(22d) A Highly Active Supported Fecuk Fischer-Tropsch Catalyst

A
highly active supported FeCuK Fischer-Tropsch catalyst

Kamyar
Keyvanloo, Calvin H. Bartholomew, and William C. Hecker

Chemical
Engineering Department, Brigham Young University, Provo, UT, 84602

Introduction:

Precipitated
FeCuK catalysts are effective for production of high molecular hydrocarbons in
fixed bed reactors [1]. Unfortunately, despite their high activity and
selectivity, they lack sufficient mechanical strength to be used in slurry
bubble column reactors (SBCRs) [2]. These catalysts undergo attrition to fine
particles leading to loss of the catalyst due to difficulty in catalyst/wax
separation. To alleviate this particle breakage problem, it is proposed that a support
material be used to create a cost effective supported Fe catalyst.  Supported
iron catalysts produced in the past have usually suffered from lower activity
and higher methane selectivity compared to precipitated iron catalysts. To
overcome this problem, an intensive study on preparation variables such as type
of support, preparation method, iron loading, and promoter type and loading is
needed.

In
this study we report our on-going effort to develop an active and selective
supported iron catalyst. To date, an active alumina-supported catalyst has been
prepared, which, compared to others reported in the literature [4-6], is one of
the most active supported iron catalysts ever made.  The activity and
selectivity are also compared to those of carbon nanotube (CNT) supported
catalysts and unsupported iron catalysts previously synthesized in the BYU
Catalysis lab [3]. To accomplish our work on the preparation of an active and
selective supported iron catalyst, silica and SiC supports, different
pretreatments, and different preparation variables/ methods will be explored.

Materials and method:

For
the alumina supported catalysts, alumina was sieved to 30-60 mesh and calcined
at 700 °C in air for 4h prior to incipient wetness impregnation. Catalysts
were prepared by co-impregnation with aqueous solutions containing desired
amounts of ferric nitrate and copper nitrate in successive steps. The sample
was dried overnight at 80°C and calcined at 300°C for about 16 h. Potassium
was then added by impregnating with potassium bicarbonate. Nominal compositions
(on a relative mass basis) of synthesized catalysts were 100Fe/ 7.5Cu/4 K/400 Al₂O₃
(Al/Fe/Cu/4K) and 100 Fe/7.5 Cu/8 K/400 Al₂O₃
(Al/Fe/Cu/8K).

The
unsupported iron catalyst (Fe/Cu/K/SiO2) was prepared from iron and copper
nitrate salts by a simple, proprietary, co-precipitation method developed by
Cosmas, Inc. Potassium (KHCO₃) and silica (Cab-O-Sil) promoters were
added to the wet precursor before the catalyst was dried. Nominal catalyst
composition was 100 Fe/5 Cu/4 K/16 SiO₂ by mass. This method was
modified to prepare a supported iron catalyst on CNT (CNT/Fe/K/Cu) that
contained 30% iron loading.

Activity
studies were performed in a 3/8 inch ID fixed-bed reactor. 0.25 g of calcined
catalyst (30 ? 60 mesh) was mixed with 1 g of quartz sand (50 ? 70 mesh) and charged
to the reactor. Catalysts were reduced in situ at 280-340°C with GHSV > 2000
h^-1 and 10% H₂/He for 10 hours followed by 100% H₂
for 6 hours. Activation and reaction conditions were 300 psig, H₂:CO=1.
Activation was at 280°C for 24-48 hours in syngas. Reaction temperatures were
varied from 230°C to 260°C. The reactor effluent was analyzed online by an HP
6890 GC with a 15 ft x 1/8 inch SS Supelco column packed with 60/80 carboxen
1000 phase.

Results and discussion:

Table 1 shows BET results and extent of reduction on four
catalysts prepared to date. The carbon nanotube-supported (CNT) catalyst had
the highest BET surface area after calcination followed by alumina-supported
and unsupported catalyst. The alumina-supported catalysts were fairly reducible
with extents of reduction of 43-45% after reduction in H₂ at 300
°C for 16 h, but only half as reduced as the unsupported catalyst.

