30WOx-800 and Ni4W-800 catalyst are shown in
Figure 9-a, indicating the formation of NiWO4 as a dominant
crystalline phase of catalysts calcined at 8008C.[28a]
Figure 8. XRD profiles of spent catalysts after THFA hydrogenolysis during
4 h at 2508C and 34 bar: (a) 10Ni, (b) 10Niꢀ2.5WOx, (c) 10Niꢀ7.5WOx, (d)
10Niꢀ15WOx, and (e) 10Niꢀ30WOx catalysts.
18 that tungsten species, regardless of their valence states,
19 dissolve in high-temperature water and form soluble hydrogen
20 tungsten bronze (HxWO3), being catalytically active in cellulose
21 hydrogenolysis. The formation of tungsten bronze in the
22 presence of leached nickel ions under hydrothermal condition
23 can form nickel tungstate species, which will be subsequently
24 deposited on the surface of the spent catalyst after cooling
25 down to room temperature and exposure to air.[19a,31] It is
26 important to mention that a blue color solution after running
27 the reactions using either WOx or NiꢀWOx catalyst is always
28 observed, indicating the presence of soluble tungsten bronze.
29 However, the blue color solution becomes transparent after
30 exposure to air, suggesting that the dissolve tungsten bronze
31 species, either in the hydrogen or nickel form, are oxidized and
32 deposited on the surface of the catalyst. Unlike the previous
33 study[18a] which identified HxWO3 phase by XRD after 6 min from
34 catalyst recovery, the NiWO4 phase is always observed in the
35 XRD pattern of spent catalyst, regardless of exposure time to
36 air. Furthermore, as reported before,[18a] the color disappearance
37 was within 3 hours, whereas it was quite fast in the current
38 study, less than 1 h depending on the W density of the catalyst
39 (Figure S10).
40
41
42 Effect of Calcination and Reduction Temperature
43
44 The characterization of the silica-supported 10Niꢀ30WOx cata-
45 lyst by XRD, XPS, and EXAFS/XANES demonstrated the
46 formation of a NiꢀW intermetallic phase, along with WOx
47 species and probably metallic phases of nickel and tungsten.
48 Moreover, a less leaching tendency of Ni and W is observed on
49 10Niꢀ30WOx catalyst, even though a higher amount of W is
50 employed. To determine whether the NiꢀW intermetallic phase
51 formation was responsible for the better stability of the catalyst
52 against leaching, we have investigated the effect of calcination
53 and reduction temperatures on both the catalytic performance
54 and the leaching of the active metals. Two catalysts,
55 10Niꢀ30WOx and Ni4W (Ni/W molar ratio=4) were prepared, as
56 reported in the material and method section, to investigate the
57 effect of calcination and reduction temperatures. The resulting
Figure 9. a) (top) XRD patterns of calcined catalysts: (a) Ni4W-800 and (b)
10Niꢀ30WOx-800. b) (middle) XRD patterns of reduced catalysts: (a)
10Niꢀ30WOx-500-500, (b) 10Niꢀ30WOx-500-800, (c) 10Niꢀ30WOx-800-800,
and (d) Ni4W-800-800. c) (bottom) XRD patterns of spent 10Niꢀ30WOx-500-
800 catalyst after THFA hydrogenolysis at 2508C, 34 bar, 4 h.
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
9
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
3
4
5
6
7
8
9
The high intensity reduction peak (ca. 8008C) on the TPR
profile of 10Niꢀ30WOx-800 sample can be also attributed to the
integrated reduction of NiO, WOx, and NiWO4 phases.[40] After
being reduced at 8008C, crystalline metallic W and Ni4W
intermetallic phases are clearly identified in the XRD patterns of
10Niꢀ30WOx, regardless of calcination temperature, as shown
in Figure 9-b (b) and (c). Additionally, new peaks assigned to
metallic Ni, are observed in the case of Ni4W-800-800 catalyst.
