Key Roles of Lewis Acid-Base Pairs on ZnxZryOz in Direct Ethanol/Acetone to Isobutene Conversion

Article abstract of DOI:10.1021/jacs.5b07401

The effects of surface acidity on the cascade ethanol-to-isobutene conversion were studied using Zn_xZr_yO_z catalysts. The ethanol-to-isobutene reaction was found to be limited by the secondary reaction of the key intermediate, acetone, namely the acetone-to-isobutene reaction. Although the catalysts with coexisting Br?nsted acidity could catalyze the rate-limiting acetone-to-isobutene reaction, the presence of Br?nsted acidity is also detrimental. First, secondary isobutene isomerization is favored, producing a mixture of butene isomers. Second, undesired polymerization and coke formation prevail, leading to rapid catalyst deactivation. Most importantly, both steady-state and kinetic reaction studies as well as FTIR analysis of adsorbed acetone-d₆ and D₂O unambiguously showed that a highly active and selective nature of balanced Lewis acid-base pairs was masked by the coexisting Br?nsted acidity in the aldolization and self-deoxygenation of acetone to isobutene. As a result, Zn_xZr_yO_z catalysts with only Lewis acid-base pairs were discovered, on which nearly a theoretical selectivity to isobutene (～88.9%) was successfully achieved, which has never been reported before. Moreover, the absence of Br?nsted acidity in such Zn_xZr_yO_z catalysts also eliminates the side isobutene isomerization and undesired polymerization/coke reactions, resulting in the production of high purity isobutene with significantly improved catalyst stability (<2% activity loss after 200 h time-on-stream). This work not only demonstrates a balanced Lewis acid-base pair for the highly active and selective cascade ethanol-to-isobutene reaction but also sheds light on the rational design of selective and robust acid-base catalyst for C-C coupling via aldolization reaction.

Full text of DOI:10.1021/jacs.5b07401

Subscriber access provided by The University of Liverpool
Article
Key Roles of Lewis Acid-base Pairs on ZnxZryOz
in Direct Ethanol/Acetone to Isobutene Conversion
Junming Sun, Rebecca A. L. Baylon, Changjun Liu, Donghai Mei,
Kevin J. Martin, Padmesh Venkitasubramanian, and Yong Wang
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b07401 • Publication Date (Web): 01 Dec 2015
Downloaded from http://pubs.acs.org on December 11, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Key Roles of Lewis Acidꢀbase Pairs on ZnxZryOz in
Direct Ethanol/Acetone to Isobutene Conversion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Junming Sun,† Rebecca A.L. Baylon, † Changjun Liu † Donghai Mei, ‡ Kevin J. Martin, §
Padmesh Venkitasubramanian,§ Yong Wang †‡*
† The Gene & 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)
§ Archer Daniels Midland Company, 1001 N Brush College Road, Decatur, IL 62521 (USA)
Corresponding author: yong.wang@pnnl.gov
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 36
1
2
3
4
Abstract
5
6
7
8
9
The effects of surface acidity on the cascade ethanolꢀtoꢀisobutene conversion were studied using
ZnxZryOz catalysts. The ethanolꢀtoꢀisobutene reaction was found to be limited by the secondary
reaction of the key intermediate, acetone, namely the acetoneꢀtoꢀisobutene reaction. Although the
catalysts with coexisting Brønsted acidity could catalyze the rateꢀlimiting acetoneꢀtoꢀisobutene
reaction, the presence of Brønsted acidity is also detrimental. First, secondary isobutene
isomerization is favored, producing a mixture of butene isomers. Second, undesired
polymerization and coke formation prevail, leading to rapid catalyst deactivation. Most
importantly, both steadyꢀstate and kinetic reaction studies as well as FTIR analysis of adsorbed
acetoneꢀd6 and D2O unambiguously showed that a highly active and selective nature of balanced
Lewis acidꢀbase pairs was masked by the coexisting Brønsted acidity in the aldolization and selfꢀ
deoxygenation of acetone to isobutene. As a result, ZnxZryOz catalysts with only Lewis acidꢀbase
pairs were discovered, on which nearly a theoretical selectivity to isobutene (~88.9%) was
successfully achieved, which has never been reported before. Moreover, the absence of Brønsted
acidity in such ZnxZryOz catalysts also eliminates the side isobutene isomerization and undesired
polymerization/coke reactions, resulting in the production of high purity isobutene with
significantly improved catalyst stability (< 2% activity loss after 200 h timeꢀonꢀstream). This
work not only demonstrates a balanced Lewis acidꢀbase pair for the highly active and selective
cascade ethanolꢀtoꢀisobutene reaction, but also sheds light on the rational design of selective and
robust acidꢀbase catalyst for CꢀC coupling via aldolization reaction.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 36
Journal of the American Chemical Society
1
2
3
4
1. Introduction
5
6
7
8
As one of the most important olefins, isobutene has found wide applications in the
production of industrial commodities such as butyl rubber and ethyl tertꢀbutyl ether (ETBE).1
Currently, the majority of isobutene is produced by steam cracking of naphtha. This process is
energy intensive and produces a mixture of different C4 isomers, and a further separation of IB
from the C4 isomers is required to obtain high purity isobutene. Together with dwindling fossil
fuel resources, it is desirable to explore sustainable feedstock to produce high purity isobutene.
Acetone aldolization and further selfꢀdeoxygenation has been widely studied to produce
isobutene on zeolite catalysts with Brønsted acidity at high temperatures (>350 °C).2ꢀ7 The
surface Brønsted acid site has been proposed to protonate acetone to form enol/enolate
intermediate, which then reacts with another acetone to produce diacetone alcohol via aldol
addition.8,9 Diacetone alcohol then dehydrates to form mesityl oxide that decomposes to
isobutene and acetic acid (Scheme 1).2 The main disadvantage of acetoneꢀtoꢀisobutene reaction
on zeolites is the strong Brønsted acid catalyzed secondary reaction of intermediates and
isobutene to aromatics, other hydrocarbons, and coke (Scheme 1).10 This leads to the rapid
deactivation of zeolite catalysts.3,11 Although alkaline metal ions have been used to tune activity
and improve stability,3 the isobutene yields are low (~55%), with notable deactivation within 10
h of timeꢀonꢀstream.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Lewis acidꢀbase pairs of (mixed) metal oxide have also been found highly active in the
aldolization of acetone,10,12ꢀ20 which mainly resulted in oxygenated adducts such as diacetone
alcohol, mesityl oxide, and phorone etc. Surface chemistry of acetone on Lewis acidꢀbase site
has been well understood, based largely on the IR probing of adsorption mode and
identification of reaction species at both low13 and high temperature.19,21ꢀ23 It is generally
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 36
1
2
3
4
5
6
7
8
9
accepted that the Lewis acidꢀbase pair plays a concerted role in the acetone aldolization.
Specifically, acetone is adsorbed on Lewis acid sites, while H abstraction from the αꢀmethyl
group occurs on the adjacent basic oxygen to form an enol/enolate intermediate.12 The
enol/enolate intermediate then reacts with another acetone to form mesityl oxide, trimeric
intermediates (e.g., phorone), and even aromatics via a complex condensation and cyclization
reactions at temperatures <200 °C (Scheme 1). At high temperature (>300 °C), however, a
parallel decomposition reaction is triggered by the oxidative surface basic hydroxyl groups and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
cus O2ꢀ, leading to the dominant decomposition products (i.e., methane and CO2) albeit minor
21
isobutene formation. Using in situ IR, Zaki et al.
studied surface reaction of acetone at
temperatures ranging from room temperature to 400 °C over a variety of metal oxides (e.g.,
TiO2, ZrO2 and CeO2). It was found that aldolization readily took place at temperatures <200
°C, while the oxidation ability of surface hydroxyl groups became significant at high
temperatures above 300 °C on Lewis acidꢀbase pair, leading to the dominant acetone
decomposition (Scheme 1).
4
ACS Paragon Plus Environment

