Direct conversion of methanol into 1,1-dimethoxymethane: Remarkably high productivity over an FeMo catalyst placed under unusual conditions

Article abstract of DOI:10.1039/c0gc00194e

We report here the highest productivity ever observed in the direct conversion of methanol into 1,1-dimethoxymethane (ca. 4.6 kg_DMM h^-1 kg_cat^-1 at 553 K), this result being obtained over an FeMo catalyst. This catalyst is industrially used to selectively convert methanol into formaldehyde but has never before been applied to the present reaction. Placing this FeMo catalyst under unusual reaction conditions, i.e., using a feed rich in methanol, completely changed its behaviour in terms of selectivity: the massive production of 1,1-dimethoxymethane, instead of formaldehyde, was observed.

Full text of DOI:10.1039/c0gc00194e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/greenchem | Green Chemistry
Direct conversion of methanol into 1,1-dimethoxymethane: remarkably high
productivity over an FeMo catalyst placed under unusual conditions
Julien Gornay,a,b,c Xavier S e´ cordel,
a,b,c
Guillaume Tesquet,
a,b,c
B e´ atrice de M e´ norval,
a,b,c
Sylvain Cristol,
a,b,c
a,b,d
a,b,c
a,b,c
a,b,e
f
Pascal Fongarland,
Franck Dumeignil*
Micka e¨ l Capron,
Louise Duhamel,
Edmond Payen,
Jean-Luc Dubois and
a,b,c
Received 8th June 2010, Accepted 27th August 2010
DOI: 10.1039/c0gc00194e
We report here the highest productivity ever observed in the
direct conversion of methanol into 1,1-dimethoxymethane
one step process. This direct reaction is strongly sensitive to
the nature of the catalytically-active sites. Consecutive redox
7
-
1
-1
(
ca. 4.6 kgDMM
h
kgcat at 553 K), this result being obtained
and acid-catalyzed pathways (Scheme 1) can give different
products, and the difficulty in selectively obtaining DMM
explains the scarcity of the literature on this subject. Never-
theless, promising results have been reported over supported
over an FeMo catalyst. This catalyst is industrially used
to selectively convert methanol into formaldehyde but has
never before been applied to the present reaction. Placing
this FeMo catalyst under unusual reaction conditions, i.e.,
using a feed rich in methanol, completely changed its
behaviour in terms of selectivity: the massive production
of 1,1-dimethoxymethane, instead of formaldehyde, was
observed.
8
Re oxide and Re-based mixed oxides, and over bulk or
9
supported molybdenum catalysts. Variations of V-TiO
2
-based
formulations recently gave rather good DMM productivities (at
-
1
-1 10
max. 0.33 ± 0.03 kgDMM h kgcat ). However, the highest DMM
productivity, before the present paper, was obtained in our
team on a Mo12
under optimized conditions.
V
3
x
W1.2Cu1.2Sb0.5O (AR01) formulation working
11
Methanol is mainly produced by the conversion of syngas,
usually obtained from methane or coal. Recently, production
of so-called ‘biomethanol’ has started in new units that process
biomass-derived syngas. Methanol can now be considered as a
sustainable platform molecule, with perspectives of downstream
chemical applications.
1
,1-Dimethoxymethane (DMM) is a chemically stable com-
pound with various possible uses (solvent, fuel additive . . . ). For
example, the selective oxidation of DMM into formaldehyde
1
(
F) enables the preparation of concentrated F solutions.
High volume applications have recently been patented: poly-
oxymethylenedimethylether (POMM) synthesised from DMM
Scheme 1 Methanol oxidation pathways (adapted from ref. 7).
2
can be blended with diesel fuels, or can advantageously replace
3
12,13
methanol in fuel cells with lower toxicity and higher efficiency.
FeMo formulations are industrially used to produce F
POMM is also an embalming agent that is much safer for an
embalmer’s health compared to F.
using reactant feeds with low concentrations of methanol
(< 7.5 mol%). A conversion of 99% is observed with a selectivity
for F of 94%, while DMM is formed in very small quantities.
However, we found that when in contact with a high methanol
concentration, the FeMo catalyst exhibited a remarkably high
DMM productivity.
