Vapor-phase catalytic dehydration of 2,3-butanediol to 3-buten-2-ol over ZrO2 modified with alkaline earth metal oxides

Article abstract of DOI:10.1016/j.apcata.2016.11.022

Vapor-phase catalytic dehydration of 2,3-butanediol (2,3-BDO) to produce 3-buten-2-ol (3B2OL) was investigated over several monoclinic ZrO₂ (m-ZrO₂) catalysts modified with alkaline earth metal oxides (MOs), such as SrO, BaO, and MgO, to compare with the previously reported CaO/m-ZrO₂. It was found that these modifiers enhanced the 3B2OL formation to the same level as CaO did by loading an appropriate MO content. Among all the tested catalysts, the BaO/m-ZrO₂ calcined at 800?°C with a low BaO content (molar ratio of BaO/ZrO₂?=?0.0452) shows the highest 2,3-BDO conversion (72.4%) and 3B2OL selectivity (74.4%) in the initial stage of 5?h at 350?°C. In order to characterize those catalysts, their catalytic activities, crystal structures, and basic properties were studied in detail. In X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) experiment, it was elucidated that highly dispersed M-O-Zr (M?=?Ca, Sr, and Ba) hetero-linkages were formed on the surface by loading these MOs onto m-ZrO₂ with an appropriate content and then calcining at 800?°C. It can be concluded that the M-O-Zr hetero-linkages generate the proper base-acid balance for the efficient formation of 3B2OL from 2,3-BDO.

Full text of DOI:10.1016/j.apcata.2016.11.022

Accepted Manuscript
Title: Vapor-phase catalytic dehydration of 2,3-butanediol to
3-buten-2-ol over ZrO2 modified with alkaline earth metal
oxides
Author: Hailing Duan Yasuhiro Yamada Shingo Kubo Satoshi
Sato
PII:
S0926-860X(16)30561-0
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.11.022
APCATA 16070
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
7-9-2016
5-11-2016
15-11-2016
Please cite this article as: Hailing Duan, Yasuhiro Yamada, Shingo Kubo, Satoshi
Sato, Vapor-phase catalytic dehydration of 2,3-butanediol to 3-buten-2-ol over
ZrO2 modified with alkaline earth metal oxides, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2016.11.022
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

A revised manuscript for the original paper submitted to the journal of Applied Catalysis A
MS# APCATA-D-16-01586
Vapor-phase catalytic dehydration of 2,3-butanediol to 3-buten-2-ol over
ZrO2 modified with alkaline earth metal oxides
Hailing Duana, Yasuhiro Yamadaa, Shingo Kubob, Satoshi Sato a,*
a Graduate School of Engineering, Chiba University, Yayoi, Inage, Chiba 263-8522, Japan
b Natural Science Center for Research and Education, Kagoshima University, 1-20-40
Korimoto, Kagoshima, 890-0065, Japan
* Corresponding author. Tel. +81-43-290-3376; fax: +81-43-290-3401
E-mail address: satoshi@faculty.chiba-u.jp (S. Sato)
1

Graphical abstract
Highlights
 Vapor-phase dehydration of 2,3-butanediol was studied over modified monoclinic
ZrO2.
 3-Buten-2-ol was preferentially formed over BaO/ZrO2 and SrO/ZrO2 as well as
CaO/ZrO2.
 The highest conversion of 72.4% was attained with the selectivity of 74.4% over
BaO/ZrO2.
 XPS study reveals the structure of active sites, M-O-Zr (M= Ca, Sr, and Ba)
hetero-linkages.
 The formation of 3-buten-2-ol proceeds via base-acid concerted mechanism.
2

Abstract
Vapor-phase catalytic dehydration of 2,3-butanediol (2,3-BDO) to produce
3-buten-2-ol (3B2OL) was investigated over several monoclinic ZrO2 (m-ZrO2) catalysts
modified with alkaline earth metal oxides (MOs), such as SrO, BaO, and MgO, to compare
with the previously reported CaO/m-ZrO2. It was found that those modifiers enhanced the
3B2OL formation to the same level as CaO did by loading an appropriate MO content.
Among all the tested catalysts, the BaO/m-ZrO2 calcined at 800oC with a low BaO content
(molar ratio of BaO/ZrO2 = 0.0452) shows the highest 2,3-BDO conversion (72.4%) and
3B2OL selectivity (74.4%) in the initial stage of 5 h at 350oC. In order to characterize those
catalysts, their catalytic activities, crystal structures, and basic properties were studied in
detail. In X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) experiment,
it was elucidated that highly dispersed M-O-Zr (M= Ca, Sr, and Ba) hetero-linkages were
formed on the surface by loading these MOs onto m-ZrO2 with an appropriate content and
then calcining at 800oC. It can be concluded that the M-O-Zr hetero-linkages generate the
proper base-acid balance for the efficient formation of 3B2OL from 2,3-BDO.
Keywords: 2,3-Butanediol; 3-Buten-2-ol; Dehydration; Alkaline earth metal oxides;
Monoclinic ZrO2; Ca-O-Zr hetero-linkage.
1. Introduction
In order to realize the low-carbon society and to diminish the usage of fossil
resources, currently, the demand for sustainable chemical compounds is growing rapidly.
Among many candidates of sustainable chemical compounds, 2,3-butanediol (2,3-BDO) is
one of the potential biomass-derived chemical compounds, and its production has been
greatly developed via fermentation with various bacteria from glucose, glycerol, sucrose,
and xylose, etc. [1-3] 2,3-BDO shows the future prospect for wide variety of application to
biomass-based chemicals via catalytic technologies [4-6]. Studies on 2,3-BDO dehydration
3

