Acid catalysis dominated suppression of xylose hydrogenation with increasing yield of 1,2-pentanediol in the acid-metal dual catalyst system

Article abstract of DOI:10.1016/j.apcata.2018.05.019

One-pot conversion of xylose to 1,2-pentanediol was investigated in a dual catalyst system composed of Ru/C and niobium phosphate as hydrogenation and acid catalysts, respectively. A series of niobium phosphate catalysts well-characterized by XRD, N₂ physisorption, FT-IR, NH₃-TPD, Py-IR and XPS were tested regarding the effect of their acid properties on product selectivity for the studied process. A systematic study was reported on the effect of reaction conditions. The combined yield of 21–27% to 1,2-pentanediol and its precursor 1-hydroxyl-2-pentanone was accomplished at 423 K under 3.0 Mpa hydrogen pressure in water-γ-valerolactone/cyclohexane biphasic system. At optimized conditions, the correlation between the product yield and the surface Lewis/Br?nsted ratio were analyzed. The results revealed that the lower apparent activation energy of xylose dehydration reaction catalyzed by Lewis acid site accounted for the high product selectivity for sugar intermediate and furfural hydrogenation processes, especially for the combined selectivity to 1,2-pentanediol and 1-hydroxyl-2-pentanone. This study lays the grounds for further design of improved solid acid catalysts with high selectivity of 1, 2-pentanediol.

Full text of DOI:10.1016/j.apcata.2018.05.019

Accepted Manuscript
Title: Acid catalysis dominated suppression of xylose
hydrogenation with increasing yield of 1,2-pentanediol in the
acid-metal dual catalyst system
Authors: Nailiang Wang, Zhipeng Chen, Licheng Liu
PII:
S0926-860X(18)30241-2
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.05.019
APCATA 16668
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
6-3-2018
14-5-2018
19-5-2018
Please cite this article as: Wang N, Chen Z, Liu L, Acid catalysis dominated
suppression of xylose hydrogenation with increasing yield of 1,2-pentanediol
in the acid-metal dual catalyst system, Applied Catalysis A, General (2018),
https://doi.org/10.1016/j.apcata.2018.05.019
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Acid catalysis dominated suppression of xylose hydrogenation with
increasing yield of 1,2-pentanediol in the acid-metal dual catalyst system
Nailiang Wanga, Zhipeng Chena,b, and Licheng Liua,*
a
CAS Key Laboratory of Bio-based Materials, Qingdao Institute of Bioenergy and Bioprocess
Technology, Chinese Academy of Sciences, Songling Road 189, Qingdao 266101, China
b Graduate University of Chinese Academy of Sciences, Beijing 100049, P.R. China
*Corresponding Author
Licheng Liu
Tel.: +86-532-80662780
E-mail:liulc@qibebt.ac.cn
1

Graphical Abstract:
Highlights:
 Biphasic system assembly with two catalysts is built for the production of 1,2-PeD.
 Lewis acid site enhances the combined selectivity to 1,2-PeD ant its precursor.
 The lower Ea of xylose dehydration accounts for selectivity improvement.
Abstract
One-pot conversion of xylose to 1,2-pentanediol was investigated in a dual catalyst system
composed of Ru/C and niobium phosphate as hydrogenation and acid catalysts, respectively. A series
of niobium phosphate catalysts well-characterized by XRD, N2 physisorption, FT-IR, NH3-TPD,
Py-IR and XPS were tested regarding the effect of their acid properties on product selectivity for the
studied process. A systematic study was reported on the effect of reaction conditions. The combined
yield of 21-27% to 1,2-pentanediol and its precursor 1-hydroxyl-2-pentanone was accomplished at
2

423 K under 3.0 Mpa hydrogen pressure in water-γ-valerolactone/cyclohexane biphasic system. At
optimized conditions, the correlation between the product yield and the surface Lewis/Brønsted ratio
were analyzed. The results revealed that the lower apparent activation energy of xylose dehydration
reaction catalyzed by Lewis acid site accounted for the high product selectivity for sugar
intermediate and furfural hydrogenation processes, especially for the combined selectivity to
1,2-pentanediol and 1-hydroxyl-2-pentanone. This study lays the grounds for further design of
improved solid acid catalysts with high selectivity of 1, 2-pentanediol.
Keyword: Acid catalysis; 1,2-Pentanediol; One-pot conversion; Xylose; Xylitol
3

1. Introduction
1,2-Pentanediol (1,2-PeD) is a high-value chemical used as intermediate in the synthesis of
cosmetics and chemicals [1, 2], especially for propiconazole [3]. Currently, various industrial processes
relying on petroleum resources are available for the production of 1,2-PeD, an example being the
[4]
oxidation of 1-pentene to 1,2-epoxypentene and subsequent hydrolysis to afford 1,2-PeD
.
However, due to the use of high-polluting oxidants such as sodium hypochlorite and peracetic acid,
increasingly serious environmental issues urge people to explore and utilize eco-friendly alternatives.
The lignocellulosic material as a potential C5 resource for the synthesis of 1,2-PeD has attracted
significant attention worldwide in the past from the viewpoint of atom economy and sustainable
development [5-8]. Using renewable xylose from hemicellulose to synthesize 1,2-PeD can reduce the
environmental pressure owing to the absence of oxidant since oxygen elements are largely reserved
in 1,2-PeD.
The multistep processes with furfural derivatives as reaction intermediate are mainly used to
produce 1,2-PeD from the monosaccharide, wherein acid catalyst transforms monosaccharide into
furfural, and subsequently with ring opening hydrogenation by the metal catalyst in molecular
hydrogen [9]. Noting that the dehydration of monosaccharide or corncob to furfural is a commercial
process, many efforts are devoted to the aqueous phase hydrogenolysis of biomass-derived furfural
and furfural alcohol with platinum, ruthenium or non-precious copper catalyst [8, 10-12]. The
ring-opening process to 1,2-PeD is well understood with reaching a high 1,2-PeD yield (about 73%)
[5]. However, the key deficiencies of this multistep strategy lie in high operating, purification cost and
quite inefficient dehydration process, achieving only ~50% of the theoretical furfural yield owing to
4

