Eth 46804 02 1

Research Collection
Doctoral Thesis
Heat transfer enhancement in a solar biomass gasifier
Author(s):
Krüsi, Michael
Publication Date:
2014
Permanent Link:
https://doi.org/10.3929/ethz-a-010256755
Rights / License:
In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For more
information please consult the Terms of use.
ETH Library
Diss. ETH No. 21821
HEAT TRANSFER ENHANCEMENT

IN A SOLAR BIOMASS GASIFIER
A thesis submitted to attain the degree of

DOCTOR OF SCIENCES of ETH ZURICH
(Dr. sc. ETH Zurich)
presented by
MICHAEL KRÜSI
MSc ETH ME
born 09.10.1985
citizen of Schönengrund (AR)
accepted on the recommendation of

Prof. Dr. Aldo Steinfeld, examiner
Prof. Dr. Claudio Augusto Oller do Nascimento, co-examiner
Dr. Zoran R. Jovanovic, co-examiner
2014
to my fiancée, Sholpan
Abstract
This thesis investigates the production of synthesis gas, a mixture of hydrogen

and carbon monoxide, via steam-based thermochemical gasification of biomass
using concentrated solar energy for process heat.
Energy and exergy analyses of the gasification of Brazilian sugarcane
bagasse revealed the potential benefits of solar-driven over conventional
autothermal gasification that include the superior quality of the syngas and the
higher yield per unit of feedstock. Theoretical upgrade factors (ratio of the
energy content of syngas produced over that of the feedstock) of up to 126%,
along with the treatment of wet feedstock and the elimination of the air
separation unit, support the potential benefits of solar-driven over autothermal
gasification.
Reaction rates for the gasification of fast pyrolyzed bagasse char were
measured by thermogravimetric analysis and a rate law based on the oxygen
exchange mechanism was formulated.
A two-zone laboratory-scale allothermal gasifier that combines drop-tube
and fixed-bed concepts was developed with the aim to provide pyrolysis and
gasification conditions yielding high carbon conversion into syngas and low
amounts of tar and gaseous hydrocarbons. In the upper drop-tube zone, a high
radiative heat flux to the dispersed particles induces their fast pyrolysis. In the
lower zone, a fixed bed provides sufficient residence time and temperature for
the char gasification and the decomposition of the other pyrolysis products.
Testing was performed in an electric furnace. Experimental runs at reactor
temperatures of 1073–1573 K and a biomass feed rate of 2.8 g/s-m2 yielded
high-quality syngas of molar ratios H2/CO = 1.6 and CO2/CO = 0.31, and with
lower heating values of 15.3–16.9 MJ/kg, resulting in an upgrade factor of
112%.
vi Abstract
The two-zone reactor concept was then further developed and a solar
reactor was built and evaluated experimentally using simulated concentrated
solar energy at a 1.5 kW solar input scale. In order to enhance the heat transfer
to the lower char gasification zone, the fixed bed was replaced by a trickle bed
established by a structured packing made of well conducting reticulate porous
ceramic (RPC) foam. The packing provides residence time for the solids and
enhances the heat transfer for the efficient char gasification and for the
decomposition of the hydrocarbons. A series of 20 min gasification
experiments comparing the two-zone reactor vs. a drop-tube reactor were
performed with a maximum particle flux of 16 g/s-m2. It has been demonstrated
that the former allows for more efficient decomposition of CH4 and C2
hydrocarbons at comparable reactor temperatures. The LHV of the product gas
was around 15.9 MJ/kg, thus significantly higher than those typically obtained
in conventional autothermal gasifiers. Solar energy was chemically stored in
the product gas resulting in an upgrade factor of 105% and a maximum energy
conversion efficiency of 21%. An analysis of the heat losses of the reactor
identified the main losses via conduction through the insulation and along the
reactor tube.
To study the heat and mass transfer in the trickle-bed zone of the two-zone
reactor, an externally heated gas-solid trickle-flow reactor with a RPC packing
was tested with beech char particles in a series of 43–51 min long experiments
achieving carbon conversions of up to 52%. A two-dimensional finite volume
model coupling chemical reaction with conduction, convection, and radiation of
heat within the porous structure was developed and tested against
experimentally observed temperatures and gasification rates. The sensitivity of
the gasification rate and reactor temperatures to variations of the RPC’s pore
diameter, porosity, thermal conductivity, and particle loading was studied.
Furthermore, a numerical comparison with a moving bed demonstrated that the
increased heat transfer via combined radiation and conduction leads to a more
uniform temperature distribution and higher gasification rates.
Zusammenfassung
Die vorliegende Doktorarbeit untersucht die Herstellung von Synthesegas, einer

Mischung aus Wasserstoff und Kohlenmonoxid, durch thermochemische
Dampfvergasung von Biomasse mithilfe konzentrierter Sonnenenergie als Pro-
zesswärme.
Eine energetische und exergetische Evaluierung der Dampfvergasung von
brasilianischer Zuckerrohr-Bagasse zeigt die potenziellen Vorteile der solar
betriebenen gegenüber der herkömmlichen autothermen Vergasung auf. Diese
umfassen die überlegene Qualität des Synthesegases sowie die höhere Ausbeute
pro Biomasseeinheit. Die theoretische Erhöhung des Energiegehalts des Syn-
thesegases gegenüber dem Ausgangsstoff von 26%, sowie die sich eröffnenden
Möglichkeiten feuchte Biomasse zu verwenden und auf eine Luftzerlegungsan-
lage zu verzichten unterstreichen die potentiellen Vorteile der solarbetriebenen
gegenüber der herkömmlichen Vergasung.
Die Reaktionsraten der Vergasung von schnell pyrolysierter Bagassekohle
wurden durch thermogravimetrische Analyse gemessen. Daraus wurde dann ein
Geschwindigkeitsgesetz auf der Grundlage des Sauerstoff-Austausch-Mecha-
nismus formuliert.
Ein Zweizonenvergaser im Labormassstab, der Fallrohr- und Festbettkon-
zepte kombiniert, wurde mit dem Ziel entwickelt, Pyrolyse- und Vergasungsbe-
dingungen zu schaffen, die für einen hohen Umsatz der Kohle zu Synthesegas
sowie für geringe Mengen Teer und gasförmiger Kohlenwasserstoffe sorgen. In
der oberen Fallrohrzone induziert ein hoher Strahlungswärmefluss die schnelle
Pyrolyse der dispergierten Biomassepartikel. In der unteren Zone stellt ein
Festbett genügend lange Feststoffverweilzeiten und genügend hohe Temperatu-
ren für die Kohlevergasung und die Zersetzung der anderen Pyrolyseprodukte
zur Verfügung. Die experimentellen Tests wurden an einem Elektroofen bei
Temperaturen im Bereich 1073–1573 K und Biomasseströmen von 2.8 g/s-m2
viii Zusammenfassung
durchgeführt. Dabei wurde qualitativ hochwertiges Synthesegas mit molaren

H2/CO- und CO2/CO-Verhältnissen von 1.6 bzw. 0.31 sowie Heizwerten von
15.3–16.9 MJ/kg produziert. Dies entspricht einer Erhöhung des Energiegehalts
gegenüber dem Ausgangsstoff um 12%.
Das Zweizonen-Reaktorkonzept wurde dann weiterentwickelt und es wurde
ein Solarreaktor gebaut und experimentell unter simulierter konzentrierter
Sonnenenergie bei Eingangsleistungen bis 1.5 kW getestet. Um den Wärme-
transport in der untenliegenden Kohlevergasungszone zu verbessern wurde das
Festbett durch ein Rieselbett ersetzt. Das Rieselbett wurde mithilfe einer
strukturierten Packung aus netzartiger poröser Keramik mit guter Wärme-
leitfähigkeit realisiert. Die Packung schafft genügend hohe Feststoffverweil-
zeiten und erhöht den Wärmetransport für eine effiziente Kohlevergasung und
die Zersetzung der Kohlenwasserstoffe. Zum Vergleich des Zweizonen-Reak-
tors mit einem Fallrohrreaktor wurde eine Serie 20-minütiger Vergasungsexpe-
rimente mit maximalen Feststoffmassenflüssen von 16 g/s-m2 durchgeführt. Es
konnte gezeigt werden, dass der erstere bei vergleichbaren Temperaturen eine
effizientere Zersetzung von CH4 und C2-Kohlenwasserstoffen erlaubt. Der
Heizwert des Produktgases war mit 15.9 MJ/kg signifikant höher als die typi-
scherweise in konventionellen autothermen Vergasern erreichten Werte. Solar-
energie konnte erfolgreich chemisch im Produktgas gespeichert werden. Die
Erhöhung des Energiegehalts des Synthesegases gegenüber dem Ausgangsstoff
betrug 5%. Dabei wurde eine maximale Energieumsatzseffizienz von 21% er-
reicht. Mit einer Wärmeverlustanalyse des Solarreaktors wurden die Wärmelei-
tung durch die Isolation sowie entlang des Reaktorrohres als die grössten Ver-
luste identifiziert.
Um den Wärme- und Stofftransport im Rieselbett des Zweizonen-Reaktors
genauer zu untersuchen wurde ein extern beheizter Gas-Feststoff-Rieselbett-
reaktor mit einer porösen Keramikpackung in einer Serie von 43–51-minütiger
Experimente mit Buchenholzkohlengreis getestet. Dabei wurde ein Kohlen-
stoffumsatz von 52% erreicht. Es wurde ein zweidimensionales Finite-
Volumen-Modell für Wärmeübergang und Stofftransport entwickelt. Dieses
koppelt die chemische Reaktion mit dem Wärmetransport innerhalb der po-
rösen Struktur durch Wärmeleitung, Konvektion und Strahlung. Die Modell-
ix
vorhersagen wurden mit den im Reaktor gemessenen Temperaturen und Ver-

gasungsraten verglichen. Eine Sensitivitätsanalyse zeigt den Einfluss von Ver-
änderungen des Porendurchmessers, der Porosität, der Wärmeleitfähigkeit und
der Partikelbeladung des RPC’s auf die Vergasungsrate und die Reaktortempe-
raturen. Ein numerischer Vergleich mit einem Wanderbettreaktor demonstrier-
te, dass die Erhöhung des Wärmetransports durch kombinierte Wärmeleitung
und -strahlung zu einer gleichmässigeren Temperaturverteilung sowie höheren
Vergasungsraten führt.
Acknowledgements
First of all, I thank Prof. Dr. Aldo Steinfeld, head of the Professorship of
Renewable Energy Carriers at ETH Zürich, for giving me the opportunity to
perform my dissertation in this interesting, challenging, and exciting field of
research, and for his supervision and support.
I am very grateful to Prof. Claudio Augusto Oller do Nascimento from the
Departamento de Engenharia Química, at the Escola Politécnica of the
Universidade de São Paulo for the collaboration in the project, the inspiring
support, and for acting as a co-examiner.
Very special thanks go to Dr. Zoran Jovanovic, my direct supervisor, for his
guidance, the many challenging and critical discussions, and for being co-
examiner. His confidence in my work and his support were extremely
important for my motivation and the success of this work.
I thank Dr. Andreas Haselbacher for his support in the modeling tasks of
this work and Dr. Hyug Chul Yoon my former advisor.
I acknowledge the support of Dr. Christian Wieckert and Alwin Frei from
the Solar Technology Laboratory at the Paul Scherrer Institute.
I thank the Bachelor and Master students Serge Zihlmann, Jules Petit,
Alexander David, Elena Cândida dos Santos, Michele Bernini, and Adrian
Ljutic for their contributions to this work and support in the experimental
campaigns.
I acknowledge the technical support in the design and construction of the
experimental setups that I have received from Philipp Haueter, Dominik
Herrmann, and Laurenz Schlumpf.
I thank all my colleagues at the Professorship of Renewable Energy Carriers
for providing a joyful research environment. Special thanks go to my longtime
office mates Thomas Cooper, Philipp Furler, and Peter Poživil for the many
fruitful personal and scientific discussions, Dr. Nicolas Piatkowski and Dr.
xii Acknowledgements
Gilles Maag who supported me with their knowledge of solar gasification, and
Dr. Jonathan Scheffe and Dr. Matt Roesle for their many inputs.
I would like to thank my parents Cornelia and Werner, my sister Alexandra,
and all friends for the support they have given me over all the years.
Last but not least I thank my fiancée Sholpan for her love, support, and
encouragement she has given me throughout my doctoral studies.
This project has been made possible by the financial support of the
Brazilian-Swiss Joint Research Programme (Grant Agreement No. BJRP
011005).
Contents
Abstract ............................................................................................................. v
Zusammenfassung .......................................................................................... vii
Acknowledgements .......................................................................................... xi
Nomenclature ................................................................................................ xvii
1 Introduction ............................................................................................. 1
1.1 Thesis Outline ................................................................................ 4
2 Background .............................................................................................. 7
2.1 Principles of Biomass Gasification................................................. 7
2.1.1 Pyrolysis ........................................................................... 8
2.1.2 Gasification ....................................................................... 9
2.1.3 Tar Formation and Reduction ......................................... 11
2.2 Conventional Biomass Gasifier Technology ................................ 12
2.3 Solar Gasifier Technology............................................................ 16
3 Feedstocks .............................................................................................. 21
3.1 Sugarcane Bagasse ....................................................................... 21
3.2 Beech Char ................................................................................... 22
3.3 Physical and Chemical Properties ................................................ 22
4 Thermodynamics of Bagasse Gasification ........................................... 25

4.1 Equilibrium Considerations .......................................................... 25
4.2 1st and 2nd Law Analyses .............................................................. 30
4.3 Conclusions .................................................................................. 36
xiv Contents
5 Gasification Kinetics of Bagasse........................................................... 39

5.1 Thermogravimetric Analysis ........................................................ 39
5.2 Rate Law ...................................................................................... 42
5.2.1 Reaction Mechanism ...................................................... 42
5.2.2 Surface Area Dependence .............................................. 44
5.2.3 Evaluation of the Rate Constants .................................... 46
5.3 Conclusions .................................................................................. 48
6 Drop-Tube Fixed-Bed Solar Gasifier Concept ................................... 49

6.1 Gasifier Concept .......................................................................... 49
6.2 Gasifier Testing in an Electric Furnace ........................................ 50
6.2.1 Experimental Setup and Procedures ............................... 50
6.2.2 Results ............................................................................ 52
6.3 Conclusions .................................................................................. 55
7 Drop-Tube Trickle-Bed Solar-Driven Gasifier ................................... 57

7.1 Gasifier Concept .......................................................................... 57
7.2 Gasifier Testing on a High Flux Solar Simulator ......................... 60
7.2.1 Experimental Setup ........................................................ 60
7.2.2 Experimental Procedure ................................................. 62
7.3 Results and Discussion................................................................. 64
7.4 Energy Balance and Heat Losses ................................................. 72
7.4.1 Heat Transfer to Reactants ............................................. 73
7.4.2 Radiation Losses ............................................................. 74
7.4.3 Convective Losses .......................................................... 75
7.4.4 Conduction through the Cavity Walls ............................ 77
7.4.5 Conduction along Reactor Tube ..................................... 77
7.4.6 Interpretation .................................................................. 78
7.5 Conclusions .................................................................................. 80
xv
8 Heat- and Mass-Transfer Analysis of a Trickle-Bed Gasifier ........... 83

8.1 Experimental Investigation ........................................................... 84
8.1.1 Experimental Setup ......................................................... 84
8.1.2 Experimental Procedure .................................................. 87
8.1.3 Experimental Results ...................................................... 88
8.2 Numerical Model.......................................................................... 90
8.2.1 Governing Equations ...................................................... 90
8.2.2 Boundary Conditions ...................................................... 94
8.2.3 Domain Properties .......................................................... 96
8.2.4 Numerical Implementation ............................................. 98
8.2.5 Code Verification ............................................................ 99
8.3 Simulation Results ...................................................................... 101
8.3.1 Model Predictions versus Experimental Results ........... 101
8.3.2 Sensitivity Analysis ...................................................... 103
8.3.3 Numerical Comparison of the Trickle Bed to a
Moving Bed .................................................................. 108
8.4 Conclusions ................................................................................ 113
9 Overall Conclusions and Outlook ...................................................... 115

9.1 Thermodynamics and Kinetics ................................................... 115
9.2 Reactor Design and Testing........................................................ 117
9.3 Modeling .................................................................................... 118
9.4 Outlook ....................................................................................... 119
Appendix A: Solar Gasification of Microalgae in a Drop-Tube ............... 121

A.1 Materials ..................................................................................... 123
A.2 Experimental Setup .................................................................... 125
A.3 Experimental Procedures ............................................................ 126
A.4 Results ........................................................................................ 128
A.5 Conclusions ................................................................................ 133
xvi Contents
Appendix B: Thermal Conductivity of a Bed of Char .............................. 135
Appendix C: Numerical Implementation of the Heat and Mass Transfer

Model ............................................................................................................. 139
C.1 Definitions.................................................................................. 139
C.2 Discretization ............................................................................. 140
C.3 Determination of the Flow Field ................................................ 143
C.4 Species Equation ........................................................................ 145
C.5 Fluid Phase Energy Equation ..................................................... 148
C.6 Solid Phase Energy Equation ..................................................... 152
C.7 Solution Strategy ........................................................................ 153
List of Figures ............................................................................................... 155
List of Tables ................................................................................................ 161
References ..................................................................................................... 163
Curriculum Vitae ......................................................................................... 177
List of Publications ....................................................................................... 179

Nomenclature
Latin Characters
A surface area, m2
A0 specific surface area, m2/m3
C solar concentration ratio, -
Cp (isobaric) heat capacity, J/kg-K
d diameter, m
dp particle size, m
D diffusion/dispersion coefficient, m2/s / thickness, m
es sensible energy, J/m3
E error, -
EA activation energy, J/mol
xdestr exergy destruction rate, W
xloss exergy loss rate, W
F configuration factor, -
g acceleration of gravity, 9.81 m/s2
h mass specific enthalpy, J/kg / height, m
h molar enthalpy, J/mol
hs mass specific sensible enthalpy, J/kg
hsf interfacial heat transfer coefficient, W/m2-K
h0f enthalpy of formation, J/mol
hR reaction enthalpy, J/kg
hR reaction enthalpy, J/mol
enthalpy flow, W
I direct normal insolation (DNI), W/m2
k reaction rate, according to rate law / thermal conductivity, W/m-K
k0 frequency factor, s-1
K extinction coefficient, m-1
xviii Nomenclature
l loading, -
L length, m
LHV lower heating value, J/kg
m mass, kg
mass flow rate, kg/s
'''
C mass consumption rate of carbon per unit volume, kg/s-m3
M molar mass, kg/mol
molar flow rate, mol/s
N number of moles, mol
p pressure / partial pressure, Pa
P perimeter, m
q heat flux, W/m2
qa order of accuracy, -
Q heat transfer rate , W
r radial coordinate , m / reaction rate, s-1
R molar gas constant, 8.31447 J/mol-K / residual, -
S mass source, kg/m3-s / surface area, m2 / molar entropy, J/mol-K
t time, s
T temperature, K
u velocity (superficial in porous domain), m/s
U upgrade factor, -
V volume, m3
vC correction velocity, m/s
wk mass fraction of species k, -
ASU power to air separation unit (ASU), W
x elemental molar ratio of H/C, -
xi mass fraction of species i in the solid, -
X conversion, -
y elemental molar ratio of O/C, -
yk mole fraction of species k, -
z axial coordinate, m
xix
Greek Characters
absorptivity, -
emissivity, -
molar exergy, J/mol
thermal efficiency, -
abs absorption efficiency, -
ex exergetic efficiency, -
surface coverage, -
optical thickness, -
dynamic viscosity, Pa-s
kinematic viscosity, m2/s
density, kg/m3
Stefan-Boltzmann’s constant, 5.67·10–8 W/m2-K4
tortuosity, -
equivalence ratio, -
porosity, -
Subscripts
parallel to flow
0 at normal conditions (T0 = 273.15 K, p0 = 101,325 Pa) / initial

apt aperture
cav cavity
ch chemical
cond conduction
conv convection
eff effective
exp experiment
f final / fluid
ins insulation
xx Nomenclature
k species index
liq liquid
med medium
nom nominal
prod products
quench quencher
rad radiation
reac reactants
rerad reradiation
s solid
sim numerical simulation
sol-chem solar to chemistry
surf surface
tot total
Dimensionless Groups
Gr Grashof number
Nu Nusselt number
Pe Péclet number
Pr Prandtl number
Ra Rayleigh number
Re Reynolds number
xxi
Abbreviations
ASU air separation unit
CPC compound parabolic concentrator
CSP concentrated solar power
DMC dichloromethane
DNI direct normal insolation
ETH Eidgenössische Technische Hochschule
(Swiss Federal Institute of Technology)
FVM finite volume method
GC gas chromatography
HEX heat exchanger
HFSS high flux solar simulator
LHV lower heating value
MB moving bed
MMS method of manufactured solutions
ppi pores per inch
PV photovoltaics
RPC reticulate porous ceramic
Chapter 1
1 Introduction
Affordable, transportable, and dispatchable energy is a key driver for a thriving

economy. Today’s world primary energy demand is around 13,100 Mtoe a
(2011) of which 82% are covered by fossil sources [1]. The energy demand is
expected to increase sharply to 17,400–18,600 Mtoe by 2035. This is an
increase of 33–43% over today’s demand. The increase is mainly due to global
population growth and expanding economic activity. The rate of consumption
of fossil fuels is higher than their formation rate which inevitably leads to a
limited supply in the future. Moreover, concerns about climate change due to
CO2 emissions call for a reduction in the usage of fossil fuels [2].
To match demand and supply in the long term while limiting CO2 emissions
and keeping energy prices at an affordable level, efforts have to be made on
both, the demand and the supply side. On the demand side, a reduction in
energy intensity is crucial. Although technological progress is made, the
projected reductions cannot offset the demand growth as the population and
economic growth are larger than the efficiency gains [1].
For the electricity market, additional supply of low carbon energy may be
provided by renewable sources such as wind, biomass, geothermal and solar
(PV and CSP), or nuclear power. For the transportation and the petrochemicals
sector, the key drivers for increasing oil demand [1], these technologies have
limited applicability. Especially for transportation, where high energy density is
critical, liquid hydrocarbon fuels are indispensable.
The use of concentrated solar energy for chemical processes can play an
important role in the transition away from fossil towards renewable
a
million tonnes of oil equivalent, 1 Mtoe = 41.868 PJ
2 Chapter 1: Introduction
transportation fuels. Although the solar energy hitting the earth’s surface is
very dilute (~1 kW/m2), it can be concentrated by focusing it with mirrors. The
concentrated solar energy can then be utilized as energy source to drive high
temperature thermochemical processes such as the production of syngas a,
mixture composed of H2 and CO [3]. Thereby, the intermittent solar energy can
be chemically stored. Syngas is a primary building block for the production of
chemicals and transportation fuels. It can be processed to H2 via water-gas shift
reaction or to liquid hydrocarbon fuels such as diesel or kerosene via Fischer-
Tropsch synthesis, or to methanol and then gasoline via the methanol-to-
gasoline process (Mobil). Alternatively, the syngas may be used for the
production of other chemicals or directly as a combustion fuel for power
generation.
The production of liquid transportation fuels has several benefits as they are
storable, transportable (from optimal production sites to where the demand is),
and do not require a change in the massive global distribution and fueling
infrastructure or in the highly developed vehicle propulsion technologies. There
are two promising routes for the production of solar fuels [4]. In the
short/midterm the path to solar syngas is via thermochemical gasification of
carbonaceous feedstocks such as coal, biomass, and carbon containing waste
materials. The long-term route is the production of syngas directly from H2O
and CO2 via solar thermochemical cycles. This work focuses on the
short/midterm pathway of solar-driven steam gasification of carbonaceous
feedstock, specifically the gasification of biomass illustrated in Figure 1-1. The
use of biomass feedstock makes the product of any of the above mentioned
conversion processes to chemicals, energy, or transportation fuels, in principle,
CO2 neutral because of the biogenic source of syngas. The biomass feedstocks
used in this work are Brazilian sugarcane bagasse, a residue from the sugar and
ethanol production, and beech char.
At temperatures of 1000 K or higher and in the presence of steam, biomass
is thermochemically converted into syngas via highly endothermic reactions.
The conversion comprises two main steps: (1) pyrolysis, producing tar, gases
and char, and (2) steam-based gasification and reforming of char, gas, and tar to
form syngas [5].
3
biomass char + gas + tar hR 0 (1.1)
char / tar / gas + H2O(g) H2 + CO hR 0 (1.2)
In conventional autothermal gasifiers such as the Texaco entrained flow

gasifier, the GTI high-pressure oxygen-blown fluidized bed gasifier, or the
gasifiers specifically studied for the gasification of bagasse [6-9], the heat
required for the gasification is supplied by combusting a significant amount of
the feedstock in-situ with a stream of air or pure O2. The combustion consumes
about 30 40% of the feedstock [10, 11], reducing its utilization and the quality
of the syngas due to combustion byproducts, i.e. increased amounts of CO2 in
the syngas, lower H2/CO ratio, and a lower calorific value. Indirectly heated
(allothermal) gasifiers such as dual fluidized bed gasifiers [12] can overcome
the reduction of quality due to combustion byproducts but they still require a
significant amount of the feedstock to be combusted to provide the reaction
enthalpy.
concentrated
solar energy
O2 solar energy
biomass
CO2 solar
biomass solar syngas
gasification
H2O production H2 + CO
( HR > 0)
H2O
liquid
direct
transportation hydrogen chemicals
combustion
fuels
H2O + CO2
Figure 1-1: Solar-driven steam gasification process flow sheet.

4 Chapter 1: Introduction
Solar-driven gasification uses highly concentrated solar radiation as source

of high-temperature process heat to drive the steam-based gasification of the
carbonaceous feedstock. Thus, solar energy in an amount equal to the enthalpy
change of the endothermic reactions (Eqs. 1.1 and 1.2) is chemically stored,
which leads to a syngas with a higher calorific content per unit of feedstock.
The smaller amount of low-density feedstock that is needed to produce the
same output reduces capital and operating costs related to transportation,
storage, pre-processing, and feeding of the biomass. Furthermore, the absence
of combustion eliminates the need for pure oxygen from an air separation plant.
Since no internal combustion products contaminate the syngas, the required
size of CO2 separation plant is reduced. All these advantages may justify the
specific cost of solar concepts brought in by the concentrating mirror system
and the syngas storage imposed by intermittent availability of solar energy.
1.1 Thesis Outline
This thesis aims at developing solar reactor technology for the thermochemical
conversion of biomass to syngas via pyrolysis and gasification. It was carried
out in the framework of a joint research project of the Universidade de São
Paulo, Brazil and ETH Zürich, Switzerland.
Chapter 2 gives a background on the principles of biomass gasification,
different gasification technologies such as auto- and allothermal gasification,
and solar gasification technology. In Chapter 3, the investigated feedstocks are
introduced and characterized. Chapter 4 studies the thermodynamic equilibrium
compositions for solar and autothermal gasification of Brazilian sugarcane
bagasse emphasizing the differences in syngas quality. Moreover, first and
second law analyses are carried out in order to assess the performance of the
solar and the autothermal gasification, to find the maximum energy and exergy
conversion efficiencies, and to identify the major sources of irreversibility. In
Chapter 5, a thermogravimetric analysis is performed and a kinetic rate law for
the gasification of bagasse char is formulated as a design basis for a solar-
driven gasifier. Chapter 6 presents a solar reactor concept for the gasification of
1.1 Thesis Outline 5
highly volatile biomass feedstock that is based on a combination of drop-tube

and fixed bed designs. Further, gasification of bagasse is experimentally tested
in a tubular externally heated gasifier. In Chapter 7, the reactor design is refined
and a solar reactor consisting of a drop-tube and a trickle-bed zone is built and
tested at ETH’s high flux solar simulator. The performance of the two-zone
reactor is experimentally examined with sugarcane bagasse particles and
compared to the performance of a drop-tube configuration. In Chapter 8, an
externally heated gas-solid trickle-flow reactor with a well-conducting
reticulate porous ceramic (RPC) packing is tested with beech char particles.
Moreover, a two-dimensional finite volume heat and mass transfer model
coupling chemical reaction with heat and mass transfer is developed. Its
predictions are compared to the experimental measurements and its sensitivity
to parameter changes is analyzed. Further, a model comparison of the trickle-
flow reactor to a moving bed reactor is discussed. Finally, the overall
conclusions and an outlook are presented in Chapter 9.
Chapter 2
2 Background
2.1 Principles of Biomass Gasification
Biomass is bulky and has a low energy density (LHV 15–19 MJ/kg) [13].
This makes its handling, storage and transportation difficult and expensive. The
gasification of the biomass is an attractive path to convert this low-value
material into high-value products. In the gasification process, a wide range of
carbonaceous materials can be converted into a combustible gas or a synthesis
gas containing mainly H2, CO, CO2, and CH4.
The thermochemical conversion of biomass into synthesis gas is usually
performed at temperatures of 1000 K or higher using air, oxygen, steam, or
CO2 as gasifying medium. Depending on the gasifier design and gasifying
medium, the gasification process may involve several of the following physical,
chemical, or thermal process steps [14]:
- drying
- thermal decomposition or pyrolysis
- partial combustion
- gasification of decomposition products
For the gasification of biomass with steam to synthesis gas, as it is proposed in

this work, the overall thermochemical conversion can be expressed by the
idealized simplified net reaction
8 Chapter 2: Background
x
CH x O y 1 y H2 O 1 y H 2 CO (2.1)
2
where x and y are the elemental molar ratios of H/C and O/C in the biomass.
2.1.1 Pyrolysis
Pyrolysis is the thermal decomposition of the feedstock into char, gases, and tar
in the total absence or a limited supply of an oxidizing agent such as O2, H2O,
or CO2. It is an essential process step in every gasifier. Figure 2-1 illustrates the
pyrolysis of a biomass particle. When the biomass is heated to temperatures
above 500 K via radiation and convection, gas such as light volatile
hydrocarbons, H2, CO, CO2, and tar are released, leaving behind char. The tar
then undergoes secondary gas-phase reactions in which the large complex
hydrocarbons break down into smaller molecules.
radiative and gas

convective heat
thermal liquid
boundary
layer biomass tar char
conduction gas
and pore char
convection
primary decomposition gas-phase secondary

reactions tar-cracking reactions
Figure 2-1: Primary and secondary reactions during pyrolysis of a biomass

