Carbohydrate Polymers 247 (2020) 116723

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

Dual roles of cellulose monolith in the continuous-flow generation and T

support of gold nanoparticles for green catalyst
Zheng-Tian Xie, Taka-Aki Asoh*, Yuta Uetake, Hidehiro Sakurai, Hiroshi Uyama*
Department of Applied Chemistry, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan

A R T I C LE I N FO A B S T R A C T

Keywords: Monolithic flow reactors are widely applied in numerous reactions due to its high efficiency and good reusa-
Cellulose monolith bility, but the green and efficient fabrication of monolithic flow catalytic system is still a challenge. Herein, the
Gold nanoparticles cellulose monolith prepared using a facile temperature-induced phase separation method was utilized to gen-
Continuous flow erate and immobilize the gold nanoparticles by a continuous-flow strategy, in which the cellulose monolith
Green catalyst
served as both reducing agent and supporting material. This process was conducted at room temperature and
avoided the tedious surface modification of cellulose. The obtained cellulose-Au monolith can be directly applied
as a green flow reactor in both water and organic solvents, and exhibited superior catalytic efficiency and good
stability. This work provides a highly efficient, scalable and sustainable strategy for developing green catalytic
system based on environmentally friendly cellulose monolith materials.

1. Introduction the chemical composition of the flow reactor and may introduce im-
purities in the final catalytic products. Recently, directly reducing metal
Monolithic porous materials have been rapidly developed and ions by the support materials has been considered as a reasonable
triggered increasing attention in the flow catalytic applications due to strategy, as it can combine the reduction, stabilization, and support into
their special three-dimensionally connected porous structure char- one process and eliminate the addition of a reducing agent and stabi-
acterized by both high surface area and good permeability (Li et al., lizer (Koczkur, Mourdikoudis, Polavarapu, & Skrabalak, 2015; Moitra
2017; Liang et al., 2015; Liguori et al., 2017; Pélisson et al., 2016; et al., 2013; Nakao et al., 2003; Turkevich, Stevenson, & Hillier, 1951;
Viklund, Svec, Fréchet, & Irgum, 1996; Yu et al., 2017). Numerous Van Rie & Thielemans, 2017; Wu et al., 2014a). This strategy also
recent efforts Bronstein, Polarz, Smarsly, & Antonietti (2001), Han et al. minimizes the remaining chemicals attached to the MNP surface after
(2019), Heiligtag et al. (2014), Moitra et al. (2013), Pelisson et al. immobilization and reduces the effect of impurities on the catalytic
(2017), Xie, Asoh, & Uyama (2019), Yang et al. (2005) have been at- performance of metal NPs. Indeed, some materials such as hydrogen
tempted to decorate the metal nanoparticles (NPs) on monoliths to silsesquioxane (Moitra et al., 2013), aniline (Nakao et al., 2003), citrate
develop high-efficiency flow catalytic systems. However, the current (Turkevich et al., 1951), chitosan (Bhumkar, Joshi, Sastry, &
strategies for preparing monolithic reactors usually involve a tedious Pokharkar, 2007; Shih, Shieh, & Twu, 2009) and cellulose (Van Rie &
preparation process or the use of toxic chemicals, which leads to in- Thielemans, 2017; Wu et al., 2014a) have already been proven to serve
creased production cost and environmental pollution (Bedford, Singh, as both reducing and stabilizing agents in the preparation of hetero-
Walton, Williams, & Davis, 2005; Hadjiivanov & Klissurski, 1996; Jal, geneous metal catalysts. Among them, cellulose as the most abundant
Patel, & Mishra, 2004; Macanás et al., 2010; White, Luque, Budarin, biomass in nature (Schurz, 1999), is considered as one of the most
Clark, & Macquarrie, 2009). From the green chemistry perspective, the promising supporting materials because of its inexpensive, biodegrad-
ecofriendly and efficient fabrication of a metal NPs-supported mono- able, nontoxic and chemical-resistant characteristics (Kaushik &
lithic catalytic system is still a challenge. Moores, 2016).
To stably anchor the metal NPs, the monolith support usually needs The current state of synthesis of gold NPs using cellulose as a re-
to be modified with functional groups such as amino and thiol groups ducing agent usually involves the reaction conditions such as introdu-
(Ashraf, Sher, Khalid, Mehmood, & Hussain, 2014; Jiang et al., 2018; cing the supercritical CO2 (Benaissi, Johnson, Walsh, & Thielemans,
Kaushik & Moores, 2016; Xie et al., 2019), but the modification changes 2010) or steps such as hydrothermal treatment (Wu et al., 2013,

⁎
Corresponding authors.
E-mail addresses: [email protected] (T.-A. Asoh), [email protected] (H. Uyama).

https://doi.org/10.1016/j.carbpol.2020.116723
Received 4 May 2020; Received in revised form 29 June 2020; Accepted 1 July 2020
Available online 03 July 2020
0144-8617/ © 2020 Elsevier Ltd. All rights reserved.
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