Page 5 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1 Reported acetone reaction network on Brønsted acid and Lewis acidꢀbase pair.
We recently reported that ZnxZryOz mixed oxides synthesized using a hard template (denoted
as ZnxZryOzꢀH in this paper) have balanced surface acidꢀbase properties, which are capable of
catalyzing a cascade reaction from ethanolꢀtoꢀisobutene via acetaldehyde and acetone
intermediates at 450 °C.24,25 It was found that sulfur in the BPꢀ2000 hard template is
incorporated into the ZnxZryOzꢀH catalysts after removing the hard template via calcination,26
leading to the formation of strong Brønsted acid sites in addition to Lewis acid sites. ZnO
addition can suppress most of the strong Brønsted acid sites and introduce basic sites, shifting the
reaction pathway from ethanol dehydration to ethanol dehydrogenation and ketonization pathway
toward acetone, followed by a rateꢀlimiting acetoneꢀtoꢀisobutene reaction catalyzed by weak
Brønsted acid sites.24,25 Very recently, a similar ZnxZryOz catalyst with weak Brønsted acid sites
was also reported by Crisci et al. for a cascade acetic acid to isobutene reaction.27 Unfortunately,
ZnxZryOzꢀH catalysts with Brønsted acidity suffer from rapid deactivation by losing active sites
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 36
1
2
for the secondary acetoneꢀtoꢀisobutene reaction.24 Given the results previously reported on the
roles of Lewis acidꢀbase pair and Brønsted acid site in the acetoneꢀtoꢀisobutene reaction,13,20,25
we further studied the acid properties of ZnxZryOzꢀH catalysts. We particularly focus on the
elucidation of the roles of acid type on the secondary rateꢀdetermining acetone conversion in the
cascade ethanolꢀtoꢀisobutene reaction. It was found that both Brønsted acid sites and Lewis acidꢀ
base pairs are active for the acetone aldolization and selfꢀdeoxygenation to produce isobutene.
However, Brønsted acid site also catalyzes undesired isobutene isomerization reaction and coke
formation, leading to the catalyst deactivation. In contrast, balanced Lewis acidꢀbase pairs
exhibit high activity and stability in the selective acetoneꢀtoꢀisobutene conversion.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2. Experimental
2.1 Materials and catalyst synthesis
Zn(NO3)2. 6H2O (SigmaꢀAldrich, reagent grade, 98%), zirconium (IV) oxynitrate hydrate
(SigmaꢀAldrich, 99%), Zr(OH)4 (MEL, XZO631/01), BPꢀ2000 (carbon black pearl 2000, Cabot
Corp.), ZnO (SigmaꢀAldrich, >99.99%), ethanol (PharmcoꢀAAPER, 200 proof), acetone (JTꢀ
baker,, 99.8%), acetoneꢀd6 (SigmaꢀAldrich, 99.9%), pyridine (JTꢀbaker, 100%), and diacetone
alcohol (SigmaꢀAldrich, 99%) were used as purchased. Two different methods were used to
synthesize the ZnxZryOz catalysts, namely the hard template method and simple incipient wetness
impregnation without using hard template. The details for the hard template synthesis can be
found elsewhere.25,28 In this paper, all the catalysts synthesized via hard template method were
denoted as ZnxZryOzꢀH. For the incipient wetness impregnation method, the Zr(OH)4 was used
as support and initially dried overnight at 105 °C to remove any excess water on the surface
before impregnation. A Zn(NO3)2 solution was then added on Zr(OH)4 to achieve wet
impregnation. After impregnation, the catalysts were dried overnight at room temperature
6
ACS Paragon Plus Environment

