A study of ZnxZryOz mixed oxides for direct conversion of ethanol to isobutene

Article abstract of DOI:10.1016/j.apcata.2013.07.011

Zn_xZr_yO_z mixed oxides were studied for direct conversion of ethanol to isobutene. Reaction conditions (temperature, residence time, ethanol molar fraction, steam to carbon ratio), catalyst composition, and pretreatment conditions were investigated, aiming at high-yield production of isobutene under industrially relevant conditions. An isobutene yield of 79% was achieved with an ethanol molar fraction of 8.3% at 475 C on fresh Zn₁Zr₈O₁₇ catalysts. Further durability and regeneration tests revealed that the catalyst exhibited very slow deactivation via coking formation with isobutene yield maintained above 75% for more than 10 h time-on-stream. More importantly, the catalysts activity in terms of isobutene yield can be readily recovered after in situ calcination in air at 550 C for 2.5 h. XRD, TPO, IR analysis of adsorbed pyridine (IR-Py), and nitrogen sorption have been used to characterize the surface physical/chemical properties to correlate the structure and performance of the catalysts.

Full text of DOI:10.1016/j.apcata.2013.07.011

Applied Catalysis A: General 467 (2013) 91–97
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
A study of Zn Zr O mixed oxides for direct conversion of ethanol
x
y
z
to isobutene
a
a,∗
a
a,b,∗∗
Changjun Liu , Junming Sun , Colin Smith , Yong Wang
a
The Gene and Linda Voiland School of Chemical Engineering and Bioengineering, Washington State University, Pullman, WA 99164, USA
Institute for Integrated Catalysis, Pacific Northwest National Laboratory, Richland, WA 99352, USA
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Zn Zr O mixed oxides were studied for direct conversion of ethanol to isobutene. Reaction conditions
x
y
z
Received 28 April 2013
Received in revised form 26 June 2013
Accepted 6 July 2013
(temperature, residence time, ethanol molar fraction, steam to carbon ratio), catalyst composition, and
pretreatment conditions were investigated, aiming at high-yield production of isobutene under industri-
ally relevant conditions. An isobutene yield of 79% was achieved with an ethanol molar fraction of 8.3% at
Available online xxx
◦
4
75 C on fresh Zn1Zr8O17 catalysts. Further durability and regeneration tests revealed that the catalyst
exhibited very slow deactivation via coking formation with isobutene yield maintained above 75% for
more than 10 h time-on-stream. More importantly, the catalysts activity in terms of isobutene yield can
be readily recovered after in situ calcination in air at 550 C for 2.5 h. XRD, TPO, IR analysis of adsorbed
pyridine (IR-Py), and nitrogen sorption have been used to characterize the surface physical/chemical
properties to correlate the structure and performance of the catalysts.
Keywords:
Isobutene
Bio-ethanol
Mixed oxide
Acid–base
◦
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
in terms of feedstock supply. However, the catalyst performance
under industrially viable conditions has yet to be examined. In this
work, the effects of reaction temperature, residence time, steam
to carbon ratio and ethanol molar fraction on isobutene yield have
been studied. The catalyst composition has also been further opti-
mized. An isobutene yield of 79% was achieved with an ethanol
With environment deterioration from global warming and the
gradual depletion of fossil resources, the importance of shifting
the modern chemical industry into a bio-based one has received
more and more attention. Biomass has been considered one of
the more promising sustainable natural resources with the poten-
tial to replace petroleum in production of chemicals and liquid
transportation fuels [1,2]. Isobutene has found wide applications
in the production of fuel additives (e.g., MTBE, ETBE and isooc-
tane), butyl polymers, antioxidants, tert-butylamine, tert-butanol
and et al. [2]. Isobutene is mainly produced by cracking crude oil
◦
molar fraction of 8.3% at 475 C on the optimized ZnxZryOz catalyst
of Zn:Zr = 1:8. Further durability and regeneration tests revealed
that the catalyst shows very slow deactivation via coke formation
with isobutene yield maintained above 75% for more than 10 h, and
that the catalysts activity can be fully recovered after calcination
◦
in air at 550 C for 2.5 h. Characterizations such as XRD, TG-TPO,
(
>10 million metric tons per year). The large demand and wide
IR-Py, and nitrogen sorption techniques have been employed to
correlate surface physical/chemical properties to the performance
of the catalysts.