the degradation and polymerization reactions of furfural, thus the overall 1,2-PeD yield could hardly
exceed the furfural yield in the multistep process.
More recently, an efficient one-pot conversion strategy without isolation of the platform
[13-15][16, 17]
intermediates has been extensively studied for optimizing the technology and route
.
Nijhuis et al. uncover one-pot synthesis of 1,2-PeD and 1,4-PeD with xylose dehydration coupled
with hydrogenation process in the biphasic system [18], but the strategy has not been deeply explored
and the obtained 1,2-PeD selectivity (< 10%) has yet to be improved. Looking into the major routes
for the transformation of xylose in the coexistence of acid and hydrogenation catalysts, several types
of reactions, namely, dehydration, hydrogenation, hydrogenolysis, degradation and polymerization
reactions were mainly involved [19]. These reactions interplay with each other, making the product
distribution of the whole reaction, from sugar alcohol and diols to furanic compounds, strongly
depend on the rate matching among them. Hence, the product selectivity control is an important and
significative issue in this complicated system.
Scheme 1 Reaction network for one-pot conversion of xylose in acid-metal dual catalyst system
According to our previous work [20], a reaction network for one-pot conversion of xylose in the
dual catalyst system is illustrated in scheme 1. When acid and hydrogenation catalysts are integrated
into a single phase system, xylitol is the main product because of its thermodynamical benefit for
xylose hydrogenation reaction. While the aqueous biphasic system can dramatically suppress xylose
5

hydrogenation and ensure xylose be converted into other hydrogenation products. Notably, 1,2-PeD
is the main product among these products when ruthenium, an active metal for
aqueous-hydrogenation of biomass-derived oxygenates to 1,2-PeD, is used as the metal catalyst [12, 21]
Herein, 1,2-PeD is targeted as an objective product for the product selectivity control in our catalyst
system.
.
As previously mentioned, acid and hydrogenation catalysts are involved in the one-pot
conversion of xylose, both can make a significant effect on product distribution. In order to simplify
research, we firstly analyze the apparent activation energy of one-pot conversion process. The results
indicate that the energy barrier for one-pot conversion process is from xylose dehydration reaction,
suggesting that acid catalyst plays the primary role in tuning product distribution. Thus, acid
catalysis dominated variation of product selectivity is emphatically investigated in the present
contribution. Herein, Ru/C is used as a metal catalyst for all reactions.
When considering acid catalysts for monosaccharide dehydration, various materials including
homogeneous and heterogeneous acids have been tested. Of these materials, niobium-based catalyst
[22-28]
is an essential part because of its excellent water tolerance and unique acidic properties
.
Furthermore, porous niobium phosphate with high activity and stability has been synthesized and
tested [27, 29]. More importantly, the Lewis and Brønsted acid site densities are tunable by simply
adjusting preparation conditions. Thus, niobium phosphates are introduced as acid catalysts in view
of tunable acid site in this work. We use hydrothermal method to prepare a series of niobium
phosphate solid acids with different surface Lewis/Brønsted ratio and address the effect of different
acid site on the combined selectivity to 1,2-pentanediol and its precursor 1-hydroxyl-2-pentanone.
2. Experimental section
6

2.1. Materials and Catalysts. The following chemicals and solvents were used without further
purification. Cyclohexane (AR), γ-valerolactone (98%), xylose (99%), xylitol (99%), furfural (98%),
1,2-pentanediol (98%), 1,4-pentanediol (98%), tetrahydrofurfuryl alcohol (98%), cyclopentanone
(97%), cyclopentanol (99%), n-pentyl alcohol (99%), and sec-pentanol (99.5%) were purchased from
Sigma-Aldrich. 1-Hydroxyl-2-pentanone was purchased from Zhejiang University of Technology.
Niobium phosphate samples were synthesized by a hydrothermal method according to a literature
method with slight modification [29]. Typically, 3.9 g of ammonium monohydric phosphate (98.5%)
and 5.4 g ammonium oxalate niobium (19-20% Nb) were dissolved in 120 ml deionized water which
was added to a three-neck flask. A 45 ml solution containing 3.0 g cetyl-trimethyl ammonium
bromide (CTAB, 99.0%) was dropwise added with stirring in the above solution. The pH value of
the final solution was respectively adjusted to 2, 5, 7 and 10 using phosphoric acid and ammonium
hydroxide. The solution was maintained at room temperature for 1 h under stirring, then aged in a
Teflon-lined autoclave for 24 h at 433 K. The resulting suspension was filtered and washed with 800
ml deionized water and subsequently dried at 373 K overnight. The product was then calcined by the
furnace at 773 K for 5 h. The synthesized solid acids at pH value of 2, 5, 7 and 10 were respectively
designated as follow: NbPO-2, NbPO-5, NbPO-7 and NbPO-10. The Nb2O5 sample purchased from
Sigma-Aldrich was also used for comparative studies. In addition, 1.0 wt.% Ru/C catalyst was
prepared by incipient wetness impregnation with ruthenium chloride (ShanXi KaiDa Chemical
Engineering CO., LTD) on active carbon. After impregnation, the activation was carried out in tube
furnace under flowing H2 at 523 K for 5 h.
2.2. Catalytic Reaction. The conversion of xylose was performed in a 100 ml stirred autoclave. In a
typical experiment, 200 mg reagent, 200 mg acid catalyst, 50 mg hydrogenation catalyst, and 32 g
7