particle. Adapted from ref. [14].
2.1 Principles of Biomass Gasification 9
The final pyrolysis products can be classified as [14]:
- solid (char or carbon)
- liquid (tar, heavy hydrocarbons, water)
- gaseous (H2, CO, CO2, CH4, and light hydrocarbons)
The biomass pyrolysis conditions such as heating rate, gas temperature, and
residence time have a strong influence on the release of tar and gases, as well as
on the formation and reactivity of the char. In general, the release of volatiles is
enhanced and char formation suppressed by rapid high-temperature pyrolysis
[15, 16]. Such pyrolysis conditions yield only low amounts of char that has a
high reactivity [17-20]. Moreover, high gas temperatures facilitate cracking of
tar and other hydrocarbons [18, 21-23]. Slower heating and longer residence
time lead to the formation of secondary char, which is produced from a reaction
between the primary char and the volatiles [14].
2.1.2 Gasification
In the gasification step, char and hydrocarbons released during the pyrolysis
react with an oxidant such as air, oxygen, steam, or CO2 to produce syngas. It is
usually performed in the temperature range 1000–1300 K and can be
exothermic or endothermic depending on the oxidant used. The primary solid
remainder of the gasification is ash, which consists of mineral matter and minor
amounts of unreacted carbon.
The gasification and reforming of the char and the hydrocarbons are
complex processes consisting of several homogeneous and heterogeneous
reactions. The principal reactions and their reaction enthalpies are given by
Eqs. 2.2–2.10 [24].
Combustion reactions
C(s) + 0.5 O2 CO h298K 111 kJ/mol (2.2)
CO + 0.5 O2 CO2 h298K 283 kJ/mol (2.3)
H2 + 0.5 O2 H2O(g) h298K 242 kJ/mol (2.4)
Water-gas reaction
C(s) + H2O(g) H2 + CO h298K 131 kJ/mol (2.5)
Boudouard reaction
C(s) + CO2 2 CO h298K 172 kJ/mol (2.6)
Hydrogasification reaction
C(s) + 2 H2 CH4 h298K 75kJ/mol (2.7)
Water-gas shift reaction
CO + H2O(g) H2 + CO2 h298K 41kJ/mol (2.8)
Steam reforming reactions
CH4 + H2O(g) 3 H2 + CO h298K 206 kJ/mol (2.9)
x
CnHx + n H2O(g) n H2 + n CO h298K 0 (2.10)
2
The rates of the heterogeneous reactions vary greatly with the combustion
reaction (Eq. 2.2) being the fastest followed by the water-gas reaction (Eq. 2.5),
the Boudouard reaction (Eq. 2.6), and the hydrogasification reaction (Eq. 2.7)
[14].
2.1 Principles of Biomass Gasification 11
RC+O2 RC+H 2 O RC+CO2 RC+H2 (2.11)
Although the production of char with high reactivity, discussed in the previous
section, is beneficial for the heterogeneous char gasification reactions, the char
gasification remains the rate limiting step of the overall process [5]. The water-
gas shift reaction (Eq. 2.8) is of particular interest for the production of
synthesis gas as it allows adjusting the CO/H2 ratio for the production of liquid
fuels.
2.1.3 Tar Formation and Reduction
Tar is an undesired byproduct of the pyrolysis step. Devi et al. [25] define it as
a complex mixture of condensable hydrocarbons, which includes single- to 5-
ring hydrocarbons, other oxygen-containing hydrocarbons, and complex
polycyclic aromatic hydrocarbons. In other definitions, all organic product gas
contaminants with molecular mass larger than 78, i.e. larger than benzene are
considered as tar [26].
Tar can condensate and plug downstream equipment, and contaminate the
product gas limiting its use. Keeping the tar levels low is necessary as most
downstream processes can only handle a certain amount of tar. Internal
combustion and diesel engines tolerate 10–100 mg/Nm3 and gas turbines 0.5–
5 mg/Nm3 [27]. For synthesis gas applications such as methanol or Fischer-
Tropsch synthesis, the tar content should not exceed 0.1 mg/Nm3 [28] as tar
poisons the catalyst.
Tar is mainly produced during the pyrolysis phase. In the temperature range
of 500–800 K, the biomass components cellulose, hemicellulose, and lignin
break down into primary tar. At temperatures above 800 K, primary tar
components start to decompose into lighter non-condensable gases (CO2, CO,
H2O, and light hydrocarbons), and longer chain hydrocarbons molecules known
as secondary tar. At temperatures higher than 1050 K, primary tar is destroyed
and tertiary tar such as benzene, naphthalene, and 3- and 4-ring aromatics are
produced. [14, 29]
The methods to reduce the amount of tar can be classified into primary and
secondary methods. The primary methods avoid or convert tar formed within
the gasifier by adjusting the operating conditions that play an important role in
the tar formation and reduction. An increase in the gasifier temperature reduces
most tar components. Only the formation of tertiary is increased tar. For many
gasifier types, a limiting factor to the increase in temperature is the risk of
sintering or melting of the ash. For the gasification with steam, an increase in
the steam to biomass ratio also leads to a reduction in the tar content. Tar and
other hydrocarbons are then reduced via steam-reforming reactions. [14]
x
CnHx + n H2O(g) n H2 + n CO (2.12)
2
The addition of a catalyst to the gasifier is another primary method for the
reduction of tar. Dolomite, olivine, alkali metal, nickel and also char are
catalytically active and enhance tar reforming within the gasifier. [25]
Secondary methods reduce the tar content downstream of the reactor via
thermal or catalytic cracking, or via mechanical separation using cyclones,
filters, etc. [25].
2.2 Conventional Biomass Gasifier Technology
Several reactor types are being studied or are industrially used for the
gasification of biomass. These include updraft, downdraft, fluidized bed, dual
fluidized bed, and entrained-flow gasifiers, of which a brief description is given
below. Figure 2-2 gives a schematic overview of these gasifier types and their
main flow streams. The colors indicate the relative reactor temperatures. Red
stands for hot, yellow for medium, and gray for cold sections of the reactor. An
overview of typical gas compositions is given in Table 2-1.
2.2 Conventional Biomass Gasifier Technology 13
a) updraft b) downdraft
biomass product gas biomass
fixed bed
fixed bed of
of biomass throat
biomass
oxidant oxidant
grate grate
ash oxidant product gas + ash
c) bubbling fluidized bed d) circulating fluidized bed

gas + ash + bed product gas
product gas material
ash freeboard cyclone
fluidized fluidized
bed bed
biomass biomass
ash
ash oxidant ash oxidant
e) dual fluidized bed product gas

p f) entrained-flow
gas + char + sand
gasifier biomass oxidant
cyclone
freeboard
flue
combustor gas dispeersed
dispersed
particles
partiicles
fluidized fluidized
flu
fl uidized
biomass bed hot bed
sand
ash product
air gas
steam ash
recycle gas
Figure 2-2: Biomass gasifier types with main flow streams: a) updraft
(countercurrent), b) downdraft (concurrent), c) bubbling fluidized bed,
d) circulating fluidized bed, e) dual fluidized bed, and f) entrained-flow. Colors
indicate relative temperatures (red = hot, yellow = medium, gray = cold).
Adapted from ref. [5].
In updraft gasifiers schematically shown in Figure 2-2a, a packed bed of

biomass supported by a grate moves downwards. The oxidant is introduced
from the bottom and the product gas is withdrawn from the top. Ash is removed
at the bottom of the gasifier. Such a design is very simple, robust, and scalable.
Due to its countercurrent arrangement it offers a high thermal efficiency, yields
a high carbon conversion, and can tolerate feedstock moisture contents of up to
60%wt. The tar levels are very high because the gaseous and liquid pyrolysis
products never pass the hottest zone of the reactor. It is thus unsuitable for the
production of clean product gas for highly volatile feedstocks. [5, 14]
Downdraft gasifiers, as depicted in Figure 2-2b, have a descending packed
bed of biomass that is supported by a throat at which air or oxygen is injected.
Both, the gas and the solids flow downwards concurrently. The tar and the
gaseous hydrocarbons released during pyrolysis in the upper zone of the
gasifier pass through the turbulent oxidation zone that is maintained at
temperatures in the range of 1273–1473 K and are cracked and reformed. The
relatively clean product gas with low tar and volatile hydrocarbon levels makes
oxygen- and air-blown downdraft gasifiers the preferred concepts for the
production of high quality syngas from highly volatile feedstocks. Due to the
throats limited size the scalability is low. The tolerated level of moisture is up
to 25%wt–30%wt. [5, 14]
Table 2-1: Typical product gas composition for different biomass gasifiers.
gas composition [%vol]
gasifier type
H2 CO CO2 CH4 N2
*
updraft (air-blown) 11 24 9 3 53
downdraft (air-blown)* 17 21 13 1 48
downdraft (oxygen-blown)* 32 48 15 2 3
fluidized bed (air-blown)* 9 14 20 7 50
dual fluidized bed† 38 27 21 11 3
*
from ref. [5]
†
from ref. [30]
2.2 Conventional Biomass Gasifier Technology 15
There are two types of single fluidized bed gasifiers used for biomass:
bubbling (Figure 2-2c) and circulating (Figure 2-2d). In the bubbling fluidized
bed gasifiers, the solids stay in reactor because the fluid velocity is low. In the
circulating fluidized bed gasifiers, the fluid velocity is higher so that a large
part of the solids are entrained in the product gas. The solids are then separated
with a cyclone and returned to the gasifier to increase the carbon conversion,
which is eventually higher than in the bubbling bed. Both fluidized bed
concepts are characterized by a good gas-solid mixing and high temperature
uniformity. The typical operating temperatures are between 1073 and 1173 K.
The tar levels in the product gas are between the levels of updraft and
downdraft gasifiers. [5, 14]
Dual fluidized bed gasifiers (Figure 2-2e) are allothermal, which allows
gasifying the biomass using pure steam. Hot sand is used to supply the reaction
enthalpy to the gasifier. The sand is heated in a separate fluidized bed
combustor that burns the char produced in the gasifier. Dual fluidized bed
gasifiers yield medium heating value gas without requiring pure O2 as oxidant.
As the gasification is done with steam, the product gas contains high amounts
of H2. Without additional fuel in the combustor, the temperatures in the gasifier
are low (~1073 K) resulting in high CH4 and moderate tar levels. The design is
very complex and costly. [5, 14, 30]
In entrained-flow gasifiers (Figure 2-2f), the solid feedstock and the oxidant
are injected into the chamber from the top or the side with a high velocity jet.
The ash and the product gas leave the chamber at the bottom and are then
separated. The reactor temperatures are usually well above 1273 K. This means
that the ash accrues as a slag. The high reactor temperatures especially at the
inlet lead to a good tar decomposition. Entrained-flow gasifiers are successfully
used for large scale coal gasification but its suitability for the commercial
gasification of biomass is questionable. The short residence times of only a few
seconds require very fine particles, which are difficult to produce from fibrous
biomass. Moreover, the molten ash is very aggressive and can corrode the
reactor lining. [5, 14]
2.3 Solar Gasifier Technology
For the steam gasification of carbonaceous feedstock reactor temperatures in

the range of 1000–1500 K have to be reached. Providing these temperatures by
solar energy requires a mean solar concentration ratio on the order of C =
1,000–2,000 sunsa. Such concentrations can be provided by two types of solar
concentrating systems, parabolic dishes (C = 1,000–10,000 suns) and tower
configurations (C = 500–5,000 suns) [3]. Solar gasification reactors are usually
designed to operate in conjunction with a solar tower system due to the larger
scale of the system. As shown in Figure 2-3a, the solar reactor is positioned on
top of a central tower that is surrounded by a large field two-axis tracking
mirrors, so-called heliostats. The incident solar radiation is being reflected by
the heliostats and concentrated at the aperture of the solar reactor to supply
high-temperature heat to the endothermic steam gasification process. For solar
reactor configurations with an aperture at the top of the reactor, beam-down
systems as shown in Figure 2-3b are used. A mirror, commonly of hyperbolic
a) b)
solar beam down
reactor mirror
solar reactor
heliostat field tower heliostat field tower
Figure 2-3: Solar tower concentrating systems: a) solar reactor on top of

tower and b) beam down design with secondary hyperbolic mirror on tower and
solar reactor at ground. Adapted from ref. [31].
a
1 sun = 1 kW/m2/DNI, DNI = direct normal irradiation [kW/m2]
2.3 Solar Gasifier Technology 17
shape, is installed at the top of the tower redirecting the concentrated solar
radiation to the reactor situated on the ground.
Several solar reactor concepts for the gasification of various types of
carbonaceous feedstock have been investigated [32, 33] and experimentally
demonstrated on a laboratory-scale [34-40]. An overview of the reactor
configurations and the feedstocks used is given in Table 2-2. Most concepts
exploit a cavity-receiver configuration, in which the concentrated solar
radiation passes through an aperture into a cavity to create a high-temperature
region within the cavity and supply high-temperature heat to the endothermic
reactions. Such configurations are suitable for solar concentrating applications
as they approach blackbody absorbers and minimize reradiation losses, while
providing a homogeneous temperature distribution through multiple internal
reflections and reradiation [3].
As for solar-driven gasification no heat is generated within the gasifier,
efficient transfer of the concentrated solar radiation to the reaction site is
critical for high productivity and favorable gas-phase selectivity. Directly-
irradiated solar gasifiers, where the solar radiation is absorbed directly by the
Table 2-2: Existing solar gasifier concepts and feedstocks used.
reference reactor concept feedstock
[32-34] directly irradiated activated charcoal, flexicoke, PD

fluidized / packed bed coke
[36] directly irradiated packed coal, activated carbon, petroleum

bed gravity-feed coke, walnut shells mixed with coal
[37] directly and indirectly cellulose

irradiated fluidized bed
[38, 39, 41] indirectly irradiated cellulose, beech char

entrained-flow
[40] indirectly irradiated scrap tire chips and powders,

packed-bed sewage and industrial sludge, fluff
feedstock at the reaction site, enable high heat transfer rates. Yet, those also
require a transparent window that has to be kept clean during operation. In
addition, the window introduces limitations in the operating pressure and the
scale-up as the window designs become very complex. [32-37, 42-44]
In indirectly-irradiated reactors the incident solar radiation impinges on the
outer wall of an opaque absorber which confines feedstock. The heat is
transferred to the inner absorber wall by conduction and from there to the
feedstock by convection and radiation. Therefore, the need for a window is
eliminated at the expense of having less efficient heat transfer. This imposes
even more stringent constraints on the materials of the absorber with regards to
its operating temperature, chemical stability, thermal conductivity, radiative
absorptance, and resistance to thermal shock. [10, 37-41, 45-47]
A commonly suggested method to achieve the required heat transfer rates in
an indirectly-irradiated gasifier is the drop-tube or entrained-flow reactor
concept. As shown in Figure 2-4a, the solar radiation is concentrated through
an aperture into either a specularly reflecting [38, 46] or absorbing [39, 41, 46,
47] cavity-receiver housing one or more vertical reactor tubes. The reactor
tubes absorb the high-flux irradiation and re-radiate the heat to the particles
flowing through. However, this reactor concept is suitable only for particles of
up to a couple of hundred microns in size for which the radiative heat transfer
mode is dominant [48] and the residence time of the order of a second is long
enough. Grinding raw biomass to this size range imposes high capital and
operating costs that often justify a partial low temperature pyrolysis of the
biomass (torrefaction) in order to improve its grindability [49]. At the same
time, as the gas is mainly convectively heated by the surface area of the
particles that are in a rather dilute flow, this kind of reactor generally does not
provide gas temperatures high enough for effective tar and methane cracking
and reforming. Alternative reactor concepts such as packed or moving beds
allow the use of coarser biomass particles by providing reaction time that is
sufficiently long for high carbon conversion. An example is shown in
Figure 2-4b. The concentrated solar radiation enters the reactor from the top, is
absorbed by an absorber-emitter plate. The plate reradiates onto the packed bed
in the lower section of the reactor that is supplied with steam from the bottom.
2.3 Solar Gasifier Technology 19
a) feedstock and steam inlet b) product absorber /

gas outlet emitter plate
packed bed
product of feedstock
gas outlet steam inlet
Figure 2-4: Solar gasifier concepts: a) tower-mounted drop-tube [41] and

b) packed bed for a beam-down configuration [40]. Arrow indicates where the
concentrated solar radiation enters the reactor.
Unfortunately, these concepts suffer from significantly impaired overall heat

transfer due to high extinction of radiation by the densely packed bed of
feedstock and ash [40, 44]. As the heat transfer becomes the limiting factor,
large temperature gradients and non-uniform reaction rates adversely impact
energy conversion efficiency. The scale-up from pilot to commercial scale is
generally driven by providing the heat transfer rates that are necessary for
target product rate and quality. Preferably, this is achieved by increasing the
number of cavities or, for tubular reactors, by increasing the number of tubes
within the cavity to adjust the heat transfer area per throughput [41, 50]. The
reasonable scale for commercial solar thermochemical plants is considered to
be in the range 10–100 MW of solar power measured at the aperture of the
cavity to the reactor [41, 50-53]. This corresponds to a heliostat area of
approximately 28,000–287,000 m2 [53].
Chapter 3
3 Feedstocksa
The feedstocks used for the experimentation and the modeling work of this
thesis are Brazilian sugarcane bagasse and beech char (Figure 3-1). In this
chapter they are briefly described and characterized.
a) b)
Figure 3-1: Biomass feedstocks used: a) Brazilian sugarcane bagasse and

b) beech char.
3.1 Sugarcane Bagasse
Sugarcane bagasse is a fibrous residue of the sugar or ethanol production from

sugarcane. It remains after crushing the sugarcane to extract their juice. Each
metric ton of sugarcane processed produces roughly 280 kg of bagasse, with a
moisture content in the order of 50%wt [54, 55]. On a dry and ash free basis,
a
Material of this chapter has been published in: M. Kruesi, Z. R. Jovanovic,
E. C. dos Santos, H. C. Yoon, and A. Steinfeld, "Solar-driven steam-based
gasification of sugarcane bagasse in a combined drop-tube and fixed-bed
reactor – Thermodynamic, kinetic, and experimental analyses", Biomass and
Bioenergy, vol. 52, pp. 173-183, 2013.
22 Chapter 3: Feedstocks
bagasse is composed of about 50%wt cellulose, 25%wt hemicellulose, and 25%wt

lignin [56].
The world production of sugarcane is 1,661 Mt/y, out of which 40% are
produced in Brazil alone [57]. The arising bagasse is mainly combusted in
boilers at low conversion efficiencies to satisfy the heat and electricity demand
of the sugar and alcohol industry. The unused quantity represents about 20% of
the total bagasse production [58]. Thus, bagasse is a vast and readily available
source of waste biomass.
3.2 Beech Char
Beech char or charcoal is a pyrolyzed biomass. As such it has high fixed carbon
and low content of volatiles. The beech char was chosen as a model feedstock
as it allows studying the steam gasification section of a solar reactor
independently from the pyrolysis section because the measurements are not
intruded by the presence of pyrolysis products.
3.3 Physical and Chemical Properties
The feedstocks were characterized by elemental composition, volatiles, fixed

carbon and ash contents, lower heating value, and particle size distribution. The
characteristic values are reported in Table 3-1. Before all experiments and
analyses the feedstocks were dried for at least 4 h at 378 K. To obtain more
homogeneous material properties, the fibrous bagasse was sieved with 1 mm
mesh size on a sieve shaker. The beech char particles were ground and sieved
to a size of 0.56–1.00 mm.
Ultimate analysis – The ultimate analysis to determine the elemental

composition was done with CHN-900 for carbon, hydrogen, and nitrogen, RO-
478 for oxygen, and CHNS-932 for sulfur (all LECO Corporation, St. Joseph,
MI). The unpyrolyzed bagasse has a carbon content of 43%wt, a hydrogen
content of 6%wt, and an oxygen content of 38%wt. The contents of nitrogen and
3.3 Physical and Chemical Properties 23
sulfur were 0.41%wt and 0.09%wt. In contrast, the already pyrolyzed beech char
has a very high carbon content of 86%wt and low hydrogen and oxygen contents
of 2.5%wt and 8%wt, respectively. Nitrogen and sulfur are in the same range as
for the bagasse. On a dry basis and after neglecting the presence of ash,
nitrogen and sulfur, the bagasse and the beech char can be represented by the
average chemical formulas CH1.665O0.663 and CH0.349O0.072, respectively.
Proximate analysis – The proximate analysis to determine the content of

volatiles, fixed carbon, and ash was done in a thermogravimetric balance
(Netzsch STA 409 CD). The samples were heated to 393 K in an Ar
atmosphere and kept at this temperature for 20 min to finalize the drying
process. Then they were heated at 20 K/min to 873 K and held there for 20 min
to pyrolyze, attributing the mass loss to the volatiles content. Finally, O2 was
introduced and the temperature was increased a rate of 20 K/min to 1273 K at
and kept there for another 20 min to combust the fixed carbon. The remaining
Table 3-1: Physical and chemical properties of bagasse and beech char:
Ultimate and proximate analyses, heating value and mean particle size.
bagasse beech char
carbon (C) [%wt] 42.51 85.94
hydrogen (H) [%wt] 5.94 2.50
oxygen (O) [%wt] 37.54 8.23
nitrogen (N) [%wt] 0.41 0.92
sulfur (S) [%wt] 0.09 0.05
H/C = x [mol/mol] 1.665 0.349
O/C = y [mol/mol] 0.663 0.072
volatiles [%wt] 77.3 17.1*
fixed carbon [%wt] 12.6 80.8*
ash [%wt] 10.2 2.1*
LHV [MJ/kg] 16.50 32.12*
dp [ m] 455 811
*
from ref. [40]
24 Chapter 3: Feedstocks
mass was attributed to ash. The bagasse has a volatiles content of around 77%wt
and a fixed carbon content of 13%wt. The beech char has a volatiles content of
only 17%wt but a fixed carbon content of more than 80%wt. The ash content of
the bagasse is with 10%wt far higher than the one of the beech with only 2%wt.
Heating value – The lower heating value (LHV) was determined with
calorimeter measurements (C7000, IKA-Werke). Due to the higher fixed
carbon content in the beech char, the LHV of the beech char (32.12 MJ/kg) is
higher than the one of the raw bagasse (16.50 MJ/kg).
Particle size distribution – The particle size plays a very important role in the
design and performance of a solar reactor as they affect the effective heating
rates during pyrolysis and influence the time needed for the gasification step.
The particle size distributions were measured using a laser diffraction particle
size analyzer (LA-950, HORIBA). Figure 3-2 depicts the particle size
distribution for the two feedstocks. The mean particle sizes dp of bagasse and
beech char are 455 and 811 µm, respectively.
beech char
volume density f (dp) ·dp3
bagasse
1 10 100 1000
particle diameter, dp [ m]
Figure 3-2: Particle size distribution functions of ground and sieved bagasse,
and beech char analyzed with LA-950 analyzer (HORIBA).
Chapter 4
4 Thermodynamics of Bagasse Gasificationa
In this chapter, thermodynamic equilibrium compositions are determined for

the solar and the autothermal gasification of Brazilian sugarcane bagasse with
steam emphasizing the differences in the syngas composition. Moreover, first
and second law analyses of the solar and the autothermal gasification are
conducted in order to determine the maximum energy and exergy conversion
efficiencies, and to identify the major sources of irreversibility.
4.1 Equilibrium Considerations
The overall thermochemical conversion can be expressed by the simplified net

reaction
x
CH x O y 1- y H 2 O 1 - y H2 CO hR > 0 (4.1)
2
where x and y are the elemental molar ratios of H/C and O/C in the bagasse,
respectively. Equilibrium compositions of the system of Eq. 4.1 with
CH1.665O0.663 were computed by the direct Gibbs energy minimization technique
using the commercial HSC code [59]. Nitrogen and sulfur were neglected as
their presence in trace amounts does not affect the equilibrium compositions.
a
Bioenergy, vol. 52, pp. 173-183, 2013.
26 Chapter 4: Thermodynamics of Bagasse Gasification
1.2
H2
equilibrium composition [mol] 1.0
0.8 CO
H2O
0.6
0.4
CO2
C(s)
0.2
CH4
0.0
600 800 1000 1200 1400 1600
temperature [K]
Figure 4-1: Equilibrium composition as a function of temperature of the

stoichiometric system of Eq. 4.1 for bagasse (CH1.665O0.663) at 1 bar.
1.2
equilibrium composition [mol]
1.0 H2
0.8
H2O CO
0.6
0.4 CO2
C(s)
0.2
CH4
0.0
600 800 1000 1200 1400 1600
temperature [K]

stoichiometric system of Eq. 4.1 for bagasse (CH1.665O0.663) at 10 bar.
4.1 Equilibrium Considerations 27
Figure 4-1 shows the equilibrium compositions as a function of temperature

at an absolute pressure of 1 bar. Product species with computed mole fractions
less than 10-5 have been omitted. CH4, CO2, H2O, and C are the
thermodynamically favored species at temperatures below 800 K. At
temperatures above 1350 K, H2 and CO are thermodynamically favored,
yielding a syngas with a molar H2/CO ratio of 1.17. According to Le
Châtelier’s principle, the thermodynamic equilibrium at elevated pressures
favors the production of CH4, CO2, H2O, and C. This can be seen in Figure 4-2
which shows the equilibrium composition at a pressure of 10 bar. As a result,
higher temperatures are required to achieve full conversion towards H2 and CO.
Figure 4-3 shows the reaction enthalpy hR and the lower heating value
(LHV) of the product gas as a function of temperature for the system CHxOy +
(1–y) H2O of Eq. 4.1 when reactants are fed at 298 K and 1 bar and products
are obtained at the indicated temperature having an equilibrium composition as
shown in Figure 4-1. The reference enthalpy of the bagasse is determined based
on its LHV and its composition.
20
15
[MJ/kg]
10
5
LHV
ΔhR
0
600 800 1000 1200 1400 1600
temperature [K]
Figure 4-3: Reaction enthalpy hR for the system CHxOy + (1–y) H2O
(reactants fed at 298 K and 1 bar, products obtained at chemical equilibrium)
and LHV of product gas as a function of temperature.
0 x x y
f hCH xOy f hH02 O(gas) f
0
hCO 2 f hO02
2 2
2 (4.2)
LHVCH x O y M CH x O y
The large increase in hR and the LHV in the temperature range 800–1100 K is
mainly attributed to the gasification of char and the steam-reforming of
methane. At temperatures above 1350 K, the heating value of the product gas is
nearly constant since the equilibrium composition approaches complete
conversion. The reaction enthalpy increases solely due to the increase in the
product temperature. Any increase in temperature above 1350 K is therefore
worthwhile only if the reaction rates are accelerated and the thermal losses of
the process can be reduced due to shorter residence time of the reactants in a
reactor.
In autothermal gasification, a part of the feedstock , also called
equivalence ratio, is combusted (Eq. 4.3) to drive the endothermic gasification
(Eq. 4.4), thereby reducing the amount of H2O that is consumed in a
stoichiometric gasification.
x y x
CH x O y 1 O2 H 2 O CO 2 (4.3)
4 2 2
x
1 CH x O y 1 y H 2O 1 y H2 CO (4.4)
2
Varying the equivalence ratio results in different adiabatic temperatures and

equilibrium compositions as shown in Figure 4-4. For consistency with the first
and second law analyses (discussed in Section 4.2), steam and oxygen are
introduced at 1100 K and bagasse at 298 K.
The quality of syngas, characterized by the H2/CO and CO2/CO molar
ratios, is shown in Figure 4-5 for both the solar-driven and autothermal
gasification. For the autothermal gasification, significant amounts of CO2
originate from the additional oxygen needed for the internal combustion,
lowering the ratio of H2/CO and the calorific value of the syngas. In contrast,
the solar-driven gasification produces a syngas of higher quality in terms of
4.1 Equilibrium Considerations 29
1.0
0.3
equilibrium composition [mol]
0.8
CO
H2
0.6 0.2
[-]
0.4 H2O
CO2 0.1
0.2
C(s)
CH4
0.0 0.0
800 1000 1200 1400 1600
temperature [K]
Figure 4-4: Equilibrium composition as a function of adiabatic temperature

for the system in Eqs. 4.3 and 4.4 with 0 0.33.
3
solar
autothermal
2 H2/CO
molar ratio [-]
1
CO2/CO
0
800 1000 1200 1400 1600
temperature [K]
Figure 4-5: Molar ratios of H2/CO and CO2/CO at equilibrium over

temperature for solar and autothermal gasification.
higher H2/CO and lower CO2/CO ratios than those obtained for the autothermal
gasification. This reduces the need for water-gas shift reaction and the effort for
separating CO2 when producing Fischer-Tropsch fuels.
4.2 1st and 2nd Law Analyses
First and second law analyses were conducted to assess the performance of the
solar and the autothermal gasification of bagasse as well by comparing the
energy and exergy efficiencies and the major sources of irreversibility. The two
investigated pathways are shown in Figure 4-6. Each includes a gasifier, an
adiabatic heat exchanger, and a quencher.
concentrated
solar energy
Qsolar
Qrerad
bagasse autothermal
solar gasifier
gasifier
syngas H2O syngas

H2 O
H2O O2 H2O
heat exchanger H2O heat exchanger
syngas syngas
H2O O2 H2 O
quencher quencher
Qquench Qquench
H2O H2O
syngas syngas autothermal

solar
gasification gasfication
Figure 4-6: Flow diagram applied in the first and second law analyses for the
solar (left) and autothermal gasification (right).
4.2 1st and 2nd Law Analyses 31
The gasifiers are considered as fed with H2O-, ash-, N-, and S-free bagasse
(CH1.665O0.663) at an ambient temperature T0 = 298 K and a total pressure p0 =
1 bar at a biomass flow rate CHxOy = 1 g/s. Steam or a steam/oxygen mixture
enters the solar or the autothermal gasifier, respectively, and is preheated in the
heat exchanger to 100 K below Treactor by the product gases leaving the reactor
in thermodynamic equilibrium at Treactor. After passing the heat exchanger, the
product gases are quenched to T0, rejecting the remaining heat Qquench to the
environment. The thermal energy required for the endothermic solar
gasification is provided by an industrial solar concentrating system such as a
solar tower with a heliostat field. The solar reactor is of cavity-type
configuration. It is assumed to behave like a blackbody ( eff = eff = 1) with
adiabatic walls. Highly concentrated solar radiation Qsolar enters the cavity
through a small aperture to minimize reradiation losses Qrerad [60]. The
absorption efficiency of the reactor is defined as [60, 61]
Qreactor,net 4
Treactor
abs 1 (4.5)
Qsolar IC
where I denotes the DNI (direct normal insolation) and C is the solar
concentration ratio. The net energy absorbed by the reactor matches the
reaction enthalpy
Qreactor, net nCH x O y hR (4.6)
The exergy destruction and the exergy loss because of absorption and
reradiation [62] are given by
T0 T0
Exdestr,abs Qsolar 1 Qsolar 1 (4.7)
Tsun Treactor
T0
Exloss,rerad Qrerad 1 (4.8)
Treactor
whereas the heat loss due to reradiation is derived from Eq. 4.5 as
Qrerad 1 abs Qsolar (4.9)
The exergy destruction within the reactor, the heat exchanger, and the
quenching unit are given by
T0
Exdestr Qnet 1 ni ,in i n j ,out j (4.10)
Tb i j
where the base temperature Tb = Treactor for the reactor and Tb = T0 for the
quencher. For the adiabatic heat exchanger Qnet = 0. The molar exergy of the
flow is defined as the exergy of the ideal gas mixture [63], where i is the sum
of the thermo-mechanical and chemical exergy of species i neglecting kinetic
and potential effects [64].
yi i RT0 yi ln yi (4.11)
i i
i hi hi ,0 T0 ( si si ,0 ) ch,i (4.12)
Chemical exergy, enthalpy, and entropy values are taken from previous studies
[65-67]. The chemical exergy of the bagasse is determined by an empirical
correlation for solid C-H-O compounds [65].
ch, CH x O y LHVCH x O y (4.13)
H O H
1.0414 0.0177 0.3328
1 0.0537
C C C O
for 2 (4.14)
O C
1 0.4021
C
In the adiabatic autothermal gasifier the heat is supplied by the partial

combustion of the feedstock with pure oxygen as described in the previous
section. The oxygen is provided by a cryogenic air separation unit (ASU) at an
energy expense of 0.245 kWh/kgO [68, 69]. The determination of the energy
and the exergy losses for the autothermal gasifier follows the procedure
described for the solar gasifier.
The energy and exergy efficiencies of the gasification processes are defined
respectively as
mprod LHVprod
(4.15)
mCH x O y LHVCH x O y Qsolar WASU
ni ,prod i ,prod
i
ex (4.16)
T0
n j ,reac j ,reac Qsolar 1 WASU
j Treactor
where ASU = 0 for the solar process and Qsolar = 0 in the autothermal reactor.
The upgrade factor U is defined as the ratio of the energy content of the syngas
produced to that of the converted feedstock,
mprod LHVprod
U (4.17)
mCHx Oy LHVCHx O y
The performance of the solar gasifier is determined by the solar-to-chemical

efficiency, defined as
ni ,prod hi ,prod n j ,reac h j ,reac

i j
sol-chem (4.18)
Qsolar
Table 4-1 shows a summary of the numerical results of the energy and exergy
calculations based on the following baseline parameters: inlet biomass mass
flow CHxOy = 1 g/s, total operating pressure ptot = 1 bar, DNI I = 1 kW/m2, solar
concentration ratio C = 2000, and sun temperature Tsun = 5780 K. Further, the
gasifiers are operated at temperatures (Treactor) of 1350 and 1100 K for the solar
and the autothermal case, respectively, yielding equilibrium concentrations of
CH4 of less than 0.1%vol and high heating values of the product gas. Higher
temperatures are only favorable if kinetics are accelerated and thermal losses of
the reactor (not considered here) can be reduced as a result of the smaller
reactor volume needed to perform the gasification.
The energy and exergy efficiencies were found to be = 0.92 and ex = 0.80
for the solar, and = 0.93 and ex = 0.78 for the autothermal gasification. The
major energy loss occurs in the quencher for both pathways. The largest exergy
loss/destruction occurs in the gasifier and accounts for 48% of the total exergy
loss/destruction in the solar gasification and 73% of that in the autothermal
gasification. In the case of autothermal gasification, around 25% of the
feedstock ( = 0.253) have to be combusted to achieve the desired Treactor =
Table 4-1: Energy and exergy analysis for solar and autothermal gasification.
Values for power refer to a biomass feed rate of CHxOy = 1 g/s.
units solar gasification autothermal gasification
Qsolar [kW] 7.12 –
Qrerad [kW] 0.67 –
Qreactor, net [kW] 6.45 0
Qquench [kW] 1.78 1.74
ASU [kW] – 0.32
abs [–] 0.91 –
[–] 0.92 0.93
ex [–] 0.80 0.78
sol-chem [–] 0.66 –
U [–] 1.26 0.95
xdestr, absorption [kW] 1.21 –
xloss, rerad [kW] 0.52 –
xdestr, reactor [kW] 2.82 3.43
xdestr, HEX [kW] 0.42 0.32
xdestr, quench [kW] 0.85 0.98
LHV [MJ/kg] 19.38 13.91
LHV [MJ/Nm3] 15.68 10.73
H2 [%vol] 53.7 45.5
CO [%vol] 46.1 42.4
CO2 [%vol] 0.2 12.0
CH4 [%vol] 0.1 0.1
1100 K. This is the reason for the significantly higher exergy destruction in the
autothermal gasifier and the resulting lower syngas quality. The solar gasifier,
which requires around 7 kW of solar radiation (Qsolar = 7.12 kW) to gasify 1 g/s
of bagasse, loses more than 9% of incoming solar radiation through reradiation.
To some extent, reradiation losses can be minimized by incorporating
secondary non-imaging concentrators (CPC) [70], which increase the solar
concentration ratio and reduce the aperture size. The exergy destruction that
occurs when absorbing the solar radiation (Eq. 4.7) can be reduced by
increasing Treactor. Since full conversion is reached at Treactor = 1350 K
(Figure 4-1), this would not affect the syngas quality and only increase the
exergy losses elsewhere.
The syngas produced by the solar-driven process shows a more favorable
composition with a higher H2/CO ratio (1.16 vs. 1.07) and less contamination
with CO2 (0.2%vol vs. 12%vol). This is also reflected in the LHV’s of 19.38 vs.
13.91 MJ/kg and the upgrade factor U = 1.26 vs. 0.98. Thus, the feedstock is
solar-upgraded by 26%, chemically storing solar energy with sol-chem = 66%; in
the autothermal case, the energy content of the feedstock is higher than that of
the product gas.
This thermodynamic analysis of both routes is verified by performing an
energy balance and by evaluating the maximum achievable Carnot efficiency
from the total available exergy and from the total solar energy input. The
energy balance confirms that
Qsolar Qrerad Qquench n j ,prod h j ,prod ni ,reac hi ,reac (4.19)

