One-step continuous process for the production of 1-butanol and 1-hexanol by catalytic conversion of bio-ethanol at its sub-/supercritical state

Article abstract of DOI:10.1039/c3ra00134b

A one-step, easy to scale-up, continuous process was developed to catalytically convert bio-ethanol into 1-butanol and 1-hexanol as chemicals or fuels, as well as some other compounds that are also useful as biogasoline. The process is novel as it employs ethanol at its sub- and supercritical states. The reactions were performed in a continuous-flow fixed-bed tubular reactor with γ-alumina supported Ni catalysts at 135-300 °C at a weight hourly space velocity of 6.4-15.6 h^-1 using neat ethanol. The process at 250 °C and 176 bar with an 8% Ni/γ-alumina catalyst led to ethanol conversion of 35%, with the highest selectivity of approximately 62% and 21% towards 1-butanol and 1-hexanol, respectively. In addition to 1-butanol and 1-hexanol, other minor biogasoline components such as butanal, and 2-pentanone, etc. were also synthesized from this process. The catalyst was found active for 18 h on stream in this study, and the regenerated catalysts retained their activities for the reactions. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3ra00134b

RSC Advances
View Article Online
View Journal | View Issue
PAPER
One-step continuous process for the production of
1
-butanol and 1-hexanol by catalytic conversion of bio-
Cite this: RSC Advances, 2013, 3,
4
271
ethanol at its sub-/supercritical state
ab
a
Hassan S. Ghaziaskar and Chunbao (Charles) Xu*
A one-step, easy to scale-up, continuous process was developed to catalytically convert bio-ethanol into
-butanol and 1-hexanol as chemicals or fuels, as well as some other compounds that are also useful as
1
biogasoline. The process is novel as it employs ethanol at its sub- and supercritical states. The reactions
were performed in a continuous-flow fixed-bed tubular reactor with c-alumina supported Ni catalysts at
21
1
35–300 uC at a weight hourly space velocity of 6.4–15.6 h using neat ethanol. The process at 250 uC and
76 bar with an 8% Ni/c-alumina catalyst led to ethanol conversion of 35%, with the highest selectivity of
1
Received 9th January 2013,
Accepted 14th January 2013
approximately 62% and 21% towards 1-butanol and 1-hexanol, respectively. In addition to 1-butanol and
-hexanol, other minor biogasoline components such as butanal, and 2-pentanone, etc. were also
1
DOI: 10.1039/c3ra00134b
synthesized from this process. The catalyst was found active for 18 h on stream in this study, and the
regenerated catalysts retained their activities for the reactions.
www.rsc.org/advances
separation as the ethanol/gasoline blend. Butanol has char-
acteristics similar to gasoline in terms of energy density (29.2
1
. Introduction
Due to the increasing concern over greenhouse gas emissions,
diminishing crude oil reserves, and the rise of crude oil prices,
there is substantially increased need for alternative green
sources for fuels and chemicals. Butanol is an important
chemical and fuel and potentially a better fuel extender than
ethanol, and it has a worldwide market of 370 million gallons
per year. The market demand is expected to increase
dramatically if green butanol can be produced economically
from bio-ethanol. Butanol over ethanol as a fuel has several
advantages, including tolerance to water contamination,
which makes it more suitable for transportation via pipelines.
The U.S. fuel market has been constrained by the fact that
ethanol, which absorbs water and therefore tends to corrode
pipelines, must be transported on trucks, trains and barges in
relatively small batches to storage terminals where it is
blended with gasoline. Another advantage of butanol is that
it can be blended into gasoline at higher concentrations than
ethanol without the need to modify the existing vehicle engine
system, and it offers better fuel economy than gasoline–
ethanol blends (as butanol has y25% more heating value
than ethanol). n-Butanol is being viewed as the next oxygenate
to replace ethanol in fuel applications. It behaves similarly as
gasoline although it is an oxygenate. It does not create phase
21
vs. 32 MJ L ), stoichiometric air–fuel ratio (11.1 vs. 14.6), heat
21
of vaporization (0.43 vs. 0.36 MJ L ), research octane number
(96 vs. 91–99), and motor octane number (78 vs. 81–89), so it
has attracted considerable attention as a substitute for
1–4
gasoline recently.
A newer butanol synthesis pathway through Guerbet
chemistry was demonstrated possible to produce butanol
5
from ethanol through bimolecular ethanol condensation.
Guerbet alcohols have been known for over 100 years since
6
Marcel Guerbet pioneered the basic chemistry in the 1890s. In
the Guerbet process, bi-functional catalyst (metal/metal oxide
catalysts) are needed to convert one alcohol into aldehyde to
activate the b-hydrogen and to combine with another ethanol
through aldol condensation after proton extraction by dehy-
dration, followed by hydrogenation of the allylic aldehyde to
produce butanol. The Guerbet reaction requires one of the
alcohols to have at least two carbons because when a metal
catalyst converts one of the alcohols into aldehyde, the
b-hydrogen becomes acidic and is ready to be removed by a
dehydration metal oxide catalyst to chemically combine the
acidic hydrogen with a hydroxyl group from another alcohol to
produce a heavier alcohol or aldehyde and water (through
aldol condensation). The aldehyde can either leave the catalyst
surface in the form of a higher molecular weight aldehyde or it
can be hydrogenated back to a higher molecular weight
alcohol over the metal catalysts.
a
Institute for Chemicals and Fuels from Alternative Resources (ICFAR), Faculty of
Engineering, Western University, London, ON, Canada, N6A 5B9.