To evaluate the performance and stability of the catalyst, THFA
10 hydrogenolysis reaction was conducted over 10Niꢀ30WOx-500-
11 800 catalyst (Table 3, entry 10). The highest selectivity to 15PDO
12 (47.3%) is obtained, along with THP (21%) and 125PTO (16%)
13 at nearly 29% THFA conversion. More importantly, it is obvious
Scheme 2. Reaction pathway for the CꢀO hydrogenolysis of THFA using 1,4-
dioxane as solvent.
14 that the peaks corresponding to metallic
W
and Ni4W
contributing in the ring-opening THFA in the absence of water.
Besides, only a small amount of 15PDO is obtained in the
hydrogenolysis reaction in 1,4-dioxane, suggesting the crucial
role of water in the formation of 15PDO from THFA.[4d,43] DRIFTS
measurements also prove the transformation of Lewis acid sites
into Brønsted acid sites in the presence of water, as shown in
Figure S3. It is clear that the concentration of Lewis acid sites
decreased while the formation of some Brønsted acid sites was
observed in the presence of water vapor. This is attributed to
the dissociative adsorption of water molecules on Lewis acid
sites to form OH groups.[44] This further confirms the direct
hydrogenolysis pathway in the ring-opening THFA under
aqueous-phase conditions.
15 intermetallic phases remain nearly unchanged in the XRD
16 pattern of spent catalyst (Figure 9-c), which is consistent with
17 the reduction of the losses by leaching (Table S3, entry 11). This
18 result suggests that the W species involved in the intermetallic
19 and metallic phases are probably more resistant to leaching
20 under hydrothermal reaction condition. Nevertheless, NiWO4
21 crystalline phase is still formed due to the dissolution of nickel
22 and tungsten into hot water and their redeposition on the
23 catalyst after cooling down to room temperature.
24
25
26 Effect of Solvent
27
Finally, the Ni and W losses by leaching in the reaction
products were determined, as well as the physical changes of
the spent catalyst. Not only trace amounts of Ni and W are
observed after the reaction using 1,4-dioxane as reaction
medium (Table S3), but also the crystalline structure of the
spent catalyst remains nearly stable, as shown by the XRD
pattern in Figure S12.
28 As discussed before, the use of water as a solvent has a
29 deactivating effect since hydrothermal degradation of the
30 catalyst often occurs due to the leaching of metallic active
31 sites.[38] Although water is a low cost and environmentally
32 friendly solvent, the degree of catalyst degradation and
33 deactivation may limit its usefulness for practical application.
34 The easiest way to manage deactivation problems caused by
35 leaching is to choose a more appropriate solvent. The selection
36 of a proper solvent is especially important since the solvent
37 affects the stability of acidic protons, transition states, and
38 products, thereby affecting the turnover frequencies, reaction
39 rates, reaction pathways, product distributions and yields.[41] It
40 has been shown that the use of an organic solvent like g-
41 valerolactone (GVL) has been beneficial in increasing reaction
42 rates and product selectivities as compared to water. Therefore,
43 the effect of an organic solvent (1,4-dioxane) in THFA hydro-
44 genolysis over 10Niꢀ30WOx catalyst was investigated and the
45 results are shown in Figure 7(c). A lower hydrogenolysis rate is
46 achieved using 1,4-dioxane as compared to the reactions in
47 water at similar reaction condition (Figure 7(b)). Similar trend
48 was observed in a previous study over IrꢀReOx catalyst using n-
49 heptane as solvent.[42] Moreover, a considerable amount of THP
50 is produced with a yield of around 54%, while the 15PDO yield
51 is less than 2% over 10Niꢀ30WOx in 1,4-dioxane. Besides, 1-
52 pentanol is the second main product which is obtained in 1,4-
53 dioxane. These results indicate that THFA participates in
54 dehydration pathway to produce dihydropyran (DHP) over acid
55 sites, followed by further hydrogenation step into THP over
56 metal sites, as proposed in Scheme 2. This shows that indirect
57 hydrogenolysis pathway (i.e., dehydration-hydrogenation) is
Conclusions
In summary, the promoting effect of WOx on the catalytic
performance of Ni catalyst in the ring–opening of THFA to
produce 15PDO was explored. It is shown that the higher
concentration of Lewis acidity provided by oxidated species of
tungsten enhances the catalytic activity of THFA hydrogenolysis
and selectivity towards 15PDO. Better mass balance and ca.
