The role of steam in selective oxidation of methacrolein over H3PMo12O40

Article abstract of DOI:10.1016/j.apcata.2018.11.007

Role of steam in selective oxidation of methacrolein with molecular oxygen over H₃PMo₁₂O₄₀ catalyst was investigated. Addition of steam to feed gas significantly enhanced both catalytic activity and selectivity to methacrylic acid, which were fivefold and twice increases, respectively, under the optimal steam pressure (P_H2O = 0.13 atm). Kinetic analysis demonstrated that the addition of steam caused 200-fold increase in the pre-exponential factor for the formation of methacrylic acid, leading to the significant increase in the activity. The steam in the feed gas varied hydrous state of H₃PMo₁₂O₄₀ under the reaction conditions, while did not alter redox property, molecular and crystalline structures, and surface area of the catalyst. In the presence of steam at 573 K, three H₂O per one H₃PMo₁₂O₄₀ were absorbed and hydrated protons like [H₃O]⁺ were formed in the bulk of H₃PMo₁₂O₄₀. Methacrolein was adsorbed on the surface of the hydrous catalyst, but not on anhydrous one at all. Based on the results, it was concluded that activation of methacrolein readily occurred on the catalyst in the presence of steam, leading to the significant increase in the pre–exponential factor. Quantum chemical calculation supported the smooth activation of methacrolein by the reaction with [H₃O]⁺ without any transition state.

Full text of DOI:10.1016/j.apcata.2018.11.007

Accepted Manuscript
Title: The role of steam in selective oxidation of methacrolein
over H PMo O
3
12 40
Authors: Shuhei Yasuda, Jun Hirata, Mitsuru Kanno, Wataru
Ninomiya, Ryoichi Otomo, Yuichi Kamiya
PII:
S0926-860X(18)30554-4
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.11.007
APCATA 16876
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
2 October 2018
31 October 2018
8 November 2018
Please cite this article as: Yasuda S, Hirata J, Kanno M, Ninomiya W, Otomo R, Kamiya
Y, The role of steam in selective oxidation of methacrolein over H PMo O , Applied
3
12 40
Catalysis A, General (2018), https://doi.org/10.1016/j.apcata.2018.11.007
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.

Revised manuscript: APCATA-D-18-01947
The role of steam in selective oxidation of methacrolein over H
3
PMo12
O
40
a
b
b
b
c
Shuhei Yasuda , Jun Hirata , Mitsuru Kanno , Wataru Ninomiya , Ryoichi Otomo , Yuichi
Kamiyac,*
a
Graduate School of Environmental Science, Hokkaido University, Kita 10 Nishi 5, Sapporo 060-
0810, Japan
b
Otake R&D Center, Mitsubishi Chemical Corporation, 20-1, Miyuki-cho, Otake-shi, Hiroshima
739-0693, Japan
c
Faculty of Environmental Earth Science, Hokkaido University, Kita 10 Nishi 5, Sapporo 060-
0810, Japan
Corresponding author:
Prof. Dr. Yuichi Kamiya
TEL: +81-11-706-2217, E-mail: kamiya@ees.hokudai.ac.jp
1

Graphical abstract
Highlights


Addition of steam to feed gas significantly enhanced activity and selectivity
Three H O per one Keggin unit were absorbed into the bulk of H
presence of steam
Increase in pre-exponential factor for methacrylic acid formation with steam due
to smooth activation of methacrolein on H
2
3
PMo12O40 in the

3
PMo12
O
40
Abstract
Role of steam in selective oxidation of methacrolein with molecular oxygen over
3
H PMo12O40 catalyst was investigated. Addition of steam to feed gas significantly enhanced both
catalytic activity and selectivity to methacrylic acid, which were fivefold and twice increases,
respectively, under the optimal steam pressure (PH2O = 0.13 atm). Kinetic analysis demonstrated
that the addition of steam caused 200-fold increase in the pre-exponential factor for the formation
2

of methacrylic acid, leading to the significant increase in the activity. The steam in the feed gas
varied hydrous state of H PMo12O40 under the reaction conditions, while did not alter redox
3
property, molecular and crystalline structures, and surface area of the catalyst. In the presence of
O]+
steam at 573 K, three H
2
O per one H
3
PMo12
O
40 were absorbed and hydrated protons like [H
3
were formed in the bulk of H PMo12O40. Methacrolein was adsorbed on the surface of the hydrous
3
catalyst, but not on anhydrous one at all. Based on the results, it was concluded that activation of
methacrolein readily occurred on the catalyst in the presence of steam, leading to the significant
increase in the pre–exponential factor. Quantum chemical calculation supported the smooth
+
activation of methacrolein by the reaction with [H
3
O] without any transition state.
Keywords: methacrolein oxidation, steam addition, kinetic analysis, pre-exponential factor,
methacrolein adsorption
1. Introduction
Polyoxometalate (POM) is a metal–oxygen cluster anion composed of metal–oxygen
tetrahedron XO
4
and octahedron YO as the structural building units, which are connected with
6
sharing the corner, edge, and face. Keggin-type POM (Keggin–POM) has the composition of
40]n–, in which the twelve octahedrons comprise four groups of three edge-shared
[
XY12O
octahedrons Y
3
O
13 surrounding the central XO tetrahedron. Keggin–POMs have attracted much
4
3