were reported even before 1938 [6,7], but the product distribution over different catalysts
relies on their acid-base properties, as shown in Scheme 1.
In the dehydration of 2,3-BDO, 2-butanone (MEK) can be produced easily in acidic
aqueous solution [8,9], over several acid solid catalysts such as bentonite [7], alumina
[10,11], silica-alumina [11], heteropoly acids [12], and zeolites [13-18]. As described in
Scheme 1, the hydroxyl group on the 2-position of 2,3-BDO can be easily captured by a
proton or an acidic site, and dissociated. Therefore, either hydrogen or methyl group in the
3-position moves to the 2-position as an electron donor, and results in the formation of MEK
and 2-metyl-propanal (MPA), respectively (group I in Scheme 1). In the work of Zhang et
al., it has been reported that MPA formation competes with MEK formation at a constant
ratio of MPA/MEK= ca. 20/68 over HZSM-5 zeolites [16]. In addition, strong acid sites
accelerate the migration of methyl group for MPA rather than hydrogen for MEK [17], while
alumina and zeolite catalysts are consistently deactivated due to their strong acidity. Nikitina
et al. also reported that the presence of Brønsted acid sites in zeolites resulted in the
formation of both MEK and MPA, as well as some heavy condensation products, while the
Lewis acid sites in alumina act mainly for the MEK formation but with less heavy products.
However, when combining the two types of acid sites together, heavy products increased
rapidly [18]. It has been also observed that high carbon balance and low coke formation can
be obtained via diluting the reactant 2,3-BDO with an appropriate amount of water [10].
1,3-Butadiene (BD) can be also produced from 2,3-BDO together with MEK over
acid catalysts such as TiO2 and γ-Al2O3 [10,19]. This is a rare case: both hydroxyl groups of
2,3-BDO must be captured by acid sites, and then dissociated simultaneously or eliminated
step by step. Thus, the two terminal hydrogens must also dissociate for the stable BD
formation, as shown in groups I and II of Scheme 1.
We have found the direct formation of BD from 2,3-BDO over Sc2O3 at high
o
reaction temperatures over 400 C [20], which resembled to ThO2 as referenced from
Winfield’s report [21]. However, Sc2O3 has weak basic sites without acidic sites [22] while
ThO2 is reported as an acid-base solid [23]. In addition, both of them can yield 3-buten-2-ol
4

(3B2OL) from 2,3-BDO at low reaction temperatures of 300-350oC [21, 24]. In these cases,
1,2-elimination of a water molecule leads to 3B2OL and then further to BD formation, as
shown in group II of Scheme 1. Since the formations of BD and 3B2OL differ in activation
energy, their distribution was successfully controlled by reaction temperature.
The 1,2-elimination of 2,3-BDO to produce BD and 3B2OL also occurred over the
base-acid bifunctional catalysts such as ZrO2 [19], CaO/ZrO2 [25], SiO2-supported sodium
phosphates [26], and SiO2-supported cesium dihydrogen phosphate [27]. Over the base-acid
bifunctional catalysts, the 2-position hydroxyl group of 2,3-BDO was captured by an acid
site and dissociated to form a carbon cation, together with the terminal hydrogen by a basic
site for a carbon anion. An appropriate base-acid balance is significant for the
1,2-elimination: the excess acid sites result in the formation of MEK, MPA, and the heavy
products, while the excess basic sites also lead to the dehydrogenation to produce acetoin
and diacetyl [19, 25, 26] or to dehydrate to 2,3-epoxybutane [28], as shown in groups II and
III of Scheme 1. On the other hand, supported phosphate catalysts can produce BD at high
reaction temperatures but cannot stop the selective formation of 3B2OL [26, 27].
Since the further dehydration of 3B2OL into BD with 94% yield proceeds even at
250oC over Al2O3 and SiO2-Al2O3 catalysts [20, 29, 30], 2,3-BDO can be converted into
3B2OL firstly and then further dehydrated into BD by Al2O3 at low temperatures. In
addition, 3B2OL is also an important chemical compound used as the intermediate for
anticancer drugs [31, 32]. In order to synthesize 3B2OL over safe and inexpensive catalysts
to take the place of ThO2 [19] and Sc2O3 [20], we have found that the base-acid bifunctional
m-ZrO2 catalysts efficiently work with a basic CaO modifier at a better base-acid balance in
the formation of 3B2OL from 2,3-BDO, and speculated the reaction mechanism (Scheme 1
of [25]).
In the present paper, our aim is to prepare more active catalysts than CaO/ZrO2 for
the 3B2OL production from 2,3-BDO. Another aim is to elucidate the active sites of catalyst
for the reaction. Thus, we investigated m-ZrO2 catalysts modified with other MOs such as
SrO, BaO, and MgO in the 2,3-BDO dehydration to compare with the results of CaO
5

