2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
Assessment of Thermochemistry and Preliminary Design
of a Carbon Looping Combustion Process (CarboLoop)
Piero SALATINO and Osvalda SENNECA
2
Dipartimento di Ingegneria Chimica, Università degli Studi di Napoli Federico II, Piazzale
Vincenzo Tecchio 80, 80125 Napoli, Italy
3
Istituto di Ricerche sulla Combustione, Consiglio Nazionale delle Ricerche, Piazzale
Vincenzo Tecchio 80, 80125 Napoli, Italy
*Corresponding Author, piero.salatino@unina.it
Abstract – The CarboLoop concept proposed by Salatino and Senneca [1]
pursues the simple idea that a carbon-based fuel can act itself as oxygen carrier
as it is cycled between an Oxidizer, where the fuel uptakes oxygen at moderate
temperature (say 300-400°C), and a Desorber where desorption of surface
oxides is accomplished in inert atmosphere at higher temperature (in the order of
700°C). Looping of the carbon fuel between the oxidizer and the desorber
enables stepwise conversion of carbon over multiple cycles yielding a nearly pure
stream of CO2, ready for sequestration, at the exhaust of the Desorber.
This study further contributes to the exploitation of the CarboLoop concept by
reporting on the thermochemistry of the oxidation and desorption stages and by
analyzing issues associated with energy integration between the Oxidizer and the
Desorber. A calorimetric study on raw and pre-oxidized carbon samples indicates
that carbon oxidation at temperatures in the order of 300-400°C is moderately
exothermic, whereas a pronounced exothermicity is displayed upon exposure of
the oxidized samples at higher temperature. The proposed mechanistic frame is
based on the formation of “metastable” complexes upon exposure to oxygen at
moderate temperature, followed by isomerization/rearrangement into “stable”
oxides and subsequent desorption at higher temperature.
Thermochemical and kinetic data provide the basis for a preliminary layout of the
CarboLoop process. Carbon inventory and circulation rate between the Oxidizer
and the Desorber and thermal loading in either reactor are assessed with
reference to an assigned thermal throughput and set of operating/design
parameters.
1
Introduction
Chemical looping combustion of solid fuels is rapidly gaining interest as a challenging
capture-ready combustion technology. Inherent limitations associated with direct oxidation of
the solid fuel by a solid oxygen carrier have been overcome by two approaches [2]. The first
one is to gasify the fuel and to introduce the syngas produced in the CLC system (syngasCLC). The second approach is based on directly feeding the solid fuel to the CLC converter
(solid fuelled-CLC). In this last case two further options may be envisaged: either the solid
fuel is gasified in-situ by H2O or CO2 supplied as fluidization agent (in-situ Gasification
Chemical-Looping Combustion, iG-CLC); or the solid fuel is converted along the reaction
1
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
FG,O=14.3kg/s
FG,D=3.7kg/s
FC,1=5kg/s
Desorber
Oxidizer
qD=24.2MW
qO=1.5MW
TD=1000K
WD=3600kg
TO=600K
FC,2=4kg/s
WO=4500kg
solid
gas
heat
FC,O=1kg/s
FG,O,in=17kg/s
Figure 1. Outline of the CarboLoop process based on a dual bed reactor.
Values refer to results of base case computations
path consisting of Oxygen Uncoupling from the carrier followed by heterogeneous oxidation
of the fuel by the released oxygen (CLOU).
An alternative method to overcome the inherent limitations of carrier-based looping
combustion of solid carbons has been proposed by Salatino and Senneca [1]. The
CarboLoop concept is based on the consideration of the well established role played by
surface oxides as metastable reaction intermediates in the combustion of solid carbons [311]. A simple semi-lumped oxidation mechanism [12] may be helpful to understand the
mechanistic basis of the CarboLoop:
1
 Cf  O2 

