Conversion of furfural to 2-methylfuran over CuNi catalysts supported on biobased carbon foams

Article abstract of DOI:10.1016/j.cattod.2020.10.027

In this study, carbon foams prepared from the by-products of the Finnish forest industry, such as tannic acid and pine bark extracts, were examined as supports for 5/5% Cu/Ni catalysts in the hydrotreatment of furfural to 2-methylfuran (MF). Experiments were conducted in a batch reactor at 503 K and 40 bar H₂. Prior to metal impregnation, the carbon foam from tannic acid was activated with steam (S1), and the carbon foam from pine bark extracts was activated with ZnCl₂ (S2) and washed with acids (HNO₃ or H₂SO₄). For comparison, a spruce-based activated carbon (AC) catalyst and two commercial AC catalysts as references were investigated. Compressive strength of the foam S2 was 30 times greater than that of S1. The highest MF selectivity of the foam-supported catalysts was 48 % (S2, washed with HNO₃) at a conversion of 91 %. According to the results, carbon foams prepared from pine bark extracts can be applied as catalyst supports.

Full text of DOI:10.1016/j.cattod.2020.10.027

Journal Pre-proof
Conversion of furfural to 2-methylfuran over CuNi catalysts supported on
biobased carbon foams
Toni Varila (Conceptualization) (Investigation) (Formal analysis)
(Writing - original draft) (Writing - review and editing) (Visualization),
Eveliina M a¨ kel a¨ (Conceptualization) (Investigation) (Formal
analysis) (Writing - original draft) (Writing - review and editing)
(
Visualization), Riikka Kupila (Conceptualization) (Writing - original
draft) (Writing - review and editing) (Visualization), Henrik Romar
Conceptualization) (Supervision) (Writing - original draft) (Writing -
(
review and editing), Tao Hu (Formal analysis) (Writing - review and
editing) (Visualization), Reetta Karinen (Supervision) (Writing -
original draft) (Writing - review and editing), Riikka L. Puurunen
(
Supervision) (Writing - original draft) (Writing - review and editing),
Ulla Lassi (Conceptualization) (Supervision) (Project administration)
Writing - original draft) (Writing - review and editing)
(
PII:
S0920-5861(20)30724-0
DOI:
https://doi.org/10.1016/j.cattod.2020.10.027
CATTOD 13192
Reference:
To appear in:
Catalysis Today
Received Date:
Revised Date:
Accepted Date:
19 November 2019
8 October 2020
14 October 2020
Please cite this article as: { doi: https://doi.org/

This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
©
2020 Published by Elsevier.

Conversion of furfural to 2-methylfuran over CuNi catalysts supported on
biobased carbon foams
Toni Varila*[a,b], Eveliina Mäkelä , Riikka Kupila , Henrik Romar , Tao Hu Reetta Karinen ,
[c]
[a,b]
[a]
[a]
[c]
[
c]
[a,b]
Riikka L. Puurunen and Ulla Lassi
a) Research Unit of Sustainable Chemistry, University of Oulu, P.O. Box 8000, 90014 Oulu, Finland
b) Kokkola University Consortium Chydenius, Applied Chemistry, University of Jyväskylä, P.O. Box
5
67, 67101 Kokkola, Finland
c) Department of Chemical and Metallurgical Engineering, Aalto University School of Chemical
Engineering, P.O. Box 16100, 00076 AALTO, Finland
*
Corresbonding author: toni.varila@chydenius.fi
Graphical abstract

Highlights

Activated carbon foams can be used as catalyst supports.

High conversion of furfural and selectivity for 2-methylfuran were obtained with Cu/Ni catalysts on
activated carbon foams.

Compressive strength of activated carbon foam materials can be enhanced with thermal treatment at
high temperatures.


Acid treatment of carbon foam supports with nitric acid increased the selectivity to 2-methylfuran.
Small metal particles on the surface of the catalyst supports increased the selectivity to 2-
methylfuran.
Abstract: In this study, carbon foams prepared from the by-products of the Finnish forest industry, such
as tannic acid and pine bark extracts, were examined as supports for 5/5% Cu/Ni catalysts in the
hydrotreatment of furfural to 2-methylfuran (MF). Experiments were conducted in a batch reactor at
5
03 K and 40 bar H
steam (S1), and the carbon foam from pine bark extracts was activated with ZnCl
acids (HNO or H SO ). For comparison, a spruce-based activated carbon (AC) catalyst and two
2
. Prior to metal impregnation, the carbon foam from tannic acid was activated with
2
(S2) and washed with
3
2
4
commercial AC catalysts as references were investigated. Compressive strength of the foam S2 was 30
times greater than that of S1. The highest MF selectivity of the foam-supported catalysts was 48% (S2,
washed with HNO ) at a conversion of 91%. According to the results, carbon foams prepared from pine
3
bark extracts can be applied as catalyst supports.
Keywords: Biobased foams, carbon, mechanical strength, furfural, Cu/Ni catalyst, 2-methylfuran.
1
. Introduction
Owing to the increase in the demand for chemicals and liquid fuels, which are primarily prepared
from depleting fossil fuel sources, the growing society requires sustainable and renewable resources to

produce value-added chemicals and liquid fuels [1,2]. Biomass has been widely investigated for
decades, and it has proven to be a good source of energy and raw material for chemicals and fuels due
to its consistency, renewability, structure, and easy availability [3]. Biomass can be applied in several
industrial and chemical processes, such as pyrolysis, gasification, fermentation, and extraction, to
produce heat, electricity, and biobased materials, such as ethanol and furfural derivatives. Developing
countries produce lignocellulosic feedstocks, which are mainly used as energy for industrial boilers [4].
In Finland, lignocellulosic feedstock, which is produced by the forest industry, is also used as an energy
source,[5] as well as in the pulping process, and it can be further developed in biorefineries to produce
other value-added products, e.g., 2-methylfuran (MF) and its derivatives [6]. As a side product, bark and
lignin are produced in a large amount, which are mainly considered as waste or low-value products [7].
Furfural is known as a platform chemical, which is obtained from lignocellulosic biomass by the
hydrolysis and dehydration of hemicellulose [8]. Hydrotreatment is a typical process for the upgrading
of furanic components; the conversion can be performed in either the gas or liquid phase [8,9]. MF is
one of the most important products obtained from the hydrotreatment of furfural, which can be used as a
solvent or biofuel when blended with gasoline [9]. MF demonstrates the potential to replace fossil-based
components in gasoline due to its beneficial fuel characteristics, e.g., high octane number and low

