CO2 capture and valorisation to
chemicals: methane production
Dissertation presented to the
University of Porto
for the degree of
Doctor of Philosophy in Chemical and Biological Engineering
by
CARLOS EDUARDO GERALDES DE VASCONCELOS MIGUEL
Supervisor: Prof. Dr. Luís Miguel Palma Madeira
Co-Supervisor: Prof. Dr. Adélio Miguel Magalhães Mendes
Department of Chemical Engineering Faculty of Engineering
University of Porto
2018
I
Abstract
In this work, a carbon capture and utilization (CCU) unit for CO2 capture
and conversion to methane was studied. A layered packed bed consisting in a
high temperature CO2 sorbent and a methanation catalyst was considered to
perform both capture and conversion in the same unit, in a perspective of
process intensification. The proof-of-concept of the sorptive reactor was
performed using commercial materials in a lab-scale experimental setup which
was built for such purpose allowing also to perform sorption or reaction
experiments.
A thermodynamic analysis of the CO2 hydrogenation using as feed a post-
combustion stream was performed. Hydrogenation to methane is considerably
less demanding in terms of pressure compared to the methanol option,
particularly in the operating temperature window of available industrial catalysts.
Regarding CO2 methanation, the negative effect caused by the presence of
oxygen on CO2 conversion recommends its removal from flue gas. Further
separation of CO2 from the remaining species present in flue gas is also
recommended to enrich the outlet stream with CH4.
Hydrotalcite sorbents were synthesized considering the partial substitution
of aluminum by gallium and the promotion with K, Cs and Sr; their effect on CO2
sorption at high temperature (200-300 ºC) was studied. Cesium and, especially,
potassium enhanced the sorption capacity of the hydrotalcite, while strontium
decreased. A kinetic model assuming two parallel contributions was proposed
for the sorption kinetics, which fitted very well the experiment data. Ga-
substitution improved the sorption capacity of the hydrotalcite that performed
well under cyclic operation. However, gallium incorporation increases the
sorbent cost, which certainly restrains its utilization at industrial scale.
The intrinsic kinetics of CO2 methanation over an industrial nickel catalyst
was determined in a temperature range of 250-350 ºC and at atmospheric
pressure. The selected kinetic model was adapted from the literature (among
others also tested) and has only 4 fitting parameters. The mechanism assumes
II
a formyl intermediate species (CHO*) obtained after H2 and CO2 dissociation,
and hydroxyl (OH*) as the most abundant species at the catalyst surface. Such
model was validated through the simulation of an isothermal plug-flow reactor
and a reasonable agreement between model predictions and the observed
values was obtained.
The integrated sorption-reaction process for CO2 capture and conversion
to CH4 was studied between 300 ºC and 350 ºC and at low pressure (P ≤ 2.5
bar). The sorptive reactor had inside a layered bed of a commercial hydrotalcite
sorbent (Pural MG30-K from Sasol) and a commercial nickel catalyst (METH 134
from Clariant). The materials proved their compatibility and stability under cyclic
operation, at least under the considered operation conditions. The sorption
equilibrium was well described by the dual-site Langmuir model, which considers
that CO2 is sorbed through physical sorption (exothermic contribution) and
chemical reaction (endothermic contribution). Reactive regeneration mode (i.e.
with a H2 stream) improved desorption kinetics, when compared to normal
regeneration (i.e. with a N2 stream), and allowed to completely recover the
sorbent capacity in each cycle. The steam produced in situ by the methanation
reaction causes this positive effect. The synthetic flue gas used was made of
CO2 and N2 only. CO2 conversion was high in batch operation (ca. 92 %) and
almost complete (99 %) when the reactor operated under continuous mode. The
major drawback found in the operation of the sorptive reactor was the need for
improving the methane purity; CO formation could be avoided by decreasing the
temperature and/or increasing the pressure.
The use of highly permeable and selective metal membranes can be
considered for feeding radially hydrogen to the methanation packed-bed reactor.
The permeance towards H2 of a 50-µm thick PdAg membrane was then
measured between 200 ºC and 300 ºC considering a pure H2 stream or binary
H2/CO or H2/CO2 mixtures. The Sieverts law was observed for the former
conditions, while for the binary mixtures the Sieverts-Langmuir model explained
the inhibitory effect on H2 permeance caused by CO and CO2. A modified
III
Sieverts-Langmuir equation was proposed to account for changes in the
permeation driving-force under high H2 recovery conditions.
V
Resumo
Neste trabalho estudou-se uma unidade de captura e utilização de
carbono (CCU) para a captura e conversão de CO2 em metano. Para realizar a
captura e conversão de CO2 na mesma unidade, foi considerado um leito
estratificado contendo um adsorvente de CO2 de alta temperatura e um
catalisador para a reação de metanação, numa perspetiva de intensificação de
processos. A prova de conceito do reator adsortivo foi efetuada utilizando
materiais comerciais e teve lugar numa unidade experimental à escala
laboratorial que foi construída para este fim, a qual permite também realizar
ensaios de adsorção e reação separadamente.
A reação de hidrogenação de CO2 foi estudada do ponto de vista
termodinâmico considerando como alimentação uma corrente de pós-
combustão. A hidrogenação de CO2 para formação de metano é
consideravelmente menos exigente em termos de pressão de operação quando
comparada com a opção do metanol, em particular na gama de temperatura de
operação dos catalisadores industriais disponíveis. Em relação à reação de
metanação, o efeito negativo causado pela presença de oxigénio na conversão
do CO2 recomenda a sua remoção do gás de exaustão. A separação do CO2
das restantes espécies presentes no gás de exaustão também é recomendada,
essencialmente para enriquecer a corrente de saída em CH4.
Foram sintetizadas hidrotalcites considerando a substituição parcial de
alumínio por Ga e a modificação com K, Cs e Sr, tendo-se estudado o seu efeito
na capacidade de adsorção de CO2 entre 200 ºC e 300 ºC. A impregnação com
césio e potássio, e deste último em particular, aumenta a capacidade de
adsorção da hidrotalcite, enquanto o estrôncio diminui. Foi proposto um modelo
para a cinética de adsorção considerando duas contribuições em paralelo, o
qual ajustou muito bem os dados experimentais. A substituição com Ga
melhorou a capacidade de adsorção e o adsorvente mostrou boa capacidade
durante operação cíclica. Contudo, a incorporação de gálio aumenta
VI
consideravelmente o custo do adsorvente, o que certamente pode restringir a
sua utilização a uma escala industrial.
A cinética verdadeira da metanação de CO2 foi determinada usando um
catalisador de níquel industrial entre 250 ºC e 350 ºC à pressão atmosférica. O
modelo cinético selecionado foi adaptado da literatura (entre outros igualmente
testados) e apresenta apenas 4 parâmetros de ajuste. O mecanismo assume a
formação de uma espécie de formilo (CHO*) intermediária obtida após a
adsorção dissociativa de H2 e CO2, bem como o grupo hidroxilo (OH*) como
espécie mais abundante na superfície do catalisador. O modelo foi validado
através da simulação de um reactor isotérmico de leito fixo incluindo a cinética
determinada e os resultados ajustam bem os dados experimentais.
O processo integrado de adsorção-reação para a captura e conversão de
CO2 em CH4 foi estudado entre 300 ºC e 350 ºC a baixa pressão (P ≤ 2.5 bar).
O leito do reator adsortivo foi estruturado em camadas intercaladas de uma
hidrotalcite comercial (Pural MG30-K da Sasol) e de um catalisador de níquel
comercial (METH 134 da Clariant). Os materiais mostraram serem compatíveis
e estáveis em operação cíclica, pelo menos nas condições de operação usadas.
O equilíbrio de adsorção foi bem descrito pelo modelo de Langmuir de dois
centros (dual-site), o qual considera que o CO2 é adsorvido através de adsorção
física (contribuição exotérmica) e reação química (contribuição endotérmica). A
regeneração reativa (i.e. com uma corrente de H2) melhorou a cinética de
dessorção, quando comparada com a regeneração normal (i.e. com N2), tendo
permitido recuperar completamente a capacidade de adsorção em cada ciclo.
O vapor de água produzido in situ pela reação de metanação provocou este
efeito positivo. A corrente de gás de exaustão considerada foi feita apenas com
CO2 e N2. A conversão de CO2 foi elevada em operação descontínua (ca. 92 %)
e quase total (99 %) quando o reator foi operado em modo contínuo. A maior
contrariedade registada na operação do reator adsortivo foi a necessidade de
aumentar-se a pureza do metano na corrente de saída; a formação de CO pode
ser evitada diminuindo a temperatura ou aumentando a pressão.
VII
O uso de membranas metálicas altamente permeavéis e seletivas pode
ser considerado para se alimentar radialmente hidrogénio ao reactor de
metanação de leito fixo. Mediu-se então a permeância ao H2 de uma membrana
de PdAg com 50 µm de espessura entre 200 ºC e 300 ºC considerando uma
corrente pura de H2 ou mistura binárias de H2/CO ou H2/CO2. A lei de Sieverts
foi observada para o primeiro caso, enquanto o modelo de Sieverts-Langmuir
descreve o efeito inibitório na permeação de H2 provocado pelo CO e CO2. Foi
proposta uma versão modificada deste modelo que considera as alterações na
força diretriz da permeação e que deve ser aplicada em condições de elevada
recuperação de H2.
IX
Acknowledgments
In this section I would like to acknowledge all of you who, in many ways,
gave me support throughout this journey.
I am grateful to my thesis supervisors, Prof. Luis Miguel Madeira and Prof.
Adélio Mendes, for their guidance and encouragement. Both showed me how to
think as a chemical engineer researcher. Thanks Professor Madeira for your
confidence on my work, your comprehension during these years and for assuring
me the best conditions to do my work. Thanks Professor Adélio for being always
available and for your valuable comments.
To Eng. Silvano Tosti, from ENEA (Frascaty, Italy), I acknowledge for his
collaboration (namely for providing the membranes) and sharing his expertise
on Pd-based membranes.
I am grateful to Prof. Vicente Rives, Prof. Miguel Angel Vicente and Prof.
Raquel Trujillano for welcoming me twice in the University of Salamanca (Spain)
and for providing me guidance and sharing their knowledge in the synthesis and
characterization of hydrotalcite materials.
I am grateful to Prof. Alírio Rodrigues for giving me the opportunity to finish
the adsorption experiments at LSRE and move forward with my work when I
faced endless problems with the gas sorption analyzer at LEPABE. I am grateful
to Dr. Alexandre Ferreira for his help on the adsorption experiments and for
sharing his knowledge during my stay at LSRE.
I would like to thank to Dr. Miguel Soria for helping me assembling the
experimental unit, for his useful advices on experimental questions and for his
scientific inputs in the works we have performed together. I am also grateful for
our Monday morning talks about Sporting CP and River Plate performances
during the weekend, which were very pleasant, especially when both teams had
won.
To all my lab mates, FEUPsal team colleagues and friends outside the
“academic and scientific world” I acknowledge for their camaraderie and
X
friendship, for the relevant and less relevant discussions which have made these
years very enjoyable.
I acknowledge companies Sasol and Clariant for supplying the materials
that I used in this work. I am grateful to Ana Taborda (Clariant) for her attention
and sympathy.
To my parents António and Clarisse and my brothers Vanessa and
Armando for their unconditional love.
To my wife Cristina for her love, patience, and generosity. This thesis is
dedicated to her… and to our son Gustavo.
I acknowledge the Portuguese Foundation for Science and Technology
(FCT) for my PhD grant (SFRH/BD/110580/2015) financed by national funds of
the Ministry of Science, Technology and Higher Education and the European
Social Fund (ESF) through the Human Capital Operational Programme (POCH).
Furthermore, I would like to acknowledge also partial financial support of:
● Project PTDC/EQU/ERQ/098730/2008 financed by the European Fund for
Regional Development (ERDF) through COMPETE - Operational Programme
for Competitiveness Factors and through national funds by FCT;
● Project NORTE-07-0124-FEDER-000026 - RL1_ Energy funded by the ERDF
through the COMPETE and ON.2 – O Novo Norte - North Portugal Regional
Operational Programme and through national funds by FCT;
● Project PEst-C/EQB/UI0511 funded by national funds through FCT;
● Portugal-Spain Cooperation Integrated Action Nº E-59/10 funded by CRUP.
● Project POCI-01-0145-FEDER-006939 (Laboratory for Process Engineering,
Environment, Biotechnology and Energy – UID/EQU/00511/2013) - funded by
the ERDF through COMPETE2020 - Programa Operacional Competitividade e
Internacionalização (POCI) – and by national funds through FCT;
● Project “LEPABE-2-ECO-INNOVATION”, with the reference NORTE‐01‐
0145‐FEDER‐000005, supported by Norte Portugal Regional Operational
Programme (NORTE 2020), under the Portugal 2020 Partnership Agreement,
through the ERDF.
XI
XIII
Contents
Abstract I
Resumo V
Acknowledgments IX
Chapter 1. General Introduction 1
1.1 Relevance and Motivation 2
1.1.1 Options for CO2 capture 3
1.1.1.1 Absorption 5
1.1.1.2 Membranes 6
1.1.1.3 Cryogenic distillation 9
1.1.1.4 Adsorption 10
1.1.2 Options for CO2 conversion 12
1.1.3 The Power-to-Gas concept 14
1.2 Overview of the Portuguese energy sector 17
1.2.1 Energy dependence 17
1.2.2 Energy from fossil fuels 19
1.2.2.1 Oil 19
1.2.2.2 Coal 20
1.2.2.3 Natural Gas 21
1.2.3 Energy from renewables 24
1.2.3.1 Geothermal 27
1.2.3.2 Photovoltaic 27
1.2.3.3 Biomass 28
1.2.3.4 Hydropower 29
1.2.3.5 Wind 31
1.2.4 Prospects for Power-to-Methane 34
XIV
1.2.4.1 Surplus renewable power 34
1.2.4.2 CO2 sources and availability 35
1.2.4.3 Natural gas grid 39
1.2.4.4 Summary 45
1.3 Objective and outline of the thesis 48
1.4 Acronyms 51
1.5 References 52
Chapter 2. Direct CO2 hydrogenation to methane and methanol from post-
combustion exhaust streams – a thermodynamic study 61
2.1 Introduction 62
2.2 Methodology 64
2.3 Results and Discussion 67
2.3.1 Strategies for CO2 valorisation: CH4 or CH3OH? 67
2.3.2 CO2 methanation: effect of pressure, temperature and H2/CO2
ratio 70
2.3.3 Direct CO2 methanation from coal-fired power plant (CF-PP)
flue gas streams 72
2.3.3.1 Effect of H2O 72
2.3.3.2 Effect of O2 74
2.3.3.3 Simultaneous effect of H2O and O2 75
2.4 Technological implementation 76
2.5 Conclusions 78
2.6 References 79
Chapter 3. High temperature CO2 sorption with gallium-substituted and
promoted hydrotalcites 83
3.1 Introduction 84
3.2 Experimental 86
3.2.1 Chemicals and gases 86
3.2.2 Sorbents preparation 86
XV
3.2.3 Sorbents characterization 88
3.3 Results and Discussion 89
3.3.1 Sorbents physicochemical characterization 89
3.3.1.1 TG / DTG 89
3.3.1.2 XRD 90
3.3.1.3 SEM / EDS 92
3.3.1.4 Physical adsorption of nitrogen at 77 K 94
3.3.1.5 FTIR 95
3.3.2 CO2 sorption experiments 97
3.3.2.1 Determination of adsorption equilibrium isotherms at 573
K – sorbent screening 97
3.3.2.2 Sorption-desorption experiments 100
3.4 Conclusions 107
3.5 References 108
Chapter 4. Intrinsic kinetics of CO2 methanation over an industrial nickel-
based catalyst 113
4.1 Introduction 114
4.2 Mechanisms for CO2 methanation 116
4.2.1 “Carbon intermediate” mechanism 116
4.2.2 “Formate intermediate” mechanism 118
4.2.3 “Formyl intermediate” mechanism 119
4.3 Experimental 121
4.3.1 Experimental setup 121
4.3.2 Kinetic experiments 122
4.3.3 Computational methods 124
4.4 Results and Discussion 125
4.4.1 Kinetic experiments 125
4.4.1.1 Isothermal regime and catalyst stability 125
4.4.1.2 Identifying the region of kinetic rate control 126
4.4.1.3 Calculation of observed reaction rates 128
XVI
4.4.2 Modelling work 129
4.4.2.1 Model discrimination and parameter estimation 129
4.4.2.2 Model validation 132
4.5 Conclusions 138
4.6 Nomenclature 139
4.7 References 141
Chapter 5. A sorptive reactor for CO2 capture and conversion to
renewable methane 145
5.1 Introduction 146
5.2 Experimental 148
5.2.1 Experimental setup and hybrid unit 148
5.2.2 Sorption-desorption/reaction experiments 151
5.2.3 Process indicators: definitions and equations 152
5.3 Results and Discussion 154
5.3.1 Sorptive reactor under discontinuous operation mode 155
5.3.1.1 CO2 sorption working capacity 155
5.3.1.2 Effect of steam produced during reactive regeneration
156
5.3.1.3 Effect of temperature 165
5.3.1.4 Effect of pressure 168
5.3.2 Sorptive reactor under continuous operation mode 171
5.3.3 Reactor design considerations 174
5.4 Conclusions 175
5.5 Nomenclature 177
5.6 References 179
Chapter 6. Effect of CO and CO2 on H2 permeation through finger-like Pd-
Ag membranes 183
6.1 Introduction 184
6.2 Theoretical background 186
XVII
6.3 Experimental 189
6.3.1 Membrane preparation and experimental set up 189
6.3.2 Gas permeation experiments 190
6.4 Results and Discussion 191
6.4.1 Pure H2 permeance experiments 191
6.4.2 Effect of CO and CO2 on H2 permeance 192
6.4.2.1 Comparison between Sieverts-Langmuir model and
rearranged equation 192
6.4.2.2 Modelling by the Sieverts-Langmuir equation 195
6.5 Conclusions 201
6.6 References 202
Chapter 7. Conclusions and Future Work 205
7.1 Conclusions 205
7.2 Future work 211
7.3 References 214
Appendix A. Supplementary information for Chapter 3 215
Appendix B. Supplementary information for Chapter 4 223
Appendix C. Supplementary information for Chapter 5 233
Appendix D. Experimental Setup 241
1
Chapter 1. General Introduction
This chapter presents the relevance and motivation behind this PhD
study, listing the available options for CO2 capture and conversion. Among
these options, the Power-to-Gas concept is introduced and its
implementation in Portugal forecasted based on the situation of the
Portuguese energy sector. The chapter ends listing the PhD work objectives
and providing the thesis outline.
Chapter 1
2
1.1 Relevance and Motivation
The increase of CO2 concentration in the atmosphere, resulting from
human activity and excessive burning of fossil fuels, is considered the major
contributor to global warming [1]. In this perspective, it is essential to develop
strategies that lead to improved technologies regarding CO2 capture,
sequestration and recycling [2]. Although there are several examples of systems
for capturing CO2 that have been used for several years [3], there are clear
technical and economic advantages in contemplating its conversion, particularly
if substances that can bring added value to the process/company can be
produced [2]. In these cases, CO2 can be seen as a raw material with null (or
even negative) cost, if used to produce chemicals and fuels, a market that opens
potentially gigaton-level applications for CO2 utilization [4].
There are several products into which CO2 can be converted: fuels, the
intermediates of such substances or other value-added products, a topic that
has been the subject of several reviews [4-8]. When choosing the technology to
use for CO2 recycling several criteria should also be considered, such as the
associated costs, the market needs or the easiness of storage and transportation
of the produced fuels/intermediates. For these reasons, the choice of the
technology for CO2 recycling fell in the synthesis of CH4, which can be easily
integrated with the existing natural gas infrastructures [4]. In this process there
is consumption of H2, and so this route should be viable when the H2 proceeds
from non-fossil sources [4, 7]. This is the case, for instance, when H2 is obtained
via water electrolysis. Additionally, the energy required for H2 production must
be also renewable (such as solar, wind, etc.) to approach a sustainable process
from the carbon footprint point of view.
This thesis addresses therefore the sustainable production of methane
while simultaneously reducing carbon dioxide emissions into the atmosphere by
recycling carbon using renewable energy for “green” hydrogen production.
General Introduction
3
1.1.1 Options for CO2 capture
The International Energy Agency (IEA) estimated that CO2 emissions
reached 32.4 Gt in 2014 and this value is expected to grow up to 37.5 Gt of CO2
by 2035 [9]. Additionally, IEA also shows that the emissions resulting from
electricity and heat generation in power plants represented in 2014 about 42 %
of the global CO2 emissions, whereas the transport and industry sectors
accounted for 23 % and 19 %, respectively [9]. Developing more efficient and
economical capture technologies is required to minimize the impact on the
environment caused by greenhouse gases (GHG) emissions when the world’s
economy still intensively relies on the combustion of fossil fuels for energy
production.
The efficiency of capture technologies depends on several conditions such
as the exhaust stream composition, temperature and pressure. CO2 content in
a post-combustion exhaust stream from a natural gas power plant can be as low
as 4 %, up to 15 % in a coal power plant, and between 20-30 % in cement and
steel production plants [10]. Higher values are obtained in pre-combustion
streams, such as those observed in syngas production processes, like
hydrocarbons reforming or gasification. Syngas is commonly produced via the
steam reforming of methane (SRM) (Eq. (1.1)) and water-gas shift (WGS)
reactions (Eq. (1.2)). In these processes, CO2 can be captured by adsorption
during reaction - the so-called Sorption-Enhanced Reaction Processes (SERP)
(e.g. [11-14]).
-1
4 2 2 298 KCH + H O CO + 3H 205.9 kJ molH (1.1)
-1
2 2 2 298 KCO + H O CO + H 41.1 kJ molH (1.2)
Since both SRM and WGS are equilibrium-limited reactions and because
CO2 (product) is being adsorbed, the reaction is shifted and proceeds in the
Chapter 1
4
direction of product formation so that the overall performance is improved.
Therefore, in theory this concept allows obtaining:
I. An higher reactant conversion when compared to a traditional packed-
bed reactor in the same operating conditions;
II. Two separate H2- and CO2-rich streams during the adsorption and
desorption stages, respectively.
In gasification processes, namely in Integrated Coal Gasification
Combined Cycle (IGCCC) systems, CO2 is typically captured by absorption with
amines prior to H2 combustion for electricity production in a gas turbine
combined-cycle [3, 15]. From a thermodynamic point of view, and for streams at
the same temperature and pressure, less work is required for the separation in
streams with high CO2 contents, which is the case found in pre-combustion
approaches as illustrated in Fig. 1.1.
Fig. 1.1 – Minimum thermodynamic work for various coal or gas-to-electricity conversions. NGCC: natural gas combined cycle; PCC: Post-combustion capture; IGCC: Integrated Coal Gasification combined cycle. Reprinted by permission from Springer Nature: Springer, Carbon Capture by Jennifer Wilcox, Copyright (2012).
General Introduction
5
Fig. 1.2 shows a diagram presenting the available CO2 separation
technologies. Each technology will be briefly addressed in the following sections.
Fig. 1.2 - Carbon dioxide separation technologies. Reprinted from Chemical Engineering Research and Design, 89, M. Wang, A. Lawal, P. Stephenson, J. Sidders, C. Ramshaw, Post-combustion CO2 capture with chemical absorption: A state-of-the-art review, 1609-1624, Copyright (2011), with permission from Elsevier.
1.1.1.1 Absorption
CO2 capture by absorption with amines is the most mature and established
solution among the various options available, taking place for over 70 years [3].
Absorption technology is based on the transfer of a solute from a gas- to a liquid-
phase and requires two connected columns. On one column the solute (i.e. CO2)
is being separated due to its affinity with the solvent (amine) and in the second
one it is released while the solvent is regenerated for further usage (Fig. 1.3).
Chapter 1
6
Fig. 1.3 - Typical CO2 absorption process scheme. Reprinted from Separation Science and Technology, 40, D. Aaron, C. Tsouris, Separation of CO2 from flue gas: a review, 321-348, Copyright (2005), with permission from Taylor & Francis.
The efficiency of the removal is based on the CO2 solubility, diffusivity and
chemical reactivity with the solvent. The most common solvent used is
monoethanolamine (MEA), particularly for low CO2 concentrations (<15 vol. %)
[16]; the industrial standard for CO2 removal using MEA is ca. 60 gCO2·kgMEA-1 (or
1.36 molCO2 kgMEA-1) [17]. The operation and maintenance cost including solvent
replacement is about $40-$70 per ton of CO2. Temperature requirements for
solvent regeneration represent ca. 70-80 % of the operating cost [16].
1.1.1.2 Membranes
Membranes can be porous or non-porous and the dominant related
transport mechanisms for each type are illustrated in Fig. 1.4.
General Introduction
7
a) b) c) d)
Fig. 1.4 – Transport mechanisms for (a-c) porous and (d) non-porous membranes: a) convective flow, b) Knudsen diffusion, c) activated transport/molecular sieving and d) solution-diffusion (adapted from [18]).
As illustrated in Fig. 1.4, separation in porous membranes relies on
molecular size difference between species present in the mixture (Fig. 1.4c)
and/or their diffusivity rates (Fig. 1.4b). In the former, only molecules having a
smaller size than the pores are able to pass through the membrane, while on the
latter the fastest molecules are firstly recovered at the permeate side. Depending
on the pore size range, porous membranes can be classified according to the
classes presented in Table 1.1.
Table 1.1 - Porous membranes classifications [19].
Classification Pore size range (nm)
Microfiltration 100-5000
Ultrafiltration 1-100
Nanofiltration 0.5-10
Reverse osmosis < 0.5
However, the kinetic diameter of species typically present in flue gas
streams are very similar (cf. Table 1.2). Thus, preparing membranes with a
narrow pore size distribution is complex and, generally, high selectivity towards
CO2 is difficult to achieve. Besides, water vapour condensation may block the
membrane porous strucuture [3].
Chapter 1
8
Table 1.2 – Kinetic diameter of species typically found in flue gas streams [18].
Molecule Kinetic diameter (nm)
CH4 0.380
CO 0.376
N2 0.364
O2 0.346
CO2 0.330
H2 0.289
H2O 0.280
Examples of porous membranes for CO2 separation includes zeolite
membranes, carbon molecular sieve and metal-organic frameworks (MOF’s)
based membranes [3]. The most common non-porous membranes are based on
polymers (e.g. poly-acetylenes, polyaniline, polyarylates, polycarbonates,
polyetherimides, polyimides, polysulfones) [20]. In the work by Robeson [21], the
relationship between permeability and separation for several gas pairs and
polymeric membranes is given; a trade-off relationship shows that, in general,
the higher the permeability, the lower is the selectivity.
Another class of non-porous membranes are the facilitated-transport
membranes (also referred as carrier mediated transport membranes). These
membranes can be divided in two classes: mobile- and fixed-carrier [22]. The
working principle is based on the reaction of CO2 with a carrier (e.g. amine) to
form a complex at the feed side, being regenerated in the permeate side (c.f.
Fig. 1.5). The remaining feed gases that do not react with the carrier agent
permeate by the solution-diffusion mechanism or remain in the retentate side
[23].
The applicability of membrane separation for CO2 capture from a power
plant exhaust stream is mostly compromised due to the simultaneous low CO2
content and pressure, not providing the required driving-force for effective
permeation [3]. For this reason, membrane systems are more suitable for
application in pre-combustion capture, where a higher CO2 fraction and pressure
can be found. However, to this technology become competitive, cheap,
selective, and stable membranes at high temperatures are required. If these
General Introduction
9
requirements are fulfilled, membrane systems present important advantages
such as the easiness of operation, low maintenance work, easy scale-up and,
eventually, the exemption of a regeneration stage.
Fig. 1.5 – Illustration showing the facilitated-transport mechanism. Reprinted with permission from Industrial & Engineering Chemistry Research, 47, J. Huang, J. Zou, W. S. Winston Ho, Carbon dioxide capture using a CO2-selective facilitated transport membrane. Copyright (2008) American Chemical Society.
1.1.1.3 Cryogenic distillation
Cryogenic distillation is used to separate CO2 from N2 (cf. Fig. 1.6). Thus,
the remaining flue gas species must be previously removed [16].
Fig. 1.6 – Cryogenic distillation process for CO2 separation. Reprinted from Separation Science and Technology, 40, D. Aaron, C. Tsouris, Separation of CO2 from flue gas: a review, 321-348, Copyright (2005), with permission from Taylor & Francis.
As depicted in Fig. 1.6, while CO2 is being liquefied at the bottom of the
column, gaseous N2 is recovered at the top. The major advantage of cryogenic
distillation is that CO2 is obtained in a liquefied state with a purity up to 99.95 %
(i.e. CO2 ready for transportation). The major disadvantage is related to the high
Chapter 1
10
energy consumption necessary to maintain the cool temperature (CO2 triple
point is 216.55 K at 5.185 atm), making this technology costly and thus less
interesting than other options available [16].
1.1.1.4 Adsorption
Adsorption processes showed already their effectiveness in the separation
of dilute streams [3]. Different categories of materials can be used for CO2
adsorption depending on the operating temperature as shown in Fig. 1.7.
Fig. 1.7 – CO2 adsorbent materials options as a function of the temperature [24].
Zeolites, activated carbons and organic-inorganic hybrids are limited to
temperatures up to 400 K. Recently, metal-organic frameworks (MOF’s) have
attracted the attention for CO2 capture. This type of sorbents are made of
transition metal ions and organic linkers resulting in highly organized and porous
structures with high surface areas, but their performance towards CO2 capture
is also substantially detrimentally affected by the presence of water vapour [25,
26].
General Introduction
11
Hydrotalcites are cheap materials that can be used at temperatures
between 473 and 673 K, a range typically found in both post-combustion and
pre-combustion streams. Although hydrotalcites use at temperatures above 673
K has also been reported [27], lithium zirconates and calcium oxides are options
more suited for CO2 capture at high temperatures (673-873 K).
Exhibiting a reasonably good adsorption capacity and stability,
hydrotalcites are easily regenerated by temperature or pressure swing [25, 28].
Moreover, water vapour benefits adsorption because it promotes the
reconstruction of their layered structure (see Fig. 1.8), a characteristic of these
materials known as memory effect [29, 30]. This is a distinctive feature of these
materials compared to other adsorbents, making them suitable for CO2 capture
in processes were water vapour is present (e.g. flue gas from fossil fuels
combustion or in sorption enhanced processes for H2 production (i.e. pre-
combustion capture)).
Fig. 1.8 – Illustration showing the layered structure of a hydrotalcite. Reprinted with permission from Chemistry of Materials, 16, N.D. Hutson, S.A. Speakman, E.A. Payzant, Structural effects on the high temperature adsorption of CO2 on a synthetic hydrotalcite. Copyright (2004) American Chemical Society.
Studies addressing some aspects that influence the hydrotalcite thermal
stability, CO2 adsorption/desorption kinetics and equilibrium such as the
Mg2+/Al3+ ratio, the substitution of the compensating anion, the calcination
temperature, alkali metal doping and the water vapour presence can be found
elsewhere (e.g. [28-33]).
Chapter 1
12
1.1.2 Options for CO2 conversion
Some reviews provide a comprehensive insight about emerging
technologies for CO2 recycling (e.g. [4, 34]). Since CO2 molecule has low
reactivity, its conversion requires high-energy reaction “partners” (e.g. H2). For
this reason, although some existing technologies for CO2 conversion are well-
established (e.g. urea, methanol or methane production), the development of
innovative routes with industrial potential will be surely closely linked to the
design of novel catalysts, particularly for methanol synthesis [2]. Fig. 1.9 shows
the options for CO2 conversion into chemical commodities along with the
corresponding market volume.
Fig. 1.9 – CO2 conversion options into chemicals and market volume. Reproduced from Ref. [35] with permission from the Royal Society of Chemistry.
Among the options presented in Fig. 1.9, the use of CO2 as precursor in
organic carboxylation reactions or in reduction reactions are the most promising
routes. In organic carboxylation reactions, CO2 is used for the production of
carbonates, acrylates, carboxylic acids and polymeric materials [36]. For
instance, polyurethane contains as much as 40 wt. % of CO2 and if used as an
General Introduction
13
insulating material (e.g. in buildings), it is claimed that can avoid up to 70 kg of
CO2 emissions for every kilogram of fixated CO2 [4]. However, it is claimed that
the plastics market is not large enough in order to restrain CO2 emissions as
needed [4].
Using CO2 to produce algae is another open topic in the literature, although
it has been reported that the production costs of algal biomass are still too high
to cultivate algae for biomass-to-energy applications and that, for now, the
balance appears unfavourable when compared to fossil fuel alternatives [2, 4].
On the other hand, CO2 reduction to produce methanol, methane, ethanol,
formic acid, syngas and dimethyl ether opens gigaton-level applications for CO2
utilization [4, 37]. However, these reactions require H2 which is widely produced
by steam methane reforming process [38]. For this reason, H2 production should
use renewable resources, or electrons as reducing agent in electrochemical
applications, so that these CO2 recycling options can be sustainable from the
global carbon footprint point of view (cf. Fig. 1.10).
Fig. 1.10 – Renewable power and CO2 storage routes. Reprinted from Renewable and Sustainable Energy Reviews, 69, M. Bailera, P. Lisbona, L. M Romeo, S. Espatolero, Power to Gas projects review: lab, pilot and demo plants for storing renewable energy and CO2, 292-312, Copyright (2017), with permission from Elsevier.
Chapter 1
14
1.1.3 The Power-to-Gas concept
The “Power-to-Gas” (PtG) concept is based on the conversion of surplus
electricity obtained from a renewable source (sun, wind, etc.) into a gas easily
transported and stored. The conversion to CH4 benefits from the existence of a
well-established natural gas infrastructure into where methane can be injected,
enabling the integration/optimization of power and gas grids; this allows the
stabilization of the power grid driven by electricity consumption through the
reconversion of the energy carrier produced, in this case methane [39].
The PtG concept can thus be subdivided in 2 domains: 1) Energetic,
aiming to improve renewable energy technologies/processes and 2)
Environmental, since promoting the efficiency of renewables and CO2 capture
and utilization, it tackles climate changes resulting from CO2 emissions and
contributes for reducing the carbon footprint. Since the current energy
infrastructure relying on fossil fuels is unlikely to change in the forthcoming
years, the production of valuable CH4 from waste CO2 can thus be an interesting
carbon management solution.
From the technological viewpoint, the process consists on the production
of CH4 from the catalytic reaction between CO2 and H2 through the methanation
(or Sabatier) reaction (Eq. (1.3)).
-12 2 4 2 298 KCO + 4H CH + 2H O = - 165 kJ molH (1.3)
The surplus renewable electricity is used for H2O electrolysis (Eq. (1.4))
to obtain “green” H2 necessary for the methanation.
-12 2 2 298 K
1H O H + O = 285.8 kJ mol
2H (1.4)
General Introduction
15
During electrolysis, the application of an electric potential to two
electrodes leads to the splitting of H2O molecule into hydrogen and oxygen at
the cathode and anode, respectively. The electrolysers can be classified
regarding to the used electrolyte as: i) alkaline water electrolysis (AEL), ii) proton
exchange membrane electrolysis (PEM) and iii) solid oxide electrolysis (SOEC),
also known as high-temperature water electrolysis [40]. Currently, AEL is the
cheapest and most reliable technology, PEM the most promising (in the near
future) for transient operation and SOEC is still in the development phase [41].
With respect to CO2 sources, several processes may provide the
necessary feedstock: coal or biomass gasification (for syngas production),
biomass anaerobic fermentation (for biogas production), or flue gas streams of
varied industrial processes (e.g. coal and natural gas power plants, refineries,
cement and steel industries, etc.), requiring however its separation from other
coexisting species (e.g. N2, H2O, O2, etc.) prior to methanation.
CO2 purification requirements may differ substantially depending on the
carbon source considered, which has direct influence on PtG process
configuration and flexibility.
In the European context, due to the increasing share of renewable energy
across the electricity mix, the demand for balancing the electricity supply and the
demand over spatial and temporal distances is increasing, constituting the major
drivers towards the implementation of PtG processes [42].
The largest PtG plant worldwide is the Audi e-gas plant in Wertle,
Germany, where 3 AEL (total electric power of 6 MW) are used to produce
hydrogen, whereas carbon dioxide comes from a biogas plant [41]. The
substitute natural gas (SNG) production capacity of this plant is, in average,
around 1.4x106 m3/year [43]. A picture of the Audi e-gas plant is shown in Fig.
1.11.
Chapter 1
16
Fig. 1.11 – Audi e-gas plant in Wertle (Germany). Reproduced from Ref. [44]. © Audi AG.
Another demonstration activity in Germany is the KIC-Project, DemoSNG,
where the methanator is fed with a volumetric feed flow rate of 10 Nm3/h [45].
The pilot plant is enclosed in a standard shipping container and is currently under
operation in a biomass gasification plant located in Köping, Sweden.
In Denmark, the BioCatProject developed by Eletrochaea uses biological
methanation as a part of the Power-to-Gas process chain. The plant uses an
advanced version of Hydrogenics’ S1000 alkaline electrolyser to produce H2 that
is fed to the biological methanation system together with CO2 from: raw biogas
from an adjacent anaerobic digester with a composition of approximately 60%
CH4 and 40% CO2, or a pure stream of CO2 supplied by an on-site biogas
upgrading system [46]. The plant is located at the BIOFOS Avedøre wastewater
treatment plant near Copenhagen.
A full list of Power-to-Gas projects in the lab, pilot and demonstration
stages was recently presented in the review by Bailera et al. [47], showing that
most activities take place in Europe, particularly because of the support of
governments from Germany, Denmark and Switzerland.
General Introduction
17
1.2 Overview of the Portuguese energy sector
1.2.1 Energy dependence
The energy dependence of a country is an indicator that expresses the
extent to which an economy relies upon imports to meet its energy needs. The
indicator is calculated as net imports of primary energy (IMP-EXP) divided by
the sum of gross inland energy consumption (GIC) plus international maritime
bunkers (IMB), which are quantities of fuels delivered to ships of all flags that
are engaged in international navigation (cf. Eq. (1.5)) [48].
100IMP EXP
Energy DependenceGIC IMB
(1.5)
Hence, a country that exports more primary energy than imports has a
negative energy dependence or, in a simpler way, it is said to be energetically
independent. The Portuguese energy dependence and of the Euro-economic
area (EU-19) and European Union countries (EU-28) are shown in Fig. 1.12.
Fig. 1.12 – Portuguese energy dependence: a) along recent years and b) comparison
with EU-19 countries in year 2015. Data taken from Eurostat online database [48].
In 2015, Portugal was the 7th country with higher energy dependence
among the EU-19 and EU-28 countries (cf. Fig. 1.12b). None of the EU-19
Chapter 1
18
countries has a negative energy dependence (cf. Fig. 1.12b), all depending on
primary energy imports to satisfy the energetic needs.
The normalized consumption of primary energy (CPE) per type of source
in Portugal is depicted in Fig. 1.13a, for the period 2000-2015. The country
situation is compared with those from EU-19 and EU-28 group countries for the
year 2015 in Fig. 1.13b.
The National Energy Strategy approved by the Portuguese government in
2005 settled strategic policies such as the energy market liberalization, the
promotion of renewable energies and technologies with improved efficiencies
[49]. Consequently, in the following decade the oil share remarkably declined
(i.e. 14.6 %), being replaced by natural gas and renewable energies, whose
values increased 4 % and 9 %, respectively, while the coal share practically
remained constant in the same period (rise of only 1.6 %) (cf. Fig. 1.13a). Still,
oil, natural gas and coal together represented 78 % of the consumed primary
energy in 2015, a value slightly above EU-19 (72 %) and EU-28 (73 %) group
countries, whose patterns are identical (cf. Fig. 1.13b). Notably, the remaining
primary energy consumed was exclusively based on renewable energies (22 %),
which turned Portugal the country with 5th highest share of renewables amongst
EU-28 countries [50]. If only the consumption of primary energy for electricity
generation is considered, then the share of renewables is even higher; in 2015,
44.6 % of all electricity produced in Portugal was obtained from renewable
resources [51].
Nuclear has almost the same weight as renewable energies (ca. 13-15 %)
in EU-19 and EU-28 groups, although it is absent in some members like
Portugal.
General Introduction
19
Fig. 1.13 – Normalized primary energy consumption per type of source in Portugal: a) from 2000 to 2015 and b) comparison with EU groups in 2015. Author calculations based on data taken from Eurostat online database [48] and excluding the negligible contribution of non-renewable waste sources.
A description of the current situation regarding fossil fuels and renewable
energy contributions to the Portuguese energy sector is given in sections 1.2.2
and 1.2.3, respectively, with particular emphasis on renewable power
production, as it is one of the main building blocks of Power-to-Gas technologies.
1.2.2 Energy from fossil fuels
1.2.2.1 Oil
Portugal does not have indigenous oil reserves although regular onshore
and offshore exploration activities have been carried out since 1940. Therefore,
all the oil consumed by the country is imported. Table 1.3 lists the top-5 supplier
countries from 2014 to 2016.
Table 1.3 – Top-5 oil suppliers and corresponding share (based on data taken from [52]).