application of isobutene make it urgent to seek alternative produc-
tion techniques based on biomass. Recently, we reported a tandem
reaction from bioethanol to isobutene with high yield over nano-
sized ZnxZryOz mixed oxide [3]. Later on, Iwamoto et al. [4,5] also
reported appreciable amounts of isobutene formed over modified
In O oxide in ethanol to propylene reactions. From an application
2. Experimental details
2
3
2.1. Catalyst preparation
point of view, the increased availability and reduced cost of biomass
derived ethanol [6] provides potential for scaling up this technique
ZnxZryOz mixed oxides were synthesized using a hard tem-
plate method [3,7]. Briefly, a calculated amount of zirconyl nitrate
hydrate (Sigma–Aldrich, >99.8%) and zinc nitrate hexahydrate
(Sigma–Aldrich, >99.8%) were dissolved in a given amount of DI
water and sonicated for 15 min to achieve a clear solution. Then
about 25 g of the resulted solution were mixed with 6.0 g of BP2000
∗
∗
Corresponding author. Tel.: +1 509335 1880; fax: +1 509335 4806.
Corresponding author at: The Gene and Linda Voiland School of Chemical Engi-
∗
neering and Bioengineering, Washington State University, Pullman, WA 99164, USA.
Tel.: +1 509335 1880; fax: +1 509335 4806.
(
black pearl 2000, Cabot Corp.) by incipient wetness impregna-
E-mail addresses: Junming.sun@wsu.edu (J. Sun), Yong.Wang@pnnl.gov
◦
(
Y. Wang).
tion. The BP 2000 was dried at 180 C overnight prior to use. The
0
926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.07.011

9
2
C. Liu et al. / Applied Catalysis A: General 467 (2013) 91–97
Table 1
resulted mixture was dried at room temperature overnight and at
◦ ◦
0 C for 4 h. The sample was then calcined at 400 C for 4 h (ramp-
BET surface area of Zn1Zr10O21 at different calcination temperatures.
8
◦
−1
◦
◦
ing rate, 3 C min ), followed by 550 C for another 20 h (ramping
Calcination temperature ( C)
500
153.6
525
162
550
148.3
575
130
600
76.01
◦
−1
Surface area (M2 g−1)
rate, 3 C min ) in air. The obtained white powders were denoted
as ZnxZryOz, where x:y:z represents the corresponding molar ratio
of each component.
of ethanol/H O solution and nitrogen gas as well as the catalysts
2
amount were varied to achieve different ethanol molar fraction
2
.2. Characterization
(
1
x
) and residence times. The effluent stream was kept above
ethanol
◦
50 C to avoid the condensation of condensable species and sam-
2.2.1. Nitrogen adsorption–desporption isotherms
pled periodically to a Shimadzu 2014 Gas Chromatography (GC)
equipped with an auto sampling valve, HP-Plot Q column (30 m,
◦
Nitrogen sorption experiments were performed at −196 C
with an automatic adsorption meter (Quantachrome instruments
Autosorb-6). The samples were pretreated at 150 C for 6 h under
vacuum. The surface areas were calculated from adsorption val-
ues for five relative pressures (P/P ) ranging from 0.10 to 0.31
using the Brunauer–Emmett–Teller (BET) method. The pore vol-
umes were determined from the total amount of N2 adsorbed
0
.53 mm, 40 ␮m), and Flame Ionization Detector (FID) for con-
◦
densable oxygenates and small alkane/alkene analysis (e.g., C –C4
1
alkane/alkenes, ethanol, acetaldehyde, acetone, acetic acid, etc.).
The condensable species in the effluent were then collected by a
cold trap and the dry gas (e.g., CH , ethylene, CO , CO and hydro-
0
4
2
gen, etc.) was sent to an online micro GC (MicroGC 3000A equipped
with molecular sieves 5A, plot U columns and Thermal Conductivity
Detectors (TCD)) for analysis. Nitrogen also served as the reference
gas. The carbon balance in the current study is more than 95% based
on the moles of ethanol fed into reactor and moles of carbon con-
taining products quantified by GCs. The proposed reaction pathway
proceeds via ethanol dehydrogenation to acetaldehyde (Eq. (1)),
acetaldehyde to acetone (Eq. (2)) via condensation and decom-
position [8–10], and acetone to isobutene (Eq. (3)) via a complex
condensation and dehydration, as well as decomposition of mesityl
oxide [11] or mesityl oxide like surface species [12,13]. Ethanol con-
version, the product selectivity, and isobutene yield were defined as
Eqs. (5)–(8), respectively. Theoretic isobutene yield is 66.7% based
on the overall reaction (Eq. (4)) [3].
between P/P = 0.05 and P/P = 0.98.