solvent were loaded into the vessel, respectively. The aqueous phase and organic solvent were used
in the mass ratio of 1:1. After sealed and pressurized with H2 to 3.0 Mpa, the reactor was heated to
423 K and stirred for 4 h, then the autoclave was cooled in an ice-water bath. Products were filtered
through a membrane filter (0.2 μm pore size) prior to analysis. Xylose and xylitol were analyzed on
an Agilent 1200 Series HPLC equipped with a refractive index detector with a Bio-Rad HPX-87H
column using an aqueous solution of H2SO4 at 5 mmol/L as mobile phase (0.7 ml/min). 1,2-PeD and
other products ( 1,4-PeD, CPO, CPL, FF, 1-H-2-P, THFA, GVL, n-pentyl alcohol, and sec-pentanol)
were quantified by FID-GC (Shimadzu 2014C with a column of DB-FFAP, 30 m). The conversion
of xylose, the selectivity and yield of products were determined from Eq. (1), Eqs. (2), and Eqs. (3),
respectively.
[xylose] −[xylose]
퐢
Conversion=
Selectivity=
퐟 × 100% (1)
× 100% (2)
(3)
[xylose]
퐢
[product]
[xylose] −[xylose]
퐢
퐟
Yield= [product] × 100%
[xylose]
퐢
Where [xylose]i and [xylose]f represented the initial and final molar quantity of xylose, respectively.
[product] represented the molar quantity of detected products by GC and HPLC. The reproducibility
of the one-pot conversion process was tested using NbPO-10 as acid catalyst. Two repetitions were
carried out and the results were shown in Table 2.
2.3. Isotope-Labeling Experiment. The experimental procedures were similar to that of common
reactions. Typically, 200 mg of xylose was dissolved in 16 g of D2O and 16 g of cyclohexane, then
the Bruker Avance 600-MHz Spectrometer was used to determine the incorporation of D atoms in
the produced furfural. The D content of each carbon was calculated from the added and measured
band areas of furfural, which was normalized by dimethyl sulfone.
8

2.4 Catalyst Characterization. The N2 adsorption-desorption isotherms of the samples were
determined by the ASAP-2020 analyzer at liquid nitrogen temperature. The surface area, pore
volume and average pore size were respectively calculated using the Brunauer–Emmett–Teller (BET)
and Barrett-Joyner-Halenda (BJH) methods. X-ray diffraction (XRD) analysis was performed on a
Bruker D8 advance instrument using Cu Kα with 0.03°(2θ) steps over 10-80°. The morphologies of
the as-prepared samples were measured by a field emission Hitachi S-4800 scanning electron
microscope (SEM). The Fourier transform infrared (FT-IR) absorption spectra were recorded on a
NICOLET 6700 spectrometer, with 32 scans at an effective resolution of 4 cm-1.
Temperature programmed desorption of ammonia (NH3-TPD) was analyzed on a Micromeritics
AutoChem II 2920 equipped with a thermal conductivity detector. For each experiment, the samples
(100 mg) were pretreated at 573 K at 10 K/min for 1 h and cooled to 313 K in Ar gas atmosphere (30
ml/min) before exposed to NH3 (10% NH3/Ar, 30 ml/min) until surface saturation. The weakly
adsorbed NH3 was removed by flushing Ar (30 ml/min) for 0.5 h. After the baseline became stable,
then desorption was programmed at 1073 K at 10 K/min.
Infrared spectra of pyridine adsorption (Py-IR) were recorded to determine the type and ratio of
acid sites with NICOLET 6700 spectrometer. The samples were pressed into thin self-supported
wafers, then placed in a sample cell coupled to a glass-circulation system. The samples were
dehydrated in situ at 673 K for 1 h under vacuum (1×10-3 Pa). After the temperature was cooled
down to 313 K, the spectrum was recorded as the background. Then pyridine molecules were
adsorbed, the spectra presented were recorded after the excess pyridine was desorbed under vacuum
for 1 h.
The X-ray photoelectron spectroscopy (XPS) spectra were acquired with a Thermo Escalab
250XI spectrometer with AlKα (1486.6 eV) as the X-ray source. Analysis chamber pressure was less
9

than 5×10-9 Torr. The energy step sizes of 0.1 eV were chosen for survey spectra. All binding
energies scale was calibrated using the Cꢀ1s peak of adventitious carbon on the analyzed sample
surface at 284.8 eV as the reference according. Surface atomic concentrations were calculated by
integrating the area under the curve after subtraction of the background.
3. Results and discussion
3.1 Characterization of catalysts
Fig. 1. XRD (A), N2 adsorption-desorption isotherms (B) and FT-IR (C) patterns of (a) Nb2O5, (b)
NbPO-2, (c) NbPO-5, (d) NbPO-7, (e) NbPO-10.
All the samples except Nb2O5 had similar XRD patterns, as shown in Fig. 1A. Nb2O5 was used
as a reference catalyst due to its high Lewis acidity, which would be illustrated in detail below. The
characteristic peaks of Nb2O5 at 2θ between 10o-60o were assigned to the monoclinic niobium oxide
without peaks belonging to other phases according to the corresponding JCPDS card (PDF 27-1311).
The NbPO samples in this study were confirmed to be amorphous by the presence of broad
diffraction peaks, which was in accordance with the previous work [30]. The characterization by N2
adsorption and EDS (Energy Dispersive Spectrometer) analysis to detect texture properties were
shown in Fig. 1B and Table 1. A decrease in BET surface area, pore diameter and pore volume were
observed when pH value varied from 2 to 10. In addition, as revealed by the bulk composition in
10

Table 1, a higher ratio of niobium to phosphorus was achieved for the final calcined products at
higher pH value despite the same ratio of niobium to phosphorus was controlled in the precursor
solution. It was reported that the niobium was assembled together to form precipitation instead of
interacting with surfactant at higher pH value, thus leading to the decrease in surface area, pore
diameter and pore volume, which could be also revealed by SEM images [29]. As seen in Fig. 2, the
morphology of the NbPO-2 and NbPO-5 samples was needle-like particles, but the NbPO-7 and
NbPO-10 samples were several micrometer particles without special morphology.
Fig. 2. SEM images of (a) Nb2O5, (b) NbPO-2, (c) NbPO-5, (d) NbPO-7, (e) NbPO-10.
The FT-IR spectra used for functional group characterization were shown in Fig. 1C. Peaks at
1020, 1630 and 2349 cm-1 could be clearly observed in all NbPO samples, but which were absent in
Nb2O5. The bands at 2349 and 1630 cm-1 were assigned to the stretching and deforming vibrations of
the (P)-OH group and the -OH group, respectively [31-33]. The bands at 1020 cm-1 corresponded to the
stretching vibration of the Nb-O-P network [29], whereas the ratio of its peak area to that of Nb-O at
630 cm-1 was observed decreasing from NbPO-2 to NbPO-10. In addition, a broad adsorption
centered around 689 cm-1 was found in the spectrum of Nb2O5 as well as NbPO samples due to the
Nb-O stretching vibration mode. The phenomenon was associated with niobium and phosphorus
contents detected by EDS analysis.
11