E-mail: cxu6@uwo.ca; Fax: +1-519-661-4016; Tel: +1-519-661-2111 ext. 86414
b
For ethanol conversion, choice of the elements for both the
metal and the support is a key factor for developing supported
Department of Chemistry, Isfahan University of Technology, Isfahan, 84156-83111,
I. R. Iran
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 4271–4280 | 4271

View Article Online
Paper
RSC Advances
7
catalysts with characteristics that can fulfill the requirements.
be an ideal reaction medium as it has higher diffusivity, lower
1
5
Different catalyst compositions have been proposed by various
researchers. Tsuchida et al. examined hydroxyapatite catalysts
with various Ca/P ratios for bimolecular conversion of ethanol
viscosity and better solubility for organic compounds. In this
work, we intended to develop a novel one-step continuous
process for the production of 1-butanol and 1-hexanol by
catalytic conversion of bio-ethanol at its sub-/supercritical
state.
8
to butanol. They obtained 20% conversion and about 70%
selectivity towards butanol at the optimum Ca/P ratio of 1.67
and 298 uC while ethanol gas was diluted to 16.4 vol% with
helium. Ndou et al. investigated the effect of several solid-base
catalysts for ethanol to butanol conversion, including MgO,
2 3
CaO, BaO, c-Al O , Na, K, and Cs on alumina as well as Mg on
9
2. Material and methods
silica. They found that MgO was the best basic oxide catalyst,
and Zn could be an additional electron-density donor to
promote ethanol conversion but it could not increase the
butanol selectivity. The MgO catalyst offered 56% ethanol
conversion with 18.4% butanol selectivity at 450 uC, while the
Zn-promoted MgO offered much higher ethanol conversion
2.1. Materials
Corn-derived anhydrous ethanol (purity .99%) was purchased
from Commercial Alcohols Inc. Methanol (purity .99%) was
used as solvent obtained from J.T. Baker, USA. c-Alumina
(Catalox1 SCCa-5/150, purity 97.6 w/w% with ,1.0%
2
21
a-Alumina, surface area 152 m
g
, pore volume 0.49 mL
2
1
(
95.6%) but with much lower butanol selectivity (2.95%). Yang
et al. investigated the performance of 8% Fe, Co, and Ni on
c-Al as the catalysts for conversion of ethanol in vapor
g
, and with a mean particle size of 85 mm) were supplied by
Sasol North America Inc. Mn O (purity .98%, density 4.5 g
2
3
2
1
O
3
mL , mesh .325) was obtained from Alfa Aesar1, MA, USA.
Manganese(II) nitrate, hydrate (purity .98 w/w% and degree of
hydration of 4–6), nickel(II) nitrate hexahydrate, ammonium
molybdate tetra hydrate, and 1-butanol (purity .99.8%) were
all purchased from Sigma-Aldrich Co., USA. Pyrex glass beads
2
1
0
phase to butanol via Guerbet chemistry at 200 uC, where the
nickel catalyst produced 19.1% ethanol conversion and 64.3%
selectivity to butanol.
A mixture of ethanol : 1-propanol (4 : 1) with helium as the
carrier gas (.80 vol%) was converted to 1-butanol, 2-methyl-1-
(20–50 mesh) were used for diluting some of the catalysts.
1
1
butanol, 1-pentanol, and 2-methyl-1-pentanol elsewhere.
2.2. Catalyst preparation
Conversion of ethanol : 1-propanol (4 : 1) to butanol through
Guerbet chemistry demonstrated by Sangi Co., Ltd. using
hydroxyapatite catalyst in a patent application (claiming a
n-butanol yield of y21% at a relatively high temperature of
The catalysts of Ni/c-Al O (or simply denoted as NiAl in this
2 3
paper) with varying nickel contents (8–27 wt%) were prepared
by the wet impregnation method using water soluble metal
nitrate salt and c-alumina support. Briefly, a pre-calculated
411 uC). In this patent, quantification of the conversion and
3 2 2
amount of c-alumina and Ni(NO ) ?6H O dissolved in distilled
product yields were qualitative as they used peak area to
quantify the amount of each component assuming same
response factors for all compounds. As the response factor of
GC-FID is different for each compound, measurements using
only peak area can introduce a large error in the results.
Metals like Ni can act as the hydrogen autotransfer
water were mixed in a flask. The mixture was shaken in a warm
bath at 40 uC for 12 h, followed by vacuum drying at 20.9 bar
and 80 uC in a rotary evaporator (150–200 rpm) for y2 h and
oven dried at 120 uC overnight. The dried solids were then
subject to calcination and reduction in a continuous-flow
tubular reactor (3/8-inch OD, 11-inch length) heated in an
electric furnace at 500 uC. High purity hydrogen with a flow
rate of 100 mL min controlled by a mass flow controller
(model FMA 5510, OMEGA, U.S.A) was passed over the dried
Ni(II) impregnated alumina for 8 h at 500 uC to reduce nickel
ions to metallic nickel. In most cases when a nickel loaded
c-alumina catalyst was only used, the reduction was performed
in situ, right prior to the reaction. In case of mixed catalysts,
after the calcination/reduction of nickel ions was completed
the catalyst was passivated by passing nitrogen gas over it for 2
h during the cooling time of the reactor, and then it was mixed
at a specific amount with metal oxide catalyst and packed into
the reactor.