2014b). Moreover, most previous studies have been conducted on under vacuum. The cellulose monolith was obtained by hydrolyzing the
powdery or fibrous cellulose, and the obtained products can hardly be CA monolith in 0.5 mol/L NaOH/methanol solution for 3 h, followed by
used as continuous-flow reactors. To the best of our knowledge, no washing with deionized water (300 mL) and vacuum drying.
study has reported the use of porous cellulose sponges to directly re-
duce and support the metals to construct a green catalytic system. This 2.3. Preparation of gold NPs on the cellulose monolith
is probably because that most porous cellulose sponge is prepared based
on the crosslinking of the modified cellulose, such as carboxymethyl The as-prepared cellulose monolith was directly applied to reduce
cellulose (CMC) (Zhang et al., 2014) and 2,2,6,6-tetramethylpiperidine- HAuCl4 and anchor the formed gold NPs. First, the cellulose monolith
1-oxyl radical (TEMPO)-oxidized cellulose (Saito, Kimura, Nishiyama, with a length of 1 cm was fixed in a heat-shrinkable tube, and then, the
& Isogai, 2007). The modification treatment and the crosslinking pro- HAuCl4 aqueous solution (0.2 mmol/L, 5 mL, pH was adjusted to 13
cess would inevitably complicate the chemical properties of cellulose, with 1 mol/L aqueous NaOH solution) was pumped to circularly flow
making the investigation of the reaction between cellulose sponge and through the monolith by using a peristaltic pump. During the con-
metal ions complex and unreliable. tinuous flow, the gold ion was gradually reduced to gold NPs. After
The cellulose monolith, as a green material with a hierarchically flowed for 3 h, the sample was washed with deionized water (200 mL),
porous structure, was developed based on a facile temperature-induced and then dried under vacuum for 24 h to obtain the cellulose-Au
phase separation (TIPS) method in our previous work (Xin et al., 2017). monolith.
Unlike the cellulose sponge fabricated by the chemical crosslinking of As a comparison, the reduction experiment was also conducted in a
surface-modified cellulose, the cellulose monolith prepared by this batch fashion, in which the cellulose monolith was immersed in the
method has a simple chemical composition that remains the original HAuCl4 aqueous solution to react under stirring. Other reaction con-
chemical structure of cellulose, thus enabling the investigation of the ditions were the same as those in the flow fashion. After reaction for 3
chemical reaction between the monolith and metal ions. The excellent h, the sample was taken out and washed with deionized water, and then
mechanical properties and chemical resistance of the cellulose monolith dried under vacuum for 24 h to be tested.
also enable its use in long-term flow applications. Based on all these
special merits, we consider the cellulose monolith as a promising can- 2.4. Catalytic experiment
didate of supporting materials for directly reducing the metals to effi-
ciently develop green and high-efficiency flow catalytic systems. The obtained cellulose-Au monolith was fixed in a shrinkable tube
Herein, the monolithic cellulose was used to reduce gold ions and to construct the flow reactor (Fig. S1). The 3 mmol/L 4-nitrophenol
stabilize the formed gold NPs at room temperature without adding any aqueous solution and freshly prepared 1 mol/L aqueous solution of
other reducing agent or surfactant. This strategy realized a combination NaBH4 were mixed, and then immediately flowed through the reactor
of the green preparation of metal NPs and their facile immobilization on at a certain flow rate to react. After flowed through the monolith, the
the renewable cellulose materials by using a simple and moderate effluent was collected and diluted by 30 times with deionized water to
method. The product was applied as a continuous-flow reactor with a be tested by UV–vis measurement. The standard solutions of 4-ni-
fast processing capacity and efficient operability. This study provides a trophenol were previously tested to build the standard curve as a re-
green and cost-effective strategy for developing high-efficiency gold ference. According to the standard curve, the concentration of 4-ni-
supported flow reactors based on the environmentally sustainable cel- trophenol in both the influent and effluent was calculated. Then, the
lulose materials. catalytic efficiency was estimated by using the concentration change
and flow rate. In this experiment, the flow rate was corrected by the
2. Experimental section real volume of the effluent outflowing per unit time. To test the reu-
sability of the obtained monolithic reactor, this experiment was re-
2.1. Materials peatedly conducted with the same condition. The 4-nitrophenol re-
actant with concentration of 3 mmol/L was continuously flowed
The cellulose acetate powder (CA, L30) used in this experiment was through the monolith at a flow rate of 3 mL/min, and the conversion
supplied by DAICEL Co., Ltd. Japan, and its acetylation degree was was tested using the UV–vis measurement based on the above method.
approximately 55 %. Hydrogen tetrachloroaurate tetrahydrate In each cycle, 30 mL 4-nitrophenoal was treated. To avoid the effect of
(HAuCl4·4H2O) with a purity of 99 % was purchased from Nacalai the remained reactants in flow reactor, the flow reactor was washed by
Tesque, Inc. Kyoto, Japan. Sodium borohydride (NaBH4) was purchased 20 mL of deionized water before reuse.
from FUJIFILM Wako Pure Chemical Corporation, Japan. 4-Nitrophenol The catalytic performance of the reactor in an organic solution was
and dimethylphenylsilane (PhMe2SiH) were purchased from Tokyo evaluated by the oxidation reaction of PhMe2SiH to dimethylphenylsi-
Chemical Industry Co., Ltd. Methylene blue and methyl orange were lanol (PhMe2SiOH) in DMF. 0.1 mmol of PhMe2SiH and 25 μL of H2O
purchased from Tokyo Chemical Industry Co., Ltd. Other chemicals were dissolved in 3 mL of DMF, and the mixture was circularly flowed
including N,N-dimethylformamide, 1-hexanol and sodium hydroxide through the reactor (3 mL/min) to take reaction at room temperature.
were produced by FUJIFILM Wako Pure Chemical Corporation, Japan, After reacting for a certain time (0.5, 1, 2 and 3 h), 20 μL of the reactant
all of which were chemically purified and commercially available. The solution was taken out and dissolved in 1 mL of deuterated dimethyl
heat-shrinkage polyolefin tube (THT-8.0 N) for fabricating the flow sulfoxide (DMSO-d6, with 5 wt % of tetramethylsilane) to be tested by
1
reactors was provided by Denka Electron Co. Ltd., Japan. H NMR. To test the reusability, the same monolithic reactor was em-
ployed to repeat the reaction and the catalytic performance was eval-
2.2. Preparation of the cellulose monolith uated. After 1 h of reaction, 20 μL of the obtained reactant solution was
taken out and dissolved in 1 mL of DMSO- d6 to be tested by 1H NMR.
The porous cellulose monolith was fabricated according to our
previous method (Xin et al., 2017). First, the CA powder (2.5 g) was 2.5. Characterization
firstly dissolved in the 10 mL of DMF at 90 °C, and then 15 mL of 1-
hexanol was added dropwise to the CA solution at the same tempera- The chemical change of the monolith was investigated through
ture under stirring. After complete dissolution, the mixture was care- Fourier transformed infrared (FTIR) measurements on a Thermo
fully transferred into cylindrical molds to achieve phase separation in a Scientific (Yokohama, Japan) Nicolet iS5 spectrometer equipped with
water bath at 25 °C for 8 h. Subsequently, the porous CA monolith an iD5 ATR attachment. A UV–vis spectrophotometer (HITACHI, U-
formed in the molds was washed with methanol (300 mL) and dried 2910) was used to monitor the concentration of gold ions in the