Fig. 3. NH3-TPD spectra (A) and Py-IR spectra (B) on (a) Nb2O5, (b) NbPO-2, (c) NbPO-5, (d)
NbPO-7, (e) NbPO-10.
The acid properties of the catalysts were studied by NH3-TPD and Py-IR. Fig. 3A showed the
results of NH3-TPD experiments where all samples presented broad profiles in the range of 393-900
K, indicating the coexistence of weak, moderate and strong acid sites [33]. In addition, the acid
properties were also quantificationally compared by NH3 uptake and acid density (Table 1). Nb2O5
exhibited the lowest acidity, while the acid amount as well as strength of NbPO samples decreased
from NbPO-2 to NbPO-10, as made evident by the significant changes in NH3 uptake and the
maximized desorption temperature. The Nb2O5 sample, seen in the inset of Fig. 3A, presented a
broad peak between 523 K and 773 K due to the medium and strong solid acidity. As for NbPO
samples, the medium strength acid sites in the range of 573-773 K, attributed to the terminal P(OH)2
sites and surface P-OH sites [34], were much higher for NbPO-2 than those of NbPO-5, NbPO-7 and
NbPO-10. It showed clearly that the acid properties of NbPO samples altered significantly with the
change of pH value. To further explore this, the surface compositions of the catalysts were analyzed
by XPS. Table 1 gave the surface concentration of niobium and phosphorus. Values given in the
table exclude carbon and nitrogen [35]. The surface ratio of Nb/P was slightly lower than that of bulk
ratio, indicating that surface depletion of niobium happened, however, the surface composition was
still a function of the bulk composition. As reported, Nb2O5 has been receiving increasing interest as
12

solid Lewis acid catalyst for sugar dehydration due to its activity in water [36]. In addition,
phosphorus could also act as Brønsted acid in the presence of OH groups. Thus the aforementioned
change of surface composition could significantly affect the acidic species of NbPO samples. The
individual concentrations of Brønsted and Lewis acid sites on NbPO samples were further quantified
using Py-IR (Fig. 3B). The two absorption peaks located at 1445 and 1540 cm-1 were due to pyridine
molecules interacting with the Lewis and Brønsted acid sites, respectively [37]. The areas of these two
peaks were used to quantify the number of acid sites, and thus the ratios of Lewis to Brønsted acid
site for NbPO samples were listed in Table 1. Obviously, the ratio of the Lewis to Brønsted acid site
increased with the ratio of Nb/P, which can be attributed to the increase of surface niobium with high
Lewis acidity (inset of Fig. 3B).
3.2 Optimization of reaction conditions
One-pot conversion of xylose under hydrogen pressure was carried out in the biphasic system to
evaluate the effect of reaction condition. In order to avoid the difference induced by various batches
of NbPO samples, the commercialized Nb2O5 was used as an acid catalyst for optimization
experiments. Such a reaction in coexistence of acid and hydrogenation catalysts was rather complex
since a multitude of compounds could be found, from sugar alcohol and diols to furanic compounds.
Owing to thermodynamically beneficial to sugar hydrogenation, the hydrogenation of xylose to
xylitol was preferred in the presence of hydrogenation catalyst. We have significantly suppressed the
xylose hydrogenation in biphasic system using with the kinetic control strategy. Thus the dehydration
of xylose, on the other hand, could occur over an acid catalyst and several sugar intermediates were
generated. The sugar intermediate could not be captured due to its instability, but it was further
dehydrated to produce furfural on acid sites or transformed to 1,2-PeD under the synergistic catalysis
of acid and hydrogenation sites. In addition, the subsequent transformation of furfural could also
obtain a variety of products such as 1,4-PeD, cyclopentanone (CPO), cyclopentanol (CPL),
13

tetrahydrofurfuryl alcohol and humins depending on the reaction pathway of C-C bond and C-O
bond. For example, ring-opening hydrogenolysis yielded 1,4-PeD, the rearrangement of furan ring
yielded CPO and CPL, the saturation of C=C bond yielded tetrahydrofurfuryl alcohol, and humins
was formed by polymerization reaction. Therefore, the influence of reaction conditions on product
distribution should be noteworthy in this complicated system.
Fig. 4. One-pot conversion of xylose in biphasic system in function of reaction temperature (A),
reaction time (B) and mass ratio of hydrogenation catalyst to acid catalyst (C). (■) xylose conversion;
(●) selectivity of xylitol; (▲) selectivity of 1,2-PeD; (▼) the sum of CPO selectivity and CPL
selectivity; (◆) selectivity of THFA; Reaction conditions: 3.0 Mpa of H2 pressure, 4 h of reaction
time, 0.2 g of xylose, 16.0 g of water, 16.0 g of cyclohexane, 0.2 g of Nb2O5.
As displayed in Fig. 4A, the conversion of xylose was only 11.7% with abundant xylitol while a
little 1,2-PeD, 1,4-PeD, CPO and CPL being obtained at 383 K. With the increase of reaction
temperature to 403 K, the conversion of xylose, selectivities of 1,2-PeD, CPO, CPL and THFA all
increased along with significant decrease of xylitol selectivity. Further increased the reaction
temperature, the 1,2-PeD selectivity increased first and then decreased, and the highest selectivity
was obtained at 423 K. The 1,2-PeD selectivity passed through a maximum with time, implying the
participation of 1,2-PeD in side reactions. However, the selectivities of 1,4-PeD, CPO and CPL
14