Top-5 2014 2015 2016
1st Angola (26.1 %) Angola (22.9 %) Angola (24.9 %) 2nd Saudi Arabia (12.6 %) Saudi Arabia (14.2 %) Russia (19.7 %) 3rd Algeria (9.9 %) Kazakhstan (10.6 %) Azerbaijan (11.1 %) 4th Kazakhstan (9.7 %) Algeria (9.5 %) Saudi Arabia (10.8 %) 5th Azerbaijan (9.2 %) Azerbaijan (9.0 %) Kazakhstan (9.3 %)
Imp. oil (106 ton) 7.5 (out of 11.17) 9.1 (out of 13.73) 10.7 (out of 14.09) Nr. oil suppliers 14 15 13
Year
2000 2005 2010 2015
CP
Ei /
CP
Eto
tal x 1
00
0
20
40
60
80
100
a)
b)
EU-28 EU-19 Portugal
CP
Ei /
CP
Eto
tal x 1
00
0
20
40
60
80
100
Oil
Natural gas
Coal
Renewables
Nuclear
Chapter 1
20
In the listed years, Portugal imported oil from 13 to 15 countries and the
top-5 oil suppliers were responsible for around 66-76 % of the total imported oil.
Angola is the major oil supplier with a contribution of ca. 25 %. Diversification of
oil suppliers along the years has contributed to assure reliable and secure
access to energy resources [53].
Final consumption of oil by activity sector is shown in Fig. 1.14. It shows
that the transportation sector is responsible for the largest share of final oil
consumption (ca. 75-79 %). The use of oil by all sectors declined in the 5-year
period, except the Agriculture/Forestry that remained practicaly constant. Within
the Industry sector, the non-metallic minerals industries (e.g. cement, glass, etc.)
are responsible for ca. 50-60 % of oil consumed by the industry sector.
Fig. 1.14 – Evolution of oil products final energy consumption in Portugal by activity
sectors from 2010 to 2015 (data taken from [54]).
1.2.2.2 Coal
After national coal production has ceased in 1994, Portugal dependence
on imported coal to secure its energy needs increased. Portugal imported 4.5 Mt
of coal from Colombia (88.1 %), United States (6.6 %), South Africa (3.5 %) and
General Introduction
21
Ukraine (1.8%) in 2014 [50]. Coal consumed in Portugal is of the bituminous type
and is used essentially for electricity generation in two coal-fired power plants
located in Sines (1250 MW) and Pego (620 MW). These plants act as a backup
system guaranteeing that power demand is fulfilled in periods of low renewable
power production. Coal consumption is particularly dependent on hydrological
conditions, namely when hydropower output is compromised in drought periods.
The remaining imported coal that is used for final energy consumption is
negligible when compared to the quantity used for electricity production. For
instance, in 2015, the quantity of coal used for electricity and for final energy
consumption was 3246 ktoe vs. 12 ktoe (136 PJ vs. 502 TJ), respectively. Only
the industry sector consumes coal for final energy purposes, namely the iron
and steel, chemical/petrochemical and non-metallic minerals industries (e.g.
cement industry), as shown in Fig. 1.15.
Fig. 1.15 – Coal use for final energy consumption by industry sub-sectors from 2010 to 2015 [54].
1.2.2.3 Natural Gas
Portugal absence of natural gas resources requires that the country needs
are completely satisfied though importation, similarly to oil and coal. Table 1.4
Year
2010 2011 2012 2013 2014 2015
Fin
al consum
ption
/ k
toe
0
10
20
30
40
Non-metallic minerals
Iron & Steel
Chemical/Petrochemical
Chapter 1
22
shows natural gas import origins and corresponding volumes regarding years
2014 and 2015.
Table 1.4 – Natural gas imports (106 m3) in the years 2014 and 2015 [55].
Origin 2014 2015
Natural gas (pipeline) 2736 3002 Algeria 2196 2111 Spain 535 891 Not specified 5
Liquefied natural gas (ships) 1523 1687 Algeria 102 210 Qatar 687 224 Nigeria 352 1166 Norway 80 80 Spain* 6 7 Trinidad and Tobago 223 Not specified 73
TOTAL 4259 4777 * LNG imported using tanker trucks.
Around 62-64 % of supplies were received through pipeline, while the
remaining part, in the liquefied state, was transported to Portugal in ships that
unload at the liquefied natural gas (LNG) terminal of Sines port, on the southern
part of Portugal. Only a negligible quantity (6-7x106 m3) was imported, using
tanker trucks, exclusively from Spain. The most important supplier is Algeria,
with a share ranging 49-54 % in the presented years, while Qatar and Nigeria
were the major suppliers of LNG.
Final energy consumption of natural gas by activity sector is shown in Fig.
1.16. The Industry sector accounts for the largest amount of natural gas
consumption (67-74 %), followed by Residential (16-19 %) and Services (13-14
%) sectors. The use of natural gas in Agriculture/Forest and Transportation
sectors is negligible and both sectors represent only ca. 1 % of the total final
consumption. Energy for transportation is secured essentially by oil products (as
shown in section 1.2.2.1) and natural gas represents a negligible quantity; for
instance, in 2015, the quantity of oil and natural gas consumed for transportation
were 6245 ktoe vs. 13 ktoe, respectively.
General Introduction
23
Fig. 1.16 - Natural gas for final energy consumption in Portugal by activity sectors from
2010 to 2015 (data taken from [54]).
Table 1.5 lists the natural gas consumption by the different industrial
sectors during 2005, 2014 and 2015.
Table 1.5 – Final energy consumption (ktoe) of natural gas by the Portuguese Industry (data taken from [54]).
Industry 2005 2014 2015 Δ
(2015/2005)
Paper, Pulp, and Print 38.1 90.2 111.6 2.93 Construction 5.8 13.1 14.5 2.48 Chemical and Petrochemical industry 64.4 142.1 152.7 2.37 Food and Tobacco 66.5 124.6 147.2 2.21 Non-ferrous metal industry 7.6 12.9 16.0 2.09 Machinery 21.3 32.6 36.0 1.69 Iron & steel industry 41.4 47.4 51.1 1.23 Textile and Leather 128.6 131.4 131.9 1.03 Non-metallic Minerals (e.g. cement) 516.6 426.3 441.6 0.85 Wood and Wood Products 9.7 9.0 7.8 0.81 Mining and Quarrying 6.3 5.1 4.6 0.74 Transport Equipment 28.8 14.9 16.6 0.58 Non-specified (Industry) 20.9 7.2 5.8 0.28
Total 956.1 1056.8 1137.4 1.19
Chapter 1
24
The non-metallic minerals sector is the most relevant consumer of natural
gas amongst the industries listed in Table 1.5. Its consumption is practically the
same as the sum of the 2nd, 3rd and 4th industrial sectors with higher natural gas
consumption. Table 1.5 highlights the growing importance of natural gas over
time and across several industrial activities. Natural gas consumption
significantly increased since 2005 in almost every industrial activity (see relative
variation in last column). Amongst them, due to their weight in current natural
gas consumption, it is noteworthy the remarkable demand increase by the
Paper, Pulp and Print (2.93), Chemical and Petrochemical (2.37) and Food and
Tobacco (2.21) industries.
Globally, the natural gas demand increased 19 % in the 2005-2015
decade, which reflects its growing importance in almost every activities of the
Portuguese industry sector.
1.2.3 Energy from renewables
The Renewable Energy Roadmap 21 sets a share for renewables
dissemination in the overall European energy mix of 20 % by 2020 [56]. Some
countries, where Portugal is included, have already reached or even surpassed
this target [41, 51]. The strategic effort to replace fossil fuels by renewable
energies has made Portugal to become one of Europe’s leaders in terms of use
of renewable energy sources (RES) [50]. Table 1.6 lists the amount (in ktoe) of
renewable energy produced in Portugal in the 2005-2015 decade.
Since 2006 that Portugal produces biodiesel which is incorporated almost
completely in diesel and only a small fraction (ca. 1 %) is directly sold to the
market. Soybean and, particularly, colza oils are the most used raw materials
[57].
General Introduction
25
Table 1.6 – Portuguese annual production of renewable energy (ktoe) from 2005 to 2015 [51].
Renewable energy type
2005 2007 2009 2011 2013 2015 Δ
(2015/2005)
Biofuels 0 162 226 330 274 321 - Electricity1 599 1265 1456 1872 2369 1927 3.2 Biomass2 2773 2891 3019 2571 2812 2781 1.0 Other renewables3 20 23 36 61 74 82 4.1
Total 3392 4342 4737 4835 5530 5110 1.5
More than a half of renewable energy produced in Portugal comes from
biomass, although that share declined from 82 % to 54 %, when comparing the
values of 2005 and 2015, respectively. The amount of biomass-based energy
produced remained practically constant along the 2005-2015 decade, while the
production of electricity increased 3.2 times, reaching a share of 38 % of the total
renewable energy produced in 2015 (cf. Table 1.6). It is note to worth that
electricity values shown in Table 1.6 consider only hydro, wind, photovoltaic and
geothermal contributions, although an important share of biomass is also used
for electricity generation in thermoelectric and co-generation plants.
Since renewable power is one of the main building blocks of the Power-to-
Gas concept, it will be specifically addressed in more detail below. The current
status of energy production from biofuels and biomass can be found elsewhere
(e.g. [50]).
The investment made on the different RES for power production is
highlighted through the analysis of the installed capacity (MW) values listed in
Table 1.7. The most established RES’s for electricity production in Portugal are
hydro and wind, both totaling 90 % of the installed capacity. Biomass is the third
RES with higher installed capacity, followed closely by photovoltaic, which
remarkably increased 150 times in the 2005-2015 decade. During this period,
1 Includes the contribution of hydro, wind, photovoltaic and geothermal power. 2 Includes the contribution of biogas. 3 Includes solar (for thermal purposes) and (low enthalpy) geothermal sources.
Chapter 1
26
wind energy installed capacity increased 4.7 times and was by far the type of
RES with the highest absolute variation (i.e. 3971 MW).
Table 1.7 – RES installed capacity (MW) for electricity production and corresponding variation in the 2005-2015 decade [51, 58].
RES 2005 2007 2009 2011 2013 2015 Δ
(2015/2005)
Geothermal 18 29 29 29 29 29 1.6 Photovoltaic 3 15 110 175 299 451 150.3 Biomass 429 449 518 712 718 726 1.7 Wind 1063 1699 3564 4378 4731 5034 4.7 Hydro 4816 4853 4883 5330 5533 6053 1.3
Total 6329 7045 9104 10624 11310 12293 1.9
Among biomass, it should be mentioned the evolution of biogas
production, whose installed capacity increased from 8 MW (in 2005) to 85 MW
(in 2015), while the capacity for energy generation from urban solid wastes only
increased 3 MW, reaching a total capacity of 89 MW in 2015.
The exploitation of the installed capacity to produce renewable power from
the different RES is provided in Table 1.8.
Table 1.8 – Annual renewable power production (GWh) [51, 58].
RES 2005 2007 2009 2011 2013 2015 Δ
(2015/2005)
Geothermal 71 201 184 210 197 204 2.9 Photovoltaic 3 24 160 282 479 799 266.3 Biomass 1651 1883 2086 2924 3052 3104 1.9 Wind 1773 4036 7577 9162 12015 11608 6.5 Hydro 5118 10449 9009 12114 14868 9800 1.9
Total 8616 16593 19016 24692 30610 25514 3.0
The values listed in Table 1.8 shows that renewable power production has
considerably increased in the 2005-2015 decade for all RES, as discussed in
the following sections.
General Introduction
27
1.2.3.1 Geothermal
Among the RES, high temperature geothermal resources are confined to
the Azores archipelago where this kind of energy plays an important role. Two
geothermal power plants in operation at S. Miguel island, corresponding to a
global installed capacity of 23 MW, are responsible for the production of 42 % of
the consumed electricity (i.e. around 22 % of the archipelago total demand).
Plans to increase the installed capacity up to 28.5 MW until 2019 have been
reported [59]. Fig. 1.17 shows the geothermal plant in Ribeira Grande.
The IEA reported that enhanced geothermal systems technology, which
uses thermal energy from high-temperature rocks (dry rocks) located at great
depths, may be suitable to explore the potential geothermal resources in
mainland and be tested in the future [50]. Still, Portugal is the 5th country among
IEA-29 members with the highest share of geothermal energy used for power
production [50].
Fig. 1.17 – Geothermal plant (13 MW) in Ribeira Grande (S. Miguel island, Azores) [60].
1.2.3.2 Photovoltaic
Power production in photovoltaic plants was negligible in 2005 but, after
an exceptional improvement, it reached 799 GWh in 2015. Portugal has the best
Chapter 1
28
yearly solar radiation in Europe after Cyprus, particularly in Alentejo region, in
the south of the territory, where the country has a current installed capacity of
162 MW (out of a total of 467 MW) [51, 61]. Moura photovoltaic plant (Fig. 1.18)
is in fact one of the biggest in world with an installed capacity of 46 MW.
Fig. 1.18 – Photovoltaic plant (46 MW) in Moura (Alentejo, south of Portugal) [62].
It is expected that solar energy will play an important role in decentralised
power production. A mini-generation programme created in 2011 has a target to
install approximately 250 MW of new capacity by 2020 in the country [50].
1.2.3.3 Biomass
Regarding biomass, the most common resources available in Portugal are
wood residues, animal waste and municipal solid waste [61]. It was estimated
that the country’s total biomass potential is 42.5 TWh/year, being municipal solid
wastes the main resource (17.0 TWh/year) [61]. Based on the estimated
potential, it was also reported that the use of municipal solid wastes, animal
manure and waste waters are still underexploited [61]. In 2015, 586 GWh of
power was generated from biogas and urban solid wastes (ca. 294 GWh each),
together representing 19 % of the total power produced from biomass (i.e. 3.10
TWh).
General Introduction
29
Power production from biomass is more developed in the center region of
the country, representing 62 % of the total power produced from biomass in
2015. Fig. 1.19 shows the largest biomass power plant of the country (30 MW)
located in Figueira da Foz.
Fig. 1.19 – Celbi biomass (forest residues) power plant (30 MW) located in Figueira da Foz (center of Portugal) [63].
1.2.3.4 Hydropower
Hydropower production takes place in 184 hydropower plants, considering
both large (≥ 10 MW) and small plants (<10 MW) [64]. Portugal mainland most
important river basins are: Lima, Cávado, Mondego, Tejo, Guadiana and,
particularly, Douro, which is responsible for more than a half of the hydropower
generated in the country [51]. Portugal has storage, run-of-the-river and
pumped-storage type hydropower plants. Storage plants accumulate large
quantities of water that can be used on the driest months, while run-of-river may
include a small storage capacity, but generally the turbines are in operation
depending on the river flow. Pumped-storage plants are conventional plants that
were modified to include a system for pumping water from a lower elevation
reservoir to a higher elevation, allowing to store energy in the form of
gravitational potential energy. It is used for load balancing, i.e. to bring back
water to the dam which was used previously in a period of high demand. This
Chapter 1
30
type of plants has recently gained a renewed attention as a way to store off-peak
electricity produced by renewable resources such as wind or photovoltaic.
Portugal has a pump-storage installed capacity of 2.44 GW [65].
Table 1.9 lists hydropower generation per type of plant and river basin in
the year of 2015. Up to the author’s knowledge, the values of power production
through pumped-storage plants per river basin are not publicly available, but a
global production of 1.16 TWh in 2015 was reported by Redes Energéticas
Nacionais (REN) [65].
Table 1.9 – Hydropower generation (GWh) by type of plant and river basin in 2015 [51]. River basin Storage Run-of-River Total %
Lima 484 5 489 5.0 Cávado 1180 29 1209 12.3 Douro 366 5422 5788 59.1 Mondego 322 88 410 4.2 Tejo 415 320 735 7.5 Guadiana 812 0 812 8.3 Others 0 355 355 3.6
Total 3581 6219 9800 100
Hydropower production in 2015 was affected by the hydrological
conditions and the amount of generated power was considerably lower than
typical values found, representing only 60 % of the power produced in 2014. Still,
Table 1.9 highlights the importance to the global hydropower production of
storage and run-of-river plants located in Cávado and Douro river basins,
respectively. Fig. 1.20 and Fig. 1.21 shows the biggest plants in Cávado and
Douro River, respectively.
General Introduction
31
Fig. 1.20 – Photograph of Caniçada plant (storage) in Cávado river (63 MW) [63].
Fig. 1.21 – Photograph of Miranda plant (run-of-river) in Douro river (369 MW) [63].
1.2.3.5 Wind
Portugal had 255 wind parks with 2604 turbines in operation in 2015,
corresponding to a total installed capacity of 5034 MW [51]. Fig. 1.22 shows
how the installed capacity, wind power production and annual equivalent hours
at full capacity (HFC) were distributed countrywide in 2015.
Chapter 1
32
Fig. 1.22 – Wind power installed capacity, generation and annual equivalent hours at full capacity (HFC) for year 2015 (adapted from [51]).
The HFC is determined dividing the generated output (MWh) by the
installed capacity (MW), providing a way to compare the performance of wind
turbines at different locations.
Fig. 1.22 highlights that wind power production is massively obtained in
the Center region, followed by the North, together representing 87% of the global
production [51]. Globally, wind power was generated in 2305 equivalent hours
General Introduction
33
at full capacity, being the North, Azores and Madeira the only regions with an
HFC lower than the global.
The performance of wind turbines can also be expressed in terms of a
capacity factor, where the number of equivalent hours at full load is normalized
by the number of hours available in a certain period, allowing to assess the
amount of power that was produced to the maximum possible. Table 1.10 lists
the capacity factor of each region for an ideal availability of 8760 hours (i.e. hours
available in one year and without excluding shutdown periods, for instance, for
maintenance).
Table 1.10 – Annual wind power capacity factor obtained in different regions of Portugal in 2015 (author calculations based on data taken from [51]).
Region Capacity factor
Center 0.44 Algarve 0.30 Alentejo 0.30 Lisbon 0.29 North 0.26 Azores 0.25 Madeira 0.19
The annual capacity factor ranged from 0.19 (Madeira) up to 0.44 (Center),
while the remaining regions had a value between 0.25 and 0.30. However, it is
noteworthy to mention that the capacity factor can be considerably different
depending on the time frame considered (year, month, etc.). Silva et al. [66] used
historic wind power generation time series (up to 5 years) and reported that the
capacity factor in Portugal can be 1.5 times higher in winter than in summer. The
analysis on a hourly basis also showed that wind generation is higher during
base-load and off-peak periods. Unfortunately, the amount of wind power that
was lost during these periods was not reported.
Chapter 1
34
1.2.4 Prospects for Power-to-Methane
1.2.4.1 Surplus renewable power
The IEA recently reported that instantaneous and daily renewable
electricity output in Portugal regularly exceeds national demand and that the
surpluses are either used in pumped-storage plants or exported [50].
Table 1.11 lists the amount of power that was consumed and produced in
the country through pumped-storage in the period between 2010 and 2015.
Table 1.11 – Pumped-storage power consumption and production (GWh) (author calculations based on data taken from [54]).
Pumped-storage power (GWh)
2010 2011 2012 2013 2014 2015
Consumed 512 737 1331 1459 1081 1460 Produced 399 575 1038 1138 843 1139
Table 1.11 shows that the amount of surplus power that was consumed in
pumped-storage plants increased 2.85 times in 2015 compared to the 2010
value; the storage round-trip efficiency is 78 %. However, it should be recalled
that pumped-storage plants represent huge capital investments and have
associated several environmental impacts that constitute important barriers
towards their adoption [67]. Besides, the storage capacity declines in periods of
high rainfall and that may compromise the ability to store all the exceeding power
being generated.
Another option is to export the surpluses to Spain. However, it has been
reported that when feed-in tariff supported technologies (like wind) generate
moments with high exports, Portugal provides cheap electricity to Spain, but
covers its high production costs internally, resulting in net costs to the
Portuguese Electricity System [68]. The feed-in tariff, a mechanism used by
countries to foster the diffusion of renewables, is always paid to power producers
independently of the power generated being used in Portugal or outside.
Moreover, the tariff was guaranteed to power producers for a 20-year period, the
longest reported among several European countries [68]. Therefore, it was
General Introduction
35
recommended that exports should be reduced at high-generation moments,
releasing the condition of feeding all renewables to the grid, and allowing for spill
of wind generation and/or invest in storage technologies [68].
Pumped-storage plants have been to date the main solution adopted in
Portugal. However, to avoid the associated limitations and low extension
potential of these large-scale plants, the use of decentralized units is foreseen
to increase the storage capacity of the country and to avoid exportations in
unfavorable terms. To this end, Power-to-Gas technologies, particularly Power-
to-Methane could be selected, offering the possibility to integrate the power and
gas grids. The methane produced in this way, often called substitute natural gas,
can thus be injected into the grid or stored in dedicated reservoirs, such as salt
caverns or in LNG tanks after being compressed, and used later when the
demand is higher.
1.2.4.2 CO2 sources and availability
The data presented in this section refers exclusively to GHG emissions of
CO2 and not to other GHG species and their corresponding CO2 equivalents.
Fig. 1.23 illustrates the evolution of CO2 emissions in the country from 2005 to
2015, using data taken from the GHG emissions inventory regularly performed
by the Portuguese Environmental Agency.
Fig. 1.23 shows that CO2 emissions are essentially divided in two
categories. The most relevant class is related to CO2 generated from the
combustion of fuels for energy production (ca. 90 % of all CO2 emitted in 2015)
and the other to CO2 produced in industrial processes. It should be recalled that
emissions from biomass combustion are excluded from the national emissions
totals since carbon released had been in fact fixed from atmosphere by the
photosynthetic process and when is burnt and returns to atmosphere does not
increase the atmospheric/biosphere CO2 pool [69]. However, CO2 emitted from
biomass combustion accounted 11.5 Mt in 2015 [69].
Chapter 1
36
Fig. 1.23 – CO2 emissions evolution resulting from energy production and industrial
processes and product use (adapted from [69]).
Regarding to the evolution of the emissions along time, while CO2
production from industrial processes remained almost constant in the reported
period, the emissions resulting from energy production declined from 69.1 Mt to
47.7 Mt between 2005 and 2014, due to a combination of increased renewable
power production and economic slowdown [50]. A slight increase of emissions
was observed in 2015 (total of 52 Mt), which was a reflex of higher primary
energy consumption for power production, namely of coal and natural gas.
Additionally, final energy consumption also increased, particularly in road
transport, natural gas and electricity [69].
Fig. 1.24 shows CO2 emissions related to energy production in 2015 by
type of sector and sub-sector. The largest point emission sources for energy
production considered were: 16 power plants, 2 oil refineries plants, 1 iron and
steel industry, 1 petrochemical unit, 1 carbon black industrial plant, 8 paper pulp
plants and 6 cement plants.
General Introduction
37
Fig. 1.24 – CO2 emissions due to energy production by sector (2015) (adapted from [69]).
The amount of CO2 emissions from Public Electricity and Heat Production
and Road Transportation stand out amongst other sub-sectors. Together, these
sub-sectors are responsible for 31 Mt of emitted CO2, which corresponds to 60
% of the national totals. Afterwards, the non-metallic minerals sub-sector was
responsible for the emission of 2981 kt of CO2, being the principal emitter
amongst the Manufacturing Industries and Construction sector. Besides energy
production, the mineral industry also stands out in what concerns CO2 emissions
resulting from industrial processes (cf. Fig. 1.25).
2015
Pu
blic
ele
ctr
icity
an
d h
ea
t p
rod
uctio
n
Pe
tro
leu
m r
efin
ing
Iro
n a
nd
ste
el
Ch
em
ica
ls
Pu
lp, p
ap
er
an
d p
rin
t
Fo
od
pro
ce
ssin
g, b
eve
rag
es a
nd
to
ba
cco
No
n-m
eta
llic m
ine
rals
Oth
er
Do
me
stic a
via
tio
n
Ro
ad
tra
nsp
ort
atio
n
Ra
ilwa
ys
Do
me
stic n
avig
atio
n
Co
mm
erc
ial/in
stitu
tio
na
l
Re
sid
en
tia
l
Ag
ricu
ltu
re/f
ore
str
y/fish
ing
CO
2 e
mis
sio
ns (
kt)
0
500
1000
1500
2000
2500
3000
14000
16000
Energy industries
Manuf. ind. & construction
Transport
Other sectors
Chapter 1
38
Fig. 1.25 – CO2 emissions from industrial processes and product use (non-energy use) in 2015 (adapted from [69]).
The mineral industry sector was responsible for the majority of emissions
(3794 kt), being cement production the most relevant activity with 2921 kt,
followed by other processes such as in fertilizers production (354 kt), lime
production (351 kt), and glass production (167 kt).
The second largest CO2 emitter sector is the Chemical Industry, namely
by the petrochemical and carbon black production activity with 650 kt. In the
past, ammonia production process was also an important contributor within this
sector, but since 2008 the only plant operating in the country was closed [69].
The non-energy products from fuels and solvent use comprises emissions
resulting from solvents, lubricants and paraffin wax uses by several industries
(e.g. plastics, wood products, rubber industry, metalworking industry, etc.). CO2
emissions associated to the metal industry come from secondary steel making.
2015
Mineral Industry
Chemical Industry
Metal Industry
Non-energy products from fuels and solvent use
CO
2 e
mis
sio
ns (
kt)
0
250
500
750
1000
3000
4000
General Introduction
39
From the analysis of Fig. 1.24 and Fig. 1.25, it can be concluded that
Portugal offers a wide variety of CO2 sources, essentially diluted in flue gas
streams, with a content between 5 vol. % (natural gas combustion) to 40 vol. %
(e.g. cement) that can be selected to couple in Power-to-Methane applications,
requiring however a previous CO2 capture/purification stage to separate it from
other contaminants. Criteria to select the best CO2 source for Power-to-Methane
applications include:
- Continuous access to CO2 in a stream having low concentrations of
severe catalyst and sorbent poisons (e.g. H2S);
- Proximity to the national natural gas grid for methane injection to
avoid/minimize storage and transportation costs;
- Proximity to renewable electricity plants that will power the water
electrolysis unit in periods where production exceeds demand, minimizing
distribution losses;
- Interest on recycling of the methane produced, for instance, if the
selected site has a natural gas co-generation plant;
- Interest on recycling oxygen produced during water electrolysis to the
process, which would further benefit the whole process from, at least, the
economic point-of-view.
1.2.4.3 Natural gas grid
Portugal has a well-established natural gas storage, transportation and
distribution infrastructure (cf. Fig. 1.26).
The main pipeline goes next to the coast from Sines until Valença do
Minho (and onwards to Spain), where the main natural gas consumption points
are located. It has several branches and 2 lines towards the interior of the
country, one of which that ends in Campomaior, making the connection to the
Spanish pipeline in Badajoz (cf. Fig. 1.26). This interconnection allows Portugal
to receive up to 3.5x109 m3/year of natural gas from Spain with origin in Algeria,
while the interconnection in the north (Valença do Minho/Tuy) has a lower
Chapter 1
40
capacity (0.8x109 m3/year). Still, the highest entry capacity to the grid is provided
by the LNG terminal in Sines (i.e. 5.3x109 m3/year) [50].
Fig. 1.26 – Portugal natural gas storage and transportation infrastructure [65].
Both interconnections with Spain are fully reversible and a third one in
Bragança/Zamora is planned. This connection (dashed line in Fig. 1.26) would
allow to comply with the European Regulation 994/2010 that establishes the
dispositions required for security of supply and the correct functioning of the
natural gas market, being identified by the European Commission as a project
of common interest [50, 55]. Currently, the national pipeline network has an
General Introduction
41
extension of 1375 km and contains 202 pipeline stations: 46 block valve stations,
70 junction stations, 84 gas regulation and metering systems and 2 custody
transfer stations [50].
Sines terminal receives LNG from large vessel ships with a capacity from
45x103 up to 216x103 m3. These ships unload into 3 LNG storage tanks having
a combined capacity of 390x103 m3, corresponding to ca. 242x106 m3 of natural
gas [50, 65] – cf. Fig. 1.27. The terminal is equipped with 5 vaporizers that use
sea water as thermal fluid to gasify LNG, which is further compressed to 77 barg
and injected into the gas grid [50]. The terminal facilities also includes a filling
station that may load up to 4500 tanker trucks a year to transport LNG to
locations which are not covered by the pipeline transmission network [50].
Fig. 1.27 – Two aerial views of the LNG reception, storage and regasification facilities at Sines terminal [70].
Another fundamental component of the national natural gas grid is the
combined underground storage capacity of 333x106 m3 provided by 6 salt
caverns located in Carriço, strategically placed in the middle of the main high-
pressure pipeline (cf. Fig. 1.26) [65, 71]. The reasons for its construction were:
1) the storage of strategic reserves and 2) to balance supply and demand,
namely it’s seasonal and daily fluctuations, thus securing natural gas supply [71].
Chapter 1
42
The caverns are located in the Lusitanian basin, in particular in the Monte
Real salt structure, which has a length of 27 km, a width of 6 km and 2 km of
height (cf. Fig. 1.28) [71].
Fig. 1.28 – Illustration of the Monte Real salt structure in the Lusitanian basin. Reproduced with permission from Ref. [71]. © KBB Underground Technologies GmbH.
Carriço underground storage facilities are positioned next to the sea
because during the construction of the caverns large amounts of brine were
produced and could be discharged to the sea. Around 60-80 % of the produced
brine was disposed to the Atlantic Ocean and the remaining purified by Renoeste
to produce salt with 99.5 % purity [71]. Fig. 1.29 shows an aerial view of the
Carriço facilities next to the coast and the situation of 4 existing salt caverns.
General Introduction
43
Fig. 1.29 – Aerial view of Carriço storage facilities and illustration showing the situation of 4 existing salt caverns (out of 6). Reproduced with permission from Ref. [71]. © KBB Underground Technologies GmbH.
Carriço storage facilities allows for gas injection and withdrawal at 110x103
Nm3/h and 300x103 Nm3/h, respectively. Before injection into the grid, the gas is
filtered, to remove solid and liquid particles, compressed and dehydrated in a
vertical absorber (the maximum final gas moisture content is 40 ppmv) [71].
The LNG terminal and Carriço salt caverns provide a total storage capacity
of 575x106 m3. Considering that consumption in 2016 was 4.6x109 m3 [65], the
existing combined capacity allows to stock the equivalent to the amount
consumed by the country in 46 days.
Chapter 1
44
The development plan of REN, the operator of the gas network, considers
the construction of 25 caverns in Carriço, which would increase the storage
capacity up to 1.25x109 m3 [72], but expansion plans were reported to be
currently under review [50].
The country underground storage capacity may be even higher based on
the available area of salt deposits (cf. Fig. 1.30), particularly next to the coast,
but up to the author knowledge few studies were done or are publicly available
where such potential is assessed (e.g. [72]).
Fig. 1.30 – European salt geological map showing salt deposits, gas storage and brine production sites. Reproduced with permission from Ref. [71]. © KBB Underground Technologies GmbH.
In the work by Nunes [72], the author assessed the underground storage
potential in caverns belonging to other salt structures of the Lusitanian basin
(see Fig. 1.28). Several criteria were adopted to choose the best locations.
General Introduction
45
These included rejecting zones that were in close distance to airports, roads and
houses, inside protected areas, far away to the sea and gas grid or not in a plain
field. Afterwards, 3 zones were elected: Nazaré, Caldas da Rainha and Peniche.
The study considered a similar volume and distance between caverns as it
occurs in Carriço and an underground storage potential of 1x109 m3 was
estimated. Still, only in one of the areas belonging to Peniche was possible to
have more than one cavern next to each other (in this case 2 caverns). In another
scenario, the minimum distance to roads was limited to highways and railways,
which changed completely the storage potential to a global value of 1.65x109
m3, through 24 caverns in Nazare (1.2x109 m3), 8 in Peniche (0.4x109 m3) and 1
in Caldas da Rainha (0.05x109 m3) [72]. Hence, a potential storage capacity of
3.14x109 m3 is envisaged, safeguarding 249 days of consumption (based on
2016 data). However, the preliminary assessment of the underground storage
potential made by Nunes [72] should be complemented with the necessary
environmental impact and economic studies.
1.2.4.4 Summary
The present analysis of the Portuguese energy sector highlights that the
country remains with a high dependence on fossil fuels to afford its energetic
needs. However, renewables contribution, particularly for electricity generation,
has made Portugal the 4th EU-18 country with the highest incorporation of
renewables in power production in 2015 (i.e. 44.6 %), a value that reached 57
% in 2016 [65].
So far, the country’s options to manage with the inevitable energy
surpluses that are generated by renewables were to invest in pumped-hydro
storage plants or through exportation to Spain. While these options have their
own pros and cons, as mentioned previously, another option would be to
chemically store those surpluses as a gas, being methane an appropriate choice
following the increasing natural gas consumption of the country (i.e. particularly
by the industry sector with an increase of 19 % from 2005 to 2015).
Chapter 1
46
Methane could be later reconverted in natural gas power plants in periods
of high electricity demand, used in the chemical industry or exported to Spain.
In the latter case, the third interconnection with Spain of the gas grid (see Fig.
1.26) would increase the exporting capacity, which could be further increased as
soon as projected connections are built to link the Iberian Peninsula with France
(see Fig. 1.31) [50].
Fig. 1.31 – Map of the European gas pipeline network and LNG terminals. Reproduced with permission from Ref. [71]. © KBB Underground Technologies GmbH.
Still, natural gas dependence for power production will increase in Portugal
following the decommissioning of Sines and Pego coal power plants by 2021
[73]. Storing surplus renewable electricity as methane would also allow to
diversify natural gas provision, minimizing the dependence and risk of shortage
supply from foreign countries, as it is advised by the Portuguese Directorate-
General for Energy and Geology [73]. Besides, Power-to-Methane applications
in Portugal would benefit from the high quantity of installed wind power capacity
General Introduction
47
within close distance to potential gas storage facilities (cf. Fig. 1.32), making
Portugal a predestined country to implement PtG technologies, as long as
adequate and nearby CO2 sources are also available [74].
Fig. 1.32 – Map showing the wind power parks (WPP) situated within 5 km of the natural gas network; salt deposit formations of the Lusitanian basin are also shown. White colour shows the most attractive power-to-gas plant positions. Reproduced with permission from Ref. [74]. © 2015 IEEE.
Chapter 1
48
1.3 Objective and outline of the thesis
The main objective of this PhD work was the study and development of a
Carbon Capture and Utilization (CCU) technology that can be inserted in a
Power-to-Gas process able to, simultaneously, remove CO2 from flue gas, and
further convert it to methane in a hybrid sorptive reactor unit. The in situ
conversion of captured CO2 avoids the costs associated with its compression
and transport to a methanation reactor.
The reactor bed contains a sorbent material to separate CO2 from other
species present in the flue gas and also a catalyst to accelerate the methanation
reaction during the desorption stage (via reactive regeneration with renewable-
based H2). For the process to run continuously, at least two reactors in parallel
working under cyclic operation are required; while one reactor is performing the
sorption stage, the other is running the reactive regeneration step. This approach
is illustrated in Fig. 1.33.
Fig. 1.33 – Power-to-Gas concept: system boundaries with proposed technological approach for CO2 capture and conversion.
General Introduction
49
Each step (sorption and reaction) was firstly studied individually and,
afterwards, merged in the same unit to proof-the-concept of such integrated
process.
The following paragraphs provide the thesis outline.
Chapter 2 presents a thermodynamic analysis performed using the
software Aspen Plus to check the limitations and optimal conditions when
producing methane (or methanol) using CO2 from flue gas exhaust streams
without previous removal of unnecessary species.
Chapter 3 addresses the synthesis, characterization and testing of
hydrotalcite-based sorbents for CO2 capture at high temperature. After a detailed
physical and chemical characterization, the best sorbent was selected through
the determination of the sorption equilibrium isotherms of all materials at 300 ºC.
Afterwards, its stability towards cyclic operation was checked and the sorption
and desorption kinetics modelled.
Chapter 4 concerns the determination of the reaction kinetics over a
commercial Ni-based catalyst supplied by Clariant. Kinetic tests were performed
using a fixed-bed reactor and an ideal feed stream of CO2 and H2 diluted in N2.
Conversion was measured against the space velocity at different temperatures
and after obtaining reaction rates under different conditions, non-linear
regression was used to fit the kinetic data to different rate expressions and to
obtain the corresponding parameters. Finally, the fixed-bed reactor was
modelled to validate the selected rate equation.
Chapter 5 deals with the proof-of-concept of the sorptive reactor for CO2
capture and conversion to methane using pellets of a commercial potassium
promoted hydrotalcite (from Sasol) and spheres of the commercial Ni-catalyst
tested before (from Clariant). The sorptive reactor was assessed in terms of CO2
sorption capacity, CO2 conversion, CH4 productivity and CH4 purity under
different operation conditions. The major drawbacks encountered were listed
and possible solutions for further reactor improvement suggested.
Chapter 1
50
Chapter 6 focuses on the assessment of the effect of CO and CO2 on H2
permeation through a dense palladium-based membrane. This kind of
membranes can be used for radial feeding of H2 along the sorptive reactor tube
length, to better control the H2/CO2 ratio during reactive regeneration (and
accelerate this stage). This was one of the suggestions made in Chapter 5 to
avoid undesirable CO formation.
Chapter 7 presents the general conclusions of this work and some
suggestions for future work.
General Introduction
51
1.4 Acronyms
Acronym Description
AEL Alkaline water electrolysis CCU Carbon capture and utilization EU-19 Euro economic zone countries EU-28 European Union countries EXP Primary energy exports GHG Greenhouse gas GIC Gross inland energy consumption HFC Annual equivalent hours at full capacity IEA International Energy Agency IGCCC Integrated Coal Gasification combined cycle IMB International Maritime bunkers IMP Primary energy imports LNG Liquefied natural gas MEA Monoethanolamine MOF Metal-organic framework NGCC Natural gas combined cycle PCC Post-combustion capture PEM Proton exchange membrane electrolysis PtG Power-to-Gas REN Redes Energéticas Nacionais RES Renewable energy source SERP Sorption-enhanced reaction processes SNG Substitute (or synthetic) natural gas SOEC Solid oxide electrolysis SRM Steam reforming of methane reaction WGS Water gas shift reaction
Chapter 1
52
1.5 References
[1] G.A. Olah, A. Goeppert, G.K.S. Prakash, Beyond Oil and Gas: The Methanol
Economy, Wiley - VCH Verlag GmbH & Co, Weinhein, 2006.
[2] G. Zakrzewska-Trznadel, European Workshop "CO2: From Waste to Value",
available at http://ec.europa.eu/research/industrial_technologies/pdf/co2-
workshop-30032011_en.pdf, date of last access: 17/01/2018.
[3] J. Wilcox, Carbon capture, Springer, New York, 2012.
[4] E.A. Quadrelli, G. Centi, J.L. Duplan, S. Perathoner, Carbon dioxide
recycling: Emerging large-scale technologies with industrial potential,
ChemSusChem, 4 (2011) 1194-1215.
[5] G. Centi, S. Perathoner, Opportunities and prospects in the chemical
recycling of carbon dioxide to fuels, Catalysis Today, 148 (2009) 191-205.
[6] J. Ma, N. Sun, X. Zhang, N. Zhao, F. Xiao, W. Wei, Y. Sun, A short review of
catalysis for CO2 conversion, Catalysis Today, 148 (2009) 221-231.
[7] C. Song, Global challenges and strategies for control, conversion and
utilization of CO2 for sustainable development involving energy, catalysis,
adsorption and chemical processing, Catalysis Today, 115 (2006) 2-32.
[8] M. Aresta, A. Dibenedetto, Utilisation of CO2 as a chemical feedstock:
opportunities and challenges, Dalton Transactions, (2007) 2975-2992.
[9] International Energy Agency, CO2 emissions from combustion of fossil fuels
: highlights 2016., 2016.
[10] E. Favre, Membrane processes and postcombustion carbon dioxide
capture: Challenges and prospects, Chemical Engineering Journal, 171 (2011)
782-793.
[11] E.R. Van Selow, P.D. Cobden, P.A. Verbraeken, J.R. Hufton, R.W. Van Den
Brink, Carbon capture by sorption-enhanced water-gas shift reaction process
using hydrotalcite-based material, Industrial and Engineering Chemistry
Research, 48 (2009) 4184-4193.
General Introduction
53
[12] E.L.G. Oliveira, C.A. Grande, A.E. Rodrigues, Effect of catalyst activity in
SMR-SERP for hydrogen production: Commercial vs. large-pore catalyst,
Chemical Engineering Science, 66 (2011) 342-354.
[13] K.B. Lee, M.G. Beaver, H.S. Caram, S. Sircar, Novel thermal-swing
sorption-enhanced reaction process concept for hydrogen production by low-
atemperature steam-methane reforming, Industrial and Engineering Chemistry
Research, 46 (2007) 5003-5014.
[14] A.E. Rodrigues, L.M. Madeira, Y.-J. Wu, R. Faria, Sorption-Enhanced
Reaction Processes, World Scientific Publishers, 2017.
[15] A.A. Olajire, CO2 capture and separation technologies for end-of-pipe
applications – A review, Energy, 35 (2010) 2610-2628.
[16] D. Aaron, C. Tsouris, Separation of CO2 from flue gas: A review, Separation
Science and Technology, 40 (2005) 321-348.
[17] H. Gupta, L.S. Fan, Carbonation-calcination cycle using high reactivity
calcium oxide for carbon dioxide separation from flue gas, Industrial and
Engineering Chemistry Research, 41 (2002) 4035-4042.
[18] R.H. Perry, D.W. Green, Perry's Chemical Engineers' Handbook, McGraw
Hill, New York, 1999.
[19] Z.F. Cui, Y. Jiang, R.W. Field, Fundamentals of pressure-driven membrane
separation processes, in: Z.F. Cui, H.S. Muralidhara (Eds.) Membrane
Technology: a practical guide to membrane technology and applications in food
and bioprocessing, Butterworth-Heinemann, Oxford, Uk, 2010.
