Selective hydrodeoxygenation of bio-oil derived products: Ketones to olefins

Article abstract of DOI:10.1039/c5cy00367a

The hydrodeoxygenation (HDO) of various ketones (acetone, methyl ethyl ketone and cyclohexanone) to olefins via hydrogenation-dehydration was conducted in a fixed bed reactor at 373-573 K under H₂. A ketone can be hydrogenated over the metal function to an alcohol intermediate that is subsequently dehydrated to an olefin over the acidic function. A preliminary study on hydrogenation of acetone to 2-propanol over metal/SiO₂ catalysts (Cr, Fe, Co, Ni, Cu and Pd) shows that Ni and Cu are active at >373 K. Although Ni possesses an activity higher than that of Cu, it promotes olefin hydrogenation and alcohol hydrogenolysis at >473 K. Hydrogenolysis of alcohol intermediate is suppressed over the Ni-Cu alloy catalyst. An optimum conversion with 100% selectivity to alcohol, can be obtained at 448 and 473 K for Ni and Cu, respectively. The dehydration of 2-propanol to propylene over proton zeolites (ZSM-5, Y, Mordenite and β) can be achieved at >398 K. The zeolites with three-dimensional pore structure (β and Y) provide relatively higher activity (>90% conversion). However, a bimolecular dehydration to ether is also promoted. Only HZSM-5 shows excellent selectivity to propylene (98%). Hydrodeoxygenation of ketones was tested with (i) a double bed of 5%Ni/SiO₂ and HZSM-5 (Si/Al ～ 13), (ii) a physical mixed bed of 5%Cu/SiO₂ and HZSM-5 (Si/Al ～ 13) and (iii) a bi-functional catalyst of 5%Cu/HZSM-5 (Si/Al ～ 250). It was found that high alkene selectivity was readily obtained at 448 K. While, over the physical mixed bed and bi-functional catalyst, the hydrogenation activity was enhanced as the alcohol intermediate was removed from the system. The reactivity of the ketone depends on its adsorption on the metal surface and steric hindrance, i.e. acetone > cyclohexanone > methyl ethyl ketone.

Full text of DOI:10.1039/c5cy00367a

View Article Online
View Journal
Catalysis
Science &
Technology
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: A.
Witsuthammakul and T. Sooknoi, Catal. Sci. Technol., 2015, DOI: 10.1039/C5CY00367A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/catalysis

Page 1 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
1
1
2
Selective Hydrodeoxygenation of Bio-Oil Derived Products:
Ketones to Olefins
3
4
Ayut Witsuthammakula and Tawan Sooknoia,b*
5
Abstract
6
7
The hydrodeoxygenation (HDO) of various ketones (acetone, methyl ethyl ketone and
cyclohexanone) to olefins via hydrogenation ꢀ dehydration was conducted in fixed bed
reactor at 373 – 573 K under H2. Ketone can be hydrogenated over the metal function to
alcohol intermediate that is subsequently dehydrated to olefin over the acidic function. A
preliminary study on hydrogenation of acetone to 2ꢀpropanol over metal/SiO2 catalysts (Cr,
Fe, Co, Ni, Cu and Pd) shows that Ni and Cu are active at > 373 K. Although Ni possesses an
activity higher than that of Cu, it promotes olefin hydrogenation and alcohol hydrogenolysis
at > 473 K. Hydrogenolysis of the alcohol intermediates can be reduced over NiꢀCu alloy. An
optimum conversion with 100% selectivity to alcohol, can be obtained at 448 and 473 K for
Ni and Cu, respectively. The dehydration of 2ꢀpropanol to propylene over proton zeolites
(ZSMꢀ5, Y, Mordenite and β) can be achieved at > 398 K. The zeolites with threeꢀ
dimensional pore structure (β and Y) provide relative higher activity (> 90% conversion).
However, a bimolecular dehydration to ether is also promoted. Only HZSMꢀ5 shows
excellence selectivity to propylene (98 %). Hydrodeoxygenation of ketones was tested with
8
9
10
11
12
13
14
15
16
17
18
19

Catalysis Science & Technology
Page 2 of 48
View Article Online
DOI: 10.1039/C5CY00367A
2
20
21
22
23
24
25
26
(i) Double bed of 5%Ni/SiO2 and HZSMꢀ5 (Si/Al~13), (ii) physical mixed bed of 5%Cu/SiO2
and HZSMꢀ5 (Si/Al ~13) and (iii) biꢀfunctional catalyst of 5% Cu/HZSMꢀ5 (Si/Al ~250). It
was found that high alkene selectivity was readily obtained at 448 K. While, over the
physical mixed bed and biꢀfunctional catalyst, the hydrogenation activity was enhanced as the
alcohol intermediate was removed from the system. The reactivity of ketone depends on their
adsorption on the metal surface and steric hindrance, i.e. acetone > cyclohexanone > methyl
ethyl ketone.
27