0
0
2.2.2. X-ray diffraction (XRD)
X-ray powder diffraction spectra were recorded using a Philips
X pert diffractometer with a copper anode (K˛ = 0.15405 nm)
ꢀ
1
operating at 40 kV and 50 mA. The scanning rates and accumulation
◦
−1
time were 0.040 s and 1.6 s per step, respectively.
2.2.3. Infrared analysis of adsorbed pyridine (IR-Py)
IR-Py spectra were recorded on a Bruker Tenser 27 FTIR spec-
trometer. Details on the equipment and experiment can be found
elsewhere [3]. Briefly, 20 mg of the catalyst was packed into an
◦
in situ cell and activated at 450 C for 1 h, then cooled down to
5
◦
−1
0 C under flowing He (50 STP mL min ). Spectra were recorded
as background of all IR spectra thereafter. Pyridine was introduced
at 50 C by flowing He (5 STP mL min ) through a bubbler for ca
(1)
◦
−1
1
0 min during which the surface will be saturated as verified by
◦
IR [3]. The spectra at 50 C were then recorded following a 30 min
purge with flowing He. The sample was then ramped to a given
temperature (10 C min ), and purged with flowing He for 30 min,
◦
−1
◦
then cooled down to 50 C for spectra collection.
(2)
(3)
2
.2.4. Thermogravimetric-temperature programmed oxidation
(
TG-TPO) analysis
TG-TPO analysis was performed on a Netsch STA 409 CD
thermal analyzer with a TASC 414/3 controller. Air flow rate of
(4)
Ethanol in − Ethanol out
Ethanol conversion =
Product selectivity =
× 100%
(5)
(6)
Ethanol in
−
1
1
5 STP mL min was adapted to perform the TG-TPO. About 20 mg
◦
of sample was loaded in a platinum crucible which was held at 40 C
Carbon in given product
× 100%
◦
◦
−1
.
for 20 min and then ramped to 800 C at 10 C min
.3. Catalytic activity test
The conversions of bio-ethanol to isobutene (ETIB) were con-
Carbon in all products
carbon in isobutene
carbon in Ethanol in
2
Actual isobutene yield =
× 100%
(7)
(8)
Actual isobutene yield
ducted in a fixed-bed stainless steel reactor (i.d. 5 mm), which
has been detailed elsewhere [3]. Briefly, a given amount of
ZnxZryOz catalyst was loaded with a K-type thermocouple placed
in the center of the catalyst bed to monitor the reaction tem-
Isobutene yield =
× 100%
Theoretic isobutene yield
3. Results and discussion
perature. Before the reaction catalysts were pre-treated in N
2
−1
◦
◦
−1
(
50 STP mL min ) at 450 C (ramping rate was 5 C min ) for 0.5 h;
3.1. Physical properties
then the ethanol/H O solution of given steam to carbon ratio was
2
◦
introduced into the evaporator (180 C) by a syringe pump and
XRD patterns (Fig. 1) of the ZnxZryOz catalysts show typical
carried into the reactor by flowing nitrogen gas. The flow rates
diffraction peaks, characteristics of tetragonal ZrO . No ZnO phase
2

C. Liu et al. / Applied Catalysis A: General 467 (2013) 91–97
93
Fig. 1. XRD patterns of ZnxZryOz.
was observed which indicates that ZnO was highly dispersed. The
nitrogen sorption results showed that the surface areas of ZnxZryOz
2
−1
mixed oxides are between 76 and 162 m g (Table 1) depending
the calcination temperature employed.
