Catalytic vapor-phase oxidation of 2,2,2-trifluoroethanol

Article abstract of DOI:10.1016/j.jfluchem.2005.12.012

The synthesis of trifluoroacetaldehyde by vapor-phase oxidation of 2,2,2-trifluoroethanol using supported vanadium catalysts was studied. Significant differences were observed in the reaction outcomes resulting from different types of catalysts. The ZrO₂- and Al₂O₃-supported catalyst demonstrated both high catalytic activity and selectivity. The addition of co-catalysts such as MoO₃ or SnO₂ improved catalytic performance (Selectivity: up to 91%; S.T.Y.: >200 g L^-1 h^-1). The experimental results on catalyst lifetime showed a marked decrease in the activity of the Al₂O₃-supported catalyst within tens of hours, while the ZrO₂-supported catalyst showed little, if any, performance alterations for 2000 h.

Full text of DOI:10.1016/j.jfluchem.2005.12.012

Journal of Fluorine Chemistry 127 (2006) 519–523
www.elsevier.com/locate/fluor
Catalytic vapor-phase oxidation of 2,2,2-trifluoroethanol
*
Hideyuki Mimura , Akio Watanabe, Kosuke Kawada
Research Laboratory, TOSOH F-TECH Inc., 4988 Kaisei-cho, Shunan, Yamaguchi 746-0006, Japan
Received 7 December 2005; accepted 12 December 2005
Available online 19 January 2006
Abstract
The synthesis of trifluoroacetaldehyde by vapor-phase oxidation of 2,2,2-trifluoroethanol using supported vanadium catalysts was studied.
Significant differences were observed in the reaction outcomes resulting from different types of catalysts. The ZrO2- and Al2O3-supported catalyst
demonstrated both high catalytic activity and selectivity. The addition of co-catalysts such as MoO3 or SnO2 improved catalytic performance
(Selectivity: up to 91%; S.T.Y.: >200 g LÀ1 hÀ1). The experimental results on catalyst lifetime showed a marked decrease in the activity of the
Al2O3-supported catalyst within tens of hours, while the ZrO2-supported catalyst showed little, if any, performance alterations for 2000 h.
# 2006 Elsevier B.V. All rights reserved.
Keywords: Vapor-phase oxidation; 2,2,2-Trifluoroethanol; Trifluoroacetaldehyde; Vanadium oxide
1. Introduction
easily with the products forming azeotropic liquids), the
separation and recycling of trifluoroethanol, the raw material,
is easier. So, utilizing this oxidation method a simplified and
advantageous process can be created.
Recently, the peculiar properties of fluorine have gained wide
industrial interest invarious fields, including the pharmaceutical,
agricultural chemicaland electronic materialsfields, andfluorine
has been incorporated into various functional organic molecules.
One of the practical methods of incorporating fluorine into
organic compounds is the conversion of building block
molecules containing fluorine into target compounds. Trifluor-
oacetaldehyde 2, which possesses a simple chemical structure
and a highly reactive aldehyde group, is one of the widely
applicable building blocks containing fluorine. In addition,
utilizing its pro-chiral property, the synthesis of CF3-containing
chiral compounds has also been explored [1–4].
For a large-scale industrial production of trifluoroacetalde-
hyde 2 using the method of oxidation of trifluoroethanol 1, the
vapor-phase continuous reaction is appropriate. Thevapor-phase
oxidation reaction has been long studied in the petrochemical
industry, but the reported cases of its application to fluorine-
containing compounds are extremely rare. One of the causes
thereof is that, in the case of compounds containing fluorine, the
electron density of the adjacent portion is low and the oxidation
reaction is not easy [9–11] due to the electronegativity of
fluorine. Regarding the vapor-phase oxidation reaction of
trifluoroethanol 1, there is a report about the supported V2O5
catalyst [8]. Even when 100% oxygen is used as the oxidizer, the
raw material conversion rate is as low as 18–41%.
