sintering and reactivity of cac o3 -based sorbents for in situ c o2 capture in fluidized beds under...

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Sintering and Reactivity of CaCO3-Based Sorbents forIn Situ CO2 Capture in Fluidized Beds under Realistic

Calcination ConditionsDennis Y. Lu1; Robin W. Hughes2; Edward J. Anthony3; and Vasilije Manovic4

Abstract: Sintering during calcination/carbonation may introduce substantial economic penalties for a CO2 looping cycle usinglimestone/dolomite-derived sorbents. Here, cyclic carbonation and calcination reactions were investigated for CO2 capture under fluidizedbed combustion �FBC� conditions. The cyclic carbonation characteristics of CaCO3-derived sorbents were compared at various calcinationtemperatures �700–925°C� and different gas stream compositions: pure N2 and a realistic calciner environment where high concentrationsof CO2�80–90% �and the presence of SO2� are expected. The conditions during carbonation employed here were 700°C and 15% CO2

in N2 and 0.18% or 0.50% SO2 in selected tests, i.e., typically expected for a carbonator. Up to 20 calcination/carbonation cycles wereconducted using a thermogravimetric analyzer �TGA� apparatus. Three Canadian limestones were tested: Kelly Rock, Havelock, andCadomin, using a prescreened particle size range of 400–650 �m. In addition, calcined Kelly Rock and Cadomin samples were hydratedby steam and examined. Sorbent reactivity was reduced whenever SO2 was introduced to either the calcining or carbonation streams. Themulticyclic capture capacity of CaO for CO2 was substantially reduced at high concentrations of CO2 during the sorbent regenerationprocess and carbonation conversion of the Kelly Rock sample obtained after 20 cycles was only 10.5%. Hydrated sorbents performedbetter for CO2 capture, but also showed significant deterioration following calcination in high CO2 gas streams. This indicates that highCO2 and SO2 levels in the gas stream lead to lower CaO conversion because of enhanced sintering and irreversible formation of CaSO4.Such effects can be reduced by separating sulfation and carbonation and by introducing steam to avoid extremely high CO2 atmospheres,albeit at a higher cost and/or increased engineering complexity.

DOI: 10.1061/�ASCE�EE.1943-7870.0000079

CE Database subject headings: Fluidized beds; Calcination; Combustion; Gas; Limestone.

Introduction

The following reversible reaction between CaO and CO2 may findapplications in a high-temperature process to control CO2 emis-sions from advanced power generation in a CO2 looping cycle�Abanades and Alvarez 2003; Abanades et al. 2004; Gupta andFan 2002�:

CaO�s� + CO2�g� ⇔ CaCO3�s� �1�

Circulating fluidized bed �CFB� systems are suitable for loopingcycle technology because they excel in transferring large amountsof solids between reactors and, hence, from one chemical envi-ronment to another. A good reaction between solids and gas

1CanmetENERGY, Natural Resources Canada, 1 Haanel Dr., OttawaON, Canada K1A 1M1.


3CanmetENERGY, Natural Resources Canada, 1 Haanel Dr., OttawaON, Canada K1A 1M1 �corresponding author�. E-mail: [email protected]


Note. This manuscript was submitted on February 28, 2008; approvedon February 2, 2009; published online on May 15, 2009. Discussionperiod open until November 1, 2009; separate discussions must be sub-mitted for individual papers. This paper is part of the Journal of Envi-ronmental Engineering, Vol. 135, No. 6, June 1, 2009. ©ASCE, ISSN
0733-9372/2009/6-404–410/$25.00.
404 / JOURNAL OF ENVIRONMENTAL ENGINEERING © ASCE / JUNE 200

J. Environ. Eng. 2009

streams is also greatly enhanced by the excellent solid mixing inCFBs, which in turn maximizes the mass/heat transfer and, hence,reaction rates �Abanades et al. 2004; Shimizu et al. 1999; Salva-dor et al. 2003�. CFBs using a limestone/dolomite sorbent canhave another advantage—simultaneous sulfur removal via the fol-lowing sulfation reaction when burning sulfur-containing solidfossil fuels:

CaO�s� + SO2�g� + 12O2�g� → CaSO4�s� �2�

However, CaO-derived sorbents from natural sources fail toachieve complete reconversion to CaCO3 and instead show arapid falloff in reversibility for Reaction �1� �Salvador et al. 2003;Hughes et al. 2004; Fennell et al. 2007�. A decrease in the mi-croporosity of the sorbents appears to be the major factor causingthe decay of sorbent activity and conversion, i.e., pore sinteringassociated with the calcination process, which depends mainly onfinal calcination temperature, heating rate, and duration. Wangand Anthony �2005� recently studied the decay behavior of CaO-derived sorbents and found that an empirical model curve describ-ing the process was almost identical to the deactivation ofcatalysts due to sintering. This supports the idea that the decay ofCO2 carrying capacity is attributable to sintering of the sorbentduring the carbonation and/or sorbent regeneration process.

