(206a) Experimental Study and Modelling of Kinetic Transitions upon Processing of Dimethyl Ether and Methanol to Gasoline (DMTG) at Fluctuating Workloads

1. Introduction

Dynamic
operation of chemical reactors is an important issue when feedstock supplies
are fluctuating. An example is hydrogen produced by electrolysis using excess
electricity from solar or wind power plants. One option for the chemical
storage of hydrogen is the production of synfuels. In this context, the
production of gasoline is of particular interest because the existing
infrastructure for storage and distribution of gasoline can be utilized.

Within
the process train comprising syngas conversion to methanol (MeOH), dehydration
to dimethyl ether (DME), water gas shift and synthesis of hydrocarbons, it is
the last stage denoted as DMTG (Dimethyl Ether/Methanol To
Gasoline), which deserves special attention.

So
far, the only DMTG transient experiments reported in the literature were
carried out by switching ¹²C/¹³C MeOH in the feed at
isothermal conditions; ¹³C incorporation in products was observed
within 1-2 minutes. These experiments were extremely helpful in elucidating the
reaction mechanism. It is widely accepted that a "hydrocarbon pool"
with autocatalytic properties is accumulated in the zeolite pores. In ZSM-5,
the pool presumably consists of alkenes and mono-aromatics which promote
methylation and cracking reactions in two coupled autocatalytic reaction cycles
(dual-cycle concept) [1]. It is important to note that the ¹²C/¹³C
switches mentioned above had no effect on the kinetic steady state because the
work load of the catalyst was not changed. This contribution, however, focuses
on DMTG system responses to variations in the reaction conditions. Furthermore,
single and periodical changes were imposed not only at isothermal but also at
polytropic operation. In the latter case, the catalyst bed was allowed to develop
temperature profiles due to the caused shift in the heat of reaction released.
A kinetic model has been elaborated which, after implementation in mass and
energy balances, provides a satisfactory description of the system performance
at fluctuating workloads.

2.
Methods

Experiments
were carried out in two series-flow fixed-bed tubular reactors at 10 bar (Fig.
1). The first reactor, filled with gamma-Al₂O₃ catalyst,
was used to produce an equilibrium DME/MeOH mixture from MeOH at 573 K.
Hydrocarbon synthesis occurred in the second reactor over H-ZSM-5/gamma-Al₂O₃at 613-653 K. Both catalysts were applied as extruded cylindrical (2x5
mm) bodies. Transient experiments were carried out by imposing single or
periodically repeated changes in e.g. flow velocity and feed concentration. For
isothermal experiments, the zeolitic catalyst particles had to be diluted with
SiC particles
of 0.2 mm diameter, and the power supply to the 3
zones of the reactor heating was continuously regulated to ensure a flat
T-profile (delta T_axial,_max = 2 K). In contrast, polytropic
conditions were installed by keeping the power supply at a constant level. Care
was taken to avoid pressure surges.

Figure 1. Plant scheme
and description of the residence time characteristics.

High
time resolution of the system responses was achieved by using a multiport valve
(MPV), in which samples taken at short intervals could be stored for on-line
GC. Measurements with non-reactive fixed beds were performed to assess and
model the residence time characteristics of the plant. A steady state kinetic
model (R) had been developed before. The transient behavior of the
catalyst was identified and quantified by analyzing the deviations between
dynamic system responses and predictions modelled by combining residence time
characteristics and steady state kinetics. Spent catalysts were analyzed to
investigate the influence of the reaction conditions on the hydrocarbon species
occluded in the samples.

Figure
2. Reaction network used for modelling of the kinetics.

As shown in
Fig. 2, the reaction network used for the kinetic model contains 7 lumps of
compounds. For solving the heat balance, heat capacities and heats of formation
of all compounds were either taken from the NIST database [2] or were
calculated using the Joback-Reid [3] and the Benson method [4], respectively.
Group values for the lumps were determined for characteristic lump compositions
using linear additivity rules. At 623 K, for instance, the resulting over-all
reaction enthalpy amounts to delta H_r^623K =  –(31.8 to 43)
kJ/mol MeOH/DME, the precise value depending on the severity of operation.