For the synthetic methods of trifluoroacetaldehyde 2, several
kinds of procedures are known. Fluorination of trichloroace-
taldehyde in the vapor-phase [5], reduction of trifluoroacetic
acid or its esters in the vapor- or liquid-phase [6,7], and oxidation
of trifluoroethanol 1 [8] are among the major conventional
methods. In the method of oxidation for trifluoroethanol 1
(Scheme 1), by-products of which separation is difficult aren’t
generated, unlike in the case of the fluorination reaction in the
vapor-phase (which prone to produce incomplete fluorinated
compounds) [5]. Also in this oxidation method, compared with
the method of reduction of trifluoroacetic acid (which combines
This study investigated the most appropriate type of catalyst
for use in the vapor-phase oxidation of trifluoroethanol 1 from
the viewpoint of industrial production.
2. Results and discussion
2.1. Catalyst screening
The vapor-phase oxidation reactions of trifluoroethanol 1,
using various types of supported V2O5 (7 wt.%) catalysts, were
studied. The investigation was carried out under conditions of
* Corresponding author. Tel.: +81 834 62 1300; fax: +81 834 62 1303.
E-mail address: hideyuki-mimura@f-techinc.co.jp (H. Mimura).
0022-1139/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2005.12.012

520
H. Mimura et al. / Journal of Fluorine Chemistry 127 (2006) 519–523
Scheme 1.
excess air, with the concentration of trifluoroethanol 1 being
controlled under the explosive limit (5.5 vol.%), using air as an
oxidizer. The generated trifluoroacetaldehyde 2 was collected
in water as hydrate 4 for analysis.
Marked differences in reaction results were observed due to
the differences in the chemical formulations of the supports
(Table 1). The SiO2-supported catalyst manifested a low activity,
while the Al2O3-supported catalyst and the ZrO2-supported
catalyst gave good results for both catalytic activity and
selectivity over a wide range of surface area. The TiO2 (anatase
type)-supported catalyst showed the highest catalytic activity,
though it produced trifluoroacetic acid 3 in large proportion.
The differing degree of dispersion of V2O5 and its
interaction with the support, which resulted from the difference
in the support chemical formulations, presumably caused
varied effects on the oxidation reaction of trifluoroethanol 1.
The SiO2-supported catalyst has been reported to have larger
particles of V2O5 [12]. On the other hand, the TiO2 (anatase
type)-supported catalyst that showed high activity has been
reported to have V2O5 in a highly dispersed and stratified
Table 1
The influence of supports
Entry
Catalyst
Support
Name
Conversiona (%)
Selectivityb (%)
S.T.Y.c (g LÀ1 hÀ1
)
Surface area (m2/g)
4
3
d
1
2
3
4
5
6
7
8
V2O5/SiO2
Q-50
80
13
12
38
51
64
67
62
68
94
40
92
67
84
83
82
69
44
Trace
Trace
Trace
Trace
7.8
20
156
152
241
249
226
211
202
e
e
e
V2O5/Al2O3
V2O5/Al2O3
V2O5/Al2O3
SA3135
SA3177
KHO24
XZ16075
XZ16052
XZ16154
CS-300S
116
140
53
d
V2O5/ZrO2
d
V2O5/ZrO2
95
5.9
d
V2O5/ZrO2
258
69
Trace
41
d
V2O5/TiO2
a
The conversion rate was calculated from the quantitative values (the internal reference method) in 1 according to 19F NMR.
The selectivity was calculated from the quantitative values (the internal reference method) in 3 and 4 according to 19F NMR.
b
c
d
e
S.T.Y. (space time yield) was calculated using the following formula. S.T.Y. = (production weight of 4 per hour: g hÀ1)/(catalyst volume: L).
The catalyst was prepared by method A.
The catalyst was prepared by method B.
