Screening of catalysts for hydrodeoxygenation of phenol as a model compound for bio-oil

Article abstract of DOI:10.1021/cs400266e

Four groups of catalysts have been tested for hydrodeoxygenation (HDO) of phenol as a model compound of bio-oil, including oxide catalysts, methanol synthesis catalysts, reduced noble metal catalysts, and reduced non-noble metal catalysts. In total, 23 different catalysts were tested at 100 bar H₂ and 275 C in a batch reactor. The experiments showed that none of the tested oxides or methanol synthesis catalysts had any significant activity for phenol HDO under the given conditions, which were linked to their inability to hydrogenate the aromatic ring of phenol. HDO of phenol over reduced metal catalysts could effectively be described by a kinetic model involving a two-step reaction in which phenol initially was hydrogenated to cyclohexanol and then subsequently deoxygenated to cyclohexane. Among reduced noble metal catalysts, ruthenium, palladium, and platinum were all found to be active, with activity decreasing in that order. Nickel was the only active non-noble metal catalyst. For nickel, the effect of support was also investigated and ZrO₂ was found to perform best. Pt/C, Ni/CeO₂, and Ni/CeO₂-ZrO ₂ were the most active catalysts for the initial hydrogenation of phenol to cyclohexanol but were not very active for the subsequent deoxygenation step. Overall, the order of activity of the best performing HDO catalysts was as follows: Ni/ZrO₂ > Ni-V₂O₅/ZrO ₂ > Ni-V₂O₅/SiO₂ > Ru/C > Ni/Al₂O₃ > Ni/SiO₂ Pd/C > Pt/C. The choice of support influenced the activity significantly. Nickel was found to be practically inactive for HDO of phenol on a carbon support but more active than the carbon-supported noble metal catalysts when supported on ZrO₂. This observation indicates that the nickel-based catalysts require a metal oxide as a carrier on which the activation of the phenol for the hydrogenation can take place through heterolytic dissociation of the O-H bond to facilitate the reaction.

Full text of DOI:10.1021/cs400266e

Subscriber access provided by DUESSELDORF LIBRARIES
Article
Screening of Catalysts for Hydrodeoxygenation
of Phenol as Model Compound for Bio-oil
Peter Mølgaard Mortensen, Jan-Dierk Grunwaldt, Peter Arendt Jensen, and Anker D. Jensen
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/cs400266e • Publication Date (Web): 01 Jul 2013
Downloaded from http://pubs.acs.org on July 2, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 36
ACS Catalysis
1
2
3
4
5
6
7
8
9
Screening of Catalysts for Hydrodeoxygenation of
Phenol as Model Compound for Bio-oil
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†
‡
†
Peter M. Mortensen, Jan-Dierk Grunwaldt, Peter A. Jensen, and
∗,†
Anker D. Jensen
Department of Chemical and Biochemical Engineering, Technical University of Denmark,
Søltofts Plads, Building 229, DK-2800 Lyngby, Denmark, and Institute for Chemical Technology
and Polymer Chemistry, Karlsruhe Institute of Technology (KIT), Engesserstrasse 20, D-79131
Karlsruhe, Germany
E-mail: aj@kt.dtu.dk
Abstract
Four groups of catalysts have been tested for hydrodeoxygenation (HDO) of phenol as a
model compound of bio-oil, including: oxide catalysts, methanol synthesis catalysts, reduced
noble metal catalysts, and reduced non-noble metal catalysts. In total 23 different catalysts
◦
were tested at 100 bar H2 and 275 C in a batch reactor. The experiments showed that none
of the tested oxides and methanol synthesis catalysts had any significant activity for phenol
HDO at the given conditions, which were linked to their inability to hydrogenate the phenol.
HDO of phenol over reduced metal catalysts could effectively be described by a kinetic model
involving a two-step reaction were phenol initially was hydrogenated to cyclohexanol and then
subsequently deoxygenated to cyclohexane. Among reduced noble metal catalysts ruthenium,
palladium, and platinum were all found to be active, with decreasing activity in that order.
∗To whom correspondence should be addressed
†Technical University of Denmark
‡Karlsruhe Institute of Technology
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 36
1
2
3
4
5
6
7
8
9
Nickel was the only active non-noble metal catalyst. For nickel, also the effect of support
was investigated and ZrO2 was found to perform best. Pt/C, Ni/CeO2, and Ni/CeO2-ZrO2
were the most active catalysts for the initial hydrogenation of phenol to cyclohexanol, but
were not very active for the subsequent deoxygenation step. Overall, the order of activity
of the best performing HDO catalysts was: Ni/ZrO2 > Ni-V2O5/ZrO2 > Ni-V2O5/SiO2 >
Ru/C > Ni/Al2O3 > Ni/SiO2 >> Pd/C > Pt/C. The choice of support influenced the activity
significantly. Nickel was found to be practically inactive for HDO of phenol on a carbon
support, but more active than the carbon supported noble metal catalysts when supported on
ZrO2. This observation indicates that the nickel based catalysts require a metal oxide as carrier
on which the activation of the phenol for the hydrogenation can take place through heterolytic
dissociation of the O-H bond to facilitate the reaction.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords: Bio-oil, Catalyst-screening, Hydrodeoxygenation, Noble-metals, Phenol, Oxides,
Reduced-metals
1 Introduction
A prospective route for production of bio-fuels in the future is the conversion of biomass into
bio-oil through flash pyrolysis followed by upgrading via hydrodeoxygenation (HDO).1 Flash py-
rolysis is advantageous in order to make a locally produced liquid which minimizes transportation
costs to larger bio-refineries.2–4 However, the bio-oil is a viscous, polar, and acidic liquid with a
low heating value, making it in most cases unsuitable as an engine fuel directly. These unfavorable
characteristics are all associated with high contents of water (10-30 wt%) and oxygen containing
organic compounds (30-40 wt% oxygen) in the oil.5
In HDO the bio-oil is treated with hydrogen at a pressure of up to 200 bar and temperatures
in the range from 200-400 ◦C. This converts the oxy-compounds to a hydrocarbon product which
separates from the water and ultimately gives a product equivalent to crude oil. The reaction can
2
ACS Paragon Plus Environment

