Chemical Engineering Science 134 (2015) 169–177

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces

Initial reaction between CaO and SO2 under carbonating

and non-carbonating conditions
Martin Hagsted Rasmussen a,n, Stig Wedel b, Kim H. Pedersen c, Jytte B. Illerup b,
Kim Dam-Johansen b
a
Research & Development, FLSmidth A/S, Vigerslev Allé 77, 2500 Valby, Denmark
b
Department of Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Lyngby, Denmark
c
Research & Development, Haldor Topsoe, Nymoellevej 55, 2800 Lyngby, Denmark

H I G H L I G H T S

External and internal mass transfer limits the initial reaction between CaO/SO2.

CO2 inhibits the CaO/SO2 reaction.

Recirculation of CaO in the cement preheater does not influence SO2 emission.

art ic l e i nf o a b s t r a c t

Article history: The initial kinetics of the CaO/SO2 reaction have been investigated for reaction times shorter than 1 s and
Received 1 September 2014 in the temperature interval between 450 and 600 1C under both carbonating and non-carbonating
Received in revised form conditions (0–20 vol% CO2) to clarify how recirculating CaO influences the emission of SO2 from a
22 April 2015
modern dry kiln preheater system for cement production. Calcined Faxe Bryozo limestone with a particle
Accepted 27 April 2015
Available online 5 May 2015
size smaller than 400 μm was utilized as CaO source. It was shown, both theoretically and experimen-
tally, that the observed reaction rates were influenced by mass transport limitations. The results showed
Keywords: that the CaO conversion with respect to SO2 declined when the CO2 concentration was increased. Under
SO2 all conditions, larger specific surface areas of CaO gave higher reaction rates with SO2. Higher
CO2
temperatures had a positive effect on the reaction between SO2 and CaO under non-carbonating
Calcium oxide
conditions, but no or even a negative effect under carbonating conditions. The results led to the
Reaction kinetics
Particle and gas reactions conclusion that SO2 released from raw meal in the upper stages of the preheater does not to any
Cement production significant extent react with CaO recirculating in the preheater tower.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction limestone and clay, is fed to the preheater tower, which consists of
4–6 cyclones and a calciner. Through the preheater tower, the raw
The annual global production of Portland cement in 2007 meal is heat exchanged, step by step and counter-currently, with
constituted about 2.3  109 tonnes (PCA, 2011), making cement the hot flue gas from the calciner/rotary kiln. In the calciner, about
one of the largest chemical products in the world. Cement is 90 w/w% of the CaCO3 is calcined before the raw meal enters the
manufactured in virtually all countries due to the geographic rotary kiln. The main components, on a dry basis, leaving the
abundance of raw materials, with China as the largest producer preheater tower, are O2, CO2, and N2, which typically comprise, on
in the world. Cement manufacturing is a highly energy-intensive a dry basis, 5 vol%, 30 vol%, and 65 vol%, respectively. The exhaust
process, and in 2001 the industry consumed 2% of the world's also contains water;however, the water content is highly depen-
energy and emitted about 5% of man-made CO2 (Worrell et al., dent on the fuel fired as well as the operation mode. A minor, but
2001). The dry process is the most common way to produce environmentally important component, is SO2, whose concentra-
cement today. In the dry process, raw meal, consisting mainly of tion in the exhaust gas is in the order of 10–3500 mg/Nm3
(Cembureau, 2000). In a dry cement plant SO2 originates from
n
both fuel and raw meal. SO2 from fuel is formed in either the kiln
Corresponding author.
E-mail addresses: [email protected] (M.H. Rasmussen),
burner or the calciner. However, the fuel SO2 does not constitute
[email protected] (S. Wedel), [email protected] (K.H. Pedersen), an emission problem in connection with the preheater exhaust gas
[email protected] (J.B. Illerup), [email protected] (K. Dam-Johansen). as the calciner and the bottom cyclone work as a dry scrubber,

http://dx.doi.org/10.1016/j.ces.2015.04.051
0009-2509/& 2015 Elsevier Ltd. All rights reserved.
170 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177