[20] C.E. Powell, G.G. Qiao, Polymeric CO2/N2 gas separation membranes for
the capture of carbon dioxide from power plant flue gases, Journal of Membrane
Science, 279 (2006) 1-49.
[21] L.M. Robeson, The upper bound revisited, Journal of Membrane Science,
320 (2008) 390-400.
[22] M. Mulder, Basic Principles of Membrane Technology, 2nd ed., Kluwer
Academic Publishers, Dordrecht, 1996.
Chapter 1
54
[23] J. Huang, J. Zou, W.S.W. Ho, Carbon dioxide capture using a CO2-selective
facilitated transport membrane, Industrial and Engineering Chemistry Research,
47 (2008) 1261-1267.
[24]
www.netl.doe.gov/File%20Library/research/coal/carbon%20capture/FE000243
8.pdf, date of last access: 17/01/2018
[25] Q. Wang, J. Luo, Z. Zhong, A. Borgna, CO2 capture by solid adsorbents and
their applications: current status and new trends, Energy and Environmental
Science, 4 (2011) 42-55.
[26] P. Li, J. Chen, J. Zhang, X. Wang, Water stability and competition effects
toward CO2 adsorption on metal organic frameworks, Separation & Purification
Reviews, 44 (2015) 19-27.
[27] M.G. Beaver, H.S. Caram, S. Sircar, Sorption enhanced reaction process
for direct production of fuel-cell grade hydrogen by low temperature catalytic
steam-methane reforming, Journal of Power Sources, 195 (2010) 1998-2002.
[28] E.L.G. Oliveira, C.A. Grande, A.E. Rodrigues, CO2 sorption on hydrotalcite
and alkali-modified (K and Cs) hydrotalcites at high temperatures, Separation
and Purification Technology, 62 (2008) 137-147.
[29] Y. Ding, E. Alpay, Equilibria and kinetics of CO2 adsorption on hydrotalcite
adsorbent, Chemical Engineering Science, 55 (2000) 3461-3474.
[30] M.K. Ram Reddy, Z.P. Xu, G.Q. Lu, J.C. Diniz Da Costa, Influence of water
on high-temperature CO2 capture using layered double hydroxide derivatives,
Industrial and Engineering Chemistry Research, 47 (2008) 2630-2635.
[31] Z. Yong, V. Mata, A.E. Rodrigues, Adsorption of carbon dioxide onto
hydrotalcite-like compounds (HTlcs) at high temperatures, Industrial and
Engineering Chemistry Research, 40 (2001) 204-209.
[32] M.K. Ram Reddy, Z.P. Xu, G.Q. Lu, J.C.D. Da Costa, Layered double
hydroxides for CO2 capture: Structure evolution and regeneration, Industrial and
Engineering Chemistry Research, 45 (2006) 7504-7509.
General Introduction
55
[33] C.T. Yavuz, B.D. Shinall, A.V. Iretskii, M.G. White, T. Golden, M. Atilhan,
P.C. Ford, G.D. Stucky, Markedly improved CO2 capture efficiency and stability
of gallium substituted hydrotalcites at elevated temperatures, Chemistry of
Materials, 21 (2009) 3473-3475.
[34] M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner, P. Markewitz, T.E.
Müller, Chemical technologies for exploiting and recycling carbon dioxide into
the value chain, ChemSusChem, 4 (2011) 1216-1240.
[35] N. MacDowell, N. Florin, A. Buchard, J. Hallett, A. Galindo, G. Jackson, C.S.
Adjiman, C.K. Williams, N. Shah, P. Fennell, An overview of CO2 capture
technologies, Energy & Environmental Science, 3 (2010) 1645-1669.
[36] E.A. Quadrelli, G. Centi, Green Carbon Dioxide, ChemSusChem, 4 (2011)
1179-1181.
[37] M. Aresta, A. Dibenedetto, A. Angelini, Catalysis for the valorization of
exhaust carbon: from CO2 to chemicals, materials, and fuels. Technological use
of CO2, Chemical Reviews, 114 (2014) 1709-1742.
[38] C. Song, Introduction to hydrogen and syngas production and purification
technologies, in: K. Liu, C. Song, V. Subramani (Eds.) Hydrogen and syngas
production and purification technologies, John Wiley & Sons, Inc., New Jersey,
2010, pp. 1-12.
[39] M. Specht, J. Brellochs, V. Frick, B. Stürmer, U. Zuberbühler, The Power to
Gas Process: Storage of Renewable Energy in the Natural Gas Grid via Fixed
Bed Methanation of CO2/H2, in: T.J. Schildhauer, S.M.A. Biollaz (Eds.) Synthetic
Natural Gas from Coal, Dry Biomass, and Power-to-Gas Applications, John
Wiley and Sons, Inc., Hoboken, New Jersey, 2016.
[40] S. Schiebahn, T. Grube, M. Robinius, V. Tietze, B. Kumar, D. Stolten, Power
to gas: technological overview, systems analysis and economic assessment for
a case study in Germany, International Journal of Hydrogen Energy, 40 (2015)
4285-4294.
Chapter 1
56
[41] M. Götz, J. Lefebvre, F. Mörs, A. McDaniel Koch, F. Graf, S. Bajohr, R.
Reimert, T. Kolb, Renewable Power-to-Gas: A technological and economic
review, Renewable Energy, 85 (2016) 1371-1390.
[42] T.J. Schildhauer, Introdutory remarks, in: T.J. Schildhauer, S.M.A. Biollaz
(Eds.) Synthetic Natural Gas from Coal, Dry Biomass, and Power-to-Gas
Applications, John Wiley and Sons, Inc., Hoboken, New Jersey, 2016.
[43] News: Germany/Power-to-gas: Clariant supplies SNG catalyst for first
commercial CO2 methanation plant, Oil Gas European Magazine, 39 (2013) 217.
[44] www.audi-mediacenter.com/en/press-releases/audi-e-gas-plant-stabilizes-
electrical-grid-4499, date of last access: 17/01/2018.
[45] http://www.kit.edu/kit/english/pi_2014_16120.php, date of last access:
17/01/2018.
[46] http://biocat-project.com, date of last access: 17/01/2018.
[47] M. Bailera, P. Lisbona, L.M. Romeo, S. Espatolero, Power to Gas projects
review: Lab, pilot and demo plants for storing renewable energy and CO2,
Renewable and Sustainable Energy Reviews, 69 (2017) 292-312.
[48] Eurostat, Environment and energy database,
http://ec.europa.eu/eurostat/data/database, date of last access 17/01/2018.
[49] International Energy Agency, Energy policies of IEA countries: Portugal
2009 Review.
[50] International Energy Agency, Energy policies of IEA countries: Portugal
2016 Review.
[51] Direção Geral de Energia e Geologia, Renováveis - Estatísticas Rápidas,
abril 2017 (en: Portuguese Directorate-General for Energy and Geology,
Renewables- brief statistics, April 2017).
[52] Direção Geral de Energia e Geologia, Fatura Energética Portuguesa 2016
(en: Portuguese Directorate-General for Energy and Geology, Portuguese
Energetic Invoice 2016).
[53] J. Amador, Energy production and consumption in Portugal: stylized facts,
Economic Bulletin and Financial Stability Report Articles, (2010) 69-83.
General Introduction
57
[54] Eurostat, Energy balances,
http://ec.europa.eu/eurostat/web/energy/data/energy-balances, date of last
access: 17/01/2018.
[55] Direção Geral de Energia e Geologia, Relatório sobre Avaliação dos Riscos
que afetam o aprovisionamento de Gás Natural em Portugal- Período 2017-
2025 (en: Portuguese Directorate-General for Energy and Geology, Assessment
Report of Natural Gas provision in Portugal - 2017-2025 period).
[56] European Commission, Renewable Energy Road Map - Renewable
energies in the 21st century: building a more sustainable future, Brussels, 2007.
[57] Entidade Nacional para o Mercado de Combústiveis, Mercado de
Biocombustíveis em Portugal: A atuação da ENMC desde abril de 2015. (en:
National Entity for Fuels Market, Biofuels market in Portugal: ENMC action since
April 2015).
[58] Direção Geral de Energia e Geologia, Renováveis - Estatísticas Rápidas,
outubro 2014 (en: Portuguese Directorate-General for Energy and Geology,
Renewables- brief statistics, October 2014).
[59] L.C. J.M. Carvalho, J.C. Nunes, M.R Carvalho, J. Garcia, R. Cerderia,
Portugal Country Update 2015, in: Proceedings World Geothermal Congress
2015 Melbourne, Australia 2015.
[60] http://siaram.azores.gov.pt/energia/geotermia/_texto.html, date of last
access: 17/01/2018.
[61] S. Ferreira, E. Monteiro, P. Brito, C. Vilarinho, Biomass resources in
Portugal: Current status and prospects, Renewable and Sustainable Energy
Reviews, 78 (2017) 1221-1235.
[62] http://www.portal-energia.com/a-maior-central-solar-do-mundo-no-
alentejo, date of last access: 17/01/2018.
[63] http://www.a-nossa-energia.edp.pt/centros_produtores, date of last access:
17/01/2018.
[64] C. Teotónio, P. Fortes, P. Roebeling, M. Rodriguez, M. Robaina-Alves,
Assessing the impacts of climate change on hydropower generation and the
Chapter 1
58
power sector in Portugal: A partial equilibrium approach, Renewable and
Sustainable Energy Reviews, 74 (2017) 788-799.
[65] Redes Eletricas Nacionais, Technical Data 2016,
http://www.centrodeinformacao.ren.pt/PT/InformacaoTecnica/DadosTecnicos/
REN%20Dados%20T%C3%A9cnicos%202016.pdf, date of last access:
17/01/2018.
[66] B. Silva, M. Marques, J.C. Matos, A. Rodrigues, R. Pereira, A. Bastos, P.
Cabral, F. Saraiva, Wind power in Portugal: from the potential to the integration
in the electric grid, in: European Wind Energy Conference and Exhibition 2010,
EWEC 2010, 2010, pp. 4508-4515.
[67] J.A.R. Marques, Hidroelectricidade e Barragens Reversíveis: Panorama
actual, in: Departamento de Geociências, Ambiente e Ordenamento do
Território, Universidade do Porto, Porto, 2015.
[68] I. Peña, I. Azevedo, L.A.F. Marcelino Ferreira, Lessons from wind policy in
Portugal, Energy Policy, 103 (2017) 193-202.
[69] T.C. Pereira, T. Seabra, A. Pina, P. Canaveira, A. Amaro, M. Borges, R.
Silva, Portuguese National Inventory Report on Greenhouse Gases, 1990-2015
submmitted under the United Nations Framework Convention on Climate
Change and the Kyoto Protocol, Portuguese Environmental Agency, Amadora,
2017.
[70] http://www.ign.ren.pt/en/terminal-de-gnl3, date of last access: 17/01/2018.
[71] http://www.kbbnet.de/wp-content/uploads/2015/10/150821-Carrico-
Brochure.pdf, date of last access: 17/01/2018.
[72] P.V.C. Nunes, Potencial de armazenamento subterrâneo em cavidades
salinas de gás natural em Portugal, in: Departamento de Engenharia de Minas
e Georrecursos, Instituto Superior Técnico, Lisboa, Portugal, 2010.
[73] Direção Geral de Energia e Geologia, Relatório de Monitorização da
Segurança de Abastecimento do Sistema Elétrico Nacional 2013-2030 (en:
Portuguese Directorate-General for Energy and Geology, Monitoring Report of
the Eletric National System Security of Supply 2013-2030).
General Introduction
59
[74] F. Heymann, R. Bessa, Power-to-Gas potential assessment of Portugal
under special consideration of LCOE, in: 2015 IEEE Eindhoven PowerTech,
PowerTech 2015, 2015.
61
Chapter 2. Direct CO2 hydrogenation to methane and methanol from post-combustion exhaust streams – a thermodynamic study
The contents of this chapter were adapted from: C.V. Miguel, M.A. Soria, A. Mendes,
L.M. Madeira, Direct CO2 hydrogenation to methane or methanol from post-combustion
exhaust streams – A thermodynamic study. Journal of Natural Gas Science and
Engineering, 2015, 22, 1-8.
The conversion/utilization of waste carbon dioxide is seen as a
complementary option to the well-known capture and storage strategies
(CCS) to substantially reduce atmospheric CO2 (environmental concern).
This approach is attractive regarding CCS strategies because CO2 can be
transformed into a valuable chemical (economic benefit). Among the
options available, methane and methanol are important chemicals that
could be obtained from CO2 hydrogenation and used for energy
production/storage or as intermediaries to other chemicals.
A thermodynamic analysis regarding the hydrogenation of CO2 into
CH4 or CH3OH was carried out. The analysis was performed to check the
limitations and optimal conditions when converting CO2 from flue gas
exhaust streams without previous removal of unnecessary species present
in significant amounts (e.g. N2, H2O and O2). The present analysis supports
that, from the thermodynamic point of view, the conversion of CO2 into CH4
is favoured in comparison to the CH3OH valorisation strategy, for the
considered pressure and temperature ranges.
Chapter 2
62
2.1 Introduction
Carbon dioxide is the end-product of the largest-volume and most globally
applied chemical reaction, the combustion of hydrocarbons and biomass, and it
is well known the growing concern about reducing CO2 emissions due to its
enormous contribution to the greenhouse effect [1]. The Kyoto Protocol has
created the market for carbon credits, a crucial mechanism for valuating CO2
emissions and thus incorporating the pollution effect in the cost structure of the
corporation’s economy [2]. However, top-polluting countries such as Canada
and USA are out of the agreement; moreover, according to the International
Energy Agency, the top 10 polluting countries represent around 2 3 of world CO2
emissions [3]. This means that any solution to solve the carbon dioxide problem
will always depend on the compromise of these countries. So, economic benefits
should be considered together with environmental concerns. In this regard, in
recent years there has been a focus on developing different possibilities for CO2
recycling as complement of the well-known capture, sequestration and storage
approaches, particularly, its conversion into added value products [4-6]. This
new paradigm considers CO2 as chemical feedstock (value) and not only as a
waste that needs to be treated (cost) [7].
Recent works provide together a comprehensive state-of-the-art of the
options available for CO2 valorisation and utilization, including the necessary
timeframe for development, the time of sequestration, the economic
perspectives, etc. [4-7]. Among the options presented in those works are CO2
hydrogenation into methane (Eq. (2.1)) or methanol (Eq. (2.2)), the first being
also known as Sabatier reaction or CO2 methanation.
-12 2 4 2 298 KCO + 4H CH + 2H O Δ = - 165 kJ molH (2.1)
-12 2 3 2 298 KCO + 3H CH OH + H O Δ = - 49.4 kJ molH (2.2)
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
63
These two options, however, require expensive H2 which, in turn, is
preferentially produced worldwide using non-renewable feedstock’s, being the
steam methane reforming the most developed and commercialized technology
[8]. In this regard, these routes should be viable in view of CO2 emissions
abatement only when H2 is produced from renewable resources, such as water
electrolysis. Moreover, the energy required for the electrolysis should be also
renewable for the global process to truly allow reducing CO2 emissions [5, 9].
So, in this case, important chemicals such as methane or methanol could be
produced using renewable resources (for H2 production) and waste CO2.
The conversion of CO2 into methanol (reaction 2) has, compared to the
methanation process, the advantage of consuming less hydrogen (see also the
stoichiometry of reaction 1). Moreover, methanol has a higher energy density, is
easier to store and can be used, for example, in the synthesis of important
chemicals such as formaldehyde, methyl tertiary-butyl ether (MTBE), among
others [10]. These advantages are pointed by various personalities as the driving
force for the conversion/recycling of CO2 into methanol, thus alleviating the
dependence on fossil fuels while simultaneously reducing the emission of
greenhouse gases. Among such personalities is the winner of the Nobel Prize in
Chemistry in 1994, Prof. Olah, who clearly supports a strategy of "Methanol
Economy" [11]. On the other hand, the “Power-to-Gas” concept can be a very
interesting way to chemically store the off-peak electricity generated in wind
power stations in the form of methane, which can be further integrated with the
already existing natural gas infrastructures [12-14], as long as the exit process
stream properties complies with the specifications required for natural gas
transport in pipelines.
The necessary CO2 is available from a large variety of emission sources.
However, the International Energy Agency reported that the majority of the world
CO2 emissions arise from post-combustion sources related to electricity and
heat production (41 % in 2010), particularly, from coal-fired power plants and the
combustion of oil or gas, respectively 43 %, 36 % and 20 % of the electricity
Chapter 2
64
related CO2 emissions. Previous works addressed, from the thermodynamic
standpoint, the CO2 valorisation into CH4 [15] or CH3OH [16]; in particular, Gao
et al. [15] studied the effect of species present in syngas produced by coal or
biomass gasification, where CO is the major species present (rather than CO2).
In this work, however, CO2 valorisation was assessed considering its direct
conversion from a real coal-fired power plant exhaust stream. Moreover, the
effect of pressure, temperature, H2/CO2 ratio and the presence of major co-
existing species present in flue gas streams (N2, O2 and H2O) was systematically
assessed regarding CO2 conversion, product yield and selectivity.
2.2 Methodology
Thermodynamic analysis was performed using the Gibbs reactor model
(RGibbs) available in the Aspen Plus software from AspenTech. RGibbs models
simultaneous phase and chemical equilibria minimizing the Gibbs free energy
and does not require the specification of the reactions involved and their
stoichiometry [17, 18].
The total Gibbs free energy of a system is given by the sum of the chemical
potential of all the N species [15, 19, 20]:
1
N
T i i
i
G n
(2.3)
where in is the number of moles of species i, which chemical potential,
i , is
given by:
0
0ln i
i fi
i
fG RT
f (2.4)
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
65
where 0
fiG is the standard Gibbs function of species i formation, R is the ideal
gas constant, T is the absolute temperature and if and 0
if stand for the fugacity
and standard fugacity, respectively. For the reaction equilibrium in the gas
phase:
i i if y P (2.5)
0 0
if P (2.6)
where iy is the mole fraction of species i, i is the fugacity coefficient, P is the
pressure of the system and P0 is the standard pressure. So, combining Eqs.
(2.3)-(2.6) and applying Lagrange multipliers, used to incorporate the constraints
related to conservation of the total amount of individual chemical elements into
the body of the problem, the constrained function to be minimized, objf , is [18]:
obj
1 1
0
01 1 1
ln
m N
T j j i ji
j i
N m Ni i
i fi j j i ji
i j i
f G b n a
y Pn G RT b n a
P
(2.7)
where j is the Lagrange multiplier, jb the total amount of element j in the
mixture, jia the number of atoms of element j in species i. Whenever the
presence of solid carbon was considered in simulations, Eq. (2.7) should be
changed to the following one [20]:
0 0
obj ( )01 1 1
lnN m N
i ii fi j j i ji C fC s
i j i
y Pf n G RT b n a n G
P
(2.8)
Chapter 2
66
Fugacity was estimated using the Soave-Redlich-Kwong (SRK) equation of state
contained in the Aspen Plus database, as suggested elsewhere for similar
conditions [15].
The starting gas composition (mixture 1 in Table 2.1) was obtained from a
typical coal-fired power station flue gas stream. The inlet gas compositions used
for calculations (mixtures 2-10, Table 2.1) were established to independently
analyse, in realistic scenarios, the effects of the H2/CO2 and H2O/CO2 molar
ratios, as well as the influence of the presence of H2O and/or O2, which is
discussed in section 3. One should consider that H2 is not originally present in
flue gas streams (mixture 1 in Table 2.1) and should be added. So, the resulting
feed stream (mixtures 2-10) has a composition different from the considered flue
gas one.
Table 2.1 – Inlet compositions (mol. %) of the Gibbs reactor used in simulations.
Mixture CO2 H2O N2 O2 H2 2
2
H
CO
2
2
N
CO
2
2
H O
CO
2
2
O
CO
1a) 13.0 20.5 63.0 3.5 0.0 0 4.8 1.6 0.3 2 10.2 0.0 49.2 0.0 40.6 4 4.8 0.0 0.0 3 11.3 0.0 54.8 0.0 33.9 3 4.8 0.0 0.0 4 12.8 0.0 61.8 0.0 25.5 2 4.8 0.0 0.0 5 8.8 13.8 42.4 0.0 35.0 4 4.8 1.6 0.0 6 9.9 0.0 47.9 2.7 39.5 4 4.8 0.0 0.3 7 8.6 13.5 41.4 2.3 34.2 4 4.8 1.6 0.3 8 9.6 15.1 46.4 0.0 28.9 3 4.8 1.6 0.3
9b) 8.8 4.8 51.0 0.0 35.4 4 5.8 0.5 0.0 10b) 9.7 5.2 56.0 0.0 29.1 3 5.8 0.5 0.0
a) Flue gas composition taken from [21]. b) Stream composition based on flue gas composition from [22].
Table 2.2 shows the main reactions considered for the analysis of our
results (reactions 1-7) as well as other possible reactions that may occur in small
extent.
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
67
Table 2.2 – Reactions considered in the present thermodynamic study.
Reaction formula ΔH298 K (kJ mol-1) Reaction description
Main reactions
1 CO2+4H2 ⇌ CH4+2H2O -165.0 CO2 hydrogenation to CH4
2 CO2+3H2 ⇌ CH3OH+H2O -49.4 CO2 hydrogenation to CH3OH
3 CO2+2H2 ⇌ C+2H2O -90.1 CO2 reduction
4 CO2+H2 ⇌ CO+H2O 41.2 Reverse water-gas shift
5 CH4+2O2 ⇌ CO2+2H2O -803.0 CH4 oxidation
6 CH4+ 1 2⁄ O2 ⇌ CO+2H2 -36.0 CH4 partial oxidation
7 H2+ 1 2⁄ O2 ⇌ H2O -241.8 H2 oxidation
Other possible reactions
8 C + 1 2⁄ O2 ⇌ CO -110.5 Coke partial oxidation
9 C+O2 ⇌ CO2 -393.5 Coke complete oxidation
10 CO + 1 2⁄ O2 ⇌ CO2 -283.0 CO oxidation
11 CH4+CO2 ⇌ 2CO+2H2 247.4 Dry reforming of CH4
12 CH4 ⇌ C+2H2 74.9 CH4 cracking
13 CO + 3H2 ⇌ CH4+H2O -206.2 CO hydrogenation to CH4
14 CO + H2 ⇌ C+H2O -131.3 CO reduction
15 2CO ⇌ CO2+C -172.5 Boudouard reaction
2.3 Results and Discussion
2.3.1 Strategies for CO2 valorisation: CH4 or CH3OH?
As stated before, thermodynamic analysis was performed using the
continuous Gibbs reactor model. Carbon dioxide equilibrium conversion (Eq.
(2.9)) was determined for hydrogenation into methane or methanol as a function
of pressure and temperature (Fig. 2.1),
2 2
2
2
100
in out
CO CO
CO in
CO
F FX
F
(2.9)
Chapter 2
68
In this equation, F stands for the molar flow rate at the inlet (in) or outlet
(out) of the Gibbs reactor. In this section only the main reactions (reactions 1
and 2, Table 2.2) were considered, which means that the occurrence of
secondary reactions was, at this stage, discarded. Broad ranges of pressure and
temperature were set for the calculations, including those found in industrial
catalytic reactors operating these reactions.
As shown in Fig. 2.1, for either route of CO2 valorisation its conversion
decreases with reaction temperature, because both processes are exothermic.
In addition, total pressure has a positive effect, because in either case there is a
decrease in the total number of moles (from reactants to products – cf. equations
1 and 2). However, data presented in Fig. 2.1 clearly evidences that CO2
conversion into CH3OH (Fig. 2.1b) requires high pressures, particularly in the
temperature range where active catalysts operate in industry (see dashed areas
in Fig. 2.1) so that significant conversions can be achieved. Thus, since post-
combustion flue gases are typically at the atmospheric pressure, the CH3OH
route requires compression of the feed stream, which increases operation costs,
when compared to the methane route. For instance, at 250 ºC the conversion of
CO2 in CH4 production is almost complete at the atmospheric pressure while for
the CH3OH route it is practically null. Based on these evidences, in the following
sections it was chosen to analyse into more detail only the valorisation strategy
of CO2 hydrogenation into CH4.
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
69
Fig. 2.1 – Carbon dioxide conversion obtained for hydrogenation reactions into: a) CH4
and b) CH3OH. Dashed areas show typical operation temperature ranges of industrial
catalysts.
Chapter 2
70
2.3.2 CO2 methanation: effect of pressure, temperature
and H2/CO2 ratio
The effects of pressure, temperature and H2/CO2 molar ratio on CO2
conversion (Eq. (2.9)), product selectivity (Eq. (2.10)) and yield (Eq. (2.11)) were
investigated on CO2 methanation. In this section, the methanation main reaction
1 and secondary reactions 3 and 4 (see Table 2.2) were considered. Reactions
5-7 (Table 2.2) were not accounted because O2 was not considered to be
present in the feed stream (this will be addressed in Section 2.3.3); other
reactions present in Table 2.2 occur in small/very small extent.
Selectivity to carbon-containing species (CH4, CO and C):
2 2CO CO
100out
ii in out
FS
F F
(2.10)
Yield of carbon-containing species (CH4, CO and C):
2
i 100out
i
in
CO
FY
F (2.11)
Since H2 is an expensive compound, the highest H2/CO2 ratio considered
was 4, corresponding to the stoichiometry of the desired reaction (Eq. (2.1)).
Fig. 2.2a shows that, for a given pressure, CO2 conversion decreases with
temperature but increases with the H2/CO2 ratio. In Fig. 2.2b it can be observed
that CH4 selectivity increases with both the temperature and the H2/CO2 ratio,
except at low pressures and for H2/CO2=2. For a H2/CO2 ratio of 4 the selectivity
is almost complete (Fig. 2.2b), except for low pressures and high temperatures
where a small fraction of carbon monoxide is formed (YCO< 1 %) through the
endothermic reverse water gas shift reaction. So, when a ratio of 4 is used, CO2
methanation is the most favoured reaction. This is also supported by the fact
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
71
that the outlet molar flow rate of H2O is twice the value of CH4, which obeys to
the stoichiometry of the CO2 methanation and means that these species are not
being produced nor consumed through other reactions.
Fig. 2.2 – Effect of the H2/CO2 ratio (mixtures 2, 3 and 4 in Table 2.1), temperature and pressure on: a) CO2 conversion, b) CH4 selectivity, c) CH4 yield and d) Carbon yield.
40
50
60
70
80
90
100
510
1520
2530
150200
250300
350
XC
O2
/ %
P /
bar
T / ºC
a)
40
50
60
70
80
90
100
H2/CO2=4
H2/CO2=3
H2/CO2=2
20
30
40
50
60
70
80
90
100
510
1520
2530
150200
250300
350
SC
H4 / %
P / ba
r
T / ºC
b)
20
30
40
50
60
70
80
90
100
H2/CO2=4
H2/CO2=3
H2/CO2=2
20
30
40
50
60
70
80
90
100
5
10
15
2025
30
150200
250300
350
YC
H4 / %
P / ba
r
T / ºC
c)
20
30
40
50
60
70
80
90
100
H2/CO2=2
H2/CO2=4
H2/CO2=3
0
10
20
30
40
50
60
5
10
15
2025
30
150200
250300
350
YC
/ %
P /
bar
T / ºC
d)
0
10
20
30
40
50
60
H2/CO2=2
H2/CO2=3
H2/CO2=4
Chapter 2
72
As mentioned above, when lower H2/CO2 ratios (i.e. 3 and 2) are used,
CO2 conversion decreases and a similar trend is observed for both ratios along
the temperature and pressure values. This was somehow expected because
more CO2 will remain available (unconverted) since H2 is not fed with the
required stoichiometry for CO2 methanation (Eq. (2.1)). Interestingly, for a ratio
of 3 it is observed that for temperatures below ca. 250 ºC the conversion of CO2
starts to increase (Fig. 2.2a), which is accompanied by a decrease of CH4
selectivity (Fig. 2.2b) and yield (Fig. 2.2c). This trend suggests that below 250
ºC another compound is being produced using CO2 as reactant; from Fig. 2.2d
it becomes clear that such species is solid carbon (coke) through CO2 reduction
(reaction 3 in Table 2.2). Its exothermic nature also supports that carbon
formation is favoured at lower temperatures, as observed. Moreover, this is also
corroborated by the fact that the H2O molar flow rate is higher than twice the
value of CH4, meaning that H2O is produced not only through CO2 methanation
(Eq. (2.1)) but also via the CO2 reduction reaction.
2.3.3 Direct CO2 methanation from coal-fired power plant
(CF-PP) flue gas streams
2.3.3.1 Effect of H2O
Water vapour is an important component present in coal-fired power plant
(CF-PP) flue gas streams. Herein, its effect was studied considering that H2O is
coming exclusively from the post-combustion stream. One should note that H2
is not present in CF-PP flue gas streams (cf. mixture 1 in Table 2.1) and so it
should be added. To observe the effect of H2O in the methanation reaction it was
considered that oxygen was absent. Still, in Section 2.3.3.3 the simultaneous
effect of H2O and O2 will be addressed.
Fig. 2.3 shows contour plots illustrating the effect of H2O content on CO2
conversion as a function of pressure, temperature and H2/CO2 ratio; H2O content
was varied to address realistic limits. For both H2/CO2 ratios of 4 and 3, the
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
73
addition of H2O decreases the equilibrium conversion, which was expected
because H2O is a product of the CO2 methanation reaction (Eq. (2.1)), thus
limiting the forward reaction. The decrease of CO2 conversion due to the
presence of H2O is more notorious for a H2/CO2 ratio of 3. Still more importantly,
CH4 was the only carbon-containing product formed (nearly 100 % selectivity
was observed) in the presence of H2O for the pressure and temperature ranges
considered when a H2/CO2 ratio of 4 or 3 is used.
Fig. 2.3 – Contour plots showing the effect caused by H2O content on CO2 conversion considering different H2/CO2 ratios of 4 (left column) and 3 (right column). For the compositions of the different mixtures, please refer to Table 2.1.
Chapter 2
74
In Fig. 2.4 it is shown that when a H2/CO2 molar ratio of 3 is used, the
presence of H2O markedly inhibits carbon formation at temperatures below 250
ºC. In this regard, industrial operation can be carried out using a H2/CO2 ratio of
3 without carbon formation if H2O is present. Obviously, this option represents a
CO2 equilibrium conversion decrease of at least 25 % as compared to the
situation of H2/CO2=4 without H2O (cf. Fig. 2.3a and Fig. 2.3f). Moreover, due to
the exothermic nature of the CO2 hydrogenation into CH4 (ΔH298 K = -165.0 kJ
mol-1), the addition/presence of H2O can be also an interesting strategy to control
the heat produced in a catalytic reactor, as suggested elsewhere [23].
Fig. 2.4 – Carbon yield (YC) obtained in the absence (mixture 3) or in the presence of H2O (mixture 8) considering a feed stream with a H2/CO2 ratio of 3. For the compositions of the different mixtures, please refer to Table 2.1.
2.3.3.2 Effect of O2
Oxygen is commonly found in flue gas streams despite in small contents
(< 5 %). In this regard it is important to assess its possible impact on CO2
methanation. A stream with 2.7 mol % of O2 (mixture 6 in Table 2.1) was
0
5
10
15
20
25
30
35
5
10
15
20
25
30
150200
250300
350
YC /
%
P /
bar
T / ºC
0
5
10
15
20
25
absence of H2O (mix. 3)
(surface above)
presence of H2O (mix. 8)
(surface below)
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
75
considered at the inlet of the Gibbs reactor. The possible oxidations (reactions
5-7) are depicted in Table 2.2. Oxygen can react with H2 and CH4, both species
participating in CO2 methanation, the first as a reactant and the latter as a
product. The presence of only 2.7 mol % of O2 in the feed stream produces a
significant decay of CO2 conversion, as shown in Fig. 2.5, which can be mainly
explained by the formation of CO2 and H2O. In fact, in certain conditions the CO2
conversion shifts from 95 % to ca. 82 %, while in the range of industrial nickel
catalysts operation (i.e. at 400 ºC and atmospheric pressure) the conversion can
be as low as 70 %.
Fig. 2.5 – Contour plots showing the effect of the presence of O2 on CO2 conversion. For the compositions of the different mixtures, please refer to Table 2.1.
2.3.3.3 Simultaneous effect of H2O and O2
Fig. 2.6 shows that 2COX decreases when both H2O and O2 are present
in the feed stream (trends for 4CHY are the same as for
2COX because methane
selectivity was nearly 100 % in both plots – data not shown). The presence of
H2O, as mentioned in Section 2.3.3.1, hinders CO2 methanation in the forward
direction (Eq. (2.1)) because it is a product of the reaction. Additionally, the
presence of O2 promotes the oxidation of species such as CH4 or H2 (reactions
Chapter 2
76
5-6 and 7 in Table 2.2, respectively), leading to the formation of H2O and CO2.
Therefore, the presence of O2, although in a small percentage (2.3 mol % - cf.
mixture 7 in Table 2.1) should be avoided because it leads to the parallel
consumption of a reactant (H2) and of the desired product (CH4). In fact, in the
conditions tested, all the O2 fed is consumed. Above 300 ºC and at the
atmospheric pressure a slight formation of CO (YCO < 1 %) through reverse water
gas shift (reaction 4 in Table 2.2) is observed (data not shown for brevity
reasons). However, CO formation can be suppressed increasing the pressure.
In this case, CO2 methanation (Eq. (2.1)) is favoured and overlaps the reverse
water-gas shift (reaction 4), which is not influenced by the pressure since the
reaction takes place without change in the number of moles.
Fig. 2.6 – Contour plots showing the effect of simultaneous presence of H2O and O2 on CO2 conversion for a H2/CO2 ratio of 4. For the compositions of the different mixtures, please refer to Table 2.1.
2.4 Technological implementation
From the technological point of view, i.e. for process implementation, two
related problems are identified: i) the presence of substances in the flue gas,
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
77
namely O2 (as above-mentioned) and N2 (that simply acts as diluents), and ii)
the existence of un-reacted CO2 in the reactor outlet, due to thermodynamic
restrictions. Integration of a methanation catalyst with a CO2-selective sorbent in
a single mixed bed is anticipated to allow overcoming all these drawbacks
simultaneously. In a first stage, the unit (operating in sorption mode) is fed with
the flue gas, so that CO2 is selectivity retained while the other species leave the
bed, up to almost sorbent saturation (in fact up to CO2 breakthrough from the
column). In the second stage, (renewable) H2 is fed to the bed, reacting with the
previous concentrated CO2, in a so-called reactive regeneration approach. To
operate on a continuous basis of CO2 capture and conversion at least two beds
are thus necessary operating in complementary stages: when one CO2-
saturated bed is being regenerated (with a hydrogen-containing stream) and
CH4 is being produced, the other one is capturing carbon dioxide; after
regeneration of the 1st column, the bed is able again to capture more carbon
dioxide (cf. Fig. 2.7).
Fig. 2.7 – Illustration of the integrated reactive regeneration process for CO2 capture and conversion to CH4 with two beds operating 180º out of phase with each other.
Chapter 2
78
The reactive regeneration concept was already proved for other
applications, namely in sorption-enhanced reactors for H2 production through
steam methane reforming [24], and the proof-of-concept towards CO2
methanation is presented in Chapter 5. Finally, it should be mentioned that water
must be removed from the exit stream to obtain a product with quality compatible
with existing natural gas infrastructures.
2.5 Conclusions
The present work compared, from the thermodynamic standpoint, the
carbon dioxide valorisation to methanol and methane. The option for CO2
conversion into CH3OH requires harsh operation conditions when compared to
the CH4 route, namely in terms of pressure. Thus, in the near term, CO2
methanation seems to be an easier pathway for CO2 valorisation, while research
on the development of active catalysts at lower temperatures for CO2
hydrogenation to CH3OH is required.
This study also allowed concluding that CO2 methanation can take place
with complete (~100%) methane selectivity and with high methane yields in the
temperature and pressure ranges of industrial catalysts operation, as long as the
H2/CO2 ratio is 4. A preliminary stage for O2 removal from post-combustion
exhaust streams is required due to its detrimental impact on CO2 conversion,
apart from security reasons. On the other hand, the effect caused by H2O is not
so pronounced as for O2. In fact, water presence can substantially inhibit coke
formation whenever a H2/CO2 ratio of 3 is used, thus opening a wider range of
operation conditions available for the catalytic conversion of CO2 into CH4.
Moreover, the addition of water can bring additional advantages regarding
temperature control of the methanation reactor due to the exothermic nature of
the Sabatier reaction.
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
79
2.6 References
[1] E.A. Quadrelli, G. Centi, Green carbon dioxide, ChemSusChem, 4 (2011)
1179-1181.
[2] Kyoto Protocol to the United Nations Framework Convention on Climate
Change, (1998).
[3] International Energy Agency, CO2 emission from fuel combustion: highlights,
OECD/IEA, 2012.
[4] E.A. Quadrelli, G. Centi, J.L. Duplan, S. Perathoner, Carbon dioxide
recycling: Emerging large-scale technologies with industrial potential,
ChemSusChem, 4 (2011) 1194-1215.
[5] M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner, P. Markewitz, T.E. Müller,
Chemical technologies for exploiting and recycling carbon dioxide into the value
chain, ChemSusChem, 4 (2011) 1216-1240.
[6] A.A. Olajire, Valorization of greenhouse carbon dioxide emissions into value-
added products by catalytic processes, Journal of CO2 Utilization, 3-4 (2013) 74-
92.
[7] M. Aresta, A. Dibenedetto, A. Angelini, The changing paradigm in CO2
utilization, Journal of CO2 Utilization, 3–4 (2013) 65-73.
[8] V. Subramani, P. Sharma, L. Zhang, K. Liu, Catalytic steam reforming
technology for the production of hydrogen and syngas, in: K. Liu, C. Song, V.
Subramani (Eds.) Hydrogen and Syngas Production and Purification
Technologies, John Wiley & Sons, Inc., 2010, pp. 14-126.
[9] S.K. Hoekman, A. Broch, C. Robbins, R. Purcell, CO2 recycling by reaction
with renewably-generated hydrogen, International Journal of Greenhouse Gas
Control, 4 (2010) 44-50.
[10] C. Higman, M.v.d. Burgt, Gasification, 2nd ed., Elsevier Inc., Burlington,
2008.
[11] G. A. Olah, A. Goeppert, G.K.S. Prakash, Beyond Oil and Gas: The
Methanol Economy, Wiley - VCH Verlag GmbH & Co, Weinhein, 2006.
Chapter 2
80
[12] L. Grond, P. Schulze, J. Holstein, Systems analyses Power to Gas: A
technology review, KEMA Nederland B.V., Groningen, 2013.
[13] News: Germany/Power-to-gas: Clariant supplies SNG catalyst for first
commercial CO2 methanation plant, Oil Gas European Magazine, 39 (2013) 217.
[14] S. Walspurger, G.D. Elzinga, J.W. Dijkstra, M. Sarić, W.G. Haije, Sorption
enhanced methanation for substitute natural gas production: Experimental
results and thermodynamic considerations, Chemical Engineering Journal, 242
(2014) 379-386.
[15] J. Gao, Y. Wang, Y. Ping, D. Hu, G. Xu, F. Gu, F. Su, A thermodynamic
analysis of methanation reactions of carbon oxides for the production of
synthetic natural gas, RSC Advances, 2 (2012) 2358-2368.
[16] W.-J. Shen, K.-W. Jun, H.-S. Choi, K.-W. Lee, Thermodynamic investigation
of methanol and dimethyl ether synthesis from CO2 Hydrogenation, Korean
Journal of Chemical Engineering, 17 (2000) 210-216.
[17] S. Adhikari, S. Fernando, S.R. Gwaltney, S.D. Filip To, R. Mark Bricka, P.H.
Steele, A. Haryanto, A thermodynamic analysis of hydrogen production by steam
reforming of glycerol, International Journal of Hydrogen Energy, 32 (2007) 2875-
2880.
[18] S.M. Walas, Phase equilibria in Chemical Engineering, Butterworth
Publishers, Boston, United States of America, 1985.
[19] M.A. Soria, C. Mateos-Pedrero, A. Guerrero-Ruiz, I. Rodríguez-Ramos,
Thermodynamic and experimental study of combined dry and steam reforming
of methane on Ru/ ZrO2-La2O3 catalyst at low temperature, International Journal
of Hydrogen Energy, 36 (2011) 15212-15220.
[20] Y. Li, Y. Wang, X. Zhang, Z. Mi, Thermodynamic analysis of autothermal
steam and CO2 reforming of methane, International Journal of Hydrogen Energy,
33 (2008) 2507-2514.
[21] K.A. Mumford, K.H. Smith, C.J. Anderson, S. Shen, W. Tao, Y.A.
Suryaputradinata, A. Qader, B. Hooper, R.A. Innocenzi, S.E. Kentish, G.W.
Stevens, Post-combustion capture of CO2: results from the solvent absorption
Direct CO2 hydrogenation to CH4 and CH3OH - a thermodynamic study
81
capture plant at Hazelwood power station using potassium carbonate solvent,
Energy & Fuels, 26 (2011) 138-146.
[22] M.T. Ho, G.W. Allinson, D.E. Wiley, Comparison of MEA capture cost for
low CO2 emissions sources in Australia, International Journal of Greenhouse
Gas Control, 5 (2011) 49-60.