modifier as reported in the previous work [25]. In addition, XPS of CaO-modified m-ZrO2
as well as XRD and temperature-programmed desorption (TPD) of CO2 was performed to
further discuss the structure of active sites.
2. Experimental
2.1 Catalyst preparation
Mg(NO3)26H2O, Ca(NO3)24H2O, Sr(NO3)2, and Ba(NO3)2 were purchased from
Wako Pure Chemical Industries, Ltd. m-ZrO2 was supplied as a Japan Reference Catalyst
(JRC-ZRO-4) by the Reference Catalyst Division, Catalyst Society of Japan. In the present
study, the m-ZrO2 support was preheated in air at 800 oC for 3 h before modification with
o
MOs since m-ZrO2 calcined at 800 C showed the highest formation rate of 3B2OL from
2,3-BDO [19]. All the MOs were loaded on m-ZrO2 by incipient wetness impregnation in
the same manner as the preparation of CaO/ZrO2 [25]. The modified ZrO2 catalysts are
expressed as MO-X-T, where M is the loaded alkaline earth metal element, X is the MO
loading molar ratio as MO/ZrO2=X, and T is the calcination temperature (oC) after
impregnation. Additionally, perovskite compounds, MZrO3 with M=Ca, Sr, Ba, and also Mg,
were prepared by amorphous citrate process [33] using their nitrates and citric acid with the
molar ratio M/Zr/citric acid= 1/1/3 in vacuum at 100oC for 2 h and then heated at 180oC for
3 h and finally calcined in air at 800oC for 3 h.
2.2 Catalytic reaction
The catalytic reaction was conducted in a fixed-bed flow reactor under atmospheric
pressure of H2 carrier gas with a flow rate of 80 cm3 min-1 at a temperature of 300-400oC.
After the catalyst (1.0 g) bed had been heated at the prescribed reaction temperature for 1 h,
the reactant 2,3-BDO was fed through the reactor top at a liquid feed rate of 11.8 mmol h-1.
The liquid effluent was collected hourly at 0oC and analyzed by gas chromatography
(FID-GC-8A, Shimadzu, Japan) with a 60-m capillary column (InertCapWAX-HT). The
products were identified by gas chromatography with
a
mass spectrometer
6

(GCMS-QP5050A, Shimadzu, Japan) with a 30-m capillary column (DB-WAX). The
catalytic activity was evaluated by averaging the conversion and selectivity, defined as
mol%, data of the initial 5 h.
2.3 Characterization
XRD patterns were measured on New D8 ADVANCE (Bruker) using Cu Kα
radiation (λ= 0.154 nm) to detect the catalysts crystal structure. Structural features of
CaO/ZrO2 catalysts were also characterized by XPS (Axis Ultra DLD, Shimadzu Corp.)
using X-ray gun of Mg at 10 mA and 10 kV and pass energy of 40 eV. The peak tops of C1s
spectra for each samples were 282.8-282.9 eV. The peak top of C1s has been utilized as a
reference to adjust binding energies of other orbitals. It indicates that the binding energies of
peak tops of other orbitals such as O1s, Ca2p, and Zr3d can be fairly compared. The specific
surface area (SA) of catalysts was measured by the Brunauer-Emmett-Teller (BET) theory
using the N2 isotherm at -196◦C. The TPD of adsorbed CO2 was measured by using a
self-made apparatus to estimate the basicity of catalysts [34]. After the sample had been
preheated at 500◦C under vacuum, CO2 was adsorbed at 25oC for 72 h. The temperature of
the sample was raised from 25 to 850oC at a temperature-rising rate of 10oC min-1. The
amount of desorbed CO2, which was bubbled into an aqueous solution of NaOH, was
monitored by the change in the conductivity of the solution. A cumulative amount of
desorbed CO2 was obtained as a function of the desorbed temperature and then
differentiated to give a TPD profile as a basic strength distribution. The amount of basic
sites was estimated from the neutralization-titration curves of diluted NaOH. The
thermogravimetry-differential thermal analysis (TG-DTA) was performed using Thermoplus
8120E2 (Rigaku, Japan).
3. Results and Discussion
3.1. Catalytic activity of ZrO2 modified with alkaline earth metal oxides.
The catalytic reaction of 2,3-BDO was investigated over the ZrO2 modified with MOs
7

such as SrO, BaO, and also MgO at 350oC to compare with the CaO/ZrO2 catalyst. Figure 1
shows the effects of MO modifier contents on both 2,3-BDO conversion and 3B2OL
selectivity. This indicates that the effects of MO modifier contents resembled closely with
each other over CaO/ZrO2, SrO/ZrO2, BaO/ZrO2, and MgO/ZrO2: the 2,3-BDO conversion
decreased greatly from ZrO2 support with increasing their modifier contents, while the
3B2OL selectivity increased greatly only at a small amount of these modifiers and almost
unchanged at the large amounts. In the MgO/ZrO2 catalysts, however, only a low content of
MgO could act as a modifier in the similar way to the other MOs for the 3B2OL formation.
Besides, the excess MgO inhibits the catalytic activity of ZrO2: both the 2,3-BDO
conversion and the 3B2OL selectivity decreased at the ratio of MgO/ZrO2 over 0.2.
Therefore, due to the high yield of 3B2OL, we choose representative samples with the most
proper modifier contents such as CaO-0.0668, SrO-0.0906, BaO-0.0452, and MgO-0.0452
for the further tests.
Figure 2 shows the effects of calcination temperature on the catalytic activities of the
modified ZrO2 catalysts in the dehydration of 2,3-BDO at 350oC. In all the modifiers,
especially BaO, the 2,3-BDO conversion was maximized at a calcination temperature of
800oC. In contrast, the 3B2OL selectivity was maintained at high levels and became rarely
sensitive to the modifiers when the calcination temperature was higher than 600oC.
In our previous work [25], we investigated the present reaction over the catalysts such
as CaO-0.0668-800, SrO-0.0668-800, BaO-0.0668-800, and MgO-0.0668-800 at the same
molar MO contents. Among them, CaO-0.0668-800 was the most active catalyst. In the
present work, however, it was found that SrO-0.0906-800, BaO-0.0452-800, and
MgO-0.0452-800 showed the best catalytic performance in each modified-ZrO2 series. In
addition, BaO-0.0452-800 exceeded CaO-0.0668-800 in the 2,3-BDO conversion.
Figure 3 shows the effect of reaction temperature over the samples of CaO-0.0668-800,
SrO-0.0906-800, BaO-0.0452-800, and MgO-0.0452-800, which were calcined at 800oC.
The 2,3-BDO conversion increased greatly with increasing the reaction temperature while
the highest 3B2OL selectivity was obtained at 350oC. At 350oC, BaO-0.0452-800 showed
8