Page 7 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
followed by 4 h at 105 °C. The catalysts were further heated to 400 °C (3 °C/min) and held for 2
h followed by a 5 °C/min ramp to the final calcination temperature (i.e., 550 °C) and held for 3 h.
The catalysts synthesized using the incipient wetness impregnation are denoted as ZnxZryOzꢀI.
High purity commercial ZnO was used directly. ZrO2 was prepared by calcining Zr(OH)4 and
using same temperature program for the ZnxZryOzꢀI catalysts. The details of the sample notation
and physicalꢀchemical properties are summarized in Table 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.2 Catalyst characterization
X-ray diffraction (XRD)
XRD patterns were collected on a Philips X’pert MPD (Model PW 3040/00) equipped with a
Cu Kα xꢀray source operating at 40 kV and 50 mA. A scanning step of 0.04° and duration time of
1.6 s per step were used at 2θ=10°ꢀ90° over all the catalysts.
Infrared analysis of adsorbed pyridine/acetone-d6/D2O/CO2 (IR-Py/acetone-d6/D2O/CO2)
IRꢀPy/acetoneꢀd6/D2O/CO2 spectra were recorded on a Bruker Tensor 27 FTIR
spectrometer. About 20 mg of the catalyst was loaded into an in situ cell and pretreated at 450 °C
for 1 h under flowing helium at 50 SCCM. The sample was cooled to 50 °C prior to taking a
background or sample scan. For the IRꢀD2O experiment, background was taken on KBr at 50°C.
Probe molecules (i.e., pyridine or acetoneꢀd6 or D2O) were introduced by flowing He (10 SCCM)
through an ice cooled bubbler for 10 minutes during which the surface was saturated as
confirmed by IR. For CO2 adsorption, 50 SCCM of 10% CO2/He was flowing through the
catalyst until the surface was saturated as confirmed by IR. The sample was then purged for 30
min using 50 SCCM He to remove any physisorbed molecules. A spectrum was then taken at 50
°C. The sample was then ramped at 10 °C/min to a given temperature, which was followed by a
30 min purge at the elevated temperature before being cooled to 50 °C and scanned. Spent
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 36
1
2
3
4
5
6
7
8
9
Zn1Zr10OzꢀH catalyst used for IRꢀPy measurements was generated as follows. Prior to the IR
experiment, reaction was first performed under reaction conditions until the end of the induction
period was identified by GC analysis. After that, the catalyst was cooled down to room
temperature in N2 flow for IR experiment. A mixture of spent Zn1Zr10OzꢀH and KBr (1/5 weight
ratio) was used to achieve a better signalꢀtoꢀnoise ratio.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Temperature Programmed Desorption of Ammonia (NH3-TPD)
NH3ꢀTPD was carried out in a Micromeretics Autochem 2920. About 100 mg of sample
was loaded and pretreated with a 10 °C/min ramp to 550 °C and held for 60 min. Ammonia was
flowed over the sample at 50 °C until it was fully saturated, at which point it was purged in
helium to remove any physisorbed ammonia. The sample was then ramped to 700 °C at 20
°C/min and the signals were monitored using mass spectroscopy (ThermoStarsTM GSD 320)
calibrated by decomposition of a calculated amount of ammonium oxalate monohydrate.
Quantification of Lewis acid sites is summarized in Table 2.
Nitrogen sorption
Nitrogen sorption experiments were carried out at liquid nitrogen temperatures (ꢀ196 °C) on
the Micromeretics TriStar 2 3020 physisorption analyzer. Prior to measurement, samples were
degassed at 350 °C for 3 h under vacuum.
2.3 Catalyst evaluation
The catalysts were evaluated in a homeꢀbuilt test unit described elsewhere.25 Briefly, a given
amount of catalyst was loaded into a microtubular fixedꢀbed reactor (i.d., 5 mm) and pretreated
at 450 °C with 50 SCCM N2 for 30 minutes. An ethanol (or acetone or diacetone alcohol)/water
solution with steamꢀtoꢀcarbon molar ratio (S/C) of 5 is typically used unless otherwise stated.
The solution was fed to an evaporator at 150 °C and carried into the reactor by flowing N2.
8
ACS Paragon Plus Environment

Page 9 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Products were analyzed online using a Shimadzu GCꢀ2014 Gas Chromatography (GC) equipped
with an auto sampling valve, HPꢀPlot Q column (30 m, 0.53 mm, 40 ꢁm), flame ionization
detector (FID), and thermal conductivity detector. The product stream was then fed through a
condenser and the dry gases were sent to an online Agilent 490 MicroGC equipped with 4 heated
blackflush channels (MSSA, PPQ, AL2O3/KCL, 5CB) for analysis. Equations 1 and 2 show the
ideal reaction for ethanolꢀtoꢀisobutene and acetoneꢀtoꢀisobutene. Based on the two equations, the
theoretical carbon selectivity to isobutene (StꢀIB) for ethanolꢀtoꢀisobutene and acetoneꢀtoꢀ
isobutene are 66.7% and 88.9%, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For the kinetic acetone/H2O and acetoneꢀd6/D2O reactions, 10ꢀ400 mg of catalyst was loaded
and the space velocity was varied to control the acetone conversion from 5% to 20%. The rate
was calculated based on the Lewis acid sites measured by NH3ꢀTPD.
For the controlled acetone/H2O experiments with preꢀadsorbed carbonate and acetate species,
the preꢀadsorption of carbonate and acetate was conducted by replicating the same procedure for
IRꢀacetone experiments. Specifically, 10.5 mg of the catalyst (diluted with 490 mg SiC) was
loaded into the reactor and pretreated at 450 °C for 1 h under flowing nitrogen at 50 SCCM. The
catalyst was then cooled down to 50 ˚C where acetone adsorption was conducted by flowing the
0.47 mol% acetone in N2 for 5 min. After the acetone adsorption, the catalyst was purged with
flowing N2 (50 SCCM) for 30 min, and then ramped to 450 ˚C in N2 (50 SCCM), where the
acetone/H2O steadyꢀstate reaction was performed.
Acetone/ethanol conversion (Xr), product selectivity (Sproductsꢀi), and isobutene theoretical
yield (YIB) were calculated as follows: Xr= (MolreactantꢀinꢀMolreactantꢀout)/Molreactantꢀin; Sproductsꢀ
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 36
1
2
3
4
i=(Molproductꢀi × αi)/(Molreacted reactant× β), αi and β refers to the carbon number in productꢀi and
5
6
reactant, respectively; YIB=Xproductsꢀi / StꢀIB
.
7
8
9
Table 1. Notation, physical/chemical properties of the (mixed) metal oxides.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Type of acidity*
Brønsted Lewis
Surface area
(m2/g)
138.0
Sample name
Method of preparation
Note
ZrO2ꢀH
Zn1Zr10OzꢀH
ꢀ
ꢀ
Hard templateꢀBP2000 ref25
Hard templateꢀBP2000 ref25
ꢀ
ꢀ
124.0
ZrO2
ꢀ(minor) ꢀ
89.0
91.0
4.0
Calcined Zr(OH)4
this work
Zn1Zr10OzꢀI
ZnO
ꢀ
ꢀ
ImpregnationꢀZr(OH)4 this work
Commercial ZnO
this work
* identified by IRꢀPy
3. Results and Discussions
3.1 Physical-chemical properties of the catalyst
Both physical and chemical properties of ZnxZryOzꢀH have been described elsewhere,25 and
summarized in Table 1. BET surface areas of the ZnxZryOzꢀI catalysts are also shown in Table 1.
ZrO2 and Zn1Zr10OzꢀI show comparable surface area of ̴
90 m2/g. Commercial ZnO exhibits a
low surface area of 4 m2/g. XRD patterns of the ZnxZryOzꢀI are shown in Figure 1. A mixture of
tetragonal and monoclinic ZrO2 phases is detected on ZrO2 prepared by calcination of Zr(OH)4 at
550 °C. With Zn addition, tetragonal ZrO2 is resolved by XRD after calcination at 550 °C. When
the Zn/Zr ratio is between 1/6 and 1/14, no diffraction peak characteristic of ZnO but only
tetragonal ZrO2 phase is detected on ZnxZryOzꢀI. This suggests that ZnO is highly dispersed and
inhibits the phase transition of ZrO2 from tetragonal to monoclinic phase. Interestingly, IR
spectra of both ZrO2 and ZnxZryOzꢀI mixed oxides (Figure 2) display the pattern of surface
hydroxyl groups that are identical to monoclinic ZrO2.29 This indicates that the surface of both
catalysts is dominant by a thin layer of monoclinic ZrO2 phase, which is beyond the detection
10
ACS Paragon Plus Environment