Sintering effects are strongly dependent on temperature, andtend to cause grain size to increase, increasing pore size and re-ducing surface area by causing mass transport at the atomic scale�Borgwardt et al. 1986; Borgwardt 1989a�. Solid surface curva-
ture associated with high-temperature capillary forces provides
9

.135:404-410.

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

the driving force for sintering. However, there is no consensus onhow to calculate the degree of sintering for any given situation.Louisiana State University �Silaban 1993; Silaban and Harrison1995� investigated the effect of the typical fluidized bed combus-tion �FBC� operating temperature range of 700–900°C at atmo-spheric pressure on CO2 carrying capacity of the sorbent used inthe multicycle process described by Reaction �1�, and concludedthat the calcination conditions ought to be as mild as possible toavoid sintering and improve carbonation capacity. Earlier studiesalso showed that solid impurities in sorbents and the presence ofgases such as CO2 and H2O could enhance sorbent sintering�Dobner et al. 1977; Borgwardt 1989b�. Borgwardt �1989b� foundthat sintering could be promoted by the addition of water vaporand carbon dioxide to simulated combustion flue gases. Each gasstrongly catalyzed the sintering process, whereas their combinedeffects were even more severe. Porosity reduction was also accel-erated by the presence of H2O or CO2 in the sintering atmosphereand porosity reduction followed the Coble logarithmic law forsintering at 800–1000°C for the onset of particle shrinkage�Borgwardt 1989b�. Although multiple sintering mechanisms arepossible in the presence of CO2 and/or H2O, the empirical modelcorrelates isothermal surface area reduction as a function of timeover the temperature range 380–1150°C and partial pressuresfrom 39 Pa to 15 kPa for the gas associated with sintering. Dob-ner et al. �1977� studied the effects of various reaction parameters�i.e., temperature, pressure, and gas composition� on the carbon-ation rate and found increasing temperature and partial pressureof CO2 for calcination gave lower carbonation activities and ac-celerated the loss of solid reactivity with increasing cycle number.Dobner et al. �1977� summarized the effect of gas composition onsintering as: SO2�CO2�H2O�O2�air�N2.

Calcination can be achieved by burning gas, liquid, or low-ashsolid fuels in pure oxygen, producing a highly pure CO2 stream��90% � suitable for direct sequestration �Hughes et al. 2005�.Such CO2 levels are much higher than used in previous studiesprimarily focusing on conditions appropriate to sulfation. Prelimi-nary results from CETC’s �CANMET Energy Technology Centre�pilot-scale miniature CFB system have demonstrated that lime-stone can be used successfully in a cyclic calcination/carbonationprocess under realistic conditions; however, lower CO2 capturewas observed compared to thermogravimetric analyzer �TGA�tests using the same sorbents �Salvador et al. 2003; Hughes et al.2004�. Here we hypothesize that CO2 from sorbent calcination ina bubbling mode calciner will be released to the emulsion phase.Thus, sorbent particles will be surrounded by a high-CO2 stream,which could be one of the reasons for lower CO2 capture. Inparticular, the high-CO2 environment during the calcination reac-tion enhances sintering, thereby negatively impacting sorbent re-activity.

The use of petroleum coke for heating the circulating fluidizedbed calciner, which is an ideal low-ash fuel with excellent calo-rific value, necessarily implies the presence of high-SO2 levels.Sun et al. �2006� recently studied the effect of SO2 addition onsorbent reversibility for absorbing CO2 under pressurized and at-mospheric FBC conditions, and demonstrated that SO2 signifi-cantly lowers the sorbent reversibility and impairs cyclical CO2

capture. However, this work was conducted using a N2 stream forthe calcination step, which does not represent a realistic calcina-tion environment. Hence, we attempt to examine the effects of theCO2 and SO2 streams during the calcination process on the CO2

carrying capacity of the sorbent using a TGA.