1
2
catalysts responsible for the ethanol dehydrogenation to
aldehydes and hydrogenation of the heavier aldehydes to
heavier alcohols, while the metal oxides are dehydration
2
1
1
3
catalysts, as was also reported by a US patent.
mechanism of hydrogen autotransfer in metal (M « MH
has been well known, and the aldehyde can be reduced to form
heavier alcohols by the H -adsorbed on a metal catalyst. Also
The
2
)
2
1
4
Alonso et al. examined the efficiency of hydrogen auto-
transfer catalysts and found that Ru worked better for
homogeneous catalysis, while Ni was good for heterogeneous
catalysis. In O’Lenick’s literature review on Guerbet chemistry,
a number of metal oxides such as copper, manganese,
chromium, cobalt, and molybdenum oxides were suggested
for the dehydration reaction, while nickel, platinum, and
palladium metals were suggested as active dehydrogenation/
For the comparison purpose, the mixture of 27% Ni/
c-alumina (2.48 g) and neat commercial Mn O powder (0.50
2
3
g) was used, corresponding to a Ni/Mn O ratio of 1 : 1 (w/w).
2
3
Mn
muffle furnace for 6 h before use. In addition, we prepared 20
wt% Mn /c-Al catalyst with a similar wet impregnation
2 3
O (98 wt%, -325 mesh powder) was calcined at 550 uC in a
6
hydrogenation catalysts.
Sub-/supercritical fluids are in a special phase achieved
when the pressure and temperature of a compound are
increased to near or above their respective critical points
2
O
3
2 3
O
method as described previously. The only difference was that
in the preparation of 20 wt% Mn O /c-Al O , no reduction is
2
3
2 3
(240.7 uC, 61.4 bar for ethanol). A sub-/supercritical fluid can
needed, while the calcination was performed in a muffle
4
272 | RSC Adv., 2013, 3, 4271–4280
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Fig. 1 Schematic diagram of the set-up used for sub- and supercritical ethanol conversion.
furnace for 6 h in air. In the reaction tests, 2.38 g of 20 wt%
the reaction. Herein the weight of nickel loaded on the support
was taken as the catalyst mass to calculate the WHSV.
In some tests the catalyst particles are mixed with 20–50
mesh Pyrex glass beads (at 1 : 1 w/w), in order to increase the
residence time of ethanol in the catalyst bed. However, it was
Mn
Mn
2
O
O
3
2 3
/c-Al O catalyst was used, equivalent to 0.5 g neat
2
3
.
2.3. Experimental apparatus and methods
Experiments were carried out in a continuous fixed-bed
tubular reactor (3/8-inch OD, 11-inch length) mounted
vertically in an electric furnace (model 21100, Barnstead
Thermolyne with Euroterm 2116 temperature controller), as
schematically shown in Fig. 1. The bottom of the reactor was
connected to a cooling jacket, a 2 mm stainless steel filter
observed that diluting 8% Ni/c-Al O or 27% Ni/c-Al O
2 3 2 3
catalyst with the Pyrex glass beads resulted in lower catalyst
efficiency. Therefore, in most of the experiments the catalysts
were not diluted with Pyrex glass beads.
For the continuous reaction system, the conversion was
calculated as percent moles of ethanol reacted to the moles of
ethanol fed to the reactor and the yield of each product was
calculated as percent moles of carbon in the product to the
moles of carbon in the fed ethanol. The selectivity of each
product was also calculated as the percent moles of carbon in
the target product divided to the total moles of carbon in the
ethanol reacted. These calculations can be given simply by the
following eqns (1)–(3):
(
(
Swagelok Co., UK), pressure gauge, a back-pressure regulator
KPB Swagelok series, 0–276 bar, UK), and a liquid–gas
separator. A piston pump (series III, Lab Alliance Co.,
Canada) was used to feed liquid ethanol continuously into
the reactor at 150–300 uC and pressure of up to 180 bar.
Gasification of ethanol was observed at the temperatures of
250 and 300 uC. The gases were collected in a gas bag and
analyzed using a micro GC and found that they were mostly
hydrogen, methane, carbon dioxide, ethane, and small
amount of propane.
Moles of ethanol reacted
Conversionð%Þ~
|100 (1)
|100(2)
Moles of ethanol fed to the reactor
The experiments were performed at different temperatures
and pressure by pumping ethanol as the only reactant with a
Moles of carbon in the product
Yield %
ð Þ~
2
1
flow rate of 0.5–1.0 mL min over 3.0–3.5 g of solid catalyst
throughout the work, corresponding to a WHSV of 6.4–15.6
Moles of carbon in ethanol fed to the reactor
2
h
1
. WHSV stands for weight hourly space velocity and
Moles of carbon in the target product
Selectivityð%Þ~
|100 (3)
calculated as feed room temperature mass flow/catalyst mass.
The feedstock reservoir was put on a top loaded balance (AND
Co.) to accurately measure the weight of ethanol fed to
Moles of carbon in ethanol reacted
2
1
calculate the actual mass flow (g min ) during the course of
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 4271–4280 | 4273

View Article Online
Paper
.4. Products analysis
RSC Advances
2
mL of them was injected into the GC-MS and the total ion
chromatogram were obtained. The results were plotted with
the ratio of the peak area of the analytes to the peak area of the
internal standard on the X-axis and the ratio of the
concentration of the analytes to the concentration of internal
standard on the Y-axis. The calibration curves were con-
structed using linear regression line fitted into the calibration
data. With this, the relative concentration of the analyte in an
unknown sample to that of the internal standard was
calculated based on the corresponding ratio of peak area of
the analyte to that of the internal standard and the calibration
factor. GC-MS analyses of mixtures of known compositions of
ethanol, 1-butanol, tetrahydrofuran, ethyl acetate, acetalde-
hyde, 1,1-diethoxyethane and 1-hexanol and toluene as an
internal standard sample (,0.5 v/v% each) showed that the
measurement accuracy by % relative standard deviation
The liquid reaction products collected at different time
intervals were analyzed with a Shimadzu GC-MS (model
QP2010 series, Kyoto, Japan), with ionization achieved by
electron impact at 70 eV, equipped with a capillary column:
SHRXI-5MS equivalent of DB-5 and HP-5 (30 m length, 0.25
mm i.d., and 0.25 mm film thickness). The operating
conditions of the GC-MS analysis were: injection port
temperature, 260 uC; interface temperature, 270 uC; column
oven temperature program: 40 uC for 2 min, ramped at 80 uC
2
1
min to 240 uC with 1 min hold; helium carrier gas (flow rate
2
1
of 1.04 mL min
injector was operated in the split mode with a split ratio of
00. At a higher split ratio higher volumes of the samples can
at 40 uC); 0.1 mL injection volume. The
2
be injected to have better analytical precision.