gradually increased with the increased temperature. The selectivity of xylitol changing in the
opposite trend within the whole range of temperature indicated that the hydrogenation of sugar
intermediates and furfural were kinetically favored over xylose reduction to xylitol when increasing
reaction temperature.
In the Fig. 4B, the evolution of the conversion and product selectivity were shown as a function
of reaction time. As expected, the xylose conversion increased with the increase of reaction time. The
slight decrease of xylitol selectivity was observed probably because of the inactivation of
hydrogenation catalyst or the involvement of other side reactions. More specifically, 5.6% yield of
xylitol was obtained after only 1 h reaction and gradually reached the yield of 8.5% with the increase
of reaction time to 6 h. 63.6% of xylitol was produced during the first one hour. This indicated that
the hydrogenation of xylose occurred at the early stage of the reaction. In contrast, the selectivities of
1,2-PeD, THFA and CPO all increased continuously as a function of reaction time, but the highest
1,2-PeD selectivity was obtained after 4 h of reaction time. Further prolonged the time, 1,2-PeD
would be converted to pentanol, thus decreasing the 1,2-PeD selectivity [9].
As the aforementioned reaction scheme, the various products except xylitol was generated by
the subsequent hydrogenation of xylose dehydration products, each of these processes proceeding
over both acid and hydrogenation catalysts. The adjustment of acid and hydrogenation sites in the
dual catalyst system could tune the product distribution. Therefore, the impact on product selectivity
with increasing mass of hydrogenation catalyst was investigated (Fig. 4C). A remarkable difference
in the conversion of xylose, selectivity of xylitol, selectivity of 1,2-PeD and the selectivity of THFA
was noticed, with selectivity varying up to 3-14 times. In the case of 1,2-PeD, the selectivity was 1.8%
and obtained 24.3% when the hydrogenation catalyst was quadrupled. The results indicated the
formation of hydrogenation products were meaningfully affected, rising with the concentration of
hydrogenation sites in the reactor. Accordingly, the intensification of hydrogenation process was
15

accomplished by using γ-valerolactone-water/cyclohexane biphasic system [20]. A higher 1,2-PeD
selectivity was obtained with almost no added hydrogenation catalyst. Hence, the optimized reaction
conditions were obtained and γ-valerolactone-water/cyclohexane system was chosen in this work for
further studies.
3.3 Catalytic performance of catalysts
Besides the remarkable differences in product selectivity induced by reaction conditions, the
influence of acid catalyst was also proved to be noteworthy. Table 2 showed the results of xylose
conversion achieved for different catalysts. The detected pentanol was not shown in Table 2 due to
the small quantity (< 0.3%). Brønsted and Lewis acid catalysts were respectively used as an acid
catalyst for one-pot conversion of xylose, the target product 1,2-PeD both appeared, which excluded
the uniqueness of acid species for producing 1,2-PeD. Xylose was almost completely consumed
(95.7%) after 4 h over phosphoric acid, while only 58.6% of xylose was converted due to the low
acidity of Nb2O5, however, the selectivity of 1,2-PeD was slightly enhanced. In the case of NbPO-2
sample with high acidity, the catalytic activity was indeed improved and 1-hydroxyl-2-pentanone
(1-H-2-P) appeared, which was considered as the precursor of 1,2-PeD. Subsequently, the catalytic
behavior of NbPO-5, NbPO-7 and NbPO-10 catalysts was also evaluated in this system, the product
selectivities were seen to be quite distinct. Before analyzing these results, we should remember well
that xylitol was obtained from the direct hydrogenation of xylose; 1,2-PeD and 1-H-2-P were from
the hydrogenation of sugar intermediates; 1,4-PeD, THFA, CPO and CPL were derived from the
hydrogenation of produced furfural. Hence, xylose hydrogenation, sugar intermediate hydrogenation
and furfural hydrogenation were the three main hydrogenation processes involved in the complicated
one-pot conversion process. In order to illustrate the effect of acid catalyst on the product selectivity,
the relationship between the product yield and the surface Lewis/Brønsted ratio were analyzed. As
shown in Fig. 5, the product yields of sugar intermediate and furfural hydrogenation processes both
16

increased, while the product yield of xylose hydrogenation process decreased approximately linearly
with the surface Lewis/Brønsted ratio, as will be discussed in detail below.
Fig. 5. Linear fitting of the product yield with the surface Lewis/Brønsted ratio; (●) xylose
hydrogenation process, the yield of xylitol; (▲) sugar intermediate hydrogenation process, the total
yield of 1,2-PeD and 1-H-2-P; (▼) furfural hydrogenation process, the total yield of 1,4-PeD, CPO,
CPL and THFA.
3.4 Discussion
Fig. 6. Arrhenius plots for xylose conversion; (□) xylose hydrogenation over Ru/C in aqueous system;
(○) xylose dehydration over Nb2O5 in aqueous system; (◊) xylose hydrogenation over Ru/C in
water-cyclohexane biphasic system; (■) one-pot conversion over Nb2O5; (▼) one-pot conversion
17

over NbPO-2; (◄) one-pot conversion over NbPO-5; (▲) one-pot conversion over NbPO-7; (●)
one-pot conversion over NbPO-10; (♦) one-pot conversion over H3PO4.
Extensive studies for biomass upgrading have been carried out by many groups to explore
efficient transforming strategies such as retro-aldol condensation, C-O hydrogenolysis, and
dehydration-hydrogenation reactions [18, 38, 39]. As for the conversion of xylose to 1,2-PeD, only the
multistep dehydration-hydrogenation reactions with furfural as the reaction intermediate were
achievable. More recently, an efficient one-pot conversion strategy in acid-metal dual catalyst system
was developed. However, whether the process was multistep or one-pot, dehydration and
hydrogenation reactions were the core steps in the network for xylose conversion. The assembly of
these two type reactions was the challenge of obtaining high 1,2-PeD selectivity. As displayed in
scheme 1, xylose could be hydrogenated to xylitol or dehydrated to sugar intermediates over Ru/C or
H3PO4 catalysts. However, which was the preferred reaction in the coexistence of these two catalysts?
The obtained apparent activation energies derived from the Arrhenius plots were respectively 13.2
KJ/mol and 56.7 KJ/mol for xylose hydrogenation and dehydration processes, revealing that xylose
hydrogenation process was more thermodynamically beneficial. Furthermore, the result that 95.5 %
selectivity of xylitol was obtained in aqueous system verified xylose hydrogenation being also
dynamically beneficial. Therefore, the primary challenge to obtain 1,2-PeD was the xylose
hydrogenation to xylitol, which was inert for the formation of 1,2-PeD in our system. In our previous
work, the biphasic system showed distinct features in assembling dehydration and hydrogenation
reactions, which was beneficial for the formation of 1,2-PeD. 34.6% yield of 1,2-PeD has been
developed using kinetic control strategy [20]. The mass transfer resistance due to low xylose solubility
of cyclohexane could significantly suppress xylose hydrogenation to xylitol, which was reflected in
the apparent activation energy. As expected, the apparent activation energy increased to 24.8 KJ/mol
in water-cyclohexane system. The results suggested that suppressing xylose hydrogenation reaction
18