3.2. Surface acidity
IR analysis of adsorbed pyridine (IR-Py) has been widely used
to identify the nature and strength of catalysts’ acidity. In this
work different calcination temperatures have been studied in cat-
alysts preparation to optimize the catalyst performance in terms of
isobutene yield. IR-Py was employed to identify the surface active
sites for isobutene formation. Fig. 2 shows the IR spectra of differ-
ent desorption temperatures over the catalysts calcined at different
calcination temperatures. All the catalysts exhibit both Lewis and
−
1
Brønsted acid sites. The peaks at 1611, 1576, 1450 cm
can be
attributed to pyridine adsorbed on Lewis acid sites, the peak at
−
1
1
493 cm is contributed to both Lewis and Brønsted acid sites, and
−1
the peak at 1541 cm can be assigned to the pyridine adsorbed on
Brønsted acid sites. Upon increasing the desorption temperature,
the peak areas corresponding to both Lewis and Brønsted acid sites
decrease, which can be used to identify the amount of acidic sites
−
1
of various strength [14–18]. Herein, peaks at 1611 and 1541 cm
were integrated to quantitatively compare the amounts of Lewis
and Brønsted acid sites, respectively (Fig. 3). The area difference
◦
between 250 and 350 C was referred to as the amount of weak
◦
acidic sites, that between 350 and 450 C was medium, and those
◦
above 450 C strong acidic sites. From Fig. 3a, it is observed that the
total Brønsted acid sites decreases significantly with the increase
of calcination temperature. Interestingly, strong Brønsted acid sites
only show a slight decrease, while the medium Brønsted acid sites
decreases significantly with calcination temperature. Almost no
medium Brønsted acid sites were observed at calcination temper-
◦
Fig. 2. IR-Py spectra of Zn1Zr10O21 at different desorption temperature (a) calcined
at 500 C; (b) calcined at 550 C and (c) calcined at 600 C.
ature of 600 C. However, a maximum of weak Brønsted acid sites
◦
◦
◦
◦
were observed at calcination temperature of 550 C. Brønsted acid
sites were known to be very sensitive to dehydration [19] indicating
◦
◦
that increasing the calcination temperature from 500 C to 600 C
could lead to different degree of dehydration of surface hydroxyl
species and result in changes to abundance and strength of the
3
3
.3. Catalysts activity
.3.1. Effect of reaction temperatures
We have shown that reaction temperature has a significant
◦
Brønsted acid sites. Calcination at 600 C leads to the least Brønsted
acid sites. Similar results were observed at different calcination
times (Fig. 4). In terms of Lewis acid sites (Fig. 3b) both medium and
strong Lewis acid sites decrease with increased calcination temper-
ature while weak Lewis acid sites increase slightly. Total Lewis acid
sites decrease slightly with an increase in calcination temperature.
effect on the products distribution in ETIB [3]. Herein, we further
investigated the effect of reaction temperature on the products dis-
tribution at a higher ethanol molar fraction (3.6%) over Zn Zr10O21
1
catalyst as shown in Fig. 5. Although the ethanol conversions
were high in the whole temperature range tested, the products

9
4
C. Liu et al. / Applied Catalysis A: General 467 (2013) 91–97
◦
Fig. 5. Effect of reaction temperature (S/C = 5). Zn1Zr10O21: 100 mg, TCal = 500 C,
−
1
W/F = 0.225 g s STP mL , xethanol = 3.6%, S/C = 5.
distribution is very sensitive to the reaction temperature. Acetone
intermediate [3] was found as the main product at lower temper-
ature range and its selectivity decreased rapidly as the reaction
temperature increased. Meanwhile, the isobutene yield increased
accordingly. Interestingly, we did not observe the formation of
other byproducts, which is very important in practical application.
◦
When increasing the reaction temperature above 500 C, acetone
selectivity further decreases. However, isobutene yield begins to
decline with the concurrent increase of selectivity to methane
and CO2 suggesting that ethanol/acetone decomposition is facile
at higher temperatures. Nakajima et al. [20] found that the conver-
sion of ethanol to acetone is highly thermally favorable and ethanol
can be readily converted into acetone over a mixed ZnO–Fe O3
2
catalyst. Our thermodynamic calculation for Eq. (3) at atmosphere
pressure (Fig. 6) suggests that acetone to isobutene conversion is
also highly favorable with equilibrium acetone conversion being
9
9.8% under current reaction conditions and even more at lower
reaction temperature. This indicates that the ETIB reaction on the
Fig. 3. Acidity distribution of Zn1Zr10O21 calcined at various temperatures. (a)
Brønsted acidity and (b) Lewis acidity.