Page 11 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
limit of XRD. Similar behavior was also observed on other heteroatom (e.g., Y2O3) doped
ZrO2.30 It should be noted that the peak representing terminal (3770 cmꢀ1) and triꢀbridged
hydroxyl groups (3677 cmꢀ1) shifted to lower and higher wave numbers after the addition of
ZnO, an indication of change of the surface chemistry over ZnxZryOzꢀI.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. XRD patterns of ZrO2 and ZnxZryOzꢀI catalysts. * marked the monoclinic phase
Figure 2. DRIFTS analysis of surface hydroxyl groups on ZrO2 (black curve) and ZnxZryOzꢀI
(red curve) catalysts. Catalysts were dehydrated at 450 °C for 1 h before IR collection. KBr was
used as background.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 36
1
2
3
4
5
6
7
8
9
IR spectra of adsorbed pyridine reveal strong Lewis acidity on ZrO2 evidenced by the bands at
1604 and 1444 cmꢀ1 (Figure 3). In addition, a weak band at 1540 cmꢀ1 is also resolved which is
identified as protonated Py species.7,21 However, the intensity of this band is much weaker than
that on ZrO2 prepared using BP2000 carbon as a hard template (ZrO2ꢀH)25 on which significant
sulfur (1.3 wt%) was detected.31 Given the fact that strong Brønsted acid sites could be generated
on sulfated zirconia,32 the weak Brønsted acid sites on ZrO2 (Table 1) are likely from a S impurit
on the Zr(OH)4 precursor (<200 ppm). After addition of ZnO (i.e., Zn1Zr10OzꢀI), however, only
Lewis acidity is observed (Figure 3), which is consistent with the previous observation that ZnO
can passivate Brønsted acid sites.25 In addition, the peak associated with pyridine adsorbed on
Lewis acid sites shifted to higher wave numbers, suggesting the strength of the Lewis acidity
becomes weaker.25 The weakened Lewis acidity can be further confirmed by our theoretical
calculations (section 3.5). The surface basicity of the catalysts was also investigated using IRꢀ
CO2 experiments, as shown in Figure S1. It was found that, after the addition of ZnO on ZrO2
(i.e., Zn1Zr10OzꢀI), the strength of both surface Lewis acidꢀbase pairs and basic cus O2ꢀ increased
signficantly, evidenced by the experiments at different desorption temperatures. It suggests that
the addition of ZnO weakened surface Lewis acidity and improved basicity, leading to the
balanced acidꢀbase properties on the Zn1Zr10OzꢀI.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12
ACS Paragon Plus Environment

Page 13 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. IRꢀPy of ZrO2 and ZnxZryOzꢀI catalysts
3.2 Acetone reaction on ZnxZryOz-H
In our previous work, different residence time experiments reveal that acetone acts as the key
intermediate in ethanolꢀtoꢀisobutene, and the reaction is limited by the acetoneꢀtoꢀisobutene
conversion.25 Further durability tests on Zn1Zr10OzꢀH catalysts show that IB selectivity
decreases slowly with timeꢀonꢀstream, meanwhile acetone selectivity increases.24 However, the
selectivity to acetone and IB remains constant, suggesting that the catalyst loses active sites for
the secondary acetone reaction during the ethanolꢀtoꢀisobutene.24 To better understand the
deactivation mechanism, we investigate the acetoneꢀtoꢀisobutene rateꢀdetermining reaction on
the catalysts with different type of acidity.
3.2.1 Acetone reaction on Zn1Zr10Oz-H
Figure 4 shows the acetoneꢀtoꢀisobutene reaction over the Zn1Zr10OzꢀH catalyst with Zn/Zr
ratio optimized for ethanolꢀtoꢀisobutene.25 Acetone conversion initially increases with timeꢀonꢀ
stream. After 10 h timeꢀonꢀstream, it decreases monotonically with timeꢀonꢀstream, due to the
loss of active sites observed in ethanolꢀtoꢀisobutene reaction. Significantly, there is an induction
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 36
1
2
3
4
5
6
7
8
9
period during which acetone conversion increases with timeꢀonꢀstream (timeꢀonꢀstream <10 h).
In addition, appreciable amounts of 1ꢀ and 2ꢀbutenes (~32%) are formed initially at the expense
of isobutene. With timeꢀonꢀstream, selectivity towards 1ꢀ and 2ꢀbutenes decreases rapidly with
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
=
the concurrent increase of isobutene selectivity. However, selectivity to total C4 olefins (1ꢀ
butene, 2ꢀbutene, and isobutene) is high and remains constant (~89.0%). After the induction
period, selectivity to 1&2ꢀbutene decreases to zero, and selectivity to isobutene reaches ~89%,
which is close to the theoretical selectivity (i.e., ~88.9%). Such dramatic changes in product
distribution during the induction period are likely related to the evolution of catalyst surface
properties.
Brønsted acidity has previously been identified as an active site for the acetoneꢀtoꢀisobutene
reaction.24,25 To better correlate product selectivity with catalyst surface properties, we further
investigated the surface acidity of the fresh catalyst and spent catalyst right after the induction
period using IR spectra of adsorbed pyridine (Figure 5). Over the fresh Zn1Zr10OzꢀH catalyst
(Figure 5A), both Lewis (1451, 1574, and 1608 cmꢀ1) and Brønsted acid sites (1544 cmꢀ1) are
observed at a desorption temperature of 150 °C. Upon increasing the desorption temperature,
pyridine adsorbed on Lewis and Brønsted acid sites decreases. However, pyridine adsorbed on
Brønsted acid sites decreases more profoundly. At 450 °C, while most pyridine on Brønsted acid
sites have desorbed, those adsorbed on Lewis acid sites are still observed in substantial amount,
suggesting stronger Lewis acidity is present on the catalyst.
Over the spent Zn1Zr10Oz catalyst right after the induction period (Figure 5B), the peaks
characteristic of both Lewis (1451, and 1608 cmꢀ1) and Brønsted acid sites (1544 cmꢀ1) are still
present at a desorption temperature of 150 °C. However, the shoulder peak at 1445 cmꢀ1
(characteristic of strong Lewis acid sites on ZrO2) almost disappears. More importantly, relative
14
ACS Paragon Plus Environment