1
solubility in water (research octane number of 131, 7 g L ) [10]. Recent studies have reported high MF
yields [11–14]. For example, Fu et al. [11] have reported an MF yield of 92% over a 10–10% CuNi
catalyst on an Al
hydrogen donor.
2
3
O support, with a furfural conversion of 100%, at 483 K using formic acid as the
Furanic compounds, such as furfural, can be adsorbed on a metal either via the furan ring or via the
oxygenated side chain or both. The adsorption mode of furfural depends on the nature of the metal.
1
Three typical adsorption modes are as follows: η -(O) adsorption mode via the aldehyde functionality,
η2-(C,O) adsorption mode, in which the furan ring is adsorbed on the metal with the C atom but also
1
2
with the O atom of the carbonyl group, and η -(C) acyl adsorption mode, which is converted from η -

(C,O) adsorption at high temperatures [15,16]. Furfural exhibits a stronger interaction with Ni (in
2
contrast to Cu), permitting the adsorption of the furan ring via the η -(C,O) adsorption mode. In case of
Cu, the Cu(111) surface, in particular, exhibits a strong repulsion for the furan ring; thus, the typical
1
adsorption mode is η -(O) [16].
Shi [17] has conducted a density functional theory (DFT) study and reported that MF can be
formed by four competitive routes. The attack of carbon from the C=O group by hydrogen affords an
alkoxy intermediate or a 2-furanyl(hydroxy)methyl intermediate. In addition, direct C=O and C–H
dissociation routes were presented. Among these pathways, the formation of the alkoxy intermediate
species was the rate-determining step [17]. In general, the Cu catalyst favored the production of furfuryl
alcohol (FA), MF, and pentanediols rather than decarbonylation and furan-ring hydrogenation products,
such as 2‐ methyltetrahydrofuran (MTHF) and THF. As the hydrogenation of the C=O bond in FA was
relatively less favorable, which is required to obtain MF [10,16,18], the activity can be increased by
utilizing a high reaction temperature, adding a second oxophilic metal to the catalyst to enhance the
deoxygenation ability, or increasing the Lewis acidity of the catalyst [16]. The presence of both Cu and
+
+
Cu in the catalyst is crucial. Metallic Cu was proposed to activate H
2
and the Cu species, which serve
as a Lewis acid or as electrophilic sites polarizing the C=O bond [19].
Owing to the higher hydrogenation activity of Ni, the distribution of products obtained over a Ni
catalyst is typically wider than obtained over a Cu catalyst. In addition to FA, MF, pentanediols,
cyclopentanone, and cyclopentanol, tetrahydrofurfuryl alcohol (THFA), MTHF, THF, and furan can be
obtained. Previously, Rodiansono et al. [20] have reported the formation of 1,4-pentanediols over Ni
catalysts. THFA is formed by the nonselective hydrogenation over most of the Ni catalysts [16]. The
MF selectivity can be improved by the application of bimetallic Ni catalysts; for example, the NiFe
alloy can reduce the decarbonylation activity and improve the hydrogenolysis of the C–O bond of FA
2
1
by hindering the transformation of the η -(C,O) adsorption mode to the η -(C) acyl adsorption mode
[
21]. Moreover, Yu et al. [22] have conducted a DFT study and proposed that the main adsorption mode

is via the aldehyde functionality on the NiFe surfaces. A similar interaction also was observed on the
NiCu(111) surface [15].
Activated carbons (ACs) are produced by thermal treatment or a combination of chemical and
thermal treatment of carbon-rich sources such as lignite or of lignocellulosic materials in the case of
biomass-based carbon. In AC production, coconut shells, peat, and wood chips or saw dust are typically
used as biobased raw materials [23,24]. The porous structure, high specific surface area (SSA), and
adsorption capacity of ACs are widely exploited in several industrial and municipal purification
processes. ACs are mainly applied for wastewater treatment, gas cleaning processes, metal removal
from waste streams, and as a catalyst [25–30]. Recently, AC has been used as a support material for
transition metals, and its activity for the hydrotreatment of furfural has been investigated. For example,
Fu et al. [12] have reported a high MF yield (91%) after a reaction time of 8 h at 473 K (formic acid as
the hydrogen donor) using the 10–10% CuNi catalyst on AC. Date et al. [13] have reported an MF yield
of 95%, with almost complete conversion, after a reaction time of 5 h at 493 K (6.9 bar H
over a 5% Ir/AC catalyst. Gong et al. [31] have reported the highest MF yield of almost 100%, with
00% furfural conversion, over a Cu/AC catalysts at 440 K and 40 bar H after a reaction time of 4 h.
One of the drawbacks of using ACs as a catalyst support material is their limited mechanical
2
pressure)
1
2
stability, especially abrasion strength [32,33]. Typically, briquetting or other techniques that can
enhance the mechanical stability of ACs are required to avoid particle shattering and a high pressure
drop in a fixed-bed reactor [34,35]. Hence, widely investigated condensed tannin-based carbon foams
[
36–38] are promising support materials due to the possibility of tailoring the mechanical strength by
high-temperature thermal treatment [39]. Thermal treatment (steam activation) typically produces a
microporous material, whereas chemical activation (e.g., using ZnCl
material [28].
2
) typically produces a mesoporous
Tannic-acid-based or pine-bark-extracts-based carbon foams (Fig. 1) can be produced using a
combination of a phenolic substance (condensed or hydrolysable tannin), which are found in plants or in

the wood bark structure, a cross-linker, catalyst, surfactant, and blowing agent. Physical properties, such
as SSA, pore size distribution (PSD), mechanical strength, electrical and thermal conductivities, fire
resistance, metal adsorption, and permeability, of tannin-based foams have been investigated previously
[
40–42]. However, to the best of our knowledge, the use of these foams as catalyst support materials has
not been attempted previously.
In this study, the goal was to investigate the use of two activated carbon foams (ACF), which were
produced from tannic acid and pine bark extracts, respectively, as catalyst supports for bimetallic Cu/Ni
(5/5 w%) catalysts in the conversion of furfural to MF. As-prepared ACF were activated by physical
and chemical activation methods. In addition, the SSA, PSD, and mechanical strength of these catalysts
were investigated. Prior to metal addition, the activated support materials were further washed with
HNO
3
and H
2
SO
4
to modify their surface properties. The combination of Cu and Ni was selected as
these metals were found to be beneficial for the hydrotreatment of furfural to produce MF [11,12,43].
2
. Material and Methods
For the first catalyst support material (S1), commercial tannic acid (95%), FA (98%), and surfactant
(Tween 85) were purchased from Acros Organics. The acid catalyst p-toluenesulfonic acid monohydrate