j i
The available work is calculated as the sum of the available exergy and the
exergy destruction and losses due to irreversibilities in the solar reactor, the
heat exchanger, and the quenching unit. Thus,
n j ,prod j ,prod ni ,reac i ,reac Exdest Exloss

j i
max (4.20)
Qsolar
This maximum efficiency must be equal to that of a Carnot heat engine

operating between Tsun and T0, i.e.
T0
max Carnot 1 0.948 (4.21)
Tsun
4.3 Conclusions
The thermodynamic equilibrium computations for the solar-driven gasification

of a stoichiometric mixture of Brazilian sugarcane bagasse and steam showed
an almost full conversion to H2 and CO at 1 bar and 1350 K. Any further
temperature increase is therefore justified only if the reaction rates are
accelerated and the thermal losses of the process can be reduced due to shorter
residence times of the reactants in the gasifier.
The thermodynamic analysis further indicated a superior syngas quality for
the solar-driven over the autothermal gasification. The syngas produced by the
solar-driven process showed a more favorable composition with a higher molar
H2/CO ratio (1.16 vs. 1.07) and less contamination or dilution with CO2
(0.2%vol vs. 12%vol).
First and second law analyses of the process showed theoretical upgrade
factors of 1.26 vs. 0.95 and heating values of 19.4 vs. 13.9 MJ/kg for the solar
and the autothermal gasification routes, respectively. For the autothermal
gasification, around 25% of the feedstock needs to be combusted to provide the
reaction enthalpy for the endothermic gasification reactions, thereby reducing
the amount of H2O consumed by a stoichiometric gasification. In contrast, the
solar gasification process allows the utilization of all feedstock. Moreover, the
feedstock is upgraded by 26% while chemically storing the solar energy with a
solar to chemical efficiency of 66%.
The exergy efficiencies for the solar and the autothermal gasification are
with 80% and 78% in a similar range. The main difference in the distribution of
the exergy losses and destruction is the location where they occur. In the solar
4.3 Conclusions 37
case, 21% occur where the solar energy is absorbed and 48% in the gasifier. In
the autothermal case the gasifier contributes to 73% of the total exergy losses
and destruction. This is due to the high exergy destruction of the combustion
reaction.
Chapter 5
5 Gasification Kinetics of Bagassea
5.1 Thermogravimetric Analysis
A kinetic analysis of the steam gasification of pyrolyzed Brazilian sugarcane

bagasse particles was performed using a thermogravimetric balance (Netzsch
STA 409 CD). The thermogravimetric balance, schematically shown in
Figure 5-1, consists of an electric furnace in which a sample is placed on a
crucible. The crucible is mounted on a thermocouple that is connected to a
balance recording the mass change of the sample. A reactive gas mixture of
H2O and Ar enters the furnace from the bottom through an annulus where it is
heated to the furnace temperature. Afterwards, the reactive gas flows
downwards past the sample. A small purge flow of Ar not influencing the gas
composition at the sample enters the furnace through the balance. The reactive
gas mixture is provided by a steam generator unit (Bronkhorst Hitec CEM) that
is connected to the furnace via a heated transfer line. The flows of Ar are
controlled with flow controllers (Vögtlin Q-FLOW). The flow of H2O to the
steam generator is controlled by electronic flow controllers (Bronkhorst LIQUI-
FLOW).
Due to the strong influence of the pyrolysis conditions on the formation and
reactivity of the char, discussed in Section 2.1.1, it is essential that the char
used for investigating the kinetics of its gasification with steam is generated
a
Bioenergy, vol. 52, pp. 173-183, 2013.
40 Chapter 5: Gasification Kinetics of Bagasse
under heating rates and temperatures that resemble those in a gasifier. Rapid
pyrolysis of the biomass feedstock is preferred for a solar reactor because of the
high release of volatiles and the low amounts of char with a high reactivity that
are produced. The char samples used in the kinetic analysis were thus rapidly
pyrolyzed bagasse particles. The rapid pyrolysis was achieved by entraining
raw bagasse particles with Ar into an electrically heated tubular furnace that
was preheated to 1373 K. During this rapid pyrolysis step, about 79%wt of the
feedstock devolatilized yielding bagasse char with an ash content of 64%wt as
product. An elemental analysis of the bagasse char is presented in Table 5-1.
The kinetic analysis in the thermogravimetric balance was performed under
atmospheric pressure and isothermal conditions as done by previous
investigators [71, 72]. Bagasse char samples of about 10 mg were preheated to
the desired temperature in a non-reactive atmosphere at a heating rate of
30 K/min. The inert atmosphere was established by flowing Ar across the
insulation
reactive gas
heating element
sample
crucible
thermocouple
purge gas
balance
Figure 5-1: Thermogravimetric balance configuration.

5.1 Thermogravimetric Analysis 41
Table 5-1: Elemental composition of raw and rapidly pyrolyzed bagasse

particles; C, H, N determined with CHN-900, O with RO-478, and S with
CHNS-932 (all LECO Corporation, St. Joseph, MI).
bagasse char
raw bagasse
(rapidly pyrolyzed)
carbon (C) [%wt] 42.51 29.50
hydrogen (H) [%wt] 5.94 0.60
oxygen (O) [%wt] 37.54 5.63
nitrogen (N) [%wt] 0.41 0.17
sulfur (S) [%wt] 0.09 0.12
ash (by difference) [%wt] 13.5 64.0
sample at a flow rate of 0.1 LN/min.b After temperature equilibration, the

atmosphere was switched from Ar to a reactive H2O/Ar mixture while keeping
the total flow rate across the sample constant (0.1 LN/min). The mass loss due
to the reaction of the char with the steam was then measured by the balance. To
eliminate the effect of buoyancy, a blank run with no sample on the crucible
was done for each condition and the recorded mass was subtracted from the
sample run.
Based on the measured mass change during the gasification of the sample,
the conversion or reaction extent of the sample was defined as
m0 m(t )
X (t ) (5.1)
m0 mf
where m(t), m0 and mf are the instantaneous, initial, and final sample mass,
respectively.
b
LN designates normal liters at 273 K and 1 atm.
5.2 Rate Law
5.2.1 Reaction Mechanism
Two basic mechanisms, namely the oxygen exchange and the hydrogen
inhibition mechanism, have been proposed for the steam gasification of char or
carbon given by the overall net reaction [73, 74]
C + H2O CO + H 2 (5.2)
The proposed mechanisms involve several of the following elementary reaction

steps occurring at the char surface;
k1
C + H2O C O + H2 (R1)
k2
C O + H2 C + H2O (R2)
k3
C O CO + C (R3)
k4
C + H2 C H 2
(R4)
k5
C H 2
C + H2 (R5)
1 k6
C + H2 C H (R6)
2
k7
1
C H C + H2 (R7)
2
where C designates an active carbon site, while C O , C H2 , and C H are

carbon-oxygen or carbon-hydrogen complexes on the char surface,
respectively.
5.2 Rate Law 43
Oxygen Exchange Mechanism
The oxygen exchange mechanism involves the elementary reaction steps R1–
R3. After introducing CO and C as the mole fractions of the surface species
covering the effective char surface area,
C O
NC O N tot 1 C
1 NC N tot (5.3)
Assuming that the ratio of the total number of sites Ntot to the effective char
surface area S does not change with the progress of reaction, i.e. Ntot / S =
constant, and relating the reaction rate to the char conversion X with respect to
the initial number of moles of char NC,0 yields
1 dNC NC,0 dX
rC rCO k3 CO
(5.4)
S dt S dt
Assuming sorption equilibrium, i.e.
rC O k1 pH2O 1 CO
k2 pH2 CO
k3 CO
0 (5.5)
the net gasification rate may be expressed in dependence of the partial

pressures of hydrogen pH and steam pH O.
k1 pH2O
rC (5.6)
k1 k2
1 pH O pH
k3 2 k3 2
where ki are Arrhenius type rate laws of the form
ki k i ,0 exp E A RT (5.7)
During the thermogravimetric experiments, the gaseous products were

constantly swept away from the reaction site, justifying the simplification pH =
0. Thus, the gasification rate depends only on the steam partial pressure and
temperature as follows.
k1 pH 2 O
rC (5.8)
1 ( k1 / k3 ) pH 2 O
Hydrogen Inhibition Mechanism
The hydrogen inhibition mechanism consists of four steps that are either R1,
R3, R4, and R5, or R1, R3, R6, and R7. Similarly to the oxygen mechanism we
can derive the net gasification rates as
k1 pH2O
rC (5.9)
k1 k4
1 pH2O pH
k3 k5 2
or
k1 pH2O
rC (5.10)
k k6 0.5
1 1 pH2O pH
k3 k7 2
Again, the partial pressure of hydrogen can be set to pH = 0 as the product

gases are swept away in the TG experiments. The resulting reaction rate is
therefore for both cases
k1 p H 2 O
rC (5.11)
1 ( k1 / k 3 ) p H 2 O
This is identical to the net gasification rates derived for the oxygen exchange
mechanism (Eq. 5.8).
5.2.2 Surface Area Dependence
With Eq. 5.8 at hand, the only remaining information required to integrate
Eq. 5.4 is the dependence of the effective char surface area on the char
conversion S = S(X ). Kimura et al. [75] have pointed out that this relationship
depends on the size distribution of solid particles reacting according to the
shrinking core model with reaction control [76]. Furthermore, Jovanovic [77]
has demonstrated that for wide particle size distributions a linear fit, which was
also applied for the gasification of cellulose and lignin [17], may be a
reasonable approximation. In the absence of information relative to the
5.2 Rate Law 45
effective size distribution of char particles, the latter was assumed as the initial
guess, i.e.
S S0 (1 X ) (5.12)
Combining Eqs. 5.4, 5.8, and 5.12, the net gasification rate becomes
k1 pH 2 O N C,0 1 dX
rC (5.13)
1 ( k1 / k3 ) pH 2 O S0 1 X dt
Because of the presence of ash and unknown ash-carbon structure, the initial
effective surface area participating in the gasification per mole of char is
unknown. Therefore, S0, NC,0 and rC are lumped into apparent reaction rate rC
with the apparent rate constants k1 and k3 as follows
S0 k1' pH2O 1 dX
rC' rC (5.14)
NC,0 1 (k1' / k3' ) pH2O 1 X dt
where
ki' ki S 0 N C,0 (5.15)
Finally, for constant temperature and steam partial pressure, Eq. 5.14 can be
expressed in the integral form as
k1' pH2O ln 1 X
rC' ' '
(5.16)
1 (k / k ) pH2O
1 3 t
where ki are Arrhenius type rate constants with a pre-exponential factor k0,i and
an activation energy EA,i.
k i' k 0,' i exp '

E A, i RT (5.17)
5.2.3 Evaluation of the Rate Constants
The temperature dependence of the reaction rate was investigated within the
range of 1173–1473 K at a steam concentration fixed to 75%vol. The effect of
the steam concentration was explored at a temperature fixed at 1273 K while
varying the steam fraction in the reactive gas mixture between 25%vol and
75%vol.
Figure 5-3 presents a test of Eq. 5.16 against selected experimental results
acquired in a 75%vol H2O/Ar mixture at three different reaction temperatures.
The reasonable linearity allows for extracting the experimental reaction rates as
slopes of the fitted straight lines. These were then used to determine the
apparent reaction constants k0,i and EA,i. that are listed in Table 5-2 by least
square fitting. The parity plot shown in Figure 5-2 demonstrates the good
agreement between model and experimental results over the whole range
investigated.
Using the rate law, the residence time required for attaining 90% char
conversion with a steam concentration of 75%vol at temperatures of 1373 and
1573 K was calculated to be 43 and 16 s, respectively. Operating a solar reactor
at temperatures of 1373 K would allow the use of metal alloys as material of
construction. This is beneficial due to their better manufacturability, higher
thermal shock resistance, and longer lifetimes. Further it has to be kept in mind,
that the kinetic analysis was performed at pH = 0 thereby neglecting the
inhibitive effect of hydrogen. In an actual solar reactor the partial pressure of
hydrogen might not be pH = 0, thus slowing down the reaction. Woodruff et al.
[78] reported that a partial pressure of pH = 0.15 bar results in a reduction of
the reaction rate by about 30% and therefore an increase of the required
residence time by about 43%.
Table 5-2: Apparent kinetic parameters for Eq. 5.16 for the steam gasification
of rapidly pyrolyzed bagasse.
k1 [1/s-bar] k3 [1/s]
k0,i (ki) 55.8 4.27·1011
EA,i [kJ/mol] 73.7 321
5.2 Rate Law 47
2.5 1373 K 1273 K 1223 K
-ln(1-X ) [-] 2.0
1.5
1.0
0.5 experiment
linear fit
0.0
0 100 200 300
time [s]
Figure 5-3: Typical experimental results and linear fits for –ln(1–X ) vs. time
for selected experiments with a 75%vol steam/Ar mixture.
0.1
r'C experiment [1/s]
0.01
25%-vol
25%vol
50%-vol
50%vol
75%-vol
75%vol
0.001
0.001 0.01 0.1
r'C model [1/s]
Figure 5-2: Apparent carbon conversion rate for the experimental

measurements vs. the model for the temperature range 1173 1473 K and
25 75%vol steam concentration.
5.3 Conclusions
A thermogravimetric analysis of the steam-based gasification of bagasse char

was performed in the temperature range of 1173 1473 K and steam
concentrations between 25%vol and 75%vol. The kinetic rate law based on the
oxygen exchange mechanism in conjunction with a linear decrease of the
effective char surface area with conversion was found to fit the experimental
data well.
The required residence times for 90% char conversion at a steam
concentration of 75%vol and a hydrogen concentration of 0%vol at temperatures
of 1373 and 1573 K were predicted to be 43 and 16 s, respectively. In an actual
solar gasifier, the presence of hydrogen might lead to a moderate increase the
required residence time. Therefore, while operating a reactor at 1373 K would
expand the option for the material of construction to metal alloys that have
much higher shock resistance and better cycle lifetimes than ceramic materials,
it will require a reactor concept that provides a residence time in the order of a
minute.
Chapter 6
6 Drop-Tube Fixed-Bed Solar Gasifier Concepta
This chapter presents a solar reactor concept that aims at providing pyrolysis
and gasification conditions for high carbon conversion of the biomass feedstock
into a syngas with low amounts of tar and gaseous hydrocarbons. The proposed
laboratory-scale gasifier provides sufficient residence time and temperature for
the char conversion and the decomposition of hydrocarbons by combining
drop-tube and fixed-bed concepts. Experimental testing was performed with
Brazilian sugarcane bagasse particles in an electrically heated furnace with the
final aim to supply heat by concentrated solar radiation.
6.1 Gasifier Concept
As discussed in Section 2.1, the heating rate, gas temperature, and residence
time have a strong influence on the release of tar and gases, as well as on the
formation and reactivity of the char. Moreover, the kinetic investigation of the
feedstock under consideration indicated that a residence time as high as 30–60 s
is required for 90% char conversion. The most preferable gasifier concept
would thus comprise two zones: a rapid, high-temperature pyrolysis zone
yielding low amounts of tar and highly reactive char [17-20] followed by a
a
Bioenergy, vol. 52, pp. 173-183, 2013.
50 Chapter 6: Drop-Tube Fixed-Bed Solar Gasifier Concept
zone providing sufficient residence time and temperature for the slow char
gasification and the decomposition of hydrocarbons.
In the solar reactor concept proposed here, the two zones are realized by
incorporating a grate into a drop-tube reactor creating a fixed-bed that increases
the residence time of the solids. Such a combined design retains the advantage
of the efficient radiative heat transfer to the particles inherent to drop-tube
reactors that is needed for fast pyrolysis, while, however, overcoming the drop-
tube’s residence time and particle size limitations that constrain cracking and
reforming of hydrocarbons. Similarly to downdraft gasifiers that yield low
amounts of tar, the pyrolysis products pass through hottest zone of the reactor
where they are decomposed.
6.2 Gasifier Testing in an Electric Furnace
6.2.1 Experimental Setup and Procedures
Figure 6-1 shows a schematic of the laboratory-scale reactor used for the
experimentation including the primary components and flows. The reactor was
assembled from a heat-resistant alumina tube (1200 mm long having an inner
diameter of 60 mm and a wall thickness of 5 mm) placed inside an electrical
tube furnace (Carbolite) which simulates the conditions of an absorbing cavity-
receiver that is heated by concentrated solar radiation. The tube was equipped
with a reticulated porous ceramic (RPC) foam (10 ppi, thickness 10 mm,
Erbisic, Erbicol S.A.) with a centered hole (diameter 10 mm) serving as a grate
at the center of the hot zone.
Dried and sieved bagasse particles (described in Chapter 3) were fed from
an Ar-purged hopper positioned above the tube via a calibrated screw feeder
and mixed at the top of the tube with N2-entrained steam generated with an
external evaporator (Bronkhorst). The flow rates of the inlet gases and water
into the evaporator were controlled with electronic mass flow controllers
(Bronkhorst). The product gas stream was cooled and filtered to remove
condensable components and particulate matter and then purified from tar using
dichloromethane (DMC) as solvent [79]. The resulting gas composition was
6.1 Gasifier Concept 51
bagasse
evaporator feeder
Ar & mixer
M
N2
electric
furnace
water tank
vent condenser
GC steam trap
filter DCM solvent
Figure 6-1: Schematic of the laboratory-scale biomass gasification apparatus

including the primary components and flows.
analyzed by gas chromatography (GC) using a two-channel Varian Micro-GC,

equipped with Molsieve-5A and Poraplot-U columns (1/120 Hz sampling
frequency) capable of determining concentrations of dry H2, N2, CO, CH4, CO2,
acetylene (C2H2), ethylene (C2H4), and ethane (C2H6). The nitrogen entraining
the steam from the evaporator was used as tracer gas to calculate the total molar
flow rate at the outlet of the reactor.
Prior to each experiment, the reactor was first purged with Ar. After
reaching a negligible O2 concentration in the system, the hopper purge Ar and
N2 flow rates were set to 0.5 and 0.1 LN/minb, respectively. The reactor was
b
then preheated to a desired set-point temperature in the range 1073 1573 K.

After equilibration, a steady flow of 17 g/h of steam was established, resulting
in an overall steam concentration of around 37%vol. At this point, the biomass
feed commenced at an average rate of 2.8 g/s-m2 or 0.48 g/min, leading to a
molar steam to biomass ratios ( H O(g)/ CHxOy) of about 0.94, corresponding to
2.8 times the stoichiometric amount of steam for the idealized net reaction
represented by:
x
CH x O y 1 y H 2O 1 y H2 CO (6.1)
2
6.2.2 Results
The experimental values for composition and the corresponding heating values
(LHV), carbon conversions (XC), and upgrade factors (U ) are reported in
Table 6-1. The tabulated mole fractions were calculated considering only molar
flow rates of the product gases integrated over 30 min; the flow rates of H2O,
N2, and Ar were not considered. The mole fractions were then used to calculate
the mass fractions wk = ykMk / (ykMk) and the reported LHV values as follows:
LHV wk LHVk (6.2)
The carbon conversion XC is defined as the amount of carbon evolved with

monitored product gases divided by the amount of carbon fed with bagasse.
X C mC,product mC,reactant (6.3)
For the interpretation of the experimental results, the definition of the upgrade
factor U given in Eq. 4.17 is extended to account for the unconverted feedstock.
mprod LHVprod
U (6.4)
mCH x O y LHVCH x O y mC,in (1 X C )LHVchar
The unconverted feedstock (char) was assumed to be pure carbon with an LHV
of 33.5 MJ/kg [80].
Table 6-1: Summary of the experimental results obtained for the gasification
of bagasse in the electrically heated two-zone reactor
T [K] 1073 1173 1273 1373 1473 1573
yH [%vol] 30 34.7 40.6 47.4 51.2 54.5
yCO [%vol] 34.3 32.0 33.5 30.7 33.7 34.0
yCH [%vol] 13.4 12.8 9.8 6.6 3.0 0.8
yCO [%vol] 16.7 17.6 15.4 15.1 11.9 10.6
yC H [%vol] 5.0 2.3 0.4 0.1 0.1 0.0
yC H [%vol] 0.3 0.1 0.0 0.0 0.0 0.0
yC H [%vol] 0.3 0.5 0.2 0.1 0.0 0.0
yH /yCO [-] 0.88 1.08 1.21 1.55 1.52 1.6
yCO/yCO [-] 2.05 1.82 2.17 2.03 2.83 3.2
LHV [MJ/kg] 16.93 15.95 15.34 15.29 15.55 15.69
LHV [MJ/m3] 16.09 14.44 12.84 11.81 11.27 10.83
XC [-] 0.65 0.61 0.67 0.76 0.82 0.84
U [-] 0.95 0.94 0.96 1.03 1.08 1.12
Figure 6-2 shows the experimentally measured mole fractions of the product
gases as well as the equilibrium composition as a function of temperature. The
production of H2 gradually increased with temperature approaching the
concentration predicted by equilibrium (Figure 4-1). The CO mole fraction
remained relatively constant over the whole temperature range investigated. At
1073 K it was higher than that predicted by equilibrium. For all other
experimental conditions, CO levels were over-predicted. The measured CO2
concentrations decreased with an increase in temperature but were significantly
higher than those predicted by equilibrium. Although the presence of CH4 is not
thermodynamically favored at above 1200 K, it was still observed (~1%vol) at
temperatures as high as 1573 K. C2-gases, especially ethylene (C2H4) were
detected in small amounts (> 0.1%vol) up to temperatures of about 1273 K.
Increased temperatures yielded H2/CO ratios of up to 1.60 and CO2/CO ratios
as low as 0.31.
0.6
0.5 H2
0.4
mole fraction [-]
0.3
CO
0.2
0.1 CO2
CH4
0.0
1000 1200 1400 1600
temperature [K]
Figure 6-2: Relative mole fractions (dry, N2 and Ar free basis) of the product
gases and the equilibrium composition (lines without markers)
The carbon conversion XC increased with temperature from 65% to 84%.

The reasons for the generally low carbon conversions may be partially
attributed to the formation of tar which was observed mainly in the experiments
at 1073 and 1173 K. Further, inadequate solids retention performance of the
ceramic grate led to particles penetrating through the grate and ending up at the
bottom of the reactor. Poor heat transfer to the packed bed of char and slow
reaction kinetics at 1073 and 1173 K led also to an accumulation of partially
reacted particle on the RPC. This could also be observed by a non-steady
product gas composition. In the experiments from 1273 to 1573 K a steady
product gas composition was observed.
All these findings imply that neither gas nor solids spent sufficient time at
temperatures required for target conversion which is to be addressed by an
improved version of this hybrid reactor concept, the solids retention grate in
particular.
Upgrade factors were found to increase with temperature and a maximum of
112% could be achieved. This increase is partly due to the increased carbon
conversion and partly due to the different composition. The heating values of
the syngas (LHV) were 15.3 16.9 MJ/kg or 10.8–16.1 MJ/Nm3. Due to the
decrease of CH4 and C2-gases with temperature, the heating value per volume
was significantly reduced. The heating values found are significantly higher
values than those generally obtained in conventional autothermal gasifiers. For
example, the fluidized-bed air gasifier by Gómez et al. [7] delivers syngas with
3.3 5.1 MJ/Nm3 at a cold gas efficiency of 29.2%, whereas the air blown
cyclone gasifier by Gabra et al. [8] produces product gas with 2.8 4.5 MJ/Nm3.
Ash melting was observed at above 1473 K, which is consistent with ash
fusion tests (ASTM, oxidizing atmosphere) of Indian [81] and Hawaiian
bagasse [54] where initial deformation temperatures of 1473 and 1510 K were
observed. For continuous operation it is thus necessary to stay below the ash
melting temperature of about 1473 K.
6.3 Conclusions
An allothermal gasifier configuration based on a combination of drop-tube and

fixed-bed concepts was proposed. The two-zone concept aims to provide
pyrolysis and gasification conditions yielding high carbon conversion into
syngas and low amounts of tar and gaseous hydrocarbons. In the upper drop-
tube zone, high radiative heat flux to the dispersed particles induces their fast
pyrolysis. In the lower zone, a fixed bed provides sufficient residence time and
temperature for the char gasification and the decomposition of the
hydrocarbons released during the pyrolysis.
A lab-scale prototype reactor was tested with bagasse particles at a biomass
feed rate of 2.8 g/s-m2 in the temperature range of 1073 1573 K. The reactor
was exposed to infrared radiation from an electric furnace simulating the
conditions present in an absorbing solar cavity receiver. The concentrations of
the gaseous species were approaching the thermodynamic equilibrium as the
reactor temperature was increased, i.e. H2 concentrations went up whereas
amounts of CO2, CH4, and C2-hydrocarbons decreased. Although the presence
of CH4 is not thermodynamically favored at above 1200 K, it was observed in
significant amounts over the whole temperature range investigated. The syngas
produced had molar H2/CO ratios of up to 1.6 and CO2/CO ratios as low as
0.31. The lower heating values were from 15.3 to 16.9 MJ/kg. The carbon
conversion of these preliminary experiments stayed behind the expectations,
which implied that further development in terms of particle retention and heat
transfer was necessary. Further, the observed ash melting occurring at
temperatures above around 1473 K gave an upper limit for the operating
temperature of this concept.
However, it could be confirmed that an upgrade factor of greater than 1 is
achievable and that syngas yields per unit feedstock and heating values are
significantly higher than those typically obtained in conventional gasifiers,
supporting the potential benefits of solar-driven gasification over conventional
autothermal gasification.
Chapter 7
7 Drop-Tube Trickle-Bed Solar-Driven Gasifiera
The externally heated gasifier concept presented in the previous chapter was
further improved to deliver higher carbon conversion and a better
decomposition of the gaseous hydrocarbons. The improved gasifier concept
comprises a drop-tube zone for fast pyrolysis and a trickle bed for the rate
limiting char gasification and the decomposition of the pyrolysis products. The
trickle bed utilizes a structured packing to control the overall porosity of the
gasification zone in order to increase the residence time of the char particles
while still allowing the radiation to penetrate through. The structure packing
thus enhances the heat transfer to both the particle and the gas phase. The drop-
tube trickle-bed concept was tested in a solar reactor that was designed and
built for operation at ETH’s high flux solar simulator. Its performance was
experimentally assessed with Brazilian sugarcane bagasse particles and
compared to the performance of the drop-tube configuration.
7.1 Gasifier Concept
The solar-driven gasifier developed is shown in Figure 7-1. It is based on a

vertical tubular reactor situated within a cavity-receiver in order to minimize
reradiation losses and provide a homogeneous temperature distribution [3]. The
tubular reactor comprises two zones (see inlay Figure 7-1): an upper drop-tube
a
and A. Steinfeld, "A two-zone solar-driven gasifier concept: Reactor design and
experimental evaluation with bagasse particles", Fuel, vol. 117, Part A, pp.
680-687, 2014.
58 Chapter 7: Drop-Tube Trickle-Bed Solar-Driven Gasifier
!#"%
+
!#*!'
!!"##"$ !&'(#')
Figure 7-1: Schematic of the solar cavity-receiver/reactor configuration with

thermocouple locations and blown up detail of the reactor tube showing the fast
pyrolysis drop-tube zone and the trickle-bed char gasification zone (RPC =
reticulate porous ceramic, CPC = compound parabolic concentrator).
pyrolysis zone and a lower trickle-bed char gasification zone consisting of a

structure packing. Bagasse particles and steam are both introduced from the
top. The raw bagasse particles are rapidly heated in the upper zone by infrared
radiation from the tube wall to undergo fast pyrolysis. This zone provides
sufficient residence time to ensure that the particles reaching the trickle bed are
not sticky and prone to clogging the structure packing.
The structure packing, depicted in Figure 7-2, is a reticulate porous ceramic
(RPC) foam. The pyrolyzed particles trickle through the RPC and undergo
gasification with concurrently flowing steam. In comparison to the commonly
used packings, such as spheres, Raschig rings, Pall rings, cylindrical screens, or
regularly stacked packings [82-87], the RPC has a higher degree of solid
connectivity and therefore higher effective thermal conductivity at the same
porosity. This is of a great importance as the heat is provided externally.
Figure 7-2: Structure packing made of 10 ppi SiSiC reticulate porous

ceramic (RPC) foam installed in the solar gasifier. Image from [88].
Moreover, the structure is less optically dense than a packed or a moving bed
hence the radiation penetrates deeper. Therefore, by combining conductive and
radiative heat-transfer modes, the structure is expected to enhance heat transfer
to both gas [89] and solid phases. Finally, by providing a resistance to the flow
of solids the structure not only increases the mean residence time of the
trickling particles but it also aids their radial dispersion. All these
enhancements are expected to provide a more homogeneous radial temperature
distribution and decrease the temperature difference between the gas and the
solids, thereby increasing reaction yields.
By transporting gas and solids downwards as done in downdraft gasifiers,
the tar and gases evolved during the pyrolysis pass through the hottest zone of
the reactor where they decompose via cracking and reforming reactions to H2,
CO, CO2, lighter hydrocarbons, and coke.
Together with the potential advantages mentioned above, the concept
proposed here introduces some limitations that need to be recognized. Although
high temperatures are desirable for efficient heat transfer to gas and solids
resulting in high reaction rates, the operating temperature of the trickle zone is
limited by the ash melting temperature of the feedstock. In addition, the

throughput of the feedstock depends on how well the particle loading and
residence time within the structured packing can be controlled.
7.2 Gasifier Testing on a High Flux Solar Simulator
7.2.1 Experimental Setup
The solar reactor (Figure 7-1) was made of a heat-resistant, well conducting
(30 W/m-K at 1500 K) pressureless-sintered silicon carbide tube (Hexoloy SE
SiC, Saint Gobain, L 700 mm, ID 41 mm, OD 51 mm) which was placed in a
cavity (200 × 86 × 86 mm) made of a 60 mm thick (40 mm at the front)
alumina/silica based insulation (65% Al2O3, 34% SiO2, Insulform 1600). The
insulation was fastened by a stainless steel case surrounding it. The cavity has a
30 mm-diameter aperture for the access of concentrated solar radiation. The
reactor tube was placed slightly towards the back of the cavity to reduce
thermal stress on the tube and minimize reradiation losses due to hotspots [46].
A 3-dimensional, water-cooled compound parabolic concentrator (CPC) [90]
was mounted as a secondary concentrator at the reactor aperture to boost the
concentration ratio, thereby allowing a smaller aperture size and thus further
reducing reradiation losses. The CPC was made of polished aluminum and
designed for a half acceptance angle of 45° with an exit diameter of 30 mm. It
was truncated to a height of 32.2 mm resulting in an inlet diameter of 42.3 mm
and a concentration ratio close to 2. To prevent overheating of the outer surface
of the assembly by spilled radiation, a water-cooled shield (300 × 300 mm) was
mounted around the CPC. The system was designed for 1.5 kWth solar radiative
input power and operation at ambient pressure and temperatures up to 1850 K.
Figure 7-3 presents an overview of the experimental setup including the
solar gasifier and auxiliary components. Bagasse particles were introduced by
an Ar-purged drum feeder positioned above the rector tube. N2 and steam
generated with an external evaporator (Bronkhorst) were injected through
annularly distributed inlets positioned just below the feeder. The flow rates of
the inlet gases and the water into the evaporator were controlled with electronic
7.2 Gasifier Testing on a High Flux Solar Simulator 61
bagasse
feeder
evaporator
Ar & mixer Xe short-arc
M lamps
N2
solar
reactor simulated solar
radiation
water tank
product
collection drum
to vent
condenser
to chiller elliptical
and GC steam trap reflectors
filter
Figure 7-3: Schematic of the solar-driven biomass gasifier, including the

high-flux solar simulator and the auxiliary components and flows.
mass flow controllers (Bronkhorst). The porous structure creating the trickle
bed is shown in Figure 7-2. It was made of a 10 ppi (pores per inch) SiSiC
(silicon infiltrated silicon carbide) reticulate porous ceramic foam (RPC,
porosity > 87%), which was placed in the hot zone of the reactor tube. The
RPC was supported by an alumina tube (Alsint 99.7, Haldenwanger, inner
diameter 30 mm, outer diameter 38 mm). The temperatures of the cavity and of
the reactor tube were measured with K-type thermocouples placed at multiple
axial locations inside the assembly (Tinlet, Ttop, Ttube, Tbottom, Toutlet). One
thermocouple, Tcenter, was inserted into the RPC with its tip at the centerline,
20 mm above the bottom of the RPC.
Ash and unreacted char were collected in the product collection drum
located below the reactor just before a condenser and a steam trap. A slip-
stream of the product gas was withdrawn after the condenser and analyzed by
gas chromatography (GC) after being filtered and chilled to remove particulate
matter and condensable components. The two-channel Varian Micro-GC
equipped with a Molsieve-5A and a Poraplot-U column (1/120 Hz sampling
frequency) was calibrated to determine the concentrations of H2, N2, CO, CH4,
CO2, C2H2, C2H4, and C2H6. A known flow rate of N2 introduced with the
steam was used as tracer gas to calculate the total molar flow rate of the product
gas.
The experiments were carried out at the high-flux solar simulator (HFSS) of
ETH Zürich. The HFSS is equipped with seven 6 kWel high-pressure Xe arcs
close-coupled to truncated elliptical specular reflectors [91]. It is capable of
delivering continuous thermal radiative power with a peak flux of up to
4500 kW/m2 and a mean flux of 3620 kW/m2 on a 30 mm aperture diameter.
Thus, the solar reactor was tested under comparable heat-transfer
characteristics of highly concentrating solar systems, such as solar dishes and
solar towers. Radiative flux intensities were adjustable by the number of Xe
arcs in operation, their power, and the position of the reactor aperture relative to
the focal plane. The radiative power input at the reactor aperture (Qsolar) was
determined optically with a calibrated CCD camera and verified by calorimetric
measurements at the CPC outlet.
7.2.2 Experimental Procedure
The power input and the temperature traces recorded during a typical
experiment are shown in Figure 7-4. At the beginning of all the experiments but
one, the nominal Ar and N2 flow rates were set to 0.5 LN/min: in the reference
experiment with pure pyrolysis (no steam injection) the nominal flow rates of
Ar and N2 were set to 0.5 LN/minb and 2.2 LN/min, respectively. In order to
achieve rapid heating, up to 5 arcs of the HFSS were then turned on
simultaneously to irradiate the reactor at high power levels (1.65 kWth,
2333 kW/m2). After approximately 30 min, the input power was reduced to the
levels ranging between 1.147 and 1.195 kWth (1622–1690 kW/m2) to
b
7.2 Gasifier Testing on a High Flux Solar Simulator 63
2.0
Tcenter 1250
input power [kW] 1.5 Ttop