could promote xylose being converted via dehydration process, thus improving the selectivity of
1,2-PeD.
After analyzing the advantage of biphasic system for assembling dehydration and
hydrogenation reactions, we emphasized considering the impact of acid catalyst in dual catalytic
system and gave an outlook for the product distribution. To explore this challenge, a first-order
reaction rate expression to the biphasic xylose conversion kinetics was also fitted, the apparent
activation energy was estimated to be 47.6 KJ/mol, considerably higher than those of xylose
hydrogenation in aqueous and biphasic systems, but approaching to that of the xylose dehydration
process (49.4 KJ/mol). Since the reaction mechanism of one-pot conversion of xylose to 1,2-PeD
was not well understood, it remains a challenge to explain the observed differences in the apparent
activation energy. However, in view of apparent activation energy, such results clearly indicated that
the energy barrier for one-pot conversion process was from xylose dehydration reaction. This
assumption agreed well with the observed remarkable difference in product selectivity induced by
acid catalyst shown in Table 2 and Fig. 5. Hence, the catalytic ability of acid catalyst was the notable
factor to adjust the product distribution. According to the aforementioned results, the NbPO samples
were equipped with different amount of surface Lewis and Brønsted sites, and the 1,2-PeD as well as
other hydrogenation product yields all increased linearly with the surface Lewis/Brønsted ratio
detected by Py-IR analysis. Recently, Lewis acid has been determined as a more active and selective
catalyst than homogeneous Brønsted catalysts [40]. For example, a cascade of reactions for xylose
dehydration using Lewis catalyst plus Brønsted catalyst was more efficient [41], probably ascribed to
the lower activation energy due to the difference in reaction mechanism [40]. However, caution was
required for associating the catalytic performance with the acid properties characterized by Py-IR
technology when considering the evolution of active site under reactive environment. Many Lewis
acid sites of inorganic oxides were generally regarded as inactive sites for reactions in water, due to
[36]
the formation of Lewis-base adducts by the coordination of water to the Lewis acid sites
.
19

Although NbO4 tetrahedra of niobium oxide could function as Lewis acid sites in water, a part of
surface OH groups which acted as Brønsted acid sites were generated in these distorted polyhedrons
[36]. These transformations eventually affected the contribution of acid sites in water.
Fig. 7. Linear fitting of the product yield with the content of H/D exchange at C1 position; (○) xylose
hydrogenation process, the yield of xylitol; (□) sugar intermediate hydrogenation process, the total
yield of 1,2-PeD and 1-H-2-P; (◊) furfural hydrogenation process, the total yield of 1,4-PeD, CPO,
CPL and THFA;(■) the apparent activation energy over NbPO-2; (◄) the apparent activation energy
over NbPO-5; () the apparent activation energy over NbPO-7; (♦) the apparent activation energy
over NbPO-10.
To further clarify the practical impact of Lewis and Brønsted acid sites under reactive
environment, the H/D exchange experiments in the cyclohexane-D2O system were evaluated to
demonstrate the contribution of different acid sites to xylose dehydration. According to dehydration
mechanism shown in Table 3 ( detailed mechanism could be found in reference [40, 42-44]), the H/D
exchange at C1 and C3 positions of the resultant furfural could be used as an indicator for elucidation
the contribution of Lewis and Brønsted acid sites for dehydration process. Notably, a content of 21.5%
deuterium atom was observed only at the C3 position over H3PO4, which coincided with the
aldose-type mechanism over Brønsted acid. In the case of NbOP-2 sample with both Lewis and
20

Brønsted acid sites, furfural with 10.8% and 35.1% deuterium incorporation at C1 and C3 positions
were respectively observed, where the H-D exchange at the C1 position was catalyzed by Lewis acid
site according to stepwise dehydration mechanism. It appeared that both Lewis and Brønsted acid
sites were involved in H-D exchange at the C3 position, and only Lewis acid site was involved at the
C1 position. Hence, the H-D exchange at the C1 position could illustrate the contribution of Lewis
acid site. Table 3 displayed the content of deuterium at the C1 position for all NbPO samples. The
results revealed that the contribution of Lewis acid site for xylose dehydration agreed well with the
acid properties detected by Py-IR. Then the correlation between the product yield and the content of
H/D exchange at the C1 position was also analyzed (Fig. 7). Three inverse linear correlations were
observed, revealing that Lewis acidity favored the high selectivities to sugar intermediate and
furfural hydrogenation derived products such as 1,2-pentanediol, 1-hydroxyl-2-pentanone,
1,4-pentanediol, tetrahydrofurfuryl alcohol, cyclopentanone and cyclopentanol. Furthermore, the
apparent activation energies of one-pot conversion over these catalysts were compared (see Table 2
and Fig. 7), it could be found that the apparent activation energies decreased with the increase of
Lewis/Brønsted ratio. Hence, the advantage of Lewis acid site was reducing the activation energy of
xylose dehydration, thus improving the selectivity of 1,2-PeD and its precursor.
4. Conclusions
The conversion of xylose in a one-pot process by using Ru/C and niobium phosphate dual
catalyst system was shown. We have prepared and characterized a series of niobium phosphate
catalysts with different surface Lewis and Brønsted acid sites and emphatically tested their effect on
product distribution. Bulk and surface composition detected by EDS and XPS analysis revealed a
higher niobium content for the niobium phosphate prepared at higher pH value, thus increasing the
21

ratio of surface Lewis to Brønsted acid site. When the ratio reaches 2.5, 14.0% selectivity of
1,2-pentanediol and 13.3% selectivity of 1-hydroxyl-2-pentanone were obtained at complete xylose
conversion. We further demonstrated the contribution of Lewis and Brønsted acid sites to xylose
dehydration under reactive environment using H-D exchange experiments, which agreed well with
the acid properties obtained by NH3-TPD and Py-IR technologies. Then the correlation between the
product yield and the content of H/D exchange at the C1 position was analyzed. The results revealed
that catalysts with high Lewis acidity favored the high selectivities to 1,2-pentanediol and
1-hydroxyl-2-pentanone, but disfavored the selectivity to xylitol. This was due to the lower apparent
activation energy of xylose dehydration reaction catalyzed by Lewis acid site.
Acknowledgements
This work was supported by the National Natural Science Foundation of China [grant No. 21676288]
22