(PTSA, 99%) for the foaming reaction and the blowing agent n-pentane (99.4%) were purchased from
Merck KGaA. For chemical activation, technical-grade ZnCl (97%) was purchased from VWR
2
Chemicals. Catalytic precursors of nickel(II) nitrate hexahydrate (20.17 w% Ni, 99% purity) and
copper(II) nitrate hemi(pentahydrate) (27.30 w% Cu, 98% purity) were purchased from Merck KGaA
and Alfa Aesar, respectively. For the second catalyst support material (S2), pine bark from northern
Finland was used as the raw material. Pine bark was extracted using ETA 700 solution (70 w% ethanol
in a water disinfectant solution), which was purchased from ETRA Oy Finland. For the third catalyst
support material (S3), fine spruce powder from northern Finland was used. For acid treatment, HNO
65 m%), HCl (32 w%), and H SO (96 w%) were purchased from Merck KGaA. Commercial AC
3
(
2
4
supports, i.e., steam-activated carbon (RB4C) and acid-washed carbon (RX3 extra), were obtained from
Norit.
For batch reactor experiments, the below-listed chemicals were purchased from Sigma-Aldrich. FA
(
98%), furan (≥99%), MF (99%), THFA (99%), MTHF (anhydrous, ≥99%), 2‐ propanol (≥99.5%), 2‐
butanol (99%), 2‐ pentanol (98%), cyclopentanol (99%), 2‐ pentanone (99.5%), and cyclopentanone
≥99%) were used without further purification. Furfural (99%) was distilled to a final molar purity of
9.8%. The gases used for batch experiments were purchased from Oy AGA Ab. The purity of the
hydrogen used for reactor experiments was 5.0, and the gases used in analytics were H (purity 5.0), He
purity 4.6), Ar (purity 5.0), synthetic air (purity 5.0), and N (purity 5.0). Two calibration gas mixtures
were utilized. The first contained 40 mol% N , 5 mol% CH , 10 mol% C , 5 mol% C , 10 mol%
, 5 mol% C , 5 mol% C , 10 mol% C 10, and 10 mol% isobutane. The second calibration
gas mixture contained 15 vol% CO, 15 vol% CO , 15 vol% H , 40 vol% N , and 15 vol% CH
(
9
2
(
2
2
4
2
H
6
2
H
4
C
3
H
8
3
H
6
2
H
2
4
H
2
2
2
4
.
2
.1. Preparation of catalyst supports S1, S2, and S3
Supports S1 and S3: The commercial tannic acid mixture was used to prepare support 1 (S1).
First, deionized water (9 g), furfuryl alcohol (14 g), and Tween 85 (2 g) were thoroughly mixed for 30 s

in a 1-L beaker. Second, tannic acid (30 g) was added to this liquid mixture and mixed with a
mechanical stirrer (2000 rpm) for 5 min. Third, pentane was added (4.5 g) to the homogenous solution,
followed by the addition of PTSA 8 g, 65% solution), and it was rapidly stirred with a mechanical
stirrer. Foaming occurred within a few seconds, and polymerization was highly exothermic. When the
initial foam stopped increasing, the beaker was placed in an oven for 24 h at 373 K by natural
convection to harden the foam.
Matured S1 was cut in pieces (1 cm ×1 cm × 1 cm) and placed in a stainless-steel tube, which was
subsequently placed in a tubular fixed-bed reactor for activation. For support S1, steam activation was

1
conducted at 1073 K using a ramp rate of 10 K min and a holding time of 4 h at the target
temperature. For support S3, spruce sawdust was activated by steam at 1073 K using a ramp rate of 6.5

1
1
K min and a holding time of 2 h. A water flow of 0.050 mL min was utilized for both activation
methods at the target temperature. Nitrogen gas was flushed through the reactor to avoid sample
oxidation. After physical activation, S3 material was washed several times with hot water to remove
impurities such as Ca, K, Mg, Na, and Fe. Next, the catalyst support material was sieved to obtain
particle sizes of 100–425 µm and further characterized and used for catalyst preparation.
Support S2: Pine bark extracts were used to prepare support 2 (S2). First, the bark was dried at
room temperature, followed by crushing and sieving to obtain a particle size of less than 1 mm. Second,

1
the bark was extracted with EtOH:H O (70:30) at a bark:solvent ratio of 75:1 (g L ). Extraction was
2
carried out using a Soxhlet extractor for 2 h at an oil bath temperature of ~383 K. The bark extracts
were dried at room temperature in a fume hood until the solvent was evaporated (~2 days). Then,
support S2 was prepared as follows.
First, pine bark extracts (30 g) were placed in a 400-mL beaker, followed by melting on a hot plate
at 373 K to enable the mixing of the extractives with a spatula. Second, deionized water (9 g), furfural
alcohol (14 g), and Tween 85 (2 g) were added to this solution, and the mixture was thoroughly mixed

for 5 min using a mechanical stirrer. Third, pentane (4.5 g) and PTSA (8 g, 65% solution) were added
and stirred for 10 s. Foaming occurred within few minutes after placing the beaker in an oven at 373K.
The foam was hardened for 24 h at this temperature before further use.
First, matured S2 was gently crushed in a mortar using a pestle for chemical activation. Technical-
grade ZnCl
2
(40 g) was dissolved in 100–200 mL of deionized water in a 600-mL beaker. Second, the
to the catalyst
support material was added to this solution so as to attain an initial ratio of 2:1 for ZnCl
2
support material. Third, the solution was stirred using a stir bar at 353 K for 3 h with the addition of
deionized water from time to time into the solution if the water level decreased considerably. Finally,
the ZnCl
2
-impregnated wet carbon foams were dried in an oven at 378 K for 48 h. Then, the dried
impregnated support material was placed in a stainless-steel tube for chemical activation, which was

1
conducted at 873 K using a ramp rate of 5 K min and a holding time of 2 h at the target temperature.
Nitrogen gas was flushed through the reactor the entire time to avoid sample oxidation.
After wet impregnation and activation, the support material was refluxed with 3 M HCl for 1 h at
3
73 K to remove the remaining ZnCl
deionized water until the pH was neutral. Then, the support material was dried in an oven at 373 K for
4 h. The dry material was then sieved, characterized, and further used to prepare the catalyst in the
same way as that carried out for S1.
2
. The catalyst support material was filtered and washed with
2
2
.2. Mineral acid treatment of S1 and S2 supports
Surfaces of S1 and S2 materials were modified by treatment with 3 M HNO
3
(A1) or 6 M H
2
SO
4
(
A2). Hereafter, these supports will be referred to as S1_A1, S1_A2 or S2_A1, S2_A2. The precise
modification method was conducted as follows: In a 100-mL two-neck round-bottom flask, 50 mL of
acid per 0.3 g of support was used. HNO treatment was conducted at 373 K for 2 h, and H SO
treatment was conducted at 353 K for 3 h. For S1 material, careful mixing was required to maintain an
3
2
4

appropriate particle size of the material. After acid treatment, supports were washed with distilled water
until the pH was neutral and subsequently placed in the oven at 373 K.
2
.3. Catalyst preparation
Prior to the dry impregnation of catalyst supports, pore volumes (PV) of all support materials were
measured as described in the next section. The amount of metal salts, i.e., Cu(NO
3
)
3
2.5H
2
O and
Ni(NO 6H O, inside the support were calculated so that the target concentration of metal was 5 w%
3
)
2
2
each. Metal salts were dissolved in an exact amount of water to fill the PV of each support. Supports
were mixed a few times with a spatula for 4–5 h and finally dried in an oven at 373 K for 24 h. All
catalysts were subjected to thermal treatment in a stainless-steel tube in a fixed-bed reactor under