Page 3 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
3
28
1. Introduction
29
30
31
32
33
34
35
36
37
38
Nowadays, the application of pyrolysis bioꢀoil is limited because approximately 20ꢀ50
% oxygen is retained in the products.1ꢀ3 Therefore, the upgrading of pyrolysis bioꢀoil is
necessary. Many researchers suggest that almost all oxygen can be removed by
hydrotreatment process. However, a large amount of hydrogen is consumed and the saturated
oxygenꢀfree compounds4ꢀ7 cannot be used as chemical feedstock. Accordingly, mild
hydrodeoxygenation is of interest to remove oxygen functional groups and produce olefins as
reactive hydrocarbon feedstock. For example, mild hydrodeoxygenation of acetone can
produce propylene that can be used to produce valuable C3 chemicals e.g. acrylic acid and
acrylates,8ꢀ10 acrylonitrile,11ꢀ12 pyridine13 propylene oxide and 1,2ꢀpropaneꢀdiol,14ꢀ16 and the
most consumed polymer, polypropylene.17ꢀ18
39
40
41
42
43
44
45
46
47
Aliphatic and cyclic ketones are common oxygenate compounds found in bioꢀoil19ꢀ23
and a lot more in the upgrading of bioꢀoil via ketonization process.24ꢀ26 The decarbonylation
as done for aldehydes is impossible for the ketones. The direct reduction usually leads to
formation of paraffin or alcohol when drastic or mild condition is used, respectively.27ꢀ30 It
seems that multiꢀstep reaction on one or more catalyst system is required. A similar idea to
produce acrylic acid from glycerol via an integration of dehydrationꢀoxidation31 can be
applied for mild hydrodeoxygenation of ketones to olefins. In this case, the ketone can be
primarily hydrogenated to alcohol that is subsequently dehydrated to olefin. However, the
nature of subꢀreactions is different. For example, hydrogenation is preferred at relatively low

Catalysis Science & Technology
Page 4 of 48
View Article Online
DOI: 10.1039/C5CY00367A
4
48
49
50
51
52
53
temperature and high pressure while dehydration is promoted at higher temperature and lower
pressure. In the catalytic point of view, the acid function should promote only dehydration of
the alcohol product, but not aldol condensation of the ketone fed. In the same view, the metal
function should be selective only for hydrogenation of the ketone to alcohol, not the olefin to
paraffin. Accordingly, several parameters have to be adjusted and controlled including
competitive adsorption of feed, intermediates and products on each individual active site.
54
55
56
57
58
59
60
61
62
In this work, a catalytic system designed for mild hydrodeoxygenation of ketone to
olefin was investigated. Hydrogenation of ketone to alcohol was accomplished over metal
catalysts (Ni, Cu, Fe, Co, Pt and their alloys) at low temperature.32ꢀ35 The alcohol produced
was then dehydrated over acidic catalysts (γꢀAl2O3, HZSMꢀ5, HY and HꢀBeta).36ꢀ39
Hydrogenation and dehydration were separately studied in order to understand the role of
each catalytic function. The integrated hydrogenationꢀdehydration over double bed, physical
mixed bed and biꢀfunctional catalyst bed were then optimized to allow only essential amount
of hydrogen consumption in the first stage. The olefins were selectively obtained and the
reactivity of different ketones were compared and discussed.
63

Page 5 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
5
64
2. Experimental procedure
65
66
67
68
69
70
71
72
73
74
75
76
77
78
As zeolites with various Si/Al are used in this study, the Si/Al of a particular sample
is designated as a number in the parenthesis; i.e. a HZSMꢀ5 sample with Si/Al~13 is referred
to as HZSMꢀ5 (13). HZSMꢀ5 (13) and HZSMꢀ5 (250) were obtained from Zeochem® and
HZSMꢀ5 (180) were obtained from Zeolyst®. NH4 ꢀBeta (14), HꢀMordenite (15), NH4 Y
+
+
+
(3.5) and HY (100) were supplied by Tohso. The NH4 ꢀzeolites were converted to H+ꢀzeolites
by calcination at 773 K under flow of air zero (60 mL/min) for 5 h with a heating rate of 2
K/min. Metal supported catalysts were simply prepared by incipient wetness impregnation.
Metal nitrate precursors (Cu(NO3)2.3H2O by Ajax Fine Chem, Ni(NO3)2.6H2O,
Cr(NO3)3.9H2O Fe(NO3)3.9H2O, and Co(NO3)2.6H2O by Carlo Erba®) were dissolved in
deionized water (~0.01 M) and slowly drop onto support (SiO2; Carlo Erba and Zeolites)
until wet. The sample was dried at 333 K in an oven for 15 min, and then the loading was
repeated until desired metal content was reached. The samples were kept to dry at 333 K
overnight and calcined in air zero (60 ml.minꢀ1) at 723 K for 5 h, then pelletized to the size of
600–850 µm.
79
80
81
82
83
Elemental composition was determined by Xꢀray fluorescence spectrometer (XRF;
Siemens). Specific surface area (BET) of catalysts was measured using nitrogen adsorption
analyzer (Quantachrome) at 77 K and 0.05–0.30 P/P0. Residue retained in the used catalysts
was analysed by thermoꢀgravimetric analysis (Perkin Elmer) under airꢀzero or N2 stream
from 323 ꢀ 1173 K at rate of 10 K/min. Reducible metal oxide species in the catalysts was

Catalysis Science & Technology
Page 6 of 48
View Article Online
DOI: 10.1039/C5CY00367A
6
84
85
86
87
88
89
90
91
92
analyzed by temperature programmed reduction (TPR). The catalysts were treated in airꢀzero
at 723 K for 5 hours prior to heating from 323 ꢀ 1173 K in 10% H2/Ar. The hydrogen
consumption was recorded with an onꢀline thermal conductivity detector (VICI).40ꢀ41 Copper
dispersion on support was also analyzed by selective surface TPR technique. Briefly after
typical TPR, the sample was in situ treated with N2O at 333 K for 2 hours. Then, the surfaceꢀ
oxidized sample was subjected to a secondary TPR.42 Moles of surface copper can be
calculated as described by Sagar et al.43 Acidity of HY and Cu/HY was quantified by NH3ꢀ
TPD. 1 % NH3/He was preꢀadsorbed at 323 K. TPD was carried out in He at 10 K.minꢀ1 from
323 – 973 K.44
93
94
The catalytic testing was conducted in a fixed bed flow reactor (6 mm i.d. Pyrex®).
The catalysts were primarily activated at 723 K (2 K.minꢀ1) under stream of air zero (30
ml.minꢀ1) for 5 h. Subsequently the metal supported catalysts were treated in H2 at 723 K for
2 h. The system was cooled down to the reaction temperature (373 – 573 K) and the ketone
feed was introduced by a syringe pump at a rate of 0.07ꢀ0.7 g.hꢀ1. The reaction was carried
out for 9 h on stream and the products were analyzed by an onꢀline GCꢀFID every 70
minutes. A Hayesep® P (1/8” X 8’) was used as separating column.
95
96
97
98
99
100