which in practice removes all SO2 originating from the fuel. SO2 product is found to be CaSO3. Between 500 1C, and 700 1C the
originating from the raw meal is released in the preheater tower reaction product, in an oxygen-free atmosphere, is a mixture of
due to oxidation of pyrite inclusions (Hansen, 2003). The oxidation CaSO3, CaSO4, and CaS. Above 700 1C, only CaSO4 and CaS are
of sulfur species will take place between 300 and 600 1C (Klotz, observed. In an atmosphere containing oxygen, only CaSO4 and
1997), corresponding to the top cyclones in a preheater tower. It is CaSO3 have been observed. The reaction of SO2 and CaO in the
often assumed that 50% of the sulfide entering along with the raw presence of O2 is found to be zero order with respect to oxygen by
meal will leave the preheater tower as SO2 (Salmento and Shenk, different authors (Allen and Hayhurst, 1996; Simons et al., 1988;
2004; Bech and Gundtoft, 1998). However, this number can in Oneill and Keairns, 1976).
practice range from 10 to 65% (Salmento and Shenk, 2004). In A number of authors (Hartman and Trnka, 1993; Tullin and
order to explain the variation observed in SO2 emissions, the Ljungstrom, 1989; Iisa et al., 1991) have been studying the CaO/SO2
reaction between SO2 and CaCO3 has been intensively investigated reaction under carbonating conditions at a time scale of minutes,
byHu et al. (2006, 2007a,b, 208a,b). However, CaO has also without observing any influence on the reaction rate by the
attracted some attention (Bech and Gundtoft, 1998), because CaO carbonating conditions. Borgwardt (1970) has reported a negative
entrained in the gas stream, flowing from the lower parts of the effect of CO2 in his experiments conducted under carbonating
preheater tower to the upper parts, can potentially can affect the conditions at a time scale of seconds. The CaO/SO2 reaction under
emission of SO2. Also, the potential use of a CaO slip stream from carbonating conditions has also been studied in connection with
the lower parts to the upper parts has been considered as a carbonate looping (Coppola et al., 2012; Grasa et al., 2008;
method for reducing SO2 emissions (Miller and Hansen, 2004). Manovic and Anthony, 2010; Ryu et al., 2006; Li et al., 2005;
Hence, deeper knowledge needs to be gained about the reaction Alonso et al., 2014). However, in these studies the scope was how
between SO2 and CaO, at conditions relevant to a preheater tower, SO2 affected the CaO/CO2 reaction. All authors observed that the
in order to fully understand the mechanisms behind emission of degree of carbonation was lowered when SO2 was added to the
SO2 from modern dry cement plants. reaction gas and that the effect become more pronounced when
the SO2 concentration was increased.
Table 1 summarizes the Gibbs free energy for the six reactions
2. Literature survey considered to be of most importance in the CaO/CaCO3/SO2/CO2
system when O2 is present in the reaction gas. The very short
Several authors (Mura et al., 1991; Borgwardt, 1970; Marsh and reaction times and kinetic limitations present in the cement pre-
Ulrichson, 1985; Borgwardt and Bruce, 1986; Tullin and Ljungstrom, heater make it likely that the reactions are far from equilibrium,
1989; Qin, 1995; Illerup et al., 1993) have studied the influence of however, the values in Table 1 show that the reactions considered
reaction time on the reaction between CaO and SO2 in a CO2-free here can take place from a thermodynamic point of view.
atmosphere with reaction times ranging from a few seconds up to The survey has revealed that despite the extensive prior work
hours. As the reaction time increased, the conversion of CaO also on the reaction between CaO and SO2 reported in the literature,
increased. The results obtained by Borgwardt and Bruce (1986) only a limited number of results have been obtained under
show that the reaction is very fast in the initial period. In their case, conditions relevant to the upper stages of a preheater tower.
two-thirds of the total conversion took place within the first 35 s, Especially the short reaction times and carbonating conditions of
while a reaction time of 1000 s or more was required in order to preheater towers have not been considered. This lack of data will
obtain the maximum conversion. be addressed in this paper.
Some authors (Borgwardt and Bruce, 1986; Milne et al., 1990;
Shi and Xu, 2001) have experienced a positive effect of increasing
temperature, whereas others (Mura et al., 1991; Qin, 1995) 3. Experimental
observed that the effect levels off or is even negative at about
1000 1C. A negative effect of temperature is often explained by the The entrained flow reactor, which is a modified version of the
sintering of the CaO particles used, because sintering is known to one used by Hu et al. (2007b), is sketched in Fig. 1.
begin well below 1000 1C. The reactor consists of a 14.95 m long tube of Fe–Cr–Ni-based
The effect of surface area was investigated by Borgwardt and Bruce high-temperature-resistant alloy. It is possible to inject SO2 and
(1986) using CaO particles with a surface area between 2.1 m2/g and CO2 at four different points, marked V1–V4. The reactor tube is
63 m2/g. They found that the conversion increased by increasing the located horizontally in the oven, with a downward slope of 1 1, and
surface area, for both short and long reaction times. The positive effect contains seven U-bends in order to fit into the oven. The pipes for
was also reported in other publications (Milne et al., 1990; Shih et al., injecting SO2 and CO2 are dimensioned such that the gases have
2004). The positive effect is due to a large reaction interface for the reached the reactor temperature when injected. The reactor oven
CaO/SO2 reaction. The reaction order for SO2 has been estimated to be is electrically heated and controlled by a thermostat, which
in the range from 0.49 to 0.7 by a number of authors (Borgwardt and ensures a constant reactor temperature. An electrical preheater
Bruce, 1986; Borgwardt et al., 1987; Yang et al., 1978), who also stated is used for preheating the main gases, which are nitrogen or
that diffusion limitations were negligible. Most values were about 0.6. compressed air. Solid material is fed from the loss-in-weight
Yang et al. (1978) found that the reaction order increased when
doping the pure CaO sources with Fe2O3. Borgwardt et al. (1987) Table 1
found the reaction order to be independent of whether the CaO Upper and lower Gibbs energies for the 6 most important reactions consider in this
precursor was CaCO3 or Ca(OH)2. In general, the conversion is reported paper. The unit is kJ/mole.
to be independent of the CaO precursor in the initial period of reaction
# Reaction Δ G(450 1C) Δ G(600 1C)
(Gullett and Bruce, 1987; Bruce et al., 1989; Borgwardt et al., 1987;
Shih et al., 2004). 1 CaOþCO2 ⇌ CaCO3  63.7  39.9
A number of authors (Munoz-Guillena, 1994a,b; Low et al., 2 CaOþ SO2 ⇌ CaSO3  92.5  64.7
1971; Ishizuka et al., 2000; Ghardashkhani and Cooper, 1990; Allen 3 CaSO3 þ 0.5 O2⇌ CaSO4  279.1  280.0
and Hayhurst, 1996; Li and Sadakata, 1999) have studied the 4 CaOþSO2 þ0.5 O2 ⇌ CaSO4  371.6  344.7
5 CaCO3 þ SO2⇌ CaSO3 þ CO2  28.8  24.9
reaction product from the reaction between CaO and SO2 using 6 CaCO3 þSO2 þ0.5 O2 ⇌CaSO4 þ CO2  307.9  304.9
X-ray and IR spectroscopy. In general, below 500 1C, the reaction
 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177 171