1
nitrogen. Thermal treatment was carried out at 773 K using a ramp rate of 5 K min and a holding time

1
of 2 h at the target temperature under a nitrogen flow (10 mL min ).
2
.4. Characterization of catalyst materials
The SSA and PSD were determined The SSA and PSD were determined for unreduced, thermally
treated catalyst samples by nitrogen physisorption isotherms measured at 77.15 K on a Micromeritics
ASAP 2020 instrument (Micromeritics Instruments, Norcross, GA, USA). Portions of each sample
(100–200 mg) were degassed at a low pressure of 0.27 kPa and at a temperature of 413 K for 2 h to
remove adsorbed gas. Adsorption isotherms were obtained by the immersion of sample tubes in liquid
nitrogen (77.15 K) to achieve constant temperature conditions and by the addition of a small dose of
gaseous nitrogen in samples.
SSAs were calculated from adsorption isotherms according to the Brunauer–Emmett–Teller (BET)
method [44]. Total PVs were calculated from adsorption isotherms at a p/p
0
ratio of 0.985 and in the
case of the DFT [45] calculation as the total PV was measured at the maximum pore width. PSD was

calculated by using the DFT algorithm under the assumption of slit-formed pores [46]. The distribution
of PV (vol%) was calculated from the individual volumes of micropores, mesopores, and macropores
with the DFT model. t-Plot calculations were conducted using the Harkins and Jura method [47]. By
using the instrumental setup, micropores down to a diameter of 1.5 nm can be measured even in case of
possible contribution from smaller pores. The micropore areas and volumes were calculated by the t-
plot algorithm. A previous study has reported that SSAs are typically measured with a precision of 5%
[
48].
Compressive strengths (σ) of S1 and S2 supports were determined before and after thermal
treatment at 1073 K for 2 h using a Zwick/Roell Z010 testing machine (Kennesaw, Georgia, with a load
cell of 10 kN). The loading speed was 0.1 mm/s, and the measurement ended when the force decreased
by 50% from the maximum or deformation reached 30%. The compressive strength of pieces, with
diameters of 1.1 cm × 1.5 cm × 0.4 cm, was measured and calculated using Equation 1,
σ = F/A,
(1)
2
where F (N) is the maximum force at the linear region of the compression curve, and A (mm ) is the
surface area. The accuracy of the instrument was ±1%.
Metal contents of the unreduced, thermally treated catalysts were measured by inductively coupled
optical emission spectrometry (ICP-OES) using a Perkin Elmer Optima 5300 DV instrument. First, 0.1–
0
4
.2 g of samples were digested in a microwave oven (MARS, CEM Corporation) with 9 mL of HNO
3
at
73 K for 10 min. Second, 3 mL of HCl was added, and the mixture was digested at 473 K for 10 min.
Finally, 1 mL of HF was added, and the mixture was digested again at 473 K for 10 min. Excess HF
was neutralized using H BO by heating at 443 K for 10 min. Next, the solution was diluted to 50 mL
with water, and the elements were analyzed by ICP-OES.
3
3

Prior to the mineral acid treatment, elemental analysis of support materials S1 and S2, i.e., carbon,
hydrogen, nitrogen, sulfur, and oxygen, was conducted on a Perkin Elmer 2400 Series II CHNS/O
device. These measurements were conducted in triplicate.
X-ray diffraction (XRD) patterns for the unreduced, thermally treated catalyst samples were
recorded on a PANalytical X’Pert Pro X-ray diffraction system using monochromatic CuKα1 radiation
(λ = 1.5406 Å) at 45 kV and 40 mA. Diffractograms were recorded in a 2θ range of 8–90° at intervals of
0
.017° and with a scan step time of 80 s. Crystalline phases and structures were analyzed by HighScore
Plus software (Version 4.0, PANalytical BV, Almelo, The Netherlands). The peaks were identified
using International Centre for Diffraction Data ( ICDD, PDF-4+ 2020).
X-ray photoelectron spectroscopy (XPS) profiles were recorded for the unreduced, thermally
treated catalyst samples on a Thermo Fisher Scientific ESCALAB 250Xi XPS System. The catalyst
samples were placed on an indium film with a pass energy of 20 eV and a spot size of 900 µm, and the
accuracy of the reported binding energies (BE) was ±0.2 eV. Ni, Cu as well as O, C, and N XPS profiles
were measured for all samples. Measurement data were analyzed by Avantage V5. The monochromatic
AlKα radiation (1486.7 eV) was utilized from an X-ray gun operated at 20 mA and 15 kV. Charge
compensation was utilized to determine the obtained spectra, and the calibration of the BE was
performed by utilizing the C1s line at 284.8 eV as the reference.
The morphology of the unreduced, thermally treated catalyst particles was observed on a JEOL
JEM-2200FS energy-filtered transmission electron microscope equipped with a scan generator
(EFTEM/SEM). Catalyst samples were dispersed in pure ethanol and pretreated in an ultrasonic bath for
several minutes to create a microemulsion. A small drop of the microemulsion was deposited on a
copper grid precoated with carbon (Lacey/Carbon 200 Mesh Copper) and evaporated in air at room
temperature. For the measurement, an accelerating voltage of 200 kV was utilized, and the resolution of

the scanning TEM (STEM) image was 0.2 nm. Particle sizes of metal and metal oxides (both Cu and
Ni) were calculated from TEM images using a PDF-Xchange Editor program.
2
.5.
Furfural hydrotreatment experiments
Furfural hydrotreatment experiments were conducted in a 50-mL batch reactor (Parker Autoclave
Engineers). In the experiments, 0.2 g (with an accuracy of four decimals) of the catalyst was reduced in
situ (523 K, 2 h, 40 bar H ), followed by the mixing of 1 mL of furfural and 15 mL of solvent (i.e., 2-
propanol) and their subsequent addition into the reactor from a pressurized feed tank under a H
2
2
pressure of 40 bar at the desired reaction temperature of 503 K. The reactor was heated to the desired
reaction temperature before adding the feed. A stirring speed of 800 rpm was used for all experiments,
and the catalyst formed a slurry with the reaction mixture. After the reaction, the reactor was cooled to
room temperature with an ice bath, and a sample was taken from the gas phase into an evacuated
container. Another sample was taken from the liquid phase for analysis. Reactions times of 30, 120, and
3
00 min were applied.
Product analysis was performed by a similar method as that reported in our previous study [49].
Gas samples were analyzed on an Agilent 6890 Series gas chromatography (GC) system with a flame
ionization detector (FID) and a thermal conductivity detector (TDC). CO, CO , H , and N were
2
2
2
analyzed by TCD, which was connected to two columns: HP‐ PLOT/Q (30 m × 0.53 mm × 40 μm) and
HP Molesieve (30 m × 0.53 mm × 25 μm) columns. Hydrocarbons were analyzed by FID, which was
connected to an HP‐ AL/KCL column (50 m × 0.32 m × 8 μm). The heating program started from