Page 7 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
7
101
3. Result and discussion
102
3.1 Catalyst characterization
103
104
105
106
107
108
109
110
111
112
113
Metal content and surface area of all catalyst samples are tabulated in Table S1
(supplementary information). All catalysts possess relatively high surface area. The copper
dispersion on SiO2 decrease from 95 to 24 % when loading was increased from 5 ꢀ 15%
(Table 1). However, the dispersion for 2% Cu/SiO2 cannot be accurately calculated,
presumably due to an incomplete surface oxidation by N2O and a deviated hydrogen
consumption measurement when the Cu loading is relatively small. The relatively lower Cu
dispersion is observed when zeolites are used as support (54 – 65 %). This is possibly
because, with this technique, part of the surface Cu oxide may well become exchangeable Cu
cation that cannot be easily reduced. The presence of exchangeable copper cation in HY can
be evidenced from an obvious increase in weak acidity, as compared to that in the parent HY
(Table 1).
114
115
116
117
118
119
Figure 1 shows that at 2 wt% loading on silica, high temperature is required for
reduction of Cr >Co~Fe >Ni >Cu oxides, respectively.45 Broad reduction peaks at the
shoulder of CuO and NiO at 650ꢀ800 K are attributed to various metal support interactions
including copper or nickel silicate.46ꢀ47 While, cobalt oxide reduction appears as two peaks at
548 and 643 K, corresponded to Co3O4 → CoO and CoO → Co, respectively.48ꢀ49 For iron
oxide, the first reduction peak of iron corresponds to two overlap stage Fe2O3 → Fe3O4 and

Catalysis Science & Technology
Page 8 of 48
View Article Online
DOI: 10.1039/C5CY00367A
8
120
121
122
123
124
125
126
Fe3O4 → FeO (603 K).50 While, fully reduction of FeO to Fe cannot be observed up to 1173
K. The formation of Fe2SiO4 is expected as a cause of incomplete reduction.51 The TPR
profile of chromium oxide corresponds to the reduction of dichromate and poly chromate to
Cr2O3 (700 K).52 However, chromium (III) is very stable and cannot be reduced in this
temperature range.53 When nickel and copper are mixed on silica, the complete alloys are
obtained at every composition as seen by single hydrogen consumption peak in Figure S1
(supplementary information).
127
128
129
130
131
The copper loaded HY (100) and HZSMꢀ5 (250) show two reduction peaks at 470 K,
corresponding to CuO aggregates on the surface (Figure 2). The higher temperature peak
(520 K) is defined as highly dispersed copper oxide in the pore of zeolites.54ꢀ55 It is noted that
higher Cu dispersion is observed on HY, as compared to HZSMꢀ5, presumably due to a better
diffusion of Cu precursor in the larger pore (peak at ~520 K).
132

Page 9 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
9
133
3.2 Hydrogenation
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
Since the successive hydrogenation of olefin product is not preferred, different metal
catalysts are tested for selective hydrogenation of ketone. As seen in Table 2, the initial
activated temperature for hydrogenation of acetone and propylene for various metal catalysts
are compared. The initial activated temperature reported in this study refers to the
temperature at which >3% conversion is obtained over the catalyst at the same contact time
(30 g.h.molꢀ1). It is clear that chromium and iron are inactive for hydrogenations at low
temperature. This is presumably due to an incomplete reduction of the metal phase under the
condition investigated (723 K). However, 10%Fe/SiO2 shows some activity for
hydrogenation of acetone and propylene at temperature higher than 473 and 573,
respectively. Co, Ni, Pd56ꢀ57 catalysts are active for both acetone and propylene
hydrogenation. Hydrogenolysis can also be promoted over these metals at relatively higher
temperature. For acetone hydrogenation, the activity appears to be in the order of Ni > Pd
>Co. In the opposite view, Pd is more active than Co and Ni for propylene hydrogenation.
Copper seems to be the best choice for mild hydrogenation of acetone. Ketone can be
hydrogenated (< 373 K) to alcohol without paraffin productions. The alloy with Ni (NiCu)
significantly reduces both hydrogenation of propylene and hydrogenolysis acivity, as
compared to the Ni alone. This is attributed an increase in oxophillicity of the surface when
Cu is alloyed.58 Hence, the interaction of C=C would be readily modified. In addition, the Cu
alloying would strongly affect a structure sensitive reaction, such as hydrogenolysis. Both Ni

Catalysis Science & Technology
Page 10 of 48
View Article Online
DOI: 10.1039/C5CY00367A
10
153
154
and Cu catalyst shows excellence catalyst stability and no carbon deposit can be observed by
TGA of the spent catalysts (Table 3).
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
As Ni and Cu selectively promote hydrogenation of ketone with minimal activity for
olefin hydrogenation, these catalysts are particularly studied for acetone hydrogenation at 373
ꢀ 573 K as shown in Figure 3. It can be seen that for both metals, acetone conversion
increased with temperature until the thermodynamic limitation59ꢀ61 is reached for the reaction
conditions used in this study. The maximum conversion was obtained at 448 K for nickel and
473 K for copper. It is clear that the nickel is more active than the copper. 100% Selectivity
to 2ꢀpropanol is obtained over the nickel catalyst at 373 ꢀ 448 K. However, hydrogenolysis to
methane can be observed at > 473 K. Accordingly, the conversion obtained at this
temperature is slightly higher than the equilibrium conversion of acetone to 2-proanol61.
Meanwhile, CꢀO single bond breaking is unusual over copper; hence, excellent selectivity to
2ꢀpropanol was obtained over the copper at temperature up to 573 K. Figure S2
(supplementary information) emphasizes the thermodynamic limitation for acetone
hydrogenation over copper catalyst. Although higher rate can be obtained at higher reaction
temperature (473 K), the conversion levels at ~ 50%. At lower temperature (~ 448 K), the
conversion can be boosted up at higher contact time. It is worth noting that 2ꢀpropanol is the
only product over the copper catalyst despite the contact time has reached equilibrium. For
further investigation, the contact time in this work was tested within 80 g.h.molꢀ1.

