ABSTRACT A simple


ABSTRACT
A simple, yet practical, steady-state kinetic model is developed for lime-enhanced biomass steam gasification in a dual fluidized bed reactor, one of the most promising technologies for sustainable production of hydrogen. The focus of this paper is on the bubbling fluidized bed (BFB) gasifier/carbonator kinetic model, which includes a generic two-step reaction kinetic mechanism for primary and secondary biomass pyrolysis, major homogeneous and heterogeneous gasification reactions, and a first-order kinetic model, sufficient to simulate the carbonation of highly cycled sorbent particles in steady state. An ideal reactor model is used for the BFB, assuming perfectly mixed solids and plug flow of the gas phase. Predictions of the product gas composition and yield are in good agreement with independent data from the literature where the effect of Ca-looping rate was investigated experimentally by others for Absorption Enhanced Reforming (AER) of biomass in a steady-state 20 kWth dual fluidized bed system. Sensitivity analyses on key process operating parameters such as reactor temperature suggests that this predictive model is useful for optimizing design and operation of dual fluidized bed biomass gasifiers with lime-based CO2 capture.
Keywords: Steam gasification; biomass; dual fluidized bed gasifier; CO2 capture; hydrogen; Adsorption Enhanced Reforming.

INTRODUCTION
A sustainable energy future is substantially desired by today’s mankind. Adequate fulfillment of the escalating world energy demand has justified large investments in research and development of new technologies with near-zero emissions of greenhouse gases (especially CO2) to the atmosphere which have been shown to contribute to climate change.
Biomass gasification is one of the renewable energy alternatives holding promise of reducing greenhouse gas emissions. The renewed interest in gasification partially originates from its potential for producing hydrogen as a non-polluting energy source 1. In particular, the integration of biomass gasification with CO2 capture and sequestration technologies enhances hydrogen production efficiency while potentially leading to net removal of CO2 from the atmosphere, i.e. “negative emissions of CO2”.
Process Description
Steam gasification of biomass produces higher hydrogen content of product gas compared to gasification with a mixture of steam and oxygen, whereas the hydrogen concentration in product gas generated by air gasification is the lowest 2 of these options. While biomass is partially oxidized with air or oxygen in situ autothermal gasifiers to derive the highly endothermic gasification reactions, the heat requirements of allothermal gasifiers are provided by external sources such as the combustion of unreacted outside the gasifier. As seen in the schematic of a dual fluidized bed gasifier of Figure 1, steam gasification of biomass occurs in a fluidized bed gasifier and the unreacted char resulting from incomplete gasification is transferred to another fluidized bed to be burnt with air at higher temperatures. The combustion of char, and if necessary an additional fuel, heats up the bed materials (e.g. inert silica sand) that are circulated back to the gasifier to provide the heat requirements of the endothermic gasification reactions. Owing to the separated gasification and combustion zones, a Dual Fluidized Bed reactor configuration prevents product gas dilution by N2 in spite of using inexpensive air for char combustion and produces product gas of higher hydrogen concentrations and heating values 3.
Figure 1. Operation of dual fluidized bed steam gasifier, adapted from 3.

In order to meet industrial requirements, areas of improvement for biomass gasification include low H2 concentrations (e.g. 40% volume, dry) and high tar contents (e.g. ~10 g/Nm3) of the product gas as well as high operating temperatures (;800oC), that are expensive to achieve with steam as the fluidizing agent 1. Although recognized as the most efficient and profitable route for hydrogen production 4, biomass gasification should be combined with other advanced technologies in order to provide a sustainable energy path. One promising technology is lime-enhanced biomass gasification in a dual fluidized bed reactor where CO2 capture via Calcium looping promotes hydrogen production from a renewable energy source, i.e. biomass 5. This is done by in situ removal of CO2 that shifts the H2-producing equilibrium reactions of biomass gasification, e.g. water-gas shift reaction,
(1)
with CO2 capture and release occurring through reversible carbonation-calcination reactions:
(2)
The heat generated by the exothermic carbonation reaction favors the endothermic biomass gasification reactions. While the carbonation is promoted when the partial pressure of CO2 is greater than its equilibrium partial pressure at the reaction temperature, the endothermic calcination takes place when the reverse condition is true. Although the carbonation rate depends on several process parameters, complete calcination is possible for a wide range of conditions.