Page 15 of 36
Journal of the American Chemical Society
1
2
area ratio of the integrated peak representing Brønsted acid sites at 1550 cmꢀ1 (A1550) to that of
Lewis acid sites at 1450 cmꢀ1 (A1450) changes significantly. At a desorption temperature of 150
°C, the A1550/A1450 decreases from 0.24 to 0.029, indicating that the surface Brønsted acid sites
reduce rapidly during the induction period. It is well known that Brønsted acid sites could
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
= 33
catalyze the isomerization of C4 . Considering the conversion of isobutene to butenes is
thermodynamically favorable at high temperature (e.g., 450 °C in this case),34 the constant and
=
high selectivity to C4 coupled with decreasing 1&2ꢀbutene/isobutene ratio with diminished
Brønsted acid sites suggest that Brønsted acid sites catalyze isomerization of isobutene to form
1&2ꢀbutene. The Brønsted acid sites could also catalyze other polymerization side reactions,
leading to coke deposition and thus, loss of Brønsted acid sites on the catalyst.
Unexpectedly, acetone conversion increases with decreased Brønsted acid site, which is
contradictory with the hypothesis that Brønsted acid sites are active for acetone conversion to
isobutene. While it is still unclear what role the Brønsted acid site plays in acetoneꢀtoꢀisobutene,
a more active and stable phase must be present on the Zn1Zr10OzꢀH catalyst. This phase is
believed to be the Lewis acidꢀbase pairs, supported by the IRꢀPy results of the spent Zn1Zr10OzꢀH
catalyst. In addition, methane and CO2 selectivities increase slowly at the expense of isobutene
with timeꢀonꢀstream, which could be due to the slightly favorable acetone decomposition on the
Lewis acidꢀbase pair. The activity for acetone condensation and decomposition on Lewis acidꢀ
base pairs is further discussed in section 3.3. It should be noted that Brønsted acid site is indeed
active for acetoneꢀtoꢀisobutene, which will be confirmed in section 3.2.2. Given the fact that both
Bronsted acid sites and Lewis acidꢀbase pairs are active, a decrease in acetone conversion is
expected with the loss of Brønsted acid sites. However, an increase in acetone conversion was
observed. We speculate that the existence of both Brønsted acid (proton) and Lewis acidꢀbasic
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 36
1
2
3
4
5
6
7
8
9
could affect each other. The proton generated from water dissociation on the strong Lewis center
of the ZnxZryOzꢀH catalyst could be possibly stabilized by the basic oxygen of the Lewis acidꢀ
base pair (Scheme S1), inhibiting the activity of the Lewis acidꢀbasic pair by suppressing the αꢀH
abstraction, a key step for acetone aldolization. After the induction period, the strong Lewis
center that generates the protons could be passivated by coking, releasing the Lewis acidꢀbase
pair after the consumption of the protons.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Acetone to isobutene reaction on Zn1Zr10OzꢀH. 100 mg catalyst, Pacetone= 0.47kPa,
S/C=5, T=450 °C, WHSV =0.37 gacetone.gcatal.hꢀ1.
16
ACS Paragon Plus Environment

Page 17 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. IRꢀPy of A) fresh Zn1Zr10OzꢀH and B) spent Zn1Zr10OzꢀH right after induction period
at various desorption temperatures.
3.2.2 Acetone reaction on ZrO2-H
To further confirm the role of Brønsted acidity, the acetone reaction on the ZrO2ꢀH catalyst
was studied and the results are shown in Figure 6. Since both Brønsted and Lewis acid sites are
present on the ZrO2ꢀH catalyst as previously reported,25 similar performances observed during
the induction period for the Zn1Zr10OzꢀH catalyst are expected for the ZrO2ꢀH catalyst. Indeed, a
similar induction period, in which acetone conversion and isobutene selectivity increase while
1&2ꢀbutene selectivity decreases with timeꢀonꢀstream, is also observed on the ZrO2ꢀH catalyst.
After the induction period, however, acetone conversion and product distribution on the ZrO2ꢀH
show totally different trends from those on Zn1Zr10OzꢀH. Acetone conversion decreases
dramatically with timeꢀonꢀstream, and selectivity to methane and CO2 increases rapidly at the
expense of isobutene on ZrO2. In addition, a small amount of propylene is also formed. These
results suggest that Lewis acidꢀbase chemistry on the ZrO2ꢀH is different from that on
Zn1Zr10OzꢀH.
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 36
1
2
Using FTꢀIR and mass spectrometer, Zaki et al21 studied acetone surface reaction on ZrO2. It
was found that acetone aldolization could occur on Lewis acidꢀbase pairs at low temperature
(<200 °C) to form mesityl oxide via diacetone alcohol intermediate. At high temperature (>350
°C), further cracking of mesityl oxide to isobutene was observed. Meanwhile, the surface
hydroxyl group becomes very oxidative, catalyzing acetone decomposition and other reactions to
form methane, CO2, and propylene. Therefore, we believe that the increasing acetone
decomposition is due to the loss of Brønsted acid sites and the emerging Lewis acidꢀbase
chemistry on the ZrO2ꢀH catalyst. The rapid deactivation together with the low carbon balance
(~71%) suggest that rapid coke formation and/or other side reactions also occur on the ZrO2ꢀH
catalyst, likely due to the stronger Lewis acidity evidenced by IRꢀPy.25. More importantly, the
poor catalytic performance of Lewis acidꢀbase pair on ZrO2ꢀH confirms that the high selectivity
to IB in the induction period should be attributed to presence of Brønsted acid site during the
acetoneꢀtoꢀisobutene reaction.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
ACS Paragon Plus Environment