Page 11 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
11
172
Effect of nickel and copper loading on SiO2 was studied (5ꢀ40 wt%) as shown in
Figure 4. When the metal loading is increased, acetone conversion is enhanced due to an
increase in number of active sites. The turnover frequency for all Cu/SiO2 catalyst is roughly
2.1ꢀ2.8 x 10ꢀ3 secꢀ1, as shown in Table S2 (supplementary information). However, the
normalized rate (mole of feed converted per hour over mole of the metal loaded) is decreased
due to agglomeration of the metal particles, which notably reduces the active metal surface
(Table 1). In case of nickel catalyst, the increased metal particle size also leads to an increase
hydrogenolysis activity, as seen from Figure 4a. The Ni catalysts with > 10% loading yield
significant amount of methane (~30%). Meanwhile, copper retains its excellent selectivity
173
174
175
176
177
178
179
180
181
182
183
towards C=O hydrogenation. This suggests that only ηꢀ1 adsorption mode is allowed for C=O
on Cu surface.58 Hence, 100% selectivity to 2ꢀpropanol is obtained over Cu catalyst up to
40% loading.
184
185
186
187
188
189
190
191
As methane was produced over the nickel catalyst at > 473 K, copper was mixed into
the Ni catalyst to retard the methane formation. Theoretically, nickel and copper are miscible
in wide range of component62, as evidenced by TPR (Figure S1). It is clear from Figure 5 that
the conversion is generally decreased with copper content in the alloy. A synergistic effect
observed at ~ 25%Cu is presumably due to a better dispersion of the metal phase. At 473 K,
no methane is observed from the catalysts with Cu content higher than 40 %. This is because
of the hydrogenalysis is a structure sensitive reaction, alloying with Cu shall readily inhibits
the adsorption mode that leads to CꢀC cleavage.63

Catalysis Science & Technology
Page 12 of 48
View Article Online
DOI: 10.1039/C5CY00367A
12
192
3.3 Dehydration of isopropanol
193
194
195
196
197
198
199
Dehydration of the hydrogenated product, 2ꢀpropanol was separately investigated
over Hꢀβ catalyst at 448 K. Figure 6 shows that 2ꢀpropanol conversion increased with the
contact time. Two products, propylene and diisopropyl ether are initially observed in parallel
from an intraꢀmolecular dehydration and interꢀmolecular dehydration, respectively. However,
the ether drops gradually at higher contact time (1.5ꢀ4.7 g.h.molꢀ1). This indicates that the
ether can be converted to propylene and 2ꢀpropanol64, as demonstrated by the reaction of
diisopropyl ether over Hꢀ β (Table 4) and below.
200
201
202
203
204
205
206
207
As also seen from Table 4, zeolite Y and β provide relatively higher activity, as
compared to other zeolites. However, the cage structure of Y can retain both feed and
products, which promotes the interꢀmolecular dehydration, as observed with high diisopropyl
ether selectivity at low contact time (3 g.h.molꢀ1). In addition to the structural effect, a closer
siteꢀproximity in HY (Si/Al ~8), as compared to that of Hꢀβ (Si/Al ~14) would enhance the
bimolecular dehydration. This also leads to a more rapid deactivation, as seen from a
relatively lower activity of HY after 6 h on stream, despite the acidity of HY is higher.

Page 13 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
13
208
At higher contact time (15 g.h.molꢀ1), it is clearly observed that the catalyst with three
dimensional pore opening (Hꢀβ) provides higher activity, as compared to those with two and
one dimensional pore opening (HZSMꢀ5 and HꢀMordenite), respectively. With a comparable
Si/Al, this suggests that diffusivity of the feed and products plays important role on
dehydration at relative low temperature (448 K). With poor mass transfer, reꢀhydration may
well take place at high contact time. Hence, mordenite gives even lower activity; as compared
to ZSMꢀ5 despite the pore size is larger. In addition to the observed activity, ZSMꢀ5 yields
only 2% diꢀisopropyl ether at 55 % conversion while β gives 10% diꢀisopropyl ether at
similar conversion. The result indicates that the medium pore size of ZSMꢀ5 somewhat
constrains the size of intermediates and products while the large pores of β provides intrinsic
space for the bimolecular reaction. Moreover, the large oneꢀdimensional pore system
particularly enhances the interꢀmolecular dehydration as the feed can be exceedingly retained
in the pore. In line with this view, thermogravimetric analysis shows that < 1 % hard coke is
detected on the used zeolite with medium pore while > 3 % is observed for the larger ones
(Table 3). The TGA under N2 defines that most of deposits in HZSMꢀ5 are the high
molecular weight compounds that cannot diffuse out of pore at the reaction temperature.
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
From Figure 7a, it can be seen that 2ꢀpropanol conversion is increased with the
temperature. Selectivity to propylene is also increased with the decrease in diꢀisopropyl
ether. This is simply because elimination of diꢀisopropyl ether to propylene is promoted at
high temperature. Again, the effect of pore size is revealed particularly at lower temperature.