80% combined selectivity to 15PDO, 125PTO and THP is
obtained over bimetallic NiꢀWOx catalysts, as compared with
the corresponding monometallic catalysts. However, like pre-
vious studies, NiꢀWOx catalytic system suffered from both Ni
and
W losses by leaching under hydrothermal reaction
condition. Interestingly, 10Niꢀ30WOx catalyst with the highest
tungsten density showed less tendency to leaching due to the
formation of NiꢀW intermetallic phases. The role of NiꢀW
intermetallic phases on the catalyst stability was further
investigated by changing the calcination and reduction temper-
atures. Up to 47% selectivity into 15PDO, along with the
highest combined pentane polyol (i.e., 15PDO and 125PTO)
selectivity of ca. 64%, was obtained over 10Niꢀ30WOx-500-800
catalyst (at around 30% conversion level). More importantly, it
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
10
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
flowing He and then cooled down to room temperature. XAS
spectra of the catalysts were then collected under He at room
temperature. The XAS data were processed using the WinXAS
software to obtain the structural parameters.
1
2
3
4
5
6
7
8
9
is found that metallic W and Ni4W intermetallic phase are more
stable against leaching and their corresponding peaks in the
XRD patterns of the spent catalysts remain nearly unchanged.
This result confirms previous findings regarding temperature-
controlled phase transfer characteristics of W-promoted cata-
lysts. It also provides additional evidence, suggesting that
intermetallic NiꢀW phases are more stable under liquid phase
reactions. Finally, 1,4-dioxane as an organic solvent was used in
THFA hydrogenolysis reaction, yielding THP with a yield of ca.
The metal particle sizes of NiꢀWOx catalysts were estimated by
transmission electron microscopy (TEM), using a Tecnai G2 TF20 S-
twin microscope (FEI Company). Prior to the observation of
reduced samples, the catalyst was treated at 5008C under H2 for
2 h, which is similar to the reduction condition for the catalytic
tests. The sample was then ultrasonically dispersed in ethanol and
spread over a perforated copper grid. More than 300 nickel
particles were counted to obtain the histograms of particle size
distribution.
10 54%. The 10Niꢀ30WOx catalyst is also stable after reaction in
11 1,4-dioxane, with only trace amount of Ni and W leaching This
12 indicates the crucial role of solvent in changing the reaction
13 pathways, product distribution and yields. Furthermore,
14 Brønsted acid sites can be generated from Lewis acid sites in
15 the presence of water, indicating the crucial role of water in the
16 formation of 15PDO from THFA.
H2-Temperature-Programed Reduction (H2ꢀTPR) measurement of
fresh catalysts was performed on TPD/R/O 1100 system (Thermo
Scientific) equipped with a thermal conductivity detector (TCD).
Prior to TPR measurement, 50 mg of catalyst was outgassed in Ar
at 2008C for 1 h to remove impurities and then cooled down to
room temperature. 5% H2/N2 gas was passed through the catalyst
bed (50 mLminꢀ1), while the temperature of the furnace was
17
18
increased from 50 to 9008C at a heating rate of 108Cminꢀ1
.
19 Experimental Section
20
21 Catalyst Preparation
22
NH3-Temperature-Programed Desorption (NH3-TPD) was also con-
ducted on TPD/R/O 1100 instrument. About 150 mg of catalyst was
loaded into the quartz tube, outgassed in Ar at 2008C for 1 h, and
then cooled down to room temperature. The catalyst was exposed
to NH3 (50 mLminꢀ1) for 1 h, and subsequently, the sample was
flushed with He for 1 h to remove physisorbed ammonia. For the
TPD measurement, the catalyst was heated from 100 to 7008C in
Supported NiꢀWOx catalysts were prepared by stepwise conven-
23
tional incipient wetness impregnation of silica support (Kanto)
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
using an aqueous solution of nickel nitrate hexahydrate (98%, Alfa
Aesar) and ammonium metatungstate (99.99%, Sigma Aldrich),
with nominal Ni and W loading of 10 and 0–30 wt%, respectively.