Page 19 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Figure 6. Acetone to isobutene reaction on ZrO2ꢀH. 100 mg, Pacetone= 0.47 KPa, S/C=5, T=450
°C, WHSV =0.18 gacetone.gcatal.hꢀ1. Carbon balance is ~80% during induction period and ~71%
afterwards.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While it is still unclear why acetone conversion increases with diminished Brønsted acid active
sites in the induction period on the ZnxZryOzꢀH catalyst, it does not affect our conclusion that
Brønsted acid site is the active site for acetone aldolization toward isobutene production.
Nevertheless, Brønsted acidity also favors secondary isobutene isomerization reaction to form
1&2ꢀbutene, and polymerization reactions (e.g., C4 olefins and other intermediates such as
mesityl oxide) to form coke with timeꢀonꢀstream, which is further confirmed by the isobutene
reactions on the Zn1Zr10OzꢀH catalyst (Figure S2). Coke deposition on Brønsted sites leaves only
Lewis acidꢀbase pair on Zn1Zr10OzꢀH, which is more active for acetone conversion but does not
catalyze isobutene isomerization (Figure S2). The strong Lewis acidꢀbase pairs on ZrO2 also
likely catalyze undesired acetone decomposition and coke formation reactions. Addition of ZnO
to ZrO2ꢀH possibly neutralizes these stronger Lewis acid sites and introduces basic sites, leading
to a highly active and selective catalyst for the acetoneꢀtoꢀisobutene reaction.
3.3 Acetone and acetone-d6 reaction on ZnxZryOz-I
Given the proposed functions of Lewis acidꢀbase pairs and the detrimental effect of Brønsted
and stronger Lewis acid sites in the acetoneꢀtoꢀisobutene reaction, the ZnxZryOzꢀI catalyst was
synthesized to improve the selectivity and stability in acetoneꢀtoꢀisobutene reaction. PyꢀIR
confirmed that only Lewis acidꢀbase pairs are present (Figure 3). The acetone reaction was
studied to validate that Lewis acidꢀbase pairs are highly active for acetone conversion to
isobutene, while mitigating isobutene isomerization and undesired coke formation. For
comparison purposes, ZrO2 and ZnO were also studied. Figure 7A shows the catalysts’
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 36
1
2
3
4
5
6
7
8
9
performance in terms of isobutene selectivity versus acetone conversion, and Figure 7B
compares the product distribution at similar acetone conversion of ~22ꢀ26%. As expected,
isobutene selectivity is constantly lower than 20% over ZrO2 (Figure 7A). Majority of
decomposition products, methane and CO2, are formed (Figure 7B). In addition, a notable
amount of propylene (<1%) is also formed (Figure 7B). ZnO shows a relatively high selectivity
to isobutene (Figure 7). But it exhibits significantly low activity, which gives an aldolization
rate constant 5ꢀfold lower than that of Zn1Zr10OzꢀI (Table 2).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Isobutene selectivity versus acetone conversion (A) and products distribution at similar
acetone conversion on mixed oxides (B). 10ꢀ400 mg, Pacetone= 0.47 KPa, T=450 °C
The addition of ZnO to ZrO2, however, significantly changed the reaction pathway from
acetone decomposition to aldolization and selfꢀdeoxygenation. Over the Zn1Zr10OzꢀI catalyst, an
isobutene selectivity of >80 % is achieved. The isobutene selectivity can even reach up to its
=
theoretical value (88.9%) at acetone conversion of ~10%. Notably, no other C4 isomers are
observed, suggesting that the Lewis acidꢀbase properties on Zn1Zr10OzꢀI is very selective in
isobutene production. In addition, the carbon balance is consistently higher than 95% on the
20
ACS Paragon Plus Environment

Page 21 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Zn1Zr10OzꢀI catalyst as opposed to ZnxZryOzꢀH (~85%), further confirming that side reactions
(e.g., coking) observed in the presence of Brønsted acidity on ZnxZryOzꢀH are mitigated.
It is worth mentioning that selectivity to decomposition products (i.e., methane and CO2)
increases at the expense of isobutene as acetone conversion increases, likely due to kinetic
inhibition of the aldolization reaction by adsorbed intermediate/product.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Acetone aldolization and decomposition rate constant on the (mixed) metal oxides.
Decomposition
rate constant #
Aldolization
KIEdecom
Lewis acid
site*
(mmol/g)
KIEaldol
rate constant #
Catalyst
p
kdecomp
1.47
kdecomp
’
kaldol
kaldol’
ZrO2
1.35
1.1
0.13
0.13
1.0
0.443
0.319
0.015
Zn Zr O
1.39
1.34
ꢀꢀ
1.0
ꢀꢀ
26.50
5.15
25.44
ꢀꢀ
1.0
ꢀꢀ
zꢀI
1
10
0.49
ZnO
* Number of Lewis acid site is quantified by NH3ꢀTPD; # Rate is normalized by the number of the Lewis acid site; k
for acetone/H2O reactions; k’ for acetoneꢀd6/D2O reactions
It is proposed that, over Zn1Zr10OzꢀI, ZnO addition could either modify Lewis acidꢀbase
properties or suppress the oxidation capability of hydroxyl groups on the ZrO2, leading to the
relatively favorable acetone aldolization and selfꢀdeoxygenation toward isobutene formation. To
confirm this hypothesis, kinetic analysis was performed (Figure S3), and the rate constants (k)
are listed in Table 2. kdecomp for acetone decomposition (1.47) is more than a factor of 10 higher
than kaldol for acetone aldolization (0.13) on ZrO2, suggesting acetone decomposition dominates
on this catalyst. After the addition of ZnO (i.e., Zn1Zr10OzꢀI), kdecomp decreases slightly from 1.47
to 1.39. However, kaldol increases dramatically up to 26.50. The kaldol/kdecomp ratio on Zn1Zr10OzꢀI
is two orders of magnitudes higher than that on ZrO2 (19.1 vs 0.1). Apparently, ZnO addition
mainly modifies Lewis acidꢀbase properties, leading to enhanced acetone aldolization and selfꢀ
deoxygenation reaction, and thus, isobutene selectivity. It should be mentioned that, despite the
low kdecomp on ZnO, its kaldol is also much lower than that of Zn1Zr10OzꢀI (Table 2). It can be
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 36
1
2
3
4
5
6
concluded that Lewis acidꢀbase pair on the Zn1Zr10OzꢀI is essential for high acetone aldolization
rate.
7
8
9
To further understand the reaction mechanism, acetoneꢀd6/D2O reaction was also studied on
both ZrO2 and Zn1Zr10OzꢀI catalysts. Kinetic isotopic effects (KIEs) for the acetone
decomposition and aldolization pathway were measured (Table 2) to identify if αꢀH abstraction
and H2O dissociation are the limiting steps in the cascade reactions. Over both Zn1Zr10OzꢀI and
ZrO2 catalysts, a KIE from 1.0 to 1.1 was observed for acetone aldolization and decomposition
pathway. It suggests that both the αꢀH abstraction and water dissociation are not the rate
determining step.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. DRIFTS analysis of adsorbed acetoneꢀd6 over A) ZrO2ꢀI and B) Zn1Zr10OzꢀI.
22
ACS Paragon Plus Environment