2
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

preparation process of gold NPs. Before the measurement, the solution preparation conditions including the initial concentration of CA and the
was adjusted to pH = 1 using 1 mol/L HCl aqueous solution. The volume ratio of DMF/1-hexanol according to our previous works (Xie
concentration of UV specimen was determined according to the peak et al., 2019). The nitrogen adsorption/desorption isotherm experiments
intensity and calibrated based on the volume change caused by the pH (Fig. S2) demonstrated that the cellulose monolith had a mesopore
adjustment. The concentration change of 4-nitrophenol in the catalytic structure with a pore size of 5–15 nm according to the non-local density
reaction was also determined by UV–vis experiments. The absorbance functional theory (NLDFT) (Tarazona, 1985; Tarazona, Marconi, &
curve was collected at the wavelengths ranging from 250 to 600 nm. X- Evans, 1987). Such a hierarchical pore structure not only endows the
ray photoelectron spectroscopy (XPS, JEOL JPS-9010MC) was applied monolith with high surface area suitable for the efficient reduction of
to study the surface chemistry consistent of the monolith, using a Au3+ and immobilization of gold NPs on the monolith, but also allows
monochromatized Al-Kα radiation (1486.6 eV). The XPS spectra were it to be applied in the flow system. Moreover, the cellulose monolith has
collected by at fixed analyzer pass energies of 160 eV and 10 eV, re- a very simple chemical composition compared to the cellulose sponge
spectively. The residual pressure inside the analysis chamber was in the prepared by the crosslinking of surface-modified cellulose. All these
10−7 Pa range. The binding energies were referred to CeH (sp3) carbon advantages make it a promising supporting material for constructing a
for the C1s peak set at 284.8 eV. The morphologies of the monoliths green catalytic flow reactor. As shown in Fig. 1d, the cellulose monolith
were observed using a scanning electron microscope (SEM, Hitachi S- was applied to reduce the HAuCl4 without adding any reducing agent.
3000 N) at 15 kV. Before the observation, the samples were slivered In a typical experiment, a monolith with a length of 1 cm is fixed in a
into sections and coated with a thin layer of gold by using an ion sputter shrinkable tube and the aqueous gold ion solution is continuously
apparatus (E-1010, Hitachi Ltd.). In addition, the size of the gold NPs flowed through the monolith using a peristaltic pump. The color of the
and their dispersion state in the monolith were detected by transmis- monolith gradually changes to reddish in the flow process, indicating
sion electron microscopy (TEM, JEOL JEM-2010) at an acceleration the formation of gold NPs (Fig. 1e).
voltage of 100 kV. The sample was deposited on a copper grid coated To more specifically investigate this process, the concentration of
with carbon (NP-C15, Okenshoji, Co., Ltd.) for testing. The crystalline gold ions in solution was monitored by UV–vis measurement using the
states of the cellulose monolith with and without gold loading were characteristic peak of gold ions around 312 nm at pH = 1 (Figs. 2a and
studied through X-ray Diffraction (XRD) experiments conducted with S3). The results indicated that over 60 % of gold ions were reduced after
an X-Pert diffractometer (SmartLab, RIGAKU) using graphite-mono- 3 h, which was in good agreement with the results obtained from ICP-
chromatized Cu-Kα radiation (λ = 1.54 Å). The generator voltage and AES measurements. After reaction, the unreacted gold ions in the flow
current were set at 45 kV and 200 mA, respectively. The atomic content solution can be recycled for fabricating other samples. The gold NPs
of gold in the samples was detected by inductively coupled plasma formation was confirmed by XRD measurements. As shown in Fig. 2b,
atomic emission spectrometry (ICP-AES) with SHIMADZU, ICPS-8100. new peaks at 2θ = 38.1° and 44.3°, corresponding to the fcc packing
Au standards with the atomic concentration of 0.5, 1, 2, 5, 10, 20, and structure of Au(111) and Au(200) planes (Teranishi, Kiyokawa, &
50 ppm were used for the calibration. The reaction conversion of the Miyake, 1998), respectively, were observed on the spectrum of the
oxidation of PhMe2SiH was determined by proton nuclear magnetic monolith after reaction with gold ions, demonstrating a successful
resonance (1H NMR) spectroscopy using DMSO-d6 as a solvent, with preparation of gold NPs on the cellulose monolith. Since no external
JNM-ECS400 (400 MHz, JEOL Ltd., Tokyo, Japan). reducing agent was added during the continuous-flow generation of
gold NPs, the cellulose monolith itself must work as a reducing agent,
which was confirmed by the FTIR spectra for the observation at 1740
3. Results and discussion cm−1, which corresponds to the C]O stretching, after the reaction
(Fig. 2c) (Wu et al., 2014a). In addition, the change in the oxidation
3.1. Preparation of gold NPs on cellulose monolith state was determined by the peak-differentiation-imitation analysis of
the C1s XP spectra. The original cellulose monolith displayed three
The cellulose monolith was prepared according to our previously typical peaks at binding energies of 284.8, 286.6, and 288.0 eV, cor-
reported procedure based on a facile temperature-induced phase se- responding to the CeC/CeH bond, CeO single bond, and OeCeO bond
paration method (Xin et al., 2017). As shown in Fig. 1a, the obtained of cellulose, respectively (Fig. 2d) (Johansson & Campbell, 2004). In
cellulose monolith possessed a uniform column shape with a diameter contrast, the cellulose monolith after the reaction exhibited a relatively
around 0.8 cm. The SEM images (Fig. 1b and c) illustrated that the lower contribution of CeO bonding and a new peak appearing at the
monolith had a three-dimensional continuous macropores with sizes of binding energy of 289.7 eV, which could be interpreted by the con-
5–30 μm, which exhibited a rough surface. In fact, the macropore size sumption of CeO bonds and the formation of OeC]O bond, suggesting
of the obtained monolith can be certainly adjusted by changing the