Prior experiments, catalysts were dried (1108C), calcined in air
(5008C), reduced in flowing H2 (5008C), and then passivated with
1%O2 in N2 (258C) before exposing to air. The resulting catalysts
were denoted as xNiꢀyWOx, where x and y indicate the nominal
metal loading of Ni and W, respectively.
flowing He (50 mLminꢀ1
108Cminꢀ1 to desorb ammonia.
) with a temperature ramp rate of
The IR spectra of chemisorbed pyridine (Pyr-IR) were recorded
using a Fourier transform infrared spectrometer (BIO-RAD Excalibur
series FTS 3000). The sample was prepared as a self-supporting
wafer and outgassed in vacuum at 3008C for 1 h, cooled down to
room temperature, followed by exposure of the sample to pyridine
vapor, and evacuation at room temperature. The IR spectra of
chemisorbed pyridine were recorded subsequently in the spectral
range of 400–4000 cmꢀ1 with a resolution of 4 cmꢀ1. 128 scans
were recorded for each spectrum. The desorption of pyridine was
carried out by heating the wafer under vacuum from room
temperature to 1008C, 2008C and 3008C and held at each
temperature for half an hour. FTIR spectra were recorded at each
temperature.
Catalyst Characterization
The Brunauer-Emmett-Teller (BET) surface areas of the catalysts
were determined by nitrogen adsorption at ꢀ1968C using a
Micromeritics ASAP-2020 adsorption apparatus. Prior to the meas-
urements, samples were degassed in vacuum at 2008C overnight.
The actual loading of Ni and W was determined by X-ray
Fluorescence (XRF) measurements using a S4 Explorer spectrometer
(Bruker AXS).
Pyr–IR experiments were also conducted in
a Bruker FTIR
Powder X-ray diffraction (XRD) patterns were recorded on a Bruker
spectrometer (VERTEX 70) equipped with a Harrick Praying Mantis
Diffuse reflectance Fourier transform infrared spectroscopy (DRIFTS)
cell fitted with KBr windows, and connected to a gas flow system.
Before DRIFT measurements, the samples were reduced at 5008C in
H2 flow for 1 h. The FTIR spectra were recorded after cooling down
to room temperature and used as a reference to carry out
background subtraction. For experiments in the presence of
adsorbed water, water vapor was passed through the cell using a
saturator at 3508C and H2 as carrier gas. The sample was then
exposed to pyridine vapor and the IR spectra of chemisorbed
pyridine were recorded in the spectral range of 400–4000 cmꢀ1 at a
resolution of 4 cmꢀ1. The desorption of pyridine was carried out
through heating the catalyst from room temperature to up to
4008C and held at each temperature for half an hour.
D8 Advance diffractometer using filtered Cu Ka radiation (l=
˚
1.5406 A) produced by an X-ray source operated at 35 kV and
40 mA. The diffraction patterns were taken in the Bragg angle (2q)
range from 208 to 808. The XRD data were evaluated by fitting the
whole pattern according to the Rietveld method (TOPAS software).
X–ray photoelectron spectroscopy (XPS) was performed on a VG
ESCALAB 250 spectrometer equipped with a monochromatic Mg
Ka radiation source. For reduced samples, the XPS data of the
catalysts were obtained after in situ H2 reduction inside the XPS
pre-chamber at 5008C for 1 h. All binding energies were calibrated
with the C1s peak at 284.6 eV as a reference.