3.
Results and discussion

At
isothermal conditions, modulations of the feed concentration, in particular,
resulted in pronounced and long-lasting kinetic transition states. For example,
Fig. 3 a) shows the system response on a single step-increase in the MeOH
concentration. It takes almost 45 minutes until the new steady state is
reached, whereas the reaction unit alone needs only about 1 minute. During the
transition, hydrocarbons are accumulated in the zeolite, the MeOH/DME
conversion is increased, and the amount of released aliphatics and C_6-9
aromatics is enhanced at cost of the olefins. TPD analyses of the spent
catalysts (Fig. 3 b) confirm that increasing feed concentrations result in
higher levels of hydrocarbon accumulation. As far as the composition of the
occluded hydrocarbon pool is concerned, it is observed that increases in feed
concentration clearly promote formation and accumulation of highly alkylated
aromatics. Both, experiments and modelled kinetic parameters indicate that accumulated
alkyl benzenes, in particular C₁₀₊ aromatics, have an inhibiting
effect and are responsible for the long transient time until the new steady
state is reached.

In
contrast, system responses at polytropic conditions are considerably faster. After
raising the feed concentration, for instance, the heat production by reaction
is enhanced, and the resulting increase in local temperature shifts the
thermodynamic equilibria such that formation of highly alkylated aromatics is
prevented and dealkylation of existing species of this type is promoted. Figure
4 shows exemplary plots of molar flows and hot spot temperatures measured upon
periodic modulations of the feed concentration at polytropic conditions. The
cycle time is only 2 minutes, but the system response is fast enough to follow
the imposed changes. In particular, these plots show that the flows of all
compounds except for alkyl benzenes reach their respective steady state values,
which are indicated by the horizontal black and grey lines.

Figure 4.  Polytropic conditions. a) Molar flows
(symbols) and hot spot temperature (red line)  vs. runtime during periodical step
changes in the molar feed fraction between y_MeOH,0 = 0,108 and y_MeOH,0
= 0,323; t_mod = 75 kg×s/m³; cycle
time = 2 min. The two horizontal lines indicate steady-state molar flows at y_MeOH,0
= 0,323 (black) and at y_MeOH,0 = 0,108 (grey).

The
kinetic model developed for steady-state operation can also be applied to
describe the performance of the catalyst at fluctuating workloads by
implementing a dimensionless "degree of filling" (theta) of the
zeolite pores with C_6-9 and C₁₀₊ aromatics, which is
correlated with the thermodynamic equilibria of the de-alkylation reactions at
the respective reaction conditions.

4. Conclusions

Modulations
of process parameters result in marked changes of size and composition of the
hydrocarbon pool occupying the pores of ZSM-5. In particular, the level of
population with highly alkylated aromatics plays a crucial role for the lengths
of the transient phases because these species have an inhibiting effect, be it
through blocking of pores or of active sites. At isothermal conditions,
increases in feed concentration promote the formation of these species, resulting
in a pronounced slow-down of the system response. A different situation arises
when the catalyst bed is allowed to change its temperature as is the case under
adiabatic or polytropic conditions. Increasing feed concentration and raising
bed temperature have opposite effects on the degree of filling with highly
alkylated aromatics. At non-isothermal conditions, the DMTG catalyst is able to
respond quickly to fluctuations in the workload.

References

[1]    
U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T.V.W. Janssens, F.
Joensen, S. Bordiga, and K.P. Lillerud, Angew. Chem. Int. Ed. 51 (2012),
5810 – 583, and references therein.

[2]    
NIST (National Institute of Standards) webbook, srd69: https://webbook.nist.gov/chemistry
(data taken on April 6, 2018).

[3]    
K. G. Joback and R. C. Reid, Chem. Eng. Commun. 57 (2007),
233-243.

[4]    
S. W. Benson, F. R. Cruickshank, D. M. Golden, G. R. Haugen, H. E.
O'Neal, A. S. Rodgers, R. Shaw and R. Walsh, Chem. Rev. 69 (1969), 279-324.