Page 3 of 36
ACS Catalysis
1
2
3
4
be generally written as (normalized to feed carbon):1
5
6
7
8
9
CH1.4O0.4 +0.7H2 −−→ 1”CH2”+0.4H2O
Here ”CH2” represents an unspecified hydrocarbon as product.
(1)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
One of the major challenges in this concept is to find a catalyst with a high activity for the
deoxygenation reaction and at the same time obtain a sufficient lifetime, as deposition of carbona-
ceous species has proven to be a severe problem.1,6 Preferably, HDO catalysts should be relatively
◦
inexpensive and function at low temperature (<300 C) and low pressure (<100 bar). Especially
low temperatures are desirable to prevent coking1,7 and therefore focus has in the current screening
◦
study been on catalysts operating at 275 C. Four categories of catalysts were tested as catalysts
for HDO of phenol:
1. Oxide catalysts
2. Methanol synthesis catalysts
3. Reduced noble metal catalysts
4. Reduced non-noble metal catalysts
Phenol was chosen as model compound of bio-oil, as phenols have been identified among the most
persistent,8 yet fairly abundant,9 compounds in the bio-oil.
Oxide catalysts have been proposed to catalyze the reaction through roughly three steps as
shown in Figure 1(a): chemisorption via the oxygen atom on a coordinatively unsaturated metal
site, proton donation from a hydroxyl group, and desorption.1,10 In this mechanism, the generation
of vacancy sites (the dotted circle in the figure) is responsible for the activation of both the oxy-
compound and hydrogen on the catalytic surface.1,10
Methanol synthesis catalysts have proven able to activate CO/CO2 through the oxygen.13–16
Thus, these catalysts could potentially be able to activate oxy-compounds through their oxygen
group and in this way enable reaction.
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 36
1
2
3
4
5
6
7
8
H2
9
H2O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
H
H
OH
H
H
(a) Oxide catalysts reaction mechanism
(b) Reduced metal catalysts reaction mechanism
Figure 1: Proposed reaction mechanisms on oxide catalysts1,10 and reduced metal catalysts.11,12
Figure 1(b) shows the key concepts of the reaction mechanism for reduced metal catalysts.1
The reaction is initiated by adsorption of the oxy-compound on the catalyst surface. Adsorbed
hydrogen on the active metal clusters reacts with the oxy-compound to facilitate the deoxygenation.
This is followed by desorption of the final product. The oxy-compound adsorption step can take
place on either the support or directly on the active metal, depending on which type of metal is
used. Specifically noble metals have been shown to be capable of activating the oxy-compound
on the metal sites.12,17 For non-noble metals, the activation is thought to occur through an oxygen
vacancy site in the support metal oxide,11,18 similar to the activation step for the oxide catalysts
shown in Figure 1(a). Hence, these four different classes of catalysts were screened in this work.
2 Experimental Section
2.1 Catalyst Synthesis
MnO/C, WO3/C, MoO3/C, V2O5/C, Cu/SiO2, NiCu/SiO2, Co/SiO2, Fe/SiO2, Ni based catalysts,
and Ni-V based catalysts were all prepared by incipient wetness impregnation of the corresponding
supports. An overview of the applied catalysts is given in Table 1. The precursors for the active ma-
terials were: Mn(C2H3O2)2 · 4 H2O (Sigma-Aldrich, ≥ 99%), (NH4)6H2W12O40 · xH2O (Sigma-
Aldrich, ≥ 99.0%), (NH4)6Mo7O24 ·4 H2O (Sigma-Aldrich, ≥ 99.0%), NH4VO3 (Sigma-Aldrich,
≥ 99.0%), Cu(NO3)2 · 3 H2O (Sigma-Aldrich, ≥ 99.0%), Fe(NO3)3 · 9 H2O (Sigma-Aldrich, ≥
4
ACS Paragon Plus Environment

Page 5 of 36
ACS Catalysis
1
2
3
4
5
6
98.0%), Fe(NO3)2 ·6 H2O (Sigma-Aldrich, ≥ 98.0%), Ni(C2H3O)2 ·4 H2O (Sigma-Aldrich, 98.0%),
and Ni(NO3)2 ·6 H2O (Sigma-Aldrich, ≥ 97.0%).
7
The active carbon was Daihope 009. The silica was supplied by Saint-Gobain NorPro, type
SS6*138 with a purity of ≥ 99.5%. Also the ZrO2 was supplied by Saint-Gobain NorPro, of the
type SZ6*152 having an impurity of 3.3% SiO2. The alumina was supplied by Sasol and was of
the type Puralox TH 100/150. The spinel was produced from the Al2O3 by mixing stoichiometric
amounts of Al2O3 and MgO and calcination of the mixture at 900 ◦C; the product was confirmed
to be a mixed oxide (MgAl2O4) by XRD. CeO2 and CeO2-ZrO2 were supplied by AMR Ltd. The
CeO2-ZrO2 was confirmed by XRD to be a mixed oxide. Before impregnation all supports were
crushed and sieved.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Incipient wetness impregnation was made by initially dissolving the corresponding amount of
precursors in de-ionized water equivalent to the pore volume of the support and then mixing with
the support. All precursors above were sufficiently water soluble. After impregnation, the samples
were dried at 110 ◦C for at least 12 h and then calcined at 400 ◦C with a heating rate of 10 ◦C/min
and a holding time of 4 h. The catalysts with carbon supports were calcined in nitrogen, while
other catalysts were calcined in air.
For example, NiCu/SiO2 was prepared by dissolving 3.80 g Cu(NO3)2 · 3 H2O and 4.95 g
Ni(NO3)2 ·6 H2O in 8 ml de-ionized water and impregnation of this solution on 8 g of dried SiO2.
The Ni-V2O5 catalysts were made by initially impregnating the support with vanadium, drying
the catalyst, and then impregnating with nickel nitrate. NH4VO3 was dissolved by using addition-
ally oxalic acid (Sigma-Aldrich, ≥ 99.0%) in a molar ratio of 1:2 (V:oxalic acid).
Cu/ZnO/Al2O3 was prepared according to the method of Baltes et al.19 An aqueous solution
of Cu(NO3)2 · 3 H2O (0.6 mol/l, Sigma-Aldrich, ≥ 99%), Zn(NO3)2 · 6 H2O (0.3 mol/l, Sigma-
Aldrich, ≥ 99%), and Al(NO3)3 ·9 H2O (0.1 mol/l, Sigma-Aldrich, ≥ 98%) was coprecipitated by
a solution of Na2CO3 (1 mol/l, Sigma-Aldrich, ≥ 99%) for 1 h. During the precipitation process,
pH was maintained at 7±0.1, and temperature was kept at 60 ◦C. Afterwards, the precipitate was
filtered and washed with demineralized water followed by drying overnight at 80 ◦C and calcining
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 36
1
2
3
4
5
6
7
8
9
Table 1: Summary of catalyst and support properties as the theoretical loading of active component,
particle size of support, specific surface area, and reduction conditions for the tested catalysts.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Catalyst/support
Loading
[wt%]
Particle Size Surface Area Reduction
[µm]
[m2/g]
1100
250
160
140
140
90
T [◦C]/t [h]
-
Carbon
SiO2
ZrO2
-
-
-
-
-
-
-
15
15
15
15
-
-
-
-
-
-
-
-
-
-
CeO2
CeO2-ZrO2
MgAl2O4
Al2O3
MnO/C
WO3/C
MoO3/C
V2O5/C
NiO-MoO3/Al2O3 a
CoO-MoO3/Al2O3
Cu/ZnO/Al2O3
NiCu/SiO2
Cu/SiO2
Ru/C
Pd/C
Pt/C
Co/SiO2
Fe/SiO2
Ni/SiO2
Ni/Al2O3
Ni/CeO2
-
-
-
150
680
790
660
310
-
-
710-1400
710-1400
710-1400
710-1400
300-600
300-600
≈ 900
≈ 900
300-600
≈ 15
None
None
None
None
None
None
300/7
350/19
380/19
400/2
400/2
400/2
550/2
550/2
400/2
550/2
400/2
550/2
550/2
550/2
550/2
550/2
550/2
550/2
a
-
-
60
45%Cu/5%Zn
10%Ni/10%Cu
180
170
1160
1100
1140
210
200
210
140
130
130
120
80
15
5
5
5
5
5
5
5
5
5
5
5
≈ 15
≈ 15
300-600
300-600
63-125
63-125
63-125
63-125
63-125
63-125
63-125
300-600
300-600
300-600
Ni/ZrO2
Ni/CeO2-ZrO2
Ni/MgAl2O4
Ni/C
5
1020
210
120
100
Ni-V O5/SiO2
5%Ni/5%V
5%Ni/5%V
10
Ni-V2O5/ZrO2
V O52/ZrO2
2
aCommercial catalyst by Haldor Topsøe A/S without detailed characterization data.
6
ACS Paragon Plus Environment