Table 1.  Textural properties of catalysts

Reaction
rates at 250°C for the 4 catalysts are given in Figure 1. The highest activity
was obtained on alumina-supported catalyst with 4 parts K with a CO depletion
rate of 38 mmol/gcat/h. On the other hand, this catalyst also had the highest
methane selectivity (12-16 mol% - CO₂-free basis) followed by
unsupported iron, Fe/Cu/K/SiO₂ (7 mol%) and CNT-supported (5.1
mol%). Higher selectivity of methane on alumina can be explained by acidic
sites on alumina compared to the inert surface of CNT, which provide cracking
sites for production of methane. To our surprise, alumina-supported catalyst
with the higher amount of potassium (8 vs. 4%) didn't change methane selectivity
significantly, while the activity decreased slightly.

Fig. 1.
CO conversion rate for different supports and unsupported catalyst.

The
alumina supported iron catalyst prepared with the impregnation method compares
favorably with catalysts described in the literature. Table 2 compares our
catalyst with one unsupported and two supported catalysts from the literature.
In order to obtain a quantitative comparison of catalyst activities to account
for different gas space velocities a first-order FTS reaction is assumed. Bukur
et al. [4-6] compared the performance of their best most active precipitated
catalysts (TAMU1) with silica-supported iron catalyst (TAMU2) and
alumina-supported catalyst (TAMU3). They reported that the silica-supported
catalyst was two-fold less active than precipitated catalyst if they compared
per gram catalyst (100 vs. 221 mmol(CO+H₂)/g cat/MPa/h). They also
showed that the alumina-supported catalyst had nearly 50% less activity than
silica-supported catalyst. Alumina-supported catalyst prepared in the current work is nearly two-fold more active than the
silica-supported catalyst reported by Bukur. Catalyst activity at 260
°C was 183 mmol (H₂+CO)/g cat/MPa/h compared with 221 mmol (H₂+CO)/g
cat/MPa/h for Bukur's most active unsupported
catalyst. In addition, catalyst productivity was 0.49 gHC/g cat/h compared with
0.51 gHC/g cat/h (TAMU1). The catalyst activity comparison was even more
favorable if they are compared per g Fe (915 vs. 370 mmol (H₂+CO)/g
cat/MPa/h). On the other hand, the methane selectivity (CO₂-free
basis) was higher than those in the literature (16 vs. 3-7 mol%). This catalyst
shows great promise and work continues on optimizing catalyst preparation,
performance, and stability.

Table 2. Comparison of catalyst performance

¹100Fe/3Cu/4K/16SiO₂

²100Fe/5Cu/6K/139SiO₂

³100Fe/5Cu/9K/139Al₂O₃

References:

[1] D.B. Bukur, X. Lang, D. Mukesh, W.H. Zimmerman,
M.P. Rosynek, and C. Li, Binder/support effects on the activity and selectivity
of iron catalysts in the Fischer-Tropsch synthesis, Ind. Eng. Chem. Res. 29
(1990) 1588.

[2] D.S. Kalakkad, M.D. Shroff, S. Kohler, N.
Jackson, and A.K. Datye, Attrition of precipitated iron
Fischer-Tropsch catalysts, Appl. Catal. A: Gen. 133 (1995)
335.

[3] K.M. Brunner, K. Keyvanloo, C.H. Bartholomew,
W.C. Hecker, A Simple and Novel Preparation Method For Iron FT Catalysts, ACS
Meeting, San Diego, March 25-29, 2012.

[4] D.B. Bukur and X. Lang, Highly active and stable
iron Fischer-Tropsch catalyst for synthesis gas conversion to liquid fuels,
Ind. Eng. Chem. Res. 38 (1999) 3270.

[5] D.B. Bukur and C. Sivaraj, Supported iron
catalysts for slurry phase Fischer-Tropsch synthesis, Appl. Catal. A: Gen. 231
(2002) 201.

[6] D.B. Bukur, X. Lang, and Y. Ding, Pretreatment
effect studies with a precipitated iron Fischer-Tropsch catalyst in a slurry
reactor, Appl. Catal. A: Gen. 186 (1999) 275.