4
On an industrial scale, DMM is produced by catalytic
5
distillation. Acetalization is performed in the liquid phase by the
6
reaction of methanol with F issued from methanol oxidation.
Obviously, realizing the direct conversion of methanol into
DMM would enable the elaboration of a more economical
In this work, we compare the performance of an industrial
FeMo catalyst [(MoO
3
–Fe
2
(MoO ) ] provided by Arkema, of
4
3
which the preparation method and the main characteristics
a
14
11,15
Univ. Lille Nord de France, F-59000, Lille, France. E-mail: franck.
are described elsewhere, with that of AR01.
A 6.9 wt%
dumeignil@univ-lille1.fr; Fax: +33 (0)3.20.43.65.61; Tel: +33
Re/TiO formulation (ReTi) was tested as a reference, as it
2
(
0)3.20.43.45.38
has also previously exhibited interesting performances, but with
drawbacks such as a high price due to its rarity and to Re
b
CNRS UMR8181, Unit e´ de Catalyse et Chimie du Solide, UCCS,
F-59655, Villeneuve d’Ascq, France
11
c
USTL, F-59655, Villeneuve d’Ascq, France
depletion under the reaction conditions. The ReTi catalyst
was obtained by grinding anatase-TiO (Sachtleben-HOMKAT
d
ECLille, F-59655, Villeneuve d’Ascq, France
2
e
ENCL, F-59652, Villeneuve d’Ascq, France
F01) and metallic Re at room temperature before calcining at
f
Arkema, 420 Rue d’Estienne d’Orves, 92705, Colombes, France
16
6
73 K in O
2
for 6 h. The 6.9 wt% Re loading was confirmed by
†
Electronic Supplementary Information (ESI) available: Preparation of
inductively coupled plasma (ICP) elemental analysis. The FeMo
catalyst and AR01 exhibited much lower specific surface areas
the FeMo catalyst, composition of the FeMo catalysts, DSC, XRD,
conversion curves, performances with time on stream, influence of
GHSV. See DOI: 10.1039/c0gc00194e
2
-1
2
-1
(< 10 m g ) than ReTi (101 m g ). We also synthesized an
1
722 | Green Chem., 2010, 12, 1722–1725
This journal is © The Royal Society of Chemistry 2010

View Article Online
Table 1 Comparison between the catalytic performances of ReTi, AR01 and FeMo at 553 K
Selectivity/equiv. carbon (mol%)
a
Catalyst
Feed
Conversion (%)
DMM
F
DME
MF
CO + CO
2
ReTi
P
P
P
R
R
R
65
57
59.9
20.2
18.1
55.7
74
89.2
2.8
79.6
67.8
89.7
1
11.9
5.1
7.3
9.6
15
7
6.1
1.2
0.6
1.2
0.9
0.1
AR01
FeMo
ReTi
AR01
FeMo
3.3
87.6
0.1
13.2
4.2
1.2
1.7
9.5
31
0.7
5.3
a
Conditions: P = poor feed; R = rich feed.
1
4
FeMo catalyst (see the preparation procedure in the ESI†). We
verified the reproducibility of the catalytic performances, which
were in the very reasonable range of ±3% compared to those of
industrial FeMo due to slight differences in composition (ESI†,
Table S1). Furthermore, differential scanning calorimetry (DSC)
analysis evidenced a small exothermal peak close to 648 K, and
two more intense peaks close to 668 and 683 K, respectively
The productivity of the catalysts in DMM is reported in Fig. 1.
The FeMo catalyst is completely atypical compared to ReTi
and AR01. ReTi and AR01 show similar DMM productivities,
with maximum values in rich feed conditions at 553 K of ca.
-
1
-1
1.5 and 1.0 kgDMM
h
kgcat , respectively. In contrast, over
the FeMo catalyst, while in poor feed conditions the DMM
-
1
-1
productivity did not exceed 0.05 kgDMM
h
kgcat , a remarkable
(
ESI†, Fig. S1). The presence of three phases was confirmed
by XRD analysis (ESI†, Fig. S2): Fe (MoO , MoO and b-
Fe (MoO . XRD recorded at a temperature close to 673 K
further confirmed the formation of crystallographic phases
increase was observed when using rich feed conditions, with
-
1
-1
2
4
)
3
3
DMM productivities ranging from 2.2 kgDMM
h
kgcat at 523 K
-
1
-1
2
4
)
3
to the remarkable value of 4.6 kgDMM
h
kgcat at 553 K; this
is by far the highest value ever reported in the literature. The
versatile behaviour of the FeMo catalyst is illustrated in Fig. 2,
with a F productivity much larger than that of AR01 and ReTi.