To prevent the interference of methanol as the solvent, the
m/z scan was started from 40 amu. The total ion current versus
time was recorded to obtain the chromatogram for quantita-
tive work by GC-MS, and the components mass spectrum
within a mass range of 40 to 400 were used for qualitative
identification purposes. Identification of compounds was
achieved using the 8th edition of Wiley mass spectra database,
and then by comparing the mass spectra and retention times
of the chromatographic peaks with those of authentic
samples. The physical properties of eight major products of
ethanol conversion over different catalysts are given in Table 1.
Some other minor compounds were also formed in the
reactions that make the composition of the liquid product
(
%RSD) was acceptably low (6.6–15.5%).
Gases produced in the runs at temperatures of 250 uC and
00 uC were collected in gas sampling bags and analyzed with
3
an Agilent micro GC 3000A (model G2801A). The micro GC-
TCD consists of two channels and two columns: Molecular
Sieve 5A PLOT (10 m, 0.32 mm, 30 mm) and Plot Q (3 m, 0.32
mm, 30 mm) with argon and helium carrier gas, respectively.
2
.5. Catalyst characterization
A surface area analyzer (Nova 2200e/TO, Quantachrome
Instruments, USA) was used to measure the surface areas of
the fresh and spent catalysts. In the measurements, the
samples were degassed at 250 uC for 2 h and the weight of the
sample was exactly measured via an analytical balance, prior to
N2 isothermal adsorption at 77 K. Powder X-ray diffraction
(XRD) patterns of fresh and spent catalysts were recorded on a
PANalytical XRD instrument with Cu-Ka1 radiation. XRD
patterns were recorded with 0.02 (2h) steps over 2h of 6–97u
angular range and
Thermogravimetric analysis (TGA) of fresh and spent catalysts
were carried out using TGAi 1000 (Instruments specialists
Inc.), where the catalyst samples were heated in 30 mL min
air, 25 uC to 900 uC at 20 uC min
5
similar to a biogasoline as reported elsewhere.
Quantification analyses were performed by injections of
standard solutions of various concentration, containing
ethanol, 1-butanol, tetrahydrofuran, ethyl acetate, acetalde-
hyde, 1,1-diethoxyethane, and 1-hexanol and toluene as an
internal standard. Multiple point internal standard method
with six analyses was used for the quantitative analyses. The
internal standard method accounts for any variances in gas
chromatograph performance. Toluene was chosen as an
internal standard since it has a predictable retention time
and area. Moreover, it had a unique retention time that was
not overlapping with other components. Each analysis
contains the internal standard whose concentration was kept
constant and the analyte of interest whose concentration
covers the range of the expected concentrations. The standard
sample solutions up to 5 mL were prepared using the
authentic standards with methanol as the solvent, and 0.1
2
1
a
vscan rate of 0.16u
s
.
2
1
2
1
.
3
. Results and discussions
A one-step continuous process was developed for the produc-
tion of 1-butanol and 1-hexanol by catalytic conversion of bio-
ethanol at its sub-/supercritical state. The experimental work
Table 1 Specification of compounds used for standard solution preparation
21
21
No.
Compound name
Retention time (min)
Molecular weight (g mol
)
Density (g mL
)
Boiling point (uC)
1
2
3
4
5
6
7
8
9
Acetaldehyde
Ethanol
Butanal
Ethyl acetate
Tetrahydrofuran
1-Butanol
2-Pentanone
1,1-Diethoxyethane
1-Hexanol
1.39
1.44
1.79
1.89
2.01
2.23
2.45
2.85
4.12
44.05
46.07
72.11
88.11
72.11
74.12
86.13
118.17
102.17
0.785
0.789
0.800
0.902
0.889
0.810
0.809
0.831
0.814
21
78
75
77
66
117
103
102
157
4
274 | RSC Adv., 2013, 3, 4271–4280
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
was observed more at low temperatures or at subcritical
conditions, while 1-butanol is promoted as the major product
at supercritical conditions. Formation of acetal could occur
when the hydroxyl group of a hemiacetal molecule becomes
1
7
protonated to lose a water molecule. The carbocation that is
produced can then be rapidly be attacked by another molecule
of alcohol.
Loss of the proton from the attached alcohol gives the
acetal. Acetals are stable compared to hemiacetals but their
formation is reversible as with esters. As a reaction to create an
acetal proceeds, water must be removed from the reaction
mixture or it will hydrolyse the product. The formation of
acetals reduces the total number of molecules present and
therefore entropy-wise is not favourable. Minor change vs.
temperature was observed for the ethyl acetate yields. At
temperatures of 300 uC and higher, more unwanted products
and mostly gases were produced. Therefore, in this work 250
uC was used in investigating the effects of pressure on ethanol
conversion.
Fig. 2 The % conversion of ethanol and the % yield of main products formed
2 3 2 3 2 3
over the solid bi-functional catalyst Mn O /c-Al O mixed with Ni/c-Al O (1 : 1
w/w) at 141 bar at different temperatures.