[23] T.T.M. Nguyen, L. Wissing, M.S. Skjøth-Rasmussen, High temperature
methanation: Catalyst considerations, Catalysis Today, 215 (2013) 233-238.
[24] G.H. Xiu, P. Li, A.E. Rodrigues, Sorption-enhanced reaction process with
reactive regeneration, Chemical Engineering Science, 57 (2002) 3893-3908.
83
Chapter 3. High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
The contents of this chapter were adapted from: C.V. Miguel, R. Trujillano, V. Rives,
M.A. Vicente, A.F.P Ferreira, A.E. Rodrigues, A. Mendes, L.M. Madeira, High
temperature CO2 sorption with gallium-substituted and promoted hydrotalcites.
Separation and Purification Technology, 2014, 127, 202-211.
Hydrotalcites (HTC) and gallium-substituted hydrotalcites (HTC-
10Ga) were prepared by co-precipitation and modified by impregnation
with alkali (K and Cs) and alkaline-earth (Sr) metals. The materials were
tested and screened for CO2 sorption at 573 K. The results indicate that
the modification with potassium greatly enhanced the sorption capacity
when compared with the original materials (both HTC and HTC-10Ga). An
outstanding sorption capacity (2.01 mmol·cm-3 at 1.08 bar) was obtained
for the sample partially substituted with gallium and modified with
potassium (HTC-10Ga-20K). Moreover, this material was also submitted
to sorption-desorption cycles towards CO2 under low-pressure conditions
(<0.0001 bar to 0.15 bar) to assess its use in cyclic operation. An average
working capacity of 0.97 mmol·cm-3 and 0.25 mmol·cm-3 was obtained after
repeated cycles at 573 K and 473 K, respectively. The sorption kinetics
was assessed through uptake measurements, showing two parallel
resistances, which were well described by the proposed model.
Chapter 3
84
3.1 Introduction
The excessive anthropogenic greenhouse gases (GHG) emission is
causing negative changes on the planet climate such as global warming,
frequent events of heavy precipitation and heat waves [1]. Policies have been
undertaken by developed countries to control, reduce and tax the emission of
GHG like CO2. Carbon dioxide is a relevant GHG; in fact, around 30 x 1012 kg of
CO2 per year are emitted to the atmosphere arising from the combustion of fossil
fuels (coal, petroleum and natural gas) [2]. In this regard, there is a great interest
in developing technological solutions for large-scale CO2 capture, sequestration
or conversion/utilization, from both the environmental and economic point of
views [3].
Depending on the operating temperature, several materials (e.g. activated
carbons, zeolites, metal-organic frameworks, hydrotalcites, etc.) may be used to
capture CO2 by adsorption [4, 5]. Among them, hydrotalcites can be used at
intermediate temperatures (473-673 K). Hydrotalcite materials are layered
double hydroxides, which general formula is:
2+ 3+ -1- 22
M M OH A H Onx x x n
y (3.1)
where M2+ and M3+ are divalent (e.g. Mg2+, Ni2+, Zn2+, etc.) and trivalent (e.g.
Al3+, Ga3+, Mn3+, etc.) metal cations, respectively. An- is a charge balancing anion
(CO32-, Cl-, SO4
2-, etc.) located in the interlayer space and x is generally between
0.2–0.4 [6], while y is the number of moles of hydration water located also in the
interlayer space. The mineral known as hydrotalcite has Mg2+, Al3+ and CO32- as
divalent and trivalent cations and balancing anion, respectively. Hydrotalcites
can be found in nature or be synthesized. Currently, several thousands of tons
of hydrotalcites are yearly produced by several chemicals companies (e.g.
BASF, SASOL, Clariant, Kisuma Chemicals, Sakai Chemicals, etc.) which found
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
85
application as catalysts, precursors, catalyst support, adsorbents, polymer
stabilizers and antacids.
Regarding their application as an adsorbent for CO2 capture, it has been
reported that hydrotalcites exhibit a reasonably good sorption capacity, stability
and easy regeneration by temperature or pressure swing [5, 7-9]. Additionally,
results found in the literature show that the presence of water vapour also
benefits carbon dioxide sorption [10, 11]. This is particularly important for CO2
capture in applications were water vapour is present, as it is the case of coal-
fired power station exhaust streams (post-combustion), or in sorption enhanced
processes for H2 production (pre-combustion) [12-14]. On the other hand, it has
been reported that SOx species, also present in those streams, have a stronger
affinity to hydrotalcites when compared to CO2, despite exhibiting slower kinetics
[15]; in fact, in the work by Beaver [16] it was reported that CO2 sorption capacity
on a K-promoted hydrotalcite decreased ca. 80 % after exposure to CO2/SO2
mixtures at the end of eight successive sorption-desorption cycles.
In the work by Yavuz et al. [8], the effect of gallium substitution and
modification with potassium was firstly addressed and a markedly improved CO2
sorption capacity was observed (1.4 mmol∙g-1) at 473 K for a CO2 partial
pressure of 0.7 atm under dry conditions.
In this work the effect on CO2 sorption capacity caused by gallium partial
substitution and modification with potassium, cesium and strontium was
evaluated and compared to that shown by the original formulation. Along with a
detailed physicochemical characterization of all the prepared materials, sorption
equilibrium isotherms were determined at 573 K, a typical temperature found in
both post- and pre-combustion streams. Additionally, the most promising
material was also submitted to sorption-desorption cycles in the low-pressure
region (< 0.0001 bar to 0.15 bar) at 473 K and at 573 K. Finally, the sorption
kinetics was assessed for this material.
Chapter 3
86
3.2 Experimental
3.2.1 Chemicals and gases
Magnesium nitrate 6-hydrate (Sigma-Aldrich, 99 %), aluminium nitrate 9-
hydrate (Panreac, 99.82 %) and gallium nitrate hydrate (Aldrich, 99.99 %) have
been used as Mg, Al and Ga precursors, respectively. Anhydrous sodium
carbonate (Panreac, 98.88 %) was used in excess to assure that the charge
balancing anion was carbonate instead of nitrate. Sodium hydroxide (Panreac,
98.93 %) was used to keep a basic pH during synthesis.
Potassium acetate (Panreac, 99 %), cesium acetate (Aldrich, 99.9 %) and
strontium acetate (Aldrich, 99.995 %) were used as chemical promoters for
modifying the prepared hydrotalcite and the 10 % (mol percentage) gallium-
substituted hydrotalcite samples.
Carbon dioxide (99.99 %) and helium (99.999 %), both from L’Air Liquide,
were used in the sorption experiments.
3.2.2 Sorbents preparation
Hydrotalcites (HTC) and hydrotalcites partially substituted with gallium
(herein called HTC-10Ga) were modified by impregnation with aqueous
solutions of alkaline or alkaline-earth metal cations such as K+, Cs+ and Sr2+.
The targeted not promoted hydrotalcites were Mg2Al(OH)6(CO3)0.5·yH2O and
Mg2(Al0.9Ga0.1)(OH)6(CO3)0.5·yH2O. The two previously mentioned samples,
respectively HTC and HTC-10Ga, were prepared by the co-precipitation method.
In the work by Yong et al. [9] the CO2 sorption capacity was determined at 573
K for three commercial hydrotalcites from SASOL having different Mg/Al ratios:
0.5, 1.27 and 3. It was observed that an Mg/Al ratio of 1.27 had the highest
sorption capacity followed by 3 and, finally, by 0.5. However, the lowest limit of
the Mg/Al ratio for obtaining crystallographically pure hydrotalcite (2.0) is
attributed to electrostatic repulsion between neighbouring, octahedrally
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
87
coordinated, trivalent metals in the layers, which is unavoidable if Mg/Al is lower
than 2 [6]. For this reason, and based on the results obtained by Yong et al,
[9],molar ratios of 2 were chosen for Mg2+/Al3+ and Mg2+/(Al3++Ga3+), while for
the Ga-containing sample an Al:Ga ratio of 9:1 was also considered (see method
details in [6]) to check the effect of gallium presence.
The solution containing the divalent and trivalent cations was drop-wise
added to the basic solution containing NaOH and Na2CO3. The resulting solution
was continuously stirred during 17 h at room temperature and then the
suspension was centrifuged and the solids washed with distilled water for
removing nitrate anions and sodium cations. The samples were then dried at
308-323 K and crushed. Both samples (HTC and HTC-10Ga) were then calcined
at 673 K during 2 h in air and the loss of mass assessed in order to further
prepare the alkali modified samples with a pre-established stoichiometry. So,
part of the remaining HTC sample was modified with K while three parts of the
HTC-10Ga sample were modified with K, Cs or Sr (cf. Table 3.1) by impregnation
to accomplish a 20 wt. % loading. This content is typically found in other works
[7, 14, 17] following the important work by Hufton et al. [18], although it has been
reported that the sorption capacity at 673 K is not appreciably changed when
using different loadings [13, 18]. Impregnation was followed by drying (313-323
K), crushing, and calcination at 673 K during 2 h in air.
Table 3.1 – List of prepared samples.
Sample M2+ M3+ An- Promoter (wt. %)
HTC Mg Al - CO3 - -
HTC-20K Mg Al - CO3 K 20
HTC-10Ga Mg Al Ga CO3 - -
HTC-10Ga-20K Mg Al Ga CO3 K 20
HTC-10Ga-20Cs Mg Al Ga CO3 Cs 20
HTC-10Ga-20Sr Mg Al Ga CO3 Sr 20
Chapter 3
88
3.2.3 Sorbents characterization
The HTC and HTC-10Ga materials were analysed by thermogravimetric
analysis (TG/DTG) in a TA Instruments apparatus, Model SDT Q600, to study
their thermal stability. Each sample was heated at 10 K∙min-1 up to 1273 K in a
dynamic air atmosphere.
Powder X-ray diffraction (PXRD) patterns were recorded in a Siemens
D500 diffractometer equipped with DIFFRACT-AT software, using Cu K
radiation (λ= 1.5405 Å), a Ni filter and a graphite monochromator. The analysis
conditions were: 40 kV, 30 mA, a step size of 0.05° and a scanning speed of 2º
(2 theta) per minute.
Fourier Transformed Infrared Spectroscopy (FTIR) (Perkin-Elmer, 1600
Series) was performed in the range 4000-400 cm-1 using the KBr technique.
Scanning Electron Microscopy (SEM) was performed on a JEOL JSM 6301F
instrument coupled with Energy Dispersive Spectroscopy (EDS) (JEOL JSM
6301F/Oxford INCA Energy 350) to investigate the materials structure and
composition. The textural properties of the prepared materials were obtained by
means of nitrogen adsorption at 77 K (Micrometrics, model Gemini II), after
degassing the samples for 2 h at 383 K in a Flowprep 060 apparatus
(Micrometrics).
High temperature CO2 sorption measurements were performed using a
magnetic suspension balance from Rubotherm. First, the sample (ca. 1 g) was
heated overnight up to 585 K under vacuum (< 0.001 bar) followed by buoyancy
determination with He at 313 K, which allowed to determine the solid density
after the regeneration step. Equations for buoyancy correction can be found in
Appendix A. Subsequently, the sample was again heated up to the desired
temperature and CO2 fed to the chamber until reaching the required pressure for
measuring the first sorption equilibrium value. The uptake curve was recorded
by an automated data acquisition system. The equilibrium was considered to be
reached when a variation of less than 0.1 mg was observed during 20 min. For
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
89
screening purposes, the sorption equilibrium values were obtained at 573 K and
for CO2 pressures of ca. 0.05, 0.15, 0.45 and 1.10 bar. Afterwards, the most
promising material was submitted to sorption-desorption cycles at 573 K and 473
K and the reversible/steady working capacity determined. Typically, the carbon
dioxide partial pressure in post-combustion flue gases is 0.15 bar and so each
cycle consisted in increasing the pressure up to 0.15 bar followed by desorption
down to vacuum, except for cycle number 3 where the pressure was increased
up to 1.1 bar to check the effect of the feed pressure on the cyclic stability.
3.3 Results and Discussion
3.3.1 Sorbents physicochemical characterization
3.3.1.1 TG / DTG
Thermogravimetric analysis was performed to study the thermal
decomposition of HTC and HTC-10Ga fresh samples (Fig. 3.1); a total mass loss
of 55 % and 47 % was found, respectively. Both samples show three DTG peaks
related to the removal of hydration water and hydroxyl groups (at 353-373 K and
463 K) and carbonate anions (at 653 K), as discussed in detail elsewhere [19,
20]. As mentioned in Section 3.2.2, the fresh samples were calcined at 673 K for
further studies, as this treatment ensures that all water existing in the fresh
samples is removed. The used calcination temperature is in line with several
works [7, 8, 21] and was also chosen following the study by Ram Reddy et al.
[22], where it was observed that samples calcined at 673 K provided a higher
CO2 sorption capacity. Moreover, this is the highest operation temperature
expected to be used in most final applications (e.g. for pre-combustion CO2
capture).
Chapter 3
90
Fig. 3.1 – Thermogravimetric (TG) (solid lines) and derivative thermogravimetry (DTG)
(dotted lines) analysis of the fresh samples HTC and HTC-10Ga.
3.3.1.2 XRD
The XRD patterns of the samples calcined at 673 K corresponded to
mostly amorphous materials, where only weak and broad signals due to MgO
were observed; the pattern for the Ga-containing sample is shown in Fig. A.0.1
presented in Appendix A. This result agrees with the TG curve (Fig. 3.1), which
showed that at 673 K most of the removable mass has been eliminated (water
vapour and carbonates), leaving the mostly amorphous solid. However, the XRD
patterns evidenced that one year after their synthesis, calcination at 673 K and
storage, all the materials have a typical hydrotalcite structure, suggesting that
upon exposure to humid air the materials have reconstructed their layered
structure [23] (Fig. 3.2a).
The partial substitution with gallium led to an increased height intensity of
the d(003) peak (see patterns of samples HTC and HTC-10Ga in Fig. 3.2a).
Impregnation with cesium (HTC-10Ga-20Cs) did not result in large enough
crystals that could be identified by XRD, meaning that cesium is well dispersed
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
91
on the material. For the case of impregnation with strontium, a strontianite
crystalline phase (SrCO3) was identified (JCPDS file 5-0418), while K2CO3∙1.5
H2O (JCPDS file 11-0655) was verified in both samples impregnated with
potassium.
After CO2 sorption the XRD patterns showed important differences.
Representative diagrams are included in Fig. 3.2b. First of all, the maxima due
to the layered hydrotalcite structure (easily observed in the sample stored for
one year) have vanished completely for sample HTC-10Ga-20Sr or decreased
markedly their intensities for other samples (although the bands due to SrCO3
remained unaltered). This behaviour is due to two complementary facts: (i) the
CO2 capture experiments were carried out at 573 K (after outgassing the sample
at 585 K), a temperature high enough to destroy the layered structure, as
concluded from the TG results above mentioned (Fig. 3.1); (ii) despite the
presence of a high pressure of CO2 would hopefully give rise to recovering the
layered structure (the so-called “memory effect”) [24, 25], it should be stressed
that the experiments were carried out in the absence of water vapour, and thus
the hydrotalcite-like structure could not be recovered. When the sample HTC-
10Ga-20K was submitted to five consecutive CO2 sorption-desorption cycles the
XRD patterns did not show any significant difference (cf. Fig. 3.2b), although
some minor sharpening of the maxima was observed when the cycles were
carried out at 573 K, but not at 473 K. Probably, this treatment at the higher
temperature in some sort of way favoured an incipient crystallization of the
existing crystalline phases.
Chapter 3
92
Fig. 3.2 – XRD patterns of samples: a) one year after being calcined at 673 K and stored, and b) after CO2 sorption.
3.3.1.3 SEM / EDS
SEM images evidenced that the samples exhibit a not defined and/or
regular shape (Fig. 3.3), excepting HTC-10Ga-20Sr (pattern e)) and, particularly,
HTC-20K and HTC-10Ga-20K (patterns c) and f)), which also revealed the
presence of zones of the needle-like (Z1) or flat surface (Z3) type. This was
already observed by Oliveira et al. [7] using commercial hydrotalcites modified
with potassium. Interestingly, EDS analysis allowed to verify that the potassium
mass content at the flat surface zones is the same for both HTC-20K and HTC-
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
93
10Ga-20K samples (26 wt. %). This content is also much higher than on the bulk
material (7 wt. %) and needle-like zones (5 wt. % vs. 14 wt. %, respectively) (see
Fig. A.0.2 in Appendix A). Although incipient wetness impregnation usually leads
to a homogeneous distribution of the active phase, potassium clearly has a
marked preference to deposit in certain zones of the supports.
Fig. 3.3 – SEM images of the prepared materials: a) HTC, b) HTC-10Ga, c) HTC-20K, d) HTC-10Ga-20Cs, e) HTC-10Ga-20Sr and f) HTC-10Ga-20K.
Chapter 3
94
3.3.1.4 Physical adsorption of nitrogen at 77 K
The specific surface areas were determined by physical adsorption of N2
at 77 K (cf. Table 3.2). Type II isotherms were obtained (see Fig. A.0.3 in
Appendix A), which means that the resulting materials are
macroporous/mesoporous solids; the determined volume of microporous (<
0.005 cm3∙g-1) can be neglected.
The specific surface area increased after calcination, particularly in the
case of the HTC sample (ca. 50 % of increase). This is related to dehydration
followed by collapse of the layered structure that occurs upon calcination at 673
K [26]. Nevertheless, the modification with alkali metals (K and Cs) resulted in
lower specific surface areas, being this fact more evident when the hydrotalcites
were modified with potassium. In fact, both samples modified with potassium
(HTC-20K and HTC-10Ga-20K) have nearly the same specific surface areas,
being the lowest observed. Probably this is related to pore blocking caused by
the K2CO3 phase, which is in agreement with SEM images (cf. Fig. 3.3c and Fig.
3.3f).The modification with strontium resulted in the material with the highest
specific surface area (SBET = 237 m2g-1). Although a straightforward explanation
is lacking for this behaviour, it might be tentatively related to subtle changes in
the basicity of the media upon addition of the doping cations.
Table 3.2 – BET specific surface area of the prepared materials determined by physical adsorption of nitrogen at 77 K.
Sample SBET [ m2∙g-1]
HTC (fresh) 122
HTC (calcined) 183
HTC-20K (calcined) 61
HTC-10Ga (fresh) 156
HTC-10Ga (calcined) 175
HTC-10Ga-20K (calcined) 58
HTC-10Ga-20Cs (calcined) 165
HTC-10Ga-20Sr (calcined) 237
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
95
3.3.1.5 FTIR
FTIR analyses obtained for the as prepared (uncalcined) HTC and HTC-
10Ga samples indicate the presence of carbonate as interlayer anion (cf. Fig.
3.4).
Fig. 3.4 - FTIR spectra of the starting and modified samples.
The spectra show an intense, broad band centered at ca. 3460 cm-1, due
to the stretching mode of the hydroxyl groups, both from the brucite-like layers
and the interlayer water molecules, as well as the water molecules physisorbed
on the external surface of the crystallites. The shoulder at 2900-3000 cm-1 has
been ascribed [27] to the stretching mode of interlayer hydroxyl groups
hydrogen-bonded to interlayer carbonate species. The bending mode of the
water molecules is responsible for the rather weak band at ca. 1630 cm-1, also
Chapter 3
96
recorded for both samples. The ν3 mode of interlayer carbonate is responsible
for the intense band at 1384 cm-1. This is a degenerated mode for the D3h
symmetry of the original carbonate anion; under this symmetry, the ν1 mode
expected at ca. 1080 cm-1 is inactive and it is not recorded. However, it can be
noticed that the band at 1384 cm-1 shows a clearly identified shoulder at 1500
cm-1, which can be considered because of the splitting of the ν3 mode. However,
as the band due to the ν1 mode is not recorded, the decrease in the symmetry
should be discarded and the two bands at 1384 and 1500 cm-1 should be
ascribed to carbonate species in two sites with different interaction with their
local environment. Weaker bands below 1000 cm-1 correspond mainly to
vibration of lattice bonds (Mg-OH, Al-OH, etc.).
When the samples were calcined at 673 K the spectra showed some
differences, but unfortunately it seems that some interlayer carbonate species
are still present (it should be noticed that at this temperature not all carbonate
has been removed, cf. the TG curve – Fig. 3.1); the hydroxyl groups have not
been completely removed either. Consequently, the spectra show the band at
ca. 3450 cm-1 due to the stretching mode of the hydroxyl groups (although the
band is much narrower than for the uncalcined samples), and the hydrogen
bonding between the hydroxyl groups and the carbonate species seem to be
absent. The remaining carbonate species show a lower symmetry and the ν3
mode has lost its original degeneration, probably on changing the carbonate
anion from a D3h to a C2v symmetry, the corresponding band is now split into two
bands at ca. 1480 and 1410 cm-1 and the band due to mode ν1 is now recorded
at 1110 cm-1.
Impregnation with Sr, K, or Cs and calcination at 673 K leads to some
changes in the spectra, concerning almost exclusively to the region below 1600
cm-1. The band due to mode ν3 of carbonate species splits in all cases and the
band due to mode ν1 of carbonate is recorded in all cases as well, indicating that
in this case a decrease in the local symmetry of the carbonate anions actually
has taken place. However, both components of the split ν3 band show a change
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
97
in their relative intensities from the monovalent (K and Cs) to the divalent (Sr)
doping cation; this might be tentatively related to the different nature (mono- or
divalent) of the doping cations. Probably, the increased basicity of the solids after
impregnation with the alkaline or alkaline-earth cations delays decomposition of
the solid and evolution of CO2, and thus remaining of carbonate species above
the temperature where decomposition, in the absence of the doping cations,
would be expected. Moreover, if carbonate species still exist after calcination at
673 K, this would account for the shoulder around 2900 cm-1 due to the νOH
stretching mode of hydroxyl groups hydrogen-bonded to the carbonate species.
The absence of bands around 1700-1750 cm-1 indicates that although
acetate salts were used for modifying the hydrotalcites with K, Cs and Sr, no ion
exchange between acetate and interlayer carbonate ions, within the hydrotalcite
structure, took place.
3.3.2 CO2 sorption experiments
3.3.2.1 Determination of adsorption equilibrium isotherms at 573 K
– sorbent screening
Fig. 3.5 shows CO2 sorption isotherms obtained at 573 K under dry
conditions for the prepared materials. The results are very promising, particularly
for the sample substituted with gallium and modified with potassium (HTC-10Ga-
20K); a sorption equilibrium value of 1.82 mmol·g-1 (i.e., 2.01 mmol·cm-3) at 1.08
bar was obtained which is, up to the authors knowledge, clearly above typical
results found in the literature under similar operation conditions (see Table 3.3)
[9, 22, 28, 29].
Chapter 3
98
Table 3.3 – CO2 sorption equilibrium value on HTC-10Ga-20K and comparison with other sorbents reported in the literature.
Material T [K] pCO2 [bar] q [mol·kg-1] Ref.
HTC-10Ga-20K 573 1.08 1.82 This work
Hydrotalcite 573 1 0.52 [29]
Hydrotalcite 573 1 0.50 [9]
Hydrotalcite 573 1.1 0.25 [22]
K-promoted hydrotalcite 673 1 0.79 [28]
Generally, the modification with alkali metals increases the sorption
capacity of the original hydrotalcite samples (HTC and HTC-10Ga). However,
adding potassium provided higher sorption values when compared to cesium
and this should be related to its higher basicity. On the contrary, the sorption
capacity decreased for the sample modified with an alkaline-earth metal (Sr).
Moreover, the sorption capacity of the materials decreased in general with the
increase of their specific BET surface area (cf. Table 3.2 and Fig. 3.5).
Fig. 3.5 – CO2 sorption equilibrium isotherms at 573 K for the prepared materials. Lines shows fitting using the Freundlich equation.
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
99
This means that the available BET surface area does not play a crucial
role on the sorption capacity, which was also reported in other works [7, 30]; the
sorption capacity is more likely to be related to the chemical nature of the
exposed surface. In the case of HTC-10Ga-20Sr it seems that the phase
obtained (strontianite), which has a completely different morphology (see Fig.
3.3e), almost does not sorb CO2, at least under the tested operation conditions.
The results herein presented are in line with the ones obtained by Yavuz
et al. [8], where they showed that the simultaneous modification with potassium
and gallium markedly improved the sorption capacity of hydrotalcites, the first
acting as a chemical promoter and the latter providing a superstructure robust
and stable at elevated temperatures (a sort of synergic effect). However, in our
work the sorption isotherm was obtained at 573 K and up to 1.1 bar. Moreover,
this promising sample (HTC-10Ga-20K) was also submitted to sorption-
desorption cycles at 573 K and 473 K to determine its reversible working
capacity, checking its suitability to be used in cyclic operation for post-
combustion CO2 capture/valorisation applications (Section 3.3.2.2).
The sorption data of each sample was fitted to the Freundlich equation
(Eq. (3.2)), which has also been proposed elsewhere to describe CO2 sorption
on hydrotalcites [21, 31]:
1
nq K p (3.2)
where q (mmol·cm-3) is the concentration of the sorbed species, and K
(mmol·cm-3·bar-(1/n)) and n (dimensionless) are Freundlich equation parameters.
Parameter n is normally greater than unity and as it increases the isotherm
approaches to the so-called irreversible isotherm. In this case, the pressure
needs to be very low in order to the sorbate start desorbing from the surface
[32].
Chapter 3
100
The parameters of the Freundlich equation (along with a 90 % interval of
confidence) are included in Table 3.4. The experimental data of the starting and
modified samples showed good adhesion to the Freundlich model within the
tested pressure range (see also Fig. 3.5).
Table 3.4 – Freundlich type isotherm parameters and their confidence limits (90 %) obtained after fitting the experimental data of Fig. 3.5 (T = 573 K) and solid density (determined with He at 313 K after regeneration) for each material.
Sample log(K) 1/n R2 ρ [g·cm-3]
HTC - - - 0.831
HTC-20K 0.26 ± 0.034 0.16 ± 0.038 0.9875 0.841
HTC-10Ga -0.30 ± 0.055 0.20 ± 0.074 0.9685 0.860
HTC-10Ga-20K 0.29 ± 0.035 0.10 ± 0.039 0.9685 1.107
HTC-10Ga-20Cs 0.14 ± 0.046 0.17 ± 0.054 0.9774 1.187
HTC-10Ga-20Sr -0.48 ± 0.11 0.25 ± 0.15 0.9251 0.762
3.3.2.2 Sorption-desorption experiments
To determine the working capacity of the best sorbent (HTC-10Ga-20K),
a series of sorption-desorption cycles at 573 K and 473 K was carried out. As it
can be seen in Fig. 3.6, in what we called cycle 0, CO2 was fed to the
microbalance chamber and the pressure allowed to increase up to 0.05 bar and
then to 0.15 bar. Each cycle ended with the sample being submitted to vacuum.
In this cycle, the sorption capacity determined at 573 K and 0.15 bar (Fig. 3.6a)
was 1.57 mmol·cm-3. However, when the microbalance chamber was evacuated
it was not possible to remove all the sorbed CO2; around 0.38 mmol·cm-3
remained sorbed – see open symbol. In the following cycles the pressure was
again switched between 0.15 bar and vacuum, except in cycle 3 where the
pressure was increased up to 1.1 bar before the regeneration stage. It is
important to note that the sorption values presented in each cycle (black
symbols) correspond to the working capacity, i.e., it was considered a starting
mass of the sample which included the previous irreversibly sorbed mass of CO2.
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
101
For instance, in cycle 1 at 573 K, the sample was able to effectively sorb only
1.13 mmol·cm-3, which is smaller than the value in cycle 0 because some carbon
dioxide remained irreversibly sorbed.
Fig. 3.6 – Working capacity of HTC-10Ga-20K towards CO2 (see guiding solid line) obtained after consecutive sorption-desorption cycles at: a) 573 and b) 473 K.
By adopting this approach, it is thus possible to determine the working
capacity in each cycle. Fig. 3.6 shows that just after cycle 2 a steady working
Chapter 3
102
capacity is mostly attained; at 573 K this working capacity is ca. 0.97 mmol∙cm-3
whereas at 473 K it is ca. 0.25 mmol∙cm-3.
However, at the end of the first cycle, around 24 % and 77 % of the total
sorption capacity is irreversibly lost at 573 K and 473 K, respectively (cf. Fig. 3.7
and Fig. 3.8). In Fig. 3.7 the irreversible fraction at the end of each cycle is
determined as the difference between the black and white bars (for instance in
cycle 0 at 573 K such value is the one mentioned above, 0.38 mmol·cm-3).
Fig. 3.7 – History of sorbed (black bars) and desorbed (white bars) amount of CO2 for each cycle at a) 573 K and b) 473 K.
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
103
The irreversible part was found to be also pressure dependent as shown
in Fig. 3.8, where the accumulated irreversible fraction is shown; it is noteworthy
that when the pressure was increased up to ~1.1 bar (i.e. during cycle 3), the
number of moles of CO2 that were irreversible sorbed increased 51 % and 72 %
at 573 K and 473 K, respectively. However, in the subsequent cycles the working
capacity remained nearly constant at each temperature.
Fig. 3.8 – Irreversible CO2 moles sorbed per cycle considering the mass of solid obtained after the first regeneration step at 473 K and 573 K. Lines are for eye guidance.
The obtained results suggest that the CO2 sorption takes place through an
activated process; otherwise, it would be expected to obtain a higher sorption
capacity at 473 K. In this regard, several authors proposed different models to
describe the sorption equilibrium of carbon dioxide over hydrotalcites [7, 28, 33],
although in the present work a good fit was obtained using a simple Freundlich
equation (Section 3.3.2.1).
The changes observed in the CO2 sorption capacity during the
implemented cyclic process in the absence of water vapour can be, in principle,
related to the structural changes observed by XRD, in which the hydrotalcite-like
structure could not be recovered. Further work should thus aim assessing the
behaviour of these materials in the presence of water vapour, which is found in
Chapter 3
104
both typical post- and pre-combustion streams. In fact, other authors found that
in presence of water the working capacity loss practically does not occur (< 10
%) as compared to that exhibited by the same material under dry conditions [12].
Regarding the sorption mechanism, Ritter and co-workers [34] observed,
from in situ FTIR spectroscopy, the carbonate transformations occurring in K-
promoted hydrotalcites during CO2 adsorption and desorption at 723 K. In their
study, it was observed that one irreversible and three reversible processes took
place simultaneously, that could somehow be represented by two distinct kinetic
contributions. The slow and irreversible process was related to the formation of
polydentate carbonate, whereas the fast and reversible uptakes were related to
the formation of bridged (surface), uni- and bidentate carbonate species. Ritter
and co-workers concluded that the formation of bidentate carbonate species is
favoured at very short times. Unidentate and bridged carbonate species start to
disappear after one hour, eventually due to their transformation into polydentate
carbonate. In order to preliminary assess the sorption kinetics of sample HTC-
10Ga-20K, based on the above mentioned findings, the following expression is
herein proposed to tentatively describe the total amount of substance entering
or leaving the sorbent particle [35]:
2 2 2 21 2
2 2
1 2
1 2
2 2 2 21 1
0
0
6 1 6 11 1
D Di t i t
r r
i i
q t q tF F F
q q t
q qe e
q qi i
(3.3)
Eq. (3.3) assumes constant surface concentration in the sorbent particle
and two kinetic contributions occurring in parallel, one responsible for the fast
uptake at the beginning and, the other one, by the slow increase observed for
longer times. The model parameters 1
2
D
r(s-1), 2
2
D
r(s-1) and 1q (mmol·cm-3) were
obtained through non-linear fitting of Eq. (3.3) to experimental data. One should
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
105
note that q (mmol·cm-3) is known, since it is equal to the difference between
the sorbed (or desorbed) quantity at equilibrium and the value for t = 0.
Additionally, the quantity related to the slow uptake, 2q (mmol·cm-3), is equal to
the difference between q (known a priori) and 1q (fitted parameter).
Fig. 3.9 shows the experimental uptake (F), downtake (1-F) and model
curves based on Eq. (3.3), where F is the sum of curves F1 and F2 (please see
Fig. A.0.4 in Appendix A for the total and partial loading uptake curves).
Fig. 3.9 – Uptake (and downtake) curves for CO2 sorption at 573 K (first row) and 473 K (second row) for sample HTC-10Ga-20K.
t [s]
0 1000 2000 3000 4000 5000 6000
1-F
0.0
0.2
0.4
0.6
0.8
1.0
1-Fexp
1-F1
1-F2
1-Fmod
Cycle 4T = 573 Kp < 0.001 bar
t [s]
0 1000 2000 3000 4000
F
0.0
0.2
0.4
0.6
0.8
1.0
Fexp
F1
F2
Fmod
Cycle 5T = 573 Kp = 0.05 bar
t [s]
0 2000 4000 6000 8000 10000
1-F
0.0
0.2
0.4
0.6
0.8
1.0
1-Fexp
1-F1
1-F2
1-Fmod
Cycle 4T = 473 Kp < 0.001 bar
t [s]
0 500 1000 1500 2000 2500 3000
F
0.0
0.2
0.4
0.6
0.8
1.0
Fexp
F1
F2
Fmod
Cycle 5
T = 473 K
p = 0.05 bar
t [s]
0 1000 2000 3000 4000
F
0.0
0.2
0.4
0.6
0.8
1.0
Fexp
F1
F2
Fmod
Cycle 0T = 573 Kp = 0.05 bar
t [s]
0 1000 2000 3000 4000 5000
F
0.0
0.2
0.4
0.6
0.8
1.0
Fexp
F1
F2
Fmod
Cycle 0T = 473 Kp = 0.05 bar
Chapter 3
106
The model predicts quite well the experimental data both during sorption
and desorption at 573 K and 473 K; the obtained parameters are shown in Table
3.5. Generally, the inverse of the time constant 1
2
D
r during CO2 sorption at 473 K
is two-fold higher than at 573 K.
Table 3.5 – Kinetic parameters of Eq. (3.3) obtained after fitting CO2 uptakes and downtakes at 573 K and 473 K on sample HTC-10Ga-20K.
Parameter
Cycle 0 Cycle 4 Cycle 5
p = 0.05 bar (uptake)
< 0.001 bar (downtake)
p = 0.05 bar (uptake)
473 K 573 K 473 K 573 K 473 K 573 K
4 11
210
Ds
r
15.55 7.57 117 1.00 9.08 4.88
6 12
210
Ds
r
7.41 12.26 0.55 2.54 6.01 9.91
100q
q
1 48 44 63 11 61 40
-3
1 mmol cmq 0.422 0.603 0.141 0.100 0.109 0.355
-3
2 mmol cmq
0.467 0.754 0.084 0.916 0.069 0.524
-3 mmol cmq
0.889 1.358 0.225 1.016 0.178 0.879
When compared the uptakes performed at the beginning of the experiment
on the reactivated sample (cycle 0) and at the end (cycle 5), it is observed that
the inverse of time constants 1
2
D
r and 2
2
D
r decreased for both temperatures.
The sorbed fraction related to the first contribution 1 q q increased
between cycles (13 percentage points) at 473 K, while at 573 K practically
remained constant. Desorption is much faster at 473 K where ca. 63 % of the
total desorbed quantity is related to the fast contribution whereas at 573 K
represents only 11 %.
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
107
3.4 Conclusions
Hydrotalcite-based materials were prepared and their sorption capacity
towards CO2 at high temperature was determined. The working capacities
obtained under low CO2 pressures and high temperatures are lower than those
typically observed in absorption units using amines operating at low
temperatures (i.e. 1.36 mmol·g-1), which however require that the flue gas
stream should be cooled (and later heated up to for the solvent regeneration
stage). Nevertheless, at 573 K the hydrotalcites capacities are clearly above the
minimum value (i.e. 0.3 mmol·g-1) considered for their application in pre-
combustion capture applications, particularly in sorption-enhanced reaction
processes. The sorption capacity was enhanced by partially substituting
aluminium with gallium and, particularly, after modifying with alkali metals (K and
Cs). The sorption equilibrium isotherms for all the materials were adequately
fitted using a simple Freundlich-type equation.
The best material, i.e. hydrotalcite substituted with gallium and promoted
with potassium (HTC-10Ga-20K), showed to be suitable to be used in cyclic
operation with regeneration under low CO2 pressure, particularly at 573 K, where
a very good working capacity (0.97 mmol·cm-3) was obtained. At 473 K, more
than 75 % of the sorption capacity was irreversibly lost and a low working
capacity is thus obtained (0.25 mmol·cm-3). However, as mentioned before, it is
known that in presence of water vapour such working capacity should be almost
not affected.
Sorption kinetics of the HTC-10Ga-20K material was preliminary assessed
and a model proposed considering two resistances in parallel. The model
showed good adherence to experimental data both during sorption and
desorption experiments; the inverse of the time constant related to the fast
uptake at 473 K is generally two-fold higher than at 573 K.
Chapter 3
108
3.5 References
[1] IPCC: Summary for Policymakers, in: T.F. Stocker, D. Qin, G.-K. Plattner, M.
Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, P.M. Midgley (Eds.)
Climate Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change, Cambridge University Press, Cambridge, United Kingdom and
New York, NY, USA, 2013.
[2] J. Wilcox, Carbon capture, Springer, New York, 2012.
[3] E.A. Quadrelli, G. Centi, J.L. Duplan, S. Perathoner, Carbon dioxide
recycling: Emerging large-scale technologies with industrial potential,
ChemSusChem, 4 (2011) 1194-1215.
[4] S. Choi, J.H. Drese, C.W. Jones, Adsorbent materials for carbon dioxide
capture from large anthropogenic point sources, ChemSusChem, 2 (2009) 796-
854.
[5] Q. Wang, J. Luo, Z. Zhong, A. Borgna, CO2 capture by solid adsorbents and
their applications: current status and new trends, Energy and Environmental
Science, 4 (2011) 42-55.
[6] A. de Roy, C. Forano, J.P. Besse, Layered double hydroxides: synthesis and
post-synthesis modification, in: V. Rives (Ed.) Layered double hydroxides:
present and future, Nova Science Publishers, Inc., New York, 2001, pp. 1-39.
[7] E.L.G. Oliveira, C.A. Grande, A.E. Rodrigues, CO2 sorption on hydrotalcite
and alkali-modified (K and Cs) hydrotalcites at high temperatures, Separation
and Purification Technology, 62 (2008) 137-147.
[8] C.T. Yavuz, B.D. Shinall, A.V. Iretskii, M.G. White, T. Golden, M. Atilhan,
P.C. Ford, G.D. Stucky, Markedly improved CO2 capture efficiency and stability
of gallium substituted hydrotalcites at elevated temperatures, Chemistry of
Materials, 21 (2009) 3473-3475.
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
109
[9] Z. Yong, V. Mata, A.E. Rodrigues, Adsorption of carbon dioxide onto
hydrotalcite-like compounds (HTlcs) at high temperatures, Industrial and
Engineering Chemistry Research, 40 (2001) 204-209.
[10] M.K. Ram Reddy, Z.P. Xu, G.Q. Lu, J.C. Diniz Da Costa, Influence of water
on high-temperature CO2 capture using layered double hydroxide derivatives,
Industrial and Engineering Chemistry Research, 47 (2008) 2630-2635.
[11] Y. Ding, E. Alpay, Equilibria and kinetics of CO2 adsorption on hydrotalcite
adsorbent, Chemical Engineering Science, 55 (2000) 3461-3474.
[12] J.R. Hufton, S. Mayorga, S. Sircar, Sorption-enhanced reaction process for
hydrogen production, AIChE Journal, 45 (1999) 248-256.
[13] H.T.J. Reijers, S.E.A. Valster-Schiermeier, P.D. Cobden, R.W. Van Den
Brink, Hydrotalcite as CO2 sorbent for sorption-enhanced steam reforming of
methane, Industrial and Engineering Chemistry Research, 45 (2006) 2522-2530.
[14] Y.J. Wu, P. Li, J.G. Yu, A.F. Cunha, A.E. Rodrigues, K-Promoted
hydrotalcites for CO2 capture in sorption enhanced reactions, Chemical
Engineering and Technology, 36 (2013) 567-574.
[15] M.K.R. Reddy, Z.P. Xu, G.Q. Lu, J.C. Diniz da Costa, Effect of SOx
adsorption on layered double hydroxides for CO2 capture, Industrial &
Engineering Chemistry Research, 47 (2008) 7357-7360.
[16] M.G. Beaver, High temperature CO2 chemisorbents: applications,
characterization, and study of the chemical nature of chemisorbent surfaces, in,
Lehigh University, 2010.
[17] E. Van Dijk, S. Walspurger, P. Cobden, R. Van Den Brink, Testing of
hydrotalcite based sorbents for CO2 and H2S capture for use in sorption
enhanced water gas shift, International Journal of Greenhouse Gas Control, 5
(2011) 505-511.
[18] J.R. Hufton, S. Mayorga, T. Gaffney, S. Nataraj, M. Rao, S. Sircar, Sorption
enhanced reaction process for production of hydrogen, Proceedings of the 1998
U.S. DOE Hydrogen Program Review., 1998.
Chapter 3
110
[19] V. Rives, Comment on "direct observation of a metastable solid phase of
Mg/Al/CO3-layered double hydroxide by means of high-temperature in situ
powder XRD and DTA/TG", Inorganic Chemistry, 38 (1999) 406-407.
[20] V. Rives, Characterisation of layered double hydroxides and their
decomposition products, Materials Chemistry and Physics, 75 (2002) 19-25.
[21] M.H. Halabi, M.H.J.M. De Croon, J. Van Der Schaaf, P.D. Cobden, J.C.