attention in applications as catalysts because of simple preparation procedure, high thermal
40]3– is one of the members in
stability, strong acidity, and high redox ability [1–10]. [PMo12
O
Keggin–POM family. H PMo12O40 as well as its salts is known to show excellent catalytic
3
performance for selective oxidation of various compounds such as saturated [1] and unsaturated
hydrocarbons [2], alcohols [1], and aldehydes [1, 2] due to high redox ability [10] in the form of
homogeneous as well as heterogeneous catalyst.
Solid Keggin–POM is an ionic crystal composed of [XY12O
40]n– and cations. The
structure of the polyanion itself, i.e. metal–oxygen cluster anion, in the solid Keggin–POM is
called a primary structure. A secondary structure is the three–dimensional arrangement of
polyanions, cations, and additional molecules like water of crystallization, forming a crystalline
structure. An unique character of solid Keggin–POMs is that small polar molecules like water
penetrate into the space between the polyanions, namely the molecules are absorbed into the bulk
+
+
of the solid, when solid Keggin–POMs have small cations like H and Li . In fact, water [11] and
lower alcohols [12] are absorbed into the bulk of solid Keggin–POMs at room temperature and
the number of absorbed molecules is more than 100 times larger than that calculated on the
assumption that adsorption occurs only on the outer surface [12]. It is known that the absorption
of polar small molecules affects catalytic performance of solid Keggin–POMs for reactions in
4

which those molecules participate as reactants [13–18], because the reaction occurs in the bulk as
well as on the surface of the catalyst particle.
In addition to the aforementioned reactions, there is another type of reactions over solid-
Keggin POMs that are enhanced by the presence of small polar molecules like water though those
molecules are not found in chemical reaction formulae. An exemplary reaction is gas-phase
oxidation of methacrolein (MAL) into methacrylic acid (MAA) over H PMo12O40 (eq. 1).
3
H
OH
1
2
+
O2
(eq. 1)
O
O
Misono et al. demonstrated that steam added in the feed gas accelerates the formation of MAA
[
[
19, 20]. In fact, both MAL consumption rate and selectivity to MAA were increased with steam
19, 20]. Based on the results of water vapor pressure-dependence and of transient response in
cessation of steam as well as of O supply, they proposed a reaction mechanism involving water
2
as follows (Scheme 1) [19,20]: In the first step, 1,1-diol is formed by the reaction of MAL with
water, which is promoted by Brønsted acids on H PMo12O40 and successively molybdate ester is
3
formed. Finally, it transforms into MAA. In this mechanism, the formation of 1,1-diol is a key for
the selective formation of MAA.
5

Another plausible role of steam is to suppress thermal decomposition of the catalyst (eq.
), because the decomposition occurs at around the reaction temperature for MAL oxidation [21].
2
H
3
PMo12
O40 → 0.5P O
2 5
+ 12MoO
3
2
+ 1.5H O
(eq. 2)
As mentioned above, solid Keggin–POM absorbs small polar molecules like water into
+,
its bulk at room temperature. Water absorbed in the bulk of H
3
PMo12
O
40 form dimeric H
5
O
2
giving [H
5
O
2
+]
3
[PMo12
O
40]3–·nH
O, under wet atmosphere at room temperature. Water in
2
[
H O
5 2
+]
3
[PMo12
O
40]3–·nH
O is released in a step-by-step manner with increase in temperature and
2
3
anhydrous H PMo12O40 is formed at about 433 K under dry atmosphere [22]. Since the reaction
temperature for MAL oxidation over H PMo12O40 is around 573 K, the catalyst must exist as an
3
anhydrous form if steam is not present in the feed gas. However, steam is normally added to the
feed gas to promote the reaction. Thus, H PMo12O40 could absorbs some water in its bulk under
3
the reaction conditions and consequently hydrous state could have an impact on the catalytic
performance as Misono et al. mentioned that steam may affect the state of the “pseudo-liquid
phase” without any direct evidence [19]. However, there is no study on quantitative evaluation of
the number of absorbed water in the presence of steam at high temperature and it is unclear how
the absorbed water in the bulk participate in the reaction over H
3
PMo12O .
40
6