3
.2. Effects of pressure
was started with conversion of ethanol over a solid catalyst
including Mn /c-Al mixed with Ni/c- Al , and Ni/
c-Al O with different nickel loadings of 8, 17, and 27 wt% at
various temperatures and pressure. Among various transition
metals, the high C–C bond-breaking activity of nickel, and its
relatively low price make it a promising active metal for
ethanol conversion. The performance of the above catalysts in
direct conversion of sub-/supercritical ethanol to higher
alcohols is detailed below.
Interestingly, our preliminary tests demonstrated that Ni/
c-Al O alone is more active than the mixed catalyst of Mn O /
c-Al O and Ni/c-Al O for the one-step conversion of ethanol
to higher alcohols. As such, effects of reaction pressures on
2
O
3
2
O
3
2 3
O
2
3
2 3
2
3
2
3
2 3
product yields were examined at 250 uC under different
pressure (3.8–183 bar) using Ni/c-Al
loading of 8 wt% and 27 wt%, diluted with Pyrex glass beads
1 : 1 w/w) in sub- and supercritical conditions. The results are
2 3
O catalyst with nickel
(
presented in Fig. 3. The typical qualitative composition of the
products in the tests at different pressure levels are displayed
in Table 2. As is shown previously in Table 2, the composition
of products at different pressures is similar. However, when
the pressure increases, the amount of 1-hexanol and 1-butanol
derivatives increases. In the tests at 250 uC, 183 bar, and 8%
3.1. Effects of temperature
The reactions were performed in the above described fixed-bed
tubular reactor at 150–300 uC using neat ethanol at a flow rate
2
1
of 1.0 mL min ethanol. After a constant temperature and
pressure was achieved, five samples were collected in incre-
ments of 14 mL and the last sample was analysed for
quantification. The percent yield (%Y) of products formed
and the percent conversion (%C) during the continuous
conversion of ethanol over the solid bi-functional catalyst
2 3
Ni/c-Al O catalyst, the %Y of the typical components is listed
2 3 2 3 2 3
Mn O /c-Al O with Ni/c-Al O (1 : 1 w/w) at different tem-
peratures at pressure of 141 bar are shown in Fig. 2. At 150 uC
extremely low ethanol conversion was obtained, while at a
higher temperature of 200 uC three major compounds of
1-butanol, ethane 1,1-diethoxy (i.e. acetal) and ethyl acetate
were formed. At 250 uC, seven compounds of 1-butanol,
1
2
-hexanol, acetal, ethyl acetate, acetaldehyde, butanal, and
-pentanone were detected. The %Y of 1-butanol and %C of
ethanol were increased drastically with increasing tempera-
ture. While the %Y of the acetal increased with increasing
temperature up to 200 uC and then reduced after that. As is
well known, the density of hot-compressed ethanol decreases
smoothly with increasing temperature up to 200 uC, while it
suddenly drops when it is close to its critical point (240.7 uC,
1
5,16
6.1 MPa).
This drop in ethanol density at sub- and
supercritical conditions may govern the product selectivity
towards the compounds which have a higher solubility in sub-
and supercritical ethanol. This might be the reason that the
formation of acetal (more polar and highly soluble in ethanol)
Fig. 3 The % yield of 1-butanol and 1-hexanol products versus pressure at 250
uC using 8% Ni/c-Al O3 (solid symbols) and 27% Ni/c-Al O3 (open symbols)
diluted with Pyrex glass beads (1 : 1 w/w).
2
2
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 4271–4280 | 4275

View Article Online
Paper
RSC Advances
Table 2 Product composition at 250 uC and different sub- and supercritical pressure using 27% Ni/c-Al
2
O
3
catalyst, diluted with Pyrex glass beads (1 : 1 w/w)
Pressure of reaction (bar)
No.
3.8
86
141
183
1
2
3
4
5
6
7
8
9
Acetaldehyde
Acetone
Diethyl ether
Butanal
Ethyl acetate
1-Butanol
2-Pentanone
Acetaldehyde
Acetone
Diethyl ether
Butanal
Ethyl acetate
1-Butanol
2-Pentanone
Ethane, 1,1-diethoxy
1-Butanal, 2-ethyl
1-Butanol, 2-ethyl
1-Hexanol
Acetaldehyde
Acetone
Diethyl ether
Butanal
Ethyl acetate
1-Butanol
2-Pentanone
Ethane, 1,1-diethoxy
1-Butanal, 2-ethyl
1-Butanol, 2-ethyl
1-Hexanol
Acetaldehyde
Acetone
Diethyl ether
Butanal
Ethyl acetate
1-Butanol
2-Pentanone
1-Butanol, 2-methyl
Ethane, 1,1-diethoxy
1-Butanal, 2-ethyl
1-Butanol, 2-ethyl
1-Hexanol
1-Ethoxy, butane
1-Butanal, 2-ethyl
1-Hexanol
1
1
1
0
1
2
Butanoic acid ethyl ester
Butane, 1,1-diethoxy
in Table 2, i.e., acetaldehyde, butanal, ethyl acetate, 1-butanol,
-pentanone, and 1-hexanol were 0.93, 4.11, 0.74, 20.16, 1.26,
and 5.8, respectively.
Pyrex glass beads a lower catalyst efficiency was achieved.
Therefore, in other experiments the catalysts were not diluted
with the Pyrex glass beads.