X–ray absorption spectra (XAS) of Ni K-edge and W L3-edge were
recorded at the XAFCA beamline at the Singapore Synchrotron
Light Source.[45] The samples were first mixed and ground
thoroughly with boron nitride, followed by pressing the sample
into a small wafer with a diameter of 1 cm. The catalysts were
reduced with H2 inside the chamber at 5008C for 2 h, purged with
The acid site densities on the catalyst were also quantified using
Isopropylamine Temperature–Programed Desorption-Mass Spec-
trometry (IPA-TPD-MS). About 50 mg of the sample was loaded in a
fixed bed reactor and pretreated in He at 5008C for 1 h to remove
remaining water. The sample was cooled down to 1008C, dosed
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
11
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
[3] a) S. Garcꢂa-Fernꢃndez, I. Gandarias, J. Requies, M. Gꢄemez, S. Bennici,
A. Auroux, P. Arias, J. Catal. 2015, 323, 65–75; b) Y. Nakagawa, M.
Tamura, K. Tomishige, J. Mater. Chem. A 2014, 2, 6688–6702; c) S. Garcꢂa-
Fernꢃndez, I. Gandarias, J. Requies, F. Soulimani, P. L. Arias, B. M.
Weckhuysen, Appl. Catal. B 2017, 204, 260–272; d) Y. Fan, S. Cheng, H.
Wang, D. Ye, S. Xie, Y. Pei, H. Hu, W. Hua, Z. H. Li, M. Qiao, Green Chem.
2017, 19, 2174–2183; e) J. ten Dam, K. Djanashvili, F. Kapteijn, U.
Hanefeld, ChemCatChem 2013, 5, 497–505; f) S. Garcꢂa-Fernꢃndez, I.
Gandarias, Y. Tejido-NfflÇez, J. Requies, P. L. Arias, ChemCatChem 2017, 9,
4508–4519.
[4] a) S. Koso, I. Furikado, A. Shimao, T. Miyazawa, K. Kunimori, K. Tomishige,
Chem. Commun. 2009, 2035–2037; b) S. Koso, N. Ueda, Y. Shinmi, K.
Okumura, T. Kizuka, K. Tomishige, J. Catal. 2009, 267, 89–92; c) S. Koso,
Y. Nakagawa, K. Tomishige, J. Catal. 2011, 280, 221–229; d) M. Chia, Y. J.
Pagan-Torres, D. Hibbitts, Q. Tan, H. N. Pham, A. K. Datye, M. Neurock,
R. J. Davis, J. A. Dumesic, J. Am. Chem. Soc. 2011, 133, 12675–12689;
e) K. Chen, K. Mori, H. Watanabe, Y. Nakagawa, K. Tomishige, J. Catal.
2012, 294, 171–183.
[5] a) K. Chen, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, ChemCatChem
2010, 2, 547–555; b) T. Buntara, S. Noel, P. H. Phua, I. Meliꢃn-Cabrera,
J. G. de Vries, H. J. Heeres, Angew. Chem. Int. Ed. 2011, 50, 7083–7087;
Angew. Chem. 2011, 123, 7221–7225; c) T. Buntara, I. Meliꢃn-Cabrera, Q.
Tan, J. L. Fierro, M. Neurock, J. G. de Vries, H. J. Heeres, Catal. Today
2013, 210, 106–116; d) A. C. Alba-Rubio, C. Sener, S. H. Hakim, T. M.
Gostanian, J. A. Dumesic, ChemCatChem 2015, 7, 3881–3886.
[6] a) K. Tomishige, Y. Nakagawa, M. Tamura, Green Chem. 2017; b) M.
Schlaf, Dalton Trans. 2006, 4645–4653.
with IPA until complete saturation, and then flushed with He for
1 h to remove physisorbed IPA. Employing mass spectrometer, the
concentration of unreacted IPA and desorbed propylene resulting
from IPA decomposition was quantified to determine the Lewis and
Brønsted acid site density, respectively.