1
3
13ꢀK (holding time of 9.5 min) at a heating rate of 10ꢀK min up to 473 K. The inlet temperature was
4
73 K. The liquid samples were analyzed on an Agilent 6890 Series GC system, equipped with an FID
and a Zebron ZB‐ wax Plus column (60 m × 0.25 mm × 0.25 μm). The inlet temperature was 503 K,
−
1
and the heating program started at 313 K at a heating rate of 5 K min until 373ꢀK and at a heating rate
−
1
of 20ꢀK min until 503 K. The injection volume was 1 μL, and the internal standard was 2‐ butanol.

Most of the compounds were calibrated. For those compounds that could not be calibrated, FID
response factors were estimated according to the study reported by Scanlon and Willis [50] and the
corrections provided by Jorgensen et al. [51]. For compound identification, Agilent GC–MS (7890‐
5
975) was employed using a similar column and temperature program as described earlier. Mass spectra
were recorded at an electron impact ionization of 70 eV.
Conversion, selectivity, and batch residence times were calculated as follows. Furfural conversion
(X) was calculated using Equation 2,
X = (CF0  CFt)/CF0
,
(2)
where CF0 is the concentration of furfural (mmol g1sample) in the feed, and CFt is the concentration of
furfural (mmol g1sample) at the reaction time t. Product selectivity (S) and yield (Y) were calculated
using Equation 3 and 4, respectively,
S
i
= Cit/(CF0  CFt),
= Cit/CF0,
(3)
(4)
Y
i
where Cit is the concentration of the product i (mmol g1sample) at the reaction time t. Batch residence
time (τ) [gcat min greactant1] was calculated according to references [52,53] and used instead of reaction
time to increase accuracy.
τ = (mcat t)/mreactant,
(5)
where mcat is the mass of the catalyst, and mreactant the mass of furfural added to the reactor.
3
. Results and Discussion
3
.1. Characterization of the AC foams and catalysts

Table 1 provides the detailed description of sample names. Two ACF were prepared, i.e., physical
activation with steam, which was microporous (denoted S1), and chemical activation with ZnCl , which
was mainly mesoporous (denoted S2), respectively. Surface properties of the foams were tailored by
acid washing with HNO (S1_A1 and S2_A1) and H SO (S1_A2 and S2_A2). In addition, as
references, two commercial AC supports were used, i.e., steam-activated (Norit_S) and acid-washed
Norit_A) supports, respectively. In addition, a biobased AC support from spruce (S3, for reference)
2
3
2
4
(
was prepared by the steam activation of sawdust.
Table 1
Detailed description of supports and catalyst used in this study.
Sample name
S1
Type
Description
Support
Support
Support
Catalyst
Catalyst
Catalyst
Support
Support
Support
Catalyst
Catalyst
Catalyst
Support
Catalyst
Support
Catalyst
Support
Catalyst
Support 1, prepared from tannic acid and activated with steam.
S1_A1
S1, treated with HNO
3
.
S1_A2
S1, treated with H SO .
2
4
S1_Cu/Ni
S1_A1_Cu/Ni
S1_A2_Cu/Ni
S2
S1, impregnated with Cu and Ni, thermally treated, ready catalyst.
S1_A1, impregnated with Cu and Ni, thermally treated, ready catalyst.
S1_A2, impregnated with Cu and Ni, thermally treated, ready catalyst.
Support 2, prepared from pine bark extracts and activated with ZnCl
S2, treated with HNO
S2, treated with H SO
S2, impregnated with Cu and Ni, thermally treated, ready catalyst.
2
.
S2_A1
3
.
S2_A2
2
4
.
S2_Cu/Ni
S2_A1_Cu/Ni
S2_A2_Cu/Ni
S3
S2_A1, impregnated with Cu and Ni, thermally treated, ready catalyst.
S2_A2, impregnated with Cu and Ni, thermally treated, ready catalyst.
Steam-activated carbon prepared from spruce sawdust.
S3_Cu/Ni
Norit_S
S3, impregnated with Cu and Ni, thermally treated, ready catalyst.
Commercial steam-activated carbon support material.
Norit_S_Cu/Ni
Norit_A
Norit_S, impregnated with Cu and Ni, thermally treated, ready catalyst.
Commercial steam-activated carbon support treated with acid.
Norit_A, impregnated with Cu and Ni, thermally treated, ready catalyst.
Norit_A_Cu/Ni
3
.2. Specific surface area and pore size distributions
Table 2 summarizes the SSA and PSD results. The prepared S1 support exhibited a moderate SSA
and PV. According to the DFT model, the S1 support mainly comprised micropores. t-Plot calculation,
which used the Harkins and Jura method [47], and the adsorption isotherm (Fig. S1 in Supplementary
Material) confirmed that the pores created in the S1 support are almost completely micropores. By the
treatment of the S1 support with acids, SSA and PV decreased, most probably due to the collapse of

pores [54]. According to the results shown in Table 2, compared to the S1 support, the S2 support
exhibited a higher SSA and PV, with a 55 vol% mesoporous structure. Similarly, acid treatment (HNO
3
)
led to the decrease in the SSA and PV of S2. However, sulfuric acid treatment did not affect the SSA or
PV of S2; the reason for this result is not clear thus far. The SSA and PV of Norit_A were similar to
those of the S2 support. A major difference between these two supports was the PSD. A similar
difference was observed between S3 and Norit_S supports: After impregnation and thermal treatment,
the SSA and PV mainly decreased.