Catalysis Science & Technology
Page 14 of 48
View Article Online
DOI: 10.1039/C5CY00367A
14
228
229
230
231
232
When Hꢀβ was used (Figure 7b), similar pattern to HꢀZSMꢀ5 was observed. As discussed
earlier, β possesses higher activity, as compared to ZSMꢀ5, but higher selectivity to diꢀ
isopropyl ether at low temperature. No higher hydrocarbon was detected for both catalysts at
this temperature range, presumably due to the strong adsorption of the feed and ether over the
acid sites.
233

Page 15 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
15
234
3.4 Ketones hydrodeoxygenation to olefins
235
236
237
238
239
240
241
In this work, hydrogenation – dehydration approach was employed for ketone
hydrodeoxygenation. To combine the two reactions with exothermic and endothermic nature
into one process, temperature has to be primarily optimized. For nickel ꢀ zeolite system,
hydrogenation with no hydrogenolysis takes place at 373 ꢀ 448 K (Figure 3). While
dehydration rate increases significantly above 423 K (Figure 7). Accordingly, only small gap
from 423 to 448 K is allowed for a consecutive hydrogenationꢀdehydration over Ni/SiO2ꢀ
zeolite system, as illustrated below
242
243
244
245
246
For Cu/SiO2 ꢀ zeolite system, no hydrogenolysis can be promoted up to 575 K (Figure
3). However, the reaction over acid site is limited at 473 K to avoid the drastic
oligomerization.65 Accordingly, the window for hydrodeoxygenation over Cu/SiO2 ꢀ zeolite
system is widely opened (423 – 473 K), as illustrated above.

Catalysis Science & Technology
Page 16 of 48
View Article Online
DOI: 10.1039/C5CY00367A
16
247
248
249
250
251
252
253
254
255
256
257
258
Therefore, the temperature at 448 K was selected for Ni/SiO2 ꢀ zeolite system to
obtain the maximum dehydration rate without hydrogenolysis. Since olefin hydrogenation
can also be promoted over Ni/SiO2 (Table 2), the hydrodeoxygenation is conducted in twoꢀ
bed system. The ketone is hydrogenated over Ni/SiO2 to an alcohol that becomes a feed for
the dehydration to olefin over the zeolite bed. On the other hand, hydrogenation of olefin
cannot be promoted over copper catalyst. Hence, single biꢀfunctional bed system can be used
for hydrodeoxygenation over Cu/SiO2 ꢀ zeolite catalyst. A higher temperature (at 473 K) was
chosen to obtain highest hydrogenation rate without oligomerization of the olefin products.
However, an undesirable condensation of acetone may take place over the acid sites.
Moreover, the ketone hydrogenation activity is relatively low, as compared to that of the
alcohol dehydration, as observed from Figure 4 and Figure 6. Accordingly, the metal
component in the biꢀfunctional bed system has to be greater than that of the zeolite.
259
260
Double bed system of Ni/SiO2 – zeolite
For this system, HZSMꢀ5 was used to minimize the ether formation (Figure 7). Table
261
262
263
264
265
266
5 shows that up to 60 % conversion can be obtained at 76+7 g.h.molꢀ1. However, 2ꢀpropanol
and diꢀisopropyl ether, which are intermediates, still remain in large amount. By increasing
the zeolite bed (76+27 g.h.molꢀ1), only 10 % 2ꢀpropanol is observed without ether. The
propylene is produced selectively (90% mol). No aldol product derived from the
condensation of acetone on zeolite is detected. This is probably because of strong adsorption
of alcohol that prevents the ketone condensation. However, a slight deactivation can be

Page 17 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
17
267
observed for the dehydration bed. This is concluded from a gradual decline in propylene
selectivity while the acetone conversion remains unchanged over 7 hours on stream (Figure
8). When a larger ketone, methyl ethyl ketone (MEK), is tested, a similar activity is obtained
(~60 %). However, relative low selectivity to the alcohol (2ꢀbutanol) is observed due to the
higher reactivity of butanol, as compared to propanol. Again, no ether and aldol product are
detected for hydrodeoxygenation of MEK. This is presumably due to a larger steric constrain
268
269
270
271
272
273
274
275
276
of any C4 species in the zeolite pore which inhibits the bi
-molecular reaction (either
etherification of the alcohol and aldol condensation of the ketone). Hence, the butanol
produced from MEK hydrogenation is selectively dehydrated to nꢀbutene (1ꢀbutene + 2ꢀ
butene; 87%), and then isomerization to i-butene (13%).
277
278
Physical Cu/SiO2 – zeolite mixed bed
For a single mixed bed, 5% Cu/SiO2 was physically mixed with the zeolites using
high metal/zeolite ratio at 473 K to prevent the condensation of ketone. From Table 6, it is
seen that 68 % conversion with excellent selectivity to propylene (96 %) is obtained over the
76+7 bed (For contact time, 76+7 refers to 76 g.h.molꢀ1 of Cu/SiO2 and 7 g.h.molꢀ1 of
HZSMꢀ5; 83 g.h.molꢀ1 in total). A significant improvement in acetone conversion over the
mixed bed, as compared to the equilibrium value (39% at 473 K, Figure 3) for acetoneꢀ
alcohol, is presumably due to a simultaneous conversion of 2ꢀpropanol to propylene over the
acid catalyst. This will noticeably reduce the rate of 2ꢀpropanol dehydrogenation to acetone,
and hence draw the conversion of the acetone beyond such equilibrium level. With this
279
280
281
282
283
284
285
286