Schouten, High capacity potassium-promoted hydrotalcite for CO2 capture in H2
production, International Journal of Hydrogen Energy, 37 (2012) 4516-4525.
[22] M.K. Ram Reddy, Z.P. Xu, G.Q. Lu, J.C.D. Da Costa, Layered double
hydroxides for CO2 capture: Structure evolution and regeneration, Industrial and
Engineering Chemistry Research, 45 (2006) 7504-7509.
[23] J. Rocha, M. Del Arco, V. Rives, M.A. Ulibarri, Reconstruction of layered
double hydroxides from calcined precursors: A powder XRD and 27Al MAS NMR
study, Journal of Materials Chemistry, 9 (1999) 2499-2503.
[24] K. Chibwe, W. Jones, Intercalation of organic and inorganic anions into
layered double hydroxides, Journal of the Chemical Society, Chemical
Communications, (1989) 926-927.
[25] T. Kwon, T.J. Pinnavaia, Pillaring of a layered double hydroxide by
polyoxometalates with Keggin-ion structures, Chemistry of Materials, 1 (1989)
381-383.
[26] M. Bellotto, B. Rebours, O. Clause, J. Lynch, D. Bazin, E. Elkaïm,
Hydrotalcite decomposition mechanism: a clue to the structure and reactivity of
spinel-like mixed oxides, The Journal of Physical Chemistry, 100 (1996) 8535-
8542.
[27] E.C. Kruissink, L.L. van Reijen, J.R.H. Ross, Coprecipitated nickel-alumina
catalysts for methanation at high temperature. Part 1.-Chemical composition and
structure of the precipitates, Journal of the Chemical Society, Faraday
Transactions 1: Physical Chemistry in Condensed Phases, 77 (1981) 649-663.
High temperature CO2 sorption with gallium-substituted and promoted hydrotalcites
111
[28] K.B. Lee, A. Verdooren, H.S. Caram, S. Sircar, Chemisorption of carbon
dioxide on potassium-carbonate-promoted hydrotalcite, Journal of Colloid and
Interface Science, 308 (2007) 30-39.
[29] Z. Yong, V. Mata, A.E. Rodrigues, Adsorption of carbon dioxide at high
temperature - A review, Separation and Purification Technology, 26 (2002) 195-
205.
[30] N.D. Hutson, B.C. Attwood, High temperature adsorption of CO2 on various
hydrotalcite-like compounds, Adsorption, 14 (2008) 781-789.
[31] J.L. Soares, G.L. Casarin, H.J. José, R.D.F.P.M. Moreira, A.E. Rodrigues,
Experimental and theoretical analysis for the CO2 adsorption on hydrotalcite,
Adsorption, 11 (2005) 237-241.
[32] D.D. Do, Adsorption analysis: equilibria and kinetics, Imperial College
Press, London, 1998.
[33] H. Du, A.D. Ebner, J.A. Ritter, Pressure dependence of the nonequilibrium
kinetic model that describes the adsorption and desorption behavior of CO2 in
K-promoted hydrotalcite like compound, Industrial and Engineering Chemistry
Research, 50 (2011) 412-418.
[34] H. Du, C.T. Williams, A.D. Ebner, J.A. Ritter, In situ FTIR spectroscopic
analysis of carbonate transformations during adsorption and desorption of CO2
in K-promoted HTlc, Chemistry of Materials, 22 (2010) 3519-3526.
[35] J. Crank, The mathematics of diffusion, Second ed., Oxford University
Press, London, 1975.
113
Chapter 4. Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
The contents of this chapter were adapted from: C.V. Miguel, A. Mendes, L.M. Madeira,
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst, Journal of
CO2 Utilization, 2018, in press, http://doi.org/10.1016/j.jcou.2018.03.011.
The intrinsic kinetics of CO2 methanation over an industrial nickel-
based catalyst was determined for a temperature range between 250 ºC
to 350 ºC. The kinetic experiments were performed operating a fixed-bed
reactor far from equilibrium conditions, in the absence of heat and mass
resistances and at the atmospheric pressure. Five mechanistic-based
models for describing the reaction kinetics were taken from literature and
used for fitting the experimental reaction rates. Model discrimination was
based on the assessment of the thermodynamic consistency and
statistical significance of inherent parameters (for 95 % confidence
level). Comparison of the adequacy of fit between accepted models was
done through the determination of the corresponding F-values to select
the best model. The selected model assumes a formyl intermediate
mechanism with a hydroxyl group being the most abundant species. The
proposed reaction kinetics was further validated by the simulation of an
isothermal plug-flow reactor operating at the same experimental
conditions employed in this work, where a reasonable agreement
between model predictions and the observed values was obtained.
Chapter 4
114
4.1 Introduction
The main application for the methanation reaction (also known as the
Sabatier reaction) has been since long ago the removal of carbon oxides traces
from hydrogen-rich feed streams in ammonia plants [1, 2] (Eq. (4.1)) and
Eq.(4.2)). It has also been proposed to produce synthetic natural gas (SNG),
particularly in the 60’s decade, because of increased natural gas demand. At
that time, SNG production from coal was envisioned as an alternative pathway
to assure security of supply in any event of natural gas shortage and research
on the topic was strongly financed, particularly by the USA [3].
298 K -12 2 4 2 rCO + 4H CH + 2H O Δ = - 165 kJ molH (4.1)
298 K -12 4 2CO + 3H CH + H O Δ = - 206 kJ molrH (4.2)
Nowadays, CO2 methanation reaction has gained a renewed interest in
the scope of Power-to-Gas applications (PtG), a concept where surplus
renewable electricity is transformed into hydrogen (via H2O electrolysis) and
afterwards into methane. This last step makes the process more flexible since
methane can be more easily stored and transported than hydrogen, enabling the
integration and balance of the power grid with the gas grid. Methane wide range
of end-use possibilities (e.g. vehicle fuel, for heat and power production,
intermediate to obtain other chemicals) contributes to process flexibility and
versatility [4].
Carbon dioxide methanation reaction (Eq. (4.1)) is thermodynamically
favoured at low temperatures and high pressures [5]. The equilibrium constant
(Keq) dependence with the absolute temperature (T) can be retrieved from the
equation provided by Lunde and Kester [6]:
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
115
21.056000 +34633 -16.4ln +0.00557 +33.165
1.987eq=
T T T T
K e
(4.3)
The heat released by the reaction is the major difficulty to handle at
industrial scale [3]; the adiabatic rise in temperature per each percent of CO2
converted is 60 ºC [7]. Hence, methanation usually takes place in a series of
adiabatic fixed-bed reactors with inter-bed gas recycling cooling or in fluidized-
bed reactors [3, 8]. Structured catalysts such as metal coated foams have also
been envisaged for this reaction due to their improved heat transfer capacity
(e.g. [9, 10]).
Ni-based catalysts supported on various solids (e.g. Al2O3, SiO2, CeO2,
etc.) are the most studied and commercialized catalysts at high temperatures
(i.e. > 250 ºC), a range where the formation of dangerous nickel carbonyl
(Ni(CO)4) is avoided, while more expensive Ru-based catalysts are best options
at low temperatures (< 200 °C) [7, 11]. Generally, catalysts which are effective
for CO methanation (Eq. (4.2)) are also effective for CO2 methanation, at least
for streams having low COx concentrations, as found in hydrogen purification
processes [11]. Table 4.1 lists main manufacturers and characteristics of some
industrial methanation catalysts.
The catalysts are available in several shapes and some are supplied in a
pre-reduced form (e.g. Katalco 11-4R or PK-7R), which simplifies process start-
up. Industrial methanation catalysts lifetime ranges from 5 to 10 years, although
some manufacturers report a period up to 24 years [1, 7]. Common poisons are
sulphur and arsenic compounds, particularly for nickel catalysts [2]. Industrial
catalysts for hydrogenation of carbon oxides has been the subject of a detailed
review by Golosman and Efremov [7], covering the available commercial
catalysts and their properties, useful information about their preparation for use,
handling instructions and safety precautions, besides operation issues of some
industrial processes where methanation intervenes.
Chapter 4
116
Table 4.1 - List of some industrial methanation catalysts and related characteristics [7, 12-15].
Commercial reference
Metal Metal cont.
(wt. %) P
(bar) T
(ºC) Shape
Size (mm)
METH 134 a) Ni 20-25 n.a. >200 Ball 3.0 - 6.0
METH 150 a) Ru n.a. n.a. <170 Ball 4.5 - 4.5
PK-7R b) Ni >23 n.a. 190-450 Ring 5.0 x 2.5
Katalco 11-4 c) Ni n.a n.a. n.a Pellet 5.4 x 3.6
Katalco 11-4R c) Ni n.a n.a. >220 Pellet 5.4 x 3.6
Katalco 11-4M c) Ni n.a n.a. n.a Pellet 3.1 x 3.6
Katalco 11-4MR c) Ni n.a n.a. >220 Pellet 3.1 x 3.6
NIAP-07-01 d) Ni 33-39 20-300 180-450 Pellet 5.5 x 4.5
NIAP-07-02 d) Ni 32-38 20-300 180-450 Pellet 5.5 x 5.0
NIAP-07-04 d) Ni 25 n.a n.a Pellet 5.5 x 4.5
TO-2M d) Ni 35-41 1-300 180-450 Pellet 5.5 x 5.0
RKM-3 d) Ru 0.3 300 150 Ball 2.6 - 3.0
Manufacturer
a) Clariant b) Haldor-Topsoe c) Johnson Matthey d) NIAP Katalizator
This work determines the intrinsic reaction kinetics over an industrial
nickel-based catalyst. Knowing the reaction kinetics is fundamental for
modelling, simulation and optimization of conventional or new reactor concepts
(e.g. sorptive reactors (e.g. [16]). To this end, mechanistic-based rate equations
available in the literature were considered and are presented in the following
section.
4.2 Mechanisms for CO2 methanation
4.2.1 “Carbon intermediate” mechanism
Dalmon and Martin [17] proposed a mechanism for CO and CO2
methanation over a Ni/SiO2 catalyst by studying the hydrogenation of
intermediate species. The authors postulated that in both reactions adsorbed
carbon monoxide (CO*) is a common intermediate (apart from O* in the latter
case). CO* then dissociates into C* and O* while CH4 is produced following C*
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
117
hydrogenation. Since CO* can be formed at lower temperatures in the case of
CO2 adsorption, a lower activation energy could explain the higher
hydrogenation rate of CO2. The higher selectivity of CO2 methanation was
tentatively explained by the lower C* concentration (because of lower CO2
adsorption capacity) and to the higher abundance of O*, which may act as a
geometric diluent decreasing the C-C formation.
Weatherbee and Bartholomew [18] also used a Ni/SiO2 catalyst and the
mechanism proposed by Dalmon and Martin adequately explained their results.
The elementary steps of the mechanism are presented in Table 4.2.
Table 4.2 – Carbon intermediate mechanism (from [18]).
Elementary step Step
2H + 2* 2H* 1
2CO + 2* CO* + O* 2
CO* CO + * 3
CO* + * C* + O* (r.d.s.) 4
C* + 4H* 4CH * + 4* 5
4CH * 4CH + * 6
O* + H* OH* + * 7
OH* + H* 2H O* + * 8
2H O* 2H O + * 9
The kinetic rate equation derived, assuming CO* dissociation to C* and O*
(step 4) as the rate determining step (r.d.s.), is given by Eq. (4.4).
2 2
4
2
2 2
2
0.5 0.5
CH 20.5
0.5 0.51 2 30.5
1
CO H
CO
CO H CO
H
kp pr
pK K p p K p
p
(4.4)
Chapter 4
118
where k is the rate constant, pi is the partial pressure of species i (CO2, H2, CO)
and K1, K2 and K3 are the constants that result from the combination of several
adsorption equilibrium and/or reaction rate constants.
4.2.2 “Formate intermediate” mechanism
Fujita et al. [19] studied the mechanism of CO and CO2 methanation over
an Ni/Al2O3 catalyst at 180 ºC. For CO methanation it was observed that
adsorbed carbon, linear and bridged CO were the predominant species,
whereas for CO2 methanation the presence of two types of adsorbed CO in
bridged structures and a formate species (HCOO) were detected. The presence
of linear CO species markedly retarded the hydrogenation of adsorbed carbon.
Since the latter species has not been observed during CO2 methanation, this
could explain the difference between the rate and selectivity of CO and CO2
methanation reactions [19, 20].
Pan et al. [21] recorded the in-situ FTIR spectra of CO2 methanation using
a Ni/Al2O3 catalyst at different temperatures (75-375 ºC). The presence of
bidentate formate intermediate was observed for temperatures above 225 ºC
and its hydrogenation was considered the main reaction step. Both Fujita et al.
[19] and Pan et al. [21] concluded that formate species was adsorbed at the
alumina support. A mechanism for CO2 hydrogenation over a commercial nickel-
alumina-calcium catalyst (ref.: NKM-4A from NIAP-Katalizator) involving a
formate intermediate was given by Ibraeva et al. [22] (cf. Table 4.3).
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
119
Table 4.3 – Formate intermediate mechanism (from [22]).
Elementary step Step
2H + 2* 2H* 1
2CO + * 2CO * 2
2CO * + H* HCOO* + * (r.d.s.) 3
HCOO* + H* HCO* + OH* 4
HCO* + H* CH* + OH* 5
CH* + 3H* 4CH * + 4* 6
OH* + H* 2H O + 2* 7
4CH * 4CH + * 8
In this mechanism, the formate species is obtained through the reaction of
molecularly adsorbed CO2 and atomic hydrogen (step 3), being this the rate
determining step of the reaction. The derived kinetic rate equation is expressed
by Eq. (4.5).
2 2
2 2
0.5
0.5
H CO
H CO
p pr k
p K p
4CH
(4.5)
where K is the adsorption equilibrium constant.
4.2.3 “Formyl intermediate” mechanism
A mechanism comprising a formyl species intermediate has been
proposed recently by Koschany et al. [23] for a coprecipitated NiAl(O)x catalyst
(cf. Table 4.4).
Chapter 4
120
Table 4.4 – Formyl intermediate mechanism (from [23]).
Elementary step Step
2H + 2* 2H* 1
2CO + 2* CO* + O* 2
CO* + H* CHO* + * (r.d.s.) 3
CHO* + * CH* + O* 4
CH* + 3H* 4CH * + 3* 5
4CH * 4CH + * 6
O* + H* OH* + * 7
OH* + H* 2H O* + * 8
2H O* 2H O + * 9
In this mechanism, the formyl species results from the reaction of adsorbed
CO (obtained after CO2 dissociation) and atomic hydrogen. Eq. (4.6) was
derived by Koschany et al. [23] assuming adsorbed CO hydrogenation (step 3)
as the rate limiting step and CO*, H* and OH* as the most abundant surface
species.
2 2 4 2
4
2 22
2 2 2
2
0 5 0 5 2
CO H CH H
CH 2 4
CO H eqH O 0 5 0 5 0 5
OH H H mix CO0 5
H
1
1
. .
O
. . .
.
k p p p pr
p p KpK K p K p
p
(4.6)
where 2
2
H O
OH 0 5
H
.
pK
p
, 2 2
0 5 0 5
H H
. .K p and 2
0 5
mix CO
.K p stand for the surface coverage of
hydroxyl OH , atomic hydrogen H and carbon monoxide CO ,
respectively. eqK is the reaction equilibrium constant determined through Eq.
(4.3).
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
121
If water is considered as one of the most abundant surface intermediates
instead of hydroxyl species, than Eq. (4.6) is modified as follows [23].
2 2 4 2
4
2 22 2 2 2 2
0 5 0 5 2
CO H CH H
CH 2 40 5 0 5 0 5
CO HH O H O H H mix CO
1
1
. .
O
. . .eq
k p p p pr
p p KK p K p K p
(4.7)
and the term 2 2H O H OK p stands for the water surface coverage
2H O .
4.3 Experimental
4.3.1 Experimental setup
An illustration of the experimental setup used in the kinetic tests is shown
in Fig. 4.1.
Fig. 4.1 – Scheme of the experimental setup used in this work.
Experiments were performed in a stainless steel packed-bed unit with 12
cm length and 7.75 mm inside diameter, which was placed inside a tubular oven
(model Split from Termolab, Fornos Eléctricos, Lda.) equipped with a 3-zone PID
temperature controller (model MR13 from Shimaden). The 3 type-K
Chapter 4
122
thermocouples used to measure and control the oven temperature were placed
in contact with the reactor wall.
CO2 (99.998 %, Air Liquide), H2 (99.9995 %, Air Liquide) and N2
(99.9995% %, Air Liquide) were fed upwardly to the reactor using mass flow
controllers (model F201 from Bronkhorst-High Tech). The flow rate of the stream
leaving the reactor was measured with a mass flow meter (model F101 from
Bronkhorst-High Tech).
Type-K thermocouples allowed recording the temperature histories in two
axial positions of the reactor. The reactor was placed inside the oven so that the
temperature in the bed was practically the same along its length (maximum
difference of 1 ºC). The pressure drop was measured by means of two pressure
transducers (model PMP 4010 from Druck) placed before and after the reactor.
The water vapour produced during the course of the reaction was condensed in
a home-assembled Peltier cold-trap located after the reactor.
The steady-state composition of the (dry) stream leaving the reactor was
measured using a gas chromatograph (model 1000 from DANI) equipped with a
micro-TCD detector (VICI) and a capillary column (Supelco Carboxen 1010 Plot,
30 m x 0.32 mm i.d. from Sigma Aldrich). Hydrogen fraction was obtained
through the mass balance. The carbon balance error was lower than 10 %.
4.3.2 Kinetic experiments
Kinetic tests were performed using a commercial nickel catalyst (METH
134). The catalyst has a NiO content of ca. 20-25 wt. % supported on calcium
aluminate [7, 24]. Axial dispersion and gas-solid wall effects can be neglected
when L/dp and Dr/dp ratios are above 50 and 10, respectively [25, 26], which was
guaranteed in the present work. Pressure drop along the bed may be assumed
negligible (< 7.4 % of inlet pressure).
Temperature was varied between 250-350 ºC, a range where commercial
catalysts are designed to operate without activity loss, even with occasional
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
123
temperature excursions up to 650 ºC [2]. The activity of commercial nickel
catalysts does not increase significantly above 350 ºC [27]. The experiments
were performed at the atmospheric pressure. Blank tests using the reactor
packed with only glass beads showed the absence of CO2 conversion in the
considered temperature range.
The first set of kinetic tests was carried at 350 ºC to check the absence of
external mass and heat transport resistances by varying the feed flow-rate while
the value of the time coordinate CO2
in
catW F was kept constant (see sections
4.4.1.1 and 4.4.1.2). In the absence of external mass/heat transfer effects and
in the presence of isothermal operation at 350 ºC, the same premises are
guaranteed at lower temperatures and lower catalyst activities [25].
After assuring the absence of external resistances, the presence of
internal diffusion was assessed using catalyst particles with different sizes (see
section 4.4.1.2). Intrinsic kinetics can thus be determined if the experiments are
conducted in the absence of both external and internal resistances. In this
regard, catalyst pellets were crushed and sieved to different sizes, namely,
between 150-250 µm and 250-350 µm; Kiendl et. al [24] used a similar catalyst
and reported that crushing has no influence on the Ni loading. The catalyst
particles were then mixed with inert glass beads with the same particle size.
Dilution of the catalyst with inert glass beads enhances the heat transfer and
minimizes temperature gradients, which was confirmed through direct
measurement of the bed temperature. Due to the highly exothermic nature of the
Sabatier reaction (Eq. (4.1)), the CO2 content in the reactant mixture was low
and balanced in N2 to minimize the temperature increase. An H2/CO2 feed ratio
of 4 was chosen for all experiments; at low temperature and atmospheric
pressure, H2/CO2 ratios lower than 4 could lead to low CH4 selectivity and
undesired carbon formation, as found in Chapter 2.
The mixture of catalyst and inert particles was loaded into the fixed-bed
tubular unit and framed in both ends by means of two stainless steel discs (mesh:
10-15 μm). At the bottom of the reactor (i.e. at the inlet) a layer of inert particles
Chapter 4
124
allows to even out the flow of the reaction mixture. Catalyst activation consisted
in the following protocol:
1. Purging the entire system with N2 at room temperature;
2. Heating the packed-bed (1 °C∙min-1) from room temperature up to 320
°C at the ambient pressure under a 12.8 % H2/N2 stream and hold 1h at
that temperature;
3. Increasing the temperature under N2 flow to the desired value for the first
kinetic experiment (i.e. 350 ºC).
A 4-way valve allowed to: 1) keep the reactor under a static N2 atmosphere
while the reactant mixture stream could be prepared and the flow stabilized and
2) feed the reactant stream to the reactor. Shutdown consisted on flushing the
reactor with N2 to remove unreacted carbon oxides and afterwards cooling down
the reactor up to room temperature.
4.3.3 Computational methods
The parameters of the kinetic equations were estimated through nonlinear
regression and statistically tested using the fitnlm function available in the Matlab
software as detailed below (cf. section 4.4.2.1). Nonlinear regression algorithm
consisted in the Levenberg-Marquardt method (details available in [28]) for the
minimization of the sum of residuals squares. Statistical testing was performed
for 95 % level of significance. The selected kinetic model was validated based
on the simulation of the fixed-bed reactor. The differential equations and
respective boundary conditions were solved numerically with a fourth-order
Runge-Kutta algorithm also using the Matlab software.
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
125
4.4 Results and Discussion
4.4.1 Kinetic experiments
4.4.1.1 Isothermal regime and catalyst stability
Reactor layouts were varied to assess the system behaviour regarding the
planned operation conditions (cf. Fig. 4.2).
Fig. 4.2 – Reactor layouts employed to determine the best configuration for kinetic tests.
Isothermal operation must be guaranteed for the determination of the
intrinsic kinetics. Due to the exothermic nature of CO2 methanation, the bed
temperature rise depends on the number of moles of CO2 present in the feed
composition which are converted per unit of time. In preliminary experiments,
the initial bed temperature increased 8.7 ºC vs. 2.8 ºC by changing the CO2 feed
fraction from 12 vol. % to 4 vol. %, respectively, using a reactor with 200 mg of
catalyst (reactor R1 in Fig. 4.2). This effect was not possible to assess for lower
CO2 percentages due to mass flow controllers limitations. In turn, the available
heat transfer area per mass of catalyst was increased changing the dilution
factor (Ψ) and using a feed composition (vol. %) of 4:16:80 (CO2:H2:N2) in all the
subsequent experiments. Isothermal regime was considered when the variation
Chapter 4
126
of the temperature along the bed length was lower than 1.5 ºC, which was
attained for the reactor with configuration R3 in Fig. 4.2, under the considered
range of total feed flow rates (30-100 mln∙min-1). The catalyst mass used in the
subsequent experiments varied between 12-40 mg.
Catalyst stability was checked under reactive conditions for a total period
of approximately 83 hours. During this time the catalyst was submitted to
reactive conditions in the temperature range of 250-350 ºC. The catalyst was
always kept in a nitrogen stream at 350 ºC overnight. The observed difference
of CH4 yield measured at the beginning and the end of this procedure (operation
conditions: T = 350 ºC, P = Patm, CO2
in
catW F = 6.22 gcat·h·mol-1) was negligible (ca.
0.61 %). Therefore, in the kinetic tests it was considered that the catalyst activity
remained unchanged since they were conducted in less time than the reference
period.
4.4.1.2 Identifying the region of kinetic rate control
The external and intraparticle mass and heat transport transfer limitations
were assessed based on the criterion listed in Table 4.5.
Table 4.5 – Criteria for negligible transport limitations in steady state kinetic studies [29].
Transport process Criterion Eq.
External mass transport obs p
b
0 05g
rCa .
k a' C
(4.8)
Intraparticle mass transport (Wheeler-Weisz criterion)
2
obs p2
eff s
10 10
2
p rr L n
.D C
(4.9)
External heat transport g b a
b b
0 05r
e e
k H C ECa Ca .
hT RT
(4.10)
Intraparticle heat transport eff s2 2a
s s
0 05r
i i
eff
D H C E.
T RT
(4.11)
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
127
The expressions employed for determining the thermodynamic and
transport properties and parameters used throughout Eqs. (4.8)-(4.11) are given
in the Appendix B (see parameters definitions and units in the Nomenclature
section). The criterion listed in Table 4.5 were always verified considering the
extreme situations, i.e.:
1) Conditions of minimum and maximum reaction rate at 350 ºC and;
2) Using either reactor feed or outlet compositions for calculating the
parameters.
The inequalities of Eqs. ((4.8)-((4.11) were all verified and thus both external and
intraparticle heat and mass transport limitations could be neglected. In addition,
the effect of external resistances was also assessed experimentally at 350 ºC
by determining CO2 conversion and varying the feed flow rate, while keeping the
time coordinate CO2
in
catW F constant (cf. Fig. 4.3).
Fig. 4.3 – Effect of total volumetric feed flow rate on CO2 conversion for constant
2
incat COW F = 3.73 g·h·mol-1 at 350 ºC. Dashed line stands for the thermodynamic
equilibrium conversion at 350 ºC, 1 bar and feed composition.
Fig. 4.3 shows the absence of external resistances for a feed flow rate
range between 30-100 mln·min-1. Internal diffusion effects can also be neglected
Qin / ml
n min
-1
0 20 40 60 80 100 120
XC
O2
/ %
0
20
40
60
80
100
150-250 m
250-350 m
Chapter 4
128
since the conversion obtained with catalyst particles of a different size was
nearly the same (Fig. 4.3).
4.4.1.3 Calculation of observed reaction rates
Fig. 4.4 shows the observed methane yield 4 4 2
out inCH CH COY F F as a
function of the ratio between the weight of catalyst catW to the CO2 molar feed
flow rate 2
in
COF at temperatures ranging from 250 ºC to 350 ºC.
Fig. 4.4 – Methane yield as a function of the time coordinate 2
incat COW F at various
temperatures. The continuous line stands for the equilibrium value of 4CHY at 350 ºC, 1
bar and feed composition (i.e. the minimum 4CHY for the temperature range considered).
The observed reaction rates were obtained using the differential method
[26].
(Wcat
/ Fin
CO2
) / g h mol-1
0 2 4 6 8 10 12 14
YC
H4 ,
ob
s
0.0
0.2
0.4
0.6
0.8
1.0
350 ºC
325 ºC
300 ºC
275 ºC
250 ºC
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
129
4
4
2
CH
CH incat CO
d
d
Yr
W F (4.12)
To this end, the relationship between 4CHY and
CO2
in
catW F was described in terms
of polynomial functions for each temperature (see dashed lines in Fig. 4.4). The
derivative of each polynomial function allowed to extract the observed reaction
rates at a given temperature and CO2
in
catW F (cf. Eq. (4.12)).
4.4.2 Modelling work
4.4.2.1 Model discrimination and parameter estimation
A scaling method consisting in temperature centering was adopted for the
estimation of rate and adsorption parameters of the proposed kinetic equations,
according to Eq. (4.13) and Eq. (4.14), respectively.
m
1 1
0
aE
R T T'k k e
(4.13)
m
1 1
0
H
R T T'K K e
(4.14)
where mT stands for the mean temperature of the considered range (i.e. 573 K)
and 0
'k and 0
'K are the constants evaluated at T = Tm. Temperature centering
reduces the correlation between the pre-exponential factor and Ea or ΔH, making
the numerical least-squares procedure more robust and less sensitive to poor
starting estimates [28]. Afterwards, the “true” pre-exponential factors are
calculated from Eq. (4.15) and Eq. (4.16), by the Arrhenius and Van’t Hoff
equations, respectively.
Chapter 4
130
a
m
0 0
E
RT'k k e
(4.15)
m
0 0
H
RT'K K e
(4.16)
Model discrimination was based on the thermodynamic consistency and
statistical validity of the estimated parameters and the overall adequacy of fitting
given by the F-value with a confidence level , as proposed by Froment and
Bischoff [26] (cf. Eq. (4.17)),
2
21
mod
k
mod obs
k
r
pF value F p,k p;
r r
k p
(4.17)
where k and p stand for the number of experiments and estimated parameters,
respectively. The statistical significance of parameters was assessed through
the t-value, which is equal to the parameter estimate minus zero (null
hypothesis) divided by the standard deviation of the parameter. If the obtained
t-value is lower than the t-value for = 0.95 and (k-p) degrees of freedom, or
has a large confidence interval including zero, it has no significant contribution
to the rate equation and can be deleted [26].
The kinetic models proposed in section 4.2 were then rejected based on
the criteria mentioned above. However, the experimental data could be quite
well described after deleting the hydrogen and carbon monoxide coverage terms
from the denominator of Eq. (4.6) and Eq. (4.7). The simplified reaction rate
equations are as follows:
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
131
2 2 4 2
4
2 22
2
0 5 0 5 2
CO H CH H
CH 2 4
CO HH O
OH 0 5
H
1
1
. .
O
eq
.
k p p p pr
p p KpK
p
(4.18)
2 2 4 2
4
2 22 2
0 5 0 5 2
CO H CH H
CH 2 4
CO HH O H O
1
1
. .
O
eq
k p p p pr
p p KK p
(4.19)
The kinetic model assuming water as most abundant species (Eq. (4.19))
was however rejected because the obtained F-value was lower than the F-value
for the model expressed by Eq. (4.18) (209 vs. 259).
Fig. 4.5 shows the parity plot of the observed reaction rates against those
obtained by the model of Eq. (4.18).
Fig. 4.5 – Parity of plot of rmod versus robs showing adequacy of fit of proposed model (Eq. (4.18)).
Table 4.6 lists the parameters estimates with a 95 % confidence level (at
reference conditions).
robs
x 100 / mol gcat
-1 h
-1
0 2 4 6 8 10 12 14
r mo
d x
10
0 /
mo
l g
ca
t-1 h
-1
0
2
4
6
8
10
12
14
350 ºC
325 ºC
300 ºC
275 ºC
250 ºC
Chapter 4
132
Table 4.6 – Parameter estimates at reference temperature (573 K) and corresponding lower (LL) and upper (UL) limits of the 95 % confidence interval.
Parameter Units Estimate t-value LL UL
0
'k -1 -1mol g h 100 0.8936 8.959 0.7938 0.9933
ER
K
14275 13.34 13205 15345
0,OH
'K kPa-0.5 0.4326 3.001 0.2885 0.5768
HR
K
7414 3.626 5369 9458
Besides the significance of the estimated parameters, the F value was
also satisfied (i.e. 259 > 2.76). The estimated activation energy was 118.7
kJ·mol-1, a value close to others reported in literature for similar nickel catalysts
(e.g. 105.9 kJ·mol-1 in [30] and 113.5 kJ·mol-1 in [31]). The adsorption enthalpy
of the hydroxyl coverage term is positive, which is in agreement with the results
by Koschany et al. [23]. The obtained hydroxyl enthalpy of adsorption was 61.6
kJ·mol-1, a value which is however 2.75 times higher than the value reported by
Koschany et al. [23].
4.4.2.2 Model validation
A one-dimensional pseudo-homogeneous plug-flow model was used to
validate the selected kinetic expression against the experimental data. The
following main assumptions have been made:
1) Isothermal operation;
2) Negligible axial and radial dispersion;
3) Negligible mass and heat transport limitations within the catalyst
(internal) and between the catalyst and the bulk gas phase (external);
4) Ideal gas behaviour.
5) Reaction takes place on the catalyst surface.
The model included the estimation of pressure (P) along the dimensionless
reactor length (z) by the Ergun equation [32]:
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
133
2
g g 2o o3 2 3
p p
1 1d150 1 75
d
PL u . u
z d d
(4.20)
where L is the reactor length, is the column porosity, pd is the particle diameter,
g is the gas mixture viscosity, ou is the fluid superficial velocity and g is the
gas mixture density.
The density of the gas mixture g was calculated locally through
g i i1
n
i
Py M
RT
(4.21)
where R is the ideal gas constant, while iy and iM stand for the mole fraction
and molar mass of species i, respectively, and n for the number of species.
The gas mixture viscosity g was estimated locally applying the Wilke
method [33], as follows (Eq. (4.22)):
i ig
j ijij
y
y
(4.22)
with parameter ij being estimated through Eq. (4.23) [33]
21 4
ji
j i
ij 1 2
i
j
1
8 1
M
M
M
M
(4.23)
Chapter 4
134
and the viscosity of the pure components i by Eq. (4.24)
B
i
2
A
C D1
T
T T
(4.24)
where A, B, C and D are coefficients that depend on the species and were
retrieved from Aspen Properties software. Table 4.7 lists the coefficients A-D
used to determine pure viscosity of each species.
Table 4.7 – List of coefficients A-D to determine pure viscosity (in Pa·s) of each species.
Species A B C D
CO2 2.1480 x 10-6 0.46000 290.000 0 H2 1.7970 x 10-7 0.68500 -0.590 140 CH4 5.2546 x 10-7 0.59006 105.670 0 H2O 1.7096 x 10-8 1.11460 0.000 0 N2 6.5592 x 10-7 0.60810 54.714 0
The fluid superficial velocity ou was estimated along z by the following
differential equation obtained from the total mass balance to the reactor in
steady-state (Eq. (4.25)):
o o cati
r
d d
d d i
u u WP RTLr
z P z V P (4.25)
where catW is the weight of catalyst, rV is the reactor volume and i is the
stoichiometric coefficient of species i (excluding N2).
The mole fraction of species i along z was estimated by the following
general equation obtained from the partial mass balance to the reactor in steady-
state (Eq. (4.26)):
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
135
o cati i ii
o r o
dd d
d d d
u Wy y yRT L Pr
z u z V P u P z (4.26)
The initial conditions for solving the differential equations (4.20), (4.25) and
(4.26) were in0z
P P
, 0
0o ozu u
and 0
i i0zy y
, where the superscript “0”
refers to feed conditions.
Fig. 4.6 shows the model predictions of methane yield compared to the
observed values. A very good agreement was found between model predicted
and observed values along the large methane yield and temperature ranges.
Fig. 4.6 – Model predicted vs. observed values of the methane yield at different temperatures.
Fig. 4.7 shows the simulation profiles of the species mole fractions and
methane formation reaction rate along the dimensionless reactor length
calculated in conditions of maximum rate (i.e. higher temperature and lower
2
incat COW F ). The predicted values at the reactor outlet (i.e. z=1) are compared
with the experimental values (see the red squares in Fig. 4.7).
YCH
4,obs
/ %
0 20 40 60 80
YC
H4,m
od /
%
0
20
40
60
80
350 ºC
325 ºC
300 ºC
275 ºC
250 ºC
Chapter 4
136
Fig. 4.7 – Simulated values along the dimensionless reactor length of: a) the species
mole fractions and b) reaction rate. T=350 ºC and 2
in -1
cat CO 1 52 g h molW F . . Red
squares stand for the experimental values.
Fig. 4.7a shows that H2 and CO2 are consumed along all the reactor length
while H2O and CH4 are continuously produced because one is still far from
equilibrium conditions. N2 fraction slightly increases because of the
stoichiometry of the reaction. The experimental values measured at the reactor
outlet agree well with the predicted values, being the most noticeable deviation
observed for hydrogen. It should be however recalled that hydrogen fraction was
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
137
determined through the molar balance and it is therefore subject to the errors of
the determination of each species composition.
Regarding to the reaction rate, it decreases steeply during the first half of
the reactor length and then starts declining slowly until a value of 0.10
-1catmol g h is reached at the exit, a value that is very close to the observed
reaction rate (0.13 -1catmol g h ) (Fig. 4.7b).
The simulated methane yield profiles along the reactor length obtained in
the temperature and reaction rate limits are shown in Fig. 4.8.
Fig. 4.8 – Simulated methane yield profile along the dimensionless reactor length for the highest and lowest robs conditions measured at 250 ºC and 350 ºC. Red squares symbols stand for the experimental values.
The model was able to capture the behaviour of the reactor performance
in the considered temperature window limits (250 ºC to 350 ºC) and in both of
minimum and maximum reaction rates conditions.
z
0.0 0.2 0.4 0.6 0.8 1.0
YC
H4
/ %
0
20
40
60
80
350 ºC
250 ºC
2
in
cat CO minW F
2
in
cat CO maxW F
2
in
cat COmax
W F
2
in
cat CO minW F
Chapter 4
138
4.5 Conclusions
The kinetics of the methanation reaction over an industrial nickel-based
catalyst was determined for the relevant 250 ºC to 350 ºC temperature window.
Three reaction mechanisms assuming different intermediate species were
selected from the literature. The mechanism which assumes hydrogen and
carbon dioxide dissociation followed by hydrogenation of adsorbed carbon
monoxide to yield a formyl species, assuming hydroxyl as the most abundant
species, showed a good fit to the experimental data.
An isothermal plug-flow model including the proposed kinetic rate equation
with estimated parameters was used to simulate the fixed-bed reactor. The
reactor model satisfactorily captured the experimental observations in limiting
conditions of temperature and CO2
in
catW F , allowing to validate the proposed
(intrinsic) kinetic rate model.
The selected kinetic rate and estimated model parameters can thus be
transferred and used for modeling, simulation and design of industrial reactors
used for CO2 utilization applications, where methanation is standing out among
the existing options.
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
139
4.6 Nomenclature
Parameter Description Units
a' Specific external area of the catalyst
particle ( p6a' d for spheres) m-1
Ca Carberry number -
bC CO2 concentration in the bulk phase mol·m-3
sC CO2 concentration at the surface of the catalyst particle
mol·m-3
pd Particle diameter m
effD Effective mass diffusivity in the catalyst m2·s-1
Dr Reactor diameter m
aE Activation energy J·mol-1
4
outCHF CH4 outlet molar feed flow rate mol·s-1
2
inCOF CO2 inlet molar feed flow rate mol·s-1
h Heat transport coefficient between gas and particle
W·m-2·K-1
k Reaction rate constant mol·g-1·s-1·Pa-n kg Mass transport coefficient m·s-1
0k Pre-exponential factor of rate constant mol·g-1·s-1·Pa-n
0K Pre-exponential factor of adsorption equilibrium constant
Pa-1 or Pa-0.5
K Adsorption equilibrium constant Pa-1 or Pa-0.5 Keq Equilibrium constant for methanation Pa-2 L Reactor length m
pL Characteristic catalyst dimension
6 for spheresp pL d m
Mi Molar weight of species i kg·mol-1 nr Reaction order - pi Partial pressure of species i Pa P Total pressure Pa Q Volumetric feed flow rate mLn·min-1
4CHr Reaction rate of methane formation mol·g-1·s-1
obsr Observed reaction rate mol·g-1·s-1
modr Model predicted reaction rate mol·g-1·s-1
R Ideal gas constant 8.314 J·mol-1·K-1 T Absolute temperature K
bT Temperature in the bulk phase K
Chapter 4
140
sT Temperature at the catalyst particle surface
K
Tm Mean temperature of range (Tm=300 ºC) K
ou Fluid superficial velocity m·s-1
rV Reactor volume m3
catW Catalyst weight kg
2COX CO2 conversion -
iy Mole fraction of species i -
4CHY Methane yield -
z Dimensionless reactor length - Greek letters
H Enthalpy of adsorption J·mol-1
rH Heat of reaction J·mol-1
e Dimensionless number for extraparticle heat transport
-
i Dimensionless number for intraparticle heat transport
-
e External Arrhenius number -
i Internal Arrhenius number -
Column porosity (0.36-0.38) -
g Density of the gas mixture kg·m-3
p Density of the catalyst particle g·m-3
Internal effectiveness factor - Generalized Thiele modulus -
2 Wheeler-Weisz modulus -
g Thermal conductivity of the gas mixture W·m-1·K-1
eff Effective thermal conductivity of the catalyst particle
W·m-1·K-1
i Viscosity of pure species i Pa·s
g Viscosity of the gas mixture Pa·s
i Stoichiometric coefficient of species i -
Ψ Inert dilution factor -
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
141
4.7 References
[1] H.F. Rase, Handbook of commercial catalysts: heterogenous catalysts, CRC
Press, Boca Raton, 2000.
[2] M.V. Twigg, Catalyst Handbook, 2nd ed., Manson Publishing Ltd, London,
1996.
[3] J. Kopyscinski, T.J. Schildhauer, S.M.A. Biollaz, Production of synthetic
natural gas (SNG) from coal and dry biomass - A technology review from 1950
to 2009, Fuel, 89 (2010) 1763-1783.
[4] M. Götz, J. Lefebvre, F. Mörs, A. McDaniel Koch, F. Graf, S. Bajohr, R.
Reimert, T. Kolb, Renewable Power-to-Gas: A technological and economic
review, Renewable Energy, 85 (2016) 1371-1390.
[5] C.V. Miguel, M.A. Soria, A. Mendes, L.M. Madeira, Direct CO2 hydrogenation
to methane or methanol from post-combustion exhaust streams - a
thermodynamic study, Journal of Natural Gas Science and Engineering, 22
(2015) 1-8.
[6] P. Lunde, F. Kester, Carbon dioxide methanation on a ruthenium catalyst
Industrial & Engineering Chemistry Process Design and Development 13 (1974)
27–33.
[7] E.Z. Golosman, V.N. Efremov, Industrial catalysts for the hydrogenation of
carbon oxides, Catal. Ind., 4 (2012) 267-283.
[8] T.T.M. Nguyen, L. Wissing, M.S. Skjøth-Rasmussen, High temperature
methanation: Catalyst considerations, Catalysis Today, 215 (2013) 233-238.
[9] Y. Li, Q. Zhang, R. Chai, G. Zhao, Y. Liu, Y. Lu, Structured Ni-CeO2-Al2O3/Ni-
foam catalyst with enhanced heat transfer for substitute natural gas production
by syngas methanation, ChemCatChem, 7 (2015) 1427-1431.