 C( O )
2
 C( O )  O 2 

CO 2  C(O)
(1)
 C( O ) 
 CO  C f
3
The very basis of the CarboLoop concept is to exploit the differences in the reaction rates of
the individual steps in scheme 1) as a function of temperature. At moderate temperature (say
300-400°C) oxygen chemisorption (step 1) progresses at an appreciable rate, whilst switchover (reaction 2) and surface oxide desorption (reaction 3) are modest, being kinetically
limited. At these temperatures it is possible to extensively uptake oxygen on carbon, with
minimal release of gaseous carbon oxides. The oxidized carbon can eventually be exposed
to higher temperatures, in the order of 700°C, so that oxygen complexes can be released as
CO/CO2, according to step 3 in scheme 1). CarboLoop pursues the simple idea that the
carbon-based fuel can act itself as oxygen carrier as it is cycled between an Oxidizer, (Fig.
1), where the fuel uptakes oxygen at moderate temperature, and a Desorber where
desorption of surface oxides is accomplished in inert atmosphere at higher temperature.
Looping of the carbon fuel between the oxidizer and the desorber enables stepwise
conversion of carbon over multiple cycles yielding an almost pure stream of CO2, ready for
sequestration, at the exhaust of the Desorber.
2
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
A successful proof-of-concept of CarboLoop has been given by Salatino and Senneca [1]
who simulated the cyclic fate of solid fuel by discontinuous experiments in a
thermogravimetric analyzer. Further assessment of the process is reported in refs [13-14],
including estimates of bed inventories per unit thermal throughput of the plant.
Full exploitation of the CarboLoop concept requires energy integration between the Oxidizer
and the Desorber, calling for detailed knowledge of the thermochemistry of the oxidation and
desorption stages. Detailed thermochemical studies of individual steps of carbon oxidation
are lacking and the subject has been recently addressed by the authors [15].
2
2.1
Experimental
Kinetics and extent of chemisorption/desorption
The CarboLoop concept has received experimental support from results of discontinuous
tests simulating alternated oxidation/desorption conditions. To this end cyclic thermal
analysis programs have been performed in a thermogravimetric analyzer Netzsch 409C
coupled with CO/CO2 IR analyser (HB URAS 3E).
Experiments have been carried out with samples of a char from a South African bituminous
coal. The sample was ground and sieved to obtain particles of average size 200µm.
Approximately 30mg of sample have been used for each test. Tests consisted of sequential
oxidation/desorption steps carried out as follows:
Step 1 (sample heating up and oxidation). The sample was initially loaded into the TGA and
heated in a flow of air (200ml/min) up to the desired oxidation temperature TO at the heating
rate of 50°C/min. The sample was then kept the flow of air for 30min.
Step 2 (desorption). The gas was swiched over from air to 100% nitrogen of chromatographic
grade while the temperature was stepwise raised to the desorption temperature TD, at the
heating rate of 50°C/min. The sample was held at TD for 30min before the temperature was
again decreased stepwise for a new oxidation cycle.
Step 3. (oxidation). Once the temperature reached pre-set oxidation temperature TO, the gas
feeding to the TG was again switched over from nitrogen to air the sample was oxidized for
30min.
Steps 2 and 3 were then iterated several times. During the experiments the weight loss as
well as the CO and CO2 concentrations at the exhaust were continuously monitored.
2.2
Thermochemistry of chemisorption/desorption
Experiments have been carried out on char from a South African bituminous coal, whose
properties are reported in Table 1. The char was prepared by pyrolyzing batches of coal for
5min in a lab-scale fluidized bed reactor at 800°C in nitrogen flow. Char samples were
ground and sieved in size range 100-150μm prior to further processing. A complete set of
experiments aimed at the characterization of the thermochemistry of oxygen chemisorption,
including Temperature programmed desorption, differential scanninc calorimetry and
peribolic calorimetry, have been carried out on samples of the char, and are reported in ref.
[15].
In the present paper only results of calorimetric measurement of the heat of combustion on
raw and on pre-oxidized char will be presented. The char sample was dehumidified in a
3
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
OX
DES
OX
DES
OX
stove at 110°C overnight. Approximately 2
g of the sample were loaded in a pyrex
reactor of 3cm ID and 5 cm height. The
reactor was continuously flushed with air
and heated up externally by electric ovens
to the temperature of 300°C. The