content is stored in a sealed container for later analysis. In the

Pa
rti
experiments presented here, an SO2 concentration of 1000 ppm

cl
N2

e
was used unless otherwise else stated.

Fe
ed
Gases (N2, CO2, and SO2) with a purity of at least 99.2% were

er
supplied from gas cylinders. Oxygen was supplied through the
N2 /Air Preheater Mixing Point
T P Reactor Oven pressurized air line in the experimental facility building as dry
Water atmospheric air.
T
Table 2 shows the composition of the Faxe Bryozo limestone
Reactor Tube T and the commercial CaO (based on Faxe Bryozo limestone) used
for experiments. The Faxe Bryozo limestone has been calcined in
SO2 order to obtain CaO containing particles for reaction with SO2 and
V1 CO2. The calcination was carried out either as flash calcination in
T
the entrained flow reactor with a calcination time of less than 1 s
22.48mm or by calcination in a fluid bed reactor for 50 min at a set point
CO2 temperature of 850 1C. Both methods resulted in lime products
V2 that were only partially calcined, meaning that the product still
T
contained some CaCO3. Table 3 summarizes the properties of the
calcined limestone. It is further noted from Tables 2 and 3 that the
Gas Outlet calcined limestone contains significantly larger surface areas than
V3 the uncalcined limestone. The significant increase in surface area
Capsule Fan
T is due to the formation of large surface area micro grains of CaO.
T
Thus, it is reasonable to assume that the main component in
Cyclone