JOURNA


Experimental Setup

Thermogravimetric Analyzer Setup

The calcination/carbonation cycles using limestone sorbents wereconducted in a TGA apparatus, described elsewhere �Salvador etal. 2003; Hughes et al. 2004�. The TGA consists of an electronicbalance �Cahn 1100, ThermoFisher Scientific, Newington, USA�,vertical furnace, reactor tube, carrier gas system, and computer-ized data acquisition system. The reactor tube is made of Inconel600 alloy �Special Metals Corp., New Hartford, N.Y.� and has aninner diameter of 24 mm and a height of 900 mm. It can be un-screwed from the TGA, thus uncovering the platinum sampleholder �10 mm in diameter, 1.5 mm in depth�. A small amount ofthe sample �20–30 mg� was preloaded into the holder prior to thetests.

The reactor was heated externally by means of an electric fur-nace that can be placed around the reactor tube and then removedrapidly to cool down the reactor to the lower temperature. A ther-mocouple was used to measure the reactor temperature near thereaction site �i.e., 10 mm below the sample holder�. The furnacewas heated from room temperature to the desired calcination tem-perature at a heating rate of 10°C /min �programmed by an elec-tronic temperature controller�. The temperature and sample massversus time data for the sample were recorded at 5-s intervalsuntil termination of the run, using a Keithley Instruments Inc.�Solon, Ohio� 2700 data acquisition system and Keithley Xlinxsoftware.

Experimental Procedure

In this work a high-CO2 �80 or 90% by volume concentration�sweep gas was used in the calcination step. A temperature inexcess of 890°C was required to ensure calcination. However,preliminary runs during this study indicated that the observedcalcination proceeded extremely slowly when the temperaturewas less than �915°C, as the backreaction �carbonation� pro-ceeds at a finite rate. Once the temperature was raised to 920°C,sorbent calcination occurred at an acceptable rate. This agreeswith observations from CETC’s miniature pilot-scale FBC work,which indicated that calcination of the limestone bed �5–7 kgmass� was excessively slow at 800–850°C, when the bed wasfluidized with air and electrically heated to avoid the presence ofCO2, especially during the first calcination cycle �Salvador et al.2003�. Elevating the temperature strongly accelerated the calcina-tion process; hence a sorbent regeneration temperature of 925°Cwas chosen for the TGA work to guarantee complete calcinationin a relatively short period, i.e., a few minutes. Carbonation testswere conducted at 700°C using 15% CO2 �N2 balance� at atmo-spheric pressure. Earlier work �Borgwardt 1989b; Curran et al.1967� indicated that the carbonation process occurred in twophases: a rapid reaction phase �kinetically controlled� taking up to4–5 min, depending on conditions, followed by a slower reactionthat can continue for more than 10–20 h for the last 5% of theconversion. As it is the rapid phase that is of practical importance,the carbonation process was carried out for 30 min and CO2 cap-ture capacity of the sample was determined by the weight gainover that period.

Limestones tested were selected from across Canada, such asHavelock from New Brunswick, Cadomin from Alberta, andKelly Rock from Nova Scotia. Their chemical compositions aregiven in Table 1. The sorbent samples were prescreened within a
particle size range of 400–650 �m. Samples were loaded into the
L OF ENVIRONMENTAL ENGINEERING © ASCE / JUNE 2009 / 405

.135:404-410.

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

reactor prior to starting each run. Nitrogen was used as a refer-ence carrier gas and a digital mass flow controller �Matheson GasProducts� regulated the flow rate at 100 mL /min for the carrierstream and 60 mL /min for the purge stream. Two levels of SO2

�5,000 and 1,800 parts per million �ppm�� were used in selectedtests during both calcination and carbonation reactions for com-parison purposes.

Results and Discussion

Calcination up to 925°C did not significantly degrade sorbentreactivity associated with the carbonation/calcination cycling be-yond what was normally expected. It should be noted that a pos-sible limitation of this work is that the TGA runs were conductedat the identical heating rate �10°C /min� regardless of the finaltemperature set point. Borgwardt et al. �1986� and Borgwardt�1989a� found that heating rate, rather than final temperature, wasa more critical factor in terms of sorbent sintering. In terms ofcarbonation, temperatures below 650°C led to reaction rates thatwere too low to achieve high-CO2 removal in a small pilot-scaleFBC operating in the bubbling bed mode �Hughes et al. 2004�.For a synthetic flue gas containing 15% CO2 at atmospheric pres-sure, the carbonation temperature was limited to �780°C, andthe best CO2 capture capacity was achieved in the carbonationtemperature range of 700–740°C, which is in reasonable agree-ment with observations by Silaban �1993� and Silaban and Harri-son �1995�.