Fig. 1. (a) Digital picture of the monolith. (b)

and (c) are SEM images of the cross-section of
the cellulose monolith under different magni-
fications. (d) Schematic diagram of the reduc-
tion of Au ions by cellulose monolith without
adding any external reducing agent. (e) Digital
pictures of the monolith after different reaction
times during the continuous-flow reaction with
HAuCl4 aqueous solution.

3
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

Fig. 2. (a) UV–vis spectra of gold ion solution before and after flowing with cellulose. Before reaction: 0.2 mmol/L, 5 mL. The inset figure is the standard curve of the
UV–vis intensity (312 nm) as a function of the concentration of gold ions. The pH of the sample was adjusted to 1 before the test. (b) XRD spectra of cellulose powder,
cellulose monolith, and cellulose-Au monolith, respectively. (c) FTIR spectra of the cellulose monolith and cellulose-Au monolith. (d) and (e) C1s XP spectra of the
cellulose monolith before and after reaction. (f) Schematic diagram of the interaction between gold NPs and the oxygen-containing groups of cellulose.

that some hydroxyl groups of the cellulose were oxidized during the spherical sphere shape without obvious aggregation. The corresponding
reaction (Fig. 2e). In general, cellulose does not show reducibility to statistical histogram illustrated that the size of the formed gold was
gold ions at room temperature, because of the high crystallinity and 12.1 ± 5.5 nm (based on over 300 counts). As a comparison, the gold
strong hydrogen bonding interaction. The reason for the high reactivity NPs synthesized by immersing the monolith in the gold ion solution
of the monolith may be attributed to the weak hydrogen bonding net- with the same pH and reaction conditions showed bigger size
work to liberate the free hydroxyl groups due to the monolith structure (26.4 ± 9.1 nm, based on over 300 counts) and wider size distribution.
as well as the alkaline conditions (Hu, Meng, Liu, Fu, & Lucia, 2017; In addition, the aggregation of the gold NPs was observed in the TEM
Zhang, Ruan, & Zhou, 2001). Meanwhile, the pore structure of the image (Fig. 3b). These results indicated that the flow method was more
cellulose monolith can act as an ideal scaffold for the seeding of gold. conducive to the homogenous formation of the gold NPs. The con-
During the continuous-flow process, the gold ions were forced to con- tinuous-flow process realized a homogeneous concentration distribu-
tact the activated hydroxyl groups on the pore surface and reduced by tion of gold ions, as well as achieves the better contact between the gold
the hydroxyl groups of the cellulose monolith. Both the large number of solution and the large surface of the monolith that bears sufficient re-
the hydroxyl groups and the newly formed carboxyl groups may con- action sites.
tribute to the immobilization of gold NPs on the surface of the monolith Furthermore, the influences of both the concentration of HAuCl4
via coordination or electrostatic adsorption (Fig. 2f). and the reaction time were investigated. As shown in Fig. 4a, the
sample prepared by the flow reaction with a low initial Au concentra-
3.2. Morphology of gold NPs formed on the monolith tion of 0.05 mmol/L (cellulose-Au-1) led to the smallest size of the gold
NPs (particles are marked in the red circle). To apply for a large
The morphology of the obtained gold NPs was observed by TEM throughput in practical use, the concentration of HAuCl4 was increased
(Fig. 3a). The gold NPs formed on the monolith showed a virtually to 0.2 mM (cellulose-Au-2) and 1 mM (cellulose-Au-3) (Fig. 4a–d). The

4
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

Fig. 3. TEM images and the size distribution of the gold NPs loaded on the monolith by (a) flow method and (b) immersion method.