1000
temperature [K]
1.0
Tbottom 750
0.5 500
0.0 250
-80 -60 -40 -20 0 20 40
time [min]
Figure 7-4: Input power (Qsolar) and reactor temperatures during a typical
experiment: steam on, biomass feed on, and biomass feed off.
equilibrate the reactor in a steady-state with the resulting temperature inside the
RPC (Tcenter) settling within 1256–1362 K range. The maximum temperature at
the top of the RPC at the tube (Ttube) was kept below 1428 K to avoid ash
slagging that has been observed at temperatures above 1473 K [54, 81, 92].
After the thermal equilibration of the system, steam preheated to about 560 K
was injected at a steady rate of 81.6 g/h resulting in a nominal inlet steam
concentration of 62.9%vol. The injection of steam was reflected by a
temperature drop in the upper part of the reactor (Ttop) and an increase towards
the bottom (Tbottom). After reaching another steady state, the biomass feed
commenced at feeding rates between 61 and 94 g/h leading to molar steam
to biomass ratios ( H O(g)/ CHxOy) of 1.37–2.09. This corresponds to 2.06–3.15
times the stoichiometric amount of steam for the idealized net reaction
represented by:
x
CH x O y 1 y H2O 1 y H2 CO (7.1)
2
After another 20 min, biomass feed, steam flow, and irradiation were all turned
off and the reactor was cooled down . Temporal gas composition,
temperature, and pressure data were acquired during the course of the
experiment. In addition, char samples were collected after the experiments and
analyzed for their elemental composition (CHN-900, LECO Corporation, St.
Joseph, MI).
7.3 Results and Discussion
Figure 7-5 shows the molar flow rates of the product gases as a function of time
during a typical experimental run. A slight increase in H2 and CO can be
observed immediately after the commencement of steam This has been
attributed to the gasification of the carbon residue in the RPC originating from
previous experiments. As the step changes in product flow rates could be
correlated with turning the biomass feed on and off , the observed
0.5
H2
0.4
molar flow rate [mmol/s]
0.3
CO
0.2
CO2
0.1
CH4
0
-20 -10 0 10 20 30
time [min]
Figure 7-5: Molar flow rates of product gases as a function of time during a
typical experimental run: steam on, biomass feed on, and biomass
feed off.
7.3 Results and Discussion 65
fluctuations in the data shown in Figure 7-5 reflect the intermittent feed rate. In
addition, the data indicated an increase in the production of H2 and CO2 and a
decrease in the production of CO, CH4 and C2H4 with time. This could be
explained by two factors: (1) the gradual increase in the temperature at the
bottom of the cavity as indicated by Figure 7-4 and (2) a suspected buildup of
char within the RPC. As a result of the char buildup, more syngas was
produced within the RPC that then underwent the water-gas shift and
hydrocarbon cracking and reforming reactions favored by a higher temperature
at the bottom of the cavity.
Experiments were carried out with different system configurations as
summarized in Table 7-1. Two-zone experimental sets A1-A3 and B1-B3 were
performed with RPCs that were 50 and 100 mm tall, respectively. As reference
runs simulating the drop-tube concept alone, free-fall pure pyrolysis (PP) and
free-fall steam-based gasification (SG) experiments were performed without
Table 7-1: Summary of the experimental results (A = 50 mm tall RPC, B =

100 mm tall RPC, PP = free-fall pure pyrolysis, SG = free-fall steam-based
gasification, all values are on a H2O, N2 and Ar free basis).
A1 A2 A3 B1 B2 B3 SG PP
Tcenter [K] 1344 1359 1362 1303 1284 1285 1343 1338
Ttop [K] 1098 1105 1108 1045 1027 1028 1080 1110
Tbottom [K] 1024 1037 1040 980 965 980 1068 1053
yH [%vol] 45.8 46.0 46.5 45.0 43.7 45.0 37.2 35.1
yCO [%vol] 33.8 34.7 34.1 34.8 34.8 34.4 39.9 46.5
yCH [%vol] 5.9 5.7 5.4 6.2 6.6 6.3 8.7 7.0
yCO [%vol] 13.2 12.5 13.0 12.8 13.2 13.2 10.9 6.2
yC H [%vol] 1.0 0.8 0.8 0.8 1.1 0.6 2.0 3.7
yC H [%vol] 0.4 0.3 0.3 0.4 0.6 0.5 1.1 1.5
yC H[%vol] 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
LHV [MJ/kg] 15.9 15.9 15.7 15.9 15.9 15.8 17.0 18.5
U [-] 1.06 1.04 1.06 1.05 1.04 1.05 0.99 1.00
XC [-] 0.89 0.80 0.90 0.89 0.90 0.88 0.69 0.64
any RPC in the system. Furthermore, the experiments were performed at two
levels of Tcenter: ~1350 K for set A, SG, and PP, and ~1290 K for set B. The
observed temperature variations within experimental sets A and B are the result
of the physical limitations to strictly control the power input to the solar
reactor.
The tabulated mole fractions were calculated considering only molar flow
rates of the product gases integrated over the course of an experiment; the flow
rates of H2O, N2, and Ar were not considered. LHV, upgrade factor (U), and
carbon conversion (XC) were calculated as defined in Eqs. 6.2–6.4.
The carbon mass balance showed that between 89% and 98% of the total
carbon fed into the system could be accounted for by the gas phase evolved
over the course of an experiment and the solids recovered from the RPC and
the product collection drum. The remainder was attributed to deposits on the
tube wall, carryover of fine particles or tar, and the overall measurement error.
The fraction of the carbon retained within the RPCs was less than 2.3% and
5.3% for the configurations A and B, respectively.
The statistical significance of the differences between the responses of the
investigated configurations has been assessed as follows:
(a) The responses of configurations A and B were compared using two-

sample t-tests at a significance level of 5% (MATLAB, The
MathWorks, Inc.). The variances of the unpaired samples were
considered to be equal as confirmed by two-sided F-tests at a 95%
confidence level.
(b) The single responses of configurations SG and PP were compared to

the responses of configurations A and B by single-sided Grubb’s
outlier tests [93] at a 5% confidence level. These tests were used to
decide with 95% confidence if the responses of configurations SG and
PP do not belong to the same normal distributions as the
corresponding responses of sets A1–A3 and B1–B3.
The results of the statistical analysis are summarized in Table 7-2. For a
comparison X vs. Y, “>” or “<” designate statistically significant higher or
lower response value for configuration X; “~” designates no statistically

significant difference.
Figure 7-6 presents a comparison of the gas compositions measured with
the different system configurations. The average gas mole fractions calculated
for the experiments A1–A3 and B1–B3 correspond to H2/CO ratios of 1.35 and
1.29, respectively, and the same CO2/CO ratios of 0.38. Although the
temperatures (Tcenter, Ttop, Tbottom) in set B were on average 59–70 K lower than
in set A, according to Table 7-2 only the mole fractions of H2 and CH4 were
found to be statistically different.
As shown in Table 7-2, the comparison of the two-zone configurations A
and B to the free-fall pure pyrolysis (PP) indicates a significant decrease in CO,
C2H2, and C2H4 as well as an increase in H2 and CO2. The comparison to the
free-fall steam-based gasification experiment (SG) indicated the following:
The amounts of CO and CH4 obtained with either of the two-zone

configurations (A, B) were lower and the amount of H2 higher.
The amounts of C2H2 and C2H4 obtained with configuration A were

lower but those obtained with configuration B showed no significant
difference.
The amounts of CO2 obtained with configuration B were higher but

the ones obtained with configuration A showed no significant
difference.
The above observations imply that the two-zone reactor concept allows for
more efficient decomposition of CH4 and C2-hydrocarbons compared to the
drop-tube gasifier operating at comparable reactor tube temperatures. At
present, it is not clear whether this should be attributed to a higher gas
temperature, an increased residence time of the gas, or both. In addition, it
appears that the two-zone reactor facilitates the water-gas shift reaction (CO +
H2O H2 + CO2) which may be attributed to the higher gas temperatures as no
effort has been made to quench the product gas. The higher extent of the
exothermic water-gas shift reaction with increased gas temperatures is to be
attributed merely to faster kinetics as the measured concentrations of H2 and
a)
H2 CO CO2 CH4
50%
40%
30%
20%
10%
0%
PP SG A B
b) 4%
C2H4 C2H4 C2H6
3%
2%
1%
0%
PP SG A B
Figure 7-6: Average gas concentrations for a) H2, CO, CO2, and CH4, and
b) C2H2, C2H4, and C2H6 calculated on H2O, N2, and Ar free basis for the
investigated configurations: A = 50 mm tall RPC, B = 100 mm tall RPC, PP =
free-fall pure pyrolysis, and SG = free-fall steam-based gasification. Error bars
indicate the span between the minimum and maximum values.
CO2 were still below the values expected from thermodynamic equilibrium
[59].
The LHVs of the syngas were 15.9 MJ/kg (12.5 MJ/Nm3) for configurations
A and B, and 18.5 and 17.0 MJ/kg (15.6 and 14.4 MJ/Nm3) for PP and SG,
respectively. The statistically higher values for PP and SG originate from the
higher content of light hydrocarbons. For all cases, however, the heating values
achieved in this work were substantially higher than those obtained in
conventional autothermal bagasse gasification. For example, the air-blown
cyclone particle-gasifier by Gabra et al. [8] and the air-blown downdraft
gasifier for bagasse pellets by Erlich and Fransson [94] produced syngas with
an LHV of 4.8–8.1 MJ/Nm3 and 9.5 MJ/Nm3, respectively. For comparison
reasons, all LHVs are reported on a H2O and N2 free basis. The higher values
obtained in the solar gasification were mainly due to the lower content of CO2.
Table 7-2: Summary of the statistical analysis for the effects of reactor
configurations: for comparison X vs. Y “>” or “<” designate statistically higher
or lower mean of a response for the configuration X; “~” designates no
statistically significant difference.
response A vs. B* A vs. SG† B vs. SG† A vs. PP† B vs. PP†
yH > > > > >
yCO ~ < < < <
yCH < < < ~ ~
yCO ~ ~ > > >
yC H ~ < ~ < <
yC H ~ < ~ < <
yC H - - - - -
LHV ~ < < < <
U ~ ~ > ~ >
XC ~ ~ > ~ >
*
two-sample t-tests at a significance level of 5% with equal variances of the
unpaired samples
†
single-sided Grubb’s outlier tests [40] at a 5% confidence level
Carbon conversions XC and upgrade factors U are presented in Figure 7-7.

The carbon conversions for configurations A and B were calculated to be
around 87% and 89%, respectively. For B, this is a statistically significant
improvement over both PP (64%) and SG (69%), in spite of the considerably
lower reactor temperatures. Remarkably, the improvement for configuration A
could not be shown to be significant by the Grubbs test due to the scatter in the
data and a small sample size. The reasons for the generally low carbon
conversion may be attributed to insufficient solids retention or particle build up
on top of the RPC that was observed in some experiments.
The upgrade factor calculated was found to be ~1.05 for cases A and B, and
less than 1 for PP and SG. This means that the biomass could be energetically
upgraded by 5% effectively storing solar energy in the syngas.
Figure 7-8 shows the RPC after one of the A experiments. The easily
removable char particles were quite uniformly distributed throughout the entire
structure, thereby implying rather volumetric char gasification. It was possible
to feed up to 77 g/h bagasse without a significant increase in the pressure drop
1 1.06
carbon conversion
upgrade factor
0.8 1.04
carbon conversion [-]
upgrade factor [-]
0.6 1.02
0.4 1.00
0.2 0.98
0 0.96
PP SG A B
Figure 7-7: Carbon conversion and upgrade factors for investigated

configurations: A = 50 mm tall RPC, B = 100 mm tall RPC, PP = free-fall pure
pyrolysis, SG = free-fall steam gasification. Error bars illustrate the span
between the minimum and maximum values.
over the RPC. This corresponds to a particle flux per cross-sectional area of
16 g/s-m2. At higher feeding rates char started to deposit on top of the RPC.
The elemental analysis of the solid residue showed lower carbon conversion for
particles collected in the product collection drum than for the particles retrieved
from the RPC. This implies that the retention of the particles by the RPC and/or
their interaction with the RPC surface is the key for increasing the overall
carbon conversion.
The energy conversion efficiency is defined as
msyngas LHVsyngas
tend
(7.2)
Qsolar dt mbagasse LHVbagasse mC,in (1 X C )LHVchar
t0
strongly depends on the feeding rate. For reporting the maximum it is thus
important to consider only the experiments in which no particle build-up was
Figure 7-8: The 50 mm tall RPC after one of the A experiments. The
particles are uniformly distributed throughout the entire RPC.
observed. Under this constraint, the highest efficiency of = 21.6% was

achieved at a feeding rate of 73 g/h (experiment A3). To overcome the
throughput constraints and increase it is necessary to develop a method for
on-line ash removal from the RPC and improve the control of the solids
residence time. Since the particles deposited on the RPC were easy to remove,
vibration and/or occasional pulse-flow of purge gas (inert or syngas) could
serve as means for achieving both of these tasks that should be addressed by
future research.
7.4 Energy Balance and Heat Losses
To gain an insight to the energy flows in the solar reactor and to understand
where energetic improvements to the reactor design are possible, an energy
balance across the reactor has been applied. The energy flows considered
Tinlet
Q̇ cond,top
Ttop
Tcav
Toutside
Q̇ rad
Q̇ walls
Q̇ conv
Ttube
Tbottom Q̇ cond,bottom
T∞
Toutlet
Figure 7-9: Overview temperature measurements and heat losses: radiation

and convection through the aperture Qrad and Qconv, conduction through the
cavity walls Qwalls and along the reactor tube towards the top Qcond,top and the
bottom Qcond,bottom.
7.4 Energy Balance and Heat Losses 73
include the heat transferred to the reactants Qnet as well as the major heat losses,
namely radiation and convection through the aperture of the cavity Qrad and
Qconv, conduction through the insulation at the cavity walls Qwalls, and
conduction along the reactor tube towards the top Qcond,top and the bottom
Qcond,bottom. An overview of the location of the heat losses and the temperature
measurements that were used for their determination is given in Figure 7-9.
7.4.1 Heat Transfer to Reactants
The net heat transfer to the reactants Qnet was determined by an energy balance
across the reactor tube.
Qnet H out H in (7.3)
where in and out are the enthalpy flows entering the reactor tube at the top
and leaving it at the bottom. Steam and N2 enter the reactor at the measured Tin,
Ar and bagasse enter at ambient temperature T = 298 K. The product gases
leave the reactor at Tout, which was measured at the bottom seal. The unreacted
char was assumed to be pure carbon and to leave the reactor at T .
H in nH 2 O hH 2 O (Tsteam ) nN 2 hN 2 (Tsteam ) nAr hAr (T0 ) nCH x O y hCH x O y (7.4)
H out ni ,out hi (Tout ) nC hC (7.5)

i
where i = N2, Ar, H2O, H2, CO, CO2, CH4, C2H2, C2H4, and C2H6. The enthalpy
of species i is defined as
T
hi (T ) h f0,i C p ,i dT (7.6)
T0
The enthalpies of the bagasse hCHxOy and the char hC are based on their LHV and
composition.
x
hCH x O y hH O(g) hCO 2 LHVCH x O y M CH x O y (7.7)
2 2
hC hCO 2 LHVchar M char (7.8)
7.4.2 Radiation Losses
The radiation losses through the aperture were computed using the net-radiation
method for enclosures [95] with the following assumptions: (a) all surfaces are
opaque-gray-diffuse, (b) uniform temperature, flux, and optical properties over
each element, and (c) non-participating medium/gas in the cavity space. The net
radiative flux qj leaving surface j was computed by solving the following
system of equations
N
ij 1 j
N
Fi j qj ij Fi j T j4 (7.9)
j 1 j j j 1
for the surfaces i = 1, 2, …, N. ij is the Kronecker delta defined as:
1 when i j
ij
(7.10)
0 when i j
The cavity was treated as an enclosure of three discrete internal surfaces (N =

3), cavity wall, reactor tube and aperture. The view factors Fi-j were calculated
by Monte-Carlo ray tracing of 109 rays using the in-house code Vegas [96].
Emissivities of = 1 for the aperture, = 0.86 for the SiC tube [97], and =
0.58 for the cavity insulation [98] were applied. The temperature of the aperture
and the reactor tube were Tapt = 0 K and the measured Ttube, respectively. The
cavity temperature Tcav,rad was defined as the quartic mean of all n measured
cavity temperatures Tcav,i.
n
1
Tcav,rad 4 Tca4v,i (7.11)
n i 1
7.4.3 Convective Losses
The convective heat losses from the cavity through the aperture were estimated
by solving the steady state mass, momentum, and thermal energy conservation
equations using the commercial computational fluid dynamics (CFD) tool
Ansys CFX [99]. The domain was divided into: 1) a cavity, 2) a cone
representing the CPC, and 3) a rectangular volume extending outwards 150 mm
from the shield representing the surroundings. No-slip conditions and uniform
temperatures were applied to the wall boundaries: Tcav (arithmetic average of all
Tcav,i) for the cavity and T0 = 298 K for the cone and the shield, which were
both water cooled. The remaining boundaries on the rectangular volume were
modeled as openings with the fluid temperature T0 and a pressure p0 = 1 atm.
The buoyancy driven flow was assumed to be laminar. This is appropriate
along the cavity walls because [100]
Tcav L3
Gr g 1 2
109 (7.12)
T0
where L is the distance from the top edge of aperture to the bottom of cavity
[101] and is the kinematic viscosity evaluated at (Tcav–T0)/2. Furthermore, it is
appropriate through the aperture because
Lu
Re 103 (7.13)
where L = dapt/2, and are the fluid density and the dynamic viscosity
evaluated at T0, and u is the maximum computed velocity. The fluid (air) was
modeled as ideal gas neglecting viscous dissipation. The values of thermal
conductivity, dynamic viscosity, and specific heat were taken from correlations
[102]. The governing equations were solved on a mesh that was refined around
the aperture and the top of the CPC inlet. A grid refinement study was
performed and it was found that the solution is grid-independent for a mesh
with ~2.2 million tetrahedron cells.
Figure 7-10 shows a contour plot of the temperature distribution in the
domain for a selected Tcav. The fluid temperature is almost uniform within the
Figure 7-10: Convection losses through aperture: Air temperature distribution

in cavity, aperture, and surrounding.
cavity except below the aperture. Cold air from the outside of the cavity flows
through the aperture towards the bottom of the reactor. The cold air is then
heated and flows out of the reactor at the top of the aperture rising towards the
top after leaving the cavity.
The convective heat loss computed for cavity temperatures Tcav between
500 and 1500 K, and was found to be well represented by the following
correlation.
2
5 2
Qconv 1.13 10 Tcav T0 6.19 10 Tcav T0 2.66 [W] (7.14)
7.4.4 Conduction through the Cavity Walls
For the estimate of the conduction losses through the cavity walls it was
assumed that the inside and outside surfaces of the cavity are at a uniform
temperature allowing the use of shape factors for the estimate of the heat
transfer.
Qwalls keff,ins S Tcav Toutside (7.15)
where keff,ins is the effective thermal conductivity of the insulation, S the total
shape factor and Toutside is the temperature measured at the outer wall of the
cavity. The total shape factor S is the sum of all individual shape factors Si. Swall
= A/D, Sedge = 0.54L, Scorner = 0.15D with the surface area of wall A, the wall
thickness D, and the length of edge L [103]. The effective thermal conductivity
of the insulation is computed as
2
kins T dT
k eff,ins 1
(7.16)
T2 T1
7.4.5 Conduction along Reactor Tube
The conduction losses along the reactor tube (Qtube) were estimated by
numerically solving the steady heat conduction equation with a 1D finite
volume model. Convective and radiative heat fluxes (Qconv,i and Qrad,i) were
determined based on the surface temperature of each cell (Tsurf,i).
4
Qrad,i P x SiC Tsurf,i (7.17)
Qconv,i P x hsf (Tsurf,i T0 ) (7.18)
where P is the perimeter of the tube, x is the element length, and is the
Stefan-Boltzmann constant. SiC = 0.86 is the emissivity of the SiC tube [97], hsf
the convective heat transfer coefficient, and Tsurf,i the surface temperature of the
element i. For the non-insulated parts of the reactor tube the surface
temperature is equal to the cell temperature (Tsurf,i = Ti). For the insulated parts,
the surface temperature was determined by setting the heat flux through the
insulation to
2 keff,ins (Ti Tsurf,i ) x

Qins,i Qrad,i Qconv,i (7.19)
ln( d ins / d tube )
The temperature dependent thermal conductivity of the insulation keff,ins was

evaluated at each finite volume. The convective heat transfer coefficient hsf was
determined by a Nusselt correlation for vertical cylinders with diameter d and
length L [104].
1/ 6
2
L
Nu 0.825 0.387 Ra f1 (Pr) 0.435 (7.20)
d
where
16/9
9/16
0.492
f1 (Pr) 1 (7.21)
Pr
Pr is the Prandtl number, and Ra the Rayleigh number.

At both ends of each domain, fixed temperature boundary conditions were
applied representing the temperatures measured at the carbon seals and where
the reactor tube passes through the cavity walls.
7.4.6 Interpretation
The resulting heat losses as well as the heat transferred to the reactants are
illustrated in Figure 7-11 and summarized in Table 7-3. The power input
calculated from the sum of all losses and the heat transfer to the reactants was
within 10% of the experimentally measured power input. It was found that the
radiative and convective heat losses through the aperture are in the order of Qrad
= 126 W and Qconv = 47 W, accounting for only 10% and 4% of the total
calculated power input. The mayor heat losses occur by conduction. The
Qwalls = 362 W, 29 %
Qrad = 126 W, 10 %
conduction walls
Qcond,bottom = 374 W, 31 %
Qconv = 47 W, 4 %
conduction tube bottom

Qcond,top = 199 W, 16 %
conduction tube top
radiation
convection
power input
Qin = 1233 W
to reactants
Qnet = 125 W, 10 %
Figure 7-11: Energy losses of solar reactor via radiation and convection
through the aperture Qrad and Qconv, conduction through the walls Qwalls and
along the tube Qcond,top and Qcond,bottom.
conduction losses through the cavity walls and along the reactor tube are 29%
and 47% of the input power (Qwalls = 362 W and Qtube = Qcond,top + Qcond,bottom =
573 W). The heat transferred to the reactants is around Qnet = 125 W or 10% of
the input power.
The energy flows calculated suggest first of all better insulation to reduce
conduction losses. The losses through the walls can easily be reduced by
applying thicker insulation. A reduction of the conduction losses along the tube
imposes a more difficult task as the preferably well conducting reactor tube
needs to be sealed at colder temperatures. More importantly they suggest
improving the heat utilization of the process by increasing the throughput and
heat transfer to the reactants. Such an improvement has a much bigger effect on
the overall performance of the reactor.
Table 7-3: Energy flows in the solar reactor.

radiation through aperture, Qrad 126 W 10.2%
convection through aperture, Qconv 47 W 3.8%
conduction
- through walls, Qwalls 362 W 29.3%
- along tube (top / bottom), Qtube 573 W 46.6% (16.2% / 30.4%)
heat to reactants, Qnet 125 W 10.2%
calculated power input, Qin 1233 W
(sum of the above)
measured power input 1174 W
7.5 Conclusions
A two-zone solar-driven biomass gasifier concept has been proposed and

evaluated at a 1.5 kWth solar radiative power input with bagasse particles. Its
first zone operates in a drop-tube mode which ensures efficient radiative heat
transfer to dispersed biomass particles that is required for fast pyrolysis. Its
second zone is designed to operate as a trickle bed consisting of a structured
packing that increases solids holdup in the hot zone, thereby providing
residence time and temperature needed for the rate limiting char gasification
and the decomposition of the pyrolysis products.
A series of 20 min gasification experiments with bagasse particles and
steam were performed on a high-flux solar simulator to compare the drop-tube
trickle-bed reactor with a drop-tube reactor. It was demonstrated that the two-
zone reactor concept decomposes CH4 and C2-hydrocarbons more efficiently
than the drop-tube gasifier operating at comparable reactor tube temperatures.
The concentrations of H2 and CO2 were higher than in the drop tube reactor,
whereas the concentrations of CO were lower, which has been attributed to a
higher extent of the water-gas shift reaction due to a higher gas temperature.
The LHV of the syngas produced by the two-zone reactor was around
15.9 MJ/kg, which is substantially higher than those typically obtained in
7.5 Conclusions 81
conventional autothermal bagasse gasification [8, 94]. The biomass was

energetically upgraded by 5%, effectively storing solar energy in the syngas.
The maximum energy conversion efficiency was 21%.
An energy balance applied to the solar reactor indicated the highest losses
are due conduction through the cavity walls and along the reactor tube
accounting for 29% and 47% of the total input power. As only 10% of the input
power was transferred to the reactants the largest efficiency gains can be
achieved by increasing the throughput and improving the heat transfer to the
reactants.
It is important to note that the concept has only been evaluated under the
very limited conditions of both short-term operation (20 min) and light particle
loadings (a maximum particle flux of up to 16 g/s-m2 was achieved without
particle build-up on the structure packing). Future research should address on-
line ash removal from the structure packing to increase the throughput and
ensure longer term operability.
Chapter 8
8 Heat- and Mass-Transfer Analysis of a Trickle-Bed

Gasifiera
The previously introduced drop-tube trickle-bed solar reactor concept that aims
at providing high heat transfer rates to the gas and solid phase showed
favorable results in terms of carbon conversion and selectivity towards H2 and
CO in comparison to a pure drop-tube design. To understand the mass and heat-
transfer phenomena leading to the more favorable performance a numerical
heat transfer model of an allothermal tubular reactor gasifying beech char
particles with steam was developed. The model considers steady-state, two-
dimensional finite-volume material and energy balances and accounts for all
three modes of heat transfer: conduction and radiation from the wall through
the RPC and convection from the RPC surface to the flowing gas. The
contribution of the chemical reaction was evaluated based on an apparent first-
order rate law after adopting a suitable value of the activation energy from the
literature and optimizing the pre-exponential factor for the best agreement
between the model predictions and experimental results obtained in a
laboratory-scale electrically heated reactor prototype. The model was then used
to assess the sensitivity of the gasification rate and the reactor temperatures to
the particle loading and the RPC properties such as pore diameter, porosity, and
thermal conductivity. In addition, the performance of the trickle-bed reactor
was numerically compared to the performance of a moving bed as an
alternative concept for increasing the residence time of the char particles.
a
Material of this chapter is published in: M. Kruesi, Z. R. Jovanovic, A.
Haselbacher, and A. Steinfeld, "Analysis of Solar-Driven Gasification of
Biochar Trickling through an Interconnected Porous Structure", submitted
84 Chapter 8: Heat- and Mass-Transfer Analysis of a Trickle-Bed Gasifier
8.1 Experimental Investigation
8.1.1 Experimental Setup
A schematic of the experimental ambient-pressure concurrent gas-solid trickle-

bed reactor setup is shown in Figure 8-1. A 1.2 m long alumina reactor tube
having an inner diameter of 0.06 m and a wall thickness of 5 mm was situated
inside an electrical tube furnace (Carbolite) to simulate the conditions of
absorbing cavity-receivers that are heated by concentrated solar radiation [41,
105]. Two 0.1 m tall RPC pieces, as the one shown in Figure 8-2 , were stacked
and inserted into the alumina tube with the stack’s top at the middle of the hot
zone and held in place by another smaller diameter supporting alumina tube.
The RPC was made of SiSiC (silicon-infiltrated silicon carbide) with a high
thermal conductivity (kSiC = 36 W/m-K at 1273 K) and an overall porosity of
particle
feeder
Ar
porous
steam structure
generator (RPC)
N2
electric
furnace
water tank particle

separator
vent
condenser
GC
steam trap
filter
Figure 8-1: Schematic of the experimental setup.