the same level of 3B2OL selectivity as CaO-0.0668-800 but the highest 2,3-BDO
conversion among all the tested catalysts.
In our previous work, it was found that 2,3-BDO conversion decreased slightly in the
initial 5 h of time on stream over CaO/ZrO2 [25] as well as ZrO2 catalysts [19] while we
have not checked their stabilities of catalyst for a long-term operation. Figure 4 shows the
catalytic activities of BaO-0.0452-800, CaO-0.0668-800, and m-ZrO2-800 catalysts for 25 h
of time on stream. The m-ZrO2-800 showed high conversion in the initial stage but
deactivated with time on stream. Decline in conversion of BaO-0.0452-800 and
CaO-0.0668-800 was also observed in a similar way to that of m-ZrO2-800. This indicates
that the deactivation should be resulted from the support itself. In contrast to the rapid
decline of 3B2OL selectivity (from 51.2% to 23.0%) over m-ZrO2-800, the 3B2OL
selectivity decreased gradually over BaO-0.0452-800 and CaO-0.0668-800 (from ca. 75% to
ca. 60%). It depicts that the MO modifiers enhanced not only the 3B2OL selectivity but also
the alleviation for the catalytic deactivation. In addition, contrary to the decline of the
3B2OL selectivity, the byproducts such as MEK, acetoin, MPA, and 2-methyl-propanol
increased. After 25 h of time on stream over m-ZrO2-800, the selectivities of MEK, acetoin,
and MPA increased from the initial 20.2, 4.9, and 5.4% to 35.5, 22.1, and 12.4%,
respectively. In contrast, the selectivities of MEK, acetoin, and MPA over BaO-0.0452-800
increased from 6.6, 3.4, and 3.3% to 11.7, 16.3, and 3.4%, respectively. This indicates that
BaO-0.0452-800 increases dehydrogenation ability to form acetoin during the gradual
deactivation. The reasons of catalytic deactivation are not clear: no clear changes were
observed in their crystal structures from the XRD patterns of these catalysts before and after
the reaction for 25 h. Besides, the deposited coke was observed less than 1.5 wt.% over the
modified ZrO2 and m-ZrO2-800 after being used for 25 h (TG-DTA data not shown). Thus,
coke should not be the main reason for deactivation.
In the ZrO2 modified with an appropriate content of MOs such as CaO, SrO, BaO, and
also MgO, the catalytic activities of m-ZrO2 were enhanced efficiently for the 3B2OL
formation. Their catalytic activities resembled closely with each other for their loaded
9

contents, calcination temperatures, reaction temperatures, and also stabilities with time on
stream (Figures 1-4). In addition, among the four modified catalysts, SrO/ZrO2 and
BaO/ZrO2 also acted well for both 2,3-BDO conversion and 3B2OL selectivity while those
of MgO/ZrO2 were always the lowest. In the following sections, in order to identify the key
factors for the enhancement of the present reaction, their structural and basic properties were
investigated.
3.2. Structural properties of ZrO2 modified with alkaline earth metal oxides.
Figure 5 shows XRD patterns of the typical samples with different calcination
temperatures and MO contents. In the samples, main diffraction peaks are attributed from
m-ZrO2 with several additional peaks of perovskite compounds such as CaZrO3, SrZrO3,
and BaZrO3. High MO contents and high calcination temperatures induced the formation of
the new perovskite phases. In addition, both the formation and the crystalline growth of
these new perovskite phases can be also observed in the order of BaZrO3 > SrZrO3 >
CaZrO3. In contrast to CaZrO3, BaZrO3, and SrZrO3 phases were detected even at low MO
contents and low calcination temperatures < 800oC. However, in MgO/ZrO2 samples, no
peaks assigned to MgZrO3 were detected, while the crystalline of m-ZrO2 phase grew
obviously with increasing both MgO content and calcination temperature. Thus, it is
reasonable that the 3B2OL selectivity was enhanced over MgO/ZrO2 due to the crystalline
growth of m-ZrO2, which is the most efficient ZrO2 crystal phase for the 3B2OL formation,
as described in our previous work [19]. Here, we have to check the catalytic activity of
perovskite compounds such as CaZrO3, SrZrO3, and BaZrO3.
In order to investigate the catalytic activities of the perovskite compounds, we
prepared CaZrO3, SrZrO3, and BaZrO3. As listed in Table 1, we succeeded in preparing
these perovskite compounds (XRD patterns in Figure 5b) but even together with a small
amount of SrCO3 and BaCO3, which are difficult to be removed during the preparation and
characterization processes in air. The results in the catalytic reaction emphasized that all the
perovskite compounds showed no catalytic activity for the formation of 3B2OL from
10