As seen in Figure 2, to enhance biomass steam gasification, instead of inert silica sand, limestone particles are used as the bed material within the dual fluidized bed reactor. Under desirable operating conditions for simultaneous biomass gasification and sorbent carbonation, limestone particles function as effective heat carrier, selective sorbent, and transporter of CO2. In situ CO2 capture shifts the dominant equilibrium reactions of biomass gasification to produce H2-rich product gas from the gasifier (carbonator). The unreacted char and calcium carbonate particles produced from incomplete biomass gasification and carbonation reactions, respectively, are transported to the combustor (calciner) where the combustion of char particles (and additional fuel) with pure oxygen or oxygen-enriched air stream heats up the calcination reaction to produce a concentrated CO2 stream and a hot regenerated sorbent (CaO) stream that is circulated back to the gasifier.

Figure 2. Schematic of dual vessel system for lime-enhanced biomass steam gasification adapted from 5.

An integrated reaction scheme for coal and biomass gasification have been studied by several researchers. In a process named ‘HyPr- RING’, Lin et al. (2004a) used a continuous pressurized reactor to study high-pressure steam gasification of coal in the presence of CaO at 650°C and up to 60 atm (steam partial pressure 30 atm) 6. They found that increasing pressure increases H2 production with a maximum concentration of 77 vol. % H2 (dry basis) at 60 atm 7. Florin and Harris 8 reported that hydrogen production from atmospheric biomass steam gasification is significantly enhanced by CaO affecting the water-gas shift reaction during primary pyrolysis of biomass. According to their Thermogravimetric Analysis/Mass Spectrometry (TGA-MS) experiments, CaO catalytic effects on tar cracking and char decomposition also contribute to increased H2 yields 8.

Pilot Plant and Industrial Applications
Large-scale Ca-looping process was first developed to eliminate CO2 and upgrade the product gas heating value from gasification processes 9. The CO2 Acceptor Process developed by the Consolidation Coal Company in the 1960s-1980s 10 utilized two interconnected bubbling fluidized beds, i.e. a steam gasifier/carbonator operating at 10 bar and 825oC and a combustor/calciner operating at atmospheric pressure and 1000oC.
Biomass steam gasification has been recently enhanced by the Adsorption Enhanced Reforming (AER) process under the European Commission’s 6th Framework Programme 11. In order to carry out pilot-scale investigations of the AER process, a “Fast Internally Circulating Fluidized Bed” has been built at the Vienna University of Technology 12 with a gasifier/carbonator operating at 600–700oC and atmospheric pressure and a 100 kWth combustor/calciner thermal power. AER process addresses the shortcomings of a commercial-scale biomass gasification process without sorbent leading to the increased H2 concentrations of the product gas from ~40 to 75% v/v and lower tar emissions due to tar cracking mechanisms catalysed by CaO that are possible at lower gasification temperatures. Theoretical and experimental studies of a demonstration-scale AER process with 8 MWth fuel input has also been performed in a Combined Heat and Power (CHP) unit in Guessing, Austria 13. Other successful examples of Bio-Energy with Carbon Capture and Storage (BECCS) with Calcium Looping at scales larger than 100 kWth include La Robla power plant (Spain) 14, Darmstadt University of Technology (Germany) 15 and ITRI (Taiwan) 16. Abanades 17 highlighted the ongoing demonstration of the Ca-looping technology at increasing pilot scales and the potential for minimum energy penalties and low capture cost. Furthermore, Pfeifer 18 reviewed the thermodynamics of sorption-enhanced gasification in detail and described experiments on laboratory and pilot scales.
Reactor Modelling
Most of the reactor models developed for integrated biomass gasification and CO2 capture are based on thermodynamic equilibrium which predict the upper limit for fuel conversion 5, 19-24. In particular, Aspen Plus models have been used to discover the range of operating conditions appropriate for lime-enhanced biomass gasification 21-24. Increased hydrogen production and energy efficiency due to in situ CO2 capture was found for steady state biomass gasification with and without limestone in a dual fluidized bed reactor using IPSEpro equation-oriented simulation software 5. PETRONAS iCON simulation of pressurized biomass gasification integrated with CO2 capture showed increased hydrogen production as a result of increased pressure, temperature and steam-to-biomass ratio 25. Inayat et al. 26 developed a MATLAB model and predicted that increased biomass gasification thermodynamic efficiency can be obtained at greater temperatures and steam-to-biomass ratios and in the presence of CaO.