V4
Gas contact with the gas phase is CaO, and that CaO is the reacting
Sampling component in the initial phase.
Fig. 2 shows a typical SO2 measurement, with the horizontal
Ga
sA lines indicating the steady-state levels used for calculations. The
na
lyz difference in SO2 concentration with and without particle injection
Particle Container er
is due to SO2 reacting with the injected material. The steady-state
Particle
Filter levels are used for calculating an average SO2 concentration before,
during, and after particle injection.
Fig. 1. Sketch of the entrained flow reactor used for studying the reactions
Eq. (1) defines the conversion of a basis Ca-species (Ca, CaO,
between calcium containing species and SO2/CO2.
etc.) to sulfur compounds (CaSO3, CaSO4, etc.) as the moles of SO2
reacted per mole of basis compound. The use of different bases is
particle feeder, which is situated in an airtight bin. Blockage of the convenient when dealing with different materials and partially
feeding string is avoided by letting nitrogen pass through the calcined particles
particle feeder. The use of nitrogen to facilitate the feeding of solid nSO2
material ensures an inert atmosphere in the sealed particle feeder X SO2 =basis ¼ ; nbasis ¼ nCa ; nCaO ; nCaCO3 ð1Þ
nbasis
bin. Preheated gas and solid material are mixed before entering
the reactor oven. In the reactor oven, gas and solid materials are In the experiments, the change of SO2 concentration, ΔxSO2 ,
brought up to reactor temperature before SO2 and CO2 are was calculated as the difference between average steady-state SO2
injected. Just before the reactor outlet, about 2% of the gas is concentration before and after particle feeding and the steady-
sampled through a filter for analysis. The filter is heated in order to state SO2 concentration with particle feeding. This allows the
ensure that water does not condense. In the reactor outlet, the conversion to be calculated from Eq. (2), where P is the reaction
remaining solid material and gas are separated and the gas is pressure, νamb is the total gas flow at ambient temperature, R is the
discarded, while the solid material is collected in a particle universal gas constant, Tamb is the ambient temperature, W is the
container for further analysis. Temperature and pressure are feed rate of solid material to the reactor, Mbasis is the molar mass of
measured at the places marked T and P, respectively. All gas flows
to the reactor are controlled by mass flow controllers. Table 2
Composition of the limestone and the commercial CaO used for SO2 experiments.
Initially, the reactor temperature is allowed to stabilize. Next,
Compositions have been obtained from suppliers, while surface area and porosity
the main gas is supplied and the preheater temperature is are measured by BET measurements and Hg-porosimetry, respectively.
adjusted so that the reactor tube is isothermal from V1 and
forward. Then CO2 and SO2 are added to the reactor through one Material composition (w/w%) Faxe Bryozo limestone CaO commercial
of the inlet valves V1–V4; the valve used depends on the required
CaO 53.20 96.00
reaction time. Next, the valve situated between the particle feeder SiO2 0.49 1.00
and the reactor tube is opened, and nitrogen from the feeder is MgO 0.43 0.90
mixed with the main gas. The SO2 concentration is adjusted to the Al2O3 0.02 0.25
correct value and allowed to stabilize. When a stable SO2 con- Fe2O3 0.03 0.25
K2O 0.01 0.01
centration has been obtained, the injection of particles starts. The Na2O 0.00 0.00
injected amount of particles depends on the amount of SO2 TiO2 0.00 0.00
reacting. The amount should be enough to minimize measurement Mn2O3 0.01 0.01
uncertainties; however, the drop in SO2 concentration must not be P2O5 0.02 0.02
SrO 0.05 0.05
so high that the SO2 concentration changes significantly. In
LOI 43.50 1.00
practice, feed rates between 0.2 and 2.5 kg/h are reasonable. When Total CaCO3 97.16 NaN
a stable SO2 concentration has been obtained for 3–5 min, particle
Surface area (m2/g) 0.8 3.2
feeding is stopped and the SO2 concentration is allowed to Porosity (%) 8.5 18
stabilize again. Finally, the particle container is emptied and the
172 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177