Furthermore, the performance of the FeMo catalyst was stable
with time on stream, which is well in line with its actual industrial
use for producing F. An example of a long-term experiment is
given in the ESI (Fig. S5).† In addition, Fig. S6 (ESI†) confirms
(
ESI†, Fig. S3).
The performances of the catalysts were evaluated at atmo-
spheric pressure in a fixed bed reactor. Two different feed
compositions were selected outside of the flammability zone
17
(
between 7.5 and 40% methanol in air): the ‘poor feed’
with CH OH/O /He = 7.5/8.5/84 and the ‘rich feed’ with
CH OH/O
GHSV) was adjusted to 22 NL h
3
2
-
1
-1
3
2
/He = 40/13/47 (mol%). Gas hour space velocity
that the GHSV of 22 NL h
formulation.
g
cat is optimal for the FeMo
-
1
-1
(
gcat . The reactants, namely
oxygen and methanol, and the products, namely dimethylether
DME), F, DMM, methyl formate (MF), CO and CO , were
(
2
analyzed using an online gas micro chromatograph (SRA3000)
equipped with two columns (plot U and molecular sieves) and
two thermal conductivity detectors (TCDs). The total carbon
balance was, in any case, over 95%, which was satisfactory. Prior
to use, AR01 and ReTi were activated at 623 K for 1 h in an
oxygen flow, while FeMo was used without pre-treatment, as in
its conventional use to yield F.
In poor feed conditions, all the catalysts exhibited quite
similar conversions, within a ±5% range (Table 1; Fig. S4†). The
conversion obviously increased with temperature to reach ca.
6
0% at 553 K, irrespective of the sample (Fig. S4†). Meanwhile,
in rich feed conditions, the conversion of ReTi and AR01
drastically decreased (less than 20%) but remained at the same
level as that observed in poor feed conditions over the FeMo
catalyst (Fig. S4†). The selectivities for the reactions performed
at 553 K are reported in Table 1. In poor feed conditions,
the FeMo catalyst is, as expected, selective for F (87.6%)
with a very low selectivity for DMM (2.8%), while ReTi and
AR01 are selective for DMM (74 and 89.2%, respectively).
In rich feed conditions, the DMM selectivity observed over
ReTi and AR01 remains high (79.6 and 67.8%, respectively)
and the FeMo catalyst, which was selective for F in poor
feed conditions, becomes very selective for DMM (ca. 90%)
with an almost unaltered conversion (59.9%—poor vs. 55.7%—
rich).
Fig. 1 DMM productivity over the various catalysts as a function of
temperature with poor (open symbols) and rich (filled symbols) feeds.
We calculated the apparent activation energy of the reaction
(E , Table 2) using Arrhenius plots. In poor feed conditions,
a
the three catalysts exhibited similar E values of ca. 41 kJ
a
-
1
mol , which corresponds to the value previously reported for
18,19
the conversion of methanol into F.
However, while the main
product over FeMo is F, it is DMM over ReTi and AR01. This
can be explained if we consider that the limiting step of the
reaction is the oxidation step of methanol into F. Then, provided
proper acid sites are present on the catalyst, F can further be
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1722–1725 | 1723

View Article Online
Table 2 Activation energy (E
a
) calculated for the various catalysts
working in the poor and the rich feed conditions
the catalyst to the reactants’ atmosphere, we observed a shift
of peaks, corresponding to an elongation of the b parameter of
a
/kJ mol-1
the MoO cell. Furthermore, Raman spectroscopy performed
3
Catalyst
Feed
E
a
Major product
under Operando conditions using the FeMo catalyst in poor
and rich feed conditions showed in both cases a concomitant
and progressive decrease with time on stream in the intensity of
ReTi
P
P
P
R
R
R
40.8
41.7
40.8
26.2
33.5
45.7
DMM
DMM
F
DMM
DMM
DMM
AR01
FeMo
ReTi
AR01
FeMo
the Raman bands corresponding to MoO
3
and Fe
2
(MoO ) .