2
Fig. 3 illustrates the %yield of 1-butanol and 1-hexanol
The variations in temperature and pressure could signifi-
cantly change the density of the reaction mixture and hence
the residence time. As expected a longer residence time would
lead to a higher conversion. Thus, at higher pressure, the
density of the reaction mixture increases, so does the
residence time, and as a result the conversion. The other
reason for the positive effects of pressure on conversion may
be simply that higher pressure promotes the reaction kinetics
by the increase in the reactant concentration.
products versus pressure at 250 uC using 8% Ni/c-Al
7% Ni/c-Al diluted with Pyrex glass beads (1 : 1 w/w). As
shown in the Figure, with increasing the reaction pressure, the
Y of 1-butanol is increased drastically and it is almost
doubled from 3.8 bar to 176–183 bar. The highest 1-butanol
Y was 12.8% at 183 bar with 27% Ni/c-Al . The selectivity
at pressure of 176–183 bar was better than 70% for 1-butanol.
While the 1-butanol yield with 8% Ni/c-Al is almost double
that of 27% Ni/c-Al , the yields of 1-hexanol with both
catalysts remained almost the same (,2%), while the
2 3
O and
2
2 3
O
%
%
2 3
O
2 3
O
2 3
O
3.3. Effects of catalyst types and nickel loading
1
-hexanol yield slightly increases with increasing pressure.
Comparing the results displayed in Fig. 2 and 3, the use of
% or 27% nickel loaded on c-alumina as the catalyst showed
Effects of type of catalysts and nickel loading on the product
yields and selectivity as well as the ethanol conversion were
2
1
8
studied at 176 bar, 250 uC, and a flow rate of 0.5 mL min
2
1
higher catalytic activity toward 1-butanol synthesis at sub- and
supercritical conditions than the bi-functional catalysts (i.e.,
mixed catalyst of Mn O /c-Al O and Ni/c-Al O ). While, the
2 3 2 3 2 3
yield of all other products excluding acetaldehyde was found
below 2%. It shall be emphasized that, in these tests the
catalysts are mixed with Pyrex glass beads (1 : 1 w/w), which
may be advantageous for a longer ethanol residence time in
the catalyst bed. However, it was also observed that when 8%
(corresponding to a WHSV of 6.4 h ). The results are
comparatively shown in Table 3. As shown in Table 3, the
highest product yield (21.6%, molar basis or 17%, weight
basis), selectivity (61.7%), and ethanol conversion (35%) were
obtained with the 8% Ni/c-Al O catalyst. At the same
2
3
conditions, the highest selectivity of 20% and yield of y5%
were also achieved for 1-hexanol. This work is thus superior to
the recently published patent on conversion of ethanol in
vapor phase at 200–350 uC and normal pressure using a
2 3 2 3
Ni/c-Al O or 27% Ni/c-Al O catalyst was diluted with the
Table 3 Performances of different catalysts in conversion of ethanol to 1-butanol at supercritical conditions (250 uC and 176 bar)
Catalyst
% Ni/c-Al
Reg. 8% Ni/c-Al
Reg. 8% Ni/c-Al
1-Butanol yield (%)
Ethanol conversion (%)
1-Butanol selectivity (%)
8
2
O
3
21.6
20.3
19.0
18.3
10.35
18.5
35.0
33.9
33.4
28.8
15.9
29.3
61.7
59.9
56.8
63.6
65.0
63.2
a
a
a
2
2
2
O
O
O
3
3
3
Reg. 8% Ni/c-Al
b
8
1
2
2
% Ni/c-Al
2
O
3
7% Ni/c-Al
7% Ni/c-Al
7% Ni/c-Al
2
2
2
O
O
O
3
3
3
Full gasification of ethanol
b
9.2
12.0
4.0
16.7
18.7
9.8
55.2
64.0
40.8
b
Mn
2
O
3
/c-Al
2
O
3
+ 27% Ni/c-Al
2
O
3
3
b
2 3 2
Neat Mn O + 27% Ni/c-Al O
a
21
b
Regenerated catalysts at flow rate of 100 mL min
particle size of 20–50 mesh (1 : 1 w/w).
H
2
and 500 uC for 4, 6, and 8 h, respectively. Dispersed with Pyrex glass beads of a
4
276 | RSC Adv., 2013, 3, 4271–4280
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
catalyst comprising strontium phosphate apatite having an Sr/
The number of compounds produced with Ni/c-Al
catalysts was found to be much less compared to the 1 : 1
mixture catalyst of (Mn /c-Al Ni/c-Al ). More
importantly, the selectivity of the reaction using 8% Ni/
c-Al catalyst is much higher than what was reported
previously for the Guerbet reactions. The superior perfor-
mance of the Ni/c-Al over the bi-functional catalyst
suggests different mechanism involved in the reaction system
of this work than the Guerbet mechanism. Direct condensa-
tion of two ethanol molecules into 1-butanol might be involved
2 3
O
1
8
P atomic ratio of 1.5–2.0 where the n-butanol yield (i.e.
4.8%) and ethanol conversion (i.e. 21.1%) were reported. The
1
2
O
3
2
O
3
+
O
2 3
reason for the superior performance of our catalytic reaction
system might be owing to our reaction conditions involving a
different catalyst and sub-/supercritical ethanol under pres-
sure. Sub-/supercritical ethanol could be a superior reaction
medium as it has higher diffusivity, lower viscosity and better
2 3
O
1
0
2 3
O
1
5
solubility for organic compounds. It was observed that when
7% Ni/c-alumina was used as a catalyst, all fed ethanol was
gasified to form CH , H and C gaseous products. In
2
H
6
2 5 4 9 2
in this process (i.e., 2C H OH A C H OH + H O). However, the
4
2
2
contrast, when the 27% Ni/c-alumina catalyst was diluted with
Pyrex glass beads gasification was not severe but the product
yields and conversion of the reaction were lower compared to
the 8% Ni/c-Al O or 17% Ni/c-Al O with or without dilution
formation of some side products as shown in Table 2, suggests
oxidation of ethanol and butanol to aldehydes, aldol con-
densation after proton extraction, dehydration of the aldol
products, and hydrogenation of the allylic aldehydes, typical of
the Guerbet reactions. Therefore, the Guerbet mechanism
2
3
2 3
of Pyrex glass beads.