Fig. 4. TEM images of gold NPs in cellulose-Au samples with different initial concentrations and reaction times: (a) Cellulose-Au-1: 0.05 mmol/L, 3 h. (particles are
indicated by red circle) (b) Cellulose-Au-2: 0.2 mmol/L, 3 h. (c) Cellulose-Au-3: 1.0 mmol/L, 3 h. (e) Cellulose-Au-4: 0.2 mmol/L, 12 h. (f) Cellulose-Au-5: 1 mmol/L,
12 h. The mean particle sizes of each sample were accounted and calculated in (d).

5
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

results showed that a larger size of gold NPs was formed by increasing cellulose-Au monolith as a catalyst of continuous-flowing system. As
the gold concentration. The effect of reaction time was also investigated shown in Fig. 5a, the cellulose-Au-2 was used as a flow reactor, and a
by extending the reaction time from 3 h to 12 h. With prolonged re- peristaltic pump was employed to flow the reactant mixture through
action time, almost all the gold ions in the solution with an initial the monolith to react under the catalysis of gold NPs. The NaBH4 re-
concentration of 0.2 mmol/L were converted to gold NPs, according to duction from 4-nitrophenol to 4-aminophenol was chosen as a model
the result of UV–vis experiment. (Fig. S4) However, the particles with reaction to evaluate the catalytic performance.
longer reaction time exhibited a significantly larger size, as shown in In a typical experiment, a mixture of 4-nitrophenol and NaBH4 was
Fig. 4e and f (compared to Fig. 4b and c, respectively). It could be flowed through the reactor using the tubing pump flow system. As a
interpreted by that the longer reaction time allowed the particles to result, the yellow color of 4-nitrophenol disappeared to become color-
grow further by using the formed particles as seeds and form bigger less when the flow rate was 3.5 mL/min, suggesting the successful re-
particles during the flow process as long as the supply of gold ions was duction of 4-nitrophenol. (Fig. 5b). We also monitored the progress of
adequate. The sample that reacted with an initial concentration of 0.2 reduction by using UV–vis spectroscopy, and found the complete dis-
mmol/L for 3 h (Cellulose-Au-2) was adopted in the later experiments appearance of the signal of 4-nitrophenol (400 nm) and the generation
due to its relatively smaller particle size and the shorter time con- of a new peak centered at 300 nm, indicating the formation of 4-ami-
sumption. nophenol. (Fig. 5c) The decrease of 4-nitrophenol at the different flow
rates was traced through the decay of the UV–vis signal at 400 nm
(Fig. 5d). Complete transformation was observed at the slower flow rate
3.3. Catalytic application of the obtained cellulose-Au monolith of up to 3.5 mL/min. According to the standard curve of 4-nitrophenol
in water (Fig. S5), the concentration of 4-nitrophenol remaining in the
Owing to the large number of oxide-containing groups near the gold effluent (C) was determined and then the conversion ((C0 − C )/ C0 ) was
NPs, the obtained gold NPs were stably anchored on the surface of calculated (Fig. 5e). Based on these results, the moles hourly space
monolith (Kaushik & Moores, 2016; Wu et al., 2014b). Even after cir- velocity (Miyamura, Suzuki, Yasukawa, & Kobayashi, 2018) (MHSV,
cularly washing the cellulose-Au monolith with 10 mL of deionized defined as the molar number of reactants catalyzed by per molar Au
water for 3 h, the amount of leached gold NPs was very low (< 1 ppm), atom for 1 h) was calculated according to the following equation:
as measured by ICP-AES. This result prompted us to investigate the

Fig. 5. (a) Digital photograph of the flow reactor; (b) The close-up photograph of monolith microreactor; (c) UV–vis spectra of the reactant solution at a flow rate of 1
mL/min. (diluted by 30 times before test) (d) The UV–vis spectra of effluent at the different flow rates. (diluted by 30 times before test)(e) Correlation between flow
rate and conversion. (f) Reusability test.