Page 7 of 36
ACS Catalysis
1
2
3
4
at 300 ◦C under air for 3 h.
5
6
7
8
NiO-MoO3/Al2O3 and CoO-MoO3/Al2O3 catalysts were obtained from Haldor Topsøe A/S.
Ru/C, Pd/C, and Pt/C were obtained from Sigma-Aldrich.
9
Table 1 also summarizes theoretical loading, particle size, specific surface area, and pretreat-
ment conditions of the different catalysts.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.2 Catalyst Testing
The screening experiments were performed in a 300 ml batch reactor (Parr, type 4566) made from
Hastelloy C steel. 1 g catalyst was loaded in the reactor and hereafter 50 g of phenol (Sigma-
Aldrich, ≥ 99%) was added. The mixture was stirred with a propeller at 380-390 rpm throughout
the experiment and heated to 275 ◦C in a hydrogen atmosphere, giving a final pressure of 100 bar.
The heating rate was ca. 12.5 ◦C/min. During the experiments, hydrogen was added continuously
to maintain the pressure. To stop the experiment, the reactor was placed into an ice bath (cooling
◦
rate of ca. 25-50 C/min). The start of the experiment was taken as the time when the heater
was turned on and the end of the experiment when the reactor was lowered into the ice bath. The
resultant product was filtered by suction filtration to separate the product liquid and catalyst.
Catalytic activity measurements over oxide and methanol synthesis catalysts were performed
with a mixture of 10 g of phenol in 40 ml of de-ionized water as feedstock, because low conversions
were observed in this series of experiments. Addition of water was required as a solvent for phenol,
as the reactant is solid at room temperature and therefore difficult to work with in the product
separation and GC analysis.
In a blank experiment without catalyst, 10 g of phenol and 40 ml of H2O was allowed to react
for 4 h at 275 ◦C and 100 bar. Almost no catalytic activity was seen in this case as a conversion
of 0.3% was observed. Thus, the reactor was hardly catalytically active and did not influence the
experiments. The blank experiment was repeated occasionally to ensure that the reactor was not
contaminated over time.
By threefold repetition of a hydrodeoxygenation experiment with Ni/SiO2, it was found that
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 36
1
2
3
4
5
6
7
8
this procedure typically had an uncertainty in the measured yields within ±2 mole%, correspond-
ing to less than 5% as relative standard deviation. Overall the repeatability of the experiments was
good.
9
In some cases shorter experiments had to be performed not to reach 100% conversion. Here the
batch reactor was initially heated without stirring. It was assumed that the extent of reaction was
low in the heating phase due to the mass transfer restriction of hydrogen in such systems.20 This
was supported by the observation that no hydrogen was consumed in the heating phase. When the
desired temperature was reached, the stirring was started at 380-390 rpm and the experiment could
be performed at close to isothermal conditions. This made it possible to measure the activity in
short term (5-30 min) experiments.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As specified in Table 1, some of the catalysts were pre-treated in hydrogen to reduce the active
metals. This was done in a continuous flow setup, where the sample was treated at the specified
temperature in a 50%/50% mixture of hydrogen and nitrogen at a total flow of 500 Nml/min. Tem-
perature programmed reduction by hydrogen (H2-TPR) was used to evaluate the required reduction
temperature and XRD confirmed that the catalysts were reduced under the specified conditions.
2.3 Product Analysis
Analysis of the liquid product was performed with a Shimadzu GCMS/FID-QP2010UltraEi fitted
with a Supelco Equity-5 column and equipped with a mass spectrometer (MS) for product identi-
fication and a flame ionization detector (FID) for quantification. External standards were prepared
for phenol, cyclohexanol, cyclohexanone, and cyclohexane using ethanol as solvent. The concen-
trations of the remaining peaks were calculated from the FID on the basis of the effective carbon
number method,21 where the concentration of a compound is found as:
Ai νeff,ref
Aref νeff,i
Ci = Cref ·
·
(2)
8
ACS Paragon Plus Environment

Page 9 of 36
ACS Catalysis
1
2
3
4
5
6
7
8
9
Here C is the concentration, A the area of the peak in the FID spectrum, and νeff the effective
carbon number. Index i refers to the compound with the unknown concentration and index ref
refers to a reference compound where the concentration is known. In all calculation with this
formula cyclohexanol was used as reference. The effective carbon number was obtained from the
review by Schofield.21
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The conversion, X, was calculated as:
ꢀ
ꢁ
nPhenol
n0,Phenol
X = 1−
·100%
(3)
Here nPhenol is the moles of phenol after reaction and n0,Phenol is the moles of phenol prior to
reaction.
The yields (Yi) of relevant products were calculated as:
νi ·ni
Yi =
·100%
(4)
6·n0,Phenol
Here ni is the moles of product i after reaction and νi is the number of carbon atoms in compound
i.
The carbon balance was evaluated in all experiments by comparing the carbon initially in the
reactor to the carbon measured in the product and on the catalyst:


νi ·ni −6·n0,Phenol
6·n0,Phenol
∑




i
∆C =
·100%
(5)
Here ∆C is the carbon deviation in % and ni is the moles of compound i. All compounds identified
in the GC analysis were included in the carbon balance. Generally the carbon balance was closed
within 10%, but in many experiments almost complete closure was achieved; this will be discussed
later.
All calculations of X, Yi, and ∆C were corrected for loss of mass from transferring and filtration
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 36
1
2
3
4
processes, which was quantified as the average of a number of blank tests.
5
6
7
8
9
2.4 Catalyst Characterization
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A Quantachrome iQ2 was used for measurement of the specific surface area on the basis of BET
theory.22 Nitrogen at its boiling point was used in the p/p0 range from 0.05-0.3 to construct a seven
point BET plot. Generally all the tested supports had surface areas of the same order of magnitude,
only the carbon support had a significantly larger surface area, see Table 1.
Temperature programmed reduction by hydrogen was performed for selected catalysts on a
Netzsch STA 449 F1 Jupiter ASC thermogravimetric analyzer to identify the required temperature
for complete reduction of the supported metal. The analysis was performed on the basis of changes
in mass and the total mass applied was in the order of 20-40 mg. 5 vol% hydrogen in nitrogen was
added at a flow of 100 Nml/min and the samples were heated at: 5 ◦C/min up to 300 ◦C, 1 ◦C/min
from 300 ◦C to 500 ◦C, and 5 ◦C/min from 500 ◦C to 700 ◦C.
NH3-TPD was performed with an Autochem II 2920 apparatus. Here ca. 0.1 g of the samples
were initially heated to 500 ◦C in 50 ml/min He and then cooled to 100 ◦C. At this point the samples
were saturated with 50 ml/min NH3 for 120 min. Desorption of NH3 was hereafter measured by
flushing with 50 ml/min He while heating at a ramp of 5 ◦C/min up to 500 ◦C.
X-ray diffraction (XRD) was measured with a PANalytical X’Pert PRO using a rotating copper
anode X-ray source at 40 kV and 30 mA, nickel filter, and automatic anti-scatter and divergence
slits. The 2Θ angle was scanned from 12◦ to 120◦ by increments of 0.00656◦ with 45.9 seconds
per step. The mean coherent-scattering domain (CSD) size was derived from reflection half-widths
using the Selyakov-Scherrer formula.23 All samples were analyzed as-prepared powders.
10
ACS Paragon Plus Environment