8.1 Experimental Investigation 85
RPC = 89% with a nominal pore diameter of dnom = 2.54 mm corresponding to

10 ppi (pores per inch).
Temperatures were measured with seven K-type thermocouples positioned
as illustrated in Figure 8-3: four along the centerline of the RPC (C1–C4) and
five along at the tube/RPC interface (W1–W5). Two movable thermocouples
(W6, W7) were used to measure the wall temperature above and below the
RPC. The z-coordinates of the thermocouples are listed in Table 8-1.
The particles were fed via a calibrated feeding system (LAMBDA
Laboratory Instruments) from an Ar-purged hopper positioned above the tube.
A sweep Ar flow was introduced through a side port below the feeding system
to prevent the steam from back-flowing and condensing in the hopper. The N2-
entrained steam generated with an external evaporator (Bronkhorst) was
introduced at the top of the alumina tube. The flow rates of the gases and water
0.1 m
Figure 8-2: Photograph of the SiSiC reticulate porous ceramic (RPC) with 10
pores per inch (ppi) used as packing in the tubular gasifier.
inlet
TW6
r TW1
TC1
TW2
TC2
z TW3
TC3 RPC Tfurnace
TW4
TC4 TW5
TW7
centerline
reactor tube
outlet supporting tube
Figure 8-3: Schematic (not to scale) of the main reactor tube, RPC, and
supporting tube assembly indicating thermocouple locations. Thermocouples
TW6 and TW7 were movable along the tube wall.
Table 8-1: Thermocouple positions: W = wall, C = centerline.

thermocouple z-coordinate [m]
W1 0.00
W2 0.02
W3 0.05
W4 0.12
W5 0.18
W6 -0.60–0.00
W7 0.20–0.60
C1 0.02
C2 0.05
C3 0.10
C4 0.18
into the evaporator were controlled with Bronkhorst mass flow controllers pre-
calibrated by Bios Definer 220.
Ash and unreacted char were removed from the gaseous product stream in
the particle separator located below the reactor just before a condenser and a
steam trap. A slip-stream of the product gas was withdrawn after the condenser
and analyzed by gas chromatography (GC) after being filtered and chilled to
remove particulate matter and condensable components. The two-channel
Varian Micro-GC was equipped with Molsieve-5A and Poraplot-U columns
(~1/120 Hz sampling frequency) and was capable of determining
concentrations of H2, N2, CO, CH4, CO2, C2H2, C2H4, and C2H6. The flow rate
of N2 used to entrain the steam was used as reference to calculate the total
molar flow rate at the outlet of the reactor.
8.1.2 Experimental Procedure
Experiments were completed to test four sets of input conditions: a steam

concentration of 15%vol at nominal furnace temperatures Tfurnace = 1173, 1223,
and 1273 K and a steam concentration of 20%vol at Tfurnace = 1223 K. Each
experimental condition was tested twice to ensure repeatability. Additionally, a
free-fall drop-tube reference experiment with no RPC in place was completed
with an inlet steam concentration of 15%vol at Tfurnace = 1273 K.
After being purged with Ar and leak-tested, the reactor was preheated to the
desired furnace temperature. The flows of Ar, N2, and steam were set to
provide a total inlet gas flow rate of 12.12 LN/minb (9 mmol/s) to the reactor.
After the temperature equilibrated, the biomass feed commenced at a nominal
rate of 1.10 g/min corresponding to a mass flux of 6.5 g/s-m2. The resulting
molar steam to biomass ratios ( H O(g)/ CHxOy) were 1.03 or 1.36, corresponding
to 1.11 or 1.47 times the stoichiometric amount of steam for the idealized net
reaction represented by
b
x
CH x O y 1 – y H2O 1 – y H2 CO (8.1)
2
Temporal gas compositions, temperature, and pressure data were recorded over
the course of each experiment (43–51 min). The reactor was operated without
any interruptions and without clogging of the RPC or increase in pressure drop
across the reactor. Ash and char deposited on the RPC were easily removed
after the experiments.
8.1.3 Experimental Results
The temporal flow rates of H2, CO, CO2, and CH4 over the duration of a typical
experiment are shown in Figure 8-4. The concentrations of C2H2, C2H4, and
C2H6 were below the detection limit of the GC. After a rapid increase following
the introduction of the particles into the hot zone , the product flow rates
slowly approached asymptotic values towards the end of the experiment .
0.8
H2
0.6
0.4
CO
0.2 CO2
CH4
0
0 10 20 30 40 50
time [min]
Figure 8-4: Temporal product gas flow rates over the course of a typical
experiment.
The product flow rates were averaged over the final 10 min of each experiment
to obtain representative steady-state test conditions.
The carbon conversion (XC) was defined as the average molar flow rate of
carbon leaving the reactor with the monitored product gases ( C,syngas) divided
by the average molar flow rate of carbon introduced into the reactor with the
char particles ( C,char), i.e.,
X C nC,syngas nC,char (8.2)
As shown by Figure 8-5, the carbon conversion in any of the experiments with
the RPC was significantly higher than in the reference free-fall drop-tube
experiment where the conversion was negligible (less than 1%). Furthermore,
the carbon conversion responded well to the furnace temperature: increasing
the furnace temperature from 1173 to 1273 K resulted in an increase in carbon
conversion from 32% to 52%.
100%
RPC / 15%vol
RPC / 15%vol
80% RPC / 15%vol
RPC / 20%vol
no RPC / 15%vol
60%
XC [-]
40%
20%
0%
1150 1200 1250 1300
Tfurnace [K]
Figure 8-5: Carbon conversion (XC) as function of furnace set-point

temperature (Tfurnace), steam concentration, and gasifier configuration (RPC
versus no RPC).
8.2 Numerical Model
To understand the origin of the experimentally observed enhancing effect of the

trickle-bed zone on char gasification, a steady-state two-dimensional numerical
model of the reactor was developed. The model provides information on the
spatial variations of temperature, steam concentration, and reaction rate within
the trickle bed as well as on the relative contribution of the heat transfer by
conduction, convection, and radiation. The model is based on a two-
dimensional axisymmetric finite-volume representation of a domain
encompassing the RPC and its pores that are considered to be filled with gas
and particles. A two-dimensional representation is imposed by the fluid flow
and heat flux entering the domain in different directions: fluid flow axially
from the top and the heat flux radially from the tube wall. The domain is treated
as a porous medium with separate solid and fluid phases. The char particles are
assigned to the solid phase and gasified according to the overall stoichiometry
given by Eq. 8.1. The mass balance for the solid phase was not solved
explicitly. Instead, a constant mass of char per unit volume was assigned to the
whole domain. With this simplification, the thermal inertia of the particles
amounting to about 10% of its gasification enthalpy is neglected.
8.2.1 Governing Equations
The mass balance of the fluid phase is
1
r ur uz Sk (8.3)
r r z k
where uz and ur are the superficial velocities in the axial and radial directions,
and is the density of the gas mixture. The mass source terms Sk account for
the mass of each species k generated by the chemical reaction.
The fluid-phase conservation equation for species k = 1, 2, …, N–1 are
given by
8.2 Numerical Model 91
1
r ur vrC wk uz vzC wk
r r z
(8.4)
1 M k yk M k yk
r Dk , Dk , Sk
r r M r z M z
where wk and yk are the mass and mole fractions of species k, Deff,k, and Deff,k,
are the effective diffusion coefficients within the porous media perpendicular
and parallel to the main flow direction, and vrC and vzC are correction velocities
to ensure global mass conservation [106].
M k yk
vrC Deff,k , (8.5)
k M r
M k yk
vzC Deff,k , (8.6)
k M z
The fluid-phase energy equation based on the sensible internal energy of the
fluid es can be written as
1
r ur es u z es
r r z
1 M k yk
r hs,k Deff,k , (8.7)
r r k M r
M k yk
hs,k Deff,k , hsf A0 Ts Tf
z k M z
where Tf and Ts are the temperatures of the fluid and solid phases, respectively,
hs,k is the sensible enthalpy, Mk is the molar mass of species k, and M is the
average molar mass of the fluid. The heat transfer between fluid and solid is
expressed by an interfacial heat transfer coefficient hsf with A0 being the
interfacial area between the two phases. Finally, as all the heat consumed by the
chemical reaction Qreact is withdrawn from the solid phase, there is no source
term contributing to the fluid-phase energy balance. The terms for fluid-phase
heat conduction, viscous dissipation, and pressure work were neglected because
ks kf, M2/Re 1, and M2 1, respectively. (M and Re were estimated for
N2 at Tf = 1173 K, uz = 0.31 m/s, = 1.33, dtube = 0.06 m, and = 4.7·10-5 Pa-s.)
The energy balance for the solid phase is expressed as
1 T Ts
0 rkeff s keff hsf A0 Tf Ts Qreact (8.8)
r r r z z
where keff is the effective thermal conductivity. The radiative heat transfer
within the RPC, an optically thick medium ( 1), is modeled with the
Rosseland diffusion approximation [95]. An additional term accounting for the
radiative contribution krad is thus included in the definition of the effective
thermal conductivity,
keff kcond krad (8.9)
where kcond is the thermal conductivity through the RPC. The heat sink due to
the heterogeneous chemical reaction (Eq. 8.1) is given by
xC,char char,bulk l
Qreact hR rC (8.10)
MC
where hR is the heat of reaction, xC,char the carbon mass fraction of the char,
char,bulk the bulk density of the char, l the particle loading , and rC the reaction
rate. The heat of reaction was computed based on the LHV and heat capacity of
the char and the enthalpy of the gaseous species. The particle loading was
defined as
mchar,domain
l (8.11)
V
char,bulk domain
where mchar,domain is the mass of char in the domain and Vdomain is the volume of
the domain. The reaction rate was represented by a first-order Arrhenius-type
rate law given by
1 dNC EA
rC k0 exp pH2 O (8.12)
NC dt RT
where rC is the molar gasification rate of carbon per total number of moles of
carbon present (NC), EA and k0 are the apparent activation energy and frequency
factor, respectively, R is the gas constant, and pH O is the partial pressure of

steam. This rate law originates from the standard definition of the gasification
rate per surface area of char particles
1 dNC EA
rC" k0" exp pH2 O (8.13)
SC dt RT
and the assumption that the carbon surface area SC decreases linearly with the
conversion [92],
SC SC,0 1 X C (8.14)
Thus, recalling that
NC N C,0 1 X C (8.15)
we have
SC SC,0
rC rC" rC" (8.16)
NC N C, 0
and
SC,0
k0 k0" (8.17)
N C,0
The apparent activation energy was set to EA = 196 kJ/mol, a typical value for
wood chars reported in the literature [107]. The pre-exponential factor k0,
however, varies strongly with the feedstock source, size, and morphology. For
the purpose of this study, k0 was set to 3.3·106 s-1bar-1, which provided good
agreement between the simulated and measured overall gasification rates
( C,syngas) and RPC temperatures (Ts).
The mass source terms Sk applied in Eqs. 8.3 and 8.4 are given by
xC,char char,bulk l
Sk k Mk rC (8.18)
MC
with
1 y for H 2 O
1 x 2 y for H 2
k 1 for CO (8.19)
0 for N 2
0 for Ar
The bulk density of the beech char was char,bulk = 280 kg/m3 [31].
The momentum equations were not solved. Instead, a profile for the axial
velocity was prescribed. Because the boundary layer of the flow through a
porous medium is rather thin, a uniform axial velocity at any cross-section
perpendicular to the axis, i.e., uz(r,z) = uz(z), was considered to be a reasonable
assumption. At the operating pressure of 1 bar, the pressure drop across the
RPC is small and was thus neglected.
8.2.2 Boundary Conditions
On the tube wall, the temperatures of fluid Tf and solid Ts were set by linear
interpolation of the measured wall temperatures and the species mass fluxes
were set to zero. At the centerline, symmetry boundary conditions were applied
for all the conservation equations. The inlet conditions for the fluid were set by
the total mass flow rate f,in, the fluid temperature Tf,in, and the inlet mass
fractions of the species wk,in. At the outlet, i.e., at the bottom of the RPC, the
diffusive fluxes of mass and energy were set to zero. The net heat fluxes across
the inlet and outlet boundaries of the solid phase were computed using the
radiosity method for enclosures [95] to account for the radiative heat exchange
of the RPC with the reactor tube and tube ends. As shown in Figure 8-6, the
tube and RPC surfaces above and below the RPC were each discretized into
100 segments and 64 rings, respectively; the tube ends were considered as
single surfaces. Assuming a non-participating medium in the void space,
opaque-gray-diffuse surfaces, and uniform temperature, flux, and optical
properties over each surface element, the radiative fluxes qi between the Nsurf =
165 surface elements were determined from the following system of equations
N surf
ij 1 j
N surf
Fi j qj ij Fi j T j4 i = 1, 2, …, Nsurf (8.20)
j 1 j j j 1
where kj is the Kronecker delta, j is the emissivity of surface j, is the Stefan-

Boltzmann constant, and Fi–j are the configuration factors that were determined
analytically [108, 109]. The temperatures of the tube elements were based on
linear interpolation of the experimentally measured temperatures. The
inlet
Tinlet
Tw-top, N
Tw-top, j
tube wall
Tw-top,1
Ts-top,1
Ts-top, j
Ts-top, N
RPC
Ts-bottom,1 Ts-bottom, j
Ts-bottom, N
Tw-bottom,1
Tw-bottom, j
tube wall
Tw-bottom, N
Toutlet
outlet
Figure 8-6: Discretization for the radiative exchange between the RPC’s top
and bottom surfaces to their surroundings. Tube sections above and below
RPC: 100 segments each, RPC surfaces: 64 rings, flanges: individual surfaces.
The smaller tube diameter below the RPC is due to the supporting tube.
emissivities of the top and bottom surfaces of the RPC were set to RPC,eff = 0.82
so that the radiative flux
4
qi ,RPC eff, RPC Ti ,RPC (8.21)
corresponds to the emission of an isotropic scattering participating media with

emissivity SiC = 0.85 [97] at a uniform temperature Ti,RPC [110]. The emissivity
of the alumina wall was set to Al O = 0.28 [97].
8.2.3 Domain Properties
The effective transport properties of the RPC such as thermal conductivity,

interfacial heat transfer coefficient, and extinction coefficient have been studied
with pore-level Monte-Carlo ray-tracing and finite-volume heat-transfer
simulations based on tomographic scans [111, 112]. An overview of the
correlations that were used in the simulations is given in Table 8-2. The
Table 8-2: Heat- and mass-transfer property correlations.

quantity symbol source
RPC specific surface A0 Suter and Haussener [111], Eq. 14
RPC tortuosity Suter and Haussener [111], Table I
RPC interfacial heat transfer coeff. hsf Suter and Haussener [111], Eq. 17
binary diffusion coefficients ij Fuller et al. [113], Eq. 4
diffusion coefficients Dk Fairbanks and Wilke [114], Eq. 3
dispersion coefficients DL, DT Petrasch [112], Eqs. C.3 and C.4
extinction coefficient K Petrasch [112], Eq. 4.24
solid thermal conductivity SiC ks Munro [115], Eq. 14a
gas viscosity Yaws [116]
gas-phase thermal conductivity kf Yaws [116]
gas-phase sensible enthalpy hs Chase [117]
gas isobaric heat capacity Cp Chase [117]
heat capacity of the char Cp,char Merrick [118], Eq. 8
radiative contribution to the effective thermal conductivity keff (Eq. 8.9) was
defined as
16
krad Ts3 (8.22)
3K
where the extinction coefficient was defined as [112]
5.5173
K (1 rad ) (8.23)
d nom
with a nominal pore diameter of dnom = 2.54 mm for the 10 ppi RPC. The char
particles deposited on the struts of the RPC were accounted for by adjusting the
RPC porosity ( RPC) by the particle loading
rad RPC l (8.24)
The conductive contribution to keff was based on a combination of parallel and

serial slabs [112]. In the present case where ks kf the correlation reduces to
kcond ks fopt 1 RPC (8.25)
where fopt = 0.3823.

The effective diffusion coefficients Deff,k, and Deff,k, take into account the
molecular diffusion through a porous medium and the dispersion induced by
the flow through the porous structure,
2
Deff,k , Dm,k eff / Dd, (8.26)
2
Deff,k , Dm,k eff / Dd, (8.27)
where Dm,k is the molecular diffusion coefficient of species k, is the tortuosity,

and Dd, and Dd, are the dispersion coefficients perpendicular and parallel to
the main flow direction, respectively.
8.2.4 Numerical Implementation
The governing equations were iteratively solved with a cell-centered finite-

volume method. The RPC section of the tube, represented as a cylinder having
diameter dtube = 0.06 m and height htube = 0.2 m was discretized into Nz by Nr
cells in the axial and radial directions, respectively. The discretization in the
axial direction was uniform with cell heights z. In the radial direction, the
mesh was refined towards the wall to resolve the large temperature and
concentration gradients. The radial cell sizes were given by
ri 1.216 Nr ri 1 (8.28)
The mass balance equation for the fluid phase (Eq. 8.3) was used to compute
the velocity field. After integrating Eq. 8.3 over each control volume, the
resulting equations were summed in the radial direction, thereby eliminating the
radial velocity component. In combination with the axial velocity at the domain
boundary and the prescribed velocity profile, this allowed sequentially solving
for the axial velocity leaving the control volumes at each axial coordinate. Once
the axial velocities were known for the whole domain, the integrated mass
balance equations could be sequentially solved for the radial velocity in all
control volumes.
The governing equations for the N–1 species, the energy of the fluid phase,
and the solid energy were discretized using the hybrid scheme [119] and
written into individual linear systems. The solutions of the sparse linear
systems were found using MATLAB, either by permutation and forward
solving or by the unsymmetric-pattern multifrontal method [120].
At a given iteration step, the velocity field is computed and the species and
the fluid and solid phase energy equations are solved sequentially, updating the
material properties after each equation. The iteration was terminated if the
convergence criterion for the residuals of Ri < 10-7 was reached for all
dependent variables (ur, uz, Ts, wk, es). The residuals Ri were defined as the root
mean square error over all cells of the i-th iteration normalized by the first
iteration. A detailed description of the numerical implementation can be found
in Appendix C.
8.2.5 Code Verification
An order-verification study was performed using the Method of Manufactured

Solutions (MMS) [121] to ensure that the implemented numerical method
solves the governing equations correctly [122]. In the MMS, a solution is first
assumed (or “manufactured”) and inserted into the governing equations. Unless
the manufactured solution happens to be an exact solution of the governing
equations, the left- and right-hand sides of the equation will differ. This
difference is then added to the right-hand side of the governing equations as a
source term to make the manufactured solution an exact solution of the
modified governing equations. By solving the modified governing equations
numerically, the order of accuracy of the numerical method can be determined
even for complicated governing equations that include variable physical
properties, for example.
In the present study, the manufactured solution was of the form
man A1 A2 cos( A3 r ) cos( A4 z ) (8.29)
where is a generic dependent variable. Physical properties such as thermal

conductivity, density, and diffusion coefficients were represented in the same
form as the dependent variable. All governing equations and their individual
terms were verified using the MMS. For brevity, only the results of the order
verification of the diffusive terms in the species equation (Eq. 8.4) are reported.
The modified equation reads
1 M k yk M k yk
0 r Deff,k , Deff,k , Lk (8.30)
r r M r z M z
where Lk is the source term to fulfill the governing equation for the
manufactured solution. The coefficients in Eq. 8.29, the dependent variable,
and the properties Deff,k, , Deff,k, , and are listed in Table 8-3. Dirichlet
boundary conditions were applied on all boundaries.
Figure 8-7a shows the evolution of E2, defined as the L2-norm of the
relative errors in each cell, as a function of the radial mesh spacing. The actual
order of accuracy qa, shown in Figure 8-7b, was calculated as the slope of E2 of
successive data points. When using the hybrid scheme, the expected order of
accuracy depends on the Péclet (Pe) number. For Eq. 8.30 Pe = 0 and the
nominal order of accuracy is 2. In the asymptotic range, the actual order of
accuracy qa is very close to the nominal order of accuracy, indicating that the
numerical implementation is correct. The numerical implementation of all the
other terms in the governing equations was also verified to be correct.
a) -2
-4
log(E2)
-6
-8
-5.0 -4.5 -4.0 -3.5 -3.0 -2.5
log(Δr)
b)
2
1
qa
0
-5.0 -4.5 -4.0 -3.5 -3.0 -2.5
log(Δr)
Figure 8-7: Results from code verification of the diffusion terms in the
species equation, a) error norm E2 and b) actual order of accuracy qa.
Table 8-3: Parameters used in order verification study (all in SI units).

variable A1 A2 A3 A4
0.1804 0.05 2.5/rmax 3.7/zmax
DH O 2.8·10–4 8.0·10–3 5/rmax 3/zmax
wH O 0.3 0.2 0.7/rmax 0.8/zmax
wN 0.7 –0.2 0.7/rmax 0.8/zmax
8.3 Simulation Results 101
8.3 Simulation Results
8.3.1 Model Predictions versus Experimental Results
The experiments were simulated using a 64×64 cell mesh that provided grid-
independent solutions. The experimentally determined particle loadings were
used as inputs for the simulations of the reactive case. As the fluid temperature
at the inlet to the RPC (Tf,in) was not measured, it was set to the measured wall
temperature 0.25 m above the RPC. This was justified by the relative
insensitivity of the carbon conversion and the solid temperatures to changes in
Tf,in. (Changing Tf,in by ± 100 K influenced the gasification rates by less than 1%
and the solid centerline temperature at z > 0.05 m by less than 1 K.)
Figure 8-8 compares the measured and simulated solid centerline
temperatures for the non-reactive (Figure 8-8a) and the reactive cases
(Figure 8-8b). The open and filled symbols indicate the thermocouple
measurements at the wall (Twall,exp) and at the centerline (Ts,center,exp),
respectively. The dashed lines indicate the wall temperatures (Twall,bc) that were
set as boundary conditions in the model. The solid lines are the centerline
temperatures (Ts,center,sim) obtained from the numerical model. The simulations
of the non-reactive conditions shown in Figure 8-8a were used to assess the
model performance in terms of heat transfer only. The predicted centerline
temperatures are seen to be in good agreement with experimental
measurements. Furthermore, the model is capable of predicting the lower
centerline temperatures at both upper and lower boundaries of the RPC that
reflect the radiation losses to the surroundings. Figure 8-8b indicates that good
agreement between the model and the experiment is also obtained for the
reactive cases. The predicted differences between solid and fluid temperatures
were in all cases smaller than 1 K for z > 0.02 m; larger differences were
observed only in the first rows of cells close to the inlet (z < 0.02 m).
a)
1273 K / 15%vol
1300
1223 K / 20%vol
1250
Ts [K]
1223 K / 15%vol
1200
1173 K / 15%vol
1150
0.00 0.05 0.10 0.15 0.20
z [m]
b)
1300 1273 K / 15%vol
1223 K / 20%vol
1250
Ts [K]
1223 K / 15%vol
1200
1173 K / 15%vol
1150
1100
0.00 0.05 0.10 0.15 0.20
z [m]
Figure 8-8: Comparison between experimental (Ts,center,exp - filled markers)

and simulated centerline solid temperature profiles (Ts,center,sim - solid lines) for
a) non-reacting conditions, and b) reacting conditions. The open markers
indicate the experimental wall temperatures (Twall,exp); the dashed lines
represent the boundary temperatures (Twall,bc) applied in the simulations.
The parity plot in Figure 8-9 indicates reasonably good agreement between the
simulated and experimentally determined molar flow rates of gasified carbon
( C,syngas) leaving the reactor. The scatter of the model results for a given
furnace temperature and steam concentration are due to differences in the
experimentally determined particle loadings.
1.0
0.8
[mmol/s]
0.6
C,syngas,exp
1173KK/ /15%vol
1173 15%vol
0.4 1223KK/ /15%vol
15%vol
1223
1273KK/ /15%vol
1273 15%vol
1223
1223KK/ /20%vol
20%vol
0.2
0.2 0.4 0.6 0.8 1.0
C,syngas,model [mmol/s]
Figure 8-9: Comparison of experimentally determined and simulated molar

flow rates of gasified carbon.
8.3.2 Sensitivity Analysis
Having demonstrated good agreement with the experimental data, the

numerical model was used to conduct a sensitivity analysis of the reactor
performance with respect to the reactor wall temperature, particle loading, and
RPC properties. The analysis was based on a 1 m tall / 0.06 m inner-diameter
domain that is considered a reasonable scale for a pilot-scale reactor on a solar
tower. The simulations were carried out on a grid with 64 and 128 cells in
radial and axial direction, respectively, yielding grid-independent solutions for
the extended domain.
Simulations performed with uniform wall temperatures of Twall = 1173,

1223, and 1273 K, a particle loading of l = 0.2, and pure steam entering the
domain at f,in = 2 g/s and Tf,in = 873 K were adopted as reference conditions.
The reference conditions were chosen to provide excess steam for all
gasification conditions, i.e. the maximum steam utilization was around 50%. As
the key performance indicators reflecting the reactor productivity and syngas
quality, the molar flow rates of the gasified carbon leaving the trickle bed
( C,syngas,RPC) and solid centerline temperatures at the outlet (Ts,center,outlet) are
presented in Table 8-4. As shown by this table, increasing the wall temperature
from 1173 to 1273 K leads to more than doubled C,syngas,RPC at the
corresponding change in Ts,center,outlet of only ~ 40 K.
Figure 8-10 presents simulated radial profiles of temperature, steam mole
fraction, and gasification rate at the mid-plane of the domain (z = 0.5 m) for
Twall = 1173, 1223, and 1273 K. As shown in Figure 8-10a, the solid
temperature and its slope are highest at the tube wall and are seen to increase
with an increase in wall temperature as expected. Conversely, the steam mole
fraction, plotted in Figure 8-10b, is the highest in the tube center and decreases
with an increase in wall temperature. These observations can be attributed to
the exponential temperature dependence of the gasification rate and the slow
radial dispersion of steam leading to the steep profiles of the gasification rate
per unit volume ( mC'" ) with the highest rates near the wall, see Figure 8-10c.
Table 8-4: Simulation results for a 1 m tall / 0.06 m inner diameter domain
with dnom = 2.54 mm, ks = kSiC, RPC = 89%, f,in = 2 g/s, Tf,in = 873 K, and l =
0.2.
wall temperature overall gasification rate average centerline temperature

Twall [K] C,syngas,RPC [mmol/s] Ts,center [K]
1173 27.0 1067
1223 42.0 1083
1273 58.9 1097
a) 1300
Twall = 1273 K
1200 Twall = 1223 K
Ts [K]
1100 Twall = 1173 K
1000
0 0.01 0.02 0.03
r [m]
b) 1
Twall = 1173 K
0.8
0.6 Twall = 1223 K

yH2O [-]
0.4 Twall = 1273 K
0.2
0
0 0.01 0.02 0.03
r [m]
c)
0.6
ṁ'''char [kg/m3-s]
Twall = 1273 K
0.4
Twall = 1223 K
0.2
Twall = 1173 K
0
0 0.01 0.02 0.03
r [m]
Figure 8-10: Simulated radial profiles of a) solid temperature, b) steam mole

fraction, and c) gasification rate at z = 0.5 m for l = 0.2 and Twall = 1173, 1223,
and 1273 K.
Figure 8-11 shows the responses of the syngas production rate ( C,syngas,RPC)
and the temperatures at the outlet to changes in the particle loading (l) and the
RPC properties (pore diameter, solid thermal conductivity, and porosity) for
three wall temperatures (Twall = 1173, 1223, and 1273 K). The changes are
reported with respect to the reference conditions discussed above. Based on the
results summarized in Figure 8-11a one could draw the following inferences
relative to the responses of syngas production rate:
1. Increasing dnom or ks at fixed RPC results in an increase in the effective

thermal conductivity of the RPC (keff) and thus an increase in the
overall gasification rate. An increase in dnom leads to an increase in the
radiative contribution due to a decrease in the extinction coefficient,
thus its effect is more pronounced at higher temperatures. Conversely,
the changes in ks have a larger influence at lower temperatures due to
the larger contribution of the conduction heat transfer to the effective
thermal conductivity of the structure.
2. An increase in the volume fraction of the RPC solid 1– RPC is

reflected by an increase in kcond (see Eq. 8.25) but impaired radiative
heat transfer due to a higher extinction of radiation. The increase in 1–
RPC is thus less beneficial at higher temperatures.
3. Of all the investigated parameters, the particle loading has the largest
effect: a 25% increase in the particle loading leads to an increase in the
overall gasification rate by 6%–11%. The increase is especially
pronounced at lower temperatures where the temperatures and
gasification rates are more uniform.
Figure 8-11b shows the responses of the temperatures at the outlet of the
domain as the difference between wall and centerline temperatures Ts,outlet =
Ts,oulet,wall – Ts,outlet,center. It can be observed that:
1. An increase in dnom, ks, or 1– RPC results in a higher effective thermal

conductivity and thus a decrease in Ts,outlet. As for the overall
a) 1.1 1173 K 1.1 1.1 1.1
1223 K
1273 K
1 1
PC
PC
PC
0.75 1 1.25 0.75 1 1.25 0.75 1.25 0.75 1.25
ṅC,syngas,RPC / ṅC,syngas,RPC,ref
0.9 0.9 0.9 0.9
dnom / dnom,ref ks / ks,ref (1 – φRPC) / (1 – φRPC)ref l / lref
b) 1.1 1.1 1.1 1.1
PC
1
PC
1
PC
0.75 1.25 0.75 1.25 0.75 1.25 0.75 1 1.25
ΔTs,outlet / ΔTs,outlet,ref
0.9 0.9 0.9 0.9
dnom / dnom,ref ks / ks,ref (1 – φRPC) / (1 – φRPC)ref l / lref
Figure 8-11: Sensitivity of a) the syngas production rate ( C,syngas,RPC) and b) the temperature difference between wall
and centerline at the outlet of the domain Ts,outlet = Ts,oulet,wall – Ts,outlet,center to changes in nominal pore diameter (dnom),
solid thermal conductivity (ks), RPC porosity ( RPC), and particle loading (l) for Twall = 1173, 1223, and 1273 K.
Reference case: dnom = 2.54 mm, ks = kSiC, RPC = 89%, and l = 0.2.
gasification rate (Figure 8-11a), the sensitivity of Ts,outlet is lower to ks

and 1– RPC as the wall temperature increases. No wall-temperature
dependence was observed for the sensitivity to dnom.
2. An increase in the particle loading leads to an increase in Ts,outlet due

to the reduced effective thermal conductivity and the increased
consumption of heat of the gasification reaction. The sensitivity is
larger at lower wall temperatures as the increase of the gasification
rate is higher and more heat is consumed.
8.3.3 Numerical Comparison of the Trickle Bed to a Moving Bed
An alternative to the incorporation of a trickle bed into a drop tube for

increasing the residence time of char would be to operate the gasifier as a
moving bed of char particles (no RPC, l = 1). To compare the performance of
the trickle bed and the moving bed in terms of temperature profiles and overall
char gasification rates, simulations were carried out for the same flow rates,
domain size, grid density, and boundary conditions described in the previous
section.
For simulation purposes, the moving bed was assumed to consist of
uniformly sized char particles (mean particle size = 0.81 mm) and to have a
porosity of MB = 0.56 [31]. As the temperature differences between the solid
and fluid phase are expected to be even smaller than in the RPC trickle bed, the
model has been reduced to a pseudo-homogeneous case, requiring only one
energy equation to be solved. The effective thermal conductivity (keff) was
based on a Yagi and Kunii model [123] with fitting parameters from
Piatkowski [31] and adjusted for the smaller particle size. The contribution of
ash to the thermal conductivity was neglected. A more detailed description of
the effective thermal conductivity model applied can be found in Appendix B.
The correlations for the diffusion and dispersion applied for the moving bed
are the following [124, 125]:
Dm,k uz d p
Deff,k , 2
(8.31)
PeT ( )
Dm,k uz d p
Deff,k , 2
(8.32)
PeL ( )
where PeT( ) 12 and PeL( ) 2. The tortuosity of a randomly packed bed

can be described by [126]
2
2 (8.33)
Figure 8-12 compares the moving bed (MB) and trickle bed (RPC) in terms of
radial profiles of temperature, steam mole fraction, and gasification rate
simulated for Twall = 1223 K at z = 0.5 m. As compared to the moving bed, both
the temperature and steam concentration profiles are more uniform across the
RPC configurations operating at particle loadings l = 0.1–0.4, see Figures 8-12a
and 8-12b. The steep temperature gradient in the moving bed close to the wall
is also consistent with previously discussed effect of particle loading. The
steam mole fraction at the wall decreases with an increase in the particle
loading thus it is the lowest for the moving bed. At the centerline of the moving
bed, the steam concentration does not differ considerably from the one at the
inlet. Due to the low core temperatures in the moving bed, the gasification rate,
plotted in Figure 8-12c, approaches zero at about one third of the radius from
the wall (~0.01 m) so the steam passes through unreacted. Conversely, the
reaction within the RPC is more uniformly distributed and proceeds at a
noticeable rate even in the center. The gasification rates in the regions closer to
the wall increase with particle loading as more char is exposed to the highest
temperatures. Therefore, the moving bed exhibited higher gasification rates at
the wall then the trickle bed.
M
y
Figure 8-12: Simulated radial profiles of a) solid temperature (Ts), b) steam

mole fraction (yH O), and c) gasification rate ( '''C) in moving bed (MB) and
RPC (l = 0.1, 0.2, and 0.4) at z = 0.5 m for Twall = 1223 K.
The higher core temperatures and reaction rates of the RPC configuration
can be attributed to the heat transfer enhancement brought in by the conduction
through the RPC material and deeper penetration of radiation through the RPC
pores. Figure 8-13 compares the effective thermal conductivities of the moving
bed with N2-filled voids and the RPC with different particle loadings. The
presence of the RPC is seen to significantly increase the effective thermal
conductivity of the domain. At lower temperatures and a particle loading of 0.4,
a decrease in the effective thermal conductivity with increasing temperature is
observed. This is due to a decrease of the conductivity of the RPC material (ks).
At lower particle loadings and higher temperatures the radiative contribution
dominates resulting in a higher effective thermal conductivity. At higher
temperatures the radiative contribution increases which leads to a higher
effective thermal conductivity. This is illustrated in Figure 8-14. The
contribution of the radiation increases from about 20% at 900 K to around 60%
at 1400 K. Increased loadings result in a lower contribution of radiation as the
voids get smaller and thus lower effective thermal conductivity.
4 l = 0.1
l = 0.2
RPC
3
l = 0.4
keff [W/m-K]
MB
0
900 1000 1100 1200 1300 1400
T [K]
Figure 8-13: Effective thermal conductivity (keff) over temperature for the
moving bed (MB) with N2 filled voids and the RPC configuration at various
loadings (l).
1.0
0.8
l = 0.1
0.6
krad/keff [-]
l = 0.2
l = 0.4
0.4
0.2
0.0
900 1000 1100 1200 1300 1400
T [K]
Figure 8-14: Contribution of radiation term krad on the effective thermal

conductivity keff for RPC configuration.
Figure 8-15 compares the overall gasification rates at Twall = 1173, 1223,
and 1273 K calculated for the moving bed ( C,syngas,MB) and the trickle bed
( C,syngas,RPC) operating at various particle loadings (l). At low particle loadings,
the overall gasification rate of the trickle bed is lower than in the moving bed,
as less reacting material is present in the reactor. With an increase in the
particle loading in the RPC, the ratio of trickle-bed to moving-bed gasification
rates increases rapidly at first and then flattens. The steep increase at low
particle loadings is almost linear with the particle loading. The flattening at
higher particle loadings occurs due to the reduction in radiative contribution
and reduced steam concentrations close to the wall. At wall temperatures of
1173 and 1273 K, the trickle bed outperforms the moving bed above particle
loadings of 0.18 and 0.32, respectively. At Twall = 1273 K, the trickle bed with l
= 0.5 gasifies ~25% more carbon than the moving bed containing twice as
much char.
Even though the gasification rates at low particle loadings are lower than in
the moving bed, an RPC might be preferred as the higher core temperatures
favor the decomposition of volatile hydrocarbons passing through this zone,
8.4 Conclusions 113
thereby improving the syngas quality. This is especially important when using
feedstocks that release high amounts of hydrocarbons and tars.
ṅC,syngas,RPC /ṅC,syngas,MB [-] 1.5
0.5 1273 K
RPC
RPC
1223 K
RPC
1173 K
0
0 0.2 0.4 0.6 0.8
l [-]
Figure 8-15: Simulation results for ratio of trickle-bed to moving-bed

gasification rates as function of the particle loading in the RPC (l) and wall
temperatures of Twall = 1173, 1223, and 1273 K.
8.4 Conclusions
A gas-solid trickle-bed reactor for the gasification of beech char particles was
evaluated experimentally and numerically. Efficient delivery of externally
supplied concentrated solar heat to the reaction site was achieved by
incorporating a high conductivity RPC as packing into a drop-tube reactor that
increased the residence time of the solids and enhanced the heat transfer to the
gas phase. A series of 43–51 min gasification experiments in the temperature
range 1173–1273 K with a mass flux of 6.5 g/s-m2 was conducted with no
pressure increase or clogging over the duration of the experiments. Compared
to the drop-tube configuration, the carbon conversions attained under same
conditions in the trickle-bed were significantly higher (< 1% versus 52%).
A two-dimensional finite-volume model of the trickle-bed reactor including

chemical reaction coupled with conduction, convection, and radiation of heat
within the porous structure was developed. Its predictions were in good
agreement with measured reactor temperatures and gasification rates. A
sensitivity analysis of the overall gasification rate and the core temperature, the
key parameters for the capacity and gas phase selectivity, indicated that any
effort to increase the effective thermal conductivity is beneficial. Reduced
particle loadings increase the core temperatures but result in lower gasification
rates.
A numerical comparison with a moving bed of char particles showed a more
uniform temperature distribution over the cross-section of the trickle-bed. This
is attributed to both high conductivity of the RPC and suppressed extinction of
the radiation due to higher open porosity of the trickle-bed. The more uniform
temperatures and thus the higher core temperatures are expected to improve
gas-phase selectivity towards the desired product gas. Furthermore, it was
shown that the gasification rates of the proposed configuration exceed those of
a moving bed for particle loadings higher than 32% for a wall temperature of
1173 K and 18% for 1273 K, respectively.
Chapter 9
9 Overall Conclusions and Outlook
This thesis was performed in the framework of a joint project of the