In the present study, to give a new insight into the role of steam in MAL oxidation over
3
H PMo12O40, firstly we conducted kinetic study for the reaction and compared the kinetic
parameters (apparent activation energy and pre–exponential factor) in the presence of steam to
those in the absence of steam. In addition, the number of water absorbed in H PMo12O40 under
3
the atmosphere with steam (wet conditions) was quantified at 573 K. Sorption amounts of MAL
in dry and wet conditions were also evaluated. Finally, we propose a role of steam for the reaction
based on the results of the kinetic study, absorption behavior of MAL under the wet conditions as
well as the changes in the chemical and physical properties of the catalyst under the dry and wet
conditions.
2. Experimental
2.1 Material
Hydrous 12–molybdophosphoric acid (H
3
PMo12
O
40·nH O) was supplied by Nippon
2
Inorganic Colour and Chemical Co., Ltd. H
3
PMo12
O
40·nH O was extracted from the aqueous
2
3
solution of H PMo12
O
40 with diethyl ether. Finally, H
3
PMo12
O
40·nH O was recrystallized in the
2
aqueous solution, which was prepared with the extracted H
3
PMo12
O
40·nH O, and dried overnight
2
at 333 K.
7

Impurities in MAL (Tokyo Chem. Ind. Co., Ltd.) were removed by extraction with water.
The purified MAL was put in a freezer to remove water as ice and stored with a small amount of
hydroquinone (Wako Pure Chem. Ind.).
2.2 Catalytic oxidation of methacrolein
Catalytic oxidation of MAL was performed in a continuous flow reactor at atmospheric
pressure. After pretreatment of the catalyst at 593 K with a mixture of O
2
(0.11 atm), H O (0.18
2
–
1
atm), and N (balance) at a total flow rate of 28 mL min for 1 h, the temperature was decreased
2
to 573 K and the mixed gas of MAL (0.03 atm), O
2
(0.06 atm), H
2
O (0.13 atm), and N (balance)
2
at a total flow rate of 72 mL min–1 was fed into the reactor to start the catalytic reaction. The
reaction products were analyzed by using on-line gas chromatographs. Acetic acid, MAL, and
MAA were analyzed by a GC (Shimadzu GC-14B) equipped with a flame ionization detector and
a capillary column (GL Sciences TC-FFAP, 0.25 mm, 50 m). For CO and CO , the other GC
2
(
Shimadzu GC-8A) equipped with a thermal conductivity detector (TCD) and two packed
columns (Molecular Sieve 5A, 2.85 mm, 3 m and Activated Carbon, 2.85 mm, 2 m) was used. To
prevent interference from organic compounds to analysis with GC-TCD, the effluent gas at outlet
of the reactor was passed through a dry-ice trap and then was analyzed by the GC. As an internal
standard for GC analysis, CH (31 vol.%) diluted with He was mixed at the outlet of the reactor.
4
8

The carbon balance was calculated with eq. 3 and was in the range of 97–103 % for whole the
reactions.
Tꢀtꢁꢂ fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf ꢁꢂꢂ pꢃꢀduꢆts
Carbon balance (%) =
× 100
(eq. 3)
Fꢅꢅd ꢃꢁtꢅ ꢀf MAL
Yield of and selectivity to each product were calculated based on carbon basis with the following
equations.
Fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf MAA
Yꢇelꢈ oꢉ ꢊꢋꢋ (%) =
× 100
(eq. 4)
(eq. 5)
(eq. 6)
(eq. 7)
(eq. 8)
(eq. 9)
Fꢅꢅd ꢃꢁtꢅ ꢀf MAL
(
Fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf ꢁꢆꢅtiꢆ ꢁꢆid)×2
Fꢅꢅd ꢃꢁtꢅ ꢀf MAL
Yꢇelꢈ oꢉ aceꢌꢇc acꢇꢈ (%) =
× 100
(
Fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf ꢍꢎ )×4
ꢏ
Yꢇelꢈ oꢉ CO (%) =
× 100
푥
Fꢅꢅd ꢃꢁtꢅ ꢀf MAL
Fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf MAA
Selecꢌꢇvꢇꢌy ꢌo ꢊꢋꢋ (%) = Rꢁtꢅ ꢀf MAL ꢆꢀꢄsumꢅd × 100
(
Fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf ꢁꢆꢅtiꢆ ꢁꢆid)×2
Rꢁtꢅ ꢀf MAL ꢆꢀꢄsumꢅd
Selecꢌꢇvꢇꢌy ꢌo aceꢌꢇc acꢇꢈ (%) =
× 100
(
Fꢀꢃmꢁtiꢀꢄ ꢃꢁtꢅ ꢀf ꢍꢎ )×4
ꢏ
Selecꢌꢇvꢇꢌy ꢌo CO (%) =
× 100
푥
Rꢁtꢅ ꢀf MAL ꢆꢀꢄsumꢅd
To observe transient response of the catalytic performance, the feed gas was promptly
switched from one to another, while the reaction temperature and the total flow rate were kept
constant.
9