6
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

n4 − NP (C − C ) × v
MHSV = = 0
t × nAu nAu

where the n4 − NP and nAu refer to the molar number of 4-nitrophenol

reduced in one flow and the gold loaded on the monolith, respectively.
Here the nAu was about 0.62 μmol according to the ICP-AES results. t is
the reaction time and the v is the flow rate. C represents the con-
centration of 4-nitrophenol in the effluent and C0 is the original con-
centration before the flow. Considering that the reaction efficiency at
the flow rate of 3.5 mL/min was about 96 %, the MHSV was calculated
to be around 970 h–1. In the practical application of supported catalysts,
the reusability of the catalyst is of crucial importance. To demonstrate
the reusability of the obtained monolithic reactor, the catalytic ex-
periment was repeatedly conducted. The reactant was continuously
flowed through the monolith with a flow rate of 3 mL/min, and the
catalytic efficiency in each cycle was tested based on the UV–vis mea- Fig. 7. (a) Schematic diagram of oxidation of PhMe2SiH in DMF; (b)
surements. From the results recorded in Fig. 5f, it can be seen that the Conversion of the reaction as a function of reaction time. Inset was the digital
catalytic efficiency still kept very high (> 90 %) even after employed picture of the circular flow process; (c) Reusability test of the monolith reactor
for 8 cycles. The reusability can be interpreted by the abundance of after 60 min.
oxygen-containing groups on the surface of monolith, which could
anchor the gold NPs (Kaushik & Moores, 2016; Wu et al., 2014b). methylene blue and methyl orange, and showed good performance
Moreover, the micro/mesopores on the skeleton of monolith provided a (Figs. S6 and S7). Although most of the bio-materials-based solid sup-
rougher surface to absorb and fix the formed particles. ports are not applicable in organic solvents due to their instability and
The good catalytic performance of the as-prepared monolithic cel- solubility, the cellulose-based materials have high organic solvent tol-
lulose reactor can be attributed to the following reasons: First, the gold erance, which is one of their attractive characteristics of them.
NPs were directly supported by the cellulose monolith without adding Moreover, since no extra modifier, reductant or surfactant was in-
any other stabilizer or surfactant, and thus the gold NPs exposed a troduced during the whole preparation of the cellulose-Au flow reactor,
larger free surface and provided more reactive sites for catalysis. the obtained reactor has a simple chemical composition and remains
Furthermore, the immobilization of gold NPs was realized by a flow the good solvent tolerance, which expands the utility of the cellulose-
method, by which the gold NPs were mainly formed at the pore surface Au catalyst. Taking this into consideration, the oxidation of silane to
along with the flow routes with a good dispersion. This allowed the silanol in an organic solvent was investigated as a demonstration
gold to take part in the reaction sufficiently and avoided the idleness of (Fig. 7a and b). The PhMe2SiH/DMF solution was circularly flowed
the catalysts during the flow process. More importantly, the monolithic through the cellulose reactor to react. After 60 min of flowing, we found
cellulose, which is different from the bulk or fibrous cellulose, has a that PhMe2SiH was successfully oxidized to afford the corresponding
special hierarchically porous structure, as shown in Fig. 6. During the silanol (PhMe2SiOH) according to the result of 1HNMR measurements
flow reaction, the three-dimensional continuous macropores endowed (Fig. S8). To illustrate the durability of the monolith reactor under the
the reactor with a good permeability and high throughput capacity, reaction conditions, the reusability test was carried out. As shown in
while the numerous micro/meso-pores, which contributed to the rough Fig. 7c, the cellulose reactor can be used at least 7 times while retaining
surface of the skeleton, enabled the reactants to be quickly diffused in the catalytic activity, showing the high reusability even in DMF. Based
the monolith and sufficiently contacted with the catalysts. Thereby the on all the analysis and discussion above, we believed that the cellulose
monolithic cellulose reactor realized a combination of good flow ability monolith is a prospective candidate for developing high-efficiency flow
and the high reaction efficiency. reactors.
Beside of the reduction of 4-nitrophenol, the reactor was also ap-
plied to other reactions such as the degradation of dyes including

Fig. 6. Schematic diagram of monolith microstructure during the continuous-flow reaction.