Universidade de São Paulo, Brazil and ETH Zürich, Switzerland. It investigates
the thermochemical steam-gasification of biomass using concentrated solar
energy for process heat. The product of the gasification is carbon neutral
synthesis gas, a mixture of hydrogen and carbon monoxide. The synthesis gas
produced can be catalytically converted to gaseous or liquid hydrocarbon fuels,
e.g. to H2 via water-gas shift reaction, to diesel or kerosene via Fischer-Tropsch
synthesis, or to methanol and then gasoline via the methanol-to-gasoline
process (Mobil). The syngas may also be used directly for power generation by
combusting it in a turbine or an internal combustion engine. Further uses
include the production of valuable chemicals.
9.1 Thermodynamics and Kinetics
Thermodynamic equilibrium compositions were calculated for the autothermal

oxygen/steam gasification and the solar-driven steam gasification of Brazilian
sugarcane bagasse to assess the differences in the syngas quality. For the solar
gasification, an almost full conversion to H2 and CO is achieved at 1 bar and
1350 K. Higher temperatures are therefore worthwhile only if the reaction rates
are accelerated and the thermal losses of the process can be reduced. Compared
to the autothermal gasification, the solar syngas shows a superior quality with a
higher H2/CO ratio (1.16 vs. 1.07) and less dilution with CO2 (0.2%vol vs.
12%vol).
116 Chapter 9: Overall Conclusions and Outlook
First and second law analyses were conducted comparing the autothermal to
the solar gasification in order to determine their performance, to find the
maximum energy and exergy conversion efficiencies, and to identify the major
sources of irreversibility. Theoretical upgrade factors (ratio of heating value of
syngas produced over that of the feedstock) of 1.26 and lower heating values of
19.4 MJ/kg can be achieved by the solar gasification, whereas the autothermal
pathway achieves only an upgrade factor of 0.95 and a lower heating value of
13.9 MJ/kg. The autothermal gasification requires about 25% of the bagasse to
be combusted to provide the reaction enthalpy for the endothermic gasification
reactions. Thereby, the amount of H2O consumed per unit feedstock in a
stoichiometric gasification is also reduced by 25% compared to the solar-driven
process. In the solar-driven process all feedstock is utilized. Moreover, the
feedstock is upgraded by 26% while chemically storing the solar energy at a
solar to chemical efficiency of 66%. In comparison to autothermal gasification,
the syngas composition is more suitable for the synthesis of liquid fuels.
Further, the possible treatment of feedstock with higher moisture content and
the elimination of the air separation unit, support the potential benefits of solar-
driven gasification.
The gasification rates of the biomass with steam and the resulting residence
time requirements play an important role in terms of efficiency of a solar
reactor and are crucial its design. Therefore, the kinetics of the rate limiting
steam gasification step of pyrolyzed bagasse particles were experimentally
studied by thermogravimetric analysis in the temperature range of
1173 1473 K and steam concentrations of 25–75%vol. A kinetic rate law based
on the oxygen exchange mechanism and a linear decrease of the effective char
surface area with conversion was found to fit the experimental data well. The
required residence time for 90% char conversion at a steam concentration of
75%vol was predicted to be 16 s at a reactor temperature of 1573 K and 43 s at
1373 K, a temperature that would allow the use of thermal shock resistant metal
alloys for the construction of a solar reactor.
9.2 Reactor Design and Testing 117
9.2 Reactor Design and Testing
In a next step, a two-zone solar reactor concept for the gasification of the highly
volatile bagasse was proposed. By combining drop-tube with fixed-bed
concepts, the proposed two-zone laboratory-scale reactor provided sufficient
solids residence time for adequate char conversion and temperatures high
enough for the decomposition of the other pyrolysis products such as tar and
gaseous hydrocarbons.
Experimental testing in the temperature range of 1073 1573 K was done by
exposing the tubular reactor to infrared radiation from an electric furnace,
thereby simulating the conditions present in an absorbing solar cavity receiver.
The preliminary results, in particular the carbon conversion and the
decomposition of CH4, stayed behind the expectations, which implied that
further development was necessary. CH4 was detected in significant amounts
over the whole temperature range investigated even though its presence was not
predicted by thermodynamic equilibrium at above 1200 K. Further, it was seen
that ash melting, which occurred at temperatures above 1473 K, gives an upper
limit to the operating temperatures of the reactor concept. It could however be
confirmed that an upgrade factor of greater than 1 is attainable and that syngas
yields per unit feedstock as well as the heating values of the syngas are
significantly higher than those typically obtained in autothermal gasifiers.
Based on the experience gained with the prototype reactor, the concept was
further developed to achieve higher the carbon conversion and a better
decomposition of the gaseous hydrocarbons. Moreover, a 1.5 kWth solar-driven
biomass gasifier consisting of a cavity-type receiver housing a tubular reactor
has been developed and evaluated with bagasse particles under simulated high
flux solar radiation. The two-zone reactor comprises a drop-tube for fast
pyrolysis and a trickle bed for the char gasification and decomposition of the
pyrolysis products. The drop-tube zone ensures efficient radiative heat transfer
to the dispersed biomass particles that is required for their fast pyrolysis. The
trickle-bed zone consists of a structured packing made of reticulate porous
ceramic (RPC), that allows controlling the overall porosity of the gasification
zone in order to increase the residence time of the char particles while still
allowing the radiation to penetrate through. Thereby, it provides residence time

and temperature needed for the rate limiting char gasification and the
decomposition of the pyrolysis products.
The comparison with a drop-tube configuration demonstrated that the two-
zone reactor concept decomposes CH4 and C2-hydrocarbons more efficiently.
Again, the syngas produced showed higher LHV’s than common in autothermal
gasification. Moreover, the carbon conversion was significantly higher than in
the drop-tube configuration. The biomass was energetically upgraded by 5%,
effectively storing solar energy in the syngas. The maximum energy conversion
efficiency was 21%.
An energy flow analysis of the solar reactor identified the largest losses as
conduction through the cavity walls and along the reactor tube. Only 10% of
the input power was transferred to the reactants. The largest efficiency gains are
thus expected by increasing the throughput of the gasifier.
9.3 Modeling
The heat transfer in the trickle-bed zone of the two-zone solar reactor was
studied using an externally heated gas-solid trickle-bed reactor with a RPC
packing. Experimental testing with beech char particles was performed in the
temperature range 1173–1273 K with a mass flux of 6.5 g/s-m2. In the 43–
51 min long experiments neither a pressure increase nor clogging of the RPC
were observed. The achieved carbon conversions were significantly higher than
for the drop-tube configuration (< 1% vs. 52%).
The predictions of reactor temperatures and gasification rates of a two-
dimensional finite volume heat and mass transfer model of the reactor including
chemical reaction and heat transfer within the porous structure agreed well with
the experimental measurements. A sensitivity analysis indicated that any effort
to increase the effective thermal conductivity of the trickle bed is beneficial for
the reactor temperature and the gasification capacity. Furthermore, a numerical
comparison with a moving-bed reactor demonstrated that the increased heat
transfer via combined radiation and conduction through the porous RPC leads
9.4 Outlook 119
to a more uniform temperature distribution and more uniform steam

concentrations across the reactor. As a result, higher core temperatures and
gasification rates than in a moving bed are achieved, the higher temperatures
being beneficial for the decomposition of pyrolysis products.
9.4 Outlook
Future research should focus on the following main experimental and

numerical topics:
long term operation
online ash removal strategies
increase of throughput
heat and mass transfer of various packings
feeding of moist feedstock
chemical processes that could use the same reactor concept
So far the steam gasification experiments were only conducted under the very
limited conditions of both short-term operation (less than an hour) and at light
particle loadings. As the goal for solar reactors is operate continuously, daily at
least from sunrise to sunset, long term experiments should be performed to
study the industrial feasibility of the concept. Testing should be done with a
variety of feedstocks, particle sizes, and feedstock preparation methods.
In conjunction with the long term experiments, the deposition of ash will
become a more pronounced issue. Therefore, the feedstocks investigated should
preferably have a high ash slagging temperature as this gives an upper limit to
the allowable operating temperatures. Moreover, time should be dedicated on
investigating online ash removal options, such as periodic inert or product gas
purge, gas pulses, or vibration of the packing.
It was seen that the throughput of the solar reactor is critical to its
efficiency. An increase in the throughput requires the knowledge and control of
the solids residence time distribution. Therefore, the maximum solids feeding
rates should be found for the different feedstocks, particle sizes, and
preparation methods. Further, methods to increase the feeding rates should be
explored. These may include the previously mentioned gas pulses or the
vibration of the packing but should also include alternative packing properties
and designs that go beyond ceramic foams. As the heat and mass transfer are
critical for the throughput of the reactor, alternative packings should therefore
be designed with that in mind.
The two zones of the gasifier concept should be studied in more depth. This
is best achieved by individually studying the pyrolysis and the gasification
zone. By doing so, experimental measurements are not impaired by the
presence of the other zone. Prior to scaling-up to a multi-tube design, it is
important to test and improve the throughput of a single tube by using a taller
reactor with a larger hot zone. The current solar reactor has a roughly 120 mm
long high temperature pyrolysis zone and an 80 mm tall trickle bed zone. An
extension of the pyrolysis zone might allow the operation with moist feedstock
as there would be sufficient residence time to evaporate the water and pyrolyze
the particles before reaching the trickle bed. The operation with wet feedstock
would reduce the effort and energy needed for the feedstock preparation and
might thus be beneficial for the economics of a solar gasification system.
Further, an increase in the height of the trickle-bed zone would increase the gas
and solid residence times and is thus expected to improve the char conversion
and the decomposition of the other pyrolysis products.
The reactor concept that has been developed is not only applicable for the
gasification of biomass and other carbonaceous feedstock. It is thus important
to also consider different high temperature gas-solid reactions that might
benefit from the proposed designs and conclusions of this work.
Appendix A:
Solar Gasification of Microalgae in a Drop-Tube
This chapter is an excursion from the previously described two-zone reactor

concept. The feedstock under consideration is microalgae (chlorella vulgaris).
Microalgae are unicellular photosynthetic micro-organisms that can grow in
practically any aquatic environment, from freshwater, sea water to even waste
water. They need a carbon source, which is usually CO2, nutrients, and light,
and grow 10 to 100 times faster than terrestrial plants [127], and are thus
considered as a potential feedstock for the large scale production of biofuels.
They can be cultivated on land that is unsuitable for agriculture or forestry and
therefore do not compromise production of food, fodder and other crop-derived
products. Further, the production yield per unit area is higher than for other
biofuel sources.
Microalgae are commonly grown in open pond or closed photobioreactors,
both of which have their merits. Open pond systems are shallow round or
raceway-style ponds that are mixed with gas bubbling through the growth
medium or with paddle wheels, respectively. They are relatively cheap to build
and to operate but suffer from high water consumption due to uncontrolled
evaporation, the risk of contamination and changes in the environment (wind,
rain, temperature fluctuations). Tubular or flat plate photobioreactors have high
productivities due to their well-controlled and easy to maintain operational
conditions. The requirements for make-up water are small. As there is no
evaporative cooling, temperature control is necessary. Gas exchange is the
limiting factor for their scalability. [128]
Apart from the growth of the microalgae, harvesting and dewatering are the
main challenges as they are very energy intensive due to the microalgae’s high
water content. One liter of ready to harvest growth medium contains only about
1–2 g (dry weight) microalgae. Commonly suggested harvesting methods, such
122 Appendix A: Solar Gasification of Microalgae
as sedimentation (~98.5%wt water), centrifugation (~80%wt water), filtration

(~80%wt water), and flocculation, do not reduce the water content significantly
and are often highly energy- or chemical-intensive. Further reducing the water
content requires thermal drying, e.g. flash drying, spray drying, sun drying, or
drying in rotary dryers. [129, 130]
Several routes for the conversion of microalgae into valuable products such
as chemicals and versatile fuels are considered in the US DOE’s algal biofuels
roadmap [131]. One of them is the gasification of microalgae. Most
conventional autothermal gasification systems use dry biomass with a moisture
content of 10%wt to 20%wt. Higher moisture contents are feasible in certain
reactor types but result in significantly reduced energy efficiency as for every
kilogram of moisture in the feedstock, a minimum of 2260 kJ of additional non-
recoverable energy from the gasifier has to be provided to vaporize the water
[14]. Thermal drying of the feedstock prior to gasifying it is often times
preferred as this is more efficient than evaporating the moisture in the gasifier
and can simplify its handling. For solar-driven gasification a moisture content
of about 25%wt represents the stoichiometric amount of water for the steam
gasification of microalgae to H2 and CO. A slight over-stoichiometry might be
beneficial for the reaction kinetics. Larger amounts of excess moisture reduce
the efficiency in the same way as in the conventional gasification.
In this work, the steam gasification of dry microalgae was investigated in a
1.5 kW indirectly-heated solar entrained flow gasifier consisting of an
absorbing cavity receiver housing a tubular reactor was experimentally
examined at ETH’s high flux solar simulator. The influence of different steam
concentrations on the composition and the heating value of the syngas, the
carbon conversion, the energetic upgrade, and on the reactor efficiency was
evaluated.
A.1 Materials 123
A.1 Materials
Figure A-1: Spray dried microalgae (chlorella vulgaris).
The microalgae used in this work are depicted in Figure A-1. They are a
chlorella vulgaris species that were produced in a tubular photobioreactor and
then spray-dried (Roquette, Klötze GmbH & Co. KG). The physical and
chemical properties given in Table A-1 were determined as described for the
bagasse and beech char in Section 3.3. Before all experiments and analyses the
microalgae were dried in a convection oven for at least 4 h at 378 K.
Table A-1: Physical and chemical properties of dry microalgae: Ultimate and
proximate analyses, heating value and mean particle size.
microalgae (chlorella vulgaris)
carbon (C) [%wt] 48.8
hydrogen (H) [%wt] 6.5
oxygen (O) [%wt] 30.7
nitrogen (N) [%wt] 8.0
sulfur (S) [%wt] 0.6
H/C = x [mol/mol] 1.597
O/C = y [mol/mol] 0.472
volatiles [%wt] 76.2
fixed carbon [%wt] 19.5
ash [%wt] 4.3
LHV [MJ/kg] 19.71
dp [ m] 59
The elemental analysis revealed an average chemical formula of

CH1.597O0.472 on a dry and ash free basis. The nitrogen content was 8%wt, which
is significantly higher than in the bagasse or the beech char and should
therefore not be neglected when analyzing the product gas as the release of the
nitrogen during gasification might be measurable.
The particle size distribution of the microalgae is shown in Figure A-2.
They had a mean particle size of 59 m, which is about an order of magnitude
smaller than for the previously investigated feedstocks.
volume density f (dp) ·dp3
1 10 100 1000
particle diameter, dp [ m]
Figure A-2: Particle size distribution of spray-dried microalgae analyzed with

LA-950 analyzer (HORIBA).
A.2 Experimental Setup 125
A.2 Experimental Setup
The smaller particle size of the microalgae compared to the bagasse previously
investigated was expected to reduce the required residence times for high
carbon conversions significantly. This might allow achieving high conversions
even without the built-in RPC. Moreover, the ash melting temperature of the
microalgae was found to be around 1273 K, so about 200 K lower than for the
bagasse, making them unsuitable for gasification using the RPC packing.
The existing 1.5 kW solar gasifier used for the bagasse gasification
experiments was thus modified to an indirectly-irradiated solar drop-tube
gasifier, similar to the concept investigated by Lichty et al. [38]. The solar
reactor and the peripheral components used for the gasification of microalgae
are schematically shown in Figure A-3. The solar reactor is identical to the one
fluidized
bed of algae Xe short-arc
lamps
evaporator Tsteam
elliptical
reflectors
Ar simulated
Tcenter solar
radiation
water tank
filter
condenser
gas analysis
steam trap
vent
Figure A-3: Schematic of the solar entrained flow gasifier including solar
simulator and the primary components and flows.
described in Section 7.2.1 except that no RPC and no supporting tube were
installed in the reactor tube.
At the top of the rector tube, dry microalgae are introduced as an aerosol
flow from an Ar fluidized bed feeder. Microalgae feeding rates are controlled
by adjusting the bypass flow of the fluidized bed keeping the total flow rate
through the aerosol inlet constant. Below the aerosol inlet Ar and steam are
injected through annularly distributed inlets. The steam is generated with an
external evaporator (Bronkhorst) and is superheated in an electrically heated
tube connecting the evaporator with the inlet. The flow rates of the inlet gases
and water into the evaporator are controlled with electronic mass flow
controllers (Bronkhorst). Temperature measurements were done with K-type
thermocouples place at the steam inlet (Tsteam) and inside the reactor tube at the
back side at the height of the aperture (Tcenter).
The particles entrained in the flow travel down the hot reactor tube, absorb
radiation from the tube, undergo rapid pyrolysis, and gasify in the steam. Tar
and gases released during the pyrolysis pass through the hottest zone of the
reactor and get at least partially decomposed. At the bottom of the reactor a
stream of the product gas is withdrawn, chilled, and filtered to remove
condensable components and particulate matter. The stream is then analyzed by
gas chromatography (Micro-GC, Varian 490 with Molsieve-5A and Poraplot-U
columns, 1/120 Hz sampling frequency, calibrated for H2, N2, CO, CH4, CO2,
C2H2, C2H4, and C2H6), infrared-based detectors for CO, CO2 and CH4
(Ultramat, Siemens), and a thermal conductivity based detector for H2 (Calomat
6, Siemens).
Experimentation was done on ETH’s high-flux solar simulator using up to 6
out of the 7 high-pressure Xe arcs lamps installed in elliptical reflectors. A
detailed description is given in Section 7.2.1.
A.3 Experimental Procedures
Before starting an experiment, the reactor was purged with Ar and tested for
leaks to ensure an O2 free environment. Up to 6 lamps of the solar simulator
A.3 Experimental Procedures 127
were then turned on to irradiate the reactor at power levels of up to

approximately Qsolar = 2 kW (at the cavity aperture) to achieve rapid heating.
Once the desired reactor temperatures were reached the flow rates of the
different gas streams were set and the power level was reduced to
approximately Qsolar = 1.2 kW to equilibrate the temperatures in the reactor.
The reactor temperature (Tcenter) was controlled to around 1430 K for all
experiments. Three different inlet gas compositions were investigated keeping
total volumetric flow rate of Ar and steam entering the reactor constant in all
experiments. The Ar flow through the aerosol inlet was set to 2.5 LN/mina and
the combined steam and Ar flow through steam inlet to 2.3 LN/min. The steam
flow rate was 0, 61 and 86 g/h resulting in an overall steam concentration of
0%vol, 27%vol, and 37%vol. The temperature of the gas entering through the
steam inlet Tsteam was around 510 K.
After temperature equilibration, microalgae (CH1.597O0.472) were fed for
15 min at an average feeding rate of 0.33 g/min (0.24–0.40 g/min) or a mass
flux of 4.2 g/s-m2. This led to molar steam to biomass ratios ( H O(g)/ CHxOy) for
the three different steam concentrations of 0, 4.5, and 6.3. This corresponded to
0, 8.5, and 12 times the stoichiometric amount of steam for the overall net
reaction given by
x
CH x O y 1 y H2O 1 y H2 CO (A.1)
2
Based on the gas composition measured the total outlet flow rates were
determined and integrated over time. To capture also the product gas that is still
in the reactor after switching off the feeding the integration included an
additional 5 min after shutdown. The instantaneous concentrations of the C2-
gases, only measured by GC, are computed based on the C2Hx/CO ratio
measured by the GC and the CO concentration measured by the infrared-based
detectors. This is justified due to the rather constant C2Hx/CO ratio over the
duration of the individual experimental runs.
a
A.4 Results
A typical experimental run is depicted in Figure A-4 showing the molar flow
rates of the major product gases. Shortly after the biomass feed is turned on
(0 min), a steep rise in the syngas production was observed. The lag between
the start of the feeding and the observed increase in the product gas flow rates
was caused by the distance the gas had to travel to reach the measurement unit.
At the end of the experiment the syngas flow rates dropped steeply. The
variations of the flow rates originated from unintended fluctuations of the
microalgae feeding rate.
A total of 10 experiments under the three previously specified conditions
were carried out. A summary of the experimental results is given in Table A-2.
The reported mole fractions are averages over all replicates that were calculated
considering only molar flow rates of the product gases integrated over the
course of each individual experiment; the flow rates of H2O and Ar were not
considered. LHV, upgrade factor (U ), and the carbon conversion (XC) were
calculated as defined in Eqs. 6.2–6.4 for each individual experiment and then
averaged over all replicates.
0.15 H2
CO
0.10
CH4
0.05
CO2
0.00
0 5 10 15 20
time [min]
Figure A-4: Molar flow rate of the product gases during a typical experiment.
A.4 Results 129
Table A-2: Summary of the experimental results (integral values averaged over
all replicates).
pure pyrolysis
27%vol steam 37%vol steam
(0%vol steam)
number of experiments 3 4 3
yH [%vol] 39.0 45.5 46.4
yCO [%vol] 32.7 25.7 24.5
yCH [%vol] 10.5 10.8 9.9
yCO [%vol] 5.0 11.4 13.5
yC H [%vol] 7.5 5.4 4.9
yC H [%vol] 5.1 1.3 0.9
yC H [%vol] 0.1 0.0 0.0
yH /yCO [-] 1.19 1.77 1.89
yCO /yCO [-] 0.15 0.44 0.55
Ttube [K] 1432 1425 1428
LHVsyngas [MJ/Nm3] 19.4 15.8 14.9
U [-] 1.02 1.05 1.06
XC [-] 0.45 0.53 0.61
Figure A-5 shows the product gas compositions for the different steam
concentrations. Changing the steam concentration at the inlet from 0%vol to
37%vol, has a significant influence on the product gas composition. An increase
in the concentrations of H2 (39.0%vol to 46.4%vol) and CO2 (5.0%vol to 13.5%vol)
and a decrease of CO (32.7%vol to 24.5%vol), C2H2 (7.5%vol to 4.9%vol) and
C2H4 (5.1%vol to 0.9%vol) were observed. CH4 concentrations (10.5%vol to
9.9%vol) were hardly influenced by the presence of steam. The amount of C2H6
was negligible under all conditions. The change in composition is caused by
more water gas shift reaction and a better decomposition of hydrocarbons. This
is also reflected by the increase in H2/CO ration (1.19 to 1.89) and CO2/CO
ratio (0.15 to 0.55).
The LHV of the product gas is shown in Figure A-6. It reduces with
increasing steam concentration from 19.4 MJ/Nm3 for pure pyrolysis to
14.9 MJ/Nm3 at a steam concentration of 37%vol. The lower values for higher
a) 0.5
H2
0.4
mole fraction [-]
0.3
CO
0.2
CH4
0.1
CO2
0.0
0% 10% 20% 30% 40%
steam concentration
b) 0.08
C2H2
0.06
mole fraction [-]
0.04
C2H4
0.02
C2H6
0.00
0% 10% 20% 30% 40%
steam concentration
Figure A-5: Product gas compositions for different volumetric steam

concentrations. Error bars indicate the span between the minimum and
maximum values.
A.4 Results 131
20
19
LHVsyngas [MJ/Nm3]
18
17
16
15
14
0% 10% 20% 30% 40%
steam concentration
Figure A-6: LHVsyngas for different volumetric steam concentrations. Error

bars indicate the span between the minimum and maximum values.
steam concentrations are due to a greater extent of the water gas shift reaction
and a better decomposition of C2H2 and C2H4. The heating values found are
significantly higher values than the 9.9–11.3 MJ/Nm3 found by Hirano et al.
[132] for the gasification of microalgae. They are also substantially higher than
what is typically achieved by conventional autothermal biomass gasification.
The carbon conversion (XC) plotted in Figure A-7 was between 45% and
61%. A significant improvement for higher steam concentrations could be
observed. This is caused by a reduction in the amount of tar not being
converted as well as char gasification. Higher carbon conversions are expected
to be achieved by increasing the length of the hot zone.
The upgrade factor (U) was greater than 1 for all investigated conditions
meaning that the biomass could be energetically upgraded and solar energy
chemically stored in the syngas. Further, an increase of the upgrade factor with
the steam concentration from 1.02 for the pure pyrolysis case to 1.06 for a
steam concentration of 37%vol could be observed.
1.0 1.10
0.9
1.05
carbon conversion [-]
upgrade factor [-]

0.8
0.7 1.00
0.6
0.95
0.5
0.4 0.90
0% 10% 20% 30% 40%
steam concentration
Figure A-7: Carbon conversion and upgrade factor for different volumetric
steam concentrations. Error bars indicate the span between the minimum and
maximum values.
The efficiency of the solar reactor is defined as
msyngas LHVsyngas
tend
(A.2)
Qsolar dt malgae LHValgae mC,in (1 X C )LHVchar
t0
where Qsolar = 1.2 kW, msyngas is the total mass of syngas produced, malgae is the
total mass of algae fed, and mC,in is the total mass of carbon fed with the algae.
Figure A-8 shows the efficiency for the different steam concentrations over
the particle feeding rate. The efficiency is between 3.8% and 7.2% and depends
strongly on the feeding rate but also increased steam concentrations result in
higher efficiencies. The main reasons for the low efficiency are conduction
losses along the tube and through the cavity walls which are inherent to lab
scale reactors and were discussed in Section 7.4. In larger reactors these losses
are significantly smaller due to the higher volume to surface ratio. Moreover,
the low particle feeding rates are responsible for the low efficiency. It is
expected that the efficiency would benefit strongly by an increase in the
particle loading.
A.5 Conclusions 133
8%
0%vol
7% 27%vol
37%vol
6%
efficiency
5%
4%
3%
0.2 0.3 0.4
feeding rate [g/min]
Figure A-8: Reactor efficiency in dependence of the feeding rate for the
three different steam concentrations.
A.5 Conclusions
The steam-gasification of dry microalgae in a 1.2 kW indirectly-heated solar

drop-tube gasifier consisting of an absorbing cavity and an absorbing tube was
experimentally examined at ETH’s high flux solar simulator at an inlet solid
mass flux of 4.2 g/s-m2. Synthesis gas with a H2/CO ratio of up to 1.89 and a
CO2/CO ratio of 0.55 was produced. The LHV of the syngas was between
19.4 MJ/Nm3 for pure pyrolysis and 14.9 MJ/Nm3 for 37%vol steam, thus
significantly higher than those typically obtained in conventional autothermal
gasification. It could also be confirmed that the biomass can be upgraded and
solar energy stored.
An investigation of the influence of the inlet steam concentration showed
that the carbon conversion, the upgrade factor, the efficiency as well as the
concentrations of H2 and CO2 in the syngas increase with increasing steam
concentration whereas the LHV of the syngas and the concentrations of CO,
C2H2, and C2H4 decreased. CH4 concentrations were hardly influenced by the
presence of steam. Further it was seen that the efficiency increases strongly
with an increase in the feeding rate.
Appendix B:
Thermal Conductivity of a Bed of Char
In the energy equation an effective thermal conductivity is applied for the heat
flux within the moving bed. A comparison of a broad variety of different
models [133] showed that the model of Yagi and Kunii [123] is suitable for
heat conduction calculations of coal with ash deposits at high temperatures. The
semi-empirical model of Yagi and Kunii that is followed in this work is a
resistance network model. By applying the diffusion approximation for
radiation within a solid medium, it includes the radiative conductivity in the
network of resistances. The heat transfer within a packed or moving bed can be
described as seven mechanisms that can be separated into two terms, a fluid
flow independent and a fluid flow dependent term.
Fluid flow independent heat transfer mechanisms:
1. Thermal conduction through the solid particles
2. Thermal conduction through the contact surface of the particles
3. Radiative heat transfer between the surfaces of the particles
4. Radiative heat transfer between neighboring voids
Fluid flow dependent heat transfer mechanisms:
5. Thermal conduction through the fluid film near the contact surface of
two particles
6. Convective heat transfer, solid-fluid-solid
7. Heat transfer by lateral mixing of fluid

136 Appendix B: Thermal Conductivity of a Bed of Char
For small Reynolds numbers the boundary layers around the solid particles are
thick and the mechanisms 1, 3, 4, and 5 are predominant. However, for large
Reynolds numbers, mechanism 7 controls the heat flux in any packed bed. Yagi
and Kunii include mechanism 5 in the fluid flow independent term since the
effect of the flow has only little effect on the mechanism. The model equation
for the effective thermal conductivity keff consisting of the conductivity for
motionless fluids k eff
0
and the conductivity due to lateral mixing of the fluid
(keff)t is
0
keff keff keff t
(B.1)
In the case of gas-filled voids the effective conductivity for motionless fluid
can be expressed as
0
keff 1 d p hrv
(B.2)
kf kf 1 kf
ks 1 d p hrs kf
The radiation heat transfer coefficient hrs through the contact surface is
3
T
hrs 1.163 0.1952 (B.3)
2 100
where is the emissivity. The radiation heat transfer coefficient through the
voids is
3
1 T
hrv 1.163 0.1952 1 (B.4)
21 100
According to existing data, parameter can be approximated as 0.04 [133].