2.3 Characterization
Specific surface area was estimated by the Brunauer–Emmett–Teller (BET) equation
applied to adsorption isotherm of nitrogen, which was acquired on a Belsorp-28SA instrument
(
Bel Japan Inc.) at 77 K. Before measurement, the sample was evacuated at 473 K for 6 h.
Average valence of Mo in the catalyst bulk was determined by redox titration using
KMnO solution. At a predetermined time point during the catalytic reaction, the feed gas was
4
switched to nitrogen and the catalyst was rapidly cooled to room temperature. A part of the
obtained catalyst (0.30 g) was dissolved in deionized water (250 mL). To a part of the solution
(
75 mL), 10 mL of 47 % H
2
SO (Wako Pure Chem. Ind.) and 20 mL of deionized water were
4
added. The solution was titrated with an aqueous solution of KMnO (Wako Pure Chem. Ind., 5
4
5+
–
mM). Mo in the solution was reacted with MnO
4
in an acidic aqueous solution in accordance
with the following equation.
5
+
–
+
6+
2+
5
Mo + MnO
4
+ 8H → 5Mo + Mn + 4H
2
O
(eq. 10)
n+
The average valence of Mo (Mo ) in the catalyst bulk was calculated by using the following
equation;
Amꢀuꢄt ꢀf Mꢀ5ꢐ iꢄ thꢅ sꢀꢂutiꢀꢄ (ꢃꢅdꢀx titꢃꢁtiꢀꢄ)
ꢊoꢄ+ = 6 −
(eq. 11)
Tꢀtꢁꢂ ꢁmꢀuꢄt ꢀf Mꢀ iꢄ thꢅ sꢀꢂutiꢀꢄ (IꢍPꢑAEꢒ)
1
0

,
in which the total amount of Mo in the solution was determined by ICP-AES analysis using a
Shimadzu ICPE-9000 spectrometer.
Oxidation state of Mo on the catalyst surface was analyzed by X-ray photoelectron
spectroscopy (XPS) using a ULVAC-PHI Quantera II (Al Kα radiation). Binding energy was
calibrated with respect to C 1s peak of a carbon tape at 284.6 eV. Thermogravimetric (TG)
analysis of H
3
PMo12
O
40·nH O was conducted under He flow on a Rigaku Thermo plus TG8120.
2
–
1
Temperature of the sample was increased at a rate of 5 K min .
Sorption of water was carried out in a continuous flow reactor (BEL-TPD, Bel Japan
Inc.) equipped with a quadrupole mass spectrometer (ANELVA Co., M-QA100F). The catalyst
(
0.5 g) was pretreated in the reactor at 573 K under a flow of He containing steam (PH2O = 0.13
–
1
atm) at a total flow rate of 40 mL min for 1 h with monitoring the concentration of steam (m/e
=
18) at the outlet of the reactor by using the mass spectrometer. After the signal of m/e =18
became constant, the gas was switched to dry He. The signal intensity was decreased and finally
became almost zero. A blank experiment with an empty reactor was conducted. The difference in
the integrated intensities between with the catalyst and with the empty reactor corresponds to the
amount of water sorbed in the catalyst under the conditions at PH2O = 0.13 atm and 573 K.
11

Sorption of MAL was carried out using the same apparatus as that for the water sorption
experiment. The catalyst (0.5 g) was pretreated in the reactor at 573 K under a flow of He
–
1
containing steam (PH2O = 0.13 atm) at a total flow rate of 40 mL min for 1 h and the temperature
was decreased to 433K while flowing the gas. Then sorption of MAL was conducted at 433 K by
injecting MAL at regular interval with monitoring the mass signals of m/e = 70 (MAL) and 18
(
water). The reason for conducting the sorption experiment at 433 K was to prevent oxidation of
MAL by the lattice oxygen of the catalyst. The pulse size of each MAL injection was 2.6 μmol
and injections were continued up to 12 times. The sorption experiment with the catalyst pretreated
under dry He at 573 K was also conducted as a comparison.
IR spectra were taken at 573 K under dry He or H O/He (PH2O = 0.13 atm) on a FT/IR–
2
2
30 spectrometer (JASCO Co., Japan). The aqueous solution of H PMo12O40 was put on a Si
3
wafer (2 cm in diameter). The solution on the wafer was dried up at 333 K and the resulting wafer
was applied to IR measurement. An IR spectrum was measured at 573 K with a flow of dry He
–1
(
15 mL min ). Then the flowing gas was switched to wet He and IR spectra were measured, while
keeping the temperature at 573 K.
3. Results and Discussion
3.1 Catalytic performances in the absence and presence of steam
1
2

Figure 1 shows dependence of the catalytic performance of H PMo12O40 on the partial
3
pressure of water vapor. Main products were MAA, acetic acid, and CO regardless of presence
x
or absence of steam. Only little CO was formed under the reaction conditions and thus CO was
2
dominant. The reaction proceeded even in the absence of steam (PH2O = 0 atm). However, the
addition of steam to the feed gas greatly enhanced the reaction rates, which is consistent with the
previous papers [19, 20]. The yields of every product increased with the increase in PH2O, but
degree of the improvement differed from one product to another. The improvement in the yield
of MAA by steam was the most prominent of all the products. Thus, the selectivity to MAA
increased with increase in PH2O (Fig. 1(b)), and inversely those to acetic acid and CO decreased.
x
Since selectivities became almost constant at PH2O = 0.13 atm or more, we decided that PH2O
=
0.13 atm was the standard partial pressure of steam for following experiments.
As Fig. 1 clearly shows, the yield of acetic acid was almost the same as that of CO
x
at
any PH2O including PH2O = 0 atm. In addition, as will be described later, apparent activation
energies for the formations of acetic acid and CO were the almost same each other for any PH2O.
x
These results suggest that one molecule of acetic acid was formed from one molecule of MAL
accompanying with the formation of two molecules of CO
from MAL is as that shown in Scheme 2.
x
. Thus, the reaction network starting
1
3