Among the few kinetic models developed for lime-enhanced biomass gasification, Hejazi et al. 5, 19 simulated an integrated biomass gasification and cyclic CO2 capture in a Dual Fluidized Bed to evaluate the influence of CO2 capture on steam gasification of wood residue. They studied the effect of various operation conditions, including temperature, pressure, steam-to-biomass ratio, sorbent circulation rate, etc., on enhanced H2 production and CO2 capture.
The objective of this study is to develop a simple kinetic reactor model to predict the performance of steam gasification of biomass with cyclic lime-based CO2 capture (i.e. Ca-looping) in a dual fluidized bed reactor. To develop a kinetic model for steam gasification of biomass in a BFB gasifier, a two-step biomass pyrolysis kinetic mechanism, major heterogeneous and homogeneous gasification reactions and an empirical kinetic model for the carbonation rate of limestone particles are coupled with an ideal reactor model. The model developed in this paper extends an earlier model 27 with the following improvements:
Added CO2 capture modelling equations as per the derivations below.

Comparison of the model predictions with literature experimental results from reference 28 on the effect of Ca-looping rate on product gas yields and composition. 
Prediction of product gas yields and composition beyond available experimental data (e.g. temperature) that are expensive to generate, but essential for designing, evaluating and improving gasifiers.

Although the other model features remain common to the ones developed in 27, 28, the good agreement of this model predictions with additional experimental data is a significant step towards model validation. This kinetic model provides a comprehensive practical, modeling approach to tackle the complicated task of modeling Ca-looping in dual fluidized bed biomass gasifiers that have previously been usually modeled assuming thermodynamic equilibrium.

MODEL DEVELOPMENT
To develop a model for steam gasification of biomass in a bubbling fluidized bed (BFB) reactor with in situ CO2 capture via Ca-looping, the following simplifying assumptions are adopted in addition to those listed previously 27:
The bed material consists of sorbent and char particles, with weak catalytic effects on tar cracking reactions 29, 30..

There is negligible loss of sorbent and char due to entrainment..

Chemical reaction is the rate-controlling mechanism for heterogeneous reactions (char gasification and carbonation) inside the bubbling fluidized bed..

Complete char combustion and complete limestone calcination occur inside the combustor/calciner 31, 32..

Reaction Kinetics
As discussed in an earlier paper 27, the generic two-step biomass pyrolysis kinetic mechanism 33-35 is adapted where biomass is primarily decomposed to non-condensable gas, tar and char via three parallel first-order Arrhenius-type reactions followed by a homgenous tar thermal cracking reaction. Details of the elemental balances as well as the kinetic rate expressions of major gasification reactions are discussed elsewhere 27, 28.

Homogeneous Grain models and Shrinking Core and Pore models are among the different modelling approaches suggested for carbonation rate of limestone particles. Despite neglecting the effects of intra-particle and transport resistances, we adapt the first-order carbonation rate expression developed by Grasa et al. 36 due to the wide range of reaction conditions, particle sizes and sorbents used to find the curve-fitting parameters:
(3)
where KS is an intrinsic kinetic constant, Save the average specific surface area available for sorbent particle carbonation, Xave the average carbonation conversion of sorbent particles, and CCO2 and CCO2,eq the actual and equilibrium CO2 concentrations. Baker 37 proposed a semi-empirical correlation for estimating the equilibrium CO2 concentration as a function of carbonation temperature and ideal gas behaviour by
(4)
Furthermore, the particle average surface area is proportional to average carbonation conversion:
(5)
where the molar volume of CaCO3 (VMCaCO3), intrinsic kinetic constant and the thickness of the CaCO3 product layer (h) used in this study are reported in Table 1.
Table 1. Parameters used to calculate sorbent conversion
32, 38.