Page 11 of 36
ACS Catalysis
1
2
3
4
5
3 Kinetics of Phenol HDO
6
7
8
9
To determine the activity of the catalysts on a quantitative basis a kinetic model was developed. In
the following this model is presented as a basis for the later discussion.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The time dependent development of the conversion of phenol and the yields of cyclohexanone,
cyclohexanol, and cyclohexane in an experiment with a Ni-V2O5/SiO2 catalyst at 250 ◦C and 100
bar is shown in Figure 2. At first, phenol was converted to primarily cyclohexanol. This yield was
maximal at approximately 3.5 h of reaction. The cyclohexane yield, on the other hand, increased
continuously throughout the experiment, and with an increasing rate toward the end.
100
X
90
80
70
YC-Hexanol
60
50
40
30
YC-Hexane
20
10
YC-Hexanone
0
0
1
2
3
4
5
6
Time [h]
Figure 2: Conversion of phenol (X) and yields of cyclohexanone, cyclohexanol and cyclohexane as
a function of time over a Ni-V2O5/SiO2 catalyst. The experiments were made with 1 g of catalyst
in 50 g phenol. T=250 ◦C, P=100 bar.
The data in Figure 2 indicate a reaction mechanism at 250 ◦C where hydrogenation of the aro-
matic ring occurs as a first step followed by deoxygenation, with cyclohexanol as the intermediate
product and cyclohexane as the final product. Also cyclohexanone and cyclohexene were observed
in small amounts. The cyclohexanone yield had a maximum of 1.7% after 1 h and then steadily
decreased (cf. Figure 2). The cyclohexene yield was even less and in the order of 0.01%.
Shuikin and Erivanskaya24 suggested that hydrogenation of phenol produces cyclohexanone
as the primary product, but the subsequent hydrogenation of cyclohexanone to cyclohexanol pro-
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 36
1
2
3
4
5
6
7
8
ceeds at a much higher rate and therefore cyclohexanone is only seen in low yields. Similarly,
cyclohexene is formed through dehydration of cyclohexanol followed by quick hydrogenation to
cyclohexane. Overall the observations from Figure 2 and previous literature for nickel catalysts,24
9
29
Pd/C,25,26 Pt/C,27 Ni/HZSM-5,28 and Ni-MoS2/Al2O3 suggest the reaction scheme depicted in
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3 with solid arrows. To verify this, measurements were performed over a Ni/ZrO2 catalyst
to quantify the rate of the four steps (see electronic supporting information (ESI)). This is shown
in Figure 3 as the turn over frequency (TOF) in each step. Hydrogenation of cyclohexanone and
cyclohexene were significantly faster reactions compared to hydrogenation of phenol and dehy-
dration of cyclohexanol. Hence, only low yields of cyclohexanone and cyclohexene should be
expected.
Figure 3: Observed reaction path for HDO of phenol. Solid arrows indicate main pathways while
the dashed arrows show the effective conversion steps in th◦e kinetic model. Indicated rates are
from measurements with a 5 wt% Ni/ZrO2 catalyst at 275 C and 100 bar in the batch reactor
setup. The estimation of the individual turn over frequencies (TOF) are supplied in the electronic
supporting information (ESI).
From the data in Figure 2 and 3 deoxygenation appears to take place more readily from a
saturated ring than from an unsaturated ring, which has also been shown elsewhere.1,8,30 This is
related to the dissociation energy of the C-O bond in alcohols,31 which decreases in the following
order:
Aromatic alcohol (469kJ/mol) >> Secondary alcohol (385kJ/mol)
≈ Primary alcohol (383kJ/mol) ≈ Tertiary alcohol (379kJ/mol)
(6)
12
ACS Paragon Plus Environment

Page 13 of 36
ACS Catalysis
1
2
3
4
5
6
7
8
9
The data clearly show that the aromatic alcohol (i.e. phenol) has ≈ 100 kJ/mol higher bond dis-
sociation energy compared to a secondary alcohol (as cyclohexane) and the deoxygenation can
therefore more readily take place from the cyclohexanol compared to the phenol as the C-O bond
is markedly weakened.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Because the rate of hydrogenation of both cyclohexanone and cyclohexene are relatively fast
reactions and the yields of these compounds are low, the effective reaction scheme can be described
by the dotted arrows in Figure 3, which results in the following two overall main reactions:
Reaction 1:
Reaction 2:
Phenol+3H2 −−→ Cyclohexanol
Cyclohexanol+H2 −−→ Cyclohexane+H2O
Assuming first order dependency of the hydrocarbon compounds leads to the following rate ex-
pressions:
−r1 = k1 ·CPhenol ·Pn
−r2 = k2 ·CC-hexanol ·Pm
(7)
(8)
H2
H2
Here ri is the rate of reaction i, ki is the rate constant of reaction i, Ci is the concentration of either
phenol or cyclohexanol, and n and m are the reaction order of hydrogen in respectively reaction 1
and 2. The assumption of first order kinetics with respect to the hydrocarbons is made in the lack
of a better estimate. As all the experiments were performed with the same and constant pressure of
hydrogen, the hydrogen pressure can be included in the rate constant and thereby give the following
lumped rate expressions:
−r1 = k10 ·CPhenol
(9)
−r2 = k20 ·CC-hexanol
(10)
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 36
1
2
3
4
5
6
7
8
9
Here k10 is the rate constant for the hydrogenation reaction and k20 is the rate constant for the de-
oxygenation reaction. The rate expressions in Eq. 9 and 10 were combined with the batch reactor
design equation32 to allow determination of k10 and k20 from the experimental data and a quantitative
comparison of the different catalysts.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A thorough description of the derivation of the model, the assumptions made, and the validation
of the model can be found in the electronic supporting information (ESI).
4 Results and Discussion
4.1 Oxide and Methanol Synthesis Catalysts
The results from testing the oxide catalysts at 275 ◦C and 100 bar are shown in Figure 4 in terms of
the yields of cyclohexanone, cyclohexanol, and cyclohexane, and the conversion of phenol. None
of the catalysts achieved a conversion above 10%. The NiO-MoO3/Al2O3 catalyst showed some
hydrogenation activity with a yield of 4% oxygenated cyclohexanes. However, the activity for
HDO over this catalyst was low, with a yield of cyclohexane of only 0.9%.
Table 2 gives a summary of the performance of the catalysts. The carbon balances of the
experiments were reasonably well closed, and in most cases were in the order of -2%. Only the
CoO-MoO3/Al2O3 catalyst had a larger carbon deviation of -8.5%. As the conversion of phenol
was 9% on this catalyst, it could indicate that this catalyst mainly was active for cracking.
Prasomsri et al.33 did observe HDO activity of bulk V2O5, Fe2O3, CuO, WO3, and MoO3
catalysts in continuous flow for acetone deoxygenation in the gas phase at atmospheric pressure
◦
and 400 C, with MoO3 performing the best. In contrast to their study where temperatures of
400 ◦C were required in order to deoxygenate anisole, much lower temperatures were used in the
current screening. The somewhat higher temperature requirement is probably linked to the catalyst
inability to hydrogenate the anisole and therefore breaking the somewhat stronger C-O bond of the
aromatic alcohol compared to the saturated, as discussed in Section 3 (see Eq. 6).
14
ACS Paragon Plus Environment