Catalysis Science & Technology
Page 18 of 48
View Article Online
DOI: 10.1039/C5CY00367A
18
287
288
289
catalyst composition and reaction conditions, no condensation products from acetone is
observed. In fact, small amount 2ꢀpropanol (3 %) is intentionally retained to inhibit the
condensation of the ketone.
290
291
292
293
294
295
296
297
298
299
300
However, when Hꢀβ (14) is primarily mixed (76+7 g.h.molꢀ1), deactivation is severe.
The acetone conversion at 6 hours on stream drops drastically, as compared to ZSMꢀ5. This
is because β possesses a larger pore size, as compared to ZSMꢀ5. This facilitates a better
diffusion and adsorprtion of acetone in the pore of β. Hence, relatively less fraction of the
acetone interact with the metal surface and hydrogenation is limited. In line with this view,
the aldol products namely mesityl oxide, acetic acid and i-butene (~ 20% in total) is instead
produced, as previously observed in literatures66ꢀ67. Since aldol condensation is largely
pronounced, high molecular weight deposits (Table 3) and a rapid deactivation can be
expected. Accordingly, reducing the contact time of β to 2 g.h.molꢀ1 can offer 92 % propylene
selectivity without the condensation products. However, the acetone conversion is still lower
than that over Cu/SiO2/HZSMꢀ5.
301
302
303
304
305
306
When larger ketone (i.e. MEK) is tested over 5% Cu/SiO2 ꢀ HZSMꢀ5 mixed catalyst
(Table 7), the conversion (56 %) is slightly lower than the acetone. This is probably due to
steric hindrance of the adsorbed ketone on the metal surface. High selectivity to 1ꢀbutene and
2ꢀbutene (82 %) was observed with some isomerization products (i-butene; 14 %) and 2ꢀ
butanol (4 %). For larger cyclic ketone, cyclohexanone is more reactive than the aliphatic
ketone. 96 % selectivity to cyclohexene is obtained with some methylcyclopentenes (4 %).

Page 19 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
19
307
This is presumably due to the stronger adsorption of cyclohexanone, as compared to those of
acetone and MEK. It is also worth noting that reducing the reaction temperature or contact
time affects the dehydration rate significantly.
308
309
310
311
Bi-functional catalyst: Cu/zeolites
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
As high selectivity to olefin can be observed over Cu/SiO2/HZSMꢀ5 system, Cu
incorporated ZSMꢀ5 is tested for acetone conversion. From Table 7, 100% acetone
conversion can be obtained at 83 g.h.molꢀ1, as compared to 67.5% from the physical mixed
bed at 76+7 g.h.molꢀ1 (Table 6). It is worth noting that a zeolite with higher Si/Al is used for
the biꢀfunctional catalyst in order to keep the number of acid sites and metal loadings the
same as in the physical mixed bed system. Despite of that, a remarkable increase in activity,
as compared to physical mixed catalyst, is probably due to the close proximity for the metal
and the acid site. The hydrogenated intermediate, 2ꢀpropanol, may well be removed
immediately by dehydration over
a
nearꢀby acid sites, diminishing reversible
dehydrogenation of the 2ꢀpropanol formed back to acetone. This successive dehydration of 2ꢀ
propanol to propylene would significantly increase the rate of acetone conversion over the Cu
catalysts. In line with this view, the turnover frequency for hydrodeoxygenation (11.8 x 10ꢀ3
secꢀ1) is approximately fourꢀfold higher than that of hydrogenation (2.6 x 10ꢀ3 secꢀ1) at 473 K
(Table S2). However, propane is also produced over the biꢀfunctional catalyst, probably via
the hydrogen transfer. It is likely that, at high contact time, propylene can be adsorbed onto
the acid site. The hydrogenation of the adsorbed propylene may well be promoted by metalꢀ
acid site interaction. In other words, the adsorbed hydrogen on the metal can be source of
hydrogen transfer to the adsorbed olefin on the proximate acid site. It should be noted that

Catalysis Science & Technology
Page 20 of 48
View Article Online
DOI: 10.1039/C5CY00367A
20
330
331
332
333
334
335
336
direct hydrogenation of propylene over copper metal surface is not the case, as evidenced and
discussed previously. The propylene hydrogenation can be diminished by the decrease in
contact time. This is because relatively small amount of propylene can be adsorbed on the
acid site when large amount of alcohol is present in the reaction stream. For a comparison,
5%Cu/HY (100) shows a higher acetone conversion. This is because the HY (100) possesses
a higher adsorption and dehydration activity, as compared to HZSMꢀ5 (250) as discussed
earlier
337
338
339
340
341
342
343
344
345
346
Over 5%Cu/HY (100), the activity is in the order of MEK < cyclohexanone <
acetone, respectively. The observed low activity for MEK is due to a rapid deactivation as
seen by relatively higher carbon deposit shown in Table 3. As discussed previously,
oligomerization is increasingly promoted for the larger ketone, particularly over the catalyst
with larger pore opening such as HY. Accordingly, as the catalyst is deactivating, the
observed activity of cyclohexanone conversion at 6th hours on stream become slightly lower
than that of acetone, despite cyclohexanone shows a higher conversion over the mixed
catalyst (Table 6). This is because the large pore size with the cage structure of Y enhances
the oligomerization of cyclohexanone, as compared to that over the medium channel of
HZSMꢀ5. The amount of coke deposit is in line with the observed activity (Table 3).
347

Page 21 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
21
348
4. Conclusion
349
350
351
352
353
354
355
356
357
358
359
360
The optimization of reaction conditions and catalyst system for both hydrogenation
and dehydration leads to a potential approach for ketone deoxygenation to olefins. Nickel
catalyst is more reactive than the copper, but it is also capable to promote hydrogenation of
the olefin produced. Accordingly, a double bed system containing Ni/SiO2 and zeolites is
ideal for ketone hydrodeoxygenation to olefins. In contrast, the mixed Cu/SiO2ꢀzeolites and
Cu/zeolites can be used for single bed system. For the latter case, copper can hydrogenate
ketone to alcohol without hydrogenolysis and it is somewhat inert for olefin hydrogenation.
While zeolite can promote dehydration of the alcohol formed at the same temperature range
(≤ 473 K). A closeꢀproximity found in Cu/zeolites synergistically promotes the activity
towards olefin formation as the alcohol formed is subsequently removed over the neighboring
acid sites. The simultaneous conversion of alcohol to alkene, observed in both single bed
systems, boost the ketone conversion to exceed the ketoneꢀalcohol equilibrium level.
361
362