Page 23 of 36
Journal of the American Chemical Society
1
2
3
4
3.4 DRIFTS analysis of adsorbed acetoneꢀd6
5
6
7
8
To further elucidate the roles of ZnO addition to ZrO2 and possible reaction intermediates
involved leading to improved isobutene selectivity, surface reaction of acetoneꢀd6 was further
studied using in situ IR. The results are shown in Figure 8. The intent of using acetoneꢀd6 was to
differentiate the activity for both αꢀH abstraction and aldolization. Figure 8A compares the IR
spectra of acetoneꢀd6 adsorbed on ZrO2 at different temperatures. At 30 °C, the νC=O/νC=C/νCꢀ
C/δCꢀH region give a strong band at 1691 cmꢀ1, a broad shoulder centered at ~1604 cmꢀ1, two
weak bands at 1467, 1343 cmꢀ1, and two moderate bands at 1288 and 1261 cmꢀ1. The strong band
at 1691cmꢀ1 can be assigned to the νC=O of acetone adsorbed on Lewis acid site,21 which is
further supported by the νCꢀC band at 1288 and 1261 cmꢀ1. The broad shoulder at 1604 cmꢀ1
could have originated from the νC=O absorption of adduct species (i.e. diacetone alcohol and
mesityl oxide), due to the minor aldolꢀcondensation of acetone.21 The weak bands at around 1467
and 1343 cmꢀ1 are due to the δCꢀH of adsorbed species (i.e., acetone, diacetone alcohol and enol)
formed from surface HꢀD exchange reaction between acetoneꢀd6 and surface hydroxyl
groups,14,35 In the νOꢀD and νCꢀD region, νCꢀD bands are observed at 2225, 2065, and 2138 cmꢀ1
(Figure S4). Unfortunately, no νOꢀD bands are resolved, but instead, a broad hump at ~2700 cmꢀ
1 is present, due to hydrogen bonded nature of OD groups.14
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Upon increasing temperature, the bands corresponding to adsorbed acetoneꢀd6 (1691, 1288,
and 1261 cmꢀ1) decrease significantly, and disappear at temperatures above 300 °C. The
decreasing acetone signal could be due to desorption and further conversion of acetone. The
former can be confirmed by the separate acetoneꢀTPD experiment, which shows acetone
desorption at ~110 °C. (Figure S5). The latter is evidenced by the increased band intensity at
1612 cmꢀ1, characteristic of νC=O of diacetone alcohol coordinated to Lewis acid site, which
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 36
1
2
3
4
5
6
7
8
9
first increases and then decreases with a maximum at 200 °C. Due to the significant oxidative
nature of hydroxyl groups at high temperature,21 most of diacetone alcohol could be quickly
converted back to acetone14 that decomposes (Scheme 2, red highlighted). This decomposition
can be further confirmed by the significant decrease of intensity of all absorption bands (Figure
8A). Meanwhile, further dehydration of diacetone alcohol to mesityl oxide is also observed,
evidenced by the further redꢀshift of νC=O band to 1595 cmꢀ1, νC=C absorption at 1525 cmꢀ1,
νCꢀD bands at 2225 cmꢀ1, and δCꢀH band at 1343 cmꢀ1.21 Notably, once mesityl oxide is formed
it is strongly adsorbed on the catalyst until 400 °C. Only minor mesityl oxide cracking occurs,
evidenced by the shoulder bands of νCOOꢀ of acetate species at 1575 (antisymetric vibration)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
21
and 1430 cmꢀ1 (symmetric vibration) and that of carbonate species at 1550, 1332 cmꢀ1.29 It
suggests that the strongly adsorbed mesityl oxide can block the Lewis acid site on ZrO2, leading
to the low aldolization activity. In addition, νOꢀD bands become resolved at 100 °C and
increases with temperature (Figure S4), suggesting enhanced HꢀD scrambling between acetoneꢀ
d6 and surface hydroxyl at higher temperature.
Over Zn1Zr10OzꢀI catalyst, other than the adsorbed acetone band at 1691 cmꢀ1, mesityl oxide
bands at 1589 cmꢀ1 and 1525 cmꢀ1 are also detected at 30 °C (Figure 8B), suggesting a facile
aldolization pathway consistent with the kinetic results in section 3.3. As temperature increases,
acetone intensity decreases, due to the acetone desorption (Figure S5) and further reaction of
acetone (Figure 8B). It should be mentioned that the extent of acetone desorption on ZrO2 and
Zn1Zr10OzꢀI is almost identical (Figure S5). However, compared to that on ZrO2, acetone
intensity decreases more rapidly with temperature, and disappears at a lower temperature (i.e.,
200 °C), which further confirms the facile aldolization pathway on Zn1Zr10OzꢀI. With the
decrease of acetone, only a small amount of diacetone alcohol is observed at 100ꢀ200 °C.
24
ACS Paragon Plus Environment