treatment was carried out for times
ranging between 5 to 600min. The sample
retrieved from the reactor after treatment
was analysed by means of a LECO CHN
2000 elemental analyser to determine its
carbon content and by means of a Parr
2000 isoperibolic calorimeter to determine
its heat of combustion (LHV).
DES
700
A
600
T, °C
500
400
300
200
100
0
5000
B
CO2, CO, ppm
4000
3000
2000
fractional carbon mass remaining, %
1000
0
100
80
60
40
3
Results and discussion
20
C
0
0
50
100
150
200
3.1
t, min
Kinetics
and
extent
chemisorption/desorption
of
Figure 2: Results of simulated looping combustion
of char from a bituminous coal.
The kinetics and extent of oxygen
uptake/carbon loss between the Oxidizer
and the Desorber have been assessed in
discontinuous experiments in which the batch of carbon was subjected to periodic cycling of
temperature and oxygen concentration in a thermogravimetric analyser. The value of TO was
set at 300°C. The desorption temperature TD varied between 600°C and 750°C.
Figure 2 reports typical results of a simulated looping combustion test. The time-temperature
history of the sample during the experiment is reported in Fig. 2A, whence step 1 and the
iterated cycles consisting of steps 2 (OX) and 3 (DES) are recognized. Figure 2B reports the
time series of CO2 and CO concentrations at the exhaust. Figure 2C reports the time-course
of the residual fractional combustible mass of the sample recorded along the experiment.
Inspection of Fig. 2C confirms that the weight change of the sample during the oxidation
(OX) stage, resulting from competition of oxygen chemisorption and desorption at the
oxidation temperature of 300°C, is very limited. On the contrary, a pronounced weight loss is
recorded during the desorption stages (DES) at 700°C, which increases to approach 10-20%
fractional carbon loss per cycle.
3.2
Thermochemistry of chemisorption/desorption
Figure 3 report results of calorimetric experiments. The rationale of this type of experiments
is based on the relationship:
H comb ,raw  H chemisorption  H comb ,preoxidized
(2)
according to which the heat of combustion of pre-oxidized samples equals the heat of
combustion of the raw char samples minus the heat involved in oxygen chemisorption at
moderate temperature. Figure 3A reports the LHV measured by the isoperibolic calorimeter
of the raw char and of char after chemisorption of oxygen at 300°C, as a function of the
duration of chemisorption. Notably the LHV per unit sample mass decreases from -28kJ/g of
4
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
the raw char to -24kJ/g after 5 minutes
of treatment. Eventually it decreases
almost linearly with the duration of
25000
treatment up to 600min, when it
reaches a value of -14kJ/g. It must be
20000
noted that prolonged heat treatment at
15000
300°C resulted not only in oxygen
uptake and formation of carbon-oxygen
10000
0
100
200
300
400
500
600
700
complexes, but also in some loss of
t,min
carbon
by
partial
oxidation.
40000
38000
Accordingly, Figure 3B reports the LHV
36000
expressed per unit of actual carbon
34000 chem
mass in the sample. The LHV of the
32000
raw char per unit mass of carbon is -36
30000
kJ/gC, slightly larger than the
28000
26000
theoretical heat of combustion of
30
40
50
60
70
80
90
graphite, consistently with the partly
C, w%
amorphous nature of carbon and with
the - limited - content of heteroatoms. A
very abrupt decay of the heat of
combustion is recorded after early
Figure 3. Heat of combustion of raw and pre- chemisorption of oxygen, entailing only
oxidized chars. Pre-oxidation performed in air at limited loss of carbon. The decay is
300°C. upper diagram) -LHV as a function of less pronouced upon further oxidation
duration of chemisorption ; lower diagram) -LHV at 300°C. A reduction of the heat of
combustion of nearly 4kJ/gC is
normalized by the actual char carbon content.
determined after prolonged oxidation of
char at 300°C, i.e. when extensive
treatments saturate the oxygen uptake
capacity of the char. According to eq. 2, this value could be regarded as the heat involved in
oxygen chemisorption at moderate temperature, though minor contributions from oxidation of
residual heteroatoms (e.g. H) present in the char after pyrolyis cannot be entirely ruled out.
The heat of reaction involved at this stage is very modest. The LHV of the fully oxygensaturated carbon approaches, instead, the much larger value of H=-31.5 kJ/gC. This must
be regarded, within a first order approximation, as the heat cumulatively associated with
processes that ultimately result into desorption of surface oxides at high temperature.
-LHV (per unit mass of C) ,J/gc
-LHV (per unit mass of char), J/g
30000
Results can be interpreted in the frame of the following reaction scheme:
1
 Cf  O2 