Page 15 of 36
ACS Catalysis
1
2
3
4
5
X
CY-CH-Heexxaannoonene
CY-CH-Heexxaannool l
CY-CH-Heexxaannee
6
10
9
8
7
6
5
4
3
2
1
0
7
8
9
Oxide
Methanol Synthesis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4: Conversion of phenol (X) and yields of cyclohexanone, cyclohexanol, and cyclohexane
from experiments with different types of oxide catalysts and methanol synthesis catalysts and for
the blank experiment. The experiments were made with 0.5 g of catalyst in 10 g phenol and 40 ml
water. T=275 ◦C, P=100 bar, reaction time=4 h.
Table 2: Overview on results from different oxide catalysts and methanol synthesis catalysts. k10
is the rate constant for hydrogenation, k20 is the rate constant for deoxygenation, and ∆C is the
deviation in the carbon balance. The experiments were made with 0.5 g of catalyst in 10 g phenol
and 40 ml water. T=275 ◦C, P=100 bar, reaction time=4 h.
Catalyst
Blank
MnO/C
WO3/C
MoO3/C
V2O5/C
NiO-MoO3/Al2O3
CoO-MoO3/Al2O3
Cu/ZnO/Al2O3
NiCu/SiO2
Cu/SiO2
k10 [ml/(kgcat min)] k20 [ml/(kgcat min)] ∆ C [%]
-
-
7
10
4
3
20
0
10
28
50
0.0
1
1
2
1
5
6
8
7
1
-1.1
-1.4
-3.1
-1.4
-2.2
-8.5
-0.2
-3.9
-0.1
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 36
1
2
3
4
5
6
7
8
9
Figure 4 also shows the results from the screening of the methanol synthesis catalysts (Cu/SiO2,
NiCu/SiO2, and Cu/ZnO/Al2O3). This group of catalysts also had a low activity. Cu/ZnO/Al2O3
had a conversion of ca. 9%, but primarily was selective towards hydrogenation, with cyclohexanol
and cyclohexanone constituting 91% of the product. NiCu/SiO2 had a conversion of 7%, but
according to the carbon balance (cf. Table 2) half of the measured conversion was not accounted
for.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Among these catalysts, Cu/ZnO/Al2O3 was the best performing, with the highest conversion
and a good closure of the carbon balance (∆C = -0.2%). However, as the product primarily was
cyclohexanol the potential of this catalyst seems limited in the context of HDO.
Ardiyanti et al.34 previously investigated Cu/Al2O3 and NiCu/Al2O3 catalysts of varying Ni/Cu
ratios for HDO of anisole at 300 ◦C and 10 bar in a continuous flow reactor. In these studies it was
observed that the pure Cu catalyst was unable to perform HDO of the anisole, as the only product
was phenol. For NiCu/Al2O3 the best catalyst was obtained with a 8:1 weight ratio of Ni:Cu. In
agreement with this, we also observed improved activity of the NiCu/SiO2 (with Ni/Cu ratio of 1)
catalyst relative to Cu/SiO2, but the NiCu/SiO2 was not optimized as in the work of Ardiyanti et
al.34 The results indicate that the Ni loading has to be markedly higher than the Cu to achieve good
HDO activity.
Summarizing, the apparent order of activity for the tested oxide and methanol catalysts was:
Cu/ZnO/Al2O3 ≈ NiO-MoO3/Al2O3 ≈ NiCu/SiO2 > Cu/SiO2
> MnO/C,WO3/C,V2O5/C,CoO-MoO3/Al2O3
Commonly, none of the catalysts achieved high conversion at the current conditions due to their
inability to hydrogenate the phenol, cf. Table 2.
16
ACS Paragon Plus Environment

Page 17 of 36
ACS Catalysis
1
2
3
4
5
4.2 Reduced Noble Metal Catalysts
6
7
8
9
In the experiments with reduced noble metal catalysts, higher conversions were observed compared
to the oxide and methanol synthesis catalysts. Thus, pure phenol was used for these experiments,
as this gave consistency. Reaction temperature and pressure were kept the same.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5 summarizes the conversion of phenol and the yields of cyclohexanol, cyclohexane,
and dicyclohexyl ether from the experiments with noble metal catalysts. All the catalysts gave
complete conversion of the phenol through a relatively fast hydrogenation reaction to cyclohexanol.
The subsequent HDO more favorably took place on the ruthenium catalyst than on the platinum
and palladium catalysts. The ruthenium catalyst provided a yield of 52% cyclohexane relative
to 1% and 11% on the platinum and palladium catalysts, respectively. The palladium catalyst
furthermore had a high yield of dicyclohexyl ether of 21%, i.e. higher than the cyclohexane yield.
Y
Y
Y
X
C-CH-Heexxaannool l
C-CH-Heexxaannee
DiDcyicycclloohheexyxlyElthEerther
100
90
80
70
60
50
40
30
20
10
0
Ru/C
Pt/C
Pd/C
Figure 5: Conversion of phenol (X) and yields of cyclohexanol, cyclohexane, and dicyclohexyl
ether from experiments with different types of reduced noble metal catalysts. The experiments
were made with 1 g of catalyst in 50 g phenol. T=275 ◦C, P=100 bar, reaction time=5 h.
Table 3 summarizes the kinetic parameters of the reactions. The hydrogenation rate constants
(k10 ) were of the same order of magnitude for Ru/C and Pd/C, but Pt/C had a rate of hydrogenation
ca. one order of magnitude higher. In general all three catalysts however were good hydrogenation
catalysts, achieving 100% conversion of the phenol in the 5 h experiments.
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 36
1
2
3
4
5
6
7
8
9
Table 3: Overview of results from the reduced noble metal catalysts. dM is the metal crystallite
size measured by XRD, k10 is the rate constant for hydrogenation, k20 is the rate constant for deoxy-
genation, and ∆C is the deviation in the carbon balance. The experiments were made with 1 g of
catalyst in 50 g phenol. T=275 ◦C, P=100 bar, reaction time=5 h.
Catalyst dM [nm] k10 [ml/(kgcat min)] k20 [ml/(kgcat min)] ∆C [%]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ru/C
Pt/C
Pd/C
7
4
6
1950a
31200b
1840c
115
1
19
-10.8
-11.8
-6.1
aDetermined from a 18 min. isothermal experiment at 275 ◦C with 50% phenol conversion.
bDetermined from a 10 min isothermal experiment at 275 ◦C with 99.8% phenol conversion.
cDetermined from a 16 min isothermal experiment at 275 ◦C with 44% phenol conversion.
Regarding the hydrodeoxygenation step (k20 ) the catalytic activity of Ru/C was an order of
magnitude higher than for Pd/C, and two orders of magnitude higher than for Pt/C. This analysis
does not take into account that the primary part of the deoxygenation for the Pd/C catalyst was
taking place through the ether forming reaction, and k20 is therefore not completely descriptive for
the affinity for deoxygenation on Pd/C. Nevertheless, the trend is clear.
The carbon balance in Table 3 shows that all the tested catalysts lacked in the order of 5-
10% carbon. Analysis of the gas phase from the Ru/C experiment by sampling in a gas bag and
then analyzing on a GC-TCD showed the presence of CO2, CH4, and CO. Furthermore, traces of
pentane, hexane, ethanol, methanol, and acetone could be identified in the oil phase. This shows
that cracking reactions take place on the Ru/C catalyst. It has previously been shown that Ru/C can
◦
produce CH4 and CO2 from guaiacol at temperatures above 250 C.35 Elliott and Hart35 linked
the production of these gases to aqueous phase reforming, which for Ru/C has been shown to be
active at similar conditions.36
Overall the apparent order of activity for deoxygenation for the tested noble metal catalysts for
HDO of phenol was found to be:
Ru/C > Pd/C > Pt/C
In agreement with this, Elliott and Hart35 found Ru/C to be a better catalyst than Pd/C for the
conversion of guaiacol in a batch reactor. Wildschut et al.37 identified Ru/C and Pd/C as better
18
ACS Paragon Plus Environment