Table 3
Characterization of partially calcined limestone used for studying the reaction between CaO and SO2. The numbers 750, 800, and 850 denote the flash calcination
temperature and the prefixes 2 and 5 denote the number of times the material has been flash calcined. FB denotes that the limestone was calcined in a fluidized bed, and
commercial means that the material was bought as CaO.

Material Calcination method xCaO=Ca Q mCaO =mtot Surface areaa Porosity

(% mol/mol) (% kg/kg) (m2/g CaO) (%)

Faxe 750 Flash 12.1 6.9 143.5 14

Faxe 800 Flash 25.8 15.4 101.9 21
Faxe 850 Flash 45.7 30.2 89.7 31
2xFaxe 850 Flash  2 60.2 42.2 – –
5xFaxe 850 Flash  5 81.8 68.9 – –
Faxe FB Fluidized bed 87.5 76.3 26.3 52
Faxe commercial Rotary Kiln 100 100 3.2 18

a
The surface area is per gram of CaO in the solid material, assuming that the CaCO3 surface area is negligible.

Fig. 3. Particle diameters at which the Weisz–Prater modulus (CWP) is equal to 0.15
Fig. 2. Example of steady-state SO2-levels measured in the entrained flow reactor. as a function of the measured CaO conversion for three different CaO materials.

the basis compound, and Q basis=mtot is the fraction of basis material

in the solid feed
 ΔxSO2  P  νamb M basis
X SO2 =basis ¼ ð2Þ
R  T amb W  Q basis=mtot

4. Results and discussion

The Weisz–Prater modulus compares the observed rate of

reaction to a characteristic reactant diffusion rate in a porous
material. As defined in Fogler (2006, p. 839), a Weisz–Prater
modulus smaller than 0.15 indicates that the results are not
influenced by internal diffusion limitations. Similar to the
Weisz–Prater modulus for evaluating internal diffusion limitations,
external diffusion limitation can be evaluated by Mears criterion.
As defined in Fogler (2006, p. 841), Mears criterion states that
external transport limitations can be neglected if Mears ratio is Fig. 4. Particle diameters at which Mears criterion (CM) is equal to 0.15 as a
smaller than 0.15. function of the measured CaO conversion for three different CaO materials.
Fig. 3 shows the particle diameter, at which the Weisz–Prater
modulus is equal to 0.15, as a function of the measured CaO 4.1. Diffusion limitations
conversion for three different CaO materials, while Fig. 4 shows a
similar plot for Mears criterion. In both cases, particle parameters From Fig. 3, it is evident that the particle diameter at which
from Table 3 have been used, and for Mears criterion, particle internal diffusion limitations are negligible is between 5 and
diameters have been calculated by assuming each particle sur- 25 μm. Fig. 4 shows that external transport limitations will begin
rounded by stagnant gas. Thus, the particle diameters shown in to influence the results at particle diameters between 10 and
Figs. 3 and 4 are the largest particle diameters in each experiment 70 μm. The particle size distributions for the three different CaO
for which diffusion limitations can be ruled out. materials (Fig. 5) show that, for at least 85% of the material, the
 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177 173

would then depend purely on the external surface area, which is

similar in all experiments with Faxe 750, 800, and 850.
Fig. 8 shows the CaO conversion as a function of surface area for
two differently calcined Faxe limestone (see Table 3). The CaO
materials were reacted with SO2 for 0.35 s or 0.55 s in a gas containing
either 20 or 25 % CO2. If these CaO conversions are compared to those
shown in Fig. 7, it is noteworthy that considerably lower CaO
conversion is seen when CO2 is present in the reaction gas. It is also
seen that a higher surface area leads to higher CaO conversion, even
though the sulfation is suppressed by CO2. Furthermore, it is

Table 4
Measured sulfur content in Faxe 850 reacted with SO2 at 517 1C for
0.08 s in atmospheric air.