KS (m4/mol/s) VMCaCO3 (m3/mol) h (m)
6.0510-10 36.910-6 5010-9
Reactor Model
As shown schematically in Figure 3, separate models are used for the gas and particle flows inside a bubbling fluidized bed reactor. While perfect mixing of solid particles is a reasonable assumption due to much larger mean solids residence time compared to their turnover time, the gas flow is closer to plug flow.

Figure 3. BFB Reactor model schematic. Perfect mixing for solids and plug flow for gas. Black circles provide schematic representations of biomass/char and limestone particles.

Assuming plug flow of gas phase and uniform distribution of drying and pyrolysis products throughout the entire dense bed height, Lbed, with uniform temperature, the one-dimensional tar mass balance at height z of the BFB is written:
(6)

Assuming that carbonation is an independent reaction occurring in parallel to other reactions and accounting for the contribution of biomass pyrolysis, tar cracking, homogeneous and heterogeneous gasification reactions, as well as lime-based CO2 capture (carbonation) in the dry gas mass balances, the one-dimensional differential equations are written as in our earlier paper 27, except for the CO2 mass balance, where the rate of selective CO2 capture due to Ca-looping is also taken into account. Therefore:
(7)
where:
(8)
and the differential gas mass balances, as well as the char balance, are given in 27.
In order to approximate the average carbonation conversion, Xave, based on the reactor design specifications and operating conditions, several issues affecting the decay of sorbent effectiveness must be taken into account, such as 19:
Sorbent loss of surface area due to multiple capture and release cycles, i.e. sintering.

Pore blockage by CaSO4 originating from the sulfur content of the biomass.

Attrition, leading to mechanical fragmentation and entrainment of sorbent from the system.

Deactivation of sorbent particles due to coke formation.

CO2 diffusion resistances due to formation of a CaCO3 product layer on the surface of the particles.

Given a mean solids residence time () from experimental measurements, no independent calculation of the average carbonation conversion is required in our model. Instead, the average carbonation conversion is estimated through an iterative procedure similar to that used above to determine the char hold-up of the bed. From a CaO balance on the dense bubbling bed at steady-state, the consumption rate of CaO in the carbonator/ gasifier is:
(9)
If MCaO is the reacting sorbent hold-up inside dense bed, then:
(10)
(11)
On the other hand, from the axial partial pressure profile of CO2 in the gas phase, the consumption rate of CaO can be approximated by:
(12)
Combining equations (9) to (12), the hold-up of sorbent available for CO2 capture is:
(13)
In order to account for the axial temperature and pressure profiles, the ODE’s for gas mole balances must be coupled with ODE’s corresponding to energy and pressure balances along the bed. However, because of the great dispersion in the dense bubbling bed, it is reasonable to assume uniform total pressure, temperature and solids hold-up 28. Another simplifying assumption is to ignore the entrainment of particles and model the solid-free freeboard section (?fb=1) with a plug flow reactor where only homogeneous reactions (i.e. tar cracking, SMR and WGS) occur. The results at top of the dense bed is then used as the boundary conditions for the ODE’s corresponding to the freeboard section.
The algorithm for simulation of lime-enhanced biomass gasification in the BFB gasifier of a dual fluidized bed reactor is provided in Appendix A where the char and sorbent hold-ups, as well as the expanded dense bed height, are calculated iteratively.