4
3
The mechanism of the reaction in conventional conditions,
i.e., poor feed, is still a matter of debate, even if several groups
have studied this reaction over various types of solids. The
unusual behaviour described in this paper—when using experi-
mental conditions radically different from the usual ones—raises
further questions.
a
Conditions: P = poor feed; R = rich feed.
Conclusion
According to our discovery, DMM, a valuable product, can
be readily and sustainably produced in existing infrastructures
dedicated to methanol conversion into formaldehyde over the
same FeMo catalysts with air as a dilutent/oxidant by using high
methanol partial pressures instead of the low partial pressures
conventionally used for formaldehyde production.
Acknowledgements
The authors gratefully acknowledge Arkema as well as the CE
for its financial support through contract number MIRG-CT-
2007-046383. The authors also want to thank Mr O. Gardoll,
Mrs N. Djelal and Mrs L. Burylo from UCCS for their precious
technical help.
Fig. 2 Formaldehyde productivity over the various catalysts as a
function of temperature with poor (open symbols) and rich (filled
symbols) feeds.
easily converted into DMM, which seems to be the case over
AR01 and ReTi.
Notes and references
In rich feed conditions, E is lower for ReTi and AR01, with
a
-
1
respective values of 26.2 and 33.5 kJ mol , while the main
product is still DMM. This can be explained if the mechanism
is not the same for the two feed conditions, as evoked by
1 JP Pat., 1186832, 1989; JP Pat., 4967014, 1990; JP Pat., 59145, 1993;
JP Pat., 2956916, 1999; J. Lojewska, J. Wasilewski, K. Terelak, T.
Lojewski and A. Kolodziej, Catal. Commun., 2008, 9(9), 1833–1837.
2
EP Pat., 1914293, 2008.
18
Edwards Density functional theory (DFT) calculations are
currently in progress in our lab to support this hypothesis of a
3 WO Pat., 084993, 2006; WO Pat., 138940, 2008.
4 WO Pat., 090294, 2008.
5
J. Masamoto and K. Matsuzaki, J. Chem. Eng. Jpn., 1994, 27(1), 1–5;
CH Pat., 688041, 1997; US Pat., 6379507B1, 2002.
WO Pat., 018209, 1993; EP Pat., 704442, 1996; US Pat., 5959456,
1999; US Pat., 6160174, 2000; US Pat., 6160186, 2000; US Pat.,
dual mechanism. In contrast, the E observed over FeMo placed
a
-
1
in rich feed conditions increased to 45.7 kJ mol , with DMM
as the main product instead of F in the poor feed conditions.
The reason for this atypical behaviour is not yet elucidated but
might originate from (1) the development of specific acid sites
over FeMo in the rich feed conditions and/or (2) a F desorption
rate lower than the formation/desorption rate of DMM in the
presence of high methanol partial pressures. The acidity of
FeMo catalysts, which is required in addition to redox properties
6
6
2
166266, 2000; 2002. US Pat., 6437195, 2002; US Pat., 6379507,
002.
7
8
J.-M. Tatibou e¨ t and J. E. Germain, J. Catal., 1981, 72, 375–378; J.-M.
Tatibou e¨ t, Appl. Catal., A, 1997, 148, 213–252; J.-M. Tatibou e¨ t and
H. Lauron-Pernot, J. Mol. Catal. A: Chem., 2001, 171, 205–216.
Y. Yuan, H. Liu, H. Imoto, T. Shido and Y. Iwasawa, Chem. Lett.,
2
000, 674–675; Y. Yuan, H. Liu, H. Imoto, T. Shido and Y. Iwasawa,
J. Catal., 2000, 195, 51–61; Y. Yuan, T. Shido and Y. Iwasawa, Chem.
Commun., 2000, 1421–1422; Y. Yuan and Y. Iwasawa, J. Phys. Chem.
B, 2002, 106, 4441–4449; Y. Yuan, K. Tsai, H. Liu and Y. Iwasawa,
Top. Catal., 2003, 22, 9–15.
