Synthesis of a porous chromium fluoride catalyst with a large surface area

Article abstract of DOI:10.1016/j.jcat.2005.01.028

A novel approach to the preparation of a porous chromium fluoride catalyst with a large surface area is reported. The pores were generated by introduction of a siliceous material into the precursor of the catalyst and then removal of the material by reaction with anhydrous hydrogen fluoride. During the reaction, the formation and escape of a volatile gas (SiF₄) from the precursor enlarged the surface area of the chromium fluoride. This process provided for the first time a porous chromium fluoride with a surface area of 187 m ²/g and a pore volume of 0.58 cm³/g. Furthermore, the porous chromium fluoride exhibited excellent chemical stability in the presence of HCl, HF, and F₂. It was catalytically active for halogen exchange in a fixed-bed fluorination reaction, and it exhibited excellent catalytic performance in mitigating the coke formation on the surface of the catalyst during vapor-phase catalytic fluorination.

Full text of DOI:10.1016/j.jcat.2005.01.028

Journal of Catalysis 231 (2005) 254–257
www.elsevier.com/locate/jcat
Synthesis of a porous chromium fluoride catalyst with a large surface area
∗
Heng-dao Quan , Hui-e Yang, Masanori Tamura, Akira Sekiya
National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 5-2, 1-1-1, Higashi, Tsukuba, Ibaraki 305-8565, Japan
Received 2 November 2004; revised 21 January 2005; accepted 24 January 2005
Abstract
A novel approach to the preparation of a porous chromium fluoride catalyst with a large surface area is reported. The pores were gener-
ated by introduction of a siliceous material into the precursor of the catalyst and then removal of the material by reaction with anhydrous
hydrogen fluoride. During the reaction, the formation and escape of a volatile gas (SiF ) from the precursor enlarged the surface area of the
4
2
chromium fluoride. This process provided for the first time a porous chromium fluoride with a surface area of 187 m /g and a pore volume
of 0.58 cm /g. Furthermore, the porous chromium fluoride exhibited excellent chemical stability in the presence of HCl, HF, and F . It was
3
2
catalytically active for halogen exchange in a fixed-bed fluorination reaction, and it exhibited excellent catalytic performance in mitigating
the coke formation on the surface of the catalyst during vapor-phase catalytic fluorination.

2005 Elsevier Inc. All rights reserved.
Keywords: Vapor-phase catalytic fluorination; Porous chromium fluoride; Porous metal fluorides; Organo-fluorine compounds; Difluoromethane (HFC-32);
Anhydrous hydrogen fluoride (AHF)
1
. Introduction
route to X-ray amorphous metal fluorides with surface areas
of approximately 206 m /g by means of the reaction of an
2
organic intermediate with HF [15,16].
Porous materials are widely used in catalysis, molecular
Chromium-based catalysts are important for the produc-
tion of several compounds [17]. Fluorinated chromia has
been a foundational pillar of vapor-phase catalytic fluori-
nation reactions, and its preparation has been investigated
extensively [18–20]. Because of the higher activity of high-
surface-area catalysts, lower reaction temperatures can be
used, which minimizes side reactions and increases selec-
tivity for the desired product [21].
separation technologies, selective reaction, optics, and solar
cell materials [1–5]. Porous materials are usually prepared
by calcination of metal salts or hydrous oxides [6], aero-
gel and sol-gel processes [7,8], and micellar systems [9,10].
Because most porous materials are composed of silicates
and metal oxides, they are unstable; the porous structure
collapses in the presence of corrosive media, such as HCl,
HF, and F2. Porous metal fluorides are inert, but existing
metal fluorides have only a small surface area [11–14]. High-
surface-area metal fluorides have been exploited for their
high activity and long catalyst life. Stacy et al. have reported
a novel method for converting zeolites to high-surface-area
AlF3 by means of plasma-assisted fluorination with nitrogen
trifluoride. Kemnitz et al. developed a nonaqueous synthetic
Here we report a facile preparation of a porous chromium
fluoride (PCrF) catalyst with a large surface area. The pores
were generated by the introduction of a siliceous material
into the precursor of the catalyst and then removal of the ma-
terial by reaction with anhydrous hydrogen fluoride (AHF).
During removal of the siliceous material, the escape of a
volatile gas (SiF4) from the precursor enlarged the surface
area of the PCrF [22]. The resulting amorphous PCrF had
2
3
a 187.0 m /g surface area and 0.58 cm /g pore volume. In
*
addition, the new PCrF exhibited excellent thermal stabil-
ity below 500 C; chemical stability in the presence of HCl,
Corresponding author. Fax: +81-298-4771.
E-mail address: Hengdao-quan@aist.go.jp (H.-d. Quan).
◦
0021-9517/$ – see front matter  2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.01.028