RESULTS AND DISCUSSION
Experimental Information
The predictions of our model are compared with experimental results of Poboss et al. 39 who studied the effect of Ca-looping ratio (defined as the ratio of the molar flow rates of regenerated sorbent (CaO) to carbon which enters the gasifier as fuel) on product gas yield and composition generated from sorption-enhanced reforming of biomass in a 20 kWth Dual Fluidized Bed reactor at the Institute of Combustion and Power Plant Technology (IFK). This DFB facility features a CFB regenerator of 12.4 m height and 70 mm diameter and a BFB gasifier of 3.5 m height and 114 mm diameter. The experiments were carried out with steam as the fluidizing/gasifying agent and wood pellets as feedstock, with their composition and lower heating value reported in Table 2. Approximately 15 kg of pre-calcined Greek limestone was used as the bed material, with characteristics shown in Table 3. The gasifier and regenerator temperatures were set at 650 and 850°C, respectively, ensuring complete combustion of char and calcination of the bed material at atmospheric pressure.
Table 2. Properties of wood pellets used in experimental study 39.

Ultimate analysis (wt%, a.r.*)
C 48.0
H 5.5
O 42.6
N ; 0.3
S ; 0.3
Moisture 3.7
Ash 0.2
LHV (MJ/kg daf) 18.8
*a.r.=as received
Table 3. Properties of bed material used in experimental study 39.

Composition (wt%)
CaO 95.53
MgO 0.94
SiO 2.30
Density (kg/m3) 1800
Diameter (m) 300-600
The gasifier and regenerator temperatures were set at 650oC and 850oC, respectively, ensuring complete combustion of char and calcination of the bed material at atmospheric pressure. The Ca-looping ratio was set at seven different levels, as given in Table 4, which also lists the most important experimental conditions during experiments 1 to 7.

Table 4. Operating conditions for experimental study 39.

Experiment number 1 2 3 4 5 6 7
Fuel flow rate (kg a.r./h) 3.3 3.4 3.6 3.9 4.0 4.0 4.1
Steam/Carbon ratio (mol/mol) 1.5 1.5 1.4 1.3 1.3 1.3 1.3
Fuel flow rate/ bed weight (kg a.r./h)/ kg 1.4 1.3 1.5 1.4 1.6 1.7 2.0
Ca/Fuel Carbon (mol/mol) 7.6 7.4 1.9 11.5 4.7 2.7 0.0
Turnover time (min) 2.5 2.5 9.3 1.5 3.5 6.1 ?
Fuel flow rate/ bed cross-sectional area
(kg a.r./h)/m² 326 332 351 379 389 393 397
Modelling Results
Figures 4 and 5 illustrate the model predictions for product gas composition and yield from biomass steam gasification in the IFK DFB gasifier, with limestone as the bed material as a function of the Ca-looping ratio. With increasing Ca-looping ratio, it is seen that more CO2 is captured in situ in the gasifier, shifting forward the WGS equilibrium towards more CO consumption and more H2 production. The model predictions are in good agreement with experimental data points from Poboss et al. 39. According to these authors, at Ca-looping ratios above 6 mole Ca/mole C, the gas concentrations tend to reach constant levels. As seen in Figure 5, with increasing Ca-looping ratio, the product gas yield decreased, mostly due to the reduced time available for biomass conversion inside the gasifier. The model overpredicts the product gas yield, perhaps due to the temperature drop in the reactor top section that is assumed to be adiabatic, given the difficulty of estimating the contribution of the electric heaters in the absence of overall heat balances in the original paper 39. Some of the simplifying assumptions adopted to develop a predictive model could also contribute to the deviation of model predictions from experimental data. For instance, the limitations of the two-step pyrolysis kinetic model for different types of biomass with varying properties significantly affects the model predictions 40. Furthermore, to close the CHO elemental balances with minimum reliance on empirical correlations, the light hydrocarbons (e.g. C2 and C3) are lumped with methane. Finally, a more elaborate reaction network for tar reforming and cracking might improve the model predictions.

Figure 4. Effect of Ca-looping ratio on dry and N2-free product gas composition for lime-enhanced steam gasification of wood pellets at reactor temperature of 650°C. Experimental data (circles) are from Poboss et al. 39. Continuous lines: Model predictions (this work).