We further investigated reaction kinetics in the presence and absence of steam for deeper
understanding of the influence of steam on the reaction. In order to suppress the conversion of
MAL to 20% or less, which corresponds to differential reaction conditions, catalyst weight to flow
–
1
rate (W/F) were adjusted to 69 and 173 g–cat h mol for the reactions in the presence and absence
of steam, respectively. All the data were taken under the steady state conditions. We assumed that
the reactions to form each product had a first–order dependence on the partial pressure of MAL.
In addition, it can be assumed that the successive oxidations of MAA and acetic acid into CO
x
were negligible since conversion of MAL was enough low. The first–order reaction rate constants
for the formations of MAA, acetic acid, and CO were estimated from the yields at each reaction
x
temperature. The natural logarithms of the rate constants were plotted against inverse of the
reaction temperature, namely Arrhenius plot, in Fig. 2. From the slopes and intercepts of the
approximation lines, apparent activation energies and pre–exponential factors were estimated
(Table 1). For all the products, the apparent activation energies in the presence of steam were
about 1.5 times larger than those in the absence of steam, suggesting that the reaction mechanism
or the rate-determining step was different for the reactions in the absence and presence of steam.
In addition, it should be noted that the addition of steam caused a substantial increase in the pre–
exponential factors (A), e.g. A for MAA formation in the presence of steam was about two orders
of magnitude greater than that in the absence of steam. The large pre–exponential factors led to
1
4

the substantial enhancement in the formation rates in the presence of steam. The increases in the
pre–exponential factors with steam suggest that i) sticking probability of MAL on the surface of
3
H PMo12O40, that is, probability of the effective collision of MAL with the surface, was increased
or ii) the number of active sites on the catalyst was increased. Those will be discussed later in
detail.
3.2 Effect of steam on redox property of the catalyst
It is reported that the selective oxidation of MAL over H PMo12O40 proceeds with Mars–
3
Van Krevelen mechanism, in which lattice oxygen of the catalyst oxidizes MAL [19, 20]. It is
possible that the amount of lattice oxygen in the catalyst involved in the redox cycle was increased
by steam, thereby the catalytic activity was increased. In order to test this hypothesis, the reaction
was carried out without any supply of O to estimate the amounts of lattice oxygen being able to
2
participate in MAL oxidation. The amount of those lattice oxygen was estimated from the total
amount of products formed after the stop of O supply. After the reaction reached near steady state
2
under the reaction conditions, supply of O to the reactor was stopped while the supply of MAL
2
was continued. The results are shown in Fig. 3, where the reactions were conducted in the different
–
1
W/F in the presence and absence of steam, being W/F= 173 and 345 g h mol , respectively, to
obtain similar conversions. As Fig. 3 clearly demonstrates, the reactions continued even after the
1
5

stop of O supply even in the absence of steam, indicating that the reaction in the absence of steam
2
proceeds with Mars–Van Krevelen mechanism as that in the presence of steam. The main products
were MAA, acetic acid and CO
x
even after the stop of O supply. The yields gradually decreased
2
with time and the reaction was completely stopped at 200 ~ 300 min after the stop of O supply.
2
–4
–4
–4
For the reaction in the presence of steam, 2.1 × 10 , 1.2 × 10 , and 2.0 × 10 mol of MAA,
acetic acid, and CO
acetic acid, and CO
x
x
, respectively, were formed after the stop of O supply. Assuming that MAA,
2
(mainly CO) were formed according to eqs. 12, 13, and 14, respectively, the
number of oxygen atoms necessary for their formations can be calculated and the sum of them
–4
–1
was 7.2 × 10 mol g , which corresponds to 1.4 Olattice per one Keggin unit. Misono et al. reported
that the about two lattice oxygens per one Keggin unit of H PMo12O40 were consumed for the
3
MAL oxidation after the stop of O
present study.
2
supply [19, 20], which was about the same number in the
C
4
4
H
6
6
O + [O] → C
4
H
6
O
2
(eq. 12)
(eq. 13)
(eq. 14)
C
H
O + 4[O] → C
2
H
4
O
2
+ 2CO + H O
2
CO + [O] → CO
2
[
O]: lattice oxygen (Olattice)
1
6