[10] Z. Wan, J. Jiang, H. Ma, Y. Li, Y. Lu, F. Cao, A study of the heat transfer
characteristics of novel Ni-foam structured catalysts, Canadian Journal of
Chemical Engineering, 94 (2016) 2225-2234.
Chapter 4
142
[11] G.H. Watson, Methanation catalysts, International Energy Agency Coal
Research, London, 1980.
[12] Chemical catalysts: catalysts for olefin processes, in: J. Matthey (Ed.), 2012.
[13] Catalysts for Syngas, in: C.I. Ltd. (Ed.), 2010.
[14] PK-7R low temperature methanation catalyst, in: H. Topsoe (Ed.), 2001.
[15] Alvigo-Matros catalysts, in: Alvigo (Ed.), 2004.
[16] C.V. Miguel, M.A. Soria, A. Mendes, L.M. Madeira, A sorptive reactor for
CO2 capture and conversion to renewable methane, Chemical Engineering
Journal, 322 (2017) 590-602.
[17] J.-A. Dalmon, G.A. Martin, Intermediates in CO and CO2 hydrogenation over
Ni catalysts, Journal of the Chemical Society, Faraday Transactions 1: Physical
Chemistry in Condensed Phases, 75 (1979) 1011-1015.
[18] G.D. Weatherbee, C.H. Bartholomew, Hydrogenation of CO2 on group VIII
metals: II. Kinetics and mechanism of CO2 hydrogenation on nickel, Journal of
Catalysis, 77 (1982) 460-472.
[19] S.i. Fujita, M. Nakamura, T. Doi, N. Takezawa, Mechanisms of methanation
of carbon dioxide and carbon monoxide over nickel/alumina catalysts, Applied
Catalysis A, General, 104 (1993) 87-100.
[20] S.I. Fujita, N. Takezawa, Difference in the selectivity of CO and CO2
methanation reactions, Chemical Engineering Journal, 68 (1997) 63-68.
[21] Q. Pan, J. Peng, T. Sun, S. Wang, S. Wang, Insight into the reaction route
of CO2 methanation: Promotion effect of medium basic sites, Catalysis
Communications, 45 (2014) 74-78.
[22] Z.A. Ibraeva, N.V. Nekrasov, B.S. Gudkov, V.I. Yakerson, Z.T.
Beisembaeva, E.Z. Golosman, S.L. Kiperman, Kinetics of methanation of carbon
dioxide on a nickel catalyst, Theoretical and Experimental Chemistry, 26 (1991)
584-588.
[23] F. Koschany, D. Schlereth, O. Hinrichsen, On the kinetics of the
methanation of carbon dioxide on coprecipitated NiAl(O)x, Applied Catalysis B:
Environmental, 181 (2016) 504-516.
Intrinsic kinetics of CO2 methanation over an industrial nickel-based catalyst
143
[24] I. Kiendl, M. Klemm, A. Clemens, A. Herrman, Dilute gas methanation of
synthesis gas from biomass gasification, Fuel, 123 (2014) 211-217.
[25] J. Horák, J. Pasek, Design of industrial chemical reactors from laboratory
data, Heyden & Son Ltd., London, 1978.
[26] G.F. Froment, K.B. Bischoff, Chemical reactor analysis and design, 2nd ed
ed., John Wiley, New York, 1990.
[27] F. Ocampo, B. Louis, L. Kiwi-Minsker, A.-C. Roger, Effect of Ce/Zr
composition and noble metal promotion on nickel based CexZr1−xO2 catalysts for
carbon dioxide methanation, Applied Catalysis A: General, 392 (2011) 36-44.
[28] B.W. Wojciechowski, N.M. Rice, Experimental methods in kinetic studies,
Revised edition ed., Elsevier Science B.V., Amsterdam, 2003.
[29] F.H.M. Dekker, A. Bliek, F. Kapteijn, J.A. Moulijn, Analysis of mass and heat
transfer in transient experiments over heterogeneous catalysts, Chemical
Engineering Science, 50 (1995) 3573-3580.
[30] T. Van Herwijnen, H. Van Doesburg, W.A. De Jong, Kinetics of the
methanation of CO and CO2 on a nickel catalyst, Journal of Catalysis, 28 (1973)
391-402.
[31] M. Swickrath, M. Anderson, The Development of Models for Carbon Dioxide
Reduction Technologies for Spacecraft Air Revitalization, in: 42nd International
Conference on Environmental Systems, American Institute of Aeronautics and
Astronautics, 2012.
[32] S. Ergun, Fluid flow through packed columns, Chemical Engineering
Progress, 48 (1952) 89-94.
[33] B.E. Poling, J.M. Prausnitz, J.P. O'Connel, The properties of gases and
liquids, 5th ed ed., The McGraw-Hill Companies, New York, 2001.
145
Chapter 5. A sorptive reactor for CO2 capture and conversion to renewable methane
The contents of this chapter were adapted from: C.V. Miguel, M.A. Soria, A. Mendes,
L.M. Madeira, A sorptive reactor for CO2 capture and conversion to renewable methane.
Chemical Engineering Journal, 2017, 352, 590-602.
The aim of this work is to contribute for the scientific
advancement of carbon capture and utilization technologies (CCU),
while exploring the integration of intermittent renewable electricity
production and Substitute Natural Gas (SNG) production, the so-called
“Power-to-Gas” concept. In this regard, a sorptive reactor for carrying
in the same unit both CO2 capture and conversion to SNG using
renewable H2 was studied, in a perspective of process intensification.
The sorptive reactor containing a layered bed of a K-promoted
hydrotalcite to capture CO2 by sorption (step 1) and a nickel-based
catalyst for CO2 hydrogenation (step 2) is operated at 300-350 ºC and
low pressure (≤2.5 bar). Integration of CO2 capture and conversion in
the same unit leads to enhanced sorption capacities and desorption
kinetics promoted by the steam produced in situ during the reactive
regeneration stage (methanation reaction).
The sorptive reactor working under continuous operation mode
allows to: i) capture ca. 0.30 moles of CO2 per kg of sorbent and per
sorption cycle, at 350 ºC and = 0.2 bar; ii) completely convert the
captured CO2 into CH4; iii) reach a productivity of ca. 2.36
; iv) avoid CO formation at 300 ºC and 1.34 bar and
v) reach a CH4 purity of 35 % at 350 ºC after N2 purge.
Chapter 5
146
5.1 Introduction
Curbing CO2 concentration in the atmosphere requires considering
simultaneously a diversity of approaches, such as the development and adoption
of carbon capture and sequestration (CCS) or utilization (CCU) technologies,
and growing policy support towards the development and penetration of energy
from renewable sources across the power market [1-3]. Regarding the adoption
of end-pipe solutions for CO2 emissions mitigation, CCU has relevant
advantages compared to CCS because it avoids the (critical) cost of CO2
transportation to a storage site and CO2 is seen as a feedstock available at zero
cost from which valuable products can be obtained [4, 5].
Among the options for CO2 utilization, its conversion to methane can offer
a way to simultaneously avoid CO2 emissions and chemically store, in the form
of methane, the surplus energy from renewable sources which is produced
during off-peak periods [6]. This concept, also known as “Power-to-Gas”, relies
on the production of H2 making use of the excess electric power for the
electrolysis of water. Further conversion of hydrogen to methane through the
Sabatier reaction (Eq. (5.1)) is envisaged because it benefits from a well-
established transmission, distribution and storage infrastructure that was
developed for natural gas, a fossil fuel which consists primarily of methane in a
concentration up to 70-90 mol % [7, 8].
-12 2 4 2 298 KCO + 4H CH + 2H O Δ = - 165 kJ molH (5.1)
Consequently, the Power-to-Gas concept brings the possibility of
connecting the power grid to the natural gas grid, thereby facilitating the
integration between the power and energy sectors. Moreover, since methane is
a low-carbon hydrogen carrier and produces the lowest amount of CO2 upon
combustion compared to other hydrocarbons, it provides a secure and efficient
way of store and supply energy from renewable sources, while simultaneously
A sorptive reactor for CO2 capture and conversion to renewable methane
147
reducing the dependence on fossil fuels and supporting the transition towards a
low-carbon economy [3, 6, 7, 9].
Direct methanation of CO2 from flue gas has been addressed in the
literature in catalytic studies using Ni-based catalysts [10] and in Chapter 2,
which showed the necessity to previously separate CO2 from other co-existent
species in the flue gas, particularly O2 (which consumes H2) and SO2 (that
causes catalyst deactivation). Sorption-enhanced methanation processes
relying on the Le Chatelier principle provide also an interesting way of producing
high grade methane through the in situ capture of H2O by an adsorbent (e.g.
zeolite 4A) during the reaction, this way shifting the reversible Sabatier reaction
(Eq. (5.1)) towards the products side [11, 12]. However, it requires a unit for
previous capture of CO2 and its transport to the sorption-enhanced methanation
reactor.
In this perspective, there is a clear advantage on integrating CO2 capture
from flue gas and its conversion in the same unit. This concept was recently
addressed using a dual function material (DFM) that contains 10 % of
nanodispersed CaO (adsorbent) and 5 % of Ru metal (catalyst) dispersed on
different commercial Al2O3 carriers [13]. However, the reported sorption capacity
per sorption cycle (11.18 2CO DFMg kg , i.e. 0.025
2CO CaOmol kg ), and CH4
productivity 4
4
CH-3CH DFM
Ru
mol3.27 g kg , i.e. 7.66 10
kg h
obtained during 10
cycles (T = 320 ºC; 2COp 0.075 atm; tcycle = 80 min) were very low [13]. In this
work, a mixed bed made of a commercial K-promoted hydrotalcite as CO2
sorbent and a commercial Ni-based methanation catalyst were employed. This
sorbent was chosen since hydrotalcites are the most suitable CO2 sorbents at
intermediate temperatures (200 ºC - 400 ºC), showing good sorption stability
under cyclic operation, reasonable kinetics and easy regeneration by pressure
swing [14-20]. Besides, hydrotalcites are also cheap materials as compared to
other available options (e.g. CaO sorbents). A nickel catalyst was selected
Chapter 5
148
because it is suitable for operation at this temperature range; moreover, Ni-
based catalysts are the most widely used materials for CO2 methanation at the
industrial scale due to their good activity and lower price when compared to
noble-based catalysts (e.g. Ru or Rh) [21, 22]. The use of a sorptive reactor for
CO2 capture and methanation is studied assessing the performance of
continuous cyclic operation, controlling the CO2 sorption capacity, CO2
conversion, CH4 productivity and CH4 purity. Given the complexity related to
processes integration, the flue gas composition is simplified to CO2 and N2,
although a pure H2 stream is used for regenerating the sorbent, as expected to
occur in real applications. Overall, this work shows the advantages of using such
a reactor for CO2 capture and utilization.
5.2 Experimental
5.2.1 Experimental setup and hybrid unit
A stainless-steel column with 15 cm length and 2.08 cm internal diameter
was placed inside a tubular oven (model Split from Termolab, Fornos Eléctricos,
Lda.) equipped with a 3-zone PID temperature controller (model MR13 from
Shimaden). The type-K thermocouples used to measure and control the oven
temperature were placed in contact with the reactor wall; the temperature
difference between the reactor wall and the bed was lower than 2 ºC. An
illustration of the setup used in this work is given in Fig. 5.1.
CO2 (99.998 %, Air Liquide), H2 (99.9995 %, Air Liquide) and N2 (99.9995
%, Air Liquide) were fed to the system using mass flow controllers (model F201
from Bronkhorst-High Tech). The flow rate of the stream leaving the system was
measured with a mass flow meter (model F101 from Bronkhorst-High Tech) and
corrected based on stream composition.
A sorptive reactor for CO2 capture and conversion to renewable methane
149
Fig. 5.1 – Illustration of the experimental setup.
The pressure was measured by means of pressure transducers (model
PMP 4010 from Druck) located before and after the packed-bed unit. The steam
produced during the reaction that left the system was condensed in a home-
assembled Peltier cold-trap.
The CO2 outlet concentration was recorded along time using an infrared
analyser (Servomex, model 4210). The concentration of other species (such as
N2, CH4 and CO) produced during the reactive regeneration stage was
determined by collecting and analysing samples along time using a 16-loop
valve (VICI, model ST) coupled to a gas chromatograph (Dani, model 1000). The
chromatograph was equipped with a micro-TCD detector (VICI, model TCD-C-
220) and a capillary column (Supelco, carboxen 1010 plot).
The reactor column was packed in alternating layers of sorbent pellets
(23.89 g) and methanation catalyst particles (7.58 g), as depicted in Fig. 5.2.
Chapter 5
150
Fig. 5.2 – Photo of the unit inside the oven (left) and a sketch of the layered bed (right).
The column porosity was 0.30. A commercial hydrotalcite and a
commercial Ni catalyst were employed in this study for the following reasons: i)
they can be purchased in quantities that allow using a larger column compatible
with the available test bench; ii) available in millimetre range size, which
minimizes the bed pressure drop and; iii) better performance uniformity of these
materials. The sorbent is synthesized through the metal alcoholate route and its
properties are listed in Table 5.1, whilst the catalyst was supplied under a non-
disclosure agreement.
Table 5.1 – Properties of the hydrotalcite sorbent.
Property Value / Info Units
Commercial reference Pural MG30K - Mg/Al ratio 0.5 wt. % / wt. %
K2CO3 content 17 wt. % Shape Cylindrical -
Dimensions (d x L) 5 x 5 mm
Bulk density b,sorb 846 kg m-3
Solid density s,sorb 2840 kg m-3
Particle porosity p,sorb 0.70 -
A sorptive reactor for CO2 capture and conversion to renewable methane
151
5.2.2 Sorption-desorption/reaction experiments
The layered bed was activated in situ increasing the temperature up to 320
ºC and holding at this temperature for 1 h under a 10 % H2/N2 stream (total flow
rate: 100 mLn∙min-1) for catalyst reduction. Bed temperature was continuously
recorded using 4 type-K thermocouples inserted laterally and both radially
centred and axially equidistant (cf. Fig. 5.2). The temperature profile along the
column length was negligible (< 2 ºC) under inert atmosphere. Sorption
experiments were performed using a typical post-combustion (dry) stream
composition obtained from coal power plants: 15 % CO2 balanced in N2. The bed
was filled with N2 prior to each sorption breakthrough, the total flow rate was kept
at 100 mLn min-1 and the pressure inside the unit was 1.34 bar. Regeneration
started when the feed was switched using a 4-way valve placed near the column
inlet to a stream containing N2 or H2 (the later for reactive regeneration), keeping
constant the total volumetric flow rate. Both streams used during sorption and
regeneration steps were fed downwardly towards the bed. Sorption-desorption
cycles were performed at 350 ºC until a steady CO2 sorption capacity was
reached and sorption-reaction cycles were carried afterwards. Two series of
sorption-reaction cycles were carried out at 350 ºC and 300 ºC. The total
pressure inside the column during the sorption step was 1.34 bar, except in
cycles 16 and 17 where it was increased to 2.00 and 2.50 bar, respectively.
Table 5.2 lists the operation conditions used in the cyclic experiments.
Table 5.2 – Operation conditions used in the experimental campaign.
Cycle(s) T (ºC) P (bar) Sorption end Regeneration method
0-5 350 1.34 Bed saturation Normal (w/ N2) 6-10 350 1.34 Bed saturation Reactive (w/ H2) 11-15 300 1.34 Bed saturation Reactive (w/ H2) 16 300 2.00 Bed saturation Reactive (w/ H2) 17 300 2.50 Bed saturation Reactive (w/ H2) 18 300 1.34 Breakthrough time Reactive (w/ H2) 19-21 350 1.34 Breakthrough time Reactive (w/ H2) 22-24 300 1.34 Breakthrough time Reactive (w/ H2)
Chapter 5
152
The system was flushed with N2 at the end of the experimental campaign
to remove carbon oxides before the temperature was decreased; this avoids the
formation of extremely toxic nickel carbonyl at temperatures below 150 ºC [21,
23]. Afterwards, since reduced Ni is pyrophoric, the bed was passivated at 50
ºC feeding a stream with increasing amounts of O2 in N2; the process ended
when the temperature remained constant, i.e., when the catalyst bed was
completely oxidised, as suggested elsewhere [21, 24]. Finally, the unit was
brought to room temperature under N2.
Scanning Electron Microscopy (SEM) was performed on a FEI Quanta
400FEG ESEM instrument coupled with Energy Dispersive Spectroscopy (EDS)
(EDAX Genesis X4M) to investigate the structure and composition of the fresh
(i.e. as supplied) and used sorbent.
5.2.3 Process indicators: definitions and equations
Carbon dioxide sorption capacity,2COq , of the reactor was determined as
follows:
,f
2
2
2
2
outCOin
CO inCO0
COads
1 dSt
S
yF t
yq
m
(5.2)
where 2
inCOF is the CO2 molar flow rate at the column inlet, adsm is the sorbent
mass,2
inCOy and
2
outCOy are the CO2 fractions at the system inlet and outlet
streams, respectively; St stands for time during the sorption stage and
,S ft for
the duration of this step.
Carbon dioxide conversion was obtained from:
A sorptive reactor for CO2 capture and conversion to renewable methane
153
2 2
2
2
sorbed desorbed/unconvertedCO , CO ,
CO , sorbedCO ,
100k k
k
k
n nX
n
(5.3)
where 2
sorbedCOn (Eq. (5.4)) and
2
desorbed/unconvertedCOn (Eq. (5.5)) are the number of
CO2 moles that were sorbed and those which left the unit during the capture and
reactive regeneration stages of a given cycle k, respectively.
,
2
2 2
2
outCOsorbed
CO CO inCO0
1 dS ft
inS
yn F t
y
(5.4)
,
2 2
desorbed/unconverted out outCO CO
0
dR ft
Rn F y t (5.5)
outF is the total molar flow rate at the column outlet, which is calculated from the
volumetric flow rate; Rt stands for time during the regeneration stage and ,R ft for
the duration of this step. Time values excluded the delay related to the tubing
volume; intra-particular amount of CO2 at the end of the sorption runs was found
to be negligible as compared to that retained in the sorbent.
Carbon balance in a sorption-reaction cycle k (Eq. (5.6)) was calculated
as follows:
2 4
2
2 2
desorbed/unconverted produced producedCO , CO,CH ,
CO , 1sorbed sorbedCO , CO , 1
100
(1 ) 100
k kk
kk
k k
n n nC
Xn n
(5.6)
and
Chapter 5
154
,produced out out
0
dR ft
i Rin F y t (5.7)
where i stands for produced species CH4 and CO. The carbon balance equation
of a given cycle k (Eq. (5.6)) includes the unconverted CO2 quantity of the
previous cycle (k-1), since it remains available to be converted in cycle k. The
error of the carbon balance was in average less than 6 % under pseudo-steady
state conditions (i.e. after 5 cycles operating in the same conditions).
The process performance was evaluated considering methane
productivity (Eq. (5.8)) and purity (dry basis, Eq. (5.9)):
4
out outCH
0
cat
d
R,ft
R
S,f R,f
F y t
Prodm t t
(5.8)
4
out outCH
0
out outj
0
d
100
d
R,f
R,f
t
R
t
Rj
F y t
Pur
F y t
(5.9)
where catm and S,f R,ft t stand for the catalyst mass and the duration of a
sorption-reaction cycle, respectively, and j for all the species present in the
system (except H2O).
5.3 Results and Discussion
In Section 5.3.1 the sorptive reactor is studied considering full saturation
of the bed before starting the regeneration step – discontinuous operation mode.
A sorptive reactor for CO2 capture and conversion to renewable methane
155
However, for the process being implemented on a continuous basis without
emitting (environmental perspective) and wasting (economical perspective) CO2,
the sorption step should end before CO2 breaks through the column; these
experiments are addressed in Section 5.3.2.
5.3.1 Sorptive reactor under discontinuous operation
mode
5.3.1.1 CO2 sorption working capacity
After the activation procedure, the layered bed was subjected to six
sorption-desorption cycles; Fig. 5.3 shows the CO2 breakthroughs and sorption
capacities calculated using (Eq. (5.2)). The sorption capacity decreased 38.5 %
with cycling (from 0.52 mmol·g-1 to 0.32 mmol·g-1), and the steady working
capacity was mostly reached at the end of the first cycle (cf. Fig. 5.3).
Fig. 5.3 – CO2 sorption breakthroughs curves (cycles 0-5) at 350 ºC and 1.34 bar. Insertion shows the sorption capacity in each cycle.
The sorption capacity decrease with sorption-desorption cycling has
already been reported for hydrotalcite-type sorbents (e.g. [16, 25]). The work by
tS
/ s
0 2000 4000 6000 8000 10000
y o
ut
C
O2
0.00
0.05
0.10
0.15
Cycle 0
Cycle 1
Cycle 2
Cycle 3
Cycle 4
Cycle 5
cycle
-1 0 1 2 3 4 5 6
qC
O2
/ m
ol kg-1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Chapter 5
156
Du et al. [26] provides an interesting insight about the transformations taking
place on the hydrotalcite during CO2 sorption and desorption. Du et al. [26] found
that under dry conditions CO2 was reversibly sorbed in a K-promoted
hydrotalcite in at least two different active sites: those that favour the formation
of bidentate carbonate (more active sites with favourable kinetics and
thermodynamics, but less abundant) and others promoting the formation of
unidentate and bridged carbonate species, also with fast sorption kinetics.
However, it was also observed that unidentate and bridged carbonates start
being slowly converted into irreversibly sorbed bulk polydentate carbonate,
which explains the sorption capacity decrease with cycling. Based on this
mechanism we have proposed on a previous work a kinetic model assuming
two active sites for CO2 sorption with parallel contributions [25]. The kinetic
model provided an excellent fit to the uptake and downtake curves of CO2
sorption on a K-promoted hydrotalcite at 200-300 ºC, clearly supporting the
validity of the sorption mechanism.
5.3.1.2 Effect of steam produced during reactive regeneration
Reactive regeneration experiments were performed after the sorbent bed
achieved its working capacity. The first experiment under these conditions was
performed in cycle 6 (Table 5.2) in which the bed was fed with a pure H2 stream
and CO2 started to desorb by purge gas stripping. Subsequently, desorbed CO2
in the gas phase became available to react with H2 when passing through the
catalyst layers.
Besides the ability to convert CO2 to CH4, a noticeable feature of the
reactive regeneration was observed: the sorbent is regenerated faster in these
conditions than through desorption with N2. Carbon dioxide outlet concentration
decreases sharply under reactive regeneration and reaches 1 % after 554 s
(cycle 6), while it takes 1309 s under non-reactive regeneration conditions (cycle
5) (cf. Fig. 5.4a-b).
A sorptive reactor for CO2 capture and conversion to renewable methane
157
Fig. 5.4 – Total pressure and CO2 and CH4 outlet fraction histories during regeneration at 350 ºC for the initial 2000 s (cycles 5-7).
Chapter 5
158
It is also noticed that the total pressure drops abruptly ca. 0.2 bar under
reactive regeneration due to H2 consumption through the Sabatier reaction, until
a plateau is reached at ca. 1.14 bar (Fig. 5.4b-c). Afterwards, the pressure
increases when unreacted H2 starts leaving the column (not shown) and a slight
overshoot occurs when CH4 fraction reaches its maximum, ending in a final value
of 1.17 bar. The pressure could not reach its initial value of 1.34 bar (as in cycle
5) because the pressure drop caused by the same flow rate of H2 or N2 is
different (also observed in blank experiments).
Besides fast desorption, reactive regeneration in cycle 6 was more
effective, so that it boosted the CO2 sorption capacity in cycle 7 (to 0.58 mmol·g-
1) and the initial bed sorption capacity in cycle 0 (0.52 mmol·g-1) was recovered
throughout cycles 8 to 10 (Fig. 5.5). However, CO2 sorption in cycle 0 was faster
than in cycle 8 (Fig. 5.6b), which might be related to changes that took place on
the hydrotalcite structure due to the presence of steam formed in the Sabatier
reaction, as observed by scanning electron microscopy analysis (SEM)
performed on the fresh and used sorbent (Fig. 5.7).
Fig. 5.5 – Evolution of CO2 capture and conversion with cycling at 350 ºC and 300 ºC.
cycle
-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
nso
rbe
d,
de
so
rbe
d /
mm
ol
0
5
10
15
20
25
XC
O2
/ %
50
60
70
80
90
100
nCO2
desorbed/unconverted (Eq. 5)
nCO2
sorbed (Eq. 4)
XCO2
(Eq. 3)
300 ºC 350 ºC
A sorptive reactor for CO2 capture and conversion to renewable methane
159
Fig. 5.6 – CO2 breakthrough curves for cycles 0 and 8 with “zoom in” at a) beginning of sorption (xx axis) and b) when reaches the saturation (yy axis).
SEM images of the aged hydrotalcite showed that the characteristic
potassium-rich needle-like zones of fresh sorbent, also reported in other studies
(e.g. [15, 25]), had disappeared (Fig. 5.7). Oliveira et al. [15] considered that the
needles might be mainly formed by K2CO3 and attributed this disappearance to
their dissolution by steam, despite that potassium was not lost and thus the same
sorption capacity was obtained here in cycles 0 and 8. Therefore, the lower
sorption kinetics of cycle 8 might be explained by the absence of these type of
sites which, based on in situ FTIR results reported by Du et al. [26] for a similar
hydrotalcite, could be responsible for CO2 sorption in the form of bidentate
carbonate.
The earlier CO2 breakthrough in cycle 8 (Fig. 5.6a) clearly supports the
evidences given by Du et al. [26] that CO2 then started to be sorbed as
unidentate and bridged carbonates, species that later are being slowly converted
into bulk polydentate carbonate, thus explaining the longer tail of the
breakthrough curve (Fig. 5.6b). Nevertheless, the earlier CO2 breakthrough in
cycle 8 compared to cycle 0 (cf. Fig. 5.6a) only represents a decrease of ca. 6.3
% of the dynamic capacity of the bed (i.e. the capacity obtained when sorption
is stopped before CO2 breakthrough - continuous operation mode, as discussed
in Section 5.3.2).
Col 8 vs Col 9
tS / s
0 500 1000 1500 2000
yC
O2
/ y
CO
2,0
0.0
0.2
0.4
0.6
0.8
1.0
tS / s
0 5000 10000 15000 20000 25000 30000
yC
O2
/ y
CO
2,0
0.85
0.90
0.95
1.00
cycle 0
cycle 8
a) b)
Chapter 5
160
Fig. 5.7 – SEM images of the fresh (first row) and used MG30K sorbent (second row) at different magnifications (10000x and 50000x, from left to right).
The enhanced desorption capacity under reactive regeneration is
highlighted through the comparison of 2COy curves of cycles 5 and 6 (Fig. 5.4a-
b); 2COy starts to decrease after 195 s in cycle 6, while under non-reactive
regeneration conditions (cycle 5) it begun after 87 s. The delay observed in cycle
6 is caused by water produced which has the ability to regenerate sorbent sites
containing irreversibly sorbed CO2 (i.e.not able to be regenerated under dry
conditions), as reported in several works dealing with sorption-enhanced
processes for H2 production with pre-combustion CO2 capture [16, 27]. Herein,
2 moles of H2O are obtained per mole of converted CO2 according to the
A sorptive reactor for CO2 capture and conversion to renewable methane
161
stoichiometry of the Sabatier reaction (Eq. (5.1)). Therefore, in the proposed
hybrid process for CO2 capture and conversion to CH4, steam is produced in
situ, thus not requiring additional costs for its industrial production, which was
reported to dominate the economy of sorption-enhanced processes [27].
Moreover, since some of the produced water molecules may be sorbed by the
hydrotalcite [28], the reaction equilibrium is shifted to the right side allowing to
produce more methane and leading to a desired methane-enrichment of the
outlet stream. The maximum methane concentration obtained at 350 ºC ranged
from 60 % (cycle 6) to 77-78% (cycles 9-10) and was reached after ca. 300 s; in
cycles 7 and 8 a slightly lower value was reached 74 % (cf. Fig. 5.4 and Fig.
C.0.1 in Appendix C).
During cycles 7-10 a roll-up effect was observed immediately before CO2
concentration starts to decrease steeply (Fig. 5.8a). In equilibrium-based
separations, the roll-up effect is observed in multicomponent mixtures when the
effluent composition of a species exceeds its inlet value, being this phenomenon
related to the displacement of the light components (weakly sorbed) by the
heavier ones (strongly sorbed) [29, 30]. Preliminary experiments allowed to
check that the sorbent does not retain CH4 (data not shown). Furthermore, the
work by Boon et al. [28] reported that this hydrotalcite has higher affinity towards
CO2 as compared to H2O. Therefore, CO2 roll-up should be based on a different
event.
Chapter 5
162
Fig. 5.8 – Effect of the cycling process at different temperatures (a-b) and pressures (c) on the roll-up effect.
tR / s
0 100 200 300 400 500 600
FC
O2
/ m
ol s
-1
0.0
2.0e-6
4.0e-6
6.0e-6
8.0e-6
1.0e-5
1.2e-5
cycle 6
cycle 7
cycle 8
cycle 9
cycle 10
tR / s
0 100 200 300 400 500 600
FC
O2
/ m
ol s
-1
0.0
2.0e-6
4.0e-6
6.0e-6
8.0e-6
1.0e-5
1.2e-5
cycle 11
cycle 12
cycle 13
cycle 14
cycle 15
350 ºC
300 ºC
tR / s
0 100 200 300 400 500 600
FC
O2
/ m
ol s
-1
0.0
2.0e-6
4.0e-6
6.0e-6
8.0e-6
1.0e-5
1.2e-5
cycle 15 (P = 1.34 bar)
cycle 16 (P = 2.00 bar)
cycle 17 (P = 2.50 bar)
300 ºC
a)
b)
c)
A sorptive reactor for CO2 capture and conversion to renewable methane
163
The analysis of the exit composition during reactive regeneration showed
that CO is being formed apart from CH4, which is expected to occur through the
parallel reverse-water gas shift reaction (RWGS) (Eq. (5.10)), as shown in
thermodynamic (e.g. [31, 32]) and kinetic studies for the methanation reaction
using a nickel catalyst (e.g. [33]).
-12 2 2 298 KCO + H CO + H O 41 kJ molH (5.10)
In Chapter 2 it was reported that CO production was slightly favoured at lower
pressures, higher temperatures and lower H2/CO2 ratios; CO production onsets,
however, for ratios as high as H2/CO2 = 4, the stoichiometric ratio of the Sabatier
reaction. In this regard, the observed roll-up effect is tentatively explained with
the production of CO2 through the water-gas shift reaction (WGS) (Eq. (5.11)),
-12 2 2 298 KCO + H O CO + H = - 41 kJ molH (5.11)
which can be catalysed by a nickel-based catalyst [34] and also by the sorbent
[16, 35]. The WGS becomes favoured along the reactor bed due to the
increasing amount of CO and H2O available. The CO2 roll-up was almost
unnoticed in cycle 6 (Fig. 5.4b and Fig. 5.8a) because less quantity of the H2O
produced was available for the WGS reaction (since it was used mainly to
recover the sorbent sorption capacity). This hypothesis is further supported
through the inspection of the temperature profiles of the regeneration step in
cycles 6 and 7-10 (Fig. 5.9a). In cycles 7-10 a small peak corresponding to a
temperature increase of ca. 1 ºC was observed in the last thermocouple placed
in the bed, whereas it was unnoticed in cycle 6, suggesting that another
(moderately exothermic) reaction, such as the WGS, also occurred.
Chapter 5
164
Fig. 5.9 – Normalized temperature histories measured in the last thermocouple (L=12 cm, z=4) during the first 200 s of CO2 regeneration at 350 ºC (a), 300 ºC (b) and different pressures (c).
A sorptive reactor for CO2 capture and conversion to renewable methane
165
Since CO is an undesired product, even at trace levels, the temperature
and pressure were varied to minimize its formation and their effects are
discussed in sections 5.3.1.3 and 5.3.1.4, respectively.
5.3.1.3 Effect of temperature
To assess the effect of the temperature on CO2 capture and conversion,
the process was tested at 350 ºC (cycles 6-10) and 300 ºC (cycles 11-15) – cf.
Table 5.2. In previous runs of CO2 sorption and regeneration with N2 (cycles 0-
5) it was observed that the shape of temperature histories profile had changed
after cycle 0 and remained similar until cycle 6 as shown in Fig. 5.10 (related
with the decrease of sorbent working capacity – Fig. 5.3).
However, the sorbent capacity was recovered in cycle 7 (Fig. 5.5) because
the bed undergone reactive regeneration with H2 for the first time in cycle 6 and
steam allowed to recover the initial bed capacity, as mentioned in the previous
section. The same profile shape was obtained in cycles performed at 300 ºC (i.e.
cycles 11-15 – cf. Fig. C.0.2 in Appendix C), although with different intensity
because both sorption and reaction kinetics are less favoured at 300 ºC. In these
cycles the temperature peak related to the WGS reaction (i.e. causing CO2 roll-
up) was also observed but in a smaller extent (Fig. 5.9b and Fig. 5.9c).
Fig. 5.5 shows the amount of sorbed and desorbed CO2 from cycles 0 to
15, besides the CO2 conversion obtained in cycles 6-15. The bed reached a
steady capacity in cycle 5 (regeneration with N2) and in the subsequent cycles
the quantity of desorbed CO2 (black bars) measured at the outlet of the bed
considerably decreases, which is related to CO2 consumption to produce CH4.
Therefore, the black bars in cycles 6-15 correspond to CO2 that was desorbed
and unconverted, besides, as mentioned previously, to the small quantity of CO2
that was produced from the WGS reaction, particularly in the later stage of the
reactor length.
Chapter 5
166
Fig. 5.10 – Normalized temperature histories in four bed positions during CO2 sorption and regeneration with N2 (cycles 0 to 5) or H2 (cycle 6); Tz,0 = 350 º C.
The sorption capacity increases after reactive regeneration experiments
(cycles 0-6 vs. cycles 7-15) and decreases with temperature decreasing (cycles
7-10 vs. cycles 11-15), as shown in Fig. 5.5. CO2 conversion slightly increases
in the first cycles at 300 ºC but then reaches a value close to the same obtained
in the end of the experimental campaign at 350 ºC (i.e. in cycle 10), ~ 91 %.
Tz -
Tz,0
/
º C
0
2
4
6
8
10
12
14
16
Tz -
Tz,0
/
º C
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
tS / s
0 500 1000 1500 2000
Tz -
Tz,0
/
º C
0
2
4
6
8
10
12
14
16
tR / s
0 500 1000 1500 2000
Tz -
Tz,0
/
º C
0
5
10
15
20
25
30
Cycle 5 - Sorption Cycle 5 - Regeneration (N2)
Cycle 6 - Sorption Cycle 6 - Regeneration (H2)
Tz -
Tz,0
/
º C
0
2
4
6
8
10
12
14
16
Tz -
Tz,0
/
º C
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
Cycle 0 - Sorption Cycle 0 - Regeneration (N2)
A sorptive reactor for CO2 capture and conversion to renewable methane
167
However, when analysing the effect of temperature on conversion it is
noteworthy to mention that a sorptive reactor is different from a conventional
reactor, where the temperature mainly influences the reaction equilibrium and
reaction kinetics. For a sorptive reactor, the temperature will also influence
sorption equilibrium and sorption kinetics, namely: i) the amount of the sorbate
previously retained in the bed, ii) the strength of the bonding between the sorbate
and the sorbent and iii) the rate in which the sorbate is transferred from the
sorbent to the gas phase and becomes available for reaction at the catalyst
active sites. In this regard, although the conversion is practically the same after
5 cycles at 350 ºC and 300 ºC, the quantity of methane produced is lower at 300
ºC (12.8 vs. 10.5 mmol per cycle), corresponding to a decrease of the maximum
methane concentration at the reactor’s outlet from 78 % to 56 % (cf. Fig. C.0.1
in Appendix C – cycles 10 and 15, respectively). Nevertheless, decreasing the
temperature also caused a remarkable drop of produced CO, as shown in Fig.
5.11.
Fig. 5.11 – Maximum CO fraction obtained during cycles at 350 ºC (6-10) and 300 ºC at 1.34 bar (11-15) and 2.0-2.5 bar (16-17).
cycle
5 6 7 8 9 10 11 12 13 14 15 16 17 18
yC
O,
má
x. x 1
00
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
nC
Opro
duced
/ m
mol
0.00
0.02
0.04
0.06
0.08
0.10 yCO,máx @ 350 ºC
yCO,máx @ 300 ºC
nCO produced
Chapter 5
168
The number of moles of CO produced decreased from 0.048 to 0.009
mmol in cycles 10 and 15, respectively. The quantity of captured CO2 in cycles
10 (350 ºC) and 11 (300 ºC) was almost the same (11.9 vs. 11.4 mmol,
respectively – Fig. 5.5), but CO formation decreased considerably. Again, this
suggests that CO is produced through the RWGS reaction, which due to its
endothermic nature is not favoured at lower temperatures; moreover, the RWGS
reaction kinetics also decreases in such conditions. The noticeable CO decline
observed from cycle 6 to 7 can be attributed to the WGS reaction, which
consumed CO to produce CO2, as discussed in the previous section.
5.3.1.4 Effect of pressure
During the sorption step, both the CO2 fraction in the feed and total
pressure were kept constant along the experimental campaign (2CO
iny =0.15 and
P=1.34 bar), except in cycles 16 and 17 where the total pressure was increased
up to ca. 2.00 and 2.50 bar, respectively (cf. Table 5.2). The obtained values for
the bed sorption capacity (2COq ) in the sorption stage were compared to those
predicted by the dual-site Langmuir sorption equilibrium equation,2
*COq ,
proposed by Oliveira et al. [15] (Eq. (5.12)).
2 2
2 2 2
2 2
0 0
* max, P max, CCO CO CO
0 01 1
sorb sorb
sorb sorb
H E
P CRT RTCO CO
H E
P CRT RTCO CO
k e p k e p
q q q
k e p k e p
(5.12)
The used model parameters can be found in Table 5.3.
A sorptive reactor for CO2 capture and conversion to renewable methane
169
Table 5.3 – List of the dual-site Langmuir model parameters taken from Oliveira et al. [16].
Parameter Units Value
2
max, PCO
q mol·kg-1 0.423
bar-1 9.07 x 10-3
kJ·mol-1 -40.0
2
max, CCO
q mol·kg-1 0.351
bar-1 1.01 x 1012
kJ·mol-1 130.8
This model assumes that sorption takes place through a combination of
(exothermic) physical sorption (superscript P) and (endothermic) chemical
reaction (superscript C), where 2
maxCOq stands for the maximum capacity (mol·kg-
1), 0k is the pre-exponential factor (bar-1),
sorbH is the sorption enthalpy
(kJ·mol-1) and sorbE the heat of reaction (kJ·mol-1). R and T stand, as usual, for
the ideal gas constant and absolute temperature, respectively.
Fig. 5.12 shows the adherence of the experimental data to Eq. (5.12)
without fitting any parameter. For cycle 0 a relative deviation between
experimental and predicted (continuous line) sorption capacity was 10 %. The
figure shows that in the first 5 cycles the experimental sorption capacity drops to
the dotted line corresponding to the physical adsorption contribution. This can
be attributed to the absence of steam in the feed stream, since in the work of
Oliveira et al. [15] the isotherms were determined using wet streams, which is
known to promote CO2 sorption [14-16]. In fact, this effect is evidenced in cycle
7 and succeeding cycles, which were performed under reactive regeneration
conditions, where steam is produced during methanation. In cycle 7 the sorption
capacity reached the predicted value of the dual-site Langmuir model
(continuous line), but dropped slightly in cycles 8-10. This difference might be
related to the absence of the WGS reaction in cycle 6 where all the steam
produced was available for sorbent regeneration, whereas in the regeneration
0Pk
sorbH
0Ck
sorbE
Chapter 5
170
stage of cycle 7 a part of the H2O produced is consumed (thus the CO2 roll-up)
and negatively influences the capacity of cycle 8; the same occurs in the
subsequent cycles.
Fig. 5.12 – Adherence of the experimental data to the dual-site Langmuir sorption
equilibrium isotherm. Model parameters taken from Oliveira et al. [15].
Cycle 11 was the first being performed at 300 ºC and here the sorption
capacity slightly exceeded the value predicted by the model (deviation of 7 %
only). In the subsequent cycles the values are lower and the decrease between
cycles is less noticeable. Since regeneration in cycle 10 was performed at 350
ºC, the sorbent had more active sites available during sorption in cycle 11 than
in the subsequent cycles. Fig. 5.12 shows also that the relative relevance of CO2
sorption via chemical reaction (dashed line) to the total capacity (continuous line)
is lower at 300 ºC than at 350 ºC. On the other side, the weight of physical
sorption (dotted line) on the total sorption capacity predicted by the model
increases from 58 % at 350 ºC to 84 % at 300 ºC (cycles 6-10 vs. cycles 11-15).
Therefore, it is expected that the enhancement of CO2 desorption using heat
released during reactive regeneration is more relevant at 300 ºC, where physical
sorption is the dominant process, whereas at 350 ºC both physical (exothermic)
A sorptive reactor for CO2 capture and conversion to renewable methane
171
and chemical (endothermic) contributions have a similar weight on the total
sorption capacity (58 % vs. 42 %, respectively).