Page 25 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Mesityl oxide is always predominant on the Zn1Zr10OzꢀI catalyst. Moreover, when temperature is
above 300 °C (i.e., 400 °C), mesityl oxide decomposition seems significant, and only a small
amount of mesityl oxide is left on the catalyst surface, evidenced by the very weak νCꢀD (2224
cmꢀ1), νC=O (1589 cmꢀ1), and νC=C (1525 cmꢀ1) bands. Instead, it is dominant with acetate and
carbonate absorption, the former being assigned to the 1572, 1467, and 1430 cmꢀ1 band, and the
latter to the 1547, 1446, 1432, and 1335 cmꢀ1 ones.21,29 This observation suggests that the
forward diacetone alcohol dehydration to mesityl oxide and mesityl oxide cracking reactions are
significantly enhanced relative to its reverse reaction to acetone on Zn1Zr10OzꢀI (Scheme 2, green
highlighted). This is consistent with the steadyꢀstate reaction data showing higher aldolꢀaddition
rate than that of decomposition. Unfortunately, the evolution of νOꢀD bands with temperature is
difficult to differentiate from that of ZrO2 (Figure S4), due to the interference (i.e., hydrogen
bond) of strongly adsorbed acetate or carbonate species. It should be mentioned that the
formation of carbonate and acetate species could block the surface active sites on the metal
oxides (e.g., surface basic oxygen). 36ꢀ38 To verify if the surface identified acetate/carbonate will
deactivate the catalyst or not, separate acetone reactions were performed on both fresh
Zn1Zr10OzꢀI and Zn1Zr10OzꢀI with preadsorbed surface carbonate species (Figure S6). No obvious
effect on the acetone has been observed, suggesting the surface carbonate does not deactivate the
catalyst under the current reaction conditions consistent with the highly stable ethanolꢀtoꢀ
isobutene reaction discussed in the section 3.6.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3.5 DRIFTS analysis of adsorbed D2O
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9 DRIFTS analysis of adsorbed D2O on ZrO2 (A) and Zn1Zr10Oz (B).
Figure 10. Evolution of the peak area with temperature representing OD groups on ZrO2 (red)
and Zn1Zr10OzꢀI (black) shown in Figure 9.
To further verify the HꢀD exchange reaction and water dissociation, D2OꢀDRIFTS was
performed on the two catalysts. Figure 9 displays the evolution of hydroxyl bands with
desorption temperatures. Both isolated OH at 3770 cmꢀ1 and triꢀbridged OH at 3770 cmꢀ1 are
observed on the two freshly pretreated samples (gray line). Upon adsorption of D2O at 50 °C,
peaks characteristic of OH groups on both supports decreases significantly and a broad hump at
3000ꢀ3400 cmꢀ1, characteristic of hydrogen bonded OH, appears simultaneously. This suggests
26
ACS Paragon Plus Environment

Page 27 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
that the adsorbed D2O interacts closely with hydroxyl groups via hydrogen bond. In addition, a
new triꢀbridged OD peak is observed on both ZrO2 (2711 cmꢀ1) and Zn1Zr10OzꢀI (2714 cmꢀ1).
However, no isolated OD was observed. It indicates that no water dissociation occurs, but HꢀD
exchange does take place on both catalysts.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The fact that the peak intensity for triꢀbridged OD is higher on Zn1Zr10OzꢀI than on ZrO2
reveals that the HꢀD exchange reaction is more active for the former. As the temperature
increases, hydrogen bonded OH groups decrease as evidenced by the decreased intensity of the
broad hump 3000ꢀ3400 cmꢀ1. Meanwhile, the peak of triꢀbridged OD increases and isolated OD
appears at 150 °C, suggesting enhanced HꢀD exchange and water dissociation might coexist. For
a better comparison, the peak area representing the amount of OD bands is further integrated to
explore the effect of temperature on the HꢀD exchange and water dissociation, as shown in
Figure 10. As temperature increases, it is clear that the amount of surface OD increases at a
much faster rate on the Zn1Zr10OzꢀI than that on ZrO2. While the amount of OD increases
monotonically as temperature is increased up to 450 °C, surface saturation appears to be
achieved at >200 °C. It should be noted that the intensity of bands associated with OH groups
also increase with temperature, which is due to the recovery of the OH groups from the hydrogen
bonded ones at 3000ꢀ3400 cmꢀ1 since most of hydrogen bonded D2O is removed. At 450 °C,
~25% and ~10% of the OH groups have been exchanged on Zn1Zr10OzꢀI and ZrO2, respectively.
It further confirms the facile HꢀD exchange on the Zn1Zr10OzꢀI catalyst. Additionally, upon D2O
adsorption and treatment at 450 °C, the total OH groups on Zn1Zr10OzꢀI and ZrO2 increased by
1.51 and 1.55 times, respectively, assuming a OD/OH response ratio of 0.51.39 The increased OH
groups are attributed to the water dissociation that occurs on both Zn1Zr10OzꢀI and ZrO2 catalyst
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 36
1
2
3
4
5
6
with the former being less active, which can be further confirmed by our theoretical calculations
(Figure S7).
7
8
9
3.6 Zn1Zr8Oz-I Catalyst Stability for ethanol-to-isobutene
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We have shown in our previous report that the ZnxZryOzꢀH (i.e., Zn1Zr8OzꢀH) catalyst suffers
deactivation in ethanolꢀtoꢀisobutene, due to the rapid loss of active sites for acetoneꢀtoꢀisobutene
reaction. After 27 h, isobutene yield dropped from ~79% to ~61% with the concurrent increase
of acetone.24 Meanwhile, selectivities to other products were constant, suggesting that the
catalyst is losing active sites for acetoneꢀtoꢀisobutene conversion. Given the high selectivity and
stability of the balanced Lewis acidꢀbase pairs in acetoneꢀtoꢀisobutene on the ZnxZryOzꢀI
catalyst, we further evaluated the stability of Zn1Zr8OzꢀI catalyst in terms of isobutene yield
under similar reaction conditions (Figure 11). It is clear that the Zn1Zr8OzꢀI catalyst shows
extremely good stability in terms of isobutene selectivity. After 200 h timeꢀonꢀstream, the loss in
isobutene yield is very minor (less than 3%). Figure 12 shows the thermo gravimetric and
differential scanning calorimetry (TGꢀDSC) analysis of the spent catalysts. Only ~0.7 wt% coke
was detected on Zn1Zr8OzꢀI after 200 h of timeꢀonꢀstream operations, whereas >5 wt% coke was
observed on Zn1Zr8OzꢀH after 27 h timeꢀonꢀstrea. It is clear that the balanced Lewis acidꢀbase
pairs are resistant to coke formation.
28
ACS Paragon Plus Environment

Page 29 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 11. Stability test for ethanolꢀtoꢀisobutene over Zn1Zr8OzꢀI(Solid mark). Dark yellow
marked isobutene yield and magenta marked acetone selectivity. Pethanol =83 KPa S/C=2.5,
WHSV=0.3 gethanol.gcatalꢀ1.hꢀ1. T=450 °C. For comparison, the ethanolꢀtoꢀisobutene over
Zn1Zr8OzꢀH is also plotted (hollow marks).
Figure 12. TGA (dark yellow) –DSC (magenta) analysis of spent Zn1Zr8OzꢀI (solid mark). For
comparison, the TGAꢀDSC over Zn1Zr8OzꢀH is also plotted (hollow marks). Spent Zn1Zr8OzꢀI
29
ACS Paragon Plus Environment