Here calcination was conducted with high concentrations ofCO2 �80 or 90% in the carrying gas�; the calcination temperaturewas set at 925°C and the carbonation occurred at 700°C. Fig. 1shows typical sorbent decomposition curves. Similar to the caseof a typical limestone decomposition process in pure inert gasstreams �such as N2�, a slight sample weight loss was seen withincreasing temperature over the first 20 min �see Fig. 1�a��. How-ever, instead of sample decomposition occurring as temperaturesrise above 500°C, as in the case of a pure N2 stream �green line�,CO2 release started at a relatively high temperature of �900°C in

Table 1. Limestone Sorbent Composition

Component Kelly Rock Havelock Cadomin

Al2O3 1.54 0.34 0.25

BaO 0.18 0.03 0.03

CaO 51.74 54.1 51.76

CO2 — — —

Fe2O3 0.36 0.3 0.3

K2O 0.36 0.06 0.12

MgO 0.58 0.29 2.18

MnO 0.16 0.07 0.01

Na2O 0.07 0.2 0.2

NiO — 0.01 0.01

P2O5 0 0.02 0.02

SiO2 5.31 1.9 2.13

SO3 0.98 0.46 0.32

SrO 0.04 0.02 0.03

TiO2 0.08 0.07 0.04

V2O5 — 0.02 0.02

LOF 43.14 42.99 43.28

SUM 102.55 99.77 101.19

the presence of 90% CO2 in the calcination stream �red line�. Fig.



1�b� shows the decomposition curves when hydrated lime �H-lime� was tested at a similar heating rate in the TGA. Again,rather than a rapid weight loss associated with Ca�OH�2 decom-position once the temperature increased to 400°C in a pure N2

stream �green line�, carbonation occurred in the presence of thehigh-concentration CO2 stream, and sample weight increasedstarting at a temperature of �350°C, until the reactor temperatureexceeded 900°C, when CO2 was released �red line�. In this case,carbonation appeared to set in at 350°C, i.e., below the tempera-ture at which decomposition of Ca�OH�2 starts when under pureN2, indicating that it was not CaO but Ca�OH�2 that reacted withCO2.

Fig. 2 shows the change in CO2 capture capacity during mul-tiple cycles of calcination and carbonation using untreated Ca-domin under a high-CO2 calcination environment. Once thecalcination had been completed at a high temperature �925°C�,the high-CO2 carrier gas was replaced by a pure N2 stream untilthe temperature dropped and stabilized at the carbonation tem-perature �700°C�, whereupon low-CO2 carbonation carrier gaswas introduced. After 30 min of carbonation, the carrier gas wasswitched back to the high-CO2 stream for calcination; simulta-neously the reactor was heated to 925°C. The peak in the sample

Fig. 1. Comparison of Cadomin limestone �25 mg� calcinationcurves with/without a high-concentration CO2 �90%� stream at car-rier gas flow rate of 100 mL /min: �a� original limestone; �b� hydratedlime

mass change increase at the beginning of each calcination period

9

.135:404-410.

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

was due to the introduction of a high-CO2 stream at low tempera-ture �700°C�. This was not observed when the calcination streamcontained lower concentrations of CO2. The curve also indicatesthat CO2 is released rapidly, once the temperature is increased tothe set point of 925°C, which is well in excess of the temperaturerequired to favour calcination at atmospheric pressure.

Table 2 compares the CO2 capture capacity for sorbents testedunder the same carbonation conditions, with and without high-concentration CO2 streams during regeneration. The three sor-bents showed slight variations in terms of CO2 capture capacity,particularly over the first few cycles; the differences appeared tobe negligible after 3 cycles. With CO2 in the calcination stream,the decay of sorbent reactivity occurred much faster than whenCO2 was absent. It can be seen in carbonation during the firstcycle that CO2 capture capacity only reached 48–58%, which istypical of the capture level after 5–6 cycles for the same sorbentcalcined in a pure N2 atmosphere. For calcination with high-CO2

atmospheres, the sorbent capture capacity after 4 cycles was de-pressed to around 25%, similar to the levels seen after15–20 cycles, without CO2 present in the calcination step.