7
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

4. Conclusions et al. (2014). Self-assembly of metal and metal oxide nanoparticles and nanowires
into a macroscopic ternary aerogel monolith with tailored photocatalytic properties.
Chemistry of Materials, 26(19), 5576–5584.
A simple, green, and cost-effective strategy was developed for Hu, Z., Meng, Q., Liu, R., Fu, S., & Lucia, L. A. (2017). Physical study of the primary and
constructing high-efficiency gold supported monolithic cellulose flow secondary photothermal events in Gold/Cellulose nanocrystals (AuNP/CNC) nano-
reactor. The cellulose monolith prepared using the facile TIPS method composites embedded in PVA matrices. ACS Sustainable Chemistry & Engineering, 5(2),
1601–1609.
can be directly used to synthesize and stabilize the gold NPs at room Jal, P. K., Patel, S., & Mishra, B. K. (2004). Chemical modification of silica surface by
temperature without any other surface modification or adding a re- immobilization of functional groups for extractive concentration of metal ions.
duction agent. The results demonstrated that the abundant hydroxide Talanta, 62(5), 1005–1028.
Jiang, X., Han, J., Cao, L., Bao, Y., Shi, J., Zhang, J., et al. (2018). A facial strategy for
groups on the cellulose endowed the monolith with reducibility to gold catalyst and reducing agent synchronous separation for AGET ATRP using thiol-
ions. Compared to the batch synthesis, the flow reduction was more grafted cellulose paper as reducing agent. Polymers, 10(1), 26.
conducive to the homogeneous preparation of gold NPs with relatively Johansson, L.-S., & Campbell, J. M. (2004). Reproducible XPS on biopolymers: Cellulose
studies. Surface and Interface Analysis, 36(8), 1018–1022.
smaller sizes. Moreover, the size of the gold NPs formed on the cellulose
Kaushik, M., & Moores, A. (2016). Nanocelluloses as versatile supports for metal nano-
monolith could be changed by controlling the reaction conditions in- particles and their applications in catalysis. Green Chemistry, 18(3), 622–637.
cluding the initial concentration of HAuCl4 and the reaction time. Koczkur, K. M., Mourdikoudis, S., Polavarapu, L., & Skrabalak, S. E. (2015).
Consequently, the obtained cellulose products could be applied as a Polyvinylpyrrolidone (PVP) in nanoparticle synthesis. Dalton Transactions, 44(41),
17883–17905.
green flow reactor in both water and organic solution and showed ex- Li, Y., Xu, L., Xu, B., Mao, Z., Xu, H., Zhong, Y., et al. (2017). Cellulose sponge supported
cellent catalytic efficiency and good reusability. This work opened a palladium nanoparticles as recyclable cross-coupling catalysts. ACS Applied Materials
new perspective for the development the green and high-efficiency flow & Interfaces, 9(20), 17155–17162.
Liang, Q., Li, Z., Yu, X., Huang, Z. H., Kang, F., & Yang, Q. H. (2015). Macroscopic 3D
reactor based on cellulose monolith. porous graphitic carbon nitride monolith for enhanced photocatalytic hydrogen
evolution. Advanced Materials, 27(31), 4634–4639.
CRediT authorship contribution statement Liguori, F., Barbaro, P., Said, B., Galarneau, A., Santo, V. D., Passaglia, E., et al. (2017).
Unconventional Pd@ sulfonated silica monoliths catalysts for selective partial hy-
drogenation reactions under continuous flow. ChemCatChem, 9(16), 3245–3258.
Zheng-Tian Xie: Investigation, Writing - original draft, Writing - Macanás, J., Ouyang, L., Bruening, M. L., Muñoz, M., Remigy, J. C., & Lahitte, J. F.
review & editing. Taka-Aki Asoh: Project administration, Writing - (2010). Development of polymeric hollow fiber membranes containing catalytic
metal nanoparticles. Catalysis Today, 156(3–4), 181–186.
review & editing, Funding acquisition. Yuta Uetake: Investigation, Miyamura, H., Suzuki, A., Yasukawa, T., & Kobayashi, S. (2018). Polysilane-immobilized
Writing - review & editing. Hidehiro Sakurai: Writing - review & Rh-Pt bimetallic nanoparticles as powerful Arene hydrogenation catalysts: Synthesis,
editing. Hiroshi Uyama: Supervision, Writing - review & editing, reactions under batch and flow conditions and reaction mechanism. Journal of the
American Chemical Society, 140(36), 11325–11334.
Funding acquisition.
Moitra, N., Kanamori, K., Shimada, T., Takeda, K., Ikuhara, Y. H., Gao, X., et al. (2013).
Synthesis of hierarchically porous hydrogen silsesquioxane monoliths and embedding
Declaration of Competing Interest of metal nanoparticles by on-site reduction. Advanced Functional Materials, 23(21),
2714–2722.
Nakao, H., Shiigi, H., Yamamoto, Y., Tokonami, S., Nagaoka, T., Sugiyama, S., et al.
The authors declare no interest conflict. (2003). Highly ordered assemblies of Au nanoparticles organized on DNA. Nano
Letters, 3(10), 1391–1394.
Pelisson, C. H., Nakanishi, T., Zhu, Y., Morisato, K., Kamei, T., Maeno, A., et al. (2017).
Acknowledgements Grafted polymethylhydrosiloxane on hierarchically porous silica monoliths: A new
path to monolith-supported palladium nanoparticles for continuous flow catalysis
This work was supported by JSPS KAKENHI Grants (Nos. applications. ACS Applied Materials & Interfaces, 9(1), 406–412.
Pélisson, C.-H., Nakanishi, T., Zhu, Y., Morisato, K., Kamei, T., Maeno, A., et al. (2016).
17H03114, 19H02778), JST-Mirai Program (Grant No. JPMJMI18E3) Grafted polymethylhydrosiloxane on hierarchically porous silica monoliths: A new
and JSPS Core-to-Core Program, B. Asia-Africa Science Platforms. We path to monolith-supported palladium nanoparticles for continuous flow catalysis
thank for the help of Prof. Susumu Kuwabata and Prof. Taro Uematsu applications. ACS Applied Materials & Interfaces, 9(1), 406–412.
Saito, T., Kimura, S., Nishiyama, Y., & Isogai, A. (2007). Cellulose nanofibers prepared by
for the observation in TEM experiments. The author Z. Xie would like to TEMPO-mediated oxidation of native cellulose. Biomacromolecules, 8(8), 2485–2491.
thank China Scholarship Council (CSC) for a scholarship support. Schurz, J. (1999). A bright future for cellulose. Progress in Polymer Science, 24(4),
481–483.
Shih, C.-M., Shieh, Y.-T., & Twu, Y.-K. (2009). Preparation of gold nanopowders and
Appendix A. Supplementary data
nanoparticles using chitosan suspensions. Carbohydrate Polymers, 78(2), 309–315.
Tarazona, P. (1985). Free-energy density functional for hard spheres. Physical Review A,
Supplementary material related to this article can be found, in the General Physics, 31(4), 2672–2679.
online version, at doi:https://doi.org/10.1016/j.carbpol.2020.116723. Tarazona, P., Marconi, U. M. B., & Evans, R. (1987). Phase equilibria of fluid interfaces
and confined fluids. Molecular Physics, 60(3), 573–595.
Teranishi, T., Kiyokawa, I., & Miyake, M. (1998). Synthesis of monodisperse gold nano-
References particles using linear polymers as protective agents. Advanced Materials, 10(8),
596–599.
Turkevich, J., Stevenson, P. C., & Hillier, J. (1951). A study of the nucleation and growth
Ashraf, S., Sher, F., Khalid, Z. M., Mehmood, M., & Hussain, I. (2014). Synthesis of cel- processes in the synthesis of colloidal gold. Discussions of the Faraday Society, 11,
lulose–metal nanoparticle composites: Development and comparison of different 55–75.
protocols. Cellulose, 21(1), 395–405. Van Rie, J., & Thielemans, W. (2017). Cellulose-gold nanoparticle hybrid materials.
Bedford, R. B., Singh, U. G., Walton, R. I., Williams, R. T., & Davis, S. A. (2005). Nanoscale, 9(25), 8525–8554.
Nanoparticulate palladium supported by covalently modified silicas: Synthesis, Viklund, C., Svec, F., Fréchet, J. M. J., & Irgum, K. (1996). Monolithic, “Molded”, porous
characterization, and application as catalysts for the suzuki coupling of aryl halides. materials with high flow characteristics for separations, catalysis, or solid-phase
Chemistry of Materials, 17(4), 701–707. chemistry: Control of porous properties during polymerization. Chemistry of Materials,
Benaissi, K., Johnson, L., Walsh, D. A., & Thielemans, W. (2010). Synthesis of platinum 8(3), 744–750.
nanoparticles using cellulosic reducing agents. Green Chemistry, 12(2), 220–222. White, R. J., Luque, R., Budarin, V. L., Clark, J. H., & Macquarrie, D. J. (2009). Supported
Bhumkar, D. R., Joshi, H. M., Sastry, M., & Pokharkar, V. B. (2007). Chitosan reduced metal nanoparticles on porous materials. Methods and applications. Chemical Society
gold nanoparticles as novel carriers for transmucosal delivery of insulin. Reviews, 38(2), 481–494.
Pharmaceutical Research, 24(8), 1415–1426. Wu, X., Lu, C., Zhang, W., Yuan, G., Xiong, R., & Zhang, X. (2013). A novel reagentless
Bronstein, L. M., Polarz, S., Smarsly, B., & Antonietti, M. (2001). Sub‐nanometer no- approach for synthesizing cellulose nanocrystal-supported palladium nanoparticles
ble‐metal particle host synthesis in porous silica monoliths. Advanced Materials, with enhanced catalytic performance. Journal of Materials Chemistry A, 1(30), 8645.
13(17), 1333–1336. Wu, X., Lu, C., Zhou, Z., Yuan, G., Xiong, R., & Zhang, X. (2014a). Green synthesis and
Hadjiivanov, K. I., & Klissurski, D. G. (1996). Surface chemistry of titania (anatase) and formation mechanism of cellulose nanocrystal-supported gold nanoparticles with
titania-supported catalysts. Chemical Society Reviews, 25(1), 61. enhanced catalytic performance. Environmental Science Nano, 1(1), 71.
Han, L., Cai, S., Gao, M., Hasegawa, J. Y., Wang, P., Zhang, J., et al. (2019). Selective Wu, X., Lu, C., Zhou, Z., Yuan, G., Xiong, R., & Zhang, X. (2014b). Green synthesis and
catalytic reduction of NOx with NH3 by using novel catalysts: State of the art and formation mechanism of cellulose nanocrystal-supported gold nanoparticles with
future prospects. Chemical Reviews, 119(19), 10916–10976. enhanced catalytic performance. Environmental Science Nano, 1(1), 71–79.
Heiligtag, F. J., Cheng, W., de Mendonça, V. R., Süess, M. J., Hametner, K., Günther, D., Xie, Z. T., Asoh, T. A., & Uyama, H. (2019). Monolithic cellulose supported metal