Page 19 of 36
ACS Catalysis
1
2
3
4
5
6
7
8
9
performing catalysts compared to Pt/C for bio-oil HDO in a batch reactor. Lee et al.38 identified
Ru/SiO2-Al2O3 as the best performing catalyst for guaiacol HDO in comparison to platinum and
palladium. Adriyanti et al.39 found Pd/ZrO2 as a better catalyst for bio-oil HDO in a batch reactor
compared to Pt/ZrO2. Thus, the present results for HDO over noble metal catalysts are in good
agreement with the literature, which all point toward ruthenium as being one of the best performing
noble metals for HDO and with Pd being more efficient than Pt. The kinetic model revealed that the
reason for this is that Pt is a relatively poor deoxygenation catalyst but a very good hydrogenation
catalyst.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The reason for ruthenium being the best performing noble metal could be correlated to the
metals affinity to bind oxygen. Nørskov et al.40 showed through density functional theory (DFT)
calculations that the binding energy of oxygen relative to water (∆EO) on Ru, Pd, and Pt, was
respectively -0.01 eV, 1.53 eV, and 1.57 eV. Thus, ruthenium has the strongest binding energy to
oxygen, and platinum the weakest, which correlates with the affinity to perform deoxygenation.
In the current work the effect of support was not investigated, but previous work has shown that
the type of support can significantly increase the activity of noble metal catalysts. Lee et al.38 tested
noble metal catalysts on C, Al2O3, and SiO2-Al2O3 for guaiacol HDO at 250 ◦C and 40-70 bar in
a batch reactor. Their work showed that the activity of all the supported noble metal catalysts was
improved in the order of increasing support acidity, i.e. an apparent order of activity of: NM/SiO2-
Al2O3 > NM/Al2O3 > NM/C, where NM is a noble metal. Similar, Foster et al.41 also observed
increasing catalytic activity for HDO of m-cresol over platinum catalysts with increasing acidity
of the support at 260 ◦C and 1 atm in a packed bed reactor setup. Both studies point toward that the
use of an acidic support can increase the activity of the noble metal catalysts. The role of support
is discussed further in Section 4.3.3.
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 36
1
2
3
4
5
4.3 Reduced Non-noble Metal Catalysts
6
7
4.3.1 Reduction Temperature
8
9
Prior to the activity tests of the non-noble metal catalysts, the required reduction temperatures
were determined by H2-TPR. Table 4 summarizes the results of these measurements, showing the
temperature intervals where the reduction took place.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4: Results from H2-TPR of the reduced non-noble metal catalysts. Start and end indicate
the temperature interval where reduction was observed. Gas: 5% H2 in N2, Flow: 100 Nml/min,
Heating ramp: 5 ◦C/min up to 300 ◦C, 1 ◦C/min from 300 ◦C to 500 ◦C, and 5 ◦C/min from 500
◦C to 700 ◦C, Sample amount: 20-40 mg.
Catalyst
Start [◦C] End [◦C] Mass loss [wt%] Theoretical mass loss
Co/SiO2
Ni/SiO2
Ni/Al2O3
Ni/CeO2
Ni/CeO2-ZrO2
Ni/ZrO2
Ni/MgAl2O4
Ni/C
Ni-V2O5/SiO2
260
305
310
280
295
310
305
360
300
520
365
505
335
500
485
500
505
390
1.0
1.4
0.9
1.3
1.2
1.4
0.6
1.1
3.1
1.4
1.4
1.4
1.4
1.4
1.4
1.4
1.4
5
For the nickel based catalysts, the theoretical mass loss is 1.4 wt% for a nickel loading of 5
wt%, assuming that all nickel is reduced from NiO to Ni. Generally the measured mass loss was
close to the theoretical mass loss. The largest deviation was seen in the case of Ni/Al2O3 and
Ni/MgAl2O4 with measured mass losses of 0.9 wt% and 0.6 wt% respectively. For these catalysts,
the XRD analyzes on the used catalysts revealed that the nickel to some extent was interacting with
the supports, forming a mixed oxide with the alumina or spinel. These oxides have been shown to
retard reduction, resulting in the observed lower degree of reduction.42
For the Ni-V2O5/SiO2 complete reduction of the nickel and vanadium into NiV/SiO2 would
result in a theoretical mass loss of 5 wt%. However only 3.1 wt% weight loss was observed,
showing that the catalyst was not completely reduced and probably still contained vanadium in the
form of oxides even at 700 ◦C.
Overall, the identified end temperatures from the TPR analyzes are judged sufficient for com-
20
ACS Paragon Plus Environment

Page 21 of 36
ACS Catalysis
1
2
3
4
5
6
plete reduction. The actual reduction temperature, which is given in Table 1, was chosen higher
for the individual samples to ensure complete reduction.
7
8
9
4.3.2 Reduced Non-noble Metals
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Three reduced non-noble metals were tested: Ni/SiO2, Co/SiO2, and Fe/SiO2. These were chosen
based on their known hydrogenation activity in other processes.43–46 Figure 6 shows the compar-
ison between the activities of these three catalysts. Ni/SiO2 was the most active catalyst with a
conversion of 80%, compared to 3.2% and 0.8% for respectively Co/SiO2 and Fe/SiO2.
C-Hexanol
YC-Hexanol
C-Hexane
YC-Hexane
X
100
90
80
70
60
50
40
30
20
10
0
Figure 6: Conversion of phenol (X) and yields of cyclohexanol and cyclohexane from experiments
with different types of reduced non-noble metal catalysts. The experiments were made with 1 g of
catalyst in 50 g phenol. T=275 ◦C, P=100 bar, reaction time=5 h.
The kinetic data for the three catalysts are shown in Table 5. The rate constants k10 of Co/SiO2
and Fe/SiO2 were respectively 5 ml/(kgcat min) and 1 ml/(kgcat min), showing that Co/SiO2 was
the better hydrogenation catalyst of the two. However, k10 was two orders of magnitude larger for
Ni/SiO2 in comparison. k20 was of the same order of magnitude for Co/SiO2 and Fe/SiO2, but
significantly larger for Ni/SiO2, as seen from Table 5.
XRD of the spent Ni/SiO2, Co/SiO2, and Fe/SiO2 showed the active metals to be reduced and
with similar metal crystallite sizes.
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 5: Overview of results from reduced non-noble metal catalysts. dM is the metal crystallite
size measured by XRD, k10 is the rate constant for hydrogenation, k20 is the rate constant for deoxy-
genation, and ∆C is the deviation in the carbon balance. NH3 ads. (adsorption) is the ammonia
saturation measured by NH3-TPD, the data is distinguished in w◦eak and strong adsorption with
weak being the signal recorded for desorption peaks below 200 and strong being above. The
experiments were made with 1 g of catalyst in 50 g phenol. T=275 ◦C, P=100 bar, reaction time=5
h.
Catalyst
dM
k10
k20
∆C NH3 ads. [µmol/gcat]
[nm] [ml/(kgcat min)] [ml/(kgcat min)] [%] Weak
Strong
Co/SiO2
Fe/SiO2
Ni/SiO2
Ni/Al2O3
Ni/CeO2
Ni/CeO2-ZrO2
Ni/ZrO2
Ni/MgAl2O4
Ni/C
12
12
9
7
8
6
7
10
14
-
5
1
230
52
64
159
100
6
-2.6
-0.3
-4.2
-0.5 210
-8.0
-6.3
-1.1 210
-4.4 140
-2.0
-0.8
-
-
40
-
-
-
3200a
21800b
10500c
1010d
6580e
3
220
400
250
210
150
-
-
-
13
361
25
121f
136
256
30
60
Ni-V O5/SiO2
1160
701
-
Ni-V2O5/ZrO2
-
-1.5 130
230
2
aDetermined from a 13 min. isothermal experiment at 275 ◦C with 54% phenol conversion.
bDetermined from a 10 min. isothermal experiment at 275 ◦C with 98.7% phenol conversion.
cDetermined from a 15 min. isothermal experiment at 275 ◦C with 95% phenol conversion.
dDetermined from a 15 min. isothermal experiment at 275 ◦C with 26% phenol conversion.
eDetermined from a 10 min. isothermal experiment at 275 ◦C with 74% phenol conversion.
fDetermined from a 4 h experiment at 275 ◦C with cyclohexanol as feed instead of phenol.
22
ACS Paragon Plus Environment