References
[1] I.C. Agostini, S, USA 6296858, October 2, 2001.
[2] H.T. Nakamura, Y. , JP 201236121, February 23, 2012.
[3] P.G. CHEN G, ZENG T, CN 102584802-A, July 18, 2012.
[4] W. Y, CN 106397112-A, February 15, 2017.
[5] R. Ma, X.-P. Wu, T. Tong, Z.-J. Shao, Y. Wang, X. Liu, Q. Xia, X.-Q. Gong, ACS Catal., 7 (2017), pp. 333-337.
[6] G.R. Jenness, W. Wan, J.G. Chen, D.G. Vlachos, ACS Catal., 6 (2016), pp. 7002-7009.
[7] Y. Nakagawa, K. Tomishige, Catal. Today, 195 (2012), pp. 136-143.
[8] T. Mizugaki, T. Yamakawa, Y. Nagatsu, Z. Maeno, T. Mitsudome, K. Jitsukawa, K. Kaneda, ACS Sustain. Chem. Eng.,
2 (2014), pp. 2243-2247.
[9] M. Chia, Y.J. Pagán-Torres, D. Hibbitts, Q. Tan, H.N. Pham, A.K. Datye, M. Neurock, R.J. Davis, J.A. Dumesic, J.
Am. Chem. Soc., 133 (2011), pp. 12675-12689.
[10] H. Liu, Z. Huang, F. Zhao, F. Cui, X. Li, C. Xia, J. Chen, Catal. Sci. Technol., 6 (2016), pp. 668-671.
[11] T. Tong, Q. Xia, X. Liu, Y. Wang, Catal. Commun., 101 (2017), pp. 129-133.
[12] B. Zhang, Y. Zhu, G. Ding, H. Zheng, Y. Li, Green Chem., 14 (2012), pp. 3402.
[13] S. De, S. Dutta, B. Saha, ChemSusChem, 5 (2012), pp. 1826-1833.
[14] G.A. Kraus, T. Guney, Green Chem., 14 (2012), pp. 1593-1596.
[15] M.L. Ribeiro, U. Schuchardt, Catal. Commun., 4 (2003), pp. 83-86.
[16] S.J. Canhaci, R.F. Perez, L.E.P. Borges, M.A. Fraga, Appl. Catal. B-Environ., 207 (2017), pp. 279-285.
[17] R.F. Perez, M.A. Fraga, Green Chem., 16 (2014), pp. 3942-3950.
[18] V.V. Ordomsky, J.C. Schouten, J. van der Schaaf, T.A. Nijhuis, Appl. Catal. A-Gen., 451 (2013), pp. 6-13.
[19] M. Zheng, J. Pang, R. Sun, A. Wang, T. Zhang, ACS Catal., 7 (2017), pp. 1939-1954.
[20] N.L. Wang, Z.P. Chen, X.D. Mu, L.C. Liu, ChemSusChem, (2018).
[21] J. Lee, Y. Xu, G.W. Huber, Appl. Catal. B-Environ., 140–141 (2013), pp. 98-107.
[22] M.T. Reche, A. Osatiashtiani, L.J. Durndell, M.A. Isaacs, A. Silva, A.F. Lee, K. Wilson, Catal. Sci. Technol., 6
(2016), pp. 7334-7341.
[23] Y. Shao, Q. Xia, L. Dong, X. Liu, X. Han, S.F. Parker, Y. Cheng, L.L. Daemen, A.J. Ramirez-Cuesta, S. Yang, Y.
Wang, Nat. Commun., 8 (2017), pp. 16104.
[24] G.S. Foo, D. Wei, D.S. Sholl, C. Sievers, ACS Catal., 4 (2014), pp. 3180-3192.
[25] C. García-Sancho, I. Agirrezabal-Telleria, M.B. Güemez, P. Maireles-Torres, Appl. Catal. B-Environ., 152-153
(2014), pp. 1-10.
[26] K. Tanabe, Catal. Today, 78 (2003), pp. 65-77.
[27] B. Pholjaroen, N. Li, Z. Wang, A. Wang, T. Zhang, J Energy Chem., 22 (2013), pp. 826-832.
[28] K. Tanabe, S. Okazaki, Appl. Catal. A-Gen., 133 (1995), pp. 191-218.
[29] Y. Zhang, J. Wang, J. Ren, X. Liu, X. Li, Y. Xia, G. Lu, Y. Wang, Catal. Sci. Technol., 2 (2012), pp. 2485-2491.
[30] R. Weingarten, Y.T. Kim, G.A. Tompsett, A. Fernández, K.S. Han, E.W. Hagaman, W.C. Conner, J.A. Dumesic, G.W.
Huber, J. Catal., 304 (2013), pp. 123-134.
[31] D. Yu, J. Qian, N. Xue, D. Zhang, C. Wang, X. Guo, W. Ding, Y. Chen, Langmuir, 23 (2007), pp. 382-386.
[32] D. Yu, C. Wu, Y. Kong, N. Xue, X. Guo, W. Ding, J. Phys. Chem. C., 111 (2007), pp. 14394-14399.
23