14,20
to yield DMM, has previously been characterized ex situ.
VI
Surface acidity is linked with the presence of Mo entities,
and the over-stoichiometry in Mo seems to have an effect on
the acid sites’ quantity/strength. However, the acidic behaviour
of this catalyst in real working conditions still needs further
investigation. The catalytic results in the present paper strongly
suggest that acidity could be drastically different, according to
the operating conditions (rich feed vs. poor feed).
9
M. Fournier, A. Aouissi and C. Rocchiccioli-Deltcheff, J. Chem.
Soc., Chem. Commun., 1994, 307–308; C. Rocchiccioli-Deltcheff, A.
Aouissi, M. M. Bettahar, S. Launay and M. Fournier, J. Catal., 1996,
1
64, 16–27; C. Rocchiccioli-Deltcheff, A. Aouissi, S. Launay and M.
Fournier, J. Mol. Catal. A: Chem., 1996, 114(1–3), 331–342; H. Liu
and E. Iglesia, J. Phys. Chem. B, 2003, 107(39), 10840–10847; H. Liu
and E. Iglesia, J. Catal., 2004, 223(1), 161–169.
Further, XRD experiments were performed using a D8
Advance apparatus (Br u¨ ker AXS) equipped with a controlled
temperature/atmosphere chamber. The FeMo catalyst was an-
alyzed in the presence of methanol and oxygen. When exposing
10 Y. Fu and J. Shen, Chem. Commun., 2007, 2172–2174; J. Liu, Y. Fu,
Q. Sun and J. Shen, Microporous Mesoporous Mater., 2008, 116, 614–
6
21; J. Liu, Q. Sun, Y. Fu and J. Shen, J. Colloid Interface Sci., 2009,
35, 216–221; H. Zhao, S. Bennici, J. Shen and A. Auroux, J. Mol.
3
Catal. A: Chem., 2009, 309(1–2), 28–34; H. Guo, D. Li, D. Jiang,
1
724 | Green Chem., 2010, 12, 1722–1725
This journal is © The Royal Society of Chemistry 2010

View Article Online
W. Li and Y. Sun, Catal. Commun., 2010, 11(5), 396–400; H. Zhao,
S. Bennici, J. Shen and A. Auroux, J. Therm. Anal. Calorim., 2010,
14 N. Pernicone, J. Less Common Met., 1974, 36, 289–297.
15 WO Pat., 034264, 2007.
16 A. A. Olsthoorn and C. Boelhouwer, J. Catal., 1976, 44(2), 197–206.
17 M. G. Zabetakis, Flammability Characteristics of Combustible Gases
and Vapors, Bulletin 627, US Bureau of Mines, Washington D.C.,
1965.
9
9(3), 843–847; H. Guo, D. Li, D. Jiang, W. Li and Y. Sun, Catal.
Lett., 2010, 135(1–2), 48–56; H. Zhao, S. Bennici, J. Shen and A.
Auroux, J. Catal., 2010, 272(1), 176–189; H. Zhao, S. Bennici, J. Cai,
J. Shen, J. and A. Auroux, A, Catal. Today, 2010, 152(1–4), 70–77.
1 S. Royer, X. S e´ cordel, M. Brandhorst, F Dumeignil, S. Cristol, C.
Dujardin, M. Capron, E. Payen and J.-L. Dubois, Chem. Commun.,
1
18 J. Edwards, J. Nicolaidis, M. B. Cutlip and C. O. Bennett, J. Catal.,
1977, 50, 24–34.
2
008, 865–867.
2 N. Pernicone, F. Lazzerin, G. Liberti and G. Lanzavecchia, J. Catal.,
969, 14, 293–302.
3 Y. Bourgeois, Techniques de l’Ing e´ nieur, 1994, J6, 340–341.
19 K. Habersberger and P. Jiru, Collect. Czech. Chem. Commun., 1972,
37, 535.
1
1
1
20 A. P. V. Soares, M. Farinha Portela, A. Kiennemann, L. Hilaire and
J. M. M. Millet, Appl. Catal., A, 2001, 206, 221–229.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1722–1725 | 1725