Particle size (μm) Sulfur content (% w/w)

0–32 4.35
32–90 1.05
90–150 0.33
Fig. 5. Particle size distribution for the three CaO materials used for evaluating Z 150 0.25
diffusion limitations.

rates obtained will be influenced by internal diffusion limitations,

and for about 80% of the material obtained, the rates will also be
influenced by external transport limitations.
In order to verify the calculated results, the reaction product
from an experiment with Faxe 850 was fractionated and the sulfur
content subsequently determined. Table 4 summarizes the results.
The table clearly indicates that only a very small part of the large
particles reacts with SO2. The actual concentration gradients inside
the particles were studied by slicing a sample of Faxe 750 particles
and subsequently mapping the cross-sectional sulfur content by
EDX (apparatus: FEI Inspect S). Fig. 6 shows the results for three
different particles, all of which had a diameter around 80 μm, and
it is observed that sulfur species are clearly only present very close
to the particle surface.
The results in Table 4 and Fig. 6 confirm the Weisz–Prater modulus
and Mears criterion calculations, showing that diffusion limitations
cannot be ruled out when interpreting experimental results.
Calculations of potential temperature gradients, caused by the
exothermic reaction, have shown that, in the worst case, the tem-
perature of the CaO materials can be about 0.5% higher than the Fig. 6. Scaled sulfur content as a function of the dimensionless particle diameter.
reactor temperature, calculated in absolute temperature. Hence, tem-
perature gradients within the particles can be neglected, and the
system can be assumed to be isothermal. That mass transport
limitations may influence the reaction between CaO and SO2 for such
relatively small particles as those used in this study has, as far as the
authors are aware, not been reported in the literature. Mass transport
limitations may be a significant reason for the varying SO2 reaction
orders previously reported in the literature.

4.2. Surface area

Fig. 7 shows CaO conversion when reacted with SO2 as a

function of the surface area for CaO materials derived from Faxe
limestone. It is evident that the CaO conversion increases with
surface area, and a positive effect of temperature is, in general, also
observed. It is also seen that an increase in surface area can
compensate for a lower reaction temperature. The positive effect
of increasing surface area is also reported in the literature
(Borgwardt and Bruce, 1986; Milne et al., 1990; Shih et al., 2004).
Previously it was shown that the observed results are heavily
influenced by diffusion limitations. The positive effect of specific Fig. 7. CaO conversion as a function of surface area for partially calcined Faxe
surface area upon CaO conversion shows that external diffusion limestone reacted for 0.36 s in atmospheric air at different temperatures,
does not exclusively limit the reaction rate, since CaO conversion C SO2 ¼ 1000 ppm.
174 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177