[33] F. Li, L.J. France, Z. Cai, Y. Li, S. Liu, H. Lou, J. Long, X. Li, Appl. Catal. B-Environ., 214 (2017), pp. 67-77.
[34] R. Weingarten, Y.T. Kim, G.A. Tompsett, A. Fernández, K.S. Han, E.W. Hagaman, W.C. Conner, J.A. Dumesic, G.W.
Huber, J. Catal., 304 (2013), pp. 123-134.
[35] H.J.M. Bosman, A.P. Pijpers, A.W.M.A. Jaspers, J. Catal., 161 (1996), pp. 551-559.
[36] K. Nakajima, Y. Baba, R. Noma, M. Kitano, J. N. Kondo, S. Hayashi, M. Hara, J. Am. Chem. Soc., 133 (2011), pp.
4224-4227.
[37] F. Liu, T. Wang, Y. Zheng, J. Wang, J. Catal., 355 (2017), pp. 17-25.
[38] N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J.G. Chen, Angew. Chem. Int. Ed., 47 (2008), pp.
8510-8513.
[39] S. Liu, Y. Okuyama, M. Tamura, Y. Nakagawa, A. Imai, K. Tomishige, Green Chem., 18 (2016), pp. 165-175.
[40] N.K. Gupta, A. Fukuoka, K. Nakajima, ACS Catal., (2017), pp. 2430-2436.
[41] V. Choudhary, S.I. Sandler, D.G. Vlachos, ACS Catal., 2 (2012), pp. 2022-2028.
[42] M.J. Antal, T. Leesomboon, W.S. Mok, G.N. Richards, Carbohyd. Res., 217 (1991), pp. 71-85.
[43] T. Ahmad, L. Kenne, K. Olsson, O. Theander, Carbohyd. Res., 276 (1995), pp. 309-320.
[44] M.J. Antal, W.S.L. Mok, G.N. Richards, Carbohyd. Res., 199 (1990), pp. 91-109.
24

Tables:
Table 1. Summary of physicochemical properties of solid acid catalysts
Bulk compositiond
Surface compositione
Pore
size
(nm) (cm3/g)
Pore
NH3
Acid
SSAa
(m2/g)
Catalyst
volume
uptakes
(μmol)b
6.2
density
(μmol/m2)
2.4
L/Bc
Nb
--
P
Nb/P
Nb
P
Nb/P
Nb2O5
35
13.8
3.7
3.5
3.4
2.4
0.14
0.40
0.29
0.20
0.14
--
--
--
20.8 --
--
473.6
315.7
234.1
225.5
NbPO-2
NbPO-5
NbPO-7
427
336
267
11.0
9.4
8.8
9.9
0.4
0.7
0.8
2.5
10.9 13.7
11.9 12.1
13.5 9.2
14.1 7.4
0.8
1.0
1.5
1.9
7.4
8.6
14.9
12.4
0.5
0.7
1.1
1.3
11.8 10.6
12.3 9.3
NbPO-10 226
a: specific surface area; b; detected by NH3-TPD; c: detected by Py-IR; d: detected by EDS analysis; e: detected by
XPS analysis
Table 2. Catalytic results of one-pot conversion of xylose in aqueous biphasic systema
Selc. (%)
Mass
Ead
Con.
Acid
W/Ob
Balance
(KJ/mol)
(%)
Xylitol 1,2-PeD 1,4-PeD 1-H-2-P THFA CPO+CPL
(%)
H3PO4
W/Cyclo
W/Cyclo
95.7
58.6
0.3
2.9
1.6
0.0
2.3
4.7
5.1
6.5
9.1
8.3
10.4
0.0
15.6
3.3
2.1
7.9
8.6
8.8
9.2
5.5
6.7
1.0
1.3
1.2
1.4
2.0
29.1
60.5
82.7
75.9
73.4
72.2
72.8
73.3
72.4
53.9
47.6
--
Nb2O5
13.5
9.1
0.0
Nb2O5
W-GVL/Cyclo 100.0 70.6
W-GVL/Cyclo 100.0 40.4
W-GVL/Cyclo 100.0 36.2
W-GVL/Cyclo 100.0 32.2
W-GVL/Cyclo 100.0 25.0
W-GVL/Cyclo 100.0 26.0
W-GVL/Cyclo 100.0 27.2
6.5
0.0
NbPO-2
NbPO-5
NbPO-7
NbPO-10
NbPO-10e
NbPO-10f
16.7
17.5
19.1
14.0
15.1
13.5
4.8
51.6
50.5
49.5
48.7
--
4.8
4.1
13.3
12.0
11.8
10.8 1.0
8.1 1.3
--
a: Reaction conditions: 0.2 g of xylose; 1.5 mmol H3PO4; 200 mg of solid acid catalyst; 50 mg of Ru/C catalyst; T = 423
K; t = 4 h; H2 pressure: 3.0 Mpa; b: W represented aqueous phase; O represented organic phase; CYH and GVL were
respectively abbreviation of cyclohexane and γ-valerolactone; 16 g of H2O and 16 g of cyclohexane were added in W/CYH
system; 4.8 g of H2O, 11.2 g of γ-valerolactone and 16 g of cyclohexane were added in W-GVL/CYH system; c: 1,2-PeD:
1,2-pentanediol; 1,4-PeD: 1,4-pentanediol; FF: furfural; 1-H-2-P: 1-hydroxyl-2-pentanone; THFA: tetrahydrofurfuryl
alcohol; CPO: cyclopentanone; CPL: cyclopentanol. d: apparent activation energy of one-pot conversion in
water-cyclohexane biphasic system. e and f:: the one-pot conversion of two repetitions were used to indicate the degree of
reproducibility.
25

Table 3. Catalytic results of xylose conversion in cyclohexane-D2O systema
D content (%)
C1
Con.
(%)
FFb
Cat
Sel. (%)
C3
H3PO4
100.0
99.8
99.4
91.5
88.6
43.4
57.6
57.1
65.2
70.5
86.2
47.4
0.0
21.5
35.1
37.5
40.6
43.9
50.9
NbPO-2
NbPO-5
NbPO-7
NbPO-10
Nb2O5
10.8
13.4
15.1
19.4
27.2
a: Reaction conditions: 0.2 g of xylose; 1.5 mmol H3PO4; 200 mg of solid acid catalyst; T = 423 K; t = 4 h; N2 pressure: 3.0
Mpa; 16 g of D2O and 16 g of cyclohexane were added in W/CYH system; b: furfural.
26