2,3-BDO: not only the extremely low 2,3-BDO conversion but also the low 3B2OL
selectivity were obtained. It can be understood that these perovskite compounds such as
CaZrO3, SrZrO3, and BaZrO3 detected in the modified m-ZrO2 could not provide the active
sites for the 3B2OL production.
Figure 6 shows the O1s, Ca2p, and Zr3d XPS spectra of CaO-0.0668-800,
CaO-0.226-800, and CaO-0.0668-600 samples together with CaO-800, ZrO2-800, and
CaZrO3-800. Firstly, from the O1s spectra of the samples, the peak position at 527.45 eV,
which is different from those of both CaO-800 (529.45 eV) and ZrO2-800 (527.83 eV), was
found to be the same as that from CaZrO3-800. In our previous work, Ca-O-Zr
hetero-linkages are also confirmed by the DRIFT spectra of CaO-X-800 samples [25]. Thus,
the new peak at 527.45 eV should be assigned to the O1s peak of Ca-O-Zr hetero-linkages,
which would attribute to the base structure of CaZrO3. The results also well agree with some
other works [35-42]. In addition, the XPS spectra of Ca2p and Zr3d also support the
Ca-O-Zr hetero-linkage formation: Ca2p1/2 shifted from 344.92 to 344.31 eV and that of
Ca2p3/2 from 348.55 to 348.03 eV, together with Zr3d3/2 shifting from 182.26 to 181.81 eV
and that of Zr3d5/2 from 179.92 to 179.47 eV. On the other hand, among the CaO/ZrO2
samples, CaO-0.0668-800 presents the highest percentage of Ca-O-Zr state via its O1s XPS
spectra, as well as the spectra of Ca2p and Zr3d. However, in comparison with
CaO-0.0668-800, the spectra of CaO-0.226-800 show lower percentages of Ca-O-Zr
hetero-linkages, but this contradicts with their XRD patterns (Figure 5a), in which stronger
peaks assigned to CaZrO3 phase were detected in CaO-0.226-800 rather than
CaO-0.0668-800. This indicates that, unlike CaO-0.226-800, Ca-O-Zr hetero-linkages are
highly dispersed on the surface of CaO-0.0668-800 at the appropriate CaO loading. In
CaO-0.226-800 with excess CaO content, most Ca-O-Zr hetero-linkages are gathered to
form CaZrO3 crystal phase, which is not active for the formation of 3B2OL from 2,3-BDO
o
(Table 1). At a low calcination temperature of 600 C, however, no proof of the Ca-O-Zr
hetero-linkage can be found from the CaO-0.0668-600 sample in its O1s, Ca2p or Zr3d XPS
spectra. In addition, the XRD pattern of CaO-0.0668-600 sample (Figure 5a), as well as the
11

samples calcined at lower than 800oC in our previous work [25], displayed no peaks
assigned to CaZrO3 phase. It indicates that high calcination temperature at 800oC or higher
is needed for the formation of Ca-O-Zr hetero-linkage and then for CaZrO3 phase.
Furthermore, some groups also reported that they succeeded in preparing CaZrO3 at high
temperatures over 800oC [37-41].
Figure 7 shows TG-DTA profiles of the alkaline earth metal nitrate loaded on m-ZrO2
samples prior to calcination. These alkaline earth metal nitrates have different difficulties in
decomposing with increasing calcination temperature, and the difficulty increases in the
order of Mg(NO3)2 < Ca(NO3)2 < Sr(NO3)2 < Ba(NO3)2. The decomposition has almost
finished at lower than 600oC (Figure 7a). The results are also well accordance with the O1s
XPS spectra shown in Figure 6, in which only the O1s XPS peaks attributed to CaO and
ZrO2 were detected from CaO-0.0668-600 sample and no O1s XPS peak attributed to
Ca(NO3)2 were observed. It is reasonable that these alkaline earth metal nitrate decomposed
at temperatures lower than 600oC for the formation of CaO, and then Ca2+ of CaO immerged
into ZrO2 to form Ca-O-Zr hetero-linkage at around 800oC, and that more Ca-O-Zr
hetero-linkages gathered and transformed to CaZrO3 phase at calcination temperatures
higher than 800oC. Judging from the XRD patterns, it is reasonable that the formation of
M-O-Zr hetero-linkage and the transformation to perovskite phases over SrO/ZrO2 and
BaO/ZrO2 samples should have occurred at lower temperatures than that of CaO/ZrO2.
However, their weight losses continued above 600oC even with a small amount, especially
the BaO-0.0452 sample seems to be further decomposed at higher than 1000oC (Figure 7b).
The weight loss after 600oC should be attributed to the decomposition of their carbonates,
which form easily in air environment containing CO2.
3.3 Basic properties of ZrO2 modified with alkaline earth metal oxides.
In our previous paper, the basic properties of CaO/ZrO2 have been reported in detail
[25]. Figure 8 shows the basic properties of MgO/ZrO2, SrO/ZrO2, and BaO/ZrO2
investigated by CO2-TPD. Among the samples, the CO2 desorption peaks attributed to weak
12