H.-d. Quan et al. / Journal of Catalysis 231 (2005) 254–257
255
HF, and F2; and high catalytic performance in mitigating the
coke formation on the surface of the catalyst.
3. Results and discussion
The results of BET analysis (Table 1) indicate that the
addition of Si increased the surface area and pore volume of
the PCrF and its precursor. When the percentage weight of Si
relative to Cr was approximately 3%, the surface area and the
2
. Experimental
2
3
The Brunauer–Emmett–Teller (BET) surface areas, pore
pore volume of precursor were 470.0 m /g and 0.79 cm /g,
volumes, and Barrett–Joyner–Halenda (BJH) adsorption
pore distributions were determined for the samples by
means of low-temperature adsorption of nitrogen with a Mi-
cromeritics ASAP 2010 instrument. The bulk crystalline
phase of the samples was determined by X-ray diffrac-
tion (XRD) with a Mac Science, MPX-18 diffractometer.
Differential scanning calorimetry and thermogravimetry pat-
terns of the samples were recorded, respectively, on Rigaku
Thermoplus DSC-8230 and Thermoplus 8120 instruments.
Scanning electron microscopy and energy-dispersive X-ray
respectively. After fluorination of the siliceous precursor by
◦
HF at 350 C for 100 h, the surface area and pore volume
2
3
of PCrF were 187.0 m /g and 0.58 cm /g, respectively. The
surface area of the chromium fluoride prepared by means of
this novel approach was twice that of chromium fluoride pre-
pared by the conventional method. This increase is ascribed
to the fact that the formed SiF4 escaped from the precur-
sor, thus enlarging the surface area during the reaction of the
siliceous precursor and HF. SiF4 is reported to be formed by
the reaction of SiO2 and gaseous HF at high temperature in
a fluidized bed reactor [22]. The surface area of PCrF de-
(
EDX) analysis of samples were performed on a Hitachi
2
S4800 field-emission scanning electron microscope (SEM).
Representative transmission electron micrographs of sam-
ples were recorded with a JEOL JEM 2010 transmission
electron microscope (TEM) operated at 200 kV. The typical
process for preparing PCrF and its precursor is as follows:
creased to 149.4 m /g when the Si weight percentage was
increased to 5%.
The pore-diameter distribution was calculated from the
adsorption isotherm branch of the nitrogen isotherms by
the BJH method. The results of the analysis indicate that
the pore size distributions of PCrF and its precursor were
wide, ranging from 20 to 300 Å. The wide pore size distri-
bution might be attributable to the fact that control of the
co-precipitating process was difficult because the chromic
hydroxide was produced rapidly in the presence of ammonia
solution. The particle size of the formed chromic hydrox-
ide varied drastically. When the calcined chromic hydroxide
was treated with HF, the SiO2 in the precursor was removed,
leaving a wide pore size distribution.
Na2SiO3 (9 g) was dissolved in 14 ml of 28% aqueous
ammonia solution, and the solution was added dropwise to
Furthermore, the SEM/EDX integrated analysis indicated
that several kinds of particles coexisted in the siliceous pre-
cursor, some containing 6.1 wt% Si and some containing
2
40 g of a rapidly stirred 30% CrCl3 solution. The pH of
the solution was approximately 7.5. The resulting slurry of
hydroxides was filtered, washed thoroughly with de-ionized
water, dried, ground, and pelleted. The pellets were calcined
3
.5 wt% Si. EDX analysis confirmed that the siliceous pre-
cipitate of chromic hydroxide was not a uniform mixture,
and this nonuniformity led to the different size pores in the
PCrF. The SEM images (Figs. 1a and 1b) show clearly that
the amorphous PCrF and its precursor were piled up in irreg-
ularly shaped particles and were composed of pores with dif-
ferent diameters. The pore size distribution was not uniform,
◦
at 400 C under nitrogen and then treated with anhydrous
◦
hydrogen fluoride (AHF) at 350 C for 100 h. Finally, the
residual AHF in the PCrF was purged with nitrogen. For the
preparation of difluoromethane (HFC-32), dichloromethane
(0.3 g/min) and vapor-phase AHF (300 ml/min) were fed
into a reactor charged with 10 ml of the prepared PCrF cata-
lyst. The product stream from the reactor was scrubbed with
H2O at 60 C and then passed first through a dryer packed
with CaCl2 and then through a gas chromatograph (Shi-
madzu GC-14 A on-line) with a Poraplot Q capillary column
Table 1
◦
Characterization of PCrF and its precursor by the BET method
a
Sample
Si
(
Surface area
(m /g)
Vp
Mean pore
radius (Å)
2
3
wt%)
(cm /g)
(
i.d. 0.32 mm; length 25 m; J & W Scientific Inc.). The col-
CrSi-0
F–CrSi-0
CrSi-1
F–CrSi-1
CrSi-3
F–CrSi-3
CrSi-5
0
298.9
90.2
0.48
0.30
0.62
0.47
0.79
0.58
0.40
0.45
32.3
66.9
31.0
51.2
33.9
55.8
60.9
60.3
◦
umn temperature program was as follows: 80 C for 15 min;
2
and the thermal conductivity detector were maintained at
2
1
yields of CH2FCl and CH2F2, their GC relative response fac-
tors (CH2F2/CH2FCl/CH2Cl2 = 1:1.23:1.67) were used in
the calculations [23].
◦ ◦
1
3
5
399.2
184.3
470.0
187.0
354.7
149.4
0 C/min to 200 C; hold for 5 min. Both the injection port
◦
00 C, and the carrier gas was He introduced at a rate of
0 ml/min. To determine the conversion of CH2Cl2 and the
F–CrSi-5
a
These values are the wt% of Si to Cr in the precursor of PCrF; F–: CrSi
fluorinated with HF.