Zn Zr10O21 is kinetically limited by acetone to isobutene conver-
1
◦
sion. Interestingly, at 400 C, 1,3-butadiene was observed to be the
major C products (78% in C products) with isobutene only making
4
4
up about 22% of C products. In addition, acetaldehyde was also pro-
4
◦
duced with high selectivity at 400 C. Only isobutene was produced
as C4 a product at reaction temperatures above 425 C where no
◦
acetaldehyde was observed. This indicates that 1,3-butadiene and
Fig. 4. Brønsted acidity distribution of Zn1Zr10O21 calcined at various time. Notes:
◦
TCal = 550 C.
Fig. 6. Thermal equilibrium constant of acetone to isobutene.

C. Liu et al. / Applied Catalysis A: General 467 (2013) 91–97
95
◦
Fig. 7. Effect of steam to carbon ratio. Zn1Zr10O21: 100 mg, TCal = 500 C,
−1
xethanol = 1.8%, T = 450 C, i-C4H8* represents isobutene
◦
◦
Cal
= 500 C,
W/F = 0.225 g s STP mL
,
Fig. 8. Effect of ethanol molar fraction. Zn Zr
O
:
100 mg,
T
1
10 21
yield.
W/F = 0.225 g s STP mL−1, T = 450 ◦C, S/C = 2.5, i-C H * represents isobutene yield.
4
8
isobutene were produced via different reaction pathways at dif-
ferent reaction temperatures. Isobutene has been confirmed to
be produced from acetone [3]. However, the formation of 1,3-
butadiene must be related to the acetaldehyde, which has been
reported on a variety of oxides catalysts with mixed acid–base
properties [21].
interest. With S/C = 2.5 and a fixed residence time, as shown in Fig. 8,
isobutene yield decreased with increasing ethanol molar fraction
while acetone selectivity increased inversely. When the ethanol
molar fraction increased from 0.6% to 11.9% the isobutene yield
dropped from 85.4% to 8.2% and the acetone selectivity increased
from 4.3% to 63.6%. The selectivity to ethylene and propene was
almost constant. It suggests that either isobutene or acetone can
be produced with high selectivity over ZnxZryOz from ethanol by
properly controlling the reaction conditions, and that acetone was
more favorable under high ethanol molar fraction.
3.3.2. Effect of steam to carbon ratio
As noted above, water has been involved in the multi-step reac-
tions from ethanol to isobutene including the ethanol dehydration
as well as the acetaldehyde and acetone condensations. The effect
of water vapor pressure on the isobutene yield was further inves-
3.5. Effect of residence time
tigated on the Zn Zr10O21 catalyst via varying the steam to carbon
1
ratios while keeping the ethanol molar fraction and other reaction
conditions constant. As shown in Fig. 7, when a lower steam to car-
bon ratio was used, isobutene yield was only 12.5%. Selectivities to
ethylene and acetone reach 17.1% and 46.6%, respectively. When
S/C ratio increased from 0.5 to 2.5 the isobutene yield jumped from
As discussed above, the ETIB conversion is kinetically limited
by acetone conversion to isobutene on the Zn Zr10O catalyst.
1
21
Higher isobutene yield will be expected at longer residence time.
Indeed, isobutene yield increased significantly from 30.6% to 72.6%
−
1
when the residence time increased from 0.225 g s STP mL
to
.04 g s STP mL , meanwhile the selectivity to acetone dropped
from 45.1% to 13.3% (Fig. 9). By further increasing the residence
1
2.5% to 69.2%. Meanwhile the selectivity to acetone decreased
−
1
1
from 46.6% to 21.1% and ethylene selectivity also decreased sig-
nificantly from 17.1% to 3.9%. After further increasing the S/C ratio
to 5, isobutene yield dropped back to 53.5% along with an increase
in the selectivity to acetone to 31.7% while ethylene selectivity fur-
ther decreased to 3.5%. It is clear that water plays a key role in
suppressing the dehydration of ethanol, which in turn promotes
the ethanol dehydrogenation and condensation pathway toward
isobutene. A suitable amount of water (i.e., S/C = 2.5 in the current
case) is pivotal to suppress the direct ethanol dehydration to eth-
ylene. Nakajima et al. [8] observed substantial acceleration of the
conversion of acetaldehyde to acetone on ZnO–CaO by increasing
steam to carbon ratio from 0.42 to 2.5 and suggested that the surface
hydroxide ion formed from dissociative adsorption of water partic-
ipated in the reaction. Although the detailed mechanism regarding
acetone to isobutene conversion is not clear, the surface hydrox-
ide and hydroxyl could also act as base and acid sites that aided
the adsorption and activation of acetone and thus enhanced the
isobutene production to a certain degree. On the other hand, water
also could be competitively adsorbed onto the active sites thus too
much water would saturate the surface and negatively affect the
activity.