Typical values for granules were chosen for and . and were both set to 1.
In the present case the fluid flow is perpendicular to the direction of the
main heat transfer. Although the Reynolds numbers are small, the effect of the
fluid flow is not negligible.
137
keff t
N peM (B.5)
kf
with NpeM = dpCpG/kf. G is the mass velocity of the fluid based on the empty
tube (G = u· ). For spheres, cylinders, pellets, or broken granules the values of
( ) cover a narrow range of 0.1 to 0.14 for the different packing conditions
[123]. Since no explicit values for the bed of char particles could be found in
literature an intermediate value of 0.12 was chosen for ( ).
For the solid thermal conductivity in Eq. B2 the correlation of Atkinson and
Merrick [134] for carbonaceous material was chosen.
3.5
ks C / 4511 T [W/m-K] (B.6)
Although this correlation is not specifically for beech char, it is considered a

reasonable approximation for chars with high carbon content. Based on a
composition-dependent correlation for subbituminous chars a true density of C
= 1279 kg/m3 the char was estimated [134] and applied with good agreement by
Piatkowski et al. [44]. The internal porosity of the particles was not considered
and is part of the error. Since the sensitivity of the effective thermal
conductivity to the solid thermal conductivity ks is very small, no further
attention was paid to this.
Since no values for emissivity of beech char could be found, the emissivity
was assumed to be 0.92, the same as the one from straw char [135].
Appendix C:
Numerical Implementation of the Heat and Mass
Transfer Model
The numerical implementation of the two-phase heat and mass transfer model
described in Section 8.2 is described in more detail here. First, some definitions
and a description of the discretization are given. Then, the discretization of the
steady-state governing equations for mass, species, and energy in fluid and
solid, which are solved iteratively using the cell centered finite volume method
(FVM), is explained. Finally, the solution strategy is illustrated.
C.1 Definitions
Porosity – The porosity of the domain is defined as the fraction of the fluid
volume Vf to the total volume of the medium Vmed comprising fluid and solid
phases.
Vf / Vmed (C.1)
Velocity – The superficial or Darcy velocity of the fluid u is defined as
u Vf / A (C.2)
where Vf is the volumetric fluid flow rate and A is the cross-sectional area.
Molar mass – The average molar mass of the fluid is defined as
M wi Mi (C.3)
i
140 Appendix C: Numerical Implementation
where wi and Mi are the mole fraction and molar mass of species i.
Mass fraction – The mass fraction of species k in dependence of the mole

fraction yk is defined as
yk M k wk M k
wk (C.4)
XiMi M
i
Density – The density of the fluid is computed using the ideal gas law.
ptot M
(C.5)
R Tf
where ptot is the total system pressure, R is the ideal gas constant, and Tf is the
fluid temperature.
Fluid temperature – The fluid temperature Tf and the sensible energy of the
mixture es are related via
RT0
es wk hs,k Tf (C.6)
k M
where hs,k is the sensible enthalpy of species k and T0 the reference temperature
of 298 K.
C.2 Discretization
The RPC section of the reactor, represented as a cylinder with diameter dtube
and height htube, was two-dimensionally discretized in cylindrical coordinates
into Nz axial by Nr radial axisymmetric finite volumes V as shown in
Figure C-1. In axial direction, the discretization was done uniformly with cell
heights z. In radial direction, the mesh was refined towards the wall to capture
C.2 Discretization 141
the large temperature and species concentration gradients. The radial cell size
was described as
ri 1.216 N r ri 1 for i = 1, 2, …, Nr –1 (C.7)
Figures C-2a and C-2b show the nomenclature in the grid-point cluster for the
finite volume cells away from the domain boundary and at the domain
boundary, respectively. P represents the cell center at which the cell properties
such as Dk, , and keff, are evaluated. The neighboring cell centers or the
domain boundary points in the north, east, south, and west are represented by
N, E, S, and W, respectively. The cell boundaries are named n, e, s, and w. The
cell centers are half way between the cell boundaries, i.e. re = rw and rn =
rs.
(1,0)
inlet (nr,1)
r
(0,1)
Δz
z (1,1)
...
(i, j) (i+1, j)
centerline
wall
Δz
(i, j+1)
Δz
(1, nz)
outlet
(nr, nz)
Δr1 Δr2 Δr3 ... Δr

nr
Figure C-1: Discretization of the RPC section of the reactor in the z-r
cylindrical coordinate system into Nr by Nz axisymmetric finite volumes V.
a) N
δzN
n
w e
δzn
W P E
δzs
s
δzS
δrw δre
S
δrW δrE
b) N
n δzN
w e
δzn
W P
E
δzS
δzs
s S
δrw δre
δrW δrE
Figure C-2: Grid-point cluster for the finite volume cell V for a) cells away
from the domain boundary and b) cells at the domain boundary (south east
corner (Nr,Nz) showed). P is the cell center. The neighboring cell centers or the
domain boundary points are labeled N, E, S, and W. The cell boundaries are
named n, e, s, and w. re = rw and rn = rs.
The property and variable values at the cell boundaries were generally
determined in the following way.
x fx P 1 fx X (C.8)
where x = n, e, s, and w, X = N, E, S, and W, and

C.3 Determination of the Flow Field 143
rx
fx 1 (C.9)
rX
The diffusion coefficients Dk and the thermal conductivities k were computed

as
1
Dk ,x (C.10)
fx 1 fx
Dk ,X Dk ,P
1
kx (C.11)
fx 1 fx
kX kP
The derivatives of the variables at the cell boundaries were defied as
E P P W
and (C.12)
r e rE r w rW
S P P N
and (C.13)
z s zS z n zN
C.3 Determination of the Flow Field
The mass conservation equation introduced in Eq. 8.3 given by
1
r ur uz Sk (C.14)
r r z k
was used to compute the full velocity field. Equation C.14 was integrated over
each control volume
zs 2 re
V(i , j ) dV rdrd dz (C.15)

V (i , j ) zn 0 rw
(i , j )
and simplified to
re2 rw2
z r ur e
r ur w
uz s
uz n
2 (i , j )
2 2
(C.16)
re r
w
z Sk
2 k (i , j )
where z = zs – zn and Sk is the cell average of the mass source of species k.

Equation C.16 is then summed up over all Nr control volumes in radial
direction while using the relations
u r , w ,( i 1, j ) u r ,e,( i , j ) (C.17)
w,(i 1, j ) e,(i , j )
(C.18)
and the boundary conditions of no flow in radial direction at the wall r = rmax
and center r = 0
ur ,w,(1, j ) ur ,e,( Nr , j ) 0 (C.19)
This leads to
nr
re2 rw2 nr
re2 rw2
uz s
uz n
z Sk (C.20)
i 1 2 (i , j ) i 1 2 k (i , j )
Applying the assumption that the axial velocity has no radial dependence
uz (r , z ) uz ( z) (C.21)
allows solving for the axial velocity at the south side of each control volume.
Nr
re2 rw2 uz n
z Sk
i 1 k (i , j )
u z ,s,j nr
(C.22)
2 2
r e r w s
(i , j )
i 1
C.4 Species Equation 145
Using the relation
u z ,n ,( i , j 1) u z ,s,( i , j ) (C.23)
and the defined inlet velocity uz,n,(i,1), uz,s and uz,n were then computed for the
whole domain in the main flow direction i.e. starting from the north going
south. To determine ur, Eq. C.16 was solved for ur,e
re2 rw2
z Sk uz uz z r ur
2 k
n s w
ur ,e,(i , j ) (C.24)
z r e
(i , j )
Eq. C.24 is then evaluated for the whole domain starting from west going east
using the ur,w,(1,j) from boundary condition at the wall shown in Eq. C.19 to start
with.
C.4 Species Equation
The conservation equations for the species k = 1, 2, …, Nk were as introduced

in Eq. 8.4 as
1
r ur vrC wk uz vzC wk
r r z
(C.25)
1 M k yk M k yk
r Deff,k , Deff,k , Sk
r r M r z M z
with the correction velocities
M k yk
vrC Deff,k , (C.26)
k M r
M k yk
vzC Deff,k , (C.27)
k M z
The derivatives of the mole fractions yk on the right hand side of Eq. C.25 are
expressed in the following way.
M k yk wk wi
Mwk (C.28)
M r r r i Mi
M k yk wk wi
Mwk (C.29)
M z z z i Mi
Equations C.28 and C.29 are then substituted into Eq. C.25. In a next step, the
resulting equation is rearranged and integrated over each control volume V
defined in Eq. C.15. The derivatives of the mass fractions wk at the cell
boundaries are expressed as described in Eqs. C.12 and C.13. The mass
fractions at the cell boundaries are evaluated as shown in Eq. C.8. This tedious
procedure leads to
wi
re z e ur vrC Deff,k , M f e wk ,P 1 f e wk ,E
r i Mi e
wi
rw z w ur vrC Deff,k , M f w wk ,P 1 f w wk ,W
r i Mi w
re2 rw2 wi
s uz vzC Deff,k , M f s wk ,P 1 fs wk ,S
2 z i Mi s
re2 rw2 wi
n uz vzC Deff,k , M f n wk ,P 1 f n wk ,N (C.30)
2 z i Mi n
wk ,E wk ,P wk ,P wk ,W
re z e Deff,k , ,e rw z w Deff,k , ,w
rE rW
re
2
r 2
w
wk ,S wk ,P r
e
2
r2
w
wk ,P wk ,N
s Deff,k , ,s n Deff,k , ,n
2 zS 2 zN
re2 rw2
z Sk ,P
2
where z = zs – zn.
C.4 Species Equation 147
The species source term Sk,P introduced in Eq. 8.18 was linearized around
wk,P*,
the value of the previous iteration of wk,P, and expressed in the following
form.
dSk
Sk ,P Sk ( wk ,P ) Sk ( wk*,P ) wk ,P Sk ,const,P Sk ,prop,P wk ,P (C.31)
dwk wk* ,P
Inserting Eq. C.31 into Eq. C.30 and denoting the convection terms as Fn, Fe,
Fs, and Fw, and the diffusion terms as Dn, De, Ds, and Dw, leads to
Fe f e wk ,P 1 f e wk ,E Fw f w wk ,P 1 f w wk ,W
Fs f s wk ,P 1 f s wk ,S Fn f n wk ,P 1 f n wk ,N
De wk ,E wk ,P Dw wk ,P wk ,W (C.32)
Ds wk ,S wk ,P Dn wk ,P wk ,N
2 2
r e rw
z S k ,P,const S k ,P,prop wk ,P
2
Eq. C.32 can now be brought into the standard form used by Patankar [136] and
obtain the coefficients aE, aW, aS, and aN.
aP wk ,P aE wk ,E a W wk ,W aS wk ,S a N wk ,N b (C.33)
To have a mathematically more stable system of equations, the hybrid scheme

[119] is applied and the coefficients aE, aW, aS, and aN are thus modified to
aE Fe , De Fe 1 fe , 0 (C.34)
aW Fw , Dw Fw 1 fw , 0 (C.35)
aS Fs , Ds Fs 1 fs , 0 (C.36)
aN Fn , Dn Fn 1 fn , 0 (C.37)
where A, B, C denotes the greatest of A, B, and C.

Equation C.33 is applied for all cells (i,j) where i = 1, 2, …, Nr and j =1, 2,
…, Nz. For the boundary nodes (i,0) for i = 1, 2, …, Nr and (0,j) for j =1, 2, …,
Nz, additional equations were defined to account for the boundary conditions.
For a defined boundary value wk,boundary the additional equation is
wk ,P wk ,boundary (C.38)
For a defined gradient boundary condition, e.g. at the south boundary, the
additional equation is
wk
wk ,P wk ,N zN (C.39)
z boundary
For each species, the resulting linear system of equations, comprising

Eqs. C.33, C.38, and C.39, was written into matrix form. The solutions were
then found using MATLAB (The MathWorks, Inc.), by either permutation and
forward solving, or by using the unsymmetric-pattern multifrontal method
described by Davis [120].
C.5 Fluid Phase Energy Equation
The fluid phase energy equation introduced in Eq. 8.7 was defined as
1
r u r es u z es
r r z
1 Mk yk
r hs,k Deff,k , (C.40)
r r k M r
Mk yk
hs,k Deff,k , hsf A0 Ts Tf
z k M z
Integrating Eq. C.40 over each control volume V defined in Eq. C.15,
expressing the derivatives of the mass fraction at the cell boundaries as
C.5 Fluid Phase Energy Equation 149
described in Eqs. C.12 and C.13 and the sensible energy at the boundaries as
shown in Eq. C.8 leads to
re z e ur ,e f e es,P 1 f e es,E
rw z u
w r ,w f w es,P 1 f w es,W
2 2
re rw
u
s z ,s fs es,P 1 fs es,S
2
re2 rw2
n uz ,n f n es,P 1 f n es,N
2
Mk yk ,E yk ,P
re z e hs , k ,e Deff,k , (C.40)
k M e rE
Mk yk ,P yk ,W
rw z w hs , k ,w Deff,k ,
k M w rW
re
2
r2
w Mk yk ,S yk ,P
s hs , k ,s Deff,k ,
2 k M s zS
re
2
r2
w Mk yk ,P yk ,N
n hs , k ,n Deff,k ,
2 k M n zN
2 2
r
e rw
z hsf A0 Ts,P Tf,P
2
where z = zs – zn. The temperature difference Ts,P – Tf,P in the term for the heat
transfer between solid and fluid can be expressed as a function of the sensible
energy es. The temperature difference is thus expanded as
Ts,P Tf,P Ts,P Tf,P* Tf,P* Tf,P (C.41)
where the symbol T*f,P is used to denote the previous iteration value of Tf,P.
Using the definition of the sensible energy
RT0 T RT0
es hs Cv dT (C.42)
M T0 M
where hs is the sensible enthalpy and Cv is the heat capacity of the mixture, we
can formulate a sensible energy difference
*
*
Tf,P RT0 Tf,P RT0 *
Tf,P
es,P es,P Cv dT Cv dT Cv dT (C.43)
T0 M T0 M Tf,P
With the approximation that Cv is constant within the small temperature

difference T*f,P – Tf,P, Eq. C.43 can be simplified and rearranged to
*
es,P es,P
Tf,P* Tf,P (C.44)
Cv (Tf,P* )
The term T*f,P – Tf,P is now expressed as a function of the sensible energy and
can be inserted into Eq. C.41 yielding
*
es,P es,P
Ts,P Tf,P Ts,P Tf,P* (C.45)
Cv (Tf*,P )
Substituting Eq. C.45 into Eq. C.40 and expressing the convection terms as Fn,
Fe, Fs, and Fw leads to
Fe f e es,P 1 f e es,E Fw f w es,P 1 f w es,W

Fs f s es,P 1 f s es,S Fn f n es,P 1 f n es,N
Mk yk ,E yk ,P
re z e hs , k ,e Deff,k ,
k M e rE
Mk yk ,P yk ,W
rw z w hs , k ,w Deff,k , (C.46)
k M w rW
re2 rw2 Mk yk ,S yk ,P
s hs , k ,s Deff,k ,
2 k M s zS
re
2
r2
w Mk yk ,P yk ,N
n hs ,k ,n Deff,k ,
2 k M n zN
*
re2 rw2 e s,P 1
z hsf A0 Ts,P Tf,P* *
es,P
2 Cv (T ) f,P Cv (Tf,P* )
The terms Dn, De, Ds, and Dw applied in the previous section are 0 here as no
diffusion terms are present. We can again bring the equation into the standard
form and determine the coefficients aE, aW, aS, and aN.
C.5 Fluid Phase Energy Equation 151
a P es,P a E es,E a W es,W a N es,N aS es,S b (C.47)
To obtain a mathematically more stable system of equations, the hybrid scheme

[119] is applied and the coefficients aE, aW, aS, and aN are thus modified to
aE Fe , De Fe 1 fe , 0 (C.48)
aW Fw , Dw Fw 1 fw , 0 (C.49)
aS Fs , Ds Fs 1 fs , 0 (C.50)
aN Fn , Dn Fn 1 fn , 0 (C.51)
Equation C.47 is applied for all cells (i,j) where i = 1, 2, …, Nr and j =1, 2, …,
Nz. For the boundary nodes (i,0) for i = 1, 2, …, Nr and (0,j) for j =1, 2, …, Nz,
additional equations were defined to account for the boundary conditions. For a
defined boundary value the additional equation is
es,P es,boundary (C.52)
For a defined gradient, shown for the south boundary here, the additional
equation is of the form
es,P es,N zN es,boundary (C.53)

z
The linear system of equations, comprising the equations for the cells and the
ones for the boundary nodes, was then written into matrix form and solved as
described in the previous section.
C.6 Solid Phase Energy Equation
The energy equation of the solid phase introduced in Eq. 8.8 is given by
1 T Ts
0 rkeff s keff hsf A0 Tf Ts Qreact (C.54)
r r r z z
Integrating Eq. C.54 over all control volumes V and applying the solid
temperature Ts derivatives at the cell boundaries as described in Eqs. C.12 and
C.13, results in
Ts,E Ts,P Ts,P Ts,W

0 z re keff,e z re keff,w
rE rW
re
2 2
r
w
Ts,S Ts,P re
2 2
r
w
Ts,P Ts,N
keff,s keff,n (C.55)
2 zS 2 zN
re2 rw2
z hsf A0 Tf,P Ts,P Qreact,P
2
where z = zs – zn.
The chemical reaction term is linearized around Ts,P*, the initial value or the
value of the previous iteration of Ts,P.
Qreact,P Qreact Ts,P

dQreact (C.56)
Qreact Ts,P* Ts,P Qreact,const,P Qreact,prop,P Ts,P
dTs *
Ts,P
Equation C.56 is inserted into Eq. C.55, that is then brought into the standard
form and the coefficients aE, aW, aS, and aN are determined.
aP Ts,P aE Ts,E a W Ts,W aSTs,S a N Ts,N b (C.57)
As the solid phase energy equation has no convective term, this is already the
final form. Applying the hybrid scheme would lead the same form.
Eq. C.57 is applied for the cells (i,j) where i = 1, 2, …, Nr and j =1, 2, …,
Nz. For the boundary nodes (i,0) for i = 1, 2, …, Nr and (0,j) for j =1, 2, …, Nz,
C.7 Solution Strategy 153
additional equations were defined to take the boundary conditions into account.
For a constant boundary temperature Ts,boundary the additional equation is
Ts,P Ts,boundary (C.58)
For a defined gradient, e.g. at the south boundary, the additional equation is of
the form
Ts,P Ts,N zN Ts,boundary (C.59)

z
The linear system of equations, comprising Eqs. C.57, C.58, and C.59, is then
written into matrix form and solved as described in Section C.4.
C.7 Solution Strategy
The solution to the set of governing equations was found iteratively by solving
them in the order shown in Figure C-3. At first, the initial conditions were used
to determine the variable properties such as keff, Deff,k, , hsf, etc. In a next step,
the full flow field (ur, uz) was computed as described in Section C.3 and the
properties were updated. Then, the species conservation equations, the fluid
phase energy equation and the solid phase energy equation were solved
sequentially while updating the properties after each equation. This yields wk ,
es, and Ts.
The values of the dependent variable = ur, uz, Ts, wk, or es were then used
to compute the residual error to check for convergence. The error of the i-th
iteration step Ei was computed by taking the root mean square of the step
change of the dependent variable , the quantity the governing equation was
solved for, over all N cells.
N
1 2
Ei n,i n,i 1 (C.60)
N n 1
The residual error was then defined as the error of the i-th iteration normalized
by the error of first iteration.
Ri Ei E1 (C.61)
The described procedure was repeated until the convergence criterion for the
residuals Ri < 10 7 was reached for all dependent variables ur, uz, Ts, wk, and es.
Once reached, the solution was found and the result returned.
boundary conditions, initial conditions, flow profile
compute variable properties
compute full flow field with mass conservation equation (ur, uz)
update variable properties
solve species conservation equations (wk )

solve fluid phase energy conservation equation (es)

solve solid phase energy equation (Ts)

all Ri < 10-7 ?

no
yes
output results (ur, uz, wk , es, Ts)
Figure C-3: Solution strategy.

List of Figures
Figure 1-1: Solar-driven steam gasification process flow sheet......................3

Figure 2-1: Primary and secondary reactions during pyrolysis of a
biomass particle. Adapted from ref. [14]. ..................................... 8
Figure 2-2: Biomass gasifier types with main flow streams: a) updraft
(countercurrent), b) downdraft (concurrent), c) bubbling
fluidized bed, d) circulating fluidized bed, e) dual fluidized
bed, and f) entrained-flow. Colors indicate relative
temperatures (red = hot, yellow = medium, gray = cold).
Adapted from ref. [5]. ................................................................ 13
Figure 2-3: Solar tower concentrating systems: a) solar reactor on top
of tower and b) beam down design with secondary
hyperbolic mirror on tower and solar reactor at ground.
Adapted from ref. [31]. .............................................................. 16
Figure 2-4: Solar gasifier concepts: a) tower-mounted drop-tube [41]
and b) packed bed for a beam-down configuration [40].
Arrow indicates where the concentrated solar radiation
enters the reactor. .......................................................................19
Figure 3-1: Biomass feedstocks used: a) Brazilian sugarcane bagasse
and b) beech char........................................................................ 21
Figure 3-2: Particle size distribution functions of ground and sieved
bagasse, and beech char analyzed with LA-950 analyzer
(HORIBA).................................................................................. 24
stoichiometric system of Eq. 4.1 for bagasse (CH1.665O0.663)
at 1 bar. ....................................................................................... 26
stoichiometric system of Eq. 4.1 for bagasse (CH1.665O0.663)
at 10 bar. ..................................................................................... 26
Figure 4-3: Reaction enthalpy hR for the system CHxOy + (1–y) H2O
(reactants fed at 298 K and 1 bar, products obtained at
156 List of Figures
chemical equilibrium) and LHV of product gas as a

function of temperature. ............................................................. 27
Figure 4-4: Equilibrium composition as a function of adiabatic
temperature for the system in Eqs. 4.3 and 4.4 with 0
0.33. ........................................................................................... 29
Figure 4-5: Molar ratios of H2/CO and CO2/CO at equilibrium over
temperature for solar and autothermal gasification. ................... 29
Figure 4-6: Flow diagram applied in the first and second law analyses
for the solar (left) and autothermal gasification (right). ............. 30
Figure 5-1: Thermogravimetric balance configuration. ................................ 40
Figure 5-2: Apparent carbon conversion rate for the experimental
measurements vs. the model for the temperature range
1173 1473 K and 25 75%vol steam concentration. ................... 47
Figure 5-3: Typical experimental results and linear fits for –ln(1–X ) vs.
time for selected experiments with a 75%vol steam/Ar
mixture. ...................................................................................... 47
Figure 6-1: Schematic of the laboratory-scale biomass gasification
apparatus including the primary components and flows. ........... 51
Figure 6-2: Relative mole fractions (dry, N2 and Ar free basis) of the
product gases and the equilibrium composition (lines
without markers) ........................................................................ 54
Figure 7-1: Schematic of the solar cavity-receiver/reactor configuration
with thermocouple locations and blown up detail of the
reactor tube showing the fast pyrolysis drop-tube zone and
the trickle-bed char gasification zone (RPC = reticulate
porous ceramic, CPC = compound parabolic concentrator). ..... 58
Figure 7-2: Structure packing made of 10 ppi SiSiC reticulate porous
ceramic (RPC) foam installed in the solar gasifier. Image
from [88]. ................................................................................... 59
Figure 7-3: Schematic of the solar-driven biomass gasifier, including
the high-flux solar simulator and the auxiliary components
and flows.................................................................................... 61
Figure 7-4: Input power (Qsolar) and reactor temperatures during a
typical experiment: steam on, biomass feed on, and
biomass feed off. ................................................................... 63
157
Figure 7-5: Molar flow rates of product gases as a function of time

during a typical experimental run: steam on, biomass
feed on, and biomass feed off. ............................................... 64
Figure 7-6: Average gas concentrations for a) H2, CO, CO2, and CH4,
and b) C2H2, C2H4, and C2H6 calculated on H2O, N2, and Ar
free basis for the investigated configurations: A = 50 mm
tall RPC, B = 100 mm tall RPC, PP = free-fall pure
pyrolysis, and SG = free-fall steam-based gasification.
Error bars indicate the span between the minimum and
maximum values. ....................................................................... 68
Figure 7-7: Carbon conversion and upgrade factors for investigated
configurations: A = 50 mm tall RPC, B = 100 mm tall RPC,
PP = free-fall pure pyrolysis, SG = free-fall steam
gasification. Error bars illustrate the span between the
minimum and maximum values. ................................................ 70
Figure 7-8: The 50 mm tall RPC after one of the A experiments. The
particles are uniformly distributed throughout the entire
RPC. ........................................................................................... 71
Figure 7-9: Overview temperature measurements and heat losses:
radiation and convection through the aperture Qrad and
Qconv, conduction through the cavity walls Qwalls and along
the reactor tube towards the top Qcond,top and the bottom
Qcond,bottom. ................................................................................... 72
Figure 7-10: Convection losses through aperture: Air temperature
distribution in cavity, aperture, and surrounding. .......................76
Figure 7-11: Energy losses of solar reactor via radiation and convection
through the aperture Qrad and Qconv, conduction through the
walls Qwalls and along the tube Qcond,top and Qcond,bottom. .............. 79
Figure 8-1: Schematic of the experimental setup. ......................................... 84
Figure 8-2: Photograph of the SiSiC reticulate porous ceramic (RPC)
with 10 pores per inch (ppi) used as packing in the tubular
gasifier. .......................................................................................85
Figure 8-3: Schematic (not to scale) of the main reactor tube, RPC, and
supporting tube assembly indicating thermocouple
locations. Thermocouples TW6 and TW7 were movable along
the tube wall. ..............................................................................86
158 List of Figures
Figure 8-4: Temporal product gas flow rates over the course of a typical
experiment. ................................................................................ 88
Figure 8-5: Carbon conversion (XC) as function of furnace set-point
temperature (Tfurnace), steam concentration, and gasifier
configuration (RPC versus no RPC). ......................................... 89
Figure 8-6: Discretization for the radiative exchange between the
RPC’s top and bottom surfaces to their surroundings. Tube
sections above and below RPC: 100 segments each, RPC
surfaces: 64 rings, flanges: individual surfaces. The smaller
tube diameter below the RPC is due to the supporting tube. ..... 95
Figure 8-7: Results from code verification of the diffusion terms in the
species equation, a) error norm E2 and b) actual order of
accuracy qa. .............................................................................. 100
Figure 8-8: Comparison between experimental (Ts,center,exp - filled
markers) and simulated centerline solid temperature
profiles (Ts,center,sim - solid lines) for a) non-reacting
conditions, and b) reacting conditions. The open markers
indicate the experimental wall temperatures (Twall,exp); the
dashed lines represent the boundary temperatures (Twall,bc)
applied in the simulations. ....................................................... 102
Figure 8-9: Comparison of experimentally determined and simulated
molar flow rates of gasified carbon.......................................... 103
Figure 8-10: Simulated radial profiles of a) solid temperature, b) steam
mole fraction, and c) gasification rate at z = 0.5 m for l =
0.2 and Twall = 1173, 1223, and 1273 K. .................................. 105
Figure 8-11: Sensitivity of a) the syngas production rate ( C,syngas,RPC) and
b) the temperature difference between wall and centerline at
the outlet of the domain Ts,outlet = Ts,oulet,wall – Ts,outlet,center to
changes in nominal pore diameter (dnom), solid thermal
conductivity (ks), RPC porosity ( RPC), and particle loading
(l) for Twall = 1173, 1223, and 1273 K. Reference case: dnom
= 2.54 mm, ks = kSiC, RPC = 89%, and l = 0.2. ......................... 107
Figure 8-12: Simulated radial profiles of a) solid temperature (Ts), b)
steam mole fraction (yH O), and c) gasification rate ( '''C) in
moving bed (MB) and RPC (l = 0.1, 0.2, and 0.4) at z =
0.5 m for Twall = 1223 K. .......................................................... 110
159
Figure 8-13: Effective thermal conductivity (keff) over temperature for

the moving bed (MB) with N2 filled voids and the RPC
configuration at various loadings (l). ....................................... 111
Figure 8-14: Contribution of radiation term krad on the effective thermal
conductivity keff for RPC configuration.................................... 112
Figure 8-15: Simulation results for ratio of trickle-bed to moving-bed
gasification rates as function of the particle loading in the
RPC (l) and wall temperatures of Twall = 1173, 1223, and
1273 K. ..................................................................................... 113
Figure A-1: Spray dried microalgae (chlorella vulgaris). ............................ 123
Figure A-2: Particle size distribution of spray-dried microalgae
analyzed with LA-950 analyzer (HORIBA)............................. 124
Figure A-3: Schematic of the solar entrained flow gasifier including
solar simulator and the primary components and flows. .......... 125
Figure A-4: Molar flow rate of the product gases during a typical
experiment. ............................................................................... 128
Figure A-5: Product gas compositions for different volumetric steam
concentrations. Error bars indicate the span between the
minimum and maximum values. .............................................. 130
Figure A-6: LHVsyngas for different volumetric steam concentrations.
Error bars indicate the span between the minimum and
maximum values. ..................................................................... 131
Figure A-7: Carbon conversion and upgrade factor for different
volumetric steam concentrations. Error bars indicate the
span between the minimum and maximum values. .................. 132
Figure A-8: Reactor efficiency in dependence of the feeding rate for
the three different steam concentrations. .................................. 133
Figure C-1: Discretization of the RPC section of the reactor in the z-r
cylindrical coordinate system into Nr by Nz axisymmetric
finite volumes V. .................................................................... 141
Figure C-2: Grid-point cluster for the finite volume cell V for a) cells
away from the domain boundary and b) cells at the domain
boundary (south east corner (Nr,Nz) showed). P is the cell
center. The neighboring cell centers or the domain
boundary points are labeled N, E, S, and W. The cell
boundaries are named n, e, s, and w. re = rw and rn = rs. ... 142
160 List of Figures
Figure C-3: Solution strategy. ..................................................................... 154

List of Tables
Table 2-1: Typical product gas composition for different biomass

gasifiers. ..................................................................................... 14
Table 2-2: Existing solar gasifier concepts and feedstocks used. ................17
Table 3-1: Physical and chemical properties of bagasse and beech
char: Ultimate and proximate analyses, heating value and
mean particle size. ...................................................................... 23
Table 4-1: Energy and exergy analysis for solar and autothermal
gasification. Values for power refer to a biomass feed rate
of CHxOy = 1 g/s. ........................................................................ 34
Table 5-1: Elemental composition of raw and rapidly pyrolyzed
bagasse particles; C, H, N determined with CHN-900, O
with RO-478, and S with CHNS-932 (all LECO
Corporation, St. Joseph, MI). .....................................................41
Table 5-2: Apparent kinetic parameters for Eq. 5.16 for the steam
gasification of rapidly pyrolyzed bagasse. ................................. 46
Table 6-1: Summary of the experimental results obtained for the
gasification of bagasse in the electrically heated two-zone
reactor ........................................................................................ 53
Table 7-1: Summary of the experimental results (A = 50 mm tall RPC,
B = 100 mm tall RPC, PP = free-fall pure pyrolysis, SG =
free-fall steam-based gasification, all values are on a H2O,
N2 and Ar free basis). ................................................................. 65
Table 7-2: Summary of the statistical analysis for the effects of reactor
configurations: for comparison X vs. Y “>” or “<”
designate statistically higher or lower mean of a response
for the configuration X; “~” designates no statistically
significant difference. ................................................................. 69
Table 7-3: Energy flows in the solar reactor. ............................................... 80
Table 8-1: Thermocouple positions: W = wall, C = centerline. ...................86
Table 8-2: Heat- and mass-transfer property correlations. .......................... 96
162 List of Tables
Table 8-3: Parameters used in order verification study (all in SI units). ... 100
Table 8-4: Simulation results for a 1 m tall / 0.06 m inner diameter
domain with dnom = 2.54 mm, ks = kSiC, RPC = 89%, f,in =
2 g/s, Tf,in = 873 K, and l = 0.2. ................................................ 104
Table A-1: Physical and chemical properties of dry microalgae:
Ultimate and proximate analyses, heating value and mean
particle size. ............................................................................. 123
Table A-2: Summary of the experimental results (integral values
averaged over all replicates). ................................................... 129
References
[1] IEA, World Energy Outlook 2013: IEA, 2013.
[2] S. Solomon, I. P. o. C. Change, and I. P. o. C. C. W. G. I., Climate

Change 2007 - The Physical Science Basis: Working Group I
Contribution to the Fourth Assessment Report of the IPCC: Cambridge
University Press, 2007.
[3] A. Steinfeld and R. Palumbo, "Solar Thermochemical Process

Technology", in Encyclopedia of Physical Science and Technology
(Third Edition), A. M. Robert, Ed., ed New York: Academic Press,
2003, pp. 237-256.
[4] A. Steinfeld, "Thermochemical production of syngas using concentrated

solar energy", Annual Review of Heat Transfer, vol. 15(15), 2012.
[5] A. V. Bridgwater, "The technical and economic feasibility of biomass

gasification for power generation", Fuel, vol. 74(5), pp. 631-653, 1995.
[6] P. De Filippis, C. Borgianni, M. Paolucci, and F. Pochetti, "Gasification

process of Cuban bagasse in a two-stage reactor", Biomass and
Bioenergy, vol. 27(3), pp. 247-252, 2004.
[7] E. O. Gómez, L. A. B. Cortez, E. S. Lora, C. G. Sanchez, and A. Bauen,

"Preliminary tests with a sugarcane bagasse fueled fluidized-bed air
gasifier", Energy Conversion and Management, vol. 40(2), pp. 205-214,
1999.
[8] M. Gabra, E. Pettersson, R. Backman, and B. Kjellström, "Evaluation of

cyclone gasifier performance for gasification of sugar cane residue –
Part 1: gasification of bagasse", Biomass and Bioenergy, vol. 21(5), pp.
351-369, 2001.
[9] L. F. Pellegrini and J. S. de Oliveira, "Exergy analysis of sugarcane

bagasse gasification", Energy, vol. 32(4), pp. 314-327, 2007.
164 References
[10] N. Piatkowski, C. Wieckert, A. W. Weimer, and A. Steinfeld, "Solar-

driven gasification of carbonaceous feedstock-a review", Energy &
Environmental Science, vol. 4(1), pp. 73-82, 2011.
[11] R. W. Taylor, R. Berjoan, and J. P. Coutures, "Solar gasification of

carbonaceous materials", Solar Energy, vol. 30(6), pp. 513-525, 1983.
[12] J. Corella, J. M. Toledo, and G. Molina, "A review on dual fluidized-

bed biomass gasifiers", Industrial & Engineering Chemistry Research,
vol. 46(21), pp. 6831-6839, 2007.
[13] P. McKendry, "Energy production from biomass (Part 1): Overview of

biomass", Bioresour Technol, vol. 83(1), pp. 37-46, 2002.
[14] P. Basu, Biomass gasification and pyrolysis: practical design and

theory. Burlington, MA: Academic Press, 2010.
[15] E. Rensfelt, G. Blomkvist, C. Ekström, S. Engström, B. Espents, and L.