basic sites were detected as shifting slightly from 77oC to 115oC (peaks α and α’), and their
amount increased properly with increasing the MO contents. Besides, the desorption peak β
attributed to new medium basic sites was also found at around 350oC from SrO/ZrO2 and
BaO/ZrO2 in the same way as CaO/ZrO2 [25]. The peaks γ at temperatures higher than
600oC were detected over SrO/ZrO2 and BaO/ZrO2 which could be resulted from the CO2
desorbed from the isolated SrO and BaO. The peaks higher than 800oC could be resulted
o
from the decomposition of their carbonates. In the actual reactions performed at 350 C, the
peak γ of the adsorption sites, which strongly adsorb CO2, could not be attributed to the
reaction.
On the other hand, new peaks generated on MgO/ZrO2 samples were detected at
around 250oC, such changes on its basic property must be related to the structural changes as
described in the previous section of XRD (Figure 5a). The desorption peaks at 335oC are
resulted from the excess MgO and it should be the reason for the decrease in the catalytic
activity of MgO/ZrO2 catalysts in Figure 1. The CO2-TPD profiles of MgO/ZrO2 resemble
closely to the reports elsewhere [43-45].
The intensity of peak β varies with increasing the MO contents. Especially, at molar
ratio MgO/ZrO2 = 0.0226, SrO/ZrO2 = 0.0906, and BaO/ZrO2 = 0.0452 as well as CaO/ZrO2
= 0.0668 in the previous report [25], the largest amount of basic sites β were detected. The
samples calcined at 800oC show the largest amount of basic sites β, but the amount
decreased with increasing calcination temperature. Of course, the catalysts with the largest
amount of basic sites β were also confirmed as the most active ones for high 2,3-BDO
conversion and high 3B2OL selectivity. The results are also well accordance with the report
of CaO/ZrO2 catalysts [25].
Figure 9 shows CO2-TPD profiles of the perovskite compounds of MZrO3 (M= Ca, Sr,
and Ba), which showed no good catalytic activities for the 3B2OL formation from 2,3-BDO,
together with the MgO-ZrO2 composed of tetragonal ZrO2 (Table 1). Over the perovskite
compounds of MZrO3, the decomposition peaks resulted from their carbonates were mainly
observed but rare basic sites appeared. Over the MgO-ZrO2 i.e. tetragonal ZrO2, large
13

desorption peaks attributed to new basic sites with medium strength were observed at
around 250oC. As reported in our previous work [19, 25], tetragonal ZrO2 was not suitable
for the formation of BD and 3B2OL from 2,3-BDO. In the present study, however, the new
medium-strength basic sites would have contributed to the formation of 3B2OL even with a
low reactant conversion.
Although the perovskite compounds of MZrO3 have no active basic sites β, MZrO3
phase was detected in the XRD patterns of modified m-ZrO2 catalysts which showed
catalytic activities for the present reaction. However, among the CaO/ZrO2 catalysts,
CaO-0.0668-800 has only weak XRD peaks attributed to CaZrO3, but it has the most
percentages of O1s XPS peak assigned to the Ca-O-Zr hetero-linkages highly dispersed on
its surface. The hetero-linkages could attribute to the most amount of basic sites β, and also
enhance the 3B2OL formation from 2,3-BDO. As discussed in the research of Zhang’s group
[46], the uneven distribution in charges of the Ca-O-Zr hetero-linkage generates new acidic
and basic sites, which is different from those of the pure CaO and ZrO2. In the XPS spectra
of the samples containing Ca-O-Zr hetero-linkage (Figure 6), it was found that not only the
O1s, but also their Ca2p and Zr3d spectra shifted to lower energy positions as compared
with those of pure CaO and ZrO2. Thus, the bond distances of Ca-O-Zr should be different
from those of pure CaO and ZrO2. Since we found that the formation rate of some
unsaturated alcohols from diols are sensitive to the ionic radii of cations in the rare earth
metal oxide catalysts [6,24], the distance between these active sites imaged as our work [25]
should also be adjusted suitably by the highly dispersed hetero-linkages for the present
reaction. In addition, the O1s spectra shifting to lower energy positions would contribute to
the new medium basic sites β, which we speculated as the O2- from Ca-O-Zr hetero-linkage.
Since the O2- of CaO has stronger base than that of ZrO2, the O2- from Ca-O-Zr results in
weaker medium basic sites β than the O2- of CaO, but still stronger than that of ZrO2. On the
other hand, their Ca2p and Zr3d spectra shifting to lower energy positions must also adjust
the acidic sites balance which should be much more suitable for the present reaction. We
have proposed that the formation of 3-buten-2-ol proceeds via base-acid concerted
14