−
1
time to 2.0 g s STP mL the isobutene yield of 79.4% was achieved
as the selectivity to acetone dropped to 5.5% accordingly. This fur-
ther confirmed that ETIB is kinetically limited by the acetone to
3.4. Effect of ethanol molar fraction
The effect of ethanol molar fraction on ETIB was further inves-
tigated to explore the reaction conditions relevant to industry
◦
Fig. 9. Effect of residence time (S/C = 2.5). Zn1Zr10O21: 200 mg, TCal = 500 C,
xethanol = 3.6%, T = 450 C, S/C = 2.5, i-C4H8 represents isobutene yield.
◦
*

9
6
C. Liu et al. / Applied Catalysis A: General 467 (2013) 91–97
◦
Fig. 12. Optimization of calcination time. Zn1Zr10O21: 100 mg, TCal = 550 C,
◦
Fig. 10. Optimization of catalyst composition. ZnxZryOz: 100 mg, TCal = 500 C,
xethanol = 0.6%, S/C = 5, W/F = 0.225 g s STP mL
isobutene yield.
−
1
◦
*
−
1
T = 450 ◦C, i-C4H8* represents
,
xethanol = 0.6%, S/C = 5, W/F = 0.225 g s STP mL
,
T = 450 C, i-C4H8 represents
isobutene yield.
isobutene conversion. In addition, with increasing residence time
the selectivity to methane also slightly increased (from 1.3% to 5.7%)
which indicated that ethanol or acetone decomposition is favored
at longer residence times.
isobutene yield dropped from 79% to 61%. Significantly, after cal-
cination at 550 C in air the activity of the spent catalyst could
◦
be readily recovered. The isobutene yield was even higher (∼82%)
and the deactivation of the regenerated catalyst was slower than
the fresh catalysts. TG-TPO analysis of the spent catalyst reveals a
◦
3.6. Catalyst optimization
weight loss accompanied by an exothermal process between 280 C
◦
and 480 C (Fig. 15) suggesting that coking was the primary reason
The acid base properties of the catalyst were critical for the
that resulted in the slow catalyst deactivation.
direct ethanol to isobutene conversion. The composition and prepa-
ration conditions were further investigated to improve the catalyst
performance. The catalyst composition of zinc to zirconium ratio
was found to be a key factor influencing the isobutene yield [3].
Our further investigation of the zinc to zirconium ratios (Fig. 10) as
well as calcination temperature and times (Figs. 11 and 12) reveal
3.7. Catalyst stability and regeneration
Acetone to isobutene reactions have been widely studied on
zeolite based catalysts [13,22–24]. The reaction pathway has been
proposed by Dolej sˇ ek et al. [12]. First, acetone is converted to diace-
tone alcohol (DAA) via aldolization which then readily dehydrated
to mesityl oxide (MSO). Finally, adsorbed MSO decomposed into
isobutene and surface acetate [12,24]. Brønsted acidic sites have
been shown to play a key role in this conversion [12,13,24]. A recent
work on Metal-Exchanged ␤ Zeolites further revealed that weak
Brønsted acid sites were more likely favorable [25] and that strong
acid sites are prone to coke formation [26]. In this work, our differ-
ent calcination temperature results show that total Brønsted and
◦
that Zn Zr O calcined at 550 C for 20 h in static air gives the best
1
8
17
performance in terms of isobutene yield under a higher ethanol
◦
molar fraction (8.3% ethanol), at 475 C reaction temperature, and
residence time of 1.98 g s STP mL
−1
(i.e., ∼79%, Fig. 13). Notably,
with a long-term time-on-stream test it showed that (Fig. 13)
the isobutene yield decreased slowly with the concurrent increase
of acetone selectivity indicating that the decrease of isobutene
yield is due to the loss of active sites which catalyzed the acetone
to isobutene conversion. After 28 h of time-on-stream operation,
Fig. 13. Activity and stability of fresh and regenerated catalyst. Zn1Zr8O17: 800 mg,
◦
−1
◦
TCal = 550 C, xethanol = 8.3%, S/C = 2.5, W/F = 1.92 g s STP mL , T = 475 C. Catalyst was
◦
Fig. 11. Optimization of calcination temperature. Zn1Zr10O21: 100 mg, xethanol = 0.6%,
regenerated by calcination at 550 C in box furnace in static air for 5 h. i-C4H8*
S/C = 5, W/F = 0.225 g s STP mL , T = 450 C, i-C4H8* represents isobutene yield.