256
H.-d. Quan et al. / Journal of Catalysis 231 (2005) 254–257
Fig. 2. TEM image of PCrF.
which is consistent with the BET results obtained by the
BJH method. The surface of the PCrF appeared to be downy,
which is ascribed to the fact that the addition of silicon in
chromia obviously enlarged the surface area of PCrF when
the siliceous precursor reacted with AHF. EDX/SEM analy-
sis indicates that the PCrF consisted mainly of Cr and F,
a certain amount of O, and trace amount of Si when the
(a)
◦
precursor was treated with AHF at 350 C for 100 h. The
presence of oxygen promotes the activity of the catalyst [24].
Fig. 2 shows TEM images of the PCrF. The TEM image can
be seen in domains of about 10 nm, which indicates that the
PCrF was composed of differently sized pores with irregu-
lar shapes. Again, the image is consistent with the pore size
distribution results obtained by BET analysis.
DSC and TG analyses showed that the thermal stability
of the PCrF was much better than that of conventionally pre-
pared chromium fluoride. The PCrF showed no obvious de-
◦
composition until 600 C, whereas the traditional chromium
(b)
◦
fluoride began to decompose at 570 C. The loss of TG wt%
Fig. 1. (a) Representative scanning electron micrograph of the siliceous pre-
of PCrF was much less than that of traditional chromium flu-
oride in the range from 50 to 500 C. We speculate that the
thermal stability of the catalyst was improved by the pres-
2
cursor with a 470.0 m /g surface area. (b) Representative scanning electron
◦
2
micrograph of PCrF with a 187.0 m /g surface area.
Fig. 3. Comparison of PCrF and traditional chromium fluoride in the reaction of dichloromethane: catalyst, 10 ml; HF, 300 ml/min; CH Cl , 80 ml/min;
2
2
−
1
GHSV, 2280 h ; contact time, 1.6 s.

