(436e) A Model-Assisted Tool for the Characterization of Char In Biomass Pyrolysis and Biomass Catalytic Pyrolysis

Coke formation
appears as a reason for catalyst deactivation and results in undesirable
product selectivity in biomass pyrolysis. However, there is no effective way to
separate coke and char in the majority of current approaches, which prevents
studying coke formation and therefore the real reason for catalyst deactivation.
The purpose of this study is to develop a model-assisted tool for quantitative
analysis of coke and char composition in the residual after biomass pyrolysis
and biomass catalytic pyrolysis.

In this
study, alpha-d-glucose was used as a biomass representative molecule and ZSM-5 from
W.R. Grace & Co. was used for catalytic pyrolysis experiments. Char samples
were prepared with biomass (catalytic) pyrolysis by devolatilizing glucose at
different temperatures in nitrogen atmosphere. The gasification of biomass chars
was carried out in a thermogravimetric analyzer (TGA) in carbon dioxide
atmosphere at different heating rates.

In order to
characterize the biomass pyrolysis char, three representative models,
volumetric model (VM), grain model (GM) and random pore model (RPM) were developed
and applied to fit the experimental data from TGA.

The following equation is the random
pore model for fluid-solid reactions developed by Bhatia and Perlmutter [1].

In the above equation, the pre-exponential
factor k₀, activation energy E, the structural
parameter ѱ can be obtained, by fitting the model to experimental data
of char/coke gasification or oxidation. Similar approaches have been used by
Miura et al. [2] and Fermoso et al. [3] in the characterization of coal/coal-biomass
blend derived chars. On the basis of these works, a method is being developed to
differentiate coke from char, using temperature programmed gasification
(reaction of char and coke with CO₂) to exploit the fact that the more
reactive char is gasified at lower temperatures than the less reactive coke. Independent
application of RPM for coke and char followed by integration for a linear ramp
in temperature yields a correlation for the total mass loss X as a
function of the composition in coke and char. The concept applies similarly on
the other two models, VM and GM. Moreover, a pore-scale effectiveness factor is
introduced to modify RPM, allowing different pore sizes to grow at different
rates [4].

The final correlations for each model are
then fitted to experimental measurements of TGA gasification at different
heating rates, resulting in a model-assisted tool for char and coke
identification. Parameter estimation is performed using GAMS global solvers.
The results of the models are also compared and verified with scanning electron
microscopy to validate the approach and generate a standard method for char and
coke characterization. Experimental results, statistical analysis of the
literature and model fitting parameters will be presented. Results illustrate
the applicability of the method for the characterization of biomass char and
coke.

References:

(1)   Bhatia, S. K.;
Perlmutter, D. D. AIChE Journal. 1980, 26, 379-386.

(2)   Miura, K.; Nakagawa,
H.; Nakai, S.; Kajitani, S. Chemical Engineering Science. 2004, 59,
5261-5268.

(3)   Fermoso, J.; Gil, M.
V.; Pevida, C.; Pis, J. J.; Rubiera, F. Chemical Engineering Journal. 2010,
161, 276-284.

(4)   Singer, S. L.;
Ghoniem, A. F. Energy & Fuels. 2011, 25, 1423-1437.