1
2
3
4
5
6
7
8
9
Catalytic Measurements
Hydrogenolysis experiments of THFA were carried out in a 100 mL
stainless-steel autoclave reactor (Parr instrument). For each reac-
tion, THFA (5 wt%), solvent, and the appropriate amount of catalyst
were loaded into the reactor. The reactor was sealed, purged three
times with hydrogen, and then pressurized to 34 bar with H2. The
reactor was then heated to the required temperature, using a
stirring rate of 500 rpm. Once the reaction time was completed, the
reactor was cooled to room temperature. The liquid products were
collected and transferred to vials and the catalyst was separated by
centrifugation and filtration. To obtain the overall carbon balance,
liquid samples were analyzed using total organic carbon (TOC)
analyzer (Shimadzu TOC-VCPH). TOC calibrations were performed
with carbon standards (SpectroPure). For recyclability experiments,
the spent catalyst was dried in vacuum at 408C, and used in the
reaction without further reduction step.
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
[7] S. Liu, Y. Amada, M. Tamura, Y. Nakagawa, K. Tomishige, Green Chem.
2014, 16, 617–626.
[8] S. Liu, Y. Amada, M. Tamura, Y. Nakagawa, K. Tomishige, Catal. Sci.
Technol. 2014, 4, 2535–2549.
Analytical Methods
Quantitative analysis of the liquid products was performed using
gas chromatography (GC, Agilent, 6890N) and high performance
liquid chromatography (HPLC, Waters). The GC was equipped with
a CPWax 52CB column and a flame ionization detector (FID). The
HPLC was equipped with a Refractive Index Detector (RID) and an
ion-exclusion column (Aminex HPX-87H, 300 mmꢁ7.8 mm), kept at
608C, with 0.005 M H2SO4 at a flow rate of 0.5 mLminꢀ1 as the
mobile phase. Inductively coupled plasma atomic emission spec-
trometry (ICP-AES) was used to detect the presence of Ni and W in
the reaction products.
[9] a) Y. Nakagawa, K. Tomishige, Catal. Today 2012, 195, 136–143; b) Y.
Nakagawa, M. Tamura, K. Tomishige, ACS Catal. 2013, 3, 2655–2668; c) Y.
Nakagawa, M. Tamura, K. Tomishige, Catal. Surv. Asia 2015, 19, 249–256.
[10] L. Schniepp, H. Geller, J. Am. Chem. Soc. 1946, 68, 1646–1648.
[11] a) Z. J. Brentzel, K. J. Barnett, K. Huang, C. T. Maravelias, J. A. Dumesic,
G. W. Huber, ChemSusChem 2017, 10, 1351–1355; b) K. Huang, Z. J.
Brentzel, K. J. Barnett, J. A. Dumesic, G. W. Huber, C. T. Maravelias, ACS
Sustainable Chem. Eng. 2017.
[12] S. Sitthisa, D. E. Resasco, Catal. Lett. 2011, 141, 784–791.
[13] J. Lee, Y. Xu, G. W. Huber, Appl. Catal. B 2013, 140–141, 98–107.
[14] a) E. Soghrati, C. Choong, C. K. Poh, S. Kawi, A. Borgna, ChemCatChem
2017, 9, 1402–1408; b) E. Soghrati, T. K. C. Ong, C. K. Poh, S. Kawi, A.
Borgna, Appl. Catal. B 2018, 235, 130–142.
[15] H. W. Wijaya, T. Kojima, T. Hara, N. Ichikuni, S. Shimazu, ChemCatChem
2017.
[16] S. De, J. Zhang, R. Luque, N. Yan, Energy Environ. Sci. 2016, 9, 3314–
3347.
35 Acknowledgements
36
37 The authors gratefully thank Singapore Agency for Science,
[17] N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J. G. Chen, Angew.
Chem. 2008, 120, 8638–8641; Angew. Chem. Int. Ed. 2008, 47, 8510–
8513.