H.-d. Quan et al. / Journal of Catalysis 231 (2005) 254–257
257
ence of residual trace amounts of Si. XRD results indicate
that PCrF and its siliceous precursor heated at 120 or 400 C
Ming-deng Wei for his help in measuring X-ray diffraction
patterns.
◦
under nitrogen were X-ray amorphous. As reported in the lit-
erature [25], amorphous chromium fluoride is advantageous
for F/Cl exchange in vapor-phase catalytic fluorination.
In the preparation of difluoromethane (HFC-32), our
PCrF catalyst exhibited higher activity than conventional
chromium fluoride (Fig. 3). We ascribe these results to the
fact that the larger surface area of PCrF was advantageous
for F/Cl exchange in the vapor-phase catalytic fluorination.
Coke formation on the catalyst surface during the prepara-
tion of HFC-32 is responsible for deactivation of the catalyst
in vapor-phase catalytic fluorination. We found that after
References
[
1] Y.F. Lu, R. Gangguli, C.A. Drewien, M.T. Anderson, C.J. Brinker,
W.L. Gong, Y.X. Guo, H. Soyez, B. Dunn, M.H. Huang, J.I. Zink,
Nature 389 (1997) 364.
[2] U. Ueno, H. Horiuchi, M. Tomita, O. Niwa, Anal. Chem. 74 (2002)
5257.
[3] H.D. Quan, M. Tamura, J. Murata, R.X. Gao, A. Sekiya, J. Fluorine
Chem. 106 (2000) 121.
[
[
4] H.S. Zhou, I. Honma, Adv. Mater. 11 (1999) 683.
5] Do Trong On, Langmuir 15 (1999) 8561.
1
00 h of reaction, the PCrF prepared from the siliceous pre-
[6] F. Schüth, K.S.W. Sing, J. Weitkamp, Handbook of Porous Solids,
vol. 1, Wiley–VCH, Weinheim, 2002, p. 4.
cursor exhibited less coke formation than did chromium flu-
oride prepared from chromia when both types of chromium
fluoride were used as process catalysts in the preparation of
HFC-32. As indicated by EDX analysis, the carbon amount
on the surface of used catalysts increased to approximately
[7] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M.
Winfield, J. Fluorine Chem. 110 (2001) 181.
[
[
8] P. Chanho, G.L. Haller, Micropor. Mesopor. Mater. 48 (2001) 165.
9] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D.
Schmitt, C.T.-W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen,
J.B. Higgins, J.I. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834.
3
wt% in PCrF and 12 wt% in conventional chromium fluo-
[
10] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Na-
ture 359 (1992) 710.
11] H.D. Quan, M. Tamura, J. Murata, R.X. Gao, A. Sekiya, J. Fluorine
Chem. 99 (1999) 167.
ride. The decrease in coke formation may be attributable to
the decreased Lewis acid strength of the Cr-based catalyst
caused by trace amounts of Si in the catalyst. In fact, weak
Lewis acid strength is advantageous for decreasing coke for-
mation, which prolongs the lifespan of the catalyst. When
the PCrF was used for 350 h in a fluorination reaction of
dichloromethane, the efficiency of the catalyst decreased to
[
[12] H.D. Quan, M. Tamura, R.X. Gao, A. Sekiya, Tetrahedron 57 (2001)
111.
4
[
[
[
[
13] M. Wojciechowska, B. Czajka, M. Pietrowski, M. Zieli n´ ski, Catal.
Lett. 66 (2000) 147.
14] H.D. Quan, M. Tamura, J. Murata, R.X. Gao, A. Sekiya, J. Fluorine
Chem. 116 (2002) 65.
15] J.L. Delattre, P.J. Chupas, C.P. Grey, A.M. Stacy, J. Am. Chem.
Soc. 123 (2001) 5364.
16] E. Kemnitz, U. Groß, S. Rüdiger, C.S. Shekar, Angew. Chem. 42
8
5% of the original value. The components (wt%) on the cat-
alyst surface determined by EDX analysis were as follows:
Fresh catalyst: C 4.2, O 6.2, F 58.4, Si 0.8, Cr 30.4, F/O 9.4,
F/Cr 1.9; Used catalyst after 350 h: C 25.0, O 3.0, F 54.9,
Si 0.2, Cr 16.9, F/O 18.4, F/Cr 3.3. The results indicate that
the decrease in the amount of oxygen on the catalyst surface
is another factor in catalyst deactivation.
(2003) 4251.
[17] C. Pak, G.L. Haller, Micropor. Mesopor. Mater. 48 (2001) 165170.
18] D.R. Coulson, P.W.J.G. Wijnen, J.J. Lerou, L.E. Manzer, J. Catal. 140
[
(1993) 103.
[
[
[
[
[
19] Swamer, F.W.U.S. Patent 3,258,500 (1966), to El du Pont Company.
20] T. Skapin, E. Kemnitz, J. Non-Crystall. Sol. 225 (1998) 163.
21] L.E. Manger, M.J. Nappa, Appl. Catal. A: Gen. 221 (2001) 267.
22] Gmelin Handbook, Si suppl. B, vol. 7, 1992, p. 117.
23] H.D. Quan, M. Tamura, Y. Matsukawa, J. Mizukado, T. Abe, A. Se-
kiya, J. Mol. Catal. A: Chem. 219 (2004) 79.
Acknowledgments
[
[
24] S.B. Brunet, D. Martin, J. Catal. 171 (1997) 284.
25] S. Akashi, I. Yoshio, K. Satoshi, EP 0514932 (1992), to Daikin Ind.
Ltd.
We thank Dr. Zhi-heng Wu and Dr. Bo Yang for their
assistance in obtaining SEM images of samples, and Dr.