 C( O 2 ) #
2
 C( O 2 ) #  O 2 

C(O)  CO 2 , CO
(3)
3
 C( O 2 ) # 

C(O)  CO 2 , CO
4
 C( O ) 

C f  CO 2 , CO
According to this mechanism, mild chemisorption at moderate temperature results in nondissociative chemisorption of oxygen, with formation of a “metastable” surface oxide. The
metastable oxide may further undergo complex switch-over in the presence of oxygen
according to reaction 2. Alternatively, metastable oxides can be rearranged or isomerized
into more energetically favourable forms, with possible release of gaseous carbon oxides,
5
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
Figure 4. Outline of the energetics of the tentative oxidation/desorption mechanism.
according to the exothermic reaction 3. Finally, reaction 4 represents the endothermic
abstraction of CO and CO2 from the oxidized carbon. A qualitative outline of the energetics of
scheme 3 is reported in Figure 4.
4
Assessment of the CarboLoop process
A preliminary assessment of the CarboLoop process has been accomplished by computing
the relevant mass/energy fluxes associated with steady operation of the process. Reference
is made to the scheme in Fig. 1. Simplified material and energy balances around the two
basic units are hereby reported for the case when a pure (ash-free) carbon fuel is processed,
using air to oxidize the fuel in the O-reactor.
Material Balance on solid carbon:
FC,1  FC,0  FC, 2  FC,0  FC,1 1  X
(4)
where X is the fractional carbon loss in the D-reactor per single pass and FC are carbon
mass flow rates on an oxygen-free bases (i.e. without accounting for uptaken oxygen),
whence:
FC,1 
FC, 0
(5)
X
Gas flow rates issuing from the O and the D reactors are related to the carbon feed rate:
FG ,O 