Table 2
The influence of co-catalysts
Entry
Catalyst
Support
Conversiona (%)
Selectivityb (%)
S.T.Y.c (g LÀ1 hÀ1
)
4
3
d
1
2
V2O5/Al2O3
d
V2O5–MoO3/Al2O3 (Mo/V = 1.0)
KHO24
64
54
97
62
63
72
74
80
67
64
85
84
91
84
82
87
80
85
82
83
91
80
Trace
1.8
6.4
5.9
9.1
6.9
5.8
7.7
7.8
7.0
15
241
219
369
226
246
254
278
293
249
257
305
KHO24
d
V2O5–SnO2/Al2O3 (Sn/V = 1.0)
3
KHO24
e
4
V2O5/ZrO2
e
V2O5–MoO3/ZrO2 (Mo/V = 1.0)
XZ16052
XZ16052
XZ16052
XZ16052
XZ16052
XZ16075
XZ16075
XZ16075
5
e
V2O5–WO3/ZrO2 (W/V = 1.0)
6
d
V2O5–TiO2/ZrO2 (Ti/V = 0.5)
7
d
V2O5–SnO2/ZrO2 (Sn/V = 1.0)
8
e
9
V2O5/ZrO2
e
V2O5–MoO3/ZrO2 (Mo/V = 1.0)
d
V2O5–SnO2/ZrO2 (Sn/V = 1.0)
10
11
a
The conversion rate was calculated from the quantitative values (the internal reference method) in 1 according to 19F NMR.
The selectivity was calculated from the quantitative values (the internal reference method) in 3 and 4 according to 19F NMR.
b
c
d
e
S.T.Y. (space time yield) was calculated using the following formula. S.T.Y. = (production weight of 4 per hour: g hÀ1)/(catalyst volume: L).
The catalyst was prepared by method B.
The catalyst was prepared by method A.

H. Mimura et al. / Journal of Fluorine Chemistry 127 (2006) 519–523
521
manner through crystallographic conformity between V2O5
(0 1 0) plane and TiO2 surface [13].
Next, the effect of co-catalyst metals was studied with regard
to the Al2O3-supported catalyst and the ZrO2-supported
catalyst (Table 2). Both catalysts showed enhanced selectivity
of trifluoroacetaldehyde hydrate 4 due to the addition of MoO3,
and increased catalytic activity due to the addition of SnO2. The
addition of WO3 or TiO2 also resulted in enhanced catalytic
activity.
The effects of the addition of these co-catalysts are also
reported for the vapor-phase partial oxidation reaction of
unfluorinated compounds (benzene [14], xylene [15], etc.). The
addition of other kinds of metal oxides such as MoO3 is
speculated to produce compound oxides [16], which affected
the reactions.
Fig. 2. Catalyst lifetime test: (*) V2O5–SnO2/Al2O3 (KHO24); (&) V2O5–
SnO2/ZrO2 (XZ16075); (~) V2O5–MoO3/ZrO2 (XZ16052); temperature: 260–
280 8C; catalyst: 20 mL; reaction tube diameter: 15 mm; concentration of 1:
2.2 vol.%; space velocity: 4300 hÀ1 (flow rate: L hÀ1/catalyst volume: L).
Since the catalytic site of the V2O5 oxidation reaction is
considered to be the double-bonded oxygen of V O [12], the
V O bonding status was investigated by measuring the infrared
absorption spectrum (Fig. 1). The results indicated absorption
at 1009 cmÀ1 for the V2O5/ZrO2 catalyst, which corresponded
to the V O stretching vibration, a shift toward a lower
wavenumber than that of unsupported V2O5 (1026 cmÀ1).
Furthermore, the V2O5–SnO2/ZrO2 catalyst which exhibited
higher catalyst activity showed absorption at 986 cmÀ1 for the
V O stretching vibration. These results suggested that the
supports and co-catalysts contributed to weakening the V O
bonding strength by interacting with V2O5, which facilitated
the oxidation process.
2.2. Catalyst lifetime
Catalyst lifetime was examined through continuous reac-
tions by using the V2O5–SnO2/Al2O3, V2O5–MoO3/ZrO2 and
V2O5–SnO2/ZrO2 catalysts which showed high performance
through catalyst screening (Fig. 2). The results indicated that
the V2O5–MoO3/ZrO2 and V2O5–SnO2/ZrO2 catalyst main-
tained catalyst performance for over 100 h. On the other hand,
the V2O5–SnO2/Al2O3 catalyst decreased the yield of
trifluoroacetaldehyde hydrate 4 as the reaction continued.