Page 23 of 36
ACS Catalysis
1
2
3
4
Summarizing, the apparent order of activity for the three non-noble metal catalysts is:
5
6
7
8
9
Ni/SiO2 ꢀ Co/SiO2 > Fe/SiO2
(11)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The better performance of Co/SiO2 compared to Fe/SiO2 is due to a better activity for hydro-
genation on this catalyst. Similar to these results, Filley and Roth47 investigated cobalt and iron
supported on alumina for HDO of guaiacol at atmospheric pressure and 350 ◦C and did not find any
notable activity. Yakovlev et al.11 investigated both cobalt and nickel catalysts for HDO of anisole
◦
at 300 C and 10 bar of hydrogen and found that Co/SiO2 was practically inactive compared to
Ni/SiO2; which is similar to the results of this work.
4.3.3 Support Effect on Nickel Catalysts
With nickel identified as the best performing of the tested non-noble metal catalysts, this was
chosen for further investigations with respect to the influence of support.
Seven different types of supports of comparable specific surface areas (cf. Table 1) were tested
and very different activities were observed, as summarized in Figure 6. Ni/C was the only catalyst
with practically no activity and a conversion of only 2%. On the other hand, Ni/CeO2, Ni/CeO2-
ZrO2, and Ni/MgAl2O4 were very active for hydrogenation with conversions of 100%, but less
active for deoxygenation with yields of cyclohexane of 4%, 8%, and 12%, respectively. Ni/Al2O3
had a high activity for both hydrogenation and deoxygenation with a conversion of 100% and a
yield of 46% cyclohexane. Ni/SiO2 gave a yield of cyclohexane in the same order of magnitude
(38%), but was significantly less active for hydrogenation with a conversion of 80%. Ni/ZrO2 was
the overall best performing catalyst under these conditions with almost complete conversion of
phenol (X = 99.8%) and a high selectivity toward cyclohexane (YC-hexane = 83%).
The kinetic data in Table 5 shows that Ni/CeO2 had the highest hydrogenation rate constant
(k10 ) followed by Ni/CeO2-ZrO2. Both catalysts showed hydrogenation rate constants an order of
magnitude higher than Ni/MgAl2O4, Ni/Al2O3, and Ni/ZrO2, decreasing in that order. Ni/SiO2
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 36
1
2
3
4
had a hydrogenation rate constant two orders of magnitude lower than Ni/CeO2.
5
6
7
8
In contrast, the Ni/CeO2 catalyst had practically no activity for deoxygenation. The deoxy-
genation activity of Ni/CeO2-ZrO2 and Ni/MgAl2O4 was 2-5 times larger than for Ni/CeO2, but
still low yields of cyclohexane were produced. Only Ni/ZrO2, Ni/SiO2, and Ni/Al2O3 had rela-
tively high rates of deoxygenation, decreasing in that order. An interesting aspect of this is that
the Ni/SiO2 had a higher rate constant for deoxygenation compared to Ni/Al2O3, but because
Ni/Al2O3 had a higher rate of hydrogenation the cyclohexane yield with the Ni/Al2O3 catalyst was
higher than with the Ni/SiO2 catalyst.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ni/SiO2 and Ni/ZrO2 were additionally tested as catalysts where the support had initially been
promoted by vanadium before impregnation with nickel. The results of these experiments are
also shown in Figure 6, while kinetic data are shown in Table 5. For Ni-V2O5/SiO2 the rate of
hydrogenation was significantly increased compared to the pure Ni/SiO2 catalyst. The cyclohexane
yield also increased, but was primarily caused by the increased rate of hydrogenation, as the rate of
deoxygenation did not change between the Ni-V2O5/SiO2 and Ni/SiO2 catalysts (cf. Table 5). The
Ni-V2O5/ZrO2 catalyst performed slightly worse compared to the non-vanadium doped catalyst,
as both the yield of cyclohexane and the rate constants were lower. Overall, it was observed that
addition of vanadium primarily influenced the hydrogenation rate of the catalyst.
The carbon balances presented in Table 5 show an adequate closure in most cases. The largest
deviation was found for the Ce containing supports, where 6-8% of the carbon was unrecovered.
This indicates that either cracking or coke deposition could take place on these catalysts. Contrary,
for the two best performing catalysts, Ni/ZrO2 and Ni/Al2O3, the carbon balance deviation was in
the order of 1% indicating that these catalysts were less prone to coke deposition. In general, better
closure of the carbon balance was found for the nickel based catalysts compared to the noble metal
catalyst (cf. Table 3 and 5) showing that these catalysts have a lower tendency toward cracking.
NH3-TPD was performed, giving information of the total amount of available acid sites on the
24
ACS Paragon Plus Environment

Page 25 of 36
ACS Catalysis
1
2
3
4
catalysts, see Table 5. The total acidity of the supports decreases in the following order:
5
6
7
8
9
Ni/Al2O3 ≥ Ni/ZrO2 ≥ Ni/CeO2 > Ni-V2O5/ZrO2 > Ni/MgAl2O4 > Ni/CeO2-ZrO2
ꢀ Ni-V2O5/SiO2 > Ni/SiO2 > Ni/C
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Comparing this to the apparent order of activity discussed above (or see Table 5), it follows that
the amount of available acid sites (as measured by NH3-TPD) and the HDO activity is not directly
correlated. Most notable is the Ni/SiO2, which practically contains no acid sites, but is found to be
one of the three best HDO catalysts (of the investigated nickel catalysts). However, comparing only
the rate constant for hydrogenation to the acidity gives some consistency as it is seen that the poor
hydrogenation catalysts (Ni/SiO2 and Ni/C) have very little acidity, while the catalysts with better
hydrogenation affinity all have higher acidity. However, the acidity measurements do not reveal
the true acidic nature of the catalyst under experimental conditions, as the reducing atmosphere
in the experiments will result in an increased formation of oxygen vacancy sites and thereby acid
sites.48–51 Better understanding of this is gained by plotting the hydrogenation activity relative to
the metal-oxygen bond strength (E(M-O)) of the support, as done in Figure 7. This reveals that
oxide supports with low metal-oxygen bond energy had the highest activity for hydrogenation.
Low M-O bond energies can be linked to the tendency for oxygen vacancy site generation in the
oxide52,53 and these vacancy sites will function as Lewis acid sites.51,53 Thus, the best performing
hydrogenation catalyst is therefore the catalyst with the highest tendency to form Lewis acid sites
in the oxide structure which is most likely on the supports with the lowest metal-oxygen bond
energy. Due to this effect, introducing V2O5 to a SiO2 support causes formation of more vacancy
sites due to the lower metal-oxygen bond energy of V2O5 (12.3 eV) compared to SiO2 (34.2 eV)
and thereby increases the hydrogenation activity (cf. Table 5). In contrast, doping ZrO2 with V2O5
does not induce any large changes as the metal-oxygen bond energies are relatively close, 14.2 eV
compared to 12.3 eV respectively.
A similar correlation to the metal-oxygen bond energy could not be made with respect to the
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 36
1
2
3
4
5
6
7
8
9
100000
10000
1000
CeO2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Al2O3
ZrO2
SiO2
100
5
10
15
20
25
30
35
40
E(M-O) [eV]
Figure 7: Hydrogenation activity as a function of the metal-oxygen bond strength (E(M-O)) of the
pure support. Only hydrogenation rate constants for nickel catalysts supported on pure oxides from
Table 5 are plotted. E(M-O) data from Idriss and Barteau.53
deoxygenation affinity, which indicates that this reaction does not rely on activation on oxygen
vacancies in the support.
Taking the yield of cyclohexane as the primary parameter for the ranking, the apparent order
of activity for HDO of phenol with nickel on different supports can be summarized as:
Ni/ZrO2 > Ni-V2O5/ZrO2 > Ni-V2O5/SiO2 > Ni/Al2O3 > Ni/SiO2
ꢀ Ni/MgAl2O4 > Ni/CeO2-ZrO2 ≈ Ni/CeO2 ꢀ Ni/C
It may be expected that the nickel crystallite particle size will influence the kinetics of the reactions,
and so it is important to note that the present ranking is made for catalysts with roughly the same
nickel crystallite size (cf. Table 5).
Comparing k20 for Ni/ZrO2, Ni-V2O5/ZrO2, Ni-V2O5/SiO2, Ni/Al2O3, and Ni/SiO2 show that
they all are in the same order of magnitude (ranging from 100-361 ml/(kgcat min)). The differ-
ences are probably related to support-metal interaction. However, Ni/MgAl2O4, Ni/CeO2-ZrO2,
and Ni/CeO2 all have lower values of k20 ’s. For the Ni/MgAl2O4 catalyst, TPR showed that it was
26
ACS Paragon Plus Environment