To investigate the situation where a high-sulfur feedstock,such as petroleum coke is employed to provide heat for both thesorbent regenerator and combustor, SO2 was introduced underdifferent conditions. It was added with CO2 to simulate a normalcombustion flue gas in the carbonation process, and also addedinto the regenerator with or without a high-concentration CO2

environment. Fig. 3 compares the typical decay of sorbent con-versions without �Fig. 3�a�� and with SO2 �Fig. 3�b�� present inthe carbonation. The curves in Fig. 3 indicate that the sorbent

Fig. 2. CO2 looping cycles for Cadomin sample �25 mg� usinghigh-CO2 carrier gas �CO2: 90%, and N2: balance� in the calcinationand low-CO2 stream �CO2: 15% and N2: balance� in the carbonation,carrier gas flow rate=100 mL /min

Table 2. Carbonation Conversions �at 700°C in 15% CO2, N2 Balance�of Various Limestones with High-CO2 Concentrations Present duringCalcination

Sorbent Calcination

Conversiona

X1 X2 X3 X4 X5

Cadomin 100% N2, 700°C 0.75 0.68 0.62 0.58 0.52

Cadomin 90% CO2 in N2, 925°C 0.54 0.31 0.27 0.23 —

Havelock 90% CO2 in N2, 925°C 0.48 0.34 0.29 0.25 0.19

Kelly Rock 90% CO2 in N2, 925°C 0.58 0.39 0.3 0.26 —a
Xn=conversion in cycle n.
JOURNA


CO2 capture capacity was dramatically reduced when SO2 wasinvolved in the carbonation stream, and the carbonation conver-sion dropped from 0.62 to 0.27 after 3 cycles �Tables 2 and 3�. Inthe test with SO2 introduction, the base sample weight signifi-cantly increased after each cycle of sorbent regeneration because

Fig. 3. Comparison of Cadomin CO2 looping cycles with/withoutadded SO2: �a� calcination in 100% N2, 700°C, carbonation in 15%CO2, N2 balance, 700°C; �b� calcination in 100% N2, 700°C, car-bonation in 15% CO2+0.5% SO2, N2 balance, 700°C

Table 3. Comparison of Cadomin Carbonation and Sulfation Conversionin Cyclic Process

Calcination Carbonation

ConversionaRatio of

conversionsb

X1 X2 X3 X1,2 X1,3

N2, 700°C 15% CO2 and0.5% SO2 inN2, 700°C

Carbonation 0.48 0.45 0.27 0.94 0.56

Sulfation 0.25 0.17 0.07 0.68 0.28

aXn=conversion in cycle n.b
Xn ,m=ratio between conversions in cycles m and n.

.135:404-410.

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

of the accumulation of CaSO4, which does not decompose attypical calcination temperatures.

The TGA curve indicates that carbonation and sulfation reac-tions compete with each other �Fig. 3�b��. Table 3 shows thecontribution of carbonation and sulfation after each cycle. Notonly did the carbonation conversion decrease with increasingcycle number, but the sulfur capture decreased as well. The lower

Fig. 4. CO2 looping cycles for Cadomin sample �25 mg� usinghigh-CO2 carrier gas �CO2: 90%, and N2: balance� in the calcinationand low-CO2 stream �CO2: 15%, SO2: 0.18%, and N2: balance� inthe carbonation, carrier gas flow rate=100 mL /min

Fig. 5. CO2 looping cycles for Cadomin sample �25 mg� usinghigh-CO2 carrier gas �CO2: 80%, SO2: 0.18% and N2: balance� in thecalcination and low-CO2 stream �CO2: 15%, SO2: 0.18%, and N2:balance� in the carbonation, carrier gas flow rate=100 mL /min

Table 4. Carbonation Conversion of Cadomin with/without SO2 Present


N2, 700°C 15% CO2 in N2, 7

N2, 700°C 15% CO2 and 0.5% SO2

90% CO2 in N2, 925°C 15% CO2 in N2, 7

90% CO2 in N2, 925°C 15% CO2 and 0.18% SO2

80% CO2 and 0.18% SO2 in N2, 925°C 15% CO2 and 0.18% SO2aXn=conversion in cycle n.b
Xn ,m=ratio between conversions in cycles m and n.


sulfate conversions, compared to carbonate conversion, are likelyattributable to a relatively lower SO2 concentration �0.5%� thanCO2 �15%� in the carbonation/sulfation stream.