Table 2
2
Specific surface area (SSA), pore volume (PV), pore size distribution (PSD), and average pore size of the prepared supports and catalysts analyzed by N adsorption isotherms.
Prepared supports/catalysts
Commercial supports/catalysts
Catalyst
Calculation method
Unit
Catalyst
S1_A1
Cu/Ni
Catalyst
S1_A2
Cu/Ni
Catalyst
S2_A1
Cu/Ni
Catalyst
S2_A2
Cu/Ni
Catalyst
S3
Catalyst
Norit_S
Cu/Ni
Support Support Support
Support Support Support
Support
S3
Support
Support
Norit_A
Cu/Ni
S1
S1_A1
S1_A2
S2
S2_A1
S2_A2
Norit_A
Norit_ S
Cu/Ni
BET
2
1
SSA
m g
560
332
381
549
469
1361
0.89
911
1395
0.91
715
1151
0.71
891
683
1381
0.63
1152
0.53
961
756
3
1
Pore volume
t-Plot
cm g
0.24
0.14
0.16
0.24
0.21
0.53
0.38
0.60
0.46
0.43
0.34
3
1
Micropore volume
Micropore area
External surface area
DFT
cm g
0.19
482
79
0.11
280
52
0.13
324
56
0.19
469
80
0.16
405
64
0.07
148
0.13
297
613
0.07
148
0.14
348
367
0.10
228
924
0.20
486
405
0.15
362
321
0.41
1023
358
0.34
859
293
0.32
810
151
0.25
630
126
2
1
m g
2
1
m g
1213
1247
Average pore size
Pore volume
Micropores
Mesopores
Macropores
Nm
<1.5
0.20
96
<1.5
0.12
93
<1.5
0.13
96
<1.5
0.19
95
<1.5
0.17
95
2.2
0.76
45
55
0
1.8
0.44
57
43
0
2.2
0.78
45
55
0
1.9
0.31
69
31
0
2.1
0.61
50
50
0
2.0
0.51
50
50
0
2.0
0.39
49
51
0
1.6
0.51
89
11
0
1.5
0.42
90
10
0
1.5
0.34
94
6
1.5
0.27
93
7
3
1
cm g
%
%
%
4
4
4
5
5
0
3
0
0
0
0
0

3
.3. Compressive strength and elemental analysis of S1 and S2 supports
Table 3 summarizes the results obtained from compressive strength measurements. A clear
difference in compressive strengths between uncarbonized (no thermal treatment) and carbonized
(high-temperature thermal treatment) S1 and S2 supports was observed, which was in good
agreement with previously published results [39]. The compressive strength of the support before
and after carbonization was compared to confirm that a more stable structure is obtained by thermal
treatment. The matured support S1 exhibited a compressive strength of 0.040 MPa, and after
thermal treatment, its compressive strength was nearly three times higher (0.142 MPa). The same
observation was made for the S2 support: After thermal treatment, the mechanical strength was two
times higher. Notably, the comparison of the compressive strengths between S1 and S2 supports
revealed significant differences. Uncarbonized S2 was almost 30 times stronger than uncarbonized
S1. In addition, the thermally treated S2 support was more than 15 times stronger than the S1
support.
Such a large difference in the compressive strength between S1 and S2 was related to the
different compositions of the raw materials. The S1 support was prepared using tannic acid, while
support S2 was prepared using pine bark extracts extracted with a 30/70 w% water/ethanol solution.
Although the exact composition of the extracts was not known, but the extracts undoubtedly
contained sugars, lignin, and different tannins. Carbon foams prepared using tannic acid are known
to be rather fragile (Table 3). According to results obtained herein, the mechanical stability of
tannic-acid-based carbon foams can be enhanced by using a complex mixture of phenolic
substances. An increased number of crosslinking molecules, such as lignin and sugars, are present
in the foam solution, which in turn increase the rigidity of foam structure during polymerization.
Table 3
Compressive strengths of matured precursor and thermally treated support materials (S1 and S2).
Support
Matured
Thermally treated

Precursor to S1
.040
Precursor to S2
1.15
Support S1
0.142
Support S2
2.29
Average compressive
strength (MPa)
0
Table 4 below summarizes the elemental analysis of the thermally treated supports S1 and S2. As
expected, both supports exhibited a high carbon content of greater than 80 w%. Sulfur and nitrogen
contents of the supports were similar, but higher contents of hydrogen (33%) and oxygen (15%)
were observed in support S1, related to the different compositions of the raw material.
Table 4
Elemental analysis for carbon, hydrogen, nitrogen, sulfur, and oxygen determined on the Perkin Elmer 2400 Series II CHNS/O
device.
Determined values
Sample
C (w%)
H (w%)
14 ± 2
N (w%)
1.1 ± 0.4
0.7 ± 0.1
S (w%)
<0.1
O (w%)
4.6 ± 0.6
3.9 ± 0.1
Support S1
Support S2
80 ± 5
86 ± 4
9.4 ± 0.5
<0.1
3
.4. Metal content
Metal contents of the catalysts were estimated by ICP-OES. Table 5 summarizes the results.
The target metal contents of Ni and Cu in the supported catalysts were 5 w%. The measured metal
contents of the impregnated and thermally treated S1_A2_Cu/Ni, S2,_A2_Cu/Ni, S3_Cu/Ni,
Norit_S_Cu/Ni, and Norit_A_Cu/Ni catalysts were close to the target values. However, the contents
of nickel and copper in S1_A1_Cu/Ni and S2_A1_Cu/Ni were higher than expected. This result
was possibly related to the acid treatment with HNO , which increased the oxygen content of the
3
supports (Table S1 in the Supplementary Material). The surface oxygen might have burned away
some carbon from the surface during thermal treatment; hence, the weight percentage of the metal
in the catalysts increases. Moreover, the content of impurities such as Zn and S, which remained
after the pretreatment of the catalysts, was measured. The limits of detection for Zn metal and sulfur
were 0.4 mg/kg and <20000 mg/kg, respectively. Zn impurity levels were relatively low, most

probably related to the reactor used for the activation of supports or thermal treatment of the
catalysts. Most of the catalysts exhibited S content of less than 2 w%. Only S2_A2_Cu/Ni exhibited
a slightly higher S content due to the sulfuric acid treatment. In addition, compared to
S1_A2_Cu/Ni, S2_A2_Cu/Ni exhibited a higher PV (Table 2), indicating that the adsorption of
sulfur-containing groups inside the pores of S2_A2_Cu/Ni is better than that of S1_A2_Cu/Ni.
Table 5
Elemental contents of the samples determined by ICP-OES.
Target values of
Cu and Ni
Determined values
Cu (w%)
Sample
(w%)
Ni (w%)
7.2
Zn (w%)
0.0060
0.014
S (w%)
<2
S1_A1_Cu/Ni
S1_A2_Cu/Ni
S2_A1_Cu/Ni
S2_A2_Cu/Ni
S3_Cu/Ni
5.0
5.0
5.0
5.0
5.0
5.0
5.0
7.0
5.0
6.6
5.7
4.3
3.8
4.3
5.1
<2
6.5
0.0030
0.013
<2
5.6
2.2
<2
4.3
0.0040
0.019
Norit_S_Cu/Ni
Norit_A_Cu/Ni
4.0
<2
4.7
0.0020
<2
3
.5. X-ray photoelectron spectroscopy and X-ray diffraction analysis
XPS (detection depth < 10 nm) was employed to characterize the surface functionalities of the
AC supports. XPS data, i.e., C1s, O1s, and N2p profiles, of untreated and acid-treated supports were
analyzed. Table S1 and S2 in Supplementary Material summarize the XPS results of the supports
and catalysts, respectively. From the C1s spectra, carbon-containing groups were deconvoluted into
five peaks based on (BE) (Table S1 and S2 and Fig. S4(a) in Supplementary Material), i.e., carbon–
carbon bonds (BE = 284.8 eV), carbon species in alcohol or ether groups (BE = 286.3–287.0 eV),
carbon in carbonyl groups (BE = 287.5–288.1 eV), carbon in carboxyl or ester groups (BE = 289.3–
2
90.0 eV), and shake-up satellites due to π–π* transition in aromatic rings (BE = 291.2–292.1 eV),