Fig. 4 Product yields and selectivity versus time at supercritical condition of 176
bar and 250 uC using 8% Ni/c-Al catalyst. The reaction was carried out for 13
2
O
3
Fig. 5 Product yields and selectivity for the reaction at supercritical condition of
h on Day 1 and then was continued for another 5 h on Day 2. The symbols in the
2 3
176 bar and 250 uC using 8% Ni/c-Al O catalyst after the third time of
figure represent 1-butanol ( ), 1-hexanol ( ), butanal ( ), acetaldehyde (%),
regeneration. The symbols in the figure represent 1-butanol ( ), 1-hexanol ( ),
2
-pentanone (#), and ethyl acetate ( ).
butanal ( ), 2-pentanone (#), and ethyl acetate ( ).
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 4271–4280 | 4277

View Article Online
Paper
RSC Advances
2
1
could also be involved in the process, mostly in the formation
of the minor side products.
min
hydrogen for 8 h, the catalyst could still retain its
activity up to 9 h on stream in the reaction, as displayed in
Fig. 5, which shows its promise in future applications.
Control tests using no nickel catalyst but fresh c-alumina
were also performed. Negligibly small conversion of ethanol to
diethyl ether, acetaldehyde, and tetrahydrofuran was observed.
While the c-alumina used was white at the start of the test, it
was fully dark brown when the reaction was finished after
about 6 h on stream, implying some coke deposition occurred
during the course of the control tests.
3
.5. Catalyst characterizations
Powder XRD analyses were performed on the c-alumina
support and the fresh/spent c-alumina supported Ni catalysts.
The obtained XRD patterns are displayed in Fig. 6. Compared
with the XRD spectra for the c-alumina support, all NiAl
catalysts showed common XRD lines corresponding to
metallic nickel (Ni) as well as c-Al
the reduction temperature of 500 uC used for the catalyst
reduction with H was sufficient to reduce the nickel species to
3.4. Stability of the catalysts
2 3
O . This evidenced that
To study the stability of the catalyst, the product yields in the
reaction were monitored for 13 h on stream at 176 bar and 250
2
uC using 8% Ni/c-Al
2
O
3
catalyst. As shown in Fig. 4, all yields
its metallic form. Moreover, the XRD patterns of spent
catalysts show that Ni metal was not converted to its oxide
forms within the course of reaction likely due to the presence
remained almost constant up to 13 h on stream after the
reaction was started on Day 1. The catalyst was kept in ethanol
in the reactor system and the reaction was resumed on Day 2
for another 5 h before completion. As shown in Fig. 4, the
catalyst maintains its activity with similar yields and selectivity
as those obtained in the Day 1 test. Therefore, the catalyst in
this study demonstrated good stability, remaining active for 18
h on stream. The fluctuation of the results as shown in the
graph was mostly due to the fluctuation in the flow rate of the
ethanol fed.
of reducing atmosphere during the reaction (H , CH and
2 4
C
2
H
6
, etc.).
TGA of the spent catalysts were carried out by temperature-
2
1
programmed oxidation under an air flow of 30 mL min
,
2
1
from 25 uC to 900 uC at a ramp rate of 20 uC min . The TGA
plots of the spent Ni/c-alumina catalysts with various Ni
loadings as well as the spent c-alumina used in ethanol
conversion reactions at 250 uC and 176 bar are displayed in
Fig. 7. The weight loss of all different spent catalysts after
being on-stream for about 10–18 h were found less than 8.2
wt%. Since, the carbon adsorbed on the surface of the catalyst
can be oxidized at the air flow, these data suggest that the
degree of coking (carbon deposition) on the NiAl catalysts
The same catalyst was in situ regenerated at 500 uC with 100
2
1
mL min hydrogen for 4 h. After the in situ regeneration, the
reaction was performed and negligible changes in the
products yields and selectivity were observed. After regenera-
tion of the catalyst for the 3rd time at 500 uC with 100 mL
Fig. 6 XRD patterns of the c-alumina support and the fresh/spent c-alumina supported Ni catalysts (NiAl).
4
278 | RSC Adv., 2013, 3, 4271–4280
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
4. Conclusions
This study developed a one-step continuous process operating
at 150–300 uC for conversion of sub-/supercritical ethanol into
butanol and 1-hexanol as chemicals or fuels, as well as some
other compounds that are also useful as biogasoline. The
process distinguishes itself from the ones in the literature as it
employs sub- and supercritical bio-ethanol as the reaction
medium and c-alumina supported Ni catalysts at an optimal
metal loading. With the continuous flow fixed-bed tubular
2
1
reactor and 8% Ni/c-Al
uC and 176 bar, the highest ethanol conversion of 35%,
-butanol selectivity of 61.7%, and 1-butanol yield of 21.6%, as
2 3
O catalyst at a WHSV of 6.4 h at 250
1
well as 1-hexanol yield of y5% with a selectivity of 20% were
obtained.
Fig. 7 TGA plots of the spent Ni/c-alumina catalysts with various Ni loadings as
well as the spent c-alumina used in supercritical ethanol conversion reactions at
The ethanol conversion and products selectivity were
strongly dependent on the operating temperature/pressure
and the nickel loading in the catalysts. Generally a higher
temperature/pressure favoured the conversion and selectivity,
while as the temperature was above 300 uC, more unwanted
products and mostly gases were produced. It was demon-
2
1
2
50 uC and 176 bar. The catalyst samples were heated in air at 30 mL min
,
2
1
from 25 uC to 900 uC at a ramp rate of 20 uC min
.
from this process was not severe, being in a good agreement
with their high stability as described in Fig. 4 and 5.