To search for the underlying causes, we analyzed the
retrieved catalysts (Table 3). The results of X-ray fluorescence
analysis revealed no compositional changes in the V2O5–
MoO3/ZrO2 catalyst, but showed a large increase in the quantity
of fluorine in the V2O5–SnO2/Al2O3 catalyst. This finding leads
to a conjecture that the fluorination of the Al2O3 support by HF,
a by-product generated during the course of the reaction, caused
performance changes. In contrast, when ZrO2 was used as a
support, the catalytic performance was maintained without
being influenced by the generated by-products.
To further grasp the potential of the V2O5–SnO2/ZrO2
catalyst, the temperature of the catalyst layer was increased to
320 8C, and continuous reactions were carried out under
conditions of high loading with an elevated conversion rate
(Fig. 3). Under these circumstances, the observed sustained
Table 3
The result of X-ray fluorescence analysis
Reaction time (h)
V2O5
ZrO2
MoO3
HfO2
F (wt.%)
V2O5–MoO3/ZrO2 (XZ16052)
6
5.9
5.8
83.0
82.9
9.5
9.6
1.6
1.7
Not detected
Not detected
266
Reaction time (h)
V2O5
Al2O3
SnO2
F (wt.%)
V2O5–SnO2/Al2O3 (KHO24)
6
8.8
7.6
88.0
77.5
3.2
2.4
Not detected
12.5
Fig. 1. Infrared absorption spectra of catalyst: (a) V2O5 (not supported); (b)
V2O5/ZrO2 (XZ16052); (c) V2O5–SnO2/ZrO2 (XZ16052).
72

522
H. Mimura et al. / Journal of Fluorine Chemistry 127 (2006) 519–523
4.2. General procedure for preparation of catalyst
Catalysts were prepared according to known methods
[17,18].
4.2.1. Method A
NH4VO3 (4.4 g, 0.038 mol) (and co-catalyst metal source)
was added to a flask containing pure water (210 g) and dissolved
byheating.Then, support(45 g)wasaddedand, afterevaporating
water while stirring, the mixturewas dried at 150–200 8C for 1 h.
The obtained catalyst was placed into a stainless steel tube of
15 mmdiameterandbakedfor3 hat450–500 8Cinanairstream.
4.2.2. Method B
NH4VO3 (4.4 g, 0.038 mol) (and co-catalyst metal source)
and oxalic acid (6.8 g, 0.076 mol) were added to a flask with pure
water (20 g) in it, and dissolved by reduction. Then, support
(45 g) was added and, after evaporating water while stirring, the
mixture was dried at 150–200 8C for 1 h. The obtained catalyst
was placed into a stainless steel tube of 15 mm diameter and
baked for 3 h at 450–500 8C in an air stream.
Fig. 3. Lifetime test of V2O5–SnO2/ZrO2 (XZ16075) catalyst: (*) conversion
of 1; (~) selectivity of 4; (&) yield of 4; temperature: 290–320 8C; catalyst:
116 mL; reaction tube diameter: 25 mm; concentration of 1: 3.2 vol.%; space
velocity: 2900 hÀ1 (flow rate: L hÀ1/catalyst volume: L).
performance lasted for 2000 h, indicating a feasibility of
industrializing the use of this catalyst.
4.3. General procedure for vapor-phase oxidation
3. Conclusion
Catalyst (20–24 mL) was placed into a 15 mm diameter
stainless steel reaction tube, and three reaction gas absorption
containers filled with 200 g of water were connected to the outlet
of the reaction tube. After setting the temperature of the catalyst
layer to 250–260 8C while in the air stream at 1.4–1.7 L/min,
2,2,2-trifluoroethanol was fed at 0.13–0.16 g/min. The reaction
elevated the temperature of the catalyst layer by 10–20 8C.
After doing the reaction for 4 h, the reaction became
stabilized. Then the gas absorption containers were replaced
and the reaction was continued for another 2 h. The gas
absorption liquid obtained for a period of 4–6 h was analyzed
by 19F NMR with acetone-d6 as the solvent and benzotri-
fluoride as the internal standard.