respectively [55]. According to C1s scans, the support S1 and S2 surfaces predominantly comprised
carbon–carbon-type bonds, with some oxygen-containing functional groups. In addition, low levels
of aromatic rings (from π–π* transitions) were detected. In addition, the total oxygen content from
O1s data was calculated (Tables S1 and S2 in Supplementary Material), which revealed some
oxygen functionalities (≤10% from surface functionalities) on the S1 and S2 supports. Moreover,
the surface-oxygen functionalities on S2 were slightly less (4%) than those on S1 (10%).
Commercial and S3 supports exhibited a similar surface composition (C1s, O1s), especially S1. The
highest amount of oxides was observed on support S1_A1 (Table S1 and Fig. S4(b)), which was
four times greater than that observed for the untreated S1 support. In addition, H
to the increase in the oxygen content on support S1, and it was two times greater than the oxygen
content for the untreated S1 support. For S2, the same effect was observed, and HNO treatment led
to the increase in the oxygen functionalities on the support in comparison with that observed for
SO treatment. Overall, compared to support S2, support S1 exhibited a higher amount of oxides
as detected by XPS. Supports exhibited a small amount of nitrogen as observed by XPS N2 scans
<0.5%); however, HNO treatment possibly led to the increase in the nitrogen content due to the
2
SO
4
treatment led
3
H
2
4
p
(
3
addition of nitrates on the surface, which were observed at 405 eV in the N12p XPS spectrum.
XPS verified the successful addition of metals in the case of impregnated (thermal treatment
and unreduced) catalysts. XPS data for the Cu2p and Ni2p spectra revealed the presence of metal
oxides (such as CuO and NiO) in thermally treated, unreduced catalysts (Table S2). The Cu2p scans
revealed peaks at ~932 and ~933 eV for all catalysts, which were attributed to Cu metal and metal
oxides of Cu
eV), XRD measurements were also carried out to identify the phases. For the S1 (HNO
SO )-supported catalyst, satellites at ~943 eV were detected, indicative of the presence of Cu(II)
Fig. S5(a) S1_A2_Cu/Ni). This result was in agreement with XRD results (Fig. S3): CuO was
2
O and CuO, respectively [56]. Owing to similar BE of Cu (932.6 eV) and Cu
2
O (932.7
3
and
H
2
4
(

present (at 2 = 35.5° and 38.7°), and a metal copper phase was absent. For S2-supported catalysts,
satellite peaks at 943 eV were not detected, indicative of the presence of Cu(I) or the Cu metal (Fig
S5(b) S2_A2_Cu/Ni); this result was also in agreement with the XRD results: CuO and Cu
were not observed. From Ni2p scans (Fig. S6), the main peak at ~854 eV and the broad satellite at
60 eV revealed the presence of nickel oxide NiO and/or Ni in all catalysts [57]. In addition, for S3
2
O peaks
8
and commercial AC-supported catalysts, the presence of copper and nickel oxides was verified.
Compared with that of S1_A2, the O1s scan of S1_A2_Cu/Ni revealed an increased peak
intensity at a BE range of 529–530 eV, indicative of the presence of M–O bonds. Compared with
the O1s (BE = 531.2 eV) of S1_A2 (5.6 atom%, Table S1), the O1s (BE = 531.2 eV) of
S1_A2_Cu/Ni was 11.7 atom% (Table S2), indicative of the increase in the number of M–O bonds
by the impregnenation of Cu and Ni. XRD analysis was conducted for seven catalyst samples. Fig
S3 in Supplementary Material shows the results. Cu (ICDD 04-002-8854) and Ni (ICDD 04-004-
2
759) metals were partly oxidized in all catalysts, except in the case of S1_A2_Cu/Ni, where total
oxidation was observed. CuO (ICDD 00-048-1548), Cu2O (ICDD 04-016-6875), and NiO (ICDD
4-023-3539) were formed. Additional peaks were observed at 41.7°, 48.6°, and 71.2° for
S1_A1_Cu/Ni, which were not identified.
0
3
.6. Electron microscopy measurements
The morphology of the thermally treated, unreduced catalyst particles was investigated by
EFTEM in the STEM mode and by scanning electron microscopy (SEM). On the surface of
supports, which were treated with acids A1 and A2, metals appeared to be quite evenly distributed
(Fig. S2 in Supplementary Material). In all catalysts, especially nickel catalysts, homogeneously
distributed particles (10–40 nm) were observed with a smaller particle size variation over the
support than that observed for copper, where larger aggregates (25–250 nm) were observed, and the

variation in the particle size was broader. On the S1 support, metal particle sizes for S1_A1_Cu/Ni
were 60–70 nm and 10–20 nm for copper and nickel, respectively, and for S1_A2_Cu/Ni, particle
sizes were 100–200 nm and 20–40 nm for copper and nickel, respectively. Particle sizes for
S2_A1_Cu/Ni (Fig. 2c) were 25–50 nm and 7–10 nm for copper and nickel, respectively. On
S2_A2_Cu/Ni (Fig. 2d), the variation in particle sizes for copper and nickel was 50–125 nm and
1
0–20 nm, respectively. Overall, the particles seemed to be smaller on support S2 than on support
S1.
Variation in particle sizes between S1 and S2 supports can be partly explained by different PV
and PSD (Table 1). The PV and pore size of S1 were less than those of the S2 support, possibly
resulting in the agglomeration of metal particles on the catalyst surface as some micropore
entrances can be blocked by large metal hydroxides [58]. In addition, notably, compared to H
treated supports, HNO -treated S1 and S2 supports exhibited smaller metal particles, indicating that
a higher surface oxygen content might be beneficial for the metal distribution; hence, the particle
size is affected [59]. From the S3-supported catalyst, particle sizes for copper and nickel were 100–
50 nm and 15–40 nm, respectively. Particle sizes for copper in Norit_S_Cu/Ni and Norit_A_Cu/Ni
were 15–45 nm and 25–80 nm, respectively; the corresponding values for nickel were 7–15 nm and
5–25 nm. Surface structures of the supports S1 and S2 were examined by SEM. The SEM images
Fig. S7 in the Supplementary Material) clearly revealed the presence of large ordered pores in both
2
SO
4
-
3
2
1
(
supports (S1 and S2). Fig. S8 in the Supplementary Material summarizes the size distribution,
including Ni and Cu particles, which is calculated from the TEM images shown in Fig. 2. As an
example, for S1_A1_Cu/Ni, 82.2% of the particle sizes were observed in the range of 8–28 nm, and
1
1.0% of the particles were observed in the range of 68–88 nm.