−
1
◦
represents isobutene yield.

C. Liu et al. / Applied Catalysis A: General 467 (2013) 91–97
97
Additionally, the propylene selectivity dropped to be ∼1% with
isobutene selectivity closing to its theoretical value (i.e., 88.9%, Eq.
(3)) suggesting the formation of propylene is from both hydrogena-
tion mechanism proposed above [3,4].
4. Conclusion
We have studied the direct conversion of ethanol to isobutene
on the ZnxZryOz mixed oxides under industrially relevant condi-
◦
tions. The catalyst of Zn Zr O calcined at 550 C in air for 20 h
1
8
17
exhibited an isobutene yield as high as 79% with an ethanol molar
◦
fraction of 8.3% at 475 C. Weak Brønsted acid sites were found
to be the active sites for actone conversion to isobutene reac-
tions, while strong Brønsted acid sites are probably responsible for
the coke formation on catalysts that leads to slow catalyst deac-
tivation with time-on-stream. After regernation, we observed an
improvement of catalysts performance in terms of isobutene yield
and stability which could be correlated to the increased number of
weak Brønsted acidic sites that coincided with the suppression of
stronger Brønsted sites.
Fig. 14. Brønsted acidity distribution of fresh and regenerated catalyst.
Acknowledgements
We gratefully acknowledge the Archer Daniels Midland (ADM)
Company for the support of this work. We also acknowledge the
financial support from the US Department of Energy (DOE), Office of
Basic Energy Sciences, Division of Chemical Sciences, Geosciences,
and Biosciences. Part of the research described in this paper was
performed in the Environmental Molecular Sciences Laboratory
(EMSL), a national scientific user facility sponsored by the DOE’s
Office of Biological and Environmental Research and located at
Pacific Northwest National Laboratory (PNNL).
References
[
[
[
1] J.C. Serrano-Ruiz, R. Luque, A. Sepulveda-Escribano, Chem. Soc. Rev. 40 (2011)
266–5281.
2] B. van Leeuwen, A. van der Wulp, I. Duijnstee, A. van Maris, A. Straathof, Appl.
Microbiol. Biotechnol. 93 (2012) 1377–1387.
5
3] J. Sun, K. Zhu, F. Gao, C. Wang, J. Liu, C.H.F. Peden, Y. Wang, J. Am. Chem. Soc.
Fig. 15. TG-TPO profile of spent Zn1Zr8O17 after 28 h time-on-stream operations.
133 (2011) 11096–11099.
[
[
4] M. Iwamoto, S. Mizuno, M. Tanaka, Chem. – Eur. J. 19 (2013) 7214–7220.
5] S. Mizuno, M. Kurosawa, M. Tanaka, M. Iwamoto, Chem. Lett. 41 (2012)
Lewis acid sites decreases monotonically as temperature increased
892–894.
◦
◦
◦
from 500 C to 600 C (Fig. 3). However, catalyst calcined at 550 C
exhibited the best catalytic performance in terms of isobutene
yield. The only sites that correlate with the catalyst performance
are the weak Brønsted acid sites, which show a maximum at the
[6] G.A. Deluga, J.R. Salge, L.D. Schmidt, X.E. Verykios, Science 303 (2004) 993–997.