The supported V2O5 catalyst produced trifluoroacetaldehyde
hydrate 4 with high selectivity in the vapor-phase oxidation of
trifluoroethanol 1 with air as the oxidation agent. Co-catalysts
and catalyst supports caused transition in the V O bonding
status, which is the catalytic site, exerting a significant effect on
the reaction process. The successful oxidation of trifluoroetha-
nol 1 requires a suitable selection of catalyst supports that are
inert to fluorination by HF, an unfavorable by-product. The
V2O5–SnO2/ZrO2 catalyst, possessing high catalytic activity
and durability, offers good prospects as an efficient industrial
catalyst.
4. Experimental
4.4. Measurement of infrared absorption spectra
4.1. Materials and apparatus
A few catalyst pellets were ground into fine powder and
mixed with KBr powder, and then measurements were
performed by diffuse reflection spectroscopy.
The Al2O3 supports were supplied by Sumitomo Chemicals
Co. Ltd. (KHO24) and Saint Gobain NorPro (SA3135, SA3177).
The ZrO2 supports (XZ16052, XZ16075, and XZ16154) were
supplied by Saint Gobain NorPro. The SiO2 support (Q-50) was
supplied by Fuji Silysia Chemical Ltd. The TiO2 support (CS-
300S) was supplied by Sakai Chemical Industry Co. Ltd.
NH4VO3 (Wako Pure Chemical Industries), (NH4)6
Mo7O24Á4H2O (Wako Pure Chemical Industries), SnO2
(Aldrich Company), (NH4)10W12O41Á5H2O (Wako Pure Che-
mical Industries), and TiO2 (Wako Pure Chemical Industries)
were used as the source of V, Mo, Sn, W, and Ti, respectively.
The 19F NMR spectrum and the infrared absorption
spectrum were recorded on the JEOL FX90Q model
(84.25 MHz) and Shimadzu FTIR-8100 model, respectively.
The X-ray fluorescence analysis was performed on JEOL JSX-
3200 model.
4.5. X-ray fluorescence analysis
A few catalyst pellets were ground into fine powder and
formed into a pellet, and then measurements were performed
(Voltage 30 kV; Live time 600 s).
References
[1] K. Mikami, Y. Itoh, M. Yamanaka, Chem. Rev. 104 (2004) 1–16.
[2] P.V. Ramachandran, K.J. Padiya, V. Rauniyar, M.V.R. Reddy, H.C. Brown,
Tetrahedron Lett. 45 (2004) 1015–1017.
[3] H. Torii, M. Nakadai, K. Ishihara, S. Saito, H. Yamamoto, Angew. Chem.
Int. Ed. Engl. 43 (2004) 1983–1986.
[4] D. Enders, K. Funabiki, Org. Lett. 3 (2001) 1575–1577.

H. Mimura et al. / Journal of Fluorine Chemistry 127 (2006) 519–523
523
[5] T. Maruta, T. Nishinomiya, Y. Tasaka, JP Patent 91-181432 (1991).
[12] A. Miyamoto, Y. Murakami, Petrotech 6 (1983) 579–583.
[13] A. Vejux, P. Courtine, J. Solid State Chem. 23 (1978) 93–103.
[6] O.R. Pierce, T.G. Kane, J. Am. Chem. Soc. 76 (1954) 300–301.
[7] R. Jacquot, WO Patent 97/17134A1 (1997).
[14] S.K. Bhattacharyya, N. Venkataraman, J. Appl. Chem. 8 (1958) 728–737.
[15] S.K. Bhattacharyya, I.B. Gulati, Ind. Eng. Chem. 50 (1958) 1719–1726.
[16] R.H. Munch, E.D. Pierron, J. Catal. 3 (1964) 406–413.
[17] M. Wada, I. Yanagisawa, M. Ninomiya, T. Ohara, JP Patent 74-11371
(1974).
[8] M. Braid, S. Township, M. County, F. Lawlor, US Patent 3038936 (1962).
´ ´
[9] V. Kesavan, D.B. Delpon, J.P. Begue, A. Srikanth, S. Chandrasekaran,
Tetrahedron Lett. 41 (2000) 3327–3330.
[10] R.J. Linderman, D.M. Graves, J. Org. Chem. 54 (1989) 661–668.
[11] P. Bernardelli, WO Patent 01/72667A2 (2001).
[18] T. Norisugi, JP Patent 53-6274 (1953).