Figure 5. Effect of Ca-looping ratio on product gas yield for lime-enhanced steam gasification of wood pellets at reactor temperature 650°C. Experimental data points (circles) are from Poboss et al. 39. Continuous line: Model predictions (this work).
The dry and N2-free product gas composition predicted by the model developed in this study is illustrated in Figure 6 as a function gasifier/carbonator temperature for typical operating conditions of the IFK Dual Fluidized Bed reactor, i.e. Ca-looping ratio=6 mole Ca/mole C, Turnover time=2.5 min, Fuel flow rate=4 kg a.r./h, T= 650-750oC, Steam/Carbon ratio= 1.3 mol/mol. As seen, increasing the reactor temperature reverses the exothermic WGS reaction toward more CO production and H2 consumption. Raising the reactor temperature also reverses the exothermic carbonation, leading to less CO2 capture and thus less enhancement of H2 production by limestone. At increased reactor temperature, carbonation no longer occurs because the partial pressure of CO2 in the gasifier falls below its equilibrium value. For temperatures below 700oC, the predicted H2 content of dry product gas for the lime-enhanced biomass gasification is more than 70 mole %.

Figure 6. Predicted dry and N2-free product gas composition for lime-enhanced steam gasification of wood pellets as a function of BFB gasifier/carbonator temperature.
CONCLUSIONS
A simple kinetic model, without fitting parameters, was developed for lime-enhanced biomass steam gasification under steady-state Ca-looping operation in a dual fluidized bed reactor. By adopting an empirical kinetic model for the carbonation rate of limestone particles, the kinetic model of bubbling fluidized bed gasifier is extended to account for selective CO2 capture. The model predictions of product gas yield and composition as a function of Ca-looping ratio are shown to be in good agreement with independent experimental data from the literature. Sensitivity analysis on gasifier/carbonator temperature clearly shows sorbent enhancement by shifting the water-gas shift reaction towards more H2 production and CO consumption.

ACKNOWLEDGMENTS
The authors gratefully acknowledge financial aid from Carbon Management Canada and the Natural Sciences and Engineering Research Council of Canada.

NOMENCLATURE
A Bubbling bed cross-sectional area, m2
Ci Concentration of species i, mol/m3
h Thickness of CaCO3 product layer, m
kj Arrhenius-type kinetic rate constant of jth reaction, s-1
KS Intrinsic kinetic rate constant for carbonation reaction, m4/mol.s
Lbed Dense bubbling bed height, m
Mass flow rate, kg/s
MCaO CaO hold-up inside dense bed, kg
VMCaCO3 Molar volume of CaCO3, m3/mol
MW Molecular weight, g/mol
r Reaction rate, s-1 or mol/m3.s
R Universal ideal gas constant, 8.314 J/mol.K
Save Average specific surface area available for sorbent particles carbonation, m2/m3
T Temperature, K
U Superficial gas velocity, m/s
Xave Average carbonation conversion, –
Average yield, kg/kg dry biomass
z Axial coordinate along reactor height, m
Greek
Stoichiometric coefficient of ith species in tar cracking reaction, –
Heat of reaction at standard condition, kJ/mol
Bed voidage at height z, –
Density, kg/m3
Mean solids residence time, s
Subscripts
B Biomass
Carb Carbonation
eq Equilibrium
fb Freeboard
G Non-condensable gas
in Input to reactor
M Moisture
Pyro Pyrolysis
T Tar
V Water vapor
ABBREVIATIONS
AER Absorption enhanced reforming
a.r. As-received
B.C. Boundary condition
BECCS Bio-Energy with carbon capture and storage
BFB Bubbling fluidized bed
CHP Combined heat and power
CSTR Continuous stirred tank reactor
DFB Dual fluidized bed
LHV Lower heating value
MS Mass Spectrometry
PFR Plug flow reactor
SMR Steam methane reforming
S/B steam-to-biomass ratio
TGA Thermogravimetric Analysis
WGS Water-gas shift