[
[
7] K.K. Zhu, J.M. Sun, H. Zhang, J. Liu, Y. Wang, J. Nat. Gas Chem. 21 (2012) 215–232.
8] T. Nakajima, K. Tanabe, T. Yamaguchi, I. Matsuzaki, S. Mishima, Appl. Catal. 52
(
1989) 237–248.
[9] T. Nishiguchi, T. Matsumoto, H. Kanai, K. Utani, Y. Matsumura, W.-J. Shen, S.
Imamura, Appl. Catal. A: Gen. 279 (2005) 273–277.
10] R.S. Murthy, P. Patnaik, P. Sidheswaran, M. Jayamani, J. Catal. 109 (1988)
◦
calcination temperature of 550 C indicating that weak Brønsted
[
acid sites might be responsible for the acetone to isobutene con-
298–302.
version on the ZnxZryOz catalysts. Further IR analysis of pyridine
[11] N. Daly, M. Gilligan, Aust. J. Chem. 24 (1971) 955–960.
[
[
[
12] Z. Dolej sˇ ek, J. Nováková, V. Bosá cˇ ek, L. Kubelková, Zeolites 11 (1991) 244–247.
13] G.J. Hutchings, P. Johnston, D.F. Lee, C.D. Williams, Catal. Lett. 21 (1993) 49–53.
14] R.W. Stevens Jr., S.S.C. Chuang, B.H. Davis, Thermochim. Acta 407 (2003) 61–71.
adsorbed on the fresh and regenerated Zn Zr O catalysts (Fig. 14)
1
8
17
also reveals that more weak Brønsted acid sites were observed
while strong Brønsted acid sites almost disappeared on the regen-
erated catalysts. This coincided with the slightly improved activity
and stability of the regenerated catalyst, further confirming that
weak Brønsted acid sites are the active sites for acetone conversion
and stronger Brønsted acid sites might result in coking and slow
deactivation of the catalysts.
It should be mentioned that other than the predominant for-
mation of isobutene small amounts of propylene have always been
observed, which can be due to the protonation of acetone over the
Brønsted acid sites followed by dehydration pathway [3]. In addi-
tion, Iwamoto et al. found that the presence of hydrogen would
also promote the hydrogenation of acetone toward propylene [4].
Indeed, hydrogen was formed during the ethanol to acetone con-
version and our separate experiments revealed that when ethanol
reactant was replaced with acetone, no hydrogen was detected.
[15] C.A. Emeis, J. Catal. 141 (1993) 347–354.
[
16] S. Damyanova, M.A. Centeno, L. Petrov, P. Grange, Spectrochim. Acta A: Mol.
Biomol. Spectrosc. 57 (2001) 2495–2501.
[
17] E. Selli, L. Forni, Microporous. Mesoporous. Mater. 31 (1999) 129–140.
[18] J. Datka, A.M. Turek, J.M. Jehng, I.E. Wachs, J. Catal. 135 (1992) 186–199.
[19] F. Babou, G. Coudurier, J.C. Vedrine, J. Catal. 152 (1995) 341–349.
[
[
[
[
20] T. Nakajima, H. Nameta, S. Mishima, I. Matsuzaki, K. Tanabe, J. Mater. Chem. 4
1994) 853–858.
21] E.V. Makshina, W. Janssens, B.F. Sels, P.A. Jacobs, Catal. Today 198 (2012)
338–344.
22] K. Zhu, J. Sun, J. Liu, L. Wang, H. Wan, J. Hu, Y. Wang, C.H.F. Peden, Z. Nie, ACS
Catal. 1 (2011) 682–690.
23] C.O. Veloso, J.L.F. Monteiro, E.F. Sousa-Aguiar, in: H.G.K.H.P.J. Weitkamp, W.
Hölderich (Eds.), Stud. Surf. Sci. Catal., Elsevier, Amsterdam, The Netherlands,
(
1994, pp. 1913–1920.
[
[
24] L. Kubelková, J. Cˇ ejka, J. Nováková, Zeolites 11 (1991) 48–53.
25] A.J. Cruz-Cabeza, D. Esquivel, C. Jiménez-Sanchidrián, F.J. Romero-Salguero,
Materials 5 (2012) 121–134.
[
26] L. Kubelková, J. Nováková, J. Mol. Catal. 75 (1992) 53–62.