For the reaction in the absence of steam, the lattice oxygen involved in the reaction was
.3 Olattice per one Keggin unit, which is almost the same as that in the presence of steam. In
1
addition, there was no difference in the average valence of Mo after the reaction, which was Mo5.8+
regardless of the presence and absence of steam. Based on these results, it was concluded that the
addition of steam did not alter the redox property of H
PMo12O .
3 40
3.3 Reversibility of effect of steam on the catalytic performance
Fig. 4 shows the transient response in the catalytic performance for the MAL oxidation
over H PMo12O40. In the initial 400 min, the reaction was conducted at PH2O = 0 atm and then the
3
feed gas was switched to that with PH2O = 0.13 atm. After the switching, the yield of MAA was
gradually increased with time and reached a steady state after 60 min of the switching. As
mentioned in the introduction, Misono et al. proposed that MAL is hydrated to produce 1,1–diol,
which is an intermediate, with the aid of Brønsted acids in the presence of steam. If the production
of 1,1-diol is the sole role of steam, the reaction must reach the steady state immediately after the
switching the feed gas. In fact, however, about 60 min was necessary to reach the steady state,
indicating that the influence of steam was related to somewhat slow process. After the reaction
reached steady state at PH2O = 0.13 atm, the feed gas was switched back to that without steam
(PH2O = 0 atm) at 760 min, whereupon the yield of MAA decreased to almost the same value as
1
7

original one at PH2O = 0 atm. We also conducted the experiment in the reverse order of the feed
gas (PH2O = 0.13 atm → 0 atm → 0.13 atm), the latter yield of MAA at PH2O = 0.13 atm was
almost the same as the initial one (Fig. S1). These results clearly indicate that the influence of
steam was reversible phenomenon.
To investigate the changes in the catalyst structure during the transient response
experiments, the reaction was quenched in the middle of the reactions under the same reaction
conditions shown in Figs. 4 and S1, and then the catalysts were analyzed by using IR (Fig. S2),
powder XRD (Fig. S3), nitrogen adsorption (Table S1), redox titration (Table S1), and XPS (Fig.
S4). However, little difference was found between the fresh catalyst and the spent ones taken in
the middle of the reaction runs. Hence, the large difference in the catalytic data between in the
presence and absence of steam was impossible to be rationalized by the structural changes in the
catalyst. Therefore, dynamic (reversible) change in the catalyst structure occurring in the presence
of steam brought about the enhancement in the catalytic performance with steam.
3
.4 Hydration of H PMo12O40 under the reaction conditions
3
Figure 5 shows TG profile of hydrous H
3
PMo12
O
40·nH O used in this study in dry He.
2
The number of water was 12 (= n in H
3
PMo12
O
40·nH O) at room temperature, and a part of water
2
was released at around 323 K to turn to be n = 8, and finally anhydrous H PMo12O40 was formed
3
1
8

at around 423 K. According to the TG profile, the catalyst must exist as anhydrous H
3
PMo12O
40
under dry atmosphere at the reaction temperature (573 K) of the MAL oxidation. However, since
anhydrous H
3
PMo12
O
40 reversibly returned back to hydrous H
3
PMo12
O
40·nH O if it was put under
2
wet atmosphere at room temperature (Fig. S5), it is presumable that the catalyst has a certain
number of water of crystallization at the reaction temperature if the feed gas contains steam.
Thus, we determined the amount of water contained in H PMo12O40 at 573 K and PH2O =
3
0.13 atm with the procedure described in the section 2.3 and the result is shown in Fig. 6. For the
blank test with an empty reactor, the signal intensity due to H O (m/e = 18) immediately decreased
2
to the baseline level after switching the gas from wet He (PH2O = 0.13 atm) to dry one. On the
other hand, when the catalyst was present in the reactor, H O had continued to desorb till about
2
1
0 min after switching to dry He. The total amount of H O desorbed from the catalyst was 1.4
2
–
1
mmol g , which corresponds to H
3
PMo12
O
40·3H O. This is the first report to demonstrate that
2
3
H PMo12O40 contains a certain amount of water in the bulk of Keggin-POM at high temperature
under wet atmosphere.
It was also confirmed from IR spectra that H PMo12O40 was present as a hydrous form
3
at 573 K in the presence of steam (PH2O = 0.13 atm). Fig. 7 shows the IR spectrum of H
3
PMo12O
40
at 573 K under dry He, and those after switching the flowing gas to wet He (PH2O = 0.13 atm)
while keeping the temperature at 573 K. Under dry He (0 min in Fig. 7), the absorption band at
1
9

–1
–1
3
250 cm due to O–H stretching vibration in H
3
PMo12
O
40 and those at 794 cm (ν Mo–O–Mo),
–1
–1
–1
8
83 cm (ν Mo–O–Mo), 982 cm (ν Mo=O), and 1063 cm (ν P–O) due to framework vibrations
3–
of [PMo12O40] were observed, which are consistent with a previous report [23]. After switching
–
1
the flowing gas to wet He, the broad absorption band at 3300 – 3400 cm assignable to O–H
stretching vibration of H O [24, 25] became intense with time (5–30 min in Fig. 7). In addition,
2
slight change in the framework vibrations was also observed after the gas switching, supporting
that the catalyst was hydrated at 573 K in the presence of steam. It is noted that the change in the
IR spectra was nearly completed in 30 min after the gas switching. This time length was almost
the same as that to reach a steady state in the transient response for the MAL oxidation shown in
Fig. 3. Those analogous behavior for time strongly suggested that hydration of H PMo12O40 under
3
the reaction conditions was responsible for the high catalytic performance in the presence of steam.
+
O]+,
We speculated that three H
2
O were coordinated to each of the three H to form monomeric [H
3
or dimeric [H
2
O---H---OH
]+, though still under investigation on how water is present in
2
H
3
PMo12
O
40·nH O (n = 3) under the reaction conditions with steam.
2
3.5 Effect of steam on adsorption property for methacrolein
The hydration of H PMo12O40 had a large impact on adsorption of MAL on the catalyst.
3
We compared the adsorption amounts of MAL at 433 K for anhydrous (H PMo12O40) and hydrous
3
2
0