The sorption capacity increased with total pressure from 0.38 mmol·g-1
(average value at 1.34 bar total pressure – cycles 12-15) to 0.43 mmol·g-1 (at
2.5 bar total pressure – cycle 17). CO2 conversion decreased slightly, from 92.5
% to 90.8 %, but the quantity of CH4 produced remained practically constant.
Therefore, the most important consequence of increasing the pressure was
decreasing CO production, as shown in Fig. 5.11. The quantity of CO formed
declined 34 % when the pressure increased from 1.34 bar to 2.50 bar.
5.3.2 Sorptive reactor under continuous operation mode
As mentioned before, in Section 5.3.1 the sorptive reactor was studied
considering full saturation of the bed before starting the regeneration step.
However, to implement the process on a continuous basis without emitting and
wasting CO2, the sorption step should end before the CO2 breakthrough time.
For this end, at least two sorptive reactors should be operated in parallel, i.e.,
while one is undergoing the capture stage, the other should be regenerated and
producing CH4. This means that under continuous operation mode the bed
sorption capacity (also-called dynamic capacity) is lower than its maximum value
(full bed saturation), which obviously has implications on the global performance
of the process. Therefore, the process performance was assessed through CH4
productivity and purity (please refer to Eqs. (5.8) and (5.9)).
Fig. 5.13 shows the outlet concentration profile of all the species (dry
basis) during the reactive regeneration stage at 350 ºC (cycle 21) and 300 ºC
(cycle 24). CO was not detected at 300 ºC (see insertion of Fig. 5.13b) and CO2
conversion was almost complete (ca. 99 %) at both temperatures since only a
small quantity of unconverted CO2 was noticed at the outlet.
Chapter 5
172
Fig. 5.13 – Outlet concentration profiles of all species during the regeneration step (dry basis) at a) 350 ºC and b) 300 ºC. Insertion shows CO2 and CO profile during the initial 500 s.
Fig. 5.14 shows the evolution of both process indicators, CH4 purity and
productivity, along the reactive regeneration time at 350 ºC and 300 ºC. It can
be observed that temperature improved CH4 purity, while the productivity was
nearly the same at both temperatures (maximum value of ca. 2.32-2.36 mol·kg-
1·h-1) because the decrease of the produced methane at 300 ºC was
counterbalanced by the decrease of the cycle time. Fig. 5.14 also shows that
productivity is still increasing when purity reaches its maximum, which in turn
takes place after H2 breaks out of the unit (please see Fig. 5.13 and Fig. 5.14).
According to the German Technical and Scientific Association for Gas and
Water, the maximum allowable concentration of H2 in the gas mixture is 5 vol. %
[6]. In this regard, if the regeneration is stopped when H2 purity reaches this
value (ca. 280 s), at 350 ºC the purity is close to the maximum value (26 % vs.
29 %), while productivity is far below (1.31 mol·kg-1·h-1 vs. 2.32 mol·kg-1·h-1).
tR / s
0 200 400 600 800 1000 1200 1400 1600
yj
0.0
0.2
0.4
0.6
0.8
1.0
1.2
tR
/ s
0 100 200 300 400 500
yj
0.000
0.001
0.002
0.003
0.004
0.005
tR / s
0 200 400 600 800 1000 1200 1400 1600
N2
CH4
H2
CO2
CO
tR
/ s
0 100 200 300 400 500
yj
0.000
0.001
0.002
0.003
0.004
0.005
Cycle 21 - 350 ºC Cycle 24 - 300 ºCa) b)
A sorptive reactor for CO2 capture and conversion to renewable methane
173
Fig. 5.14 – CH4 purity (circles) and productivity (squares) evolution during regeneration time at 350 ºC (full symbols) - cycle 21 - and 300 ºC (open symbols) – cycle 24.
Nitrogen is the species that breaks first during the regeneration step (as
shown in Fig. 5.13) and it is possible to purge ca. 30-35 % of the total amount
before methane breaks. Nitrogen purge increases methane purity, as shown in
Fig. 5.15. For instance, at 350 ºC, when the H2 purity reaches 5 %, the CH4 purity
increases from 26 % to 35 % (Fig. 5.15a).
Fig. 5.15 – Effect of N2 purge on CH4 purity evolution during regeneration time at a) 350 ºC and b) 300 ºC.
tR / s
0 200 400 600 800 1000 1200 1400 1600
Pu
r /
%
0
10
20
30
40
50
0.0
0.5
1.0
1.5
2.0
2.5
Pro
d /
mo
l kg
-1 h
-1
tR / s
0 200 400 600 800 1000 1200 1400 1600
Pu
r /
%
0
10
20
30
40
N2 purge (YES)
N2 purge (NO)
tR / s
0 200 400 600 800 1000 1200 1400 1600
Pu
r / %
0
10
20
30
40
N2 purge (YES)
N2 purge (NO)
Cycle 21 - 350 ºC Cycle 24 - 300 ºCa) b)
Chapter 5
174
5.3.3 Reactor design considerations
The results obtained when the sorptive reactor was operated under
discontinuous (Section 5.3.1) or continuous mode (Section 5.3.2) provided
insight about the operations conditions, particularly temperature and pressure,
which could be elected for further optimization of the reactor.
Generally, CO formation through parallel RWGS and low CH4 purity level
(affected by all species leaving the reactor and not only by CO) are the main
drawbacks that need to be tackled from the reactor design point of view.
Controlling the temperature and pressure proved their suitability to surpass
these limitations, but not entirely.
CH4 integral selectivity (i.e. nCH4/nCO ratio) obtained in discontinuous
operation mode improves by a factor of 4.1 when the operation temperature was
decreased from 350 ºC to 300 ºC. For this reason, the temperature was kept at
300 ºC in subsequent cycles while the total pressure was varied (cycles 16 and
17). Increasing the total pressure from 1.34 bar to 2.50 bar enhanced the CO2
sorption capacity (by 13 %) and CH4 integral selectivity (by 50 %). However, in
practical application, flue gas is delivered at atmospheric pressure and, if
possible, its compression should be avoided for economic reasons. In this
regard, the pressure was kept at 1.34 bar when the reactor was operated in
continuous mode (cycles 19-24) and CO was not detected in cycles performed
at 300 ºC. During reactive regeneration of these cycles, the amount of H2
available in the gas phase per mole of desorbed CO2 presumably increased
when compared to discontinuous operation, favouring the Sabatier reaction in
detriment of the RWGS (H2/CO2=4 vs. H2/CO2=1 stoichiometric ratios,
respectively). Nevertheless, the maximum CH4 purity level reached at the
reactor outlet is still unsatisfactory (30 %). It is anticipated that further
optimization of CH4 purity and productivity can be achieved by tuning the bed
properties (e.g. catalyst/sorbent ratio, bed layout, H2 feed mode, etc.). However,
A sorptive reactor for CO2 capture and conversion to renewable methane
175
special care regarding the influence of other species present in the flue gas on
reactor performance should be also systematically evaluated in future work.
5.4 Conclusions
A sorptive reactor was successfully piloted at lab-scale to capture and
convert CO2 into CH4. The process was studied in a temperature window
between 300-350 ºC based on the typical range reported for both CO2 sorption
and methanation using K-promoted hydrotalcites and Ni-based catalysts,
respectively. The process was studied in a low-pressure range (≤ 2.5 bar)
because post-combustion streams are generally at the atmospheric pressure.
This type of reactor allows that both processes (i.e. sorption and reaction)
take place in the same unit and has the economic advantage that it does not
require CO2 transportation from a capture unit to a (methanation) reactor nor to
a storage site. Process intensification allowed enhancing CO2 capture due to
steam formed in situ during the reactive regeneration step of the previous cycle.
The dual-site Langmuir model could well describe the sorption equilibrium.
Reactive regeneration resulted in a synergetic effect taking place along the
reactor length: CH4 formation (desired product) is favoured by the presence of
H2O, which in turn participates on CO2 desorption, making it faster and removing
more CO2 than under dry conditions. Consequently, the sorption capacity in the
subsequent cycle is enhanced until a pseudo steady-state is reached.
Furthermore, since steam may be sorbed by the hydrotalcite, the reaction
equilibrium can be shifted to the right side to produce more CH4 (Eq. (5.1)). CO2
conversion was high in batch operation (ca. 92 %) and almost complete (99 %)
when the reactor was operated under continuous mode.
Although methane concentration in the outlet stream can be as high as 78
% (dry basis), the maximum CH4 purity obtained at 350 ºC was 37 % if N2 is
purged at the beginning of the regeneration step.
Chapter 5
176
The temperature effect on productivity was found to be negligible in the
range studied (maximum value of 2.32 vs. 2.36 mol·kg-1·h-1 at 350 ºC and
300 ºC, respectively). The undesired formation of CO through the RWGS
reaction was observed and could be minimized by decreasing the temperature
and/or increasing the operating total pressure.
The materials employed in this work proved their compatibility and activity
for this application, besides showing good stability under cyclic operation (which
also allowed to assess repeatability). In this regard, the envisaged process
scale-up would benefit by the fact that both materials are already at a
commercial level. Still, the process conditions and reactor design (e.g. bed
layout) should be further studied to improve methane purity up to the level
required for injection in the natural gas grid.
A sorptive reactor for CO2 capture and conversion to renewable methane
177
5.5 Nomenclature
Parameter Description Units
C Carbon molar balance %
sorbE Heat of chemical reaction of CO2 with the sorbent
kJ·mol-1
2COF CO2 molar flow-rate mol·s-1
F Total molar flow-rate mol·s-1
sorbH Heat of physical sorption kJ·mol-1
0Ck
Arrhenius pre-exponential factor of CO2 sorption equilibrium constant through chemical reaction
bar-1
0Pk
Arrhenius pre-exponential factor of CO2 sorption equilibrium constant through physical adsorption
bar-1
L Sorptive reactor length cm
adsm Sorbent mass kg
catm Catalyst mass kg
2
sorbedCOn Moles of sorbed CO2 mol
2
desorbed/unconvertedCOn Moles of desorbed and unconverted CO2 mol
producedin
Moles of carbon-containing produced species “i”
mol
2COp CO2 partial pressure bar
Pout Total pressure at the outlet of the sorptive reactor
bar
Prod CH4 productivity mol·kg-1·h-1
Pur CH4 purity %
2COq CO2 sorption capacity (experimental) mol·kg-1
2
*COq
CO2 sorption capacity (dual-site Langmuir model)
mol·kg-1
2
max, CCO
q Maximum CO2 sorption capacity for endothermic chemical reaction
mol·kg-1
2
max, PCO
q Maximum CO2 sorption capacity for exothermic physical adsorption
mol·kg-1
R Ideal gas constant (8.314 x 10-3) kJ·mol-1·K-1
T Temperature in the sorptive reactor K
Rt Time during regeneration step s
,R ft Duration of regeneration step s
Chapter 5
178
St Time during sorption step s
,S ft Duration of sorption step s
2COX CO2 conversion %
2COy CO2 molar fraction
Subscripts & Superscripts
i Carbon-containing species formed: CH4 and CO
in Inlet of the sorptive reactor
j Species N2, H2, CO2, CH4 and CO
k Sorption-reaction cycle
out Outlet of the sorptive reactor
R Regeneration step
S Sorption step
z Axial position of the sorptive reactor (1-4)
A sorptive reactor for CO2 capture and conversion to renewable methane
179
5.6 References
[1] E.A. Quadrelli, G. Centi, J.L. Duplan, S. Perathoner, Carbon dioxide
recycling: Emerging large-scale technologies with industrial potential,
ChemSusChem, 4 (2011) 1194-1215.
[2] M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner, P. Markewitz, T.E. Müller,
Chemical technologies for exploiting and recycling carbon dioxide into the value
chain, ChemSusChem, 4 (2011) 1216-1240.
[3] S. Kitasei, Powering the low-carbon economy: the once and future roles of
renewable energy and natural gas, Worldwatch Institute, Washington D.C.,
2010.
[4] E.A. Quadrelli, G. Centi, Green Carbon Dioxide, ChemSusChem, 4 (2011)
1179-1181.
[5] M. Aresta, A. Dibenedetto, A. Angelini, Catalysis for the valorization of
exhaust carbon: from CO2 to chemicals, materials, and fuels. Technological use
of CO2, Chemical Reviews, 114 (2014) 1709-1742.
[6] S. Schiebahn, T. Grube, M. Robinius, V. Tietze, B. Kumar, D. Stolten, Power
to gas: technological overview, systems analysis and economic assessment for
a case study in Germany, International Journal of Hydrogen Energy, 40 (2015)
4285-4294.
[7] S. Walspurger, W.G. Haije, B. Louis, CO2 reduction to substitute natural gas:
toward a global low carbon energy system, Israel Journal of Chemistry, 54
(2014) 1432-1442.
[8] J.G. Speight, Natural gas: A Basic Handbook, Gulf Publishing Company,
Houston, 2007.
[9] N. Mellquist, M. Fulton, Natural Gas and Renewables: A Secure Low Carbon
Future Energy Plan for the United States Deutsche Bank Group, 2010.
[10] K. Müller, M. Fleige, F. Rachow, D. Schmeißer, Sabatier based CO2-
methanation of flue gas emitted by conventional power plants, Energy Procedia,
40 (2013) 240-248.
Chapter 5
180
[11] A. Borgschulte, N. Gallandat, B. Probst, R. Suter, E. Callini, D. Ferri, Y.
Arroyo, R. Erni, H. Geerlings, A. Zuttel, Sorption enhanced CO2 methanation,
Physical Chemistry Chemical Physics, 15 (2013) 9620-9625.
[12] S. Walspurger, G.D. Elzinga, J.W. Dijkstra, M. Sarić, W.G. Haije, Sorption
enhanced methanation for substitute natural gas production: experimental
results and thermodynamic considerations, Chemical Engineering Journal, 242
(2014) 379-386.
[13] Q. Zheng, R. Farrauto, A. Chau Nguyen, Adsorption and methanation of
flue gas CO2 with dual functional catalytic materials: a parametric study,
Industrial & Engineering Chemistry Research, 55 (2016) 6768-6776.
[14] M.K. Ram Reddy, Z.P. Xu, J.C. Diniz da Costa, Influence of water on high-
temperature CO2 capture using layered double hydroxide derivatives, Industrial
& Engineering Chemistry Research, 47 (2008) 2630-2635.
[15] E.L.G. Oliveira, C.A. Grande, A.E. Rodrigues, CO2 sorption on hydrotalcite
and alkali-modified (K and Cs) hydrotalcites at high temperatures, Separation
and Purification Technology, 62 (2008) 137-147.
[16] M.A. Soria, S. Tosti, A. Mendes, L.M. Madeira, Enhancing the low
temperature water-gas shift reaction through a hybrid sorption-enhanced
membrane reactor for high-purity hydrogen production, Fuel, 159 (2015) 854-
863.
[17] Q. Wang, J. Luo, Z. Zhong, A. Borgna, CO2 capture by solid adsorbents and
their applications: current status and new trends, Energy and Environmental
Science, 4 (2011) 42-55.
[18] S. Choi, J.H. Drese, C.W. Jones, Adsorbent materials for carbon dioxide
capture from large anthropogenic point sources, ChemSusChem, 2 (2009) 796-
854.
[19] M.K.R. Reddy, Z.P. Xu, G.Q. Lu, J.C. Diniz da Costa, Effect of SOx
adsorption on layered double hydroxides for CO2 capture, Industrial &
Engineering Chemistry Research, 47 (2008) 7357-7360.
A sorptive reactor for CO2 capture and conversion to renewable methane
181
[20] M.G. Beaver, High temperature CO2 chemisorbents: applications,
characterization, and study of the chemical nature of chemisorbent surfaces, in,
Lehigh University, 2010.
[21] E.Z. Golosman, V.N. Efremov, Industrial catalysts for the hydrogenation of
carbon oxides, Catalysis in Industry, 4 (2012) 267-283.
[22] W. Wang, S. Wang, X. Ma, J. Gong, Recent advances in catalytic
hydrogenation of carbon dioxide, Chemical Society Reviews, 40 (2011) 3703-
3727.
[23] H.F. Rase, Handbook of Commercial Catalysts: Heterogeneous Catalysts,
CRC Press, Boca Raton, , 2000.
[24] A. Gil, A. Diaz, M. Montes, Passivation and reactivation of nickel catalysts,
Journal of the Chemical Society, Faraday Transactions, 87 (1991) 791-795.
[25] C.V. Miguel, R. Trujillano, V. Rives, M.A. Vicente, A.F.P. Ferreira, A.E.
Rodrigues, A. Mendes, L.M. Madeira, High temperature CO2 sorption with
gallium-substituted and promoted hydrotalcites, Separation and Purification
Technology, 127 (2014) 202-211.
[26] H. Du, C.T. Williams, A.D. Ebner, J.A. Ritter, In situ FTIR spectroscopic
analysis of carbonate transformations during adsorption and desorption of CO2
in K-promoted HTlc, Chemistry of Materials, 22 (2010) 3519-3526.
[27] J. Boon, P.D. Cobden, H.A.J. van Dijk, M. van Sint Annaland, High-
temperature pressure swing adsorption cycle design for sorption-enhanced
water-gas shift, Chemical Engineering Science, 122 (2015) 219-231.
[28] J. Boon, P.D. Cobden, H.A.J. van Dijk, C. Hoogland, E.R. van Selow, M.
van Sint Annaland, Isotherm model for high-temperature, high-pressure
adsorption of CO2 and H2O on K-promoted hydrotalcite, Chemical Engineering
Journal, 248 (2014) 406-414.
[29] G. Li, P. Xiao, D. Xu, P.A. Webley, Dual mode roll-up effect in
multicomponent non-isothermal adsorption processes with multilayered bed
packing, Chemical Engineering Science, 66 (2011) 1825-1834.
Chapter 5
182
[30] A.F.P. Ferreira, A.M. Ribeiro, S. Kulaç, A.E. Rodrigues, Methane
purification by adsorptive processes on MIL-53(Al), Chemical Engineering
Science, 124 (2015) 79-95.
[31] C.V. Miguel, M.A. Soria, A. Mendes, L.M. Madeira, Direct CO2
hydrogenation to methane or methanol from post-combustion exhaust streams
- A thermodynamic study, Journal of Natural Gas Science and Engineering, 22
(2015) 1-8.
[32] J. Gao, Y. Wang, Y. Ping, D. Hu, G. Xu, F. Gu, F. Su, A thermodynamic
analysis of methanation reactions of carbon oxides for the production of
synthetic natural gas, RSC Advances, 2 (2012) 2358-2368.
[33] J. Xu, G.F. Froment, Methane steam reforming, methanation and water-gas
shift: I. Intrinsic kinetics, AIChE Journal, 35 (1989) 88-96.
[34] K. Liu, C. Song, V. Subramani, Hydrogen and syngas production and
purification technologies, Wiley; AIChE, New York, 2010.
[35] M. Maroño, Y. Torreiro, D. Cillero, J.M. Sánchez, Experimental studies of
CO2 capture by a hybrid catalyst/adsorbent system applicable to IGCC
processes, Applied Thermal Engineering, 74 (2015) 28-35.
183
Chapter 6. Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
The contents of this chapter were adapted from: C.V. Miguel, A. Mendes, S. Tosti, L.M.
Madeira, Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes.
International Journal of Hydrogen Energy, 2012, 37(17), 12680-12687.
The permeance of a 50 µm thick PdAg (25 wt.% of silver) tubular
membrane towards hydrogen was evaluated for hydrogen feed streams or
binary mixtures of industrial relevance, namely of H2/CO or H2/CO2.
It was proposed a rearrangement of the Sieverts-Langmuir (SL)
equation to account for the inhibition effect on the membrane permeance
towards H2 due to the presence of CO or CO2. Such modification is based
on the average partial pressure of CO or CO2 and the logarithm-mean
driving-force to account for concentration gradients along the axial length
of the tubular membrane. Comparison with the original SL equation
evidenced relative deviations < 10.7 %. The experimental data showed
good agreement with the SL model.
It was found that CO has a much stronger inhibition effect on H2
permeation than CO2. Considering the values obtained for the adsorption
enthalpy of each gas, it was concluded that both CO and CO2 physically
adsorb on the membrane surface.
Chapter 6
184
6.1 Introduction
Most of the hydrogen production worldwide is based on the conversion of
fossil fuels, being steam reforming of methane (SRM) (Eq. (6.1)) in combination
with water-gas shift (WGS) (Eq. (6.2)) reactions the most commercialized
process [1].
-14 2 2 298 KCH + H O CO + 3H Δ = 205.9 kJ molH (6.1)
-12 2 2 298 KCO + H O CO + H Δ = - 41.1 kJ molH (6.2)
Recently, hydrogen production through water electrolysis (Eq. (6.3)) has
gained a renewed interest as an important building block in Power-to-Gas (PtG)
applications (as addressed in Chapter 2 and Chapter 5).
-12 2 2 298 KH O H +1 2O Δ =285.8 kJ molH (6.3)
In PtG, exceeding renewable power is chemically stored as hydrogen, or
another energy carrier such as methane or methanol, which are more easily
stored and transported. The hydrogen obtained from water electrolysis is further
used in hydrogenation reactions of a carbon source, such as CO2, to produce
methane (Eq. (6.4)) or methanol (Eq. (6.5)).
-12 2 4 2 298 KCO + 4H CH + 2H O Δ = - 165 kJ molH (6.4)
-12 2 3 2 298 KCO + 3H CH OH + H O Δ = - 49.4 kJ molH (6.5)
The importance of keeping an H2/CO2 stoichiometric ratio of 4 to reach
high CO2 conversion and CH4 selectivity was observed in the methane pathway
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
185
(Eq. (6.4)), from the thermodynamic point of view, in Chapter 2. In Chapter 5,
when the sorptive reactor was presented and tested, it was observed
experimentally the formation of carbon monoxide through the reverse water-gas
shift (inverse reaction of Eq. (6.2)). One possible explanation may be the
inexistence of an ideal H2/CO2 ratio of 4 along the layered bed length. It was
suggested that the hydrogen feed mode could be an option to minimize this
effect, for instance, using an H2 permselective material for the reactor wall,
allowing to feed the sorptive reactor along its length with the desired H2 flux so
that the ideal ratio is verified in every axial position of the reactor. The option for
a membrane feeding system would benefit of high pressure hydrogen delivery
from the electrolyser, which commonly can reach up to 30 bar.
The hydrogen-palladium system has been extensively studied along the
years (see for instance [2, 3]), and it is commonly accepted that Pd-based
materials are the most suitable for H2 separation/delivery with membranes,
despite its price. Palladium membranes are known to exhibit high H2 selectivity
and permeability and when alloyed with other metals (e.g. Cu, Ag, etc.) can lead
to materials less susceptible to hydrogen embrittlement and poisoning [4].
Additionally, long term tests, namely thermal and hydrogenation cycles, also
evidenced that self-supported dense and thin wall PdAg23 wt.% tubular
membranes with a “finger-like” configuration have high durability and reliability
(complete selectivity remained constant for at least one year) [5]. The use of this
“finger-like” configuration allows the free elongation and contraction of the
membrane, thus avoiding any mechanical stress that may damage the thin wall
tube. Besides, this kind of membranes allows obtaining ultrapure hydrogen due
to their nearly infinite permselectivity towards H2 [5, 6]. This is particularly
important when using membranes for selective removal of hydrogen, a product,
in thermodynamic equilibrium-limited reactions such as SRM (Eq. (6.1)) and
WGS (Eq. (6.2)), enabling to overcome the equilibrium conversion, which is the
limit of a correspondent conventional fixed bed reactor operated at the same
Chapter 6
186
conditions of the membrane reactor. A pure/ultrapure hydrogen stream can be
obtained, exempting further separation processes [7, 8].
However, for the implementation of these membranes for the above-
mentioned applications (H2 feed or removal), it is crucial to measure permeability
under different conditions, partial pressures, temperatures, besides assessing
the effect on the permeance caused by species such as CO or CO2 which are
present in reactions given by Eqs. (6.1)-(6.2) and Eqs. (6.4)-(6.5). This topic has
been reported by Gallucci et al. [9], whom measured the permeance of a 60 µm
thick finger like membrane towards H2 in binary mixtures of H2 and CO or CO2.
The authors observed that CO2 presence has no significant inhibition effect (at
523 K), whereas CO presence greatly reduced the H2 permeating flux. Other
authors found that inhibition by methane and steam was negligible as compared
to CO and CO2 [10, 11].
In the present chapter, the experimental tests and the development of an
accurate model that describe the effect of CO and CO2 on the hydrogen
permeability is described. The model has been validated through the
experimental results. In this work a rearranged Sieverts-Langmuir equation is
proposed for the first time to account for the change of the permeability driving
force along the membrane length caused by H2 removal in binary mixtures
containing CO or CO2.
6.2 Theoretical background
Generally, it is considered that Pd-based (defect-free) membranes
permeability to hydrogen follows a solution-diffusion mechanism [3, 12] and the
permeating hydrogen flux, 2HJ , can be described by [12, 13]:
2
2 2 2
H n nH H , feed H , permeate
LJ p p
(6.6)
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
187
where 2HL is the palladium alloy permeability to hydrogen, is the membrane
thickness, 2
nH , feedp and
2
nH , permeatep are the hydrogen partial pressures in the feed
and permeate sides, respectively, and n is the pressure exponent that varies
between 0.5 and 1.0. If diffusion of atomic hydrogen through the metal lattice of
the membrane is the limiting step, the flux can be expressed by the Sieverts Law
[2]:
2
2 2 2
H 1 2 1 2SievH H , feed H , permeate
LJ p p
(6.7)
However, the presence of surface adsorbed chemical species, such as CO
or CO2 that are present in the water gas shift or reforming reactors, may lead to
deviations to the Sieverts law. In this regard, Barbieri et al. [14] proposed a new
equation where the Sieverts law and the Langmuir isotherm were coupled to
describe the decrease on H2 permeation in the presence of CO as follows:
2 2
SL SievCO COH H
CO CO
11
K pJ J
K p
(6.8)
Where 2
SLHJ is the permeated hydrogen molar flux in the presence of CO,
2
SievHJ
is the hydrogen flux assessed accordingly to the Sieverts law (Eq. (6.7)),COp is
the carbon monoxide partial pressure, COK is the adsorption equilibrium
constant, and CO CO
CO CO1
K pCF
K p
is a dimensionless correction factor due to the
adsorption of CO on the membrane surface that depends on CO COCO
CO CO1
K p
K p
,
which stands for the fraction of the membrane surface covered by CO. Finally,
is a parameter that depends only on the temperature and accounts for
Chapter 6
188
additional effects of the adsorbed gas, for instance in the dissociation of H2
molecules. Before, Eq. (6.8) has been reported with reference to CO [14]: similar
considerations and a similar expression is herein proposed for the first time and
written for CO2.
However, since hydrogen is being removed, the driving-force for
permeability changes along the membrane length and consequently so the CO
(or CO2) partial pressure. In this regard, the SL equation (Eq. (6.8)) should be
rearranged as follows
2
2
HSL i iH ln
i i
11
LK pJ P
K p
(6.9)
where ip is the average partial pressure of species i (CO, in this study extended
also to CO2) between the feed and retentate sides, and lnP is the logarithmic
mean driving-force (calculated based on heat-exchanger theory for parallel flow
[15]), expressed as:
2 2 2 2
2 2
2 2
1 2 1 2 1 2 1 2H , feed H , permeate H , retentate H , permeate
ln 1 2 1 2H , feed H , permeate
1 2 1 2H , retentate H , permeate
ln
p p p pP
p p
p p
(6.10)
In this work, it was studied the inhibition effect caused by CO and CO2 on
a 50 µm thick PdAg25 wt.% membrane permeability to H2 in the temperature
range of 473-573 K. The Sieverts-Langmuir model was used to describe not only
H2/CO but also H2/CO2 permeation experiments. Furthermore, Eq. (6.10) that
accounts for the driving-force change along the membrane length was
introduced by improving significantly the accuracy of the analysis. Finally, the
adsorption enthalpies for both CO and CO2 adsorption were determined.
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
189
6.3 Experimental
6.3.1 Membrane preparation and experimental set up
The membrane used in this work was prepared by diffusion bonding of
cold-rolled Pd-Ag foils, providing a thin wall tubular membrane with a finger-like
configuration (cf. Fig. 6.1).
Fig. 6.1 – Sketch of the module assembly and the cold rolled PdAg25 wt.% membrane with a finger-like configuration inside.
In the preparation, it was used a commercial 50 µm thick PdAg25 wt.% foil
(from Johnson Matthey); detailed information regarding this technique can be
found elsewhere [16]. The permeation module (Fig. 6.1) was placed inside an
electrical oven (Memmert, model UNE200).
The gases (H2, N2, CO and CO2) were fed into the membrane lumen using
four mass controllers (Bronkhorst High-Tech, model F201) and preheated inside
the oven before entering the membrane module; permeate and retentate stream
flow rates were measured with mass flow meters (Bronkhorst High-Tech, model
F201).
Pressure transducers (Druck, ref. 4010) were used to measure the
pressure in the feed and permeate sides; the pressure in the feed side was set
using a back-pressure regulator (Swagelok).
Chapter 6
190
6.3.2 Gas permeation experiments
Before the experiments with pure hydrogen, the palladium/silver
membrane was firstly activated by reduction with hydrogen following a two-step
protocol: i) while the temperature increased until 573 K, nitrogen was fed to the
feed and permeate sides; ii) after reaching the end temperature, H2 was
introduced in the feed side (H2:N2 molar ratio of 1:1) with a total flow rate of 90
mLNmin-1. The activation step ended when maximum and stable H2 permeating
flux was measured, i.e., when steady state conditions were reached.
Pure hydrogen permeation experiments were performed at different
temperatures (473-573 K) and feed pressures (150-223 kPa) for a constant feed
flow rate of 90 mLNmin-1 of H2. Higher pressures were not used due to the
mechanical resistance limitation of the thin wall membrane (maximum operating
total pressure difference across the membrane should not exceed ca. 200 kPa).
The permeate side was kept at the atmospheric pressure in all experiments.
Additionally, permeation experiments were performed with membranes
being submitted to feed streams comprising binary mixtures of H2/CO or H2/CO2;
the total volumetric flow rate was kept constant while the molar ratio of CO/H2 or
of CO2/H2 was varied. The Sieverts driving force (Eq. (6.7)) was kept
approximately constant (at 130 Pa0.5) by adjusting the total feed pressure (up to
a maximum of 253 kPa) when changing the feed composition. These
experiments were performed at low H2 recoveries: the maximum recovery
obtained was ca. 34% in CO2/H2 experiments. The pressure drop along the
membrane tube was negligible and therefore it was assumed that the total
pressure remained constant, being equal in the feed and retentate streams.
After a set of experiments with binary mixtures, the membrane was
poisoned and consequently needing a recovery of the permeating behaviour
before a new test. A regeneration strategy was then implemented to remove the
adsorbed contaminant gases. The membrane regeneration was performed by
increasing the temperature up to 573 K and feeding synthetic air or nitrogen, to
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
191
both sides of the membrane, after experiments with CO or CO2, respectively. In
the case of treatment with synthetic air aimed at oxidizing the CO to CO2, for
safety reasons the membrane was first flushed with N2 for H2 removal.
6.4 Results and Discussion
6.4.1 Pure H2 permeance experiments
The experimental data for pure H2 permeation were fitted to the Sieverts
law, Eq. (6.7) (Fig. 6.2); in all cases R2 was found to be higher than 0.993.
Fig. 6.2 – Hydrogen molar flux through the membrane as a function of the difference between square root of H2 partial pressure in the feed and permeate sides - Eq. (6.7).
Since permeability follows an Arrhenius-type equation, the activation
energy was determined from the linear fit between 2H
L and the inverse of
temperature (c.f. Fig. 6.3).
P (Pa0.5)
0 20 40 60 80 100 120 140 160
JH
2
(m
ol m
-2 s
-1)
0.000
0.005
0.010
0.015
0.020
0.025
0.030573 K (R2=0.9954)
548 K (R2=0.9932)
523 K (R2=0.9938)
498 K (R2=0.9939)
473 K (R2=0.9936)
Chapter 6
192
Fig. 6.3 – Hydrogen permeability as a function of the inverse of temperature.
According to Yun and Oyama [12], low activation energy values (< 30
kJmol-1) indicate that diffusion is the limiting step, which is observed in the
present case (17.41 kJ∙mol-1). This value and also the pre-exponential factor
(1.16 x 10-5 mol∙m∙m-2∙s-1∙kPa-0.5) were found to be within the ranges reported in
the literature for this kind of membranes, 12.48-48.50 kJ∙mol-1 and 0.38-9.33 x
10-5 mol∙m∙m-2∙s-1∙kPa-0.5, respectively [13].
6.4.2 Effect of CO and CO2 on H2 permeance
6.4.2.1 Comparison between Sieverts-Langmuir model and
rearranged equation
The Sieverts-Langmuir equation (Eq. (6.8)) and the new herein proposed
rearranged one (Eq. (6.9)) were used to predict the membrane permeability to
hydrogen in the presence of carbon monoxide or carbon dioxide. Eq. (6.9) is in
principle more accurate, because it considers the influence of the concentration
gradients along the membrane length, making use of a logarithm-mean pressure
1/T (K-1)
0.0017 0.0018 0.0019 0.0020 0.0021 0.0022
-ln
(LH
2
)
18.4
18.6
18.8
19.0
19.2
19.4
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
193
difference. However, in practical applications and for design its use is not
straightforward, as it requires information about the exit composition.
Comparison of H2 fluxes predictions by both equations is shown in Fig. 6.4a and
Fig. 6.4b for CO and CO2, respectively.
The results of Fig. 6.4 clearly indicate an almost linear relationship
between the hydrogen fluxes calculated with Eq. (6.8) and Eq. (6.9) for
experiments in the presence of CO or CO2, in the temperature range studied.
However, Fig. 6.4b clearly indicates a systematic deviation. This trend is related
to the CO2 partial pressure that is underestimated in the case of the original
Sieverts-Langmuir equation since it is based on feed conditions. Once hydrogen
permeates through the membrane, the retentate side becomes richer in the non-
permeating species (CO2 in this case) along the tube. Therefore, the carbon
dioxide partial pressure increases, which is only considered in Eq. (6.9) – an
average partial pressure is considered for CO2, as well as a corrected driving
force for H2, based on the logarithmic mean P (Eq. (6.10)). The deviation seen
for CO2 (average deviation of 4.8 %) is almost not observed in the case of CO
(deviation < 0.9 %) because:
- the experiments were performed with low CO concentrations (max. 5 %
for CO vs. 20 % for CO2, as described below),
- the hydrogen recovery was lower for the runs with carbon monoxide (up
to 24% vs. up to 34% with carbon dioxide), thus reducing the need to correct the
gas concentrations along the membrane axis.
Chapter 6
194
Fig. 6.4 – Parity plot for comparing the H2 flux determined with the Sieverts-Langmuir model (Eq. (6.8)) with the H2 flux determined with the rearranged Sieverts-Langmuir model (Eq. (6.9)) in the presence of CO (a) or CO2 (b).
Nevertheless, Eq. (6.9) does not allow predicting H2 flux reduction due to
the presence of CO or CO2 because the hydrogen partial pressure in the
retentate side is not previously known. Therefore, using the conditions based on
the feed side (Eq. (6.8)) is more convenient for design or modelling purposes,
JH
2
SL (mol m-2 s-1), Eq. (6.9)
0.000 0.004 0.008 0.012 0.016 0.020
JH
2
SL (
mol m
-2 s
-1),
Eq.
(6.8
)
0.000
0.004
0.008
0.012
0.016
0.020573 K
548 K
523 K
498 K
a)
JH
2
SL (mol m-2 s-1), Eq. (6.9)
0.000 0.004 0.008 0.012 0.016 0.020
JH
2
SL (
mol m
-2 s
-1),
Eq.
(6.8
)
0.000
0.004
0.008
0.012
0.016
0.020573 K
548 K
523 K
498 K
473 K
b)
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
195
even if some relaxation/inaccuracy has to be admitted, which might be
compromised at higher recoveries (e.g. with thinner membranes or higher
permeation driving forces). Even so, in the conditions of Fig. 6.4, the deviation
between both equations is below 10.7% (average deviation of 4.8%).
6.4.2.2 Modelling by the Sieverts-Langmuir equation
With the operation conditions used in this work, it was observed that CO
significantly reduces H2 permeability (cf. Fig. 6.5), even at low concentrations.
Fig. 6.5 – Normalized H2 flux as a function of the CO concentration in the feed for different temperatures. The fitted function (dashed lines) based on SL model (Eq. (6.8)) is also shown.
For a CO content of only 5 %, hydrogen permeation flux decreases 57 %
at 573 K, being this effect even more remarkable at lower temperatures – at 498
K, hydrogen permeation flux is less than 18 % of that obtained in pure H2
streams, for the same driving force. Gallucci et. al [9] also reported a significant
inhibition effect of CO on the same type of membranes. They refer that such
effect is due to a combination of surface (at low CO content) and dilution effects
% (v/v) CO
0 1 2 3 4 5 6
JH
2*
0.0
0.2
0.4
0.6
0.8
1.0 573 K
548 K
523 K
498 K
Chapter 6
196
(as the CO content increases). Also, Fig. 6.5 highlights that the experimental
data show good agreement with the Sieverts-Langmuir model (SL) – Eq. (6.8),
for all conditions tested (see the fitting lines).
In the study performed by Barbieri et al. [14], it was stated that CO2 acts
as an in inert and thus H2 permeance is not affected: however, these authors
used a 60 µm thick PdAg23 %wt. membrane, 10 % of CO2 in the feed stream
and operated at higher temperature (647 K). Differently, for a thicker membrane
(250 µm) Chabot et al. [17] reported that a mixture containing 9 % of CO2 and
9.5 % of H2 balanced with He slightly affects H2 permeation (although much less
significantly than CO), particularly at temperatures below 423 K. In this work, it
was found that H2 permeance could be considerably reduced due to CO2
presence; as shown in Fig. 6.6, the hydrogen flux can be reduced by as much
as 40 %, for a CO2 content in the feed of 20 % at 473 K. In this case, however,
the effect of temperature is not as pronounced as noticed for CO (c.f. Fig. 6.5).
Furthermore, the SL model also predicted quite well the experimental data
obtained for CO2, suggesting that this compound also adsorbs on the membrane
surface (Fig. 6.6).
The results reported in Fig. 6.6 are particularly important because CO2 is
a reaction product in most applications concerning hydrogen
production/purification. Therefore, its content will increase along the membrane
axial direction due to both CO2 formation (in membrane reactors) and H2
permeation, thus enriching the retentate stream in non-permeating species. In
this regard, it is noteworthy that at 473 K the H2 flux decreased by ~41 % for a
feed comprising 20 % of CO2 (see Fig. 6.6).
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
197
Fig. 6.6 – Normalized H2 flux as a function of the CO2 concentration in the feed for different temperatures. The fitted function (dashed lines) based on SL model (Eq. (6.8)) is also shown.
Despite the results found in the literature concerning the effects of these
species on membrane permeance towards H2 (see for instance the review of
Paglieri and Way [4]), it is difficult to compare them due to the variety of operating
conditions used, types of membrane (supported or unsupported, Pd or Pd-alloy,
range of thickness, etc.) and preparation techniques. Therefore, for assessing
the performance of a membrane it is better to use normalized conditions, for
instance using the normalized hydrogen flux, i.e. in the presence of
contaminants vs. pure hydrogen.
Some results taken from literature concerning the effect of CO and CO2 on
the H2 flux reduction are shown on Fig. 6.7 for comparison with those obtained
in the present work. In this figure is shown that the temperature effect on H2
permeation (when in the presence of CO or CO2) depends on the membrane
type (Pd/Pd-alloy), thickness and preparation technique. For instance, in the
works of Mejdell et al. [18] and Gielens et al. [19] were employed membranes
prepared by magnetron sputtering (MS). In the former case, 5 % CO in the feed
% (v/v) CO2
0 5 10 15 20 25
JH
2
*
0.4
0.5
0.6
0.7
0.8
0.9
1.0 573 K
548 K
523 K
498 K
473 K
Chapter 6
198
(with a constant total feed pressure of 3 bar) caused the most drastic H2 flux
reduction, whereas in the latter, it is noteworthy that a feed with 20 % of CO2 at
673 K is responsible for the same decrease on H2 flux as 5 % of CO at 573 K for
the membrane used in this work (~55 %).
Fig. 6.7 – H2 flux reduction as a function of temperature for different CO or CO2 feed concentrations and at constant Sieverts driving force. Values taken from literature for several Pd-based membranes are shown for comparison (the lines are for eye guidance).
As shown in Fig. 6.5 and Fig. 6.6, the temperature effect on the membrane
permeation to hydrogen is more notorious in the case of CO than in the case of
CO2, for which parameter α in the Sieverts Langmuir model, Eq. (6.8), remains
nearly constant and equal to 0.45 (see Table 6.1).
Table 6.1 – Sieverts-Langmuir parameters obtained by fitting the experimental data of Fig. 6.5 and Fig. 6.6 to Eq. (6.8).
T / K CO CO2
α K / kPa-1 α K / kPa-1
573 0.63 0.698 0.38 0.109 548 0.68 0.720 0.41 0.116 523 0.77 1.146 0.45 0.121 498 0.86 1.443 0.45 0.128 473 - - 0.45 0.140
T (K)
448 473 498 523 548 573 598 623 648 673 698 723 748
(1-
JH
2*)
x 1
00 (
%)
10
20
30
40
50
60
70
80
90
1005% (v/v) CO2 - 50 m PdAg25 - CR - This work
20% (v/v) CO2 - 0.9 m Pd - MS - [19]
20% (v/v) CO2 - 50 m PdAg25 - CR - This work
5% (v/v) CO - 50 m PdAg25 - CR - This work
5 % (v/v) CO - 3 m PdAg23 - MS - [18]
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
199
In the case of CO, however, parameter α decreases considerably when
the temperature increases. For both gases, the adsorption constant decreases
with T, due to the exothermal nature of the process. Besides, these constants
are, as expected, much higher for carbon monoxide than for carbon dioxide.