8
Z.-T. Xie, et al. Carbohydrate Polymers 247 (2020) 116723

nanoparticles as green flow reactor with high catalytic efficiency. Carbohydrate gold sponges for efficient and stable catalysis in a continuous-flow microreactor.
Polymers, 214, 195–203. Materials Chemistry Frontiers, 1(3), 482–486.
Xin, Y., Xiong, Q., Bai, Q., Miyamoto, M., Li, C., Shen, Y., et al. (2017). A hierarchically Zhang, L., Ruan, D., & Zhou, J. (2001). Structure and properties of regenerated cellulose
porous cellulose monolith: A template-free fabricated, morphology-tunable, and ea- films prepared from cotton linters in NaOH/Urea aqueous solution. Industrial &
sily functionalizable platform. Carbohydrate Polymers, 157, 429–437. Engineering Chemistry Research, 40(25), 5923–5928.
Yang, H., Lu, Q., Gao, F., Shi, Q., Yan, Y., Zhang, F., et al. (2005). One-step synthesis of Zhang, Y., Liu, Y., Wang, X., Sun, Z., Ma, J., Wu, T., et al. (2014). Porous graphene oxide/
highly ordered mesoporous silica monoliths with metal oxide nanocrystals in their carboxymethyl cellulose monoliths, with high metal ion adsorption. Carbohydrate
channels. Advanced Functional Materials, 15(8), 1377–1384. Polymers, 101, 392–400.
Yu, Y., Xiao, W., Zhou, T., Zhang, P., Yan, C., & Zheng, Z. (2017). Monolithic hierarchical