3
.7. Conversion of furfural to 2-methylfuran
Scheme 1a and Scheme 1b show the furfural hydrotreatment and observed condensation
products, respectively. The hydrotreatment of furfural to produce MF as the target product was
investigated with the prepared 5–5% CuNi catalysts on ACF (S1_Cu/Ni, S1_A1_Cu/Ni,
S1_A2_Cu/Ni, S2_Cu/Ni, S2_A1_Cu/Ni, and S2_A2_Cu/Ni). In addition, 5–5% CuNi catalysts
were prepared on two commercial AC supports (i.e., Norit_S_Cu/Ni and Norit_A_Cu/Ni,
respectively) and on biobased AC from spruce (S3_Cu/Ni) and tested in the reaction just for
comparison. The main interest is in the carbon foams. In addition to MF, FA, THFA, MTHF, and
furan were obtained. Moreover, furfural ring-opening products, namely 5-hydroxy-2-pentanone
(HPN), 1,4-pentanediol (PDOL), and 2-pentanone (PN), were observed. In addition, acetone was
formed, which can be produced by 2-propanol via dehydrogenation or hydrogenation transfer,
where 2-propanol acts as a hydrogen donor and concurrently forms acetone [60]. Two condensation
products, i.e., furanmethanol acetate (FMA) and 2-(2-furylmethyl)-5-methylfuran (FMMF), were
also obtained.

Fig. 3a shows the conversion of furfural as a function of the batch residence time (τ, defined as
the catalyst mass divided by the reagent mass and multiplied with the reaction time) for the ACF-
supported catalysts. Fig. 3b shows the conversion of furfural over the acid-washed ACF catalysts,
and Fig. 3c shows the conversion over the reference catalysts. From the ACF supports, S1 was more
active, leading to a furfural conversion of 100% after a reaction time of 300 min. The higher
activity of the S1_Cu/Ni catalyst compared with that of the S2_Cu/Ni catalyst was slightly
surprising as the surface area of S1 was significantly lower (Table 2), and most of the pores in S1
were micropores, where the reaction can suffer from mass-transfer limitations; however, this was
not observed for the reference catalysts. Overall, the selectivity of MF over ACF-based catalysts
was not desirable.

Acid washing exhibited an interesting effect on the catalytic activity (Fig. 3b). With HNO
3
washing, the activity of the S1-supported catalyst decreased in the beginning of the reaction; this
activity decrease was possibly related to the adverse effect of the acid treatment on the SSA and PV
of the S1 catalyst; hence, the catalytic activity decreases. However, with support S2, a similar trend
was not observed. For the support S2, HNO
throughout the range of the batch residence time. On the other hand, washing with H
significant decrease in the activity of the S1 supported catalyst. Acid washing with H
3
washing led to the increase in the catalytic activity
2
SO
4
led to the
can
2
SO
4
create anchoring sites for metal particles, leading to large agglomerates; the formation of these
agglomerates could also decrease the activity. The loss in activity is also speculated to result from
the sulphur-containing groups from the support after the washing step. In case of the S2 support,
washing with H
washing, washing with HNO
For comparison, the activities of CuNi on commercial supports and biobased AC were
2
SO
4
did not affect the catalytic activity. For both supports, compared to H
2
SO
4
3
led to a more active catalyst.
measured. Compared to most of the ACF-based catalysts, the reference catalysts exhibited higher
activity (Fig. 3). Only CuNi supported on S1 (without acid washing) exhibited activity similar to
those of the reference catalysts. Norit_A_Cu/Ni was the most active reference catalyst at low batch
residence times, which is an acid-washed AC. At high batch residence times, a furfural conversion
of almost 100% was achieved over all of the reference catalysts.
Fig. 4 shows the MF selectivity as a function of conversion. Compared to the S1_Cu/Ni
catalyst, the S2_Cu/Ni catalyst exhibited higher MF selectivity even though the catalyst was less
active (Fig. 3a). In general, the acid-washed catalysts were more selective toward MF. Among the
acid-washed catalysts, HNO
similar conversion of ~90% (Fig. 3b). However, for the H
60% (Fig. 3b) led to a similar MF selectivity. At a conversion of greater than 90%, the
3
-washed S2 exhibited the highest selectivity for MF as evidenced by a
2
SO -washed catalysts, a conversion of
4
~

S2_A1_Cu/Ni catalyst exhibited MF selectivity almost similar to those of the reference catalysts. In
addition, the S2_A2_Cu/Ni catalyst exhibited an MF selectivity of >40%, but as the catalyst was
less active, the conversion was less than 90%.
Table 6
Furfural conversion at the highest observed MF yield, the corresponding reaction time, and the main side-product yields.[a]
Catalyst
X
t
τ
Y
[%]
FA
Y
[%]
MF
Y
[%]
THFA
Y
[%]
MTHF
Y
[%]
Furan
Y
FMA
[%]
[gcat greactant
1
[
%]
[min]
min]
S1_Cu/Ni
99
86
93
91
66
85
96
98
99
300
300
300
300
300
300
120
120
300
60
60
60
60
60
61
23
24
61
71
14
18
37
48
20
38
61
57
58
18
0.0
17
4.9
0.0
3.8
0.7
0.2
0.1
4.1
5.1
4.5
0.1
6.3
4.2
5.3
5.6
6.1
0.2
0.3
0.0
S2_Cu/Ni
5.1
16
6.6
6.6
3.3
1.7
1.3
7.2
3.2
8.1
S1_A1_Cu/Ni
S2_A1_Cu/Ni
S1_A2_Cu/Ni
S2_A2_Cu/Ni
Norit_S_Cu/Ni
Norit_A_Cu/Ni
S3_Cu/Ni
3.4
2.9
1.9
2.3
2.2
10
3.6
13
5.5
8.0
1.1
2.9
3.0