Catalysis Science & Technology
Page 22 of 48
View Article Online
DOI: 10.1039/C5CY00367A
22
363
Acknowledgement
364
365
366
367
368
The authors are grateful for a financial support from the Thailand Research Fund
through the Royal Golden Jubilee Ph.D. Program (Grant No. PHD/0137/2553) and assistance
from PTT Public Co. Ltd.,.
Note
369
370
aDepartment of Chemistry, Faculty of Science, King Mongkut’s Institute of Technology
Ladkrabang, Chalongkrung Road, Bangkok 10520, Thailand
371
372
bCatalytic Chemistry Research Unit, Faculty of Science, King Mongkut’s Institute of
Technology Ladkrabang, Chalongkrung Road, Bangkok 10520,
373
E-mail: kstawan@kmitl.ac.th; Fax: +662-326-4415; Tel: +6681-929-8288
374
375

Page 23 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
23
376
References
1 C. A. Mullen and A. A. Boateng, Energy Fuels, 2008, 22, 2104ꢀ2109.
2 M. F. Demirbas, Energy Sources Part A, 2010, 32, 909ꢀ916.
3 S. Suttibak, Proc. Sci. Eng., 2013, 2013, 338ꢀ343.
4 D. C. Elliott, T. R. Hart, G. G. Neuenschwander, L. J. Rotness and A. H. Zacher, Environ.
Prog. Sustainable Energy, 2009, 28, 441ꢀ449.
5 E. G. Baker and D. C. Elliott, Catalytic Hydrotreating of Biomass-Derived Oils, first ed.,
American Chemical Society, 1988.
6 J. Wildschut, I. MeliánꢀCabrera and H. J. Heeres, Appl. Catal., B, 2010, 99, 298ꢀ306.
7 C. A. Fisk, T. Morgan, Y. Ji, M. Crocker, C. Crofcheck and S. A. Lewis, Appl. Catal., A,
2009, 358, 150–156.
8 M. A. Naylor, US patent, 3271447, 1966.
9 E. Sert and F. S. Atalay, Ind. Eng. Chem. Res., 2012, 51, 6666ꢀ6671.
10 C. T. Kautter and U. Baumann, US patent, 3458561, 1969.
11 L. Wu, G. Wang and X. Chen, US patent, 6420307B1, 2002.
12 P. G. Menon, J. Catal., 1979, 50, 314ꢀ316.
13 L. Forni, M. Tescari and P. Zambelli, J. Catal., 1980, 65, 470ꢀ474.
14 Y. Liu, K. Murata, M. Inaba and N. Mimura, Catal. Lett., 2003, 89, 49ꢀ53.
15 S. Ghosh, S. S. Acharyya, R. Tiwari, B. Sarkar, R. K. Singha, C. Pendem, T. Sasaki and
R. Bal, ACS Catal., 2014, 4, 2169−2174.

Catalysis Science & Technology
Page 24 of 48
View Article Online
DOI: 10.1039/C5CY00367A
24
16 M. O. Robeson and T. P. Webb, US patent, 2623909, 1952.
17 L. P. Silva and E. F. Barbosa, Polypropylene, Nova Science Publishers, Inc., 2013.
18 L. Y. Zhang, G. Q. Fan, C. Y. Guo, J. Y. Dong, Y. L. Hu and M. B. Huang, Eur. Polym.
J., 2006, 42, 1043–1050.
19 P. Sun, M. Heng, S. H. Sun and J. Chen, Energy Convers. Manage., 2011, 52, 924–933.
20 A. Demirbas, Energy Convers. Manage., 2000, 41, 633ꢀ646.
21 M. A. Serio, E. Kroo and M. A. Wójtowicz, Prepr. Pap. Am. Chem. Soc., Div. Fuel
Chem., 2003, 48, 584ꢀ589.
22 K. Ross and G. Mazza, World J. Agric. Sci., 2011, 7, 763ꢀ776.
23 C. Song, A. Pawlowski, J. Ji, S. Shan and Y. Cao, Environ. Prot. Eng., 2014, 40, 35ꢀ43.
24 T. N. Pham, T. Sooknoi, S. P. Crossley and D. E. Resasco, ACS Catal., 2013, 3, 2456ꢀ
2473.
25 C. A. Gaertner, J. C. SerranoꢀRuiz, D. J. Braden and J. A. Dumesic, J. Catal., 2009, 266,
71ꢀ78.
26 T. N. Pham, D. Shi and D. E. Resasco, J. Catal., 2014, 314, 149ꢀ158.
27 M. Acke and M. Anteunis, Bull. SOC. Chim. Belges,, 1965, 74, 41ꢀ51.
28 A. F. Trasarti, N. M. Bertero, C. R. Apesteguía and A. J. Marchi, Appl. Catal., A, 2014,
475, 282–291.
29 G. D. Yadav and R. K. Mewada, Catal. Today, 2012, 198, 330– 337.
30 J. Deng, C. Lu, G. Yang and Z. Chen, React. Funct. Polym., 2012, 72, 378–382.