Fig. 4 shows the cyclic sorbent performance when the calci-nation occurred in 90% CO2 at 925°C, with 0.18% SO2 added to15% CO2 in the carbonation stream at 700°C. Fig. 5 shows theresults of introducing 0.18% SO2 into both the carbonator with15% CO2 and sorbent regenerator with 80% CO2. Compared tothe results shown in Fig. 3�b�, sorbent regeneration is much fasterat the higher temperature; however, sorbent CO2 absorption ca-pacity degenerates as process conditions become more severe.Table 4 compares results from the Cadomin TGA tests associatedwith and without the presence of SO2. As expected, CO2 conver-sion capacity clearly decreased with higher SO2. These resultssupport the importance of avoiding high-SO2 concentrations inthe flue gas when using the CO2 looping cycle with lime-basedsorbents.

It is also of interest to observe �Fig. 6� that the sudden increasein the sample weight in the period of transition between carbon-ation and calcination seemed to be eliminated when SO2 waspresent in the calcination carrier gas. This can be explained by theoccurrence of sulfation on the surface of the sorbent particlesduring the initial calcination, preventing carbonation on the sor-bent surface, which is clearly represented as the sharp spike whenthe process switched from carbonation to calcination using thehigh-CO2 stream without SO2 at lower temperatures �Figs. 2 and4�.

Sorbent hydration, to improve performance when high-sulfurfuels are employed to achieve calcination, was also explored. Thesamples chosen were commercial quicklime �designated as

Conversiona Ratio of conversionsb

X1 X3 X5 X1,2 X1,3

0.70 0.6 0.45 0.88 0.86

700°C 0.58 0.32 — 0.72 0.56

0.54 0.27 — 0.57 0.49

, 700°C 0.33 0.13 — 0.54 0.41

, 700°C 0.44 0.13 — 0.50 0.30

Fig. 6. CO2 looping cycles for hydrated Cadomin sample �25 mg�using high-CO2 carrier gas �CO2: 90%, and N2: balance� in the cal-cination and low-CO2 stream �CO2: 15%, SO2: 0.18%, and N2: bal-ance� in the carbonation, carrier gas flow rate=100 mL /min

00°C

in N2,

00°C

in N2

in N2

9

.135:404-410.

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

H-lime� and hydrated limes prepared from Cadomin and KellyRock limestones in our laboratory. Details of the hydration pro-cedure can be found elsewhere �Wu et al. 2004; Anthony et al.2005�. Some of the improvement seen previously �Hughes et al.2004� clearly occurs because previous tests avoided the severeconditions studied here; however, in general, hydrated sorbentstill showed much better results in terms of carbonation conver-sion and maintained higher sorbent reversibility after a number ofcycles, particularly in the absence of SO2. Table 5 presents resultsfor commercial hydrated lime and hydrated lime obtained fromCadomin. The presence of high concentrations of CO2 in the cal-cination cycle lowered the CO2 capture capacity dramatically, butsorbent reactivity remained excellent for at least 5 cycles withoutelevated CO2. However, once SO2 was added to the carbonationstream, both the sorbent reactivity with increasing cycles and thecapacity for CO2 capture significantly decreased �see Fig. 6�.

The influence of realistic conditions, as expected for CO2

looping cycle FBC systems, was examined here, typically for5 cycles. During this period, sorbent capture capacity fell rapidlyand this tendency may be expected for longer series of cycles.This was confirmed with original and hydrated Kelly Rocksamples �Figs. 7 and 8�. The experiments were performed in 90%CO2 in the calcination stage, and it can be seen that carbonationconversion in the 21st cycle was only 10.5% for the originalsample. Unfortunately, the hydrated sorbent achieved only 2%higher conversion, leading to the conclusion that sorbent activatedby hydration also loses activity much faster under realistic condi-tions than under those typically investigated. This observation isin agreement with an earlier study by Borgwardt �1989b�, who

Table 5. Conversion of Hydrated Cadomin and Commercial Hydrated L


N2, 700°C 15% CO2 in N2, 700°C

90% CO2 in N2, 925°C 15% CO2 in N2, 700°C

90% CO2 in N2, 925°C �H-lime� 15% CO2 and 0.18% SO2 in NaXn=conversion rate �%�, n=cycle number.bXn ,m=conversion ratio between cycle numbers of m and n.