38 Technology and Research (A*STAR), and NEA (ETRP 1501 103) for
39 their generous financial supports. The help of Dr. Liu Dapeng and
40 Mr. Jason Chen is also greatly acknowledged. One of us (E.S.) is
41 grateful to A*STAR for the Singapore International Graduate
42 Award (SINGA) scholarship.
43
[18] a) Z. Tai, J. Zhang, A. Wang, M. Zheng, T. Zhang, Chem. Commun. 2012,
48, 7052–7054; b) Z. Tai, J. Zhang, A. Wang, J. Pang, M. Zheng, T. Zhang,
ChemSusChem 2013, 6, 652–658.
ˇ
[19] a) K. Fabicovicovꢃ, O. Malter, M. Lucas, P. Claus, Green Chem. 2014, 16,
3580–3588; b) S. J. You, I. G. Baek, E. D. Park, Appl. Catal. A 2013, 466,
161–168.
44
[20] a) Y. Liu, C. Luo, H. Liu, Angew. Chem. 2012, 124, 3303–3307; Angew.
Chem. Int. Ed. 2012, 51, 3249–3253; b) N. Li, Y. Zheng, L. Wei, H. Teng, J.
Zhou, Green Chem. 2017, 19, 682–691; c) Y. Cao, J. Wang, M. Kang, Y.
Zhu, RSC Adv. 2015, 5, 90904–90912.
45 Conflict of Interest
46
47 The authors declare no conflict of interest.
[21] C. Liu, C. Zhang, S. Sun, K. Liu, S. Hao, J. Xu, Y. Zhu, Y. Li, ACS Catal.
2015, 5, 4612–4623.
48
49
50
51
52
[22] a) A. Wang, T. Zhang, Acc. Chem. Res. 2013, 46, 1377–1386; b) M. Zheng,
J. Pang, R. Sun, A. Wang, T. Zhang, ACS Catal. 2017, 7, 1939–1954.
[23] X. Zuwei, Z. Ning, S. Yu, L. Kunlan, Science 2001, 292, 1139–1141.
[24] M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso,
J. Rouquerol, K. S. Sing, Pure Appl. Chem. 2015, 87, 1051–1069.
[25] T. Kim, A. Burrows, C. J. Kiely, I.E. Wachs, J. Catal. 2007, 246, 370–381.
[26] a) I.E. Wachs, T. Kim, E. I. Ross, Catal. Today 2006, 116, 162–168; b) C. D.
Baertsch, K. T. Komala, Y.-H. Chua, E. Iglesia, J. Catal. 2002, 205, 44–57;
c) J. Macht, C. D. Baertsch, M. May-Lozano, S. L. Soled, Y. Wang, E.
Iglesia, J. Catal. 2004, 227, 479–491.
Keywords: Hydrogenolysis · Tetrahydrofurfuryl alcohol · 1,5-
pentanediol · Nickel-Tungsten · Intermetallic
53
[1] a) M. Besson, P. Gallezot, C. Pinel, Chem. Rev. 2013, 114, 1827–1870;
54
55
56
57
b) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007,
46, 7164–7183; Angew. Chem. 2007, 119, 7298–7318.
[2] a) A. M. Ruppert, K. Weinberg, R. Palkovits, Angew. Chem. Int. Ed. 2012,
51, 2564–2601; Angew. Chem. 2012, 124, 2614–2654; b) P. Yazdani, B.
Wang, Y. Du, S. Kawi, A. Borgna, Catal. Sci. Technol. 2017, 7, 4680–4690.
[27] R. Gorte, Catal. Lett. 1999, 62, 1–13.
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
12
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
[28] a) S. Chen, L. Zeng, H. Tian, X. Li, J. Gong, ACS Catal. 2017, 7, 3548–
3559; b) J. H. Kwak, G. Y. Han, K. J. Yoon, Int. J. Hydrogen Energy 2013,
38, 8293–8305.
[29] Y. Cao, J. Wang, M. Kang, Y. Zhu, J. Mol. Catal. A 2014, 381, 46–53.