32  32 79 
 1  e  1  FC, 0


12  28 21 

FG ,D 
;
44
 FC, 0
12
(6)
where e is the excess air factor. Simplified energy balances on O and D reactors, assuming
constant specific heats in the process temperature range and assuming ambient temperature
Ti as the reference temperature in expressions for enthalpy of gaseous and solid streams,
read:
q O  FC, 2 cTD  Ti   FC,1cTO  Ti   FG,O cG,O TO  Ti   FC,0  HO
6
(7)
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
q D  FC,1cTO  Ti   FC, 2 cTD  Ti   FG,D cG,D TD  Ti   FC,0  H D
(8)
The carbon loading required in the reactors is computed on the basis of the assigned values
of average carbon residence time in either reactor:
WO   O  FC,1
WD   D  FC, 2
(9)
Base case computations of the relevant process variables have been carried out by selecting
likely values for the model input variables. Reference was made to a 1kg/s carbon feed rate
(LHV=30MJ/kg) corresponding to about 30MW nominal thermal power input. X has been
set at 20%. Enthalpies of oxidation/desorption have been computed as: HD=f·LHV and
HO=(1-f)·LHV, taking f=0.85. Excess air was fixed at e=1.2. Reactor temperatures are
TO=600K and TD=1000K. The average carbon residence time  is set at 15min in both
reactors.
Results of the computations are illustrated in Fig. 1. Thermal balance on the two reactors
indicates that carbon oxidation in the Oxidizer is selfsustained, with moderate heat extraction
(qO=1.5MW), whereas most of the heat release takes place in the desorber (qO=22.4MW). Of
course this result is crucially influenced by the energetics of oxidation and desorption, and
more specifically by the balance between the enthalpy of mild oxygen chemisorption at
moderate temperature and that associated with reactions occurring at high temperature, like
the stabilization by surface oxide rearrangement of the metastable complexes, and,
eventually, desorption of surface oxides as CO and CO2.
5
Conclusions
Looping combustion of solid carbon may be accomplished, according to the CarboLoop
concept, by alternating carbon oxidation with air at moderate temperature and desorption of
carbon-oxygen complexes at higher temperature in a dual interconnected fluidized/entrained
bed reactor. A key issue is energy integration between the oxidation and desorption stages,
which calls for detailed characterization of the thermochemistry of each reaction step.
Experiments based on a combination of thermoanalytical and calorimetric techniques
indicate that mild chemisorption at low temperature is moderately exothermic. Reactions
occurring as preoxidized carbon is exposed at higher temperatures are also overall
exothermic. This is consistent with a mechanism that assumes that oxygen is uptaken in a
metastable form at moderate temperature, and that metastable surface oxides may be
stabilized at higher temperatures into more energetically favourable carbon-oxygen
complexes. Stable surface oxides are eventually desorbed at high temperature into CO and
CO2.
A simple model of the CarboLoop process has been developed, based on simplified material
and energy balances on the oxidizer and the desorber. Model computations have been
performed to determine key process variables, like the solid recirculation rate, the carbon
loading in both reactors, the thermal throughput at the oxidizer and at the desorber.
6
Acknowledgments
This study has been carried out in the framework of the MISE-CNR Agreement “Tecnologie
innovative per migliorare le prestazioni ambientali delle centrali a polverino di carbone”. The
support of Ms Letizia Romano is gratefully acknowledged.
7
2nd International Conference on Chemical Looping, 26-28 September 2012, Darmstadt, Germany
References
[1]
Salatino P., Senneca O. (2009), Ind. Eng. Chem. Res. 48(1), 102.
[2]
Adanez J., Abad A., Garcia-Labiano F., Gayan P., de Diego L.F. (2012), Progr. Energy
Combust. Sci. 38, 215
[3]
Zhuang Q.L., Kyotani T., Tomita A. (1995), Energy & Fuels 9, 630.
[4]
Pan Z., Yang R.T. (1992), Ind. & Eng. Chem. Res. 31, 2675.
[5]
Lear A.E., Brown T.C., Haynes B.S. (1991), Proc. Combust. Inst. 23, 1191.
[6]
Brown T.C., Lear A.E., Haynes B.S. (1992), Proc. Combust. Inst. 24, 1199.
[7]
Du Z., Sarofim A.F., Longwell J.P. (1991), Energy & Fuels 5, 214.
[8]
Feng B., Bhatia S.K. (2000), Chem. Eng. Sci. 55, 6187.
[9]
B.S. Haynes, Combust. Flame 126 (2001) 1421.
[10] Senneca O., Salatino P., Masi S. (2004), Proc. Combust. Inst. 30, 2223.
[11] Senneca O., Salatino P., Menghini D. (2007), Proc. Combust. Inst. 31, 1889.
[12] Hurt R.H., Calo J.M. (2001), Combust. Flame 125, 1138.
[13] Salatino P., Senneca O. (2010) Int. Patent application WO2010/026259A2.
[14] Salatino P., Senneca O., Cortese L. (2010), CarboLoop: a novel concept of looping
combustion of carbon, Proc. of the 1st International Conference on Chemical Looping,
Lyon (FR).
[15] Salatino P., Senneca O., Cortese L. (2012), Assessment of the thermochemistry of
oxygen chemisorption and surface oxide desorption during looping combustion of coal
char, Proc. Comb. Inst. 34, in press.
8