0

10

20

30

40

50

60

70

80

90

100

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time, min

Carbonationconversion,%

0

100

200

300

400

500

600

700

800

900

1000

Temperature,o C

10.5%

Fig. 7. CO2 looping cycles for Kelly Rock �25 mg� using high-CO2

carrier gas �CO2: 90%, and N2: balance� in the calcination at 925°Cand low-CO2 stream �CO2: 15% and N2: balance� in the carbonationat 700°C, carrier gas flow rate=100 mL /min

JOURNA


found hydrated-CaO sintered faster than calcined-CaO. These re-sults obtained for 20-cycle tests additionally confirm those for5-cycle tests presented in this paper.

It seems reasonable to suppose that the main contribution tothe loss of sorbent capture capacity is carbonation that continuesto occur �above dashed lines in Figs. 7 and 8�, simultaneously thetemperature is being raised from that for carbonation to the levelnecessary for calcination under high-CO2 concentrations. Duringthis period, formation of CaCO3 �and after that it begins its de-composition� occurs at higher temperatures, which enhance bulkmass transfer and sintering. Our experimental study on CO2 cap-ture capacity �Manovic and Anthony 2008� showed thatformation/decomposition of CaCO3 is the critical step for sinter-ing and that this is enhanced at higher temperatures. Further, itmay be expected that elimination of the carbonation peak at thebeginning of the calcination stage may help to reduce the loss ofcapture capacity. For example, this can be achieved by increasingthe rate of rise of sample temperature �to that needed for calcina-tion�, i.e., shorter exposure to carbonation conditions at the be-ginning of the calcination stage. Experimentally, in TGA systemsthis means increasing the heating rate at the beginning of thecalcination stage to the upper equipment performance limit. Prac-tically, this is more easily achieved in FBC systems as sorbentcirculates between carbonator and calciner and much more rapidheat transfer is experienced than that in a TGA. In other words, itmay be expected that real FBC systems will be subjected to aslower drop in sorbent capture capacity than what was obtained inthis study using TGA equipment.

ith SO2 Present in Carbonation Gas

Conversiona Conversion ratiob

X1 X3 X5 X1,2 X1,3

0.76 0.72 0.69 0.97 0.95

0.45 0.41 — 0.96 0.92

°C 0.41 0.30 — 0.73 0.73

0

10

20

30

40

50

60

70

80

90

100

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time, min

Carbonationconversion,%

0

100

200

300

400

500

600

700

800

900

1000

Temperature,o C

12.5%

Fig. 8. CO2 looping cycles for hydrated Kelly Rock �25 mg� usinghigh-CO2 carrier gas �CO2: 90%, and N2: balance� in the calcinationat 925°C and low-CO2 stream �CO2: 15% and N2: balance� in thecarbonation at 700°C, carrier gas flow rate=100 mL /min

ime w

2, 700


.135:404-410.

Dow

nloa

ded

from

asc

elib

rary

.org

by

KA

NSA

S ST

AT

E U

NIV

LIB

RA

RIE

S on

05/

28/1

4. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Conclusions

High concentrations of CO2 in the sorbent calcination environ-ment lead to significant sintering of sorbent particles, and conse-quent lowering of sorbent reactivity for CO2 conversion capacityand sorbent reversibility. This phenomenon is inevitable for theCO2 looping FBC process explored, where high-CO2 concentra-tions ��90–95% � are produced in the sorbent regenerator. Thedata obtained support the contention that cycling studies done onlimestones should employ realistic calcination conditions in termsof likely off-gas compositions. Also, as noted previously withdifferent limestones, sulfate formation blocks active pores/surfacearea, resulting in lower CO2 capture and causes a faster decline insorbent reversibility, although hydration may somewhat reducethese effects, suggesting that the fuel used in the regeneratorshould ideally contain low sulfur concentrations. High CO2 andSO2 in the CaO /CaCO3 formation/decomposition gas stream willlead to enhanced sorbent sintering and irreversible CaSO4 forma-tion, consequently lowering CO2 conversion. These effects can beexpected to be reduced in practice by separating sulfation andcarbonation in the two stages of the carbonator and introducingsteam to avoid extremely high-CO2 atmospheres.

References

Abanades, J. C., and Alvarez, D. �2003�. “Conversion limits in the reac-tion of CO2 with lime.” Energy Fuels, 17�2�, 308–315.

