(640g) Ethylene Oligomerization Catalyst Optimization Using Fundamental Kinetic Modeling

Ethylene Oligomerization Catalyst
Optimization Using Fundamental Kinetic Modeling

Kenneth Toch¹, Joris W. Thybaut^1,*,
Maria A. Arribas², Agustin Martinez², Guy B. Marin¹

¹
Laboratory for Chemical Technology, Ghent University, Krijgslaan 281 ? S5, 9000
Ghent, Belgium

² Instituto
de Tecnologica Quimica, Universidad Politecnica de Valencia, Av. de los
Naranjos, E-46022 Valencia, Spain

Homogeneously
catalyzed ethylene oligomerization is an established industrial process, mainly
resulting in even carbon atom numbered alpha alkenes. Heterogeneous catalysis
is preferred, however, e.g., because of environmental reasons and for avoiding
a catalyst separation step from the product. Moreover, heterogeneous catalysis
offers more opportunities to tune the product distribution and corresponding
yields to the market demands, i.e., even-numbered, linear α-alkenes versus
a high octane fuel blend hydrocarbon mixture [1].

The present work
aims at developing an active, stable and selective catalyst for ethylene
oligomerization. This challenge is addressed via fundamental modeling using
Single-Event MicroKinetics (SEMK) [2]. Model parameters are determined which
can be classified as kinetic and catalyst descriptors. The first type of
parameters is specific to the reaction family considered and independent from
the catalyst, while the latter specifically accounts for the effect of catalyst
properties, such as Si/Al ratio, metal-ion site concentration? on the kinetics,
e.g., via chemisorption enthalpies, sterical constraint factors,? Through simulation,
the catalyst descriptors can be varied in order to identify optimal catalyst
behavior defined in terms of product selectivity or yield.

Experimental
ethylene oligomerization data have been gathered on three different catalysts,
i.e. 1.8wt% Ni-SiO₂-Al₂O₃, 2.7wt% Ni-Beta and 1.7wt%
Ni-USY. At 393 K, a total pressure of 3.5 MPa and a space time equal to 48 kg_cat
s mol_eth^-1, see Figure 1 a, stable behavior up to 9h time
on stream was exhibited by Ni-SiO₂-Al₂O₃ and
Ni-Beta, whereas Ni-USY suffered a significant activity loss. Ni-SiO₂-Al₂O₃
was more active than Ni-Beta and, hence, was selected as the benchmark catalyst
for modelling purposes. As observed in Figure 1b, reaction products found
mainly comprised even-numbered olefins in the C₄-C₁₂
range. Only traces of odd-numbered alkenes were founded at the reaction
conditions studied. A different product distribution was observed over the
silica-alumina-based and microporous Beta-based catalysts. A Schulz Flory
distribution was obtained with the Ni-SiO₂-Al₂O₃
catalyst. The C₁₀₊ fraction products (diesel products) was 38 wt%
and 17wt % for the Ni-Beta and Ni-SiO₂-Al₂O₃ catalysts,
respectively, probably because the oligomerization activity over the acid sites
is more important over the more acidic zeolite-based catalyst.

A more extended
ethylene oligomerization kinetic data set has been acquired on the 1.8wt%Ni-SiO₂-Al₂O₃
within a temperature range from 323 K to 503 K, total pressures between 1.0 and
3.5 MPa with a molar ethylene content in the feed of 60%. The space time was
varied from 3.0 to 45.0 kg_cat s mol_eth^-1. At
these conditions, intrinsic kinetics are measured. The
reaction network considered in the SEMK model was limited to molecules with a
maximum carbon number of 12 and contained over 3000 species and over 10000
elementary steps. Physical adsorption of the components inside the catalyst pores,
prior to any chemical elementary step was also accounted for.

As previously reported
in the literature [1;3], two distinct activity regions were observed as a function
of the temperature, see Figure 2. At low temperatures mainly linear α-alkenes
are obtained through a coordinated ethylene insertion mechanism on an active Ni
cation site. Between 373 and 403 K, in the presence of ethylene, an
irreversible transformation of Ni occurs which leads to a significantly lower
activity of these Ni sites [1]. As a result, higher temperatures, i.e., above
423 K, are required reach the same ethylene conversion through reaction on the
acid sites after an initial dimerization on the Ni cation sites. The absence of
odd carbon numbered alkenes in the product spectrum shows that no cracking
occurred, see Figure 1b.

Figure
2: Experimental conversion of ethylene as function of temperature at 3.5MPa and
45.0 kg_cat s mol_eth^-1.

Because these higher
operating temperatures offer more possibilities for tuning the product distribution,
the corresponding data are assessed first using the microkinetic model. Simulations
have been performed using initial parameter values based on literature data [1].
The C₄ fraction obtained is entirely built up out of linear alkenes.
C₆ and heavier fractions also contain branched isomers. Metal-ion
catalyzed oligomerization leads to Anderson Schulz Flory product distribution.
The acid catalyzed oligomerization disturbs this ASF distribution, however, see
Figure 3, left. This is more pronounced with increasing temperatures, see
Figure 3 right. The decrease with the temperature of the linear alkene content
in the C₈ fraction from 43.2 % at 443 K to 25.2 % at 503 K, also
illustrates the increase of the relative importance of acid catalyzed reactions
at higher temperatures.

Figure
3: Left: natural logarithm of the molar outlet flow rates as function of the carbon
number; Right: product distribution as a function of temperature, acid catalyzed
oligomerization included at 3.5 MPa and 6.7 kg_cat s mol_eth^-1.

The research leading to these results has
received funding from the European Community's Seventh Framework Program
FP7/2007-2013 under grant agreement n° 228953.

[1]        J. Heveling, C.P. Nicolaides, and M.S. Scurrell, Applied
Catalysis a-General 173 (1998) 1-9.

[2]        J.W. Thybaut, I.R. Choudhury, J.F. Denayer, G.V. Baron,
P.A. Jacobs, J.A. Martens, and G.B. Marin, Topics in Catalysis 52 (2009)
1251-1260.

[3]        J.R. Sohn, Catalysis Surveys from Asia 8 (2004) 249-263.