mechanism (Scheme 1 of [25]). The present MO/m-ZrO2 catalysts must efficiently work
with the basic modifiers in a fine base-acid balance for the formation of 3B2OL from
2,3-BDO. Since Zhang groups detected the changes in acidity of CaO/ZrO2 via NH3-TPD
[46], we will perform further work to identify these acid sites of the modified ZrO2 catalysts.
4. Conclusions
Vapor-phase catalytic dehydration of 2,3-BDO to 3B2OL was investigated over
several modified m-ZrO2 catalysts such as MgO/m-ZrO2, SrO/m-ZrO2, and BaO/m-ZrO2 to
compare with the previously reported CaO/m-ZrO2. In the dehydration of 2,3-BDO, it was
found that 3-buten-2-ol was preferentially formed over BaO/ZrO2 and SrO/ZrO2 as well as
CaO/ZrO2. Especially, BaO-0.0452-800 showed the highest 2,3-BDO conversion of 72.4%
and the 3B2OL selectivity of 74.4% in the initial stage of 5 h at 350oC. However, all these
catalysts declined the catalytic activity in a long-term operation of 25 h. Compared with the
rapid decline of 3B2OL formation over m-ZrO2-800 support, the decline over m-ZrO2
modified with alkaline earth metal oxides, BaO-0.0452-800 and CaO-0.0668, was gradual.
XRD and XPS studies elucidated the active structure of modified ZrO2 catalysts.
M-O-Zr (M= Ca, Sr, and Ba) hetero-linkages were highly dispersed on the surface by
loading these MOs onto m-ZrO2 with an appropriate content and then calcining at 800oC.
Perovskite compounds such as CaZrO3, which was newly produced in the catalysts, was
inactive for the reaction. It can be concluded that the highly dispersed M-O-Zr (M= Ca, Sr,
and Ba) hetero-linkages on the surface is formed to adjust the appropriate distance of these
active sites and also to generate the proper base-acid balance for the efficient formation of
3B2OL from 2,3-BDO.
Acknowledgements
This research was partly supported by JST, Strategic International Collaborative Research
Program, SICORP.
15

References
[1] S. Lee, B. Kim, K. Park, Y. Um, J. Lee, Appl. Biochem. Biotechnol. 166 (2012) 1801-1813.
[2] X. Ji, H. Huang, P. Ouyang, Biotechnol. Adv. 29 (2011) 351-364.
[3] A. Zeng, W.Sabra, Curr. Opin. Biotechnol. 22 (2011) 749-757.
[4] P. Lanzafame, G. Centi, S. Perathoner, Chem. Soc. Rev. 43 (2014) 7561-7580.
[5] E.V. Makshina, M. Dusselier, W. Janssens, J. Degre’ve, P.A. Jacobs, B.F. Sels, Chem. Soc. Rev.
43 (2014) 7917-7953.
[6] H. Duan, Y. Yamada, S. Sato, Chem. Lett. 45 (2016) 1036-1047.
[7] A.N. Bouns, R.V.V. Nicholls, Can. J. Res. 25B (1947) 80-89.
[8] R.R. Emerson, M.C. Flickinger, G.T. Tsao, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 473-477.
[9] A.C. Neish, V.C. Haskell, F.J. MacDonald, Can. J. Res. 23B (1945) 281-289.
[10] X. Liu, V. Fabos, S. Taylor, D.W. Knight, K. Whiston; G.J. Hutchings, Chem. Eur. J. 22 (2016)
12290-12294.
[11] A.V. Tran, R.P. Chambers, Biotechnol. Bioeng. 29 (1987) 343-351.
[12] B. Török, I. Pálinkó, Á. Molnár, M. Rózsa-Tarjáni, J. Mol. Struct. 482–483 (1999) 329-332.
[13] M.A. Nikitina, V.L. Sushkevich, I.I. Ivanova, Petrol. Chem. 56 (2016) 230-236.
[14] C.P. Bezouhanova, F.A. Jabur, J. Mol. Catal. 87 (1994) 39-46.
[15] A. Multer, N. McGraw, K. Hohn, P. Vadlani, Ind. Eng. Chem. Res. 52 (2013) 56-60.
[16] W. Zhang, D. Yu, X. Jia, H. Huang, Green Chem. 14 (2012) 3441-3450.
[17] J. Zhao, D. Yu, W. Zhang, Y. Hu, T. Jiang, J. Fu, H. Huang, RSC Adv. 6 (2016) 16988-16995.
[18] M.A. Nikitina, I.I. Ivanova, ChemCatChem. 8 (2016) 1346-1353.
[19] H. Duan, D. Sun, Y. Yamada, S. Sato, Catal. Commun. 48 (2014) 1-4.
[20] H. Duan, Y. Yamada, S. Sato, Appl. Catal. A: Gen. 491 (2015) 163-169.
[21] M.E. Winfield, J. Council Sci. Ind. Res. 18 (1945) 412-423.
[22] S. Sato, R. Takahashi, M. Kobune, H. Gotoh, Appl. Catal. A: Gen. 356 (2009) 57-63.
[23] A. Auroux, A. Gervasini, J. Phys. Chem. 94 (1990) 6371-6379.
[24] H. Duan, Y. Yamada, S. Sato, Chem. Lett. 43 (2014) 1773-1775.
[25] H. Duan, Y. Yamada, S. Sato, Appl. Catal. A: Gen. 487 (2014) 226-233.
16

