(156d) A Process Modeling Framework for Performance Assessment of Pd-Based Water-Gas Shift Membrane Reactors

Composite Pd and Pd/Alloy
membranes supported on porous metal supports have high H₂ flux and
stable selectivity. These membranes welded
from both ends with non-porous stainless steel tubes, can be very easily
scaled-up and integrated into high-pressure and high-temperature H₂ separation and production applications [1,2]. Steady state
mathematical models of Pd-based membrane reactors under isothermal and
adiabatic conditions were developed and used with process analysis and process
intensification objectives. The performance of the Pd-based composite membrane
reactor for the production of extra pure H₂ via the high temperature
water-gas shift reaction was characterized through the simulation studies  considering
the effects of the permeability, feed temperature (for the adiabatic reactor)
and bulk catalyst density (ρ_Bulk).  The reaction conditions
used in the simulations were as follows: P_Reaction = 15 atm, P_Permeate
= 1 atm, T_Range=300-450°C and the slurry-feed coal-derived syngas
composition as the feed [3]. The dimensions of a tubular 0.5"OD×2.5"L
membrane mounted at the center of a coaxial cylindrical 1"ID shell were
utilized in the model. No sweep gas was used to avoid the need for a subsequent
H₂-sweep gas separation unit. The Pd-based membrane was assumed to
have infinite H₂ selectivity.

The minimum permeability
that a Pd-based composite membrane should have in order to reach the process
goals in terms of CO conversion (X_CO) and hydrogen recovery (R_H2)
can be determined with the aid of the proposed modeling framework. It was
assumed that the 5 mm thick Pd composite membranes exhibited 20-100% of the
permeability of the Pd foil with the same thickness and the associated CO
conversions and H₂ recoveries were compared. The isothermal
simulations showed that the membranes having 60% and higher of the permeability
of the Pd foil with the same thickness were not affected by the gas hourly
space velocity within the range of 1600-8000 h^-1; while the CO
conversion and H₂ recovery were constant around 97% and 90% at
450°C, respectively. The permeability had a more prevalent effect on the
adiabatic reactor performance than the performance of the isothermal reactor,
especially on the H₂ recovery front. The adiabatic reactor could
achieve high X_CO and R_H2 only at low inlet flow rates due
to low feed temperatures in order to protect the membrane. (X_CO =
85% and R_H2 = 90% at GHSV = ?1600h^-1 at T_Feed=
260°C).Finally, a judicious compromise between high CO conversion
demands and the need to prevent the deterioration of the membrane and catalyst
properties by hot spots in the adiabatic membrane reactor was made.

The influence of the bulk
catalyst density and the feed temperature on system performance was
investigated by using the properties of a previously characterized 10 μm
thick Pd-based membrane. Furthermore, the effect of the feed temperature on the
adiabatic membrane reactor performance over the range of 300-450°C was
addressed. It was suggested by the simulation results that the feed temperature
in this range was not affecting the CO conversion and pure H₂ flow
rate significantly (X_CO = 90-94% and R_H2 = 90%). The feed
at a GHSV of 1243 h^-1 and at 300°C caused the reactor temperature to
rise up to 545°C which did not exceed the annealing temperature of 550°C.

The efficient use of the
total membrane area and the total catalyst weight was also evaluated to avoid
the excess use of the high temperature water gas shift catalyst and palladium.
The simulations were run with 2-100% of the maximum catalyst bulk density (ρ_Bulk,Max
: maximum catalyst weight/ reactor volume). At the isothermal reactor
temperature of 450°C, the same CO conversion level of 97% at the reactor exit
was attained for all of the catalyst bulk density values, indicating that
packing the reactor with less catalyst would not affect the production
specifications and would reduce the cost. For the adiabatic membrane reactor
operating at T_Feed= 300°C; the minimum ρ_Bulk was
determined as 8% of ρ_Bulk,Max which will result in almost the
same product specifications with 100% of ρ_Bulk,Max. The
adiabatic membrane reactor could achieve  ̃94% X_CO and 
̃90% R_H2 with the feed temperature of 300°C for both ρ_Bulk,Max
andρ_Bulk,Min. Moreover, a more controlled
temperature rise along the length of the adiabatic membrane reactor was attained
when the catalyst weight was diluted to 8% of ρ_Bulk,Max. The
portion of the total reactor volume where the H₂ permeation rate was
high was increased from 40% to 80% due to controlled adiabatic reactor
temperature profiles.

References

1. Ma, Y. H., Engwall;
E. E. and Mardilovich, I. P.; Fuel Chemistry Division Preprints, 2003, 48(1), 333.

2. Ayturk, M.E. ;
Kazantzis, N. K. and Ma, Y. H.; Energy Environ. Sci., 2009, 2, 430-438.

3. Hla, S. S.; Park, D.;
Duffy, G. J.; Edwards, J. H.; Roberts, D. G.; Ilyushechkin, A.; Morpeth, L. D.;
Nguyen, T. ;Chemical Engineering Journal, 146(2009), 148-154.