Page 27 of 36
ACS Catalysis
1
2
3
4
5
6
7
8
9
hard to reduce (cf. Table 4), indicating that part of the nickel is interacting with the support, which
could be the reason for the poor performance of this catalyst. For the catalysts with Ce contain-
ing supports the carbon balance indicated tendency to cracking/coke formation and therefore the
poor deoxygenation performance of these catalysts could be related to potential deactivation of the
deoxygenating sites. No further investigation into these issues have been performed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Obviously, Ni/C behaved markedly different than the other supports. Hence, this catalyst was
tested in an experiment where 50 g of cyclohexanol was used as feed with 1 g of catalyst at 100
◦
bar and 275 C for 4 h. In this experiment a conversion of 47% was observed with a yield of
42% cyclohexane, the remaining product being primarily dicyclohexyl ether (4% yield). Thus,
despite of not being active for hydrogenation of the phenol, this catalyst was found to be active
for deoxygenation. Calculating the kinetic constant for this reaction (corresponding to k20 ) further
showed that the deoxygenation capabilities of this catalyst was in the same order of magnitude as
the other nickel catalysts, as shown in Table 5.
Hence, all catalysts supported on oxides are active for both hydrogenation and deoxygenation,
but with different rates. In contrast, the carbon supported catalyst is only active for the deoxygena-
tion reaction. In continuation of the discussion in the introduction it follows that the dual action
of the catalyst primarily is required in the hydrogenation of complex molecules as phenol. Con-
trary, the deoxygenation can take place on the nickel surfaces. Thus, we suggest a mechanism as
sketched in Figure 8 for the HDO of phenol over nickel catalysts. The coordinatively unsaturated
metal sites in the oxide surface often behave like Lewis acids and the surface oxygen behaves like
Lewis bases.51 Thus, the activation of phenol on the oxide can take place through heterolytic dis-
sociation of the O-H bond on the oxide where the hydrogen from the phenol is adsorbed on an
oxygen site in the oxide surface layer and a metal vacancy site stabilizes the phenoxide ion. Gen-
erally, alcohols have been found to almost always adsorb through heterolytic dissociation on metal
oxides.51 Supporting this, Popov et al.54 observed phenoxide formation on Al2O3 when studying
adsorption of phenol with IR spectroscopy. Liu et al.55 concluded that when phenol is adsorbed at
a Lewis acid site, one of the carbons in the aromatic ring becomes highly nucleophile and thereby
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 36
1
2
3
4
5
6
7
8
9
receptive for electrophiles, as H+. Thus, the phenoxide will interact with a nickel crystallite where
hydrogen has been adsorbed, thereby facilitating the saturation of the double bounds and produc-
ing cyclohexanone. These steps are all illustrated in Figure 8. The cooperative effect of support
and active metal has also been found for the phenol hydrogenation on palladium catalysts where
the activity for hydrogenation can be significantly increased by the introduction of a Lewis acid
due to the activation of the phenol on the Lewis sites.55–57
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8: Proposed reaction mechanism for HDO of phenol over an oxide supported nickel catalyst
(here Ni/ZrO2). Grey spheres represent nickel atoms.
The formed cyclohexanol can directly react with the exposed metallic nickel surfaces as shown
in Figure 8 through adsorption at the OH-group. The actual deoxygenation step probably takes
place through a dehydration reaction, forming cyclohexene, as proposed in the reaction scheme of
Figure 3. The cyclohexene can subsequently be hydrogenated into cyclohexane as the final product.
This reaction mechanism is more varied than that discussed in the introduction and Figure 1(b).
In summary, among all tested catalysts, only those catalysts that showed both good hydrogena-
tion and deoxygenation capabilities performed well, as a key step in the HDO of phenol at the
28
ACS Paragon Plus Environment

Page 29 of 36
ACS Catalysis
1
2
3
4
5
6
7
8
9
tested conditions is the weakening of the C-O bond obtained by hydrogenation of the aromatic
ring, as described in Section 3. Hydrogenation of an aromatic hydrocarbon is however not trivial
as these are considered to bind weakly to metal catalysts compared to conventional alkenes,58,59
but adsorption on Lewis acid sites on the support appears to aid in this task.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5 Conclusions
Screening experiments have been performed to identify efficient catalysts for HDO of phenol at
◦
intermediate temperature and pressure (275 C and 100 bar) in a batch reactor. Four groups of
catalysts were investigated: oxide catalysts, methanol synthesis catalysts, reduced noble metal
catalysts, and reduced non-noble metal catalysts, totaling 23 catalytic systems.
Across the series of investigated catalysts distinct differences were observed in the catalytic
activity. Listing the apparent order of activity for all the tested catalysts, the following is found:
Ni/ZrO2 > Ni-V2O5/ZrO2 > Ni-V2O5/SiO2 > Ru/C > Ni/Al2O3 > Ni/SiO2
ꢀ Pd/C > Ni/MgAl2O4 > Ni/CeO2-ZrO2 ≈ Ni/CeO2 > Pt/C
ꢀ Cu/ZnO/Al2O3 ≈ NiO-MoO3/Al2O3 ≈ NiCu/SiO2 > Cu/SiO2
> Co/SiO2 > Ni/C > Fe/SiO2,MnO/C,WO3/C,V2O5/C,CoO-MoO3/Al2O3
The tested oxide and methanol synthesis catalysts had a low activity for HDO of phenol at the
given conditions, which was linked to their inability to hydrogenate the phenol ring. Reduced metal
catalysts of both noble and non-noble metals were significantly more active. On these catalysts,
HDO of phenol proceeds through an initial hydrogenation to cyclohexanone, which is rapidly
hydrogenated to cyclohexanol. This can then in a second step be dehydrated to cyclohexene,
which is rapidly hydrogenated to cyclohexane. The hydrogenation of the aromatic ring causes a
weakening of the C-O bond strength, making the deoxygenation more favorable from cyclohexanol
compared to phenol.
29
ACS Paragon Plus Environment