[30] M. Y. Zheng, A. Q. Wang, N. Ji, J. F. Pang, X. D. Wang, T. Zhang,
ChemSusChem 2010, 3, 63–66.
[38] I. Sꢃdaba, M. L. Granados, A. Riisager, E. Taarning, Green Chem. 2015, 17,
4133–4145.
1
2
3
4
5
[39] B. Scheffer, P. Molhoek, J. Moulijn, Appl. Catal. 1989, 46, 11–30.
[40] N. Siddiqui, A. S. Roy, R. Goyal, R. Khatun, C. Pendem, A. N. Chokkapu, A.
Bordoloi, R. Bal, Sus. Energy Fuels 2018, 2, 191–198.
[41] L. Shuai, J. Luterbacher, ChemSusChem 2016, 9, 133–155.
[42] Y. Nakagawa, K. Mori, K. Chen, Y. Amada, M. Tamura, K. Tomishige, Appl.
Catal. A 2013, 468, 418–425.
[31] G.-Y. Pan, Y.-L. Ma, X.-X. Ma, Y.-G. Sun, J.-M. Lv, J.-L. Zhang, Chem. Eng. J.
2016, 299, 386–392.
6
[43] R. Ma, X.-P. Wu, T. Tong, Z.-J. Shao, Y. Wang, X. Liu, Q. Xia, X.-Q. Gong,
ACS Catal. 2016, 7, 333–337.
[44] a) G. S. Foo, D. Wei, D. S. Sholl, C. Sievers, ACS Catal. 2014, 4, 3180–
3192; b) A. Mehdad, N. S. Gould, B. Xu, R. F. Lobo, Catal. Sci. Technol.
2018, 8, 358–366.
7
8
9
[32] J. Wang, X. Zhao, N. Lei, L. Li, L. Zhang, S. Xu, S. Miao, X. Pan, A. Wang,
T. Zhang, ChemSusChem 2016, 784–790.
[33] T. Nasu, M. Sakurai, T. Kamiyama, T. Usuki, O. Uemura, T. Yamasaki, J.
Non-Cryst. Solids 2002, 312, 319–322.
[34] a) R. Cury, J.-M. Joubert, S. Tusseau-Nenez, E. Leroy, A. Allavena-Valette,
Intermetallics 2009, 17, 174–178; b) Y. Chen, J. Shen, N.-x. Chen, Solid
State Commun. 2009, 149, 121–125.
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
[45] Y. Du, Y. Zhu, S. Xi, P. Yang, H. O. Moser, M. B. Breese, A. Borgna, J.
Synchrotron Radiat. 2015, 22, 839–843.
[35] P. V. P. s. Handbook, D. Edition, Amsterdam: ASM International 1997.
[36] J. Song, Z.-F. Huang, L. Pan, J.-J. Zou, X. Zhang, L. Wang, ACS Catal.
2015, 5, 6594–6599.
[37] M. Falco, S. Canavese, R. Comelli, N. Fıgoli, Appl. Catal. A 2000, 201, 37–
43.
Manuscript received: May 14, 2018
Version of record online: &&&, &&&&
ChemCatChem 2018, 10, 1–14
www.chemcatchem.org
13
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

Full Papers
1
2
FULL PAPERS
3
4
5
6
Intermetallic decoration: Bi-func-
tional NiꢀWOx catalysts exhibit high
catalytic activity towards the
E. Soghrati, Dr. C. Kok Poh, Dr. Y. Du,
Dr. F. Gao, Prof. S. Kawi*, Dr. A.
Borgna*
selective hydrogenolysis of tetrahy-
drofurfuryl alcohol to 1,5-pentane-
diol, along with 1,2,5-pentanetriol
and tetrahydropyran. Ni4W interme-
tallic phase formation offers a new
approach to stabilize the catalyst
against leaching.
7
8
9
1 – 14
CꢀO Hydrogenolysis of Tetrahydro-
furfuryl Alcohol to 1,5-Pentanediol
Over Bi-functional Nickel-Tungsten
Catalysts
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