(
3
H PMo12
O
40·3H O) catalysts by a pulse method. The results are shown in Fig. 8. On the hydrous
2
catalyst, MAL was adsorbed up to 7 pulses without any ejection of water from the catalyst. The
–
1
total amount of adsorbed MAL was 10.4 μmol g , corresponding to 0.019 MAL per one Keggin
unit in the bulk. The powder XRD pattern of the catalyst after the experiment showed no change
in the lattice constant. The surface area of the catalyst estimated from nitrogen adsorption
2
–1
2
isotherm was 10 m g . Since one Keggin unit occupies about 1.44 nm [26], the number of
–
1
H PMo12O40 exposed on the surface was calculated to be 12 μmol g . This value was almost the
3
same as the adsorption amount of MAL estimated by the pulse experiment. Based on these results,
it is concluded that MAL adsorbed only on the outer surface of the catalyst particle but was not
absorbed in the bulk. In contrast to the hydrous catalyst, anhydrous H
MAL at all (Fig. 8b).
3
PMo12O40 did not adsorb
3
.6 Role of steam in the oxidation of methacrolein over H
3
PMo12O
40
Based on the results mentioned above, we propose a mechanism for the enhancement
in the MAA formation over H PMo12O40 by steam as follows (Scheme 3). Under the reaction
3
conditions in the presence of steam, three H
2 3
O per one H PMo12O40 is absorbed into the bulk of
+
H
3
PMo12
O
40 to form hydrated proton like monomeric [H
3
O] (A in Scheme 3). MAL is activated
by the hydrated proton (B), forming 1,1–diol on the surface (C). Then, it is oxidized with the
2
1

lattice oxygen of H PMo12O40 to produce MAA and consequently hydrated proton is regenerated
3
(
D). Finally, the reduced catalyst is re-oxidized with gas phase O to complete the catalyst cycle.
2
As the MAL adsorption experiment demonstrated, adsorption of MAL on the surface, which may
form 1,1–diol, proceeds smoothly in the presence of steam, leading to the significant increase in
the pre–exponential factor. Quantum chemical calculation suggests that the activation of MAL
+
through the reaction with [H
3
O] proceeds without any high energy state (Fig. S6), supporting the
mechanism proposed above.
Conversely, in the absence of steam, the intermediates of B and C do not form on the
catalyst and thus the lattice oxygen would directly oxidize MAL. It is plausible that this direct
oxidation is hard to occur comparing with the activation of MAL on the hydrated catalyst in the
presence of steam. Thus pre–exponential factor was low and non–selective oxidation occurred
in the absence of steam.
In selective oxidation of acrolein to acrylic acid over Mo-V-W-O catalyst, it is
x
proposed that steam promote to hydrolyze acrylic acid ester (M–O(C=O)C
2
3
H ) on the catalyst
surface to give acrylic acid accompanying with the regeneration of surface hydroxyl group
(
Scheme 4) [27]. If the MAL oxidation over H PMo12O40 proceeds with the same reaction
3
mechanism as the oxidation of acrolein, formation of MAA is promoted with steam by hydrolysis
of methacrylic acid ester on the surface, that is, desorption of MAA from the surface is facilitated
2
2

with steam. Such smooth desorption of product MAA may lower the chance for overoxidation of
the surface species due to less prolonged stay on the catalyst, while it is still unknow how acetic
acid and CO are formed. Consequently, steam promotes the MAA formation much more than
x
those of acetic acid and CO
presence of steam.
x
, leading to the improved selectivity for MAA formation in the
4. Conclusion
Addition of steam with optimal partial pressure (PH2O = 0.13 atm) to the feed gas led to
fivefold and twice increases in the yield of and selectivity to MAA, respectively. Kinetic analysis
demonstrated that the addition of steam brought about 200-fold increase in the pre-exponential
factor for MAA formation, leading to the significant increase in the formation rate of MAA. The
steam in the feed gas varied hydrous state of H PMo12O40 under the reaction conditions, while did
3
not alter the redox property, molecular and crystalline structures, and surface area of the catalyst.
In the presence of steam at 573 K, three H
2
O per one H PMo12O40 were absorbed into the bulk of
3
+
the catalyst to form hydrated proton like [H
3
O] . On the hydrous catalyst, activation of
methacrolein readily occurred with a formation of 1,1-diol on the surface. The promotion of the
MAL activation was the reason for the significant increase in the pre–exponential factor and
consequently improvement in the catalytic performance in the presence of steam.
2
3