The enthalpy of adsorption ( iH ) for each gas was determined using the
Van’t Hoff equation, which is expressed as follows:
i idln
d 1
K H
T R
(6.11)
where R stands for the gas constant. Fig. 6.8 shows the fittings obtained for both
species along with data taken from [18] for comparison.
Fig. 6.8 – Linear regression between the determined adsorption constants of CO and CO2 versus the inverse of temperature.
The adsorption enthalpy obtained for CO2 (-5.41 kJmol-1) clearly indicates
a weak (or physical) adsorption of this compound on the membrane surface.
Physical adsorption is normally attributed to van der Waals [20, 21] and
1000/T (K-1
)
1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2
ln(K
i)
-10
-9
-8
-7
-6
-5
-4
CO - This work
CO2 - This work
CO - [18]
Chapter 6
200
electrostatic forces [21]. In the case of CO, a higher value (-25.23 kJmol-1) was
obtained. Still, this value is inferior than the enthalpy value calculated with the
results obtained by Mejdell et al. [18] (74.07 kJmol-1) for a 3 µm thick Pd/Ag23
wt.% membrane with a microchannel configuration. Nevertheless, in all cases
the enthalpy values are far away the region considered for chemical adsorption
(200 kJmol-1) [20, 22].
At the end of the experiments with these gases, i.e. after shutting off the
flow of CO or CO2, the membrane was heated up to 573 K (the last experiment
was always carried at the lowest temperature studied where adsorption is
higher) and it was checked if the initial hydrogen flux was recovered. In the case
of CO2 it was possible to recover completely the initial flux by passing N2 on both
sides of the membrane, while after using CO a treatment with synthetic air was
required. In the latter case, 50 mLNmin-1 of air was fed to each side for 2 hours,
after which the membrane permeance to H2 was restored. It appears that the
reaction of oxygen with the adsorbed CO produced CO2 that is purged, leading
to complete membrane regeneration. Further insight about heat treatment with
air is given in a work by Mejdell et al. [23] whom observed that performing a heat
treatment with air at 573 K allowed to considerably reduce CO inhibition effect
on H2 permeability at 573 and 623 K. The authors considered that this could be
related to changes on the electronic properties of the membrane surface, where
grain growth and surface roughening occurred.
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
201
6.5 Conclusions
Permeation experiments were performed to assess in which extent PdAg
membrane performance for H2 permeation could be influenced by CO and CO2
presence. A modified Sieverts-Langmuir (SL) equation was proposed based on
the concentration gradients along the membrane length. Although being more
accurate, deviations to SL model with the rearranged equation are negligible in
the case of CO and are below 10.7 % for CO2, in the used conditions.
It was found that CO has a strong inhibition effect on H2 permeation due
to its co-adsorption on the metallic surface. Still, for the used operation
conditions the inhibition effect caused by CO2 exposure must not be neglected
and clearly CO2 cannot be considered as an inert.
The Sieverts-Langmuir equation adequately fitted the experimental data
for CO and, interestingly, also described quite well the results obtained for CO2.
Based on the values of the adsorption enthalpy and for the operating
temperature range employed (473 K – 573 K) it was concluded that CO and CO2
physically adsorb on the membrane. The results here found are important for
modeling purposes since allow predicting the influence caused by these species
in H2 permeance (or recovery).
Finally, it was possible to completely restore H2 permeance by increasing
the temperature from 473 K to 573 K and performing an oxidation procedure with
synthetic air for CO, or simply flowing nitrogen in the case of CO2 exposure.
Chapter 6
202
6.6 References
[1] V. Subramani, P. Sharma, L. Zhang, K. Liu, Catalytic Steam Reforming
Technology for the Production of Hydrogen and Syngas, in: K. Liu, C. Song, V.
Subramani (Eds.) Hydrogen and Syngas Production and Purification
Technologies, John Wiley & Sons, Inc., 2010, pp. 14-126.
[2] G.L. Holleck, Diffusion and solubility of hydrogen in palladium and palladium-
silver alloys, Journal of Physical Chemistry, 74 (1970) 503-511.
[3] E. Wicke, H. Brodowsky, H. Züchner, Hydrogen in palladium and palladium
alloys, in: G. Alefeld, J. Völkl (Eds.) Hydrogen in Metals II, Springer
Berlin/Heidelberg, 1978, pp. 73-155.
[4] S.N. Paglieri, J.D. Way, Innovations in palladium membrane research,
Separation and Purification Methods, 31 (2002) 1-169.
[5] S. Tosti, A. Basile, L. Bettinali, F. Borgognoni, F. Chiaravalloti, F. Gallucci,
Long-term tests of Pd–Ag thin wall permeator tube, Journal of Membrane
Science, 284 (2006) 393-397.
[6] S. Tosti, A. Basile, L. Bettinali, F. Borgognoni, F. Gallucci, C. Rizzello, Design
and process study of Pd membrane reactors, International Journal of Hydrogen
Energy, 33 (2008) 5098-5105.
[7] D. Mendes, V. Chibante, J.M. Zheng, S. Tosti, F. Borgognoni, A. Mendes,
L.M. Madeira, Enhancing the production of hydrogen via water-gas shift reaction
using Pd-based membrane reactors, International Journal of Hydrogen Energy,
35 (2010) 12596-12608.
[8] G. Barbieri, A. Brunetti, G. Tricoli, E. Drioli, An innovative configuration of a
Pd-based membrane reactor for the production of pure hydrogen: Experimental
analysis of water gas shift, Journal of Power Sources, 182 (2008) 160-167.
[9] F. Gallucci, F. Chiaravalloti, S. Tosti, E. Drioli, A. Basile, The effect of mixture
gas on hydrogen permeation through a palladium membrane: Experimental
study and theoretical approach, International Journal of Hydrogen Energy, 32
(2007) 1837-1845.
Effect of CO and CO2 on H2 permeation through finger-like Pd-Ag membranes
203
[10] T.A. Peters, M. Stange, H. Klette, R. Bredesen, High pressure performance
of thin Pd-23%Ag/stainless steel composite membranes in water gas shift gas
mixtures; influence of dilution, mass transfer and surface effects on the hydrogen
flux, Journal of Membrane Science, 316 (2008) 119-127.
[11] European Commission, Renewable Energy Road Map - Renewable
energies in the 21st century: building a more sustainable future, Brussels, 2007.
[12] S. Yun, S. Ted Oyama, Correlations in palladium membranes for hydrogen
separation: A review, Journal of Membrane Science, 375 (2011) 28-45.
[13] A. Basile, F. Gallucci, S. Tosti, Synthesis, Characterization, and
Applications of Palladium Membranes, in: M. Reyes, M. Miguel (Eds.) Membrane
Science and Technology, Elsevier, Amsterdam, The Netherlands, 2008, pp.
255-323.
[14] G. Barbieri, F. Scura, F. Lentini, G. De Luca, E. Drioli, A novel model
equation for the permeation of hydrogen in mixture with carbon monoxide
through Pd-Ag membranes, Separation and Purification Technology, 61 (2008)
217-224.
[15] F.P. Incropera, D.P. DeWitt, Fundamentals of Heat and Mass Transfer, 4th
ed., John Wiley & Sons, Inc., 1996.
[16] S. Tosti, L. Bettinali, Diffusion bonding of Pd-Ag rolled membranes, Journal
of Materials Science, 39 (2004) 3041-3046.
[17] J. Chabot, J. Lecomte, C. Grumet, J. Sannier, Fuel clean-up system:
poisoning of palladium-silver membranes by gaseous impurities, Fusion
Technology, 14 (1988) 614-618.
[18] A.L. Mejdell, M. Jøndahl, T.A. Peters, R. Bredesen, H.J. Venvik, Effects of
CO and CO2 on hydrogen permeation through a ~3 μm Pd/Ag 23 wt.%
membrane employed in a microchannel membrane configuration, Sep. Purif.
Technol., 68 (2009) 178-184.
[19] F.C. Gielens, R.J.J. Knibbeler, P.F.J. Duysinx, H.D. Tong, M.A.G.
Vorstman, J.T.F. Keurentjes, Influence of steam and carbon dioxide on the
Chapter 6
204
hydrogen flux through thin Pd/Ag and Pd membranes, Journal of Membrane
Science, 279 (2006) 176-185.
[20] P. Atkins, J. De Paula, Physical Chemistry, 8th ed., Oxford University Press,
2006.
[21] D.M. Ruthven, Principles of Adsorption and Adsorption Processes, 1st ed.,
John Wiley and Sons Inc., 1984.
[22] J.U. Keller, R. Staudt, Gas adsorption equilibria: experimental methods and
adsorptive isotherms, Springer, 2005.
[23] A.L. Mejdell, D. Chen, T.A. Peters, R. Bredesen, H.J. Venvik, The effect of
heat treatment in air on CO inhibition of a ~3 μm Pd-Ag (23 wt.%) membrane, J.
Membr. Sci., 350 (2010) 371-377.
Conclusions and Future Work
205
Chapter 7. Conclusions and Future Work
7.1 Conclusions
This thesis focused on the study and development of a carbon capture and
utilization (CCU) process able to, simultaneously, remove CO2 from flue gas and
afterwards convert it to methane in a sorptive reactor. The proof-of-concept of
such intensified process was performed after gathering important insight
regarding thermodynamics, adsorption and reaction fundamentals, which were
individually studied.
This work reports for the first time the characterization of CO2 sorption on
hydrotalcites, CO2 methanation over nickel-based catalysts, and testing a
sorptive reactor with reactive regeneration at LEPABE. Work was also done
regarding H2 permeation through Pd-based membranes, a topic however
already addressed previously (e.g. [1]).
The first task consisted on the design and equipment selection followed by
the assembly of an experimental setup suitable to perform the sorption, reaction
and sorption-reaction experiments. Some modifications to the original design
had to be made later to comply with certain operation issues during the sorption-
reaction experiments (see photo of the lab-scale bench in Appendix D). Overall,
this task was time consuming, but the reasoning made for designing the setup
allowed being previously aware of the operational limits/difficulties when
planning the experiments. The H2 permeation tests with PdAg membranes were
performed in an existing experimental setup in the group.
In the thermodynamic study (Chapter 2), two routes for flue gas-derived
CO2 valorisation, based on its direct hydrogenation to methane or methanol,
were compared. The study showed that methanol pathway requires harsher
operation conditions than the option of CO2 reduction to methane, particularly in
terms of pressure. In the case of CO2 valorisation to methane, it was concluded
that CH4 selectivity increases with the H2/CO2 molar feed ratio, being near 100
Chapter 7
206
% (for 150 ºC ≤ T ≤ 400 ºC and 1 bar ≤ T ≤ 30 bar) if a stoichiometric H2/CO2
molar feed ratio of 4 is used; a small fraction of CO can however be formed
through the reverse water gas shift reaction, being this pathway favoured at low
pressure and high temperature. Carbon formation is avoided for a H2/CO2 ratio
of 4, but it can become significant for lower ratios. In this case, carbon formation
can be inhibited if the water present in the flue gas is not removed from the
methanation reactor feed, although it decreases CO2 conversion because it is a
product of the methanation reaction.
Oxygen content in flue gas is generally low (< 5 vol. %), but it can cause a
considerable decrease of CO2 conversion due to the consumption of H2 and CH4.
The removal of the heat released by these extremely exothermic oxidation
reactions is another problem, particularly for safety reasons and to prevent
temperature excursions up to a level that can cause catalyst sintering. Moreover,
the presence of O2 leads to the oxidation of the methanation catalyst, thus
requiring additional H2 consumption for catalyst reactivation [2]. For these
reasons, a previous step of flue gas processing for O2 removal is recommended
before being admitted to a methanation reactor.
The aim of the high temperature study of CO2 sorption with hydrotalcites
(Chapter 3) was to synthesize materials with good CO2 sorption capacity and
fast kinetics at high temperature that could be used in the cyclic sorption-reaction
process for CO2 capture and conversion. Hydrotalcite-based materials were
selected based on their ability to operate in the temperature range compatible
with methanation catalysts and their enhanced performance in wet streams (e.g.
[3-5]). Following the works of Yavuz et al. [6] and Oliveira et al. [5], gallium-
substituted hydrotalcites were prepared by co-precipitation and promoted by
impregnation with ca. 20 wt. % of K, Cs or Sr. The sorbents were screened based
on their sorption capacity at 300 ºC under dry conditions. The sorption
equilibrium isotherms of all sorbents were determined from 0.05 bar to 1.1 bar
of CO2 and were well described by the Freundlich equation. Promotion with Cs
Conclusions and Future Work
207
and particularly with K improved the sorption capacity, while Sr decreased
(compared to the unpromoted hydrotalcite).
Based on its higher sorption capacity, the gallium-substituted and K-
promoted hydrotalcite was selected and submitted to sorption-desorption cycles
to determine its working capacity under low pressure conditions (<0.0001 to 0.15
bar), as found in post-combustion CO2 capture applications. The sorbent’s
working capacity obtained at 300 ºC was 0.97 mmol·cm-3, representing 60 % of
the initial capacity, while the remaining 40 % could not be removed even under
vacuum. The working capacity at 300 ºC was almost fourfold higher than at 200
ºC, which suggests that CO2 sorption takes place through an activated process.
A kinetic model based on the CO2 sorption mechanism on hydrotalcites
suggested by Du et al. [7] was proposed to describe the sorption and desorption
kinetics. The model assumes two parallel contributions, one responsible for the
initial fast uptake and the other by the slow uptake observed for longer times.
The former is related to the reversible fraction of CO2 that is sorbed as bridged,
unidentate and bidentate carbonate, while the latter is related to the fraction of
irreversible sorbed CO2 in the form of polydentate carbonate. The kinetic model
considers 3 fitting parameters 2 2
1 2 1, and D r D r q q , and adequately fitted
the experimental uptake and downtake curves at 200 ºC and 300 ºC. The
temperature did not affect the initial relative contribution of the reversible amount
of sorbed CO2, being nearly the same at 200 ºC and 300 ºC (i.e. 48 and 44 %,
respectively). However, the temperature influences the values of the diffusion
coefficients. The amount of reversibly sorbed CO2 reaches the equilibrium faster
at 200 ºC than at 300 ºC, with the contrary being observed for the irreversible
fraction of sorbed CO2.
Sorption-desorption cycling leads to sorption kinetics slowdown of both the
reversible and irreversible contributions, independently of the temperature.
Globally, desorption was much faster at 200 ºC than at 300 ºC, especially for
removing the CO2 which was reversible sorbed.
Chapter 7
208
The intrinsic methanation reaction kinetics was determined in Chapter 4
over an industrial nickel catalyst (METH 134 from Clariant) that was later
employed for testing the integrated sorption-reaction process. The catalyst
activity remained stable for (at least) 83 hours under reactive conditions.
Extraparticle and intraparticle mass and heat transport resistances were found
negligible based on theoretical criteria and experimental observations. The
kinetic tests were conducted in a relevant temperature window (250-350 ºC) and
the pressure was the atmospheric. An ideal H2/CO2 molar ratio of 4 was used to
minimize CO and carbon formation, based on previous findings of the
thermodynamic study.
The observed reaction rates were extracted applying the differential
method and adjusted through non-linear fitting towards six kinetic models taken
(and adapted) from literature assuming three types of reaction mechanisms:
carbon intermediate, formate intermediate and formyl intermediate. Model
discrimination consisted on assessing the thermodynamic consistency and
statistical testing of model parameters. The statistical F-test allowed to choose
between rival models.
The chosen model has only four fitting parameters and is based on the
formyl intermediate mechanism, which considers that CO2 methanation begins
with H2 and CO2 dissociation and assumes hydroxyl as the most abundant
species at the catalyst surface. The kinetic equation was also successfully
validated through the modelling of an isothermal plug-flow fixed-bed reactor to
describe the experimental data.
In this thesis, the major contribution and novelty was related to the proof-
of-concept of a sorption-reaction process featuring reactive regeneration for
carbon dioxide capture by adsorption, followed of its conversion to methane; this
is addressed in Chapter 5. The sorptive reactor used for the proof-of-concept
was packed with a commercial hydrotalcite sorbent (Pural MG30-K from Sasol)
and a commercial nickel catalyst (METH 134 from Clariant), which are supplied
in millimetre size and in quantities compatible with the test bench. For these two
Conclusions and Future Work
209
reasons, the gallium-substituted and K-promoted hydrotalcite synthesized and
characterized in Chapter 3 could not be used in this stage, although showing a
higher sorption capacity than the commercial Pural MG30-K employed.
The sorptive reactor was operated between 300 ºC and 350 ºC, the
pressure was low (≤ 2.5 bar), since flue gas is generally at the atmospheric
pressure, and the feed for the capture step consisted in 15 % vol. of CO2
balanced in N2. Given the complexity of the sorption and reaction processes
integration, other species present in flue gas like H2O and O2 were not included
in the feed stream at this stage.
The CO2 sorbent working capacity measured at 350 ºC after 6 sorption-
desorption cycles (purge with N2) was 0.32 mmol·g-1. When a H2 stream was
used for the desorption step instead of N2 (i.e. reactive regeneration mode), the
initial sorption capacity of the sorbent was recovered (which was not the case
with N2), and the desorption kinetics were enhanced because of the steam
produced in situ by the methanation reaction taking place at the catalyst surface.
The sorption equilibrium of the Pural MG30-K hydrotalcite was well
described by the dual-site Langmuir model using the parameters proposed by
Oliveira et al. [5] for the same sorbent. The endothermic term of the dual-site
Langmuir model explains the increased sorption capacity with temperature, as
also observed previously with the in-house prepared gallium-substituted and K-
promoted hydrotalcite. Knowing the dependence with temperature of the
physical (exothermic) and chemical (endothermic) contributions relative to the
total amount of sorbed CO2 is crucial to select the optimum conditions for
operating the cyclic process, namely to find the best compromise in terms of
sorption capacity and desorption kinetics. In addition, it was shown that
controlling the temperature and pressure could help minimizing the CO formed
though the parallel reverse water shift reaction.
The temperature did not affect methane productivity in the hybrid
multifunctional reactor, at least under the considered operation conditions. On
the contrary, methane purity was slightly higher at 350 ºC than at 300 ºC. The
Chapter 7
210
maximum CH4 purity obtained was 37 % and further reactor design should be
considered to improve this value up to the level required for injection in the
natural gas grid. Among the reactor design options, which will be presented and
discussed in the next section, one consists in changing the H2 feed system to
assure that a constant H2/CO2 ratio of 4 (the stoichiometric one) is achieved
along the reactor length during the regeneration stage, for instance through
radial feeding employing a H2-selective membrane. In this respect, the final part
of the thesis (Chapter 6) focused on the evaluation of the permeance of a 50 µm
thick PdAg membrane for a pure H2 stream and for binary mixtures of industrial
relevance, namely H2/CO and H2/CO2, in the temperature range of 200-300 ºC.
The H2 molar flux permeating through the membrane was well described
by the Sieverts law for a pure H2 feed. For a binary H2/CO or H2CO2 feed, the
Sieverts-Langmuir equation proved better. A modified Sieverts-Langmuir
equation assuming a logarithmic mean driving force to take into consideration
the concentration gradients along the membrane length due to H2 permeation
was proposed. This new equation founds relevance in conditions of high H2
recovery, but for design and modelling purposes, it was concluded that the
simpler Sieverts-Langmuir equation is preferred, even if some inaccuracy must
be admitted, because it does not require knowing beforehand the H2 partial
pressure on the retentate side.
It was found that CO has a strong inhibitory effect on H2 permeance
through the Pd-based membrane, even at low feed contents (< 5 vol. %), being
significantly worse than CO2. The inhibition was caused by physical adsorption
on the membrane, by blocking the available surface area for hydrogen
permeation that occurs through the solution-diffusion mechanism. The inhibitory
effect on H2 permeance decreases with temperature, but while CO2 completely
desorbs by simply increasing the temperature and flushing the system with N2,
membrane regeneration to remove adsorbed CO required membrane oxidation
with air.
Conclusions and Future Work
211
Overall, the work performed in this thesis showed the technical feasibility
of a sorptive reactor for CO2 capture and conversion to methane. The good
performance of the commercial sorbent and catalyst, besides their compatibility
and stability when operating under cyclic operation, allow to be optimistic
regarding the possibility to scale-up this Carbon Capture and Utilization
technology and its potential relevance within the Power-to-Gas applications field.
7.2 Future work
The proof-of-concept of a sorptive reactor for CO2 capture and conversion
to methane was performed in this thesis. Several issues that can influence the
sorptive reactor performance in terms of CO2 capture capacity, CH4 productivity
and purity were identified and therefore future work is suggested.
First, the effect of other species present on flue gas should be
systematically assessed (e.g. steam and O2) and, afterwards, the sorptive
reactor must be tested employing a real flue gas stream. For instance, it is
expected that steam will enhance the CO2 capture step if a hydrotalcite sorbent
is used but, on the other side, steam will probably decrease methane productivity
during the CO2 conversion step. When considering a real flue gas composition,
the study will also provide evidence of the sorbent and catalyst tolerance to
certain impurities (e.g. SOx). This is crucial for process design, namely for
establishing previous flue gas processing requirements and to conclude about
the suitability of the considered materials. Depending on the effect caused by
other flue gas species on performance indicators, it could be considered to test
the sorptive reactor employing other sorbents and/or catalysts, or to find
optimum operation conditions to minimize potential negative effects. In this
respect, it would be interesting to test the concept using Ru-based catalysts,
which exhibit good activity and selectivity towards methane at lower
temperatures. In this way, both undesired CO and carbon formation reactions
would be disfavoured. Besides, some Ru-based catalysts for methanation are
Chapter 7
212
already commercially available, as presented in Chapter 4. Another option could
be including a water-gas shift catalyst layer (e.g. Cu-based) at the end of the
reactor to act as guard bed, converting CO traces and preventing its presence
in the outlet stream.
Other suggestions for future work include testing different H2 feeding
systems (e.g. countercurrent feed flow, radial feed flow with multiple entrances
along the bed length or through a Pd-based membrane), different bed layouts
(layered, mixed, etc.) and different sorbent-to-catalyst ratios. The use of
multifunctional materials (i.e., CO2 sorbents impregnated or somehow mixed
with methanation catalysts) is another strategy to pursue, apart from those
herein developed (namely the modified hydrotalcite) after appropriate
pelletization.
Testing different H2 feeding systems is important since CO2 desorbs during
reactive regeneration by purge gas stripping. Generally, purge gas stripping
method is performed with a non-adsorbing inert gas only when the adsorbate is
weakly held, otherwise it requires a high quantity of purge gas [8]. In this case,
the purge gas being used is H2, which is not retained by the sorbent, but is
consumed during the methanation reaction. Thus, the H2 feeding system should
deliver an H2 flow along the reactor length that allows to, simultaneously, achieve
complete CO2 desorption from the sorbent and the existence of an H2/CO2 molar
ratio of 4 in the gas phase near the catalyst.
Finding the optimum sorbent-to-catalyst ratio and bed layout is also
important to avoid packing the bed with an unnecessary weight of catalyst for
processing the amount of captured CO2.
Finally, it would be important to implement the cyclic batch process,
perhaps starting with a basic two-bed configuration and a modified Skarstrom
cycle to include reactive regeneration. Afterwards, modelling and simulation of
the cyclic process (with experimental validation) would allow for process design
and optimization. For the materials used in this work, some model parameters
were estimated (e.g. intrinsic methanation kinetics over METH 134), while for
Conclusions and Future Work
213
the sorbent both sorption equilibrium and kinetic parameters are available in
literature (e.g. [5]).
Chapter 7
214
7.3 References
[1] D. Mendes, V. Chibante, J.M. Zheng, S. Tosti, F. Borgognoni, A. Mendes,
L.M. Madeira, Enhancing the production of hydrogen via water-gas shift reaction
using Pd-based membrane reactors, International Journal of Hydrogen Energy,
35 (2010) 12596-12608.
[2] Q. Zheng, R. Farrauto, A. Chau Nguyen, Adsorption and methanation of flue
gas CO2 with dual functional catalytic materials: a parametric study, Industrial &
Engineering Chemistry Research, 55 (2016) 6768-6776.
[3] Y. Ding, E. Alpay, Equilibria and kinetics of CO2 adsorption on hydrotalcite
adsorbent, Chemical Engineering Science, 55 (2000) 3461-3474.
[4] M.K. Ram Reddy, Z.P. Xu, J.C. Diniz da Costa, Influence of water on high-
temperature CO2 capture using layered double hydroxide derivatives, Industrial
& Engineering Chemistry Research, 47 (2008) 2630-2635.
[5] E.L.G. Oliveira, C.A. Grande, A.E. Rodrigues, CO2 sorption on hydrotalcite
and alkali-modified (K and Cs) hydrotalcites at high temperatures, Separation
and Purification Technology, 62 (2008) 137-147.
[6] C.T. Yavuz, B.D. Shinall, A.V. Iretskii, M.G. White, T. Golden, M. Atilhan,
P.C. Ford, G.D. Stucky, Markedly improved CO2 capture efficiency and stability
of gallium substituted hydrotalcites at elevated temperatures, Chemistry of
Materials, 21 (2009) 3473-3475.
[7] H. Du, C.T. Williams, A.D. Ebner, J.A. Ritter, In situ FTIR spectroscopic
analysis of carbonate transformations during adsorption and desorption of CO2
in K-promoted HTlc, Chemistry of Materials, 22 (2010) 3519-3526.
[8] D.M. Ruthven, Principles of Adsorption and Adsorption Processes, 1st ed.,
John Wiley and Sons Inc., 1984.
Supplementary information for Chapter 3
215
Appendix A. Supplementary information for Chapter 3
Fig. A.0.1 – XRD patterns of samples HTC-10Ga as prepared and after calcination at 673 K.
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Inte
nsi
ty[au]
diffraction angle (2θ Cu Kα)
HTC-10Ga (calcined)
HTC-10Ga
Appendix A
216
Fig. A.0.2 – EDS analyses for samples HTC-20K (first row) and HTC-10Ga-20K (second row) showing the K, Mg, Al and Ga distribution across
different zones.
Supplementary information for Chapter 3
217
Fig. A.0.3 – Adsorption/desorption isotherms of nitrogen over calcined samples: a) HTC, b) HTC-10Ga, c) HTC-20K, d) HTC-10Ga-
20Cs, e) HTC-10Ga-20Sr and f) HTC-10Ga-20K.
Appendix A
218
Fig. A.0.4 – Uptake (and downtake) curves (expressed in mmol·cm-3) for CO2 sorption at 573 K (first row) and 473 K (second row) for sample HTC-
10Ga-20K.
Supplementary information for Chapter 3
219
Buoyancy correction
The specific amount of gas adsorbed (q) can be calculated from:
ads
s
adsorbedmoles mq
mass of adsorbent m M (A.1)
where mads is the mass of adsorbed gas, M is the molecular weight of the
adsorbate, and ms is the mass of the activated sample.
Buoyancy is the upward force exerted over an object immersed in a fluid
and it is equal to the weight of the fluid displaced by the object. Buoyancy does
not actually affect the mass of an object, however it can affect its measure and
correction might be necessary.
So, for buoyancy correction, the weight of fluid displaced by the activated
sample ( s gV ), the adsorbed gas ( ads gV ) and the sample container i gV must be
determined and is equal to:
( )s ads i gMass of displaced gas V V V (A.2)
where Vs is the volume of the solid adsorbent, Vads is the volume of adsorbed
gas, Vi is the volume of the measuring cell (axis, basket and glass wool) and
g is the density of the gas at measuring conditions: g
PMz
T
, where z is
the acentric factor obtained from the virial equation of state [1].
Then, the real mass of the sample (mreal), i.e., corrected by the buoyancy
effect is equal to sum of the mass given by the magnetic balance (m) and the
mass of displaced gas ( ( )s ads i gV V V ):
giadss VVVmm )(real (A.3)
Appendix A
220
The adsorbed mass is calculated from the mass difference between the
real mass (Eq. (A.3)) and the sum of the initial mass of the activated sample (
sm ) and the measuring cell ( im )
ads s is ads i gm m (V V V ) (m m ) (A.4)
Assuming that the adsorbed phase presents a density similar to that of the
liquid phase:
ads adsads
ads l
m mV
(A. 5)
and
l
M
v (A.6)
where v is the molar volume calculated by the Gunn-Yamada method [1], then:
adsads s i
l
( ) ( )s i g
mm m m m V V
(A.7)
Solving Eq. (A.7) as function of mads the following expression is obtained:
lads
l
s i g
g
m m (V V )
(A.8)
where Δm represents the mass difference between the mass given by the
microbalance (m) and the sum of the mass of the activated sample (sm ) and
the mass of the measuring cell (axis, basket and glass wool) ( im ).
Finally, the adsorption capacity (q) can be calculated from:
Supplementary information for Chapter 3
221
s i
s
g l
l g
m (V V )q
m M
(A.9)
The volume of the basket with (Vi+Vs) and without (Vi) the sample, and
the respective mass, mi+ms and mi, were calculated by helium picnometry.
References
[1] B.E. Poling, J.M. Prausnitz, J.P. O'Connell, The properties of gases and
liquids, 5th ed., McGraw-Hill, New York ; London, 2000.
Supplementary information for Chapter 4
223
Appendix B. Supplementary information for Chapter 4
Calculation of transport and thermodynamic
parameters
Table B.1 lists some thermodynamic properties of the species considered
in this study.
Table B.1 – Properties of species used in this study [1].
Property CO2 H2 H2O CH4 N2
Molar weight, M (g·mol-1) 44.01 2.02 18.02 16.04 28.01 Critical temperature, Tc (K) 304.2 33.19 647.1 190.7 126.2 Critical pressure, Pc (bar) 73.83 13.13 220.55 46.00 34.00
Heat capacity of the gas mixture
The specific heat capacity of the gas mixture assuming ideal gas state,
p,mixC (J·kg-1·K-1), was calculated as follows [1]:
i p,i
i 1 i
n
p,mix
yCC
M
(B.1)
where yi and Mi (in kg·mol-1) stand for the mole fraction and molar weight of
species i, respectively, while the ideal gas state molar heat capacity of species
i, p,iC (J·mol-1·K-1), was retrieved from [1]:
p,i 2 2A B C DC
T T TR
(B.2)
Appendix B
224
where A-D are constants that vary for each species and are listed in Table B.2
and R is the ideal gas constant (8.314 J·mol-1·K-1 or m3·Pa·mol-1·K-1).
Table B.2 – Coefficients to estimate the heat capacity of the pure species.
Species A 103 B 106 C 10-5 D
CO2 5.457 1.045 0.000 -1.157 H2 3.249 0.422 0.000 0.083 H2O 3.470 1.450 0.000 0.121 CH4 1.702 9.081 -2.164 0.000 N2 3.280 0.593 0.000 0.040
Heat of reaction
The heat of reaction, rH (J·mol-1), at a given temperature T (K) was
estimated according to [1, 2]:
0
298
298 K
TpK
r r
CH T H R dT
R
(B.3)
where 298K
rH (J·mol-1) stands for the standard heat of the methanation reaction.
The second term of Eq. (B.3) was estimated as follows [1]:
0
2 2 3 3
0 0 0298
0
B C D 1A 1 1 1
2 3
T p
K
CdT T T T
R T
(B.4)
with 0 298 T K , 0T T and 0
pC is the standard heat capacity change of the
gas mixture given in J·mol-1·K-1. Coefficients A, B, C and D are listed in Table
B.2.
Supplementary information for Chapter 4
225
Density of the gas mixture
Assuming ideal gas behavior, the density of the gas mixture, g (kg·m-3),
was calculated according to:
g i i
1
n
i
Py M
RT
(B.5)
where P stands for the total pressure (Pa).
Viscosity of the gas mixture
The viscosity of the gas mixture, g (Pa·s), can be estimated using the
Wilke method [2] through Eq. (B.6)),
i ig
1j ij
=1
n
ni
j
y
y
(B.6)
where n is the number of species, and the binary term, ij , is calculated through
Eq. (B.7).
20 5 0 25
i j ij
ij 0 5
i j
1
8 1
. .
.
M M
M M
(B.7)
The viscosity of the pure species, i (Pa·s), is estimated through Eq. (B.8) [1]:
Appendix B
226
B
i
2
A
C D1
T
T T
(B.8)
where A, B, C and D are coefficients that depend on the species and were
retrieved from Aspen Properties software. Table B.3 lists the coefficients A-D
used to determine pure viscosity of each species.
Table B.3– List of coefficients A-D of each species to determine pure viscosity in Pa·s.
Species A B C D
CO2 2.1480 x 10-6 0.46000 290.000 0 H2 1.7970 x 10-7 0.68500 -0.590 140 CH4 5.2546 x 10-7 0.59006 105.670 0 H2O 1.7096 x 10-8 1.11460 0.000 0 N2 6.5592 x 10-7 0.60810 54.714 0
Molecular diffusivity of the gas mixture
The molecular diffusivity of species i in the gas mixture, i,mixD (m2·s-1), was
estimated through Eq. (B.9), as proposed by Fairbanks and Wilke [3]:
,mix
1
1 ii n
j
j ijj
yD
y
D
(B.9)
where ijD (cm2·s-1) is the binary diffusivity of species i when diffusing into a
component j. To estimate ijD , the Fuller correlation was considered (cf. Eq.
(B.10)) [4].
Supplementary information for Chapter 4
227
0 5
3 1 75
21 3 1 3
1 110
.
.
i j
ij
i j
TM M
D
P
(B.10)
with T in K, P in atmospheres and Mi in g·mol-1. The diffusion volume of a given
molecule is equal to the sum of the atomic diffusion volumes (i.e. ). The
values for the molecules considered in this study are given in Table B.4.
Table B.4 – Diffusion volumes for simple molecules, (cm3·mol-1), of species
considered in this work [1].
Molecule Diffusion volume,
CO2 26.9 H2 7.07 H2O 12.7 CH4 24.4 N2 17.9
The binary diffusivities calculated by Eq. (B.10) must be converted to SI
units before substitution in Eq. (B.9).
Effective mass diffusivity in the catalyst
particle
The effective mass diffusivity in the catalyst particle, effD (m2·s-1), was
calculated through Eq. (B.11) [1]:
2CO ,mix p
eff
p
DD
(B.11)
Appendix B
228
where 2CO ,mixD (m2·s-1) is the CO2 diffusivity in the gas mixture calculated from
Eq. (B.9); p and p are the porosity and tortuosity of the catalyst particle,
respectively. The particle porosity was assumed equal to 0.4. The tortuosity
generally is ranged between 2 and 6 [5]. For a catalyst particle, it is
recommended to consider a tortuosity value of 4 when no other information is
available [1, 6].
Thermal conductivity of the gas mixture
The thermal conductivity of the gas mixture, g (W·m-1·K-1), was estimated
following the Wassiljewa method (cf. Eq. (B.12)) [2]:
g,
g
i 1
1
ni i
n
j ij
j
y
y A
(B.12)
where ijA is the binary interaction parameter calculated by Eq. (B.13), as
suggested by Mason and Saxena [2],
20 5 0 25
0 5
1
8 1
. .
tr ,i tr , j i j
ij .
i j
M MA
M M
(B.13)
with tr ,i tr , j being the ratio of the translational thermal conductivities
calculated from Eq. (B.14) [2].
Supplementary information for Chapter 4
229
exp 0 0464 exp 0 2412
exp 0 0464 exp 0 2412
j r ,i r ,itr ,i
tr , j i r , j r , j
. T . T
. T . T
(B.14)
r ,i c ,iT T T stands for the reduced temperature and i (m·K·W-1) is the reduced,
inversed thermal conductivity of species i (or j) according to Eq. (B.15) [2],
1 63
4
210 c,i i
i
c ,i
T M
P
(B.15)
where c ,iT (K), c ,iP (bar) and Mi (kg·mol-1) are the critical temperature, critical
pressure and molar weight of species i (see Table B.1).
The thermal conductivity of species i, g,i (W·m-1·K-1), was calculated
through the Eucken model (Eq. (B.16)) [2]:
g,
5
4
p,i
i i
i i
CR
M M
(B.16)
with i in Pa·s (cf. Eq. (B.8)), Mi in kg·mol-1 (cf. Table B.1) and p,iC in J·mol-1·K-
1 (cf. Eq. (B.2)).
Mass and heat transport coefficients
The mass and heat transport coefficients between the gas and the catalyst
particle, kg (m·s-1) and h (W·m-2·K-1), respectively, were estimated from a
correlation expressed in terms of the Colburn J factor analysis [7]:
Appendix B
230
2 3g
D
o
kJ Sc
u (B.17)
2 3
, g
PrH
p mix o
hJ
C u (B.18)
2 3
p1.09ReD HJ J
(B.19)
where ou stands for the fluid superficial velocity (m·s-1); the Schmidt (Sc),
Prandtl (Pr) and Reynolds pRe numbers are given by Eqs. (B.20)-(B.22),
respectively.
2
g
g ,CO mix
ScD
(B.20)
, g
g
Prp mixC
(B.21)
g
g
Reo p
p
u d (B.22)
Effective thermal conductivity of the catalyst
The effective thermal conductivity of the catalyst, eff (W·m-1·K-1), was retrieved
using the correlation expressed by Eq. (B.23) [8], where s (W·m-1·K-1) is the
solid thermal conductivity, which was estimated from information found
elsewhere [8-10].
1
sg
g
p
eff
(B.23)
Supplementary information for Chapter 4
231
Parameters used to assess mass and heat
transport limitations (Section 4.1.2)
Table B.5 – Parameters used to assess mass and heat transport limitations*.
Parameter Value Units
n 1
pd 256 x 10-6 m
,p mixC 3263.3 J·kg-1·K-1
rH -179.7 kJ·mol-1
g 0.47262 kg·m-3
g 2.9943 x 10-5 Pa·s
aE 118.0 kJ·mol-1
p 4
p 0.4
2 ,CO mixD 6.8062 x 10-5 m2·s-1
effD 6.8062 x 10-6 m2·s-1
gk 0.138 m·s-1
h 161.6 W·m-2·K-1
g 6.946 x 10-2 W·m-1·K-1
s 13.9 W·m-1·K-1
eff 1.667 W·m-1·K-1
* Parameters estimated based on feed conditions at the higher temperature (350
ºC) and higher 2CO
in
catW F .
Appendix B
232
References
[1] R.H. Perry, D.W. Green, Perry's Chemical Engineers' Handbook, McGraw
Hill, New York, 1999.
[2] B.E. Poling, J.M. Prausnitz, J.P. O'Connel, The properties of gases and
liquids, 5th ed ed., The McGraw-Hill Companies, New York, 2001.
[3] D.F. Fairbanks, C.R. Wilke, Diffusion Coefficients in Multicomponent Gas
Mixtures, Industrial & Engineering Chemistry, 42 (1950) 471-475.
[4] J.R. Welty, C.E. Wicks, R.E. Wilson, G.L. Rorrer, Fundamentals of
Momentum, Heat and Mass Transfer, John Wiley & Sons, Inc., 2008.
[5] D.M. Ruthven, Principles of Adsorption and Adsorption Processes, 1st ed.,
John Wiley and Sons Inc., 1984.
[6] C.N. Satterfield, Heterogeneous Catalysis in Practice
McGraw-Hill, 1980.
[7] A.E. Rodrigues, Scientific basis for the design of two phase catalytic reactors,
in: A.E. Rodrigues, J.E. Calo, N.H. Sweed (Eds.) Multiphase Chemical Reactors,
Volume II – Design Methods, Sijthoff & Noordhoff International Publishers B.V.,
Alphen aan den Rijn, The Netherlands, 1981.
[8] M. Soomro, R. Hughes, The thermal conductivity of porous catalyst pellets,
The Canadian Journal of Chemical Engineering, 57 (1979) 24-28.
[9] P. Auerkari, Mechanical and physical properties of engineering alumina
ceramics, Technical Research Centre of Finland, VTT Manufacturing
Technology, Research Notes 1792, 3-36, Espoo, 1996.
[10] J.E. Keem, J.M. Honig, Selected electrical and thermal properties of
undoped nickel oxide, Center for Information and Numerical Data Analysis and
Synthesis, Purdue University, Cindas Report 52, West Lafayette, Indiana, 1978.
Supplementary information for Chapter 5
233
Appendix C. Supplementary information for Chapter 5
(cont.)
Appendix C
234
(cont.)
Supplementary information for Chapter 5
235
Fig. C.0.1 – History of pressure, CO2 and CH4 outlet fraction during regeneration at
350ºC (cycles 8-10) and 300 ºC (cycles 11-15) for the initial 2000 s.
Appendix C
236
(cont.)
Supplementary information for Chapter 5
237
Fig. C.0.2 – Normalized temperature histories in four bed positions during CO2 sorption and regeneration with H2 (cycles 11 to 15); Tz,0 = 300 º C.
Appendix C
238
(cont.)
Supplementary information for Chapter 5
239
Fig. C.0.3 – Normalized temperature histories in four bed positions during CO2 sorption and regeneration with H2 (cycles 7 to 10); Tz,0 = 350 º C.
Experimental Setup
241
Appendix D. Experimental Setup
Fig. D.0.1 – Assembled experimental setup for the PhD work.