Linnanki, "Basic gasification studies for development of biomass
medium-BTU gasification processes", presented at the Energy from
Biomass and Wastes, Washington, D.C., 1978.
[16] F. Shafizadeh, "Introduction to Pyrolysis of Biomass", in Specialists'

workshop on fast pyrolysis of biomass, Copper Mountain, CO, pp. 79-
103, 1980.
[17] C. Fushimi, K. Araki, Y. Yamaguchi, and A. Tsutsumi, "Effect of

Heating Rate on Steam Gasification of Biomass. 1. Reactivity of Char",
Industrial & Engineering Chemistry Research, vol. 42(17), pp. 3922-
3928, 2003.
[18] R. Zanzi, K. Sjöström, and E. Björnbom, "Rapid high-temperature

pyrolysis of biomass in a free-fall reactor", Fuel, vol. 75(5), pp. 545-550,
1996.
[19] F. Mermoud, S. Salvador, L. Van de Steene, and F. Golfier, "Influence

of the pyrolysis heating rate on the steam gasification rate of large wood
char particles", Fuel, vol. 85(10-11), pp. 1473-1482, 2006.
[20] I. Milosavljevic, V. Oja, and E. M. Suuberg, "Thermal effects in

cellulose pyrolysis: Relationship to char formation processes",
165
Industrial & Engineering Chemistry Research, vol. 35(3), pp. 653-662,

1996.
[21] J. F. Stubington and S. Aiman, "Pyrolysis kinetics of bagasse at high

heating rates", Energy & Fuels, vol. 8(1), pp. 194-203, 1994.
[22] Y. Zhang, S. Kajitani, M. Ashizawa, and K. Miura, "Peculiarities of

Rapid Pyrolysis of Biomass Covering Medium- and High-Temperature
Ranges", Energy & Fuels, vol. 20(6), pp. 2705-2712, 2006.
[23] Y. Zhang, S. Kajitani, M. Ashizawa, and Y. Oki, "Tar destruction and

coke formation during rapid pyrolysis and gasification of biomass in a
drop-tube furnace", Fuel, vol. 89(2), pp. 302-309, 2010.
[24] C. Higman and M. v. d. Burgt, Gasification, 2nd ed. Amsterdam ;

Boston: Gulf Professional Pub./Elsevier Science, 2008.
[25] L. Devi, K. J. Ptasinski, and F. J. J. G. Janssen, "A review of the

primary measures for tar elimination in biomass gasification processes",
Biomass & Bioenergy, vol. 24(2), pp. 125-140, 2003.
[26] H. Knoef and J. Ahrenfeldt, Handbook biomass gasification: AE

Enschede: btg, 2005.
[27] T. A. Milne, R. J. Evans, and N. Abatzaglou, "Biomass Gasifier ''Tars'':

Their Nature, Formation, and Conversion", National Renewable Energy
Laboratory, Golden, CO NREL/TP-570-25357, 1998.
[28] R. G. Graham and R. Bain, Biomass gasification hot-gas clean-up: IEA

biomass gasification working group, 1993.
[29] R. Evans and T. Milne, "Chemistry of Tar Formation and Maturation in

the Thermochemical Conversion of Biomass", in Developments in
Thermochemical Biomass Conversion, A. V. Bridgwater and D. G. B.
Boocock, Eds., ed: Springer Netherlands, 1997, pp. 803-816.
[30] H. Hofbauer, R. Rauch, G. Loeffler, S. Kaiser, E. Fercher, and H.

Tremmel, "Six years experience with the FICFB-gasification process",
in 12th European Conference and Technology Exhibition on Biomass
for Energy, Industry and Climate Protection, pp. 982-985, 2002.
166 References
[31] N. Piatkowski, "Solar driven steam gasification of carbonaceous

feedstocks: Feedstock characterization to pilot facility testing", Doctor
of Sciences, ETH Zurich, 2011.
[32] D. Trommer, F. Noembrini, M. Fasciana, D. Rodriguez, A. Morales, M.

Romero, and A. Steinfeld, "Hydrogen production by steam-gasification
of petroleum coke using concentrated solar power – I. Thermodynamic
and kinetic analyses", International Journal of Hydrogen Energy, vol.
30(6), pp. 605-618, 2005.
[33] R. Müller, P. von Zedtwitz, A. Wokaun, and A. Steinfeld, "Kinetic

investigation on steam gasification of charcoal under direct high-flux
irradiation", Chemical Engineering Science, vol. 58(22), pp. 5111-5119,
2003.
[34] P. von Zedtwitz and A. Steinfeld, "Steam-Gasification of Coal in a

Fluidized-Bed/Packed-Bed Reactor Exposed to Concentrated Thermal
Radiation – Modeling and Experimental Validation", Industrial &
Engineering Chemistry Research, vol. 44(11), pp. 3852-3861, 2005.
[35] A. Z'Graggen, P. Haueter, D. Trommer, M. Romero, J. De Jesus, and A.

Steinfeld, "Hydrogen production by steam-gasification of petroleum
coke using concentrated solar power – II Reactor design, testing, and
modeling", International Journal of Hydrogen Energy, vol. 31(6), pp.
797-811, 2006.
[36] D. W. Gregg, R. W. Taylor, J. H. Campbell, J. R. Taylor, and A. Cotton,

"Solar gasification of coal, activated carbon, coke and coal and biomass
mixtures", Solar Energy, vol. 25(4), pp. 353-364, 1980.
[37] J. P. Murray and E. A. Fletcher, "Reaction of steam with cellulose in a

fluidized bed using concentrated sunlight", Energy, vol. 19(10), pp.
1083-1098, 1994.
[38] P. Lichty, C. Perkins, B. Woodruff, C. Bingham, and A. Weimer,

"Rapid High Temperature Solar Thermal Biomass Gasification in a
Prototype Cavity Reactor", Journal of Solar Energy Engineering, vol.
132(1), pp. 0110121-0110127, 2010.
[39] T. Melchior, C. Perkins, P. Lichty, A. W. Weimer, and A. Steinfeld,

"Solar-driven biochar gasification in a particle-flow reactor", Chemical
167
Engineering and Processing: Process Intensification, vol. 48(8), pp.

1279-1287, 2009.
[40] N. Piatkowski, C. Wieckert, and A. Steinfeld, "Experimental

investigation of a packed-bed solar reactor for the steam-gasification of
carbonaceous feedstocks", Fuel Processing Technology, vol. 90(3), pp.
360-366, 2009.
[41] C. Perkins, Z. Jovanovic, S. Strand, C. Hilton, and D. Kelley, "Systems

and methods for an indirect radiation driven gasifier reactor & receiver
configuration", US 2010/0303692 A1 Patent US 2010/0303692 A1,
2010.
[42] A. Z'Graggen, P. Haueter, G. Maag, A. Vidal, M. Romero, and A.

Steinfeld, "Hydrogen production by steam-gasification of petroleum
coke using concentrated solar power – III. Reactor experimentation with
slurry feeding", International Journal of Hydrogen Energy, vol. 32(8),
pp. 992-996, 2007.
[43] A. Z'Graggen, P. Haueter, G. Maag, M. Romero, and A. Steinfeld,

"Hydrogen production by steam-gasification of carbonaceous materials
using concentrated solar energy – IV. Reactor experimentation with
vacuum residue", International Journal of Hydrogen Energy, vol. 33(2),
pp. 679-684, 2008.
[44] N. Piatkowski and A. Steinfeld, "Solar-Driven Coal Gasification in a

Thermally Irradiated Packed-Bed Reactor", Energy & Fuels, vol. 22(3),
pp. 2043-2052, 2008.
[45] J. Lede, F. Verzaro, B. Antoine, and J. Villermaux, "Flash pyrolysis of

wood in a cyclone reactor", Chemical Engineering and Processing:
Process Intensification, vol. 20(6), pp. 309-317, 1986.
[46] T. Melchior and A. Steinfeld, "Radiative Transfer Within a Cylindrical

Cavity With Diffusely/Specularly Reflecting Inner Walls Containing an
Array of Tubular Absorbers", Journal of Solar Energy Engineering, vol.
130(2), pp. 0210131-0210137, 2008.
[47] T. Melchior, C. Perkins, A. W. Weimer, and A. Steinfeld, "A cavity-

receiver containing a tubular absorber for high-temperature
thermochemical processing using concentrated solar energy",
168 References
International Journal of Thermal Sciences, vol. 47(11), pp. 1496-1503,

2008.
[48] A. Z'Graggen and A. Steinfeld, "Hydrogen production by steam-

gasification of carbonaceous materials using concentrated solar energy –
V. Reactor modeling, optimization, and scale-up", International Journal
of Hydrogen Energy, vol. 33(20), pp. 5484-5492, 2008.
[49] K. Svoboda, M. Poho elý, M. Hartman, and J. Martinec, "Pretreatment

and feeding of biomass for pressurized entrained flow gasification",
Fuel Processing Technology, vol. 90(5), pp. 629-635, 2009.
[50] G. Maag and A. Steinfeld, "Design of a 10 MW Particle-Flow Reactor

for Syngas Production by Steam-Gasification of Carbonaceous
Feedstock Using Concentrated Solar Energy", Energy & Fuels, vol.
24(12), pp. 6540-6547, 2010.
[51] A. Steinfeld, "Solar hydrogen production via a two-step water-splitting

thermochemical cycle based on Zn/ZnO redox reactions", International
Journal of Hydrogen Energy, vol. 27(6), pp. 611-619, 2002.
[52] A. Meier, N. Gremaud, and A. Steinfeld, "Economic evaluation of the

industrial solar production of lime", Energy Conversion and
Management, vol. 46(6), pp. 905-926, 2005.
[53] R. Pitz-Paal, N. B. Botero, and A. Steinfeld, "Heliostat field layout

optimization for high-temperature solar thermochemical processing",
Solar Energy, vol. 85(2), pp. 334-343, 2011.
[54] S. Q. Turn, B. M. Jenkins, L. A. Jakeway, L. G. Blevins, R. B. Williams,

G. Rubenstein, and C. M. Kinoshita, "Test results from sugar cane
bagasse and high fiber cane co-fired with fossil fuels", Biomass and
Bioenergy, vol. 30(6), pp. 565-574, 2006.
[55] R. P. Beeharry, "Extended sugarcane biomass utilisation for exportable

electricity production in mauritius", Biomass & Bioenergy, vol. 11(6),
pp. 441-449, 1996.
[56] A. Pandey, C. R. Soccol, P. Nigam, and V. T. Soccol, "Biotechnological

potential of agro-industrial residues. I: sugarcane bagasse", Bioresource
Technology, vol. 74(1), pp. 69-80, 2000.
169
[57] FAOSTAT. (2011, 09/21/2011). FAOSTAT database. Available:

http://faostat.fao.org/site/567/default.aspx#ancor
[58] F. Teixeira, "Bagaço de cana-de-açúcar na alimentação de bovinos",

Revista eletrónica de veterinária, vol. 8(6), pp. 1695-1704, 2007.
[59] A. Roine, "Outokumpu HSC Chemistry", ed. Pori, Finland: Outokumpu

Research, 1997.
[60] A. Steinfeld, C. Larson, R. Palumbo, and M. Foley, "Thermodynamic

analysis of the co-production of zinc and synthesis gas using solar
process heat", Energy, vol. 21(3), pp. 205-222, 1996.
[61] E. A. Fletcher and R. L. Moen, "Hydrogen-and Oxygen from Water",

Science, vol. 197(4308), pp. 1050-1056, 1977.
[62] A. Bejan, Entropy generation through heat and fluid flow: Wiley New
York, 1982.
[63] N. Sato, Chemical energy and exergy: An introduction to chemical

thermodynamics for engineers. Amsterdam; Boston: Elsevier, 2004.
[64] M. J. Moran and H. N. Shapiro, Fundamentals of engineering

thermodynamics: SI version, 5th ed. Hoboken, NJ: Wiley, 2006.
[65] J. Szargut, D. R. Morris, and F. R. Steward, Exergy analysis of thermal,

chemical, and metallurgical processes. New York, NY: Hemisphere
Publishing, 1988.
[66] C. C. M. Wu, T. Yosenick, C. Kim, M. Kahn, and M. Chase,

"Continuous poling of PZT bars", Manufacturing of Electronic
Materials and Components, vol. 90, pp. 21-31, 1998.
[67] M. Holmgren, "XSteam", ed: www.x-eng.com, 2006.
[68] J. Hong, G. Chaudhry, J. G. Brisson, R. Field, M. Gazzino, and A. F.

Ghoniem, "Analysis of oxy-fuel combustion power cycle utilizing a
pressurized coal combustor", Energy, vol. 34(9), pp. 1332-1340, 2009.
[69] I. Pfaff and A. Kather, "Comparative thermodynamic analysis and

integration issues of CCS steam power plants based on oxy-combustion
170 References
with cryogenic or membrane based air separation", Energy Procedia,

vol. 1(1), pp. 495-502, 2009.
[70] W. T. Welford and R. Winston, High collection nonimaging optics. San

Diego: Academic Press, 1989.
[71] M. Barrio, B. Gøbel, H. Rimes, U. Henriksen, J. E. Hustad, and L. H.

Sørensen, "Steam Gasification of Wood Char and the Effect of
Hydrogen Inhibition on the Chemical Kinetics", in Progress in
Thermochemical Biomass Conversion, ed: Blackwell Science Ltd, 2008,
pp. 32-46.
[72] A. Moilanen and H.-J. Mühlen, "Characterization of gasification

reactivity of peat char in pressurized conditions: Effect of product gas
inhibition and inorganic material", Fuel, vol. 75(11), pp. 1279-1285,
1996.
[73] C. Di Blasi, "Combustion and gasification rates of lignocellulosic chars",

Progress in Energy and Combustion Science, vol. 35(2), pp. 121-140,
2009.
[74] K. J. Hüttinger and W. F. Merdes, "The carbon-steam reaction at

elevated pressure: Formations of product gases and hydrogen
inhibitions", Carbon, vol. 30(6), pp. 883-894, 1992.
[75] S. Kimura, S. Tone, and T. Otake, "Reaction order in the grain model
with grain size distribution", J. Chem. Engng Japan, vol. 14(6), pp. 491-
493, 1981.
[76] O. Levenspiel, Chemical reaction engineering, 3rd ed. New York:

Wiley, 1999.
[77] Z. R. Jovanovic, "Kinetics of direct nitridation of pelletized silicon

grains in a fluidized bed: experiment, mechanism and modelling",
Journal of Materials Science, vol. 33(9), pp. 2339-2355, 1998.
[78] R. Bryan Woodruff and A. W. Weimer, "A novel technique for

measuring the kinetics of high-temperature gasification of biomass char
with steam", Fuel, vol. 103(0), pp. 749-757, 2013.
[79] P. Simell, P. Ståhlberg, E. Kurkela, J. Albrecht, S. Deutsch, and K.

Sjöström, "Provisional protocol for the sampling and anlaysis of tar and
171
particulates in the gas from large-scale biomass gasifiers. Version 1998",

Biomass and Bioenergy, vol. 18(1), pp. 19-38, 2000.
[80] A. Demirbas, D. Gullu, A. ÇAglar, and F. Akdeniz, "Estimation of

Calorific Values of Fuels from Lignocellulosics", Energy Sources, vol.
19(8), pp. 765-770, 1997.
[81] E. Natarajan, M. Öhman, M. Gabra, A. Nordin, T. Liliedahl, and A. N.

Rao, "Experimental determination of bed agglomeration tendencies of
some common agricultural residues in fluidized bed combustion and
gasification", Biomass and Bioenergy, vol. 15(2), pp. 163-169, 1998.
[82] L. V. Barysheva, E. S. Borisova, V. M. Khanaev, V. A. Kuzmin, I. A.

Zolotarskii, N. A. Pakhomov, and A. S. Noskov, "Motion of particles
through the fixed bed in a gas–solid–solid downflow reactor", Chemical
Engineering Journal, vol. 91(2–3), pp. 219-225, 2003.
[83] A. W. M. Roes and W. P. M. Van Swaaij, "Hydrodynamic behaviour of

a gas—solid counter-current packed column at trickle flow", The
Chemical Engineering Journal, vol. 17(2), pp. 81-89, 1979.
[84] E. Saatdjian and J. F. Large, "Heat transfer simulation in a raining

packed bed exchanger", Chemical Engineering Science, vol. 40(5), pp.
693-697, 1985.
[85] K. R. Westerterp and M. Kuczynski, "Gas-solid trickle flow

hydrodynamics in a packed column", Chemical Engineering Science,
vol. 42(7), pp. 1539-1551, 1987.
[86] J. H. A. Kiel, W. Prins, and W. P. M. Van Swaaij, "Modelling of non-

catalytic reactions in a gas-solid trickle flow reactor: dry, regenerative
flue gas desulphurisation using a silica-supported copper oxide sorbent",
Chemical Engineering Science, vol. 47(17–18), pp. 4271-4286, 1992.
[87] A. B. Verver and W. P. M. Vanswaaij, "The Heat-Transfer Performance

of Gas Solid Trickle Flow over a Regularly Stacked Packing", Powder
Technology, vol. 45(2), pp. 133-144, 1986.
[88] A. David, "Experimental demonstration of a novel drop-tube packed bed

solar gasifier for sugarcane bagasse", Semester project, Department of
Mechanical and Process Engineering, ETH Zürich, 2012.
172 References
[89] I. Hischier, P. Leumann, and A. Steinfeld, "Experimental and Numerical

Analyses of a Pressurized Air Receiver for Solar-Driven Gas Turbines",
Journal of Solar Energy Engineering, vol. 134(2), pp. 021003-8, 2012.
[90] W. T. Welford and R. Winston, High collection nonimaging optics. San

Diego: Academic Press, 1989.
[91] J. Petrasch, P. Coray, A. Meier, M. Brack, P. Haberling, D. Wuillemin,

and A. Steinfeld, "A Novel 50 kW 11,000 suns High-Flux Solar
Simulator Based on an Array of Xenon Arc Lamps", Journal of Solar
Energy Engineering, vol. 129(4), pp. 405-411, 2007.
[92] M. Kruesi, Z. R. Jovanovic, E. C. dos Santos, H. C. Yoon, and A.

Steinfeld, "Solar-driven steam-based gasification of sugarcane bagasse
in a combined drop-tube and fixed-bed reactor – Thermodynamic,
kinetic, and experimental analyses", Biomass and Bioenergy, vol. 52, pp.
173-183, 2013.
[93] F. E. Grubbs, "Procedures for Detecting Outlying Observations in

Samples", Technometrics, vol. 11(1), pp. 1-21, 1969.
[94] C. Erlich and T. H. Fransson, "Downdraft gasification of pellets made

of wood, palm-oil residues respective bagasse: Experimental study",
Applied Energy, vol. 88(3), pp. 899-908, 2011.
[95] R. Siegel and J. R. Howell, Thermal radiation heat transfer, 4th ed.
New York: Taylor & Francis, 2002.
[96] J. Petrasch, "A Free and Open Source Monte Carlo Ray Tracing
Program for Concentrating Solar Energy Research", ASME Conference
Proceedings, vol. 2010(43956), pp. 125-132, 2010.
[97] Y. S. Touloukian and D. P. DeWitt, Thermal radiative properties:

nonmetallic solids. New York,: IFI/Plenum, 1972.
[98] M. F. Modest, Radiative heat transfer, 2nd ed. Amsterdam ; Boston:

Academic Press, 2003.
[99] ANSYS Inc., "CFX 13.0", 13.0 ed, 2010.
[100] Y. A. Çengel, Heat Transfer: A Practical Approach: McGraw-Hill,

2003.
173
[101] A. M. Clausing, "Convective Losses from Cavity Solar Receivers -

Comparisons between Analytical Predictions and Experimental Results",
Journal of Solar Energy Engineering-Transactions of the Asme, vol.
105(1), pp. 29-33, 1983.
[102] Design Institute for Physical Properties, Sponsored by AIChE, "DIPPR

Project 801 - Full Version", ed: Design Institute for Physical Property
Research/AIChE, 2010.
[103] J. P. Holman, Heat transfer, 10th ed. Boston: McGraw Hill Higher
Education, 2010.
[104] VDI-Wärmeatlas, 10 ed. Berlin: Verein Deutscher Ingenieure VDI-

Gesellschaft Verfahrenstechnik und Chemieingenieurwesen (GVC),
2006.
[105] M. Kruesi, Z. R. Jovanovic, and A. Steinfeld, "A two-zone solar-driven

gasifier concept: Reactor design and experimental evaluation with
bagasse particles", Fuel, vol. 117, Part A, pp. 680-687, 2014.
[106] T. Poinsot and D. Veynante, Theoretical and numerical combustion,

2nd ed. Philadelphia: Edwards, 2005.
[107] A. Moilanen, K. Saviharju, and T. Harju, "Steam Gasification

Reactivities of Various Fuel Chars", in Advances in Thermochemical
Biomass Conversion, A. V. Bridgwater, Ed., ed: Springer Netherlands,
1993, pp. 131-141.
[108] R. H. Leuenberger and R. A. Person, "Compilation of Radiation Shape

Factors for Cylindrical Assemblies", presented at the ASME paper 56-
A-144, ASME Annual Meeting, New York, 1956.
[109] A. J. Buschman and C. M. Pittman, "Configuration factors for exchange

of radiant energy between axisymmetrical sections of cylinders, cones
and hemispheres and their bases", NASA TN D-944, 1961.
[110] R. Siegel, "Transient Radiative Cooling of a Droplet-Filled Layer",

Journal of Heat Transfer-Transactions of the ASME, vol. 109(1), pp.
159-164, 1987.
[111] S. Suter and S. Haussener, "Morphology Engineering of Porous Media

for Enhanced Solar Fuel and Power Production", JOM, pp. 1-8, 2013.
174 References
[112] J. Petrasch, "Multi-scale analyses of reactive flow in porous media",

Doctor of Sciences, ETH Zurich, 2007.
[113] E. N. Fuller, Schettle.Pd, and J. C. Giddings, "A New Method for

Prediction of Binary Gas-Phase Diffusion Coeffecients", Industrial and
Engineering Chemistry, vol. 58(5), pp. 19-27, 1966.
[114] D. F. Fairbanks and C. R. Wilke, "Diffusion Coefficients in

Multicomponent Gas Mixtures", Industrial & Engineering Chemistry,
vol. 42(3), pp. 471-475, 1950.
[115] R. G. Munro, "Material properties of a sintered alpha-SiC", Journal of

Physical and Chemical Reference Data, vol. 26(5), pp. 1195-1203, 1997.
[116] C. L. Yaws, Transport Properties of Chemicals and Hydrocarbons:

Viscosity, Thermal Conductivity, and Diffusivity for more than 7800
Hydrocarbons and Chemicals, Including C1 to C100 Organics and Ac
to Zr Inorganics: Elsevier Science, 2009.
[117] M. W. Chase and National Institute of Standards and Technology (U.S.),

NIST-JANAF thermochemical tables, 4th ed.: American Chemical
Society, Washington, DC; American Institute of Physics for the
National Institute of Standards and Technology, New York, 1998.
[118] D. Merrick, "Mathematical-Models of the Thermal-Decomposition of

Coal .2. Specific-Heats and Heats of Reaction", Fuel, vol. 62(5), pp.
540-546, 1983.
[119] D. B. Spalding, "A novel finite difference formulation for differential

expressions involving both first and second derivatives", International
Journal for Numerical Methods in Engineering, vol. 4(4), pp. 551-559,
1972.
[120] T. A. Davis, "Algorithm 832: UMFPACK V4.3 – an unsymmetric-

pattern multifrontal method", ACM Trans. Math. Softw., vol. 30(2), pp.
196-199, 2004.
[121] P. J. Roache, "Code verification by the method of manufactured

solutions", Journal of Fluids Engineering-Transactions of the Asme, vol.
124(1), pp. 4-10, 2002.
175
[122] P. Knupp and K. Salari, Verification of Computer Codes in

Computational Science and Engineering: Chapman and Hall/CRC, 2002.
[123] S. Yagi and D. Kunii, "Studies on Effective Thermal Conductivities in

Packed Beds", AIChE Journal, vol. 3(3), pp. 373-381, 1957.
[124] M. A. N. Coelho and J. R. F. Guedes de Carvalho, "Transverse

dispersion in granular beds. I: Mass transfer from a wall and the
dispersion coefficient in packed beds", Chemical Engineering Research
& Design, vol. 66(2), 1988.
[125] J. M. P. Q. Delgado, "A critical review of dispersion in packed beds",

Heat and Mass Transfer, vol. 42(4), pp. 279-310, 2006.
[126] P. C. Carman, "Fluid Flow Through Granular Beds", Transactions -

Institution of Chemical Engineeres, vol. 15, pp. 150-166, 1937.
[127] A. Gopalratnam, "Algal Biofuels: The Backstory", Chemical

Engineering Progress, vol. 107(3), pp. 35-36, 2011.
[128] A. Demirbas, "Use of algae as biofuel sources", Energy Conversion and

Management, vol. 51(12), pp. 2738-2749, 2010.
[129] G. Shelef, A. Sukenik, and M. Green, "Microalgae harvesting and

processing: a literature review", SERI/STR-231-2396; Other: ON:
TI84013036 United States10.2172/6204677Other: ON: TI84013036Thu
Mar 06 08:02:44 EST 2008NTIS, PC A04.NREL; EDB-84-
189417English, 1984.
[130] E. P. Knoshaug and A. Darzins, "Algal Biofuels: The Process",

Chemical Engineering Progress, vol. 107(3), pp. 37-47, 2011.
[131] US DOE, "National algal biofuels technology roadmap", US

Department of Energy, Office of Energy Efficiency and Renewable
Energy, Biomass Program, 2010.
[132] A. Hirano, K. Hon-Nami, S. Kunito, M. Hada, and Y. Ogushi,

"Temperature effect on continuous gasification of microalgal biomass:
theoretical yield of methanol production and its energy balance",
Catalysis Today, vol. 45(1-4), pp. 399-404, 1998.
176 References
[133] A. Zbogar, F. J. Frandsen, P. A. Jensen, and P. Glarborg, "Heat transfer

in ash deposits: A modelling tool-box", Progress in Energy and
Combustion Science, vol. 31(5-6), pp. 371-421, 2005.
[134] B. Atkinson and D. Merrick, "Mathematical-Models of the Thermal-

Decomposition of Coal, 4. Heat-Transfer and Temperature Profiles in a
Coke-Oven Charge", Fuel, vol. 62(5), pp. 553-561, 1983.
[135] J. Fjellerup, U. Henriksen, A. D. Jensen, P. A. Jensen, and P. Glarborg,

"Heat transfer in a fixed bed of straw char", Energy & Fuels, vol. 17(5),
pp. 1251-1258, 2003.
[136] S. V. Patankar, Numerical Heat Transfer and Fluid Flow: Hemisphere

Publishing Corporation, 1980.
List of Publications
Journal Papers:
M. Kruesi, M. E. Galvez, M. Halmann, and A. Steinfeld, "Solar Aluminum

Production by Vacuum Carbothermal Reduction of Alumina—Thermodynamic
and Experimental Analyses", Metallurgical and Materials Transactions B, vol.
42, pp. 254-260, 2011.
M. Kruesi, Z. R. Jovanovic, E. C. dos Santos, H. C. Yoon, and A. Steinfeld,

"Solar-driven steam-based gasification of sugarcane bagasse in a combined
drop-tube and fixed-bed reactor – Thermodynamic, kinetic, and experimental
analyses", Biomass and Bioenergy, vol. 52, pp. 173-183, 2013.
M. Kruesi, Z. R. Jovanovic, and A. Steinfeld, "A two-zone solar-driven gasifier

concept: Reactor design and experimental evaluation with bagasse particles",
Fuel, vol. 117, Part A, pp. 680-687, 2014.
M. Kruesi, Z. R. Jovanovic, A. Haselbacher, and A. Steinfeld, "Analysis of

Solar-Driven Gasification of Biochar Trickling through an Interconnected
Porous Structure", submitted

Eth 46804 02 1

Uploaded by

Copyright:

Available Formats

Eth 46804 02 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eth 46804 02 1

Uploaded by

Copyright:

Available Formats

Research Collection

Heat transfer enhancement in a solar biomass gasifier

HEAT TRANSFER ENHANCEMENT

A thesis submitted to attain the degree of

accepted on the recommendation of

This thesis investigates the production of synthesis gas, a mixture of hydrogen

Die vorliegende Doktorarbeit untersucht die Herstellung von Synthesegas, einer

durchgeführt. Dabei wurde qualitativ hochwertiges Synthesegas mit molaren

vorhersagen wurden mit den im Reaktor gemessenen Temperaturen und Ver-

Zusammenfassung .......................................................................................... vii

Nomenclature ................................................................................................ xvii

4 Thermodynamics of Bagasse Gasification ........................................... 25

5 Gasification Kinetics of Bagasse........................................................... 39

6 Drop-Tube Fixed-Bed Solar Gasifier Concept ................................... 49

7 Drop-Tube Trickle-Bed Solar-Driven Gasifier ................................... 57

8 Heat- and Mass-Transfer Analysis of a Trickle-Bed Gasifier ........... 83

9 Overall Conclusions and Outlook ...................................................... 115

Appendix A: Solar Gasification of Microalgae in a Drop-Tube ............... 121

Appendix B: Thermal Conductivity of a Bed of Char .............................. 135

Appendix C: Numerical Implementation of the Heat and Mass Transfer

List of Figures ............................................................................................... 155

List of Tables ................................................................................................ 161

References ..................................................................................................... 163

Curriculum Vitae ......................................................................................... 177

List of Publications ....................................................................................... 179

0 at normal conditions (T0 = 273.15 K, p0 = 101,325 Pa) / initial

Affordable, transportable, and dispatchable energy is a key driver for a thriving

biomass char + gas + tar hR 0 (1.1)

char / tar / gas + H2O(g) H2 + CO hR 0 (1.2)

In conventional autothermal gasifiers such as the Texaco entrained flow

Figure 1-1: Solar-driven steam gasification process flow sheet.

Solar-driven gasification uses highly concentrated solar radiation as source

1.1 Thesis Outline

highly volatile biomass feedstock that is based on a combination of drop-tube

2.1 Principles of Biomass Gasification

- thermal decomposition or pyrolysis

- gasification of decomposition products

For the gasification of biomass with steam to synthesis gas, as it is proposed in

radiative and gas

primary decomposition gas-phase secondary

Figure 2-1: Primary and secondary reactions during pyrolysis of a biomass

The final pyrolysis products can be classified as [14]:

- solid (char or carbon)

- liquid (tar, heavy hydrocarbons, water)

- gaseous (H2, CO, CO2, CH4, and light hydrocarbons)

C(s) + 0.5 O2 CO h298K 111 kJ/mol (2.2)

CO + 0.5 O2 CO2 h298K 283 kJ/mol (2.3)

H2 + 0.5 O2 H2O(g) h298K 242 kJ/mol (2.4)

C(s) + H2O(g) H2 + CO h298K 131 kJ/mol (2.5)

C(s) + CO2 2 CO h298K 172 kJ/mol (2.6)

C(s) + 2 H2 CH4 h298K 75kJ/mol (2.7)

Water-gas shift reaction

CO + H2O(g) H2 + CO2 h298K 41kJ/mol (2.8)

Steam reforming reactions

CH4 + H2O(g) 3 H2 + CO h298K 206 kJ/mol (2.9)

RC+O2 RC+H 2 O RC+CO2 RC+H2 (2.11)

2.1.3 Tar Formation and Reduction

2.2 Conventional Biomass Gasifier Technology

c) bubbling fluidized bed d) circulating fluidized bed