Abanades, J. C., Anthony, E. J., Lu, Y., Salvador, C., and Alvarez, D.�2004�. “Capture of CO2 from combustion gases in a fluidized bed ofCaO.” AIChE J., 50�7�, 1614–1622.

Anthony, E. J., Jia, L., and Wu, Y. �2005�. “CFBC ash hydration studies.”Fuel, 84�11�, 1393–1397.

Borgwardt, R. H. �1989a�. “Sintering of nascent calcium oxide.” Chem.Eng. Sci., 44�1�, 53–60.

Borgwardt, R. H. �1989b�. “Calcium oxide sintering in atmospheres con-taining water and carbon dioxide.” Ind. Eng. Chem. Res., 28�4�, 493–500.

Borgwardt, R. H., Roache, N. F., and Bruce, K. R. �1986�. “Method forvariation of grain size in studies of gas-solid reactions involving cal-cium oxide.” Ind. Eng. Chem. Fundam., 25�1�, 165–169.

Curran, G. P., Fink, C. E., and Gorin, E. �1967�. “Carbon dioxide acceptorgasification process-studies of acceptor properties.” Adv. Chem. Ser.,



69�10�, 141–165.Dobner, S., Sterns, L., Graff, R. A., and Squires, A. M. �1977�. “Cyclic

calcination and recarbonation of calcined dolomite.” Ind. Eng. Chem.Process Des. Dev., 16�4�, 479–486.

Fennell, P. S., Pacciani, R. P., Dennis, J. S., Davidson, J. F., and Hayhurst,A. N. �2007�. “The effects of repeated cycles of calcination and car-bonation on a variety of different limestones, as measured in a hotfluidized bed of sand.” Energy Fuels, 21�4�, 2072–2081.

Gupta, H., and Fan, L.-S. �2002�. “Carbonation-calcination cycle usinghigh reactivity calcium oxide for carbon dioxide separation from fluegas.” Ind. Eng. Chem. Res., 41�16�, 4035–4042.

Hughes, R., Lu, D., Anthony, E. J., and Macchi, A. �2005�. “Design,process simulation and construction of an atmospheric dual fluidizedbed combustion system for in situ CO2 capture using high-temperature sorbents.” Fuel Process. Technol., 86�14–15�, 1523–1531.

Hughes, R., Lu, D., Anthony, E. J., and Wu, Y. �2004�. “Improved long-term conversion of limestone-derived sorbents for in situ capture ofCO2 in a fluidized bed combustor.” Ind. Eng. Chem. Res., 43�18�,5529–5539.

Manovic, V., and Anthony, E. J. �2008�. “Parametric study on CO2 cap-ture capacity of CaO-based sorbents in looping cycles.” Energy Fuels,22�3�, 1851–1857.

Salvador, C., Lu, D., Anthony, E. J., and Abanades, J. C. �2003�. “En-hancement of CaO for CO2 capture in an FBC environment.” Chem.Eng. J., 96�1–3�, 187–195.

Shimizu, T., Hirama, T., Hosoda, H., Kitano, K., Inagaki, M., and Tejima,K. �1999�. “A twin fluid-bed reactor for removal of CO2 from com-bustion processes.” Chem. Eng. Res. Des., 77�1�, 62–68.

Silaban, A. �1993�. “High-temperature high-pressure CO2 removal fromcoal gas.” Ph.D. thesis, Dept. of Chemical Engineering, LouisianaState Univ., Baton Rouge, La.

Silaban, A., and Harrison, P. �1995�. “High-temperature capture of carbondioxide: Characteristics of the reversible reaction between CaO�s� andCO2�g�.” Chem. Eng. Commun., 137�1�, 177–190.

Sun, P., Grace, J. R., Lim, C. J., and Anthony, E. J. �2006�. “Removal ofCO2 by calcium-based sorbents in the presence of SO2.” EnergyFuels, 21�1�, 163–170.

Wang, J., and Anthony, E. J. �2005�. “On the decay behavior of the CO2

absorption capacity of CaO-based sorbents.” Ind. Eng. Chem. Res.,44�3�, 627–629.

Wu, Y., Anthony, E. J., and Jia, L. �2004�. “Steam hydration of CFBC ashand the effect of hydration conditions on reactivation.” Fuel, 83�10�,1357–1370.

9

.135:404-410.

sintering and reactivity of cac o3 -based sorbents for in situ c o2 capture in fluidized beds under...

Documents