Page 25 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
25
31 A. Witsuthammakul and T. Sooknoi, Appl. Catal., A, 2012, 413– 414, 109– 116.
32 A. Rahman, Bull. Chem. Reac. Eng. Catal., 2010, 5, 113 – 126.
33 C. T. H. Stoddart and C. Kemball, J. Colloid Sci., 1956, 11, 532ꢀ542.
34 S. Qia, B. A. Cheney, R. Zheng, W. W. Lonergan, W. Yub and J. G. Chen, Appl. Catal.,
A, 2011, 393, 44–49.
35 T. M. Yurieva, L. M. Plyasova, O. V. Makarova and T. A. Krieger, J. Mol. Catal. A:
Chem., 1996, 113, 455ꢀ468.
36 R. Vidruk, M. V. Landau, M. Herskowitz, V. Ezersky and A. Goldbourt, J. Catal., 2011,
282, 215–227.
37 W. H. Wade, S. Teranishi and J. L. Durham, J. Colloid Interface Sci., 1966, 21, 349ꢀ357.
38 P. A. Jacobs, W. J. Mortier and J. B. Uytterhoeven, J. Inorg, Nucl. Chem., 1987, 40, 1919ꢀ
1923.
39 F. S. Stone and A. L. Agudo, Z Phys. Chem. Neue. Fol., 1969, 64, 161—170.
40 A. Ausavasukhi and T. Sooknoi, Appl. Catal., A, 2009, 361, 93–98.
41 N. Peamaroon and T. Sooknoi, Pet. Sci. Technol., 2012, 30, 1647–1655.
42 D. L. Hoang, T. T. H. Dang, J. Engeldinger, M. Schneider, J. Radnik, M. Richter and A.
Martin, J. Solid State Chem., 2011, 184, 1915–1923.
43 G. V. Sagar, P. V. R. Rao, C. S. Srikanth and K. V. R. Chary, J. Phys. Chem. B, 2006,
110, 13881ꢀ13888.

Catalysis Science & Technology
Page 26 of 48
View Article Online
DOI: 10.1039/C5CY00367A
26
44 A. Ausavasukhi, S. Suwannaran, J. Limtrakul and T. Sooknoi, Appl. Catal., A, 2008, 345,
89–96.
45 W. S. Chen, F. W. Chang, L. S. Roselin, T. C. Ou and S. C. Lai, J. Mol. Catal. A: Chem.,
2010, 318, 36–43.
46 H. S. Roh, W. S. Dong, K. W. Jun, Z. W. Liu, S. E. Park and Y. S. Oh, Bull. Korean
Chem. Soc., 2002, 23, 669ꢀ673.
47 N. Yao, H. Ma, Y. Shao, C. Yuan, D. Li and X. Li, J. Mater. Chem., 2011, 21, 17403ꢀ
17412.
48 S. Esposito, S. D. Vecchio, G. Ramis, M. Bevilacqua, M. Turco, G. Bagnasco, C.
Cammarano, A. Aronne and P. Pernice, Chem. Eng. Trans., 2007, 11, 83ꢀ88.
49 C. W. Tang, C. B. Wang and S. H. Chien, Thermochim. Acta, 2008, 473, 68–73.
50 N. A. M. Nasir, N. A. M. Zabidi and C. F. Kait, J. Appl. Sci., 2011, 11, 1391ꢀ1395.
51 C. H. Zhang, H. J. Wan, Y. Yang, H. W. Xiang and Y. W. Li, Catal. Commun., 2006, 7,
733–738.
52 X. Ge, H. Zou, J. Wang and J. Shen, React. Kinet. Catal. Lett., 2005, 85, 253ꢀ260.
53 C. S. Kim and S. I. Woo, J. Mol. Catal., 1992, 73, 249ꢀ263.
54 Z. Strassberger, A. H. Alberts, M. J. Louwerse, S. Tanase and G. Rothenberg, Green
Chem., 2013, 15, 768ꢀ774.
55 C. Huo, J. Ouyang and H. Yang, Sci. Rep., 2014, 4, 1ꢀ9.
56 R. S. Mann and T. R. Lien, J. Catal., 1969, 15, 1ꢀ7.

Page 27 of 48
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00367A
27
57 L. Brandao, D. Fritsch, L. M. Madeira and A. M. Mendes, Chem. Eng. J., 2004, 103, 89–
97.
58 S. Sitthisa and D. E. Resasco, Catal. Lett., 2011, 141, 784–791.
59 X. Fang, X. Min, L. X. Feng and H. X. Lan, J. Therm. Sci., 2013, 22, 613ꢀ618.
60 W. Mooksuwan and S. Kumar, Int. J. Energy Res., 2000, 24, 1109ꢀ1122.
61 E. Buckley an E.F.G. Herington, Trans. Faraday Soc., 1965, 61, 1618ꢀ1625.
62 P. Nash, Phase diagrams of binary nickel alloys, ASM International, 1991.
63 J. A. Rodriguez and D. W. Goodman, J. Phys. Chem., 1990, 94, 5342ꢀ5347.
64 H. J. Solomon, H. Bliss and J. B. Butt, Ind. Eng. Chem. Fundamen., 1967, 6, 325–333.
65 M. L. Occelli, J. T. Hsu and L. G. Galya, J. Mol. Catal., 1985, 32, 377 – 390.
66 T. Tago, H. Konno, S. Ikeda, S. Yamazaki, W. Ninomiya, Y. Nakasaka and T. Masuda,
Catal. Today, 2011, 164, 158–162.
67 A. J. CruzꢀCabeza, D. Esquivel, C. JiménezꢀSanchidrián and F. J. RomeroꢀSalguero,
Mater., 2012, 5, 121ꢀ134.

Catalysis Science & Technology
Page 28 of 48
Selective Hydrodeoxygenation of Bio-Oil Derived Products:
Ketones to Olefins
Ayut Witsuthammakul and Tawan Sooknoi*
Graphical Abstract

Page 29 of 48
Catalysis Science & Technology
X
H+
H2O
O
O
O
Si
Al
Cu
Hydrogenation of a ketone primarily takes place on the metal surface resulting in a
corresponding alcohol that can be dehydrated to olefin over the acid function. A rapid
dehydration synergistically prevent reversible dehydrogenation of alcohol while excessive
olefin hydrogenation can be limited over selected metal.