References
[
[
[
[
[
[
[
[
[
[
[
1] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199 – 218.
2] M. Misono, N. Nojiri, Appl. Catal. 64 (1990) 1 – 30.
3] M. Misono, Chem. Commun. (2001) 1141 – 1152.
4] A. Corma, Chem. Rev. 95 (1995) 559 – 614.
5] C.L. Hill, C.M. Prosser-McCartha, Coord. Chem. Rev. 143 (1995) 407 – 455.
6] M.M. Lin, Appl. Catal. A 207 (2001) 1 – 16.
7] A. Micek-Ilnicka, J. Mol. Catal. A 308 (2009) 1 – 14.
8] D.-L. Long, R. Tsunashima, L. Cronin, Angew. Chem. Int. Ed. 49 (2010) 1736 – 1758.
9] J. C. Védrine, Catalysts 7 (2017) 341; https://doi.org/10.3390/catal7110341.
10] I.V. Kozhevnikov, Chem. Rev. 96 (1998) 171 – 198.
11] G.M. Brown, M.-R. Noe-Spirlet, W.R. Busing, H.A. Levy, Acta Cryst. B33 (1977) 1038 –
1046.
[
[
12] T. Okuhara, S. Tatematsu, K.Y. Lee, M. Misono, Bull. Chem. Soc. Jpn. 62 (1989) 717 – 723.
13] T. Okuhara, T. Hashimoto, M. Misono, Y. Yonedea, H. Niiyama, Y. Saito, E. Echigoya, Chem.
Lett. (1983) 573 – 576.
[
14] T. Okuhara, A. Kasai, N. Hayakawa, M. Misono, Y. Yoneda, Chem. Lett. (1981) 391 – 394.
2
4

[
15] M. Misono, T. Okuhara, T. Ichiki, T. Arai, Y. Kanda, J. Am. Chem. Soc. 109 (1987) 5535 –
5536.
[
[
[
[
[
16] Y. Saito, H. Niiyama, J. Catal. 106 (1987) 329 – 336.
17] T. Okuhara, T. Arai, T. Ichiki, K. Y. Lee, M. Misono, J. Mol. Catal. 55 (1989) 293 – 301.
18] T. Okuhara, A. Kasai, N. Hayakawa, Y. Yoneda, M. Misono, J. Catal. 83 (1983) 121 – 130.
19] Y. Konishi, K. Sakata, M. Misono, Y. Yoneda, J. Catal. 77 (1982) 169 – 179.
20] M. Misono, K. Sakata, Y. Yoneda and W. Y. Lee, in “Proceedings on 7th International
Congress on Catalysis, Tokyo, 1980, p.1047-1059.
[
[
21] M. Furuta, K. Sakata, M. Misono, Y. Yoneda, Chem. Lett. (1979) 31 – 34.
22] M. Fournier, C. Feumi-Jantou, C. Rabia, G. Hervé, S. Launay, J. Mater. Chem. 2 (1992) 971
–
978.
[
23] A. Bielański, A. Malecka, L. Kubelkova, J. Chem. Soc., Faraday Trans. 1 85 (1989) 2847 –
2856.
[
[
[
[
24] C. Pazé, S. Bordiga, A Zecchina, Langmuir 16 (2000) 8139 – 8144.
25] M. Hashimoto, G. Koyano, N. Mizuno, J. Phys. Chem. B 108 (2004) 12368 – 12374.
26] K. Brukman, J. Haber, E.M. Serwicka, Faraday Discuss. Chem. Soc. 87 (1989) 173 – 187.
27] A. Drochner, D. Ohlig, S. Knoche, N. Gora, M. Heid, N. Menning, T. Petzold, H. Vogel, Top.
Catal. 59 (2016) 1518 – 1532.
2
5

Table 1 Apparent activation energies and pre–exponential factors estimated from Arrhenius plots
in the presence or absence of steam.
a
–1
b
–1
E / kJ mol
A / h
Product
MAAc
P
H2O/atm
PH2O/atm
0
.13
0.00
0.13
0.00
86
64
3.4×104
1.1×1012
1.2×1011
1.4×102
6.5×106
1.6×107
Acetic acid 170
120
120
CO
x
160
a
b
c
Apparent activation energy
Pre-exponential factor
Methacrylic acid
2
6

+
-
H O
2
H
OH
OH
Mo
O
OH
Mo
H O
or
2
O
OH
O
O
O
Mo
Scheme 1 Reaction mechanism involving water proposed in [19, 20].
2
7

OH
O
H
(
MAA)
O
(MAL)
OH
+
2 COx
O
Scheme 2 Possible reaction routes in the oxidation of methacrolein (MAL) over H
3
PMo12
O
40
2
8

Scheme 3 Proposed reaction mechanism for selective oxidation of methacrolein over H
3
PMo12O
40
in the presence of steam.
2
9