[26] W. Kim, W. Shin, K. Lee, H. Song, H. Kim, D. Seung, I. Filimonov, Appl. Catal. A: Gen. 511
(2016) 156-167.
[27] D. Tsukamoto, S. Sakami, M. Ito, K. Yamada, T. Yonehara, Chem. Lett. 45 (2016) 831-833.
[28] T. Kim, J. Baek, C. Song, Y. Yun, D. Park, W. Kim, J. Han, J. Yi, J. Catal. 323 (2015) 85-99.
[29] N. Ichikawa, S. Sato, R. Takahashi, T. Sodesawa, J. Mol. Catal. A: Chem. 256 (2006) 106-112.
[30] E.V. Makshina, M. Dusselier, W. Janssens, J. Degre`ve, P.A. Jacobsa, B.F. Sels, Chem. Soc. Rev.
43 (2014) 7917-7953.
[31] J.W. Bode, E.M. Carreira, J. Org. Chem. 66 (2001) 6410-6424.
[32] J.W. Bode, E.M. Carreira, J. Am. Chem. Soc. 123 (2001) 3611-3612.
[33] D.J. Anderton, F.R. Sale, Powder Metallurgy 22 (1979) 14-21.
[34] S. Sato, K. Koizumi, F. Nozaki, J. Catal. 178 (1998) 264-274.
[35] R. Koirala, K.R. Gunugunuri, S.E. Pratsinis, P.G. Smirniotis, J. Phys. Chem. C 115 (2011)
24804-24812.
[36] H. Guo, S. Wang, C. Li, Y. Zhao, Q. Sun, X. Ma, Ind. Eng. Chem. Res. 55 (2016) 7873-7879.
[37] G. Cabello, L. Lillo, C. Caro, G.E. Buono-Core, B. Chornik, M. Flores, C. Carrasco, C.A.
Rodriguezd, Ceram.Int. 40 (2014) 7761-7768.
[38] X.Y. Qiu, H. W. Liu, F. Fang, M.J. Ha, X.H. Zhou, J.M. Liu, Appl. Phys. A 81 (2005)
1431-1434.
[39] H. Wang, S. Liu, W. Zhang, N. Zhao, W. Wei, Y. Sun, Acta Chim. Sinica 64 (2006) 2409-2413.
[40] A.Y. Teterin, K.I. Maslakov, Y.A. Teterin, K.E. Ivanov, S.V. Yudintsev, S.V. Stefanovsky, T.S.
Livshits, M.I. Lapina, Nucl. Technol. Radiat. 25 (2010) 157-163.
[41] J. Chen, S. Wu, F. Zhang, S. Lü, Y. Mao, Mater. Chem. Phys. 172 (2016) 87-97
[42] S. Yan, M. Kim, S. Mohan, S.O. Salley, K.Y. Simon Ng, Appl. Catal. A: Gen. 373 (2010)
104-111.
[43] Y. Yanagisawa, K. Takaoka, S. Yamabe, T. Ito, J. Phys. Chem. 99 (1995) 3704-3710.
[44] J.I. Di Cosimo, V.K. Díez, C. Ferretti, C.R. Apesteguía, Catalysis 26 (2014) 1-28.
[45] S. Liu, X. Zhang, J. Li, N. Zhao, W. Wei, Y. Sun, Catal. Commun. 9 (2008) 1527-1532.
[46] Q. Zhang, Y. Zhang, H. Li, C. Gao, Y. Zhao, Appl. Catal. A: Gen. 466 (2013) 233-239.
17

Figure 1 Effects of alkaline earth metal oxide contents on the catalytic activities of modified ZrO2
catalysts in the dehydration of 2,3-BDO to 3B2OL at 350oC. All the catalysts were calcined at 800oC.
Figure 2 Effects of calcination temperature on the catalytic activities of modified ZrO2 catalysts in the
dehydration of 2,3-BDO to 3B2OL at 350oC.
18

Figure 3 The catalytic activities of modified ZrO2 catalysts in the dehydration of 2,3-BDO to 3B2OL as a
function of reaction temperature.
Figure 4 Catalytic stability of modified ZrO2 catalysts with time on stream at 350oC.
19

Figure 5 (a) XRD patterns of modified ZrO2 catalysts: effects of calcination temperature and the alkaline
earth metal oxide contents on their crystal structures.
20

Figure 5 (b) XRD patterns of perovskite compounds (MZrO3) aimed samples: the names in black are their
aimed compositions, and the terms in red and blue are the phases detected from these samples.
21

Figure 6 O1s, Ca2p, and Zr3d XPS spectra of CaO-800, ZrO2-800, CaZrO3-800, CaO-0.0668-800,
CaO-0.226-800, and CaO-0.0668-600.
22

Figure 7 TG-DTA profiles of the alkaline earth metal nitrate loaded on m-ZrO2 samples.
23

Figure 8 Effects of calcination temperature and alkaline earth metal oxide contents on the basic properties
(CO2-TPD) of the modified catalysts.
Figure 9 CO2-TPD profiles of the perovskite compounds shown in Fig. 5(b), MgO-ZrO2 and m-ZrO2.
24

Scheme 1 Possible chemical compounds from 2,3-BDO catalyzed by acid and base catalysts.
25

Table 1 Reaction activities of perovskite compounds (MZrO3), MgO-ZrO2 and m-ZrO2 in the dehydration
of 2,3-butanediol at 350oC.
Catalyst
S.A.
Crystal phase
Conversion
(mol%)
1.7
Selectivity (mol%)
(1.0 g) (m2 g-1)
3B2OL
7.5
MEK
12.9
Acetoin
47.2
Others
32.4
Blank test
16.5
Tetragonal
ZrO2
16.4
52.6
9.4
20.6
17.4
MgO-ZrO2
CaO-ZrO2
7.4
< 1
CaZrO3
4.5
3.3
0
11.8
11.0
64.9
76.0
23.3
5.9
SrZrO3+SrCO3
7.1
SrO-ZrO2
BaO-ZrO2
m-ZrO2
< 1
BaZrO3+BaCO3
0.9
5.9
13.6
22.7
69.8
1.7
10.7
31.8
26.3
Monoclinic
ZrO2
97.6
43.9
All the catalysts were calcined at 800oC during the preparation procedure.
26