2 3
strated that 8% Ni/c-Al O is the most active catalyst among all
catalysts tested including the Ni/c-Al O catalysts with 8–27
BET surface area were also measured on the fresh/spent
c-alumina used in the control tests, and the NiAl catalysts with
various nickel loadings, as shown in Table 4. The fresh
c-alumina as the support material has the highest surface area
2 3
wt% Ni loading, and mixed catalysts of (Mn
Ni/c-Al ) and (neat Mn + 27% Ni/c-Al
c-Al catalyst in this study demonstrated outstanding
2
O
3
/c-Al
2 3
O + 27%
2
O
3
2
O
3
2 3
O ). The 8% Ni/
2 3
O
2
21
and the BET surface areas reduced from 128 to 111 m g as
expected for the fresh NiAl catalysts as nickel loading increases
from 8% to 27%. The BET surface area for the spent c-alumina
stability, remaining active for 18 h on stream. After regenera-
tion of the catalyst for multiple times at 500 uC with 100 mL
2
1
min hydrogen for 4–8 h, the catalyst could still maintain its
activity for ethanol conversion to 1-butanol and 1-hexanol.
2
21
used in the control test has the lowest surface area, 76 m g
,
which was likely due to collapse of the pore structures of the
catalyst during the reaction or the carbon deposition on the
catalyst, blocking the micropores of the alumina sample. The
carbon deposition/coking on the spent catalysts, not severe
though, may be evidenced in Fig. 7. Interestingly, the spent 8%
Acknowledgements
The authors are grateful for the funding support from NSERC
and CFI through the Discovery Grant and Leaders Opportunity
Fund, respectively, awarded to Dr Xu. One of the authors
(H.S.G.) would also like to acknowledge the IUT financial
support for his sabbatical and Western University for the
support of his research work at ICFAR.
2
21
Ni/c-alumina had an increased BET surface (143 m g ) than
2
21
the fresh one (128 m g ), which might be due to Ni leaching
during the course of the reaction in supercritical condition.
The metal leaching is a common challenge for heterogeneous
catalysts in liquid phase reaction systems. As such, Ni leaching
would occur in the process, while interestingly nickel leaching
did not negatively affect the catalyst’s activity as the catalyst
demonstrated activity for more than 18 h on stream and could
be regenerated.
References
1
2
3
S. Atsumi, T. Hanai and J. Liao, Nature, 2008, 451, 86–90.
K. R. Szulczyk, Int. Nat. J. Energy Environ., 2010, 1, 501–512.
Butanol fuel, Wikipedia, http://en.wikipedia.org/wiki/
Butanol_fuel, downloaded on Aug. 8, 2012.
4
G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106,
Table 4 BET surface areas of various fresh/spent catalysts
4044–4098.
2
21
5 E. Hemo, R. Virduk, M. V. Landau and M. Herskowitz,
Catalyst samples
BET surface area (m g
)
Chem. Eng. Transact., 2010, 21, 1243–1248.
Fresh c-Al
Fresh 8% Ni/c-Al
Fresh 17% Ni/c-Al
Fresh 27% Ni/c-Al
2
O
3
152
128
119
111
123
76
6
7
A. J. O’Lenick Jr., J. Surfactants Deterg., 2001, 4, 311–315.
M. C. S ´a nchez-S ´a nchez, R. M. Navarro and J. L. G. Fierro,
Catal. Today, 2007, 129, 336–345.
8 T. Tsuchida, J. Kubo, T. Yoshioka, S. Sakuma, T. Takeguchi
and W. Ueda, J. Catal., 2008, 259, 183–189.
2
O
3
2
O
3
3
2
O
Fresh Mn
2
O
3
powder and 27% Ni/c-Al
2
O
3
Spent c-alumina
Spent 8% Ni/c-alumina
143
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 4271–4280 | 4279

View Article Online
Paper
RSC Advances
9
A. S. Ndou, N. Plint and N. J. Coville, Appl. Catal., A, 2003,
51, 337–345.
0 K. W. Yang, X. Z. Jiang and W. C. Zhang, Chin. Chem. Lett.,
004, 15, 1497–1500.
1 T. Tsuchida, S. Sakuma, T. Yoshioka and J. Kubo, EP 2 080
49 A1, 2009.
2 F. Alonso, P. Riente, J. A. Sirvent and M. Yus, Appl. Catal., A,
010, 378, 42–51.
14 F. Alonso, F. Foubelo, J. Gonzalez-Gomez, R. Martinez,
D. Ramon, P. Riente and M. Yus, Mol. Diversity, 2009, 14,
411–424.
15 J. Lu, E. C. Boughner, C. L. Liotta and C. A. Eckert, Fluid
Phase Equilib., 2002, 198, 37–49.
16 H. E. Dillon and S. G. Penoncello, Int. J. Thermophys., 2004,
25, 321–335.
17 G. Solomons and C. Fryhle, Organic Chemistry, 7th edn,
John Wiley & Sons, Inc., NY, USA, pp. 776–794.
18 A. Onda, S. Ogo and K. Yanagisawa, US 0190553 A1, 2011.
2
1
1
1
1
2
7
2
3 F. Engelhardt and W. Schritt, Method of Producing Beta-
branched Aldehydes, US 3558716, 1971.
4
280 | RSC Adv., 2013, 3, 4271–4280
This journal is ß The Royal Society of Chemistry 2013