noticeable that an increase in reaction time from 0.35 s to 0.55 s only to 0.4 s  1 for Faxe 850. Thus, all CaO would be converted in less
has a limited effect on CaO conversion, indicating that most of the than a second for Faxe 750 if the initial reaction rate persisted for
reaction takes place within less than 0.35 s. the entire CaO conversion.
Borgwardt and Bruce (1986) conducted experiments at 800 1C
4.3. Reaction time using CaO particles with a diameter of 1 μm and a BET surface area
of 63 m3/g. They performed their first measurement after 3 s of
Fig. 9 shows that the CaO conversion increases with reaction reaction. From their data, an apparent initial conversion rate of
time. CaO is used as the basis compound since it comprises by far 6:2  10  2 s  1 was calculated. The initial reaction rates observed in
the majority of the reactive surface area. The difference in CaO our experiments are 6.5–17 times higher than what Borgwardt and
conversion among the three calcines is due to the difference in Bruce (1986) observed at a reaction temperature about 300 1C
surface area in the starting material (see Table 3). The slopes of the higher. This ratio between short-time and longer-time rates
lines interconnecting the measurement points can be interpreted indicates that the initial rate can be significantly underestimated
as average apparent reaction rates. It is clear that, in all cases, the if determined from experimental data obtained after more than
reaction rate is highest within the first 0.08 s. Hereafter, the rate 1 s. Furthermore, our data make it reasonable to assume that
levels off and the increase in conversion appears linear from 0.18 s diffusion limitations will be present in the initial phase of any SO2/
and forward. That the initial reaction rate between SO2 and CaO is CaO experiment, unless using extreme conditions such as very low
high is further illustrated by more than 50% of the observed CaO temperatures and/or extremely small particles. Hence, true kinetic
conversion taking place within the first 0.18 s, which is less than initial rates are unlikely to be measured by experiments based on
one-third of the total reaction time. From the data in Fig. 9, the bulk quantities of SO2 and CaO particles.
apparent rate can be calculated as CaO conversion per second. The full line graphs in Fig. 10 show the conversion of CaO into
Within the first 0.08 s, the rate ranges from 1.07 s  1 for Faxe 750 sulfur species (left axis), while the dotted graph shows the
unconverted fraction of the initial CaO (right axis), all three plotted
as a function of reaction time. The experiments were conducted in
25% CO2.
The CaO conversion in CO2 is considerably lower than in
atmospheric air. Furthermore, a slight negative effect of increasing
temperature is observed in Fig. 10. From the experiments in
atmospheric air, a linear increase in CaO conversion is observed
after 0.18 s. Here, the reaction rate seems to decrease continuously
with increasing reaction time. In fact, the reaction is almost
completed after 0.35 s at 534 1C . The fraction of unconverted
CaO in the partially calcined limestone decreases remarkably fast,
reaching a constant level of about 55% after 0.35 s. TGA measure-
ments (apparatus: Netzsch STA 409 C) have shown that the CaO is
converted mainly into CaCO3, since the uptake of CO2 is about 25
times larger than the SO2 uptake. Furthermore, BET measurements
(apparatus: QuantaChrome Autosorb iQ2) have shown that, after
0.35 s, only 10% of the internal surface area found in the starting
material is left. From these results, it is obvious that free CaO will
react extremely fast with CO2 and form CaCO3 if exposed to a CO2
Fig. 8. CaO conversion as a function of surface area for partially calcined Faxe
partial pressure higher than the equilibrium pressure. The partially
limestone reacted in a CO2 atmosphere containing 20–25% CO2 at 535 1C. In all calcined limestone used has a bipartite pore system with a main
cases C SO2 ¼ 1000 ppm. pore diameter of about 500 nm and a secondary pore diameter of
about 10–20 nm. In such a bipartite pore system, the majority of
the surface area is situated in the secondary pore system.

Fig. 9. CaO conversion as a function of the reaction time for partially calcined Faxe Fig. 10. CaO conversion (left axis) and fraction of unconverted CaO after reaction
limestone with different amounts of CaO. The experiments were conducted in (right axis), as a function of the reaction time for partially calcined Faxe limestone,
atmospheric air, C SO2 ¼ 1000 ppm, T ¼ 5177 2 1C. when the reaction gas contains 25% CO2 and 1000 ppm SO2.
 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177 175

Furthermore, both the CaO/CO2 and the CaO/SO2 reaction rates The results presented here contradict those found in the
will increase with temperature, but since the molar volume of literature (Hartman and Trnka, 1993; Tullin and Ljungstrom,
CaCO3 is about 2.5 times that of CaO, the secondary pore system 1989; Iisa et al., 1991; Borgwardt, 1970), where no CO2 effect was
blocks up faster at higher temperature levels, so the surface area observed. The main reason for this discrepancy may be the
available for reaction with SO2 will decrease even faster. The reaction time used, which was significantly longer in the experi-
reduction in surface area with a decreasing fraction of uncon- ments reported in the literature. At the short reaction times used
verted CaO is supported by only 10% of the initial BET area being here, CO2 will occupy sites that otherwise could react quickly with
left after 0.35 s. The negative effect of higher temperatures is a SO2, and CO2 thereby slows down the reaction. Conversely, at long
consequence of the fast uptake of CO2. The large uptake of CO2 reaction times, the controlling mechanisms will be diffusion and
begs the question whether SO2 reacts with CaO or CaCO3. This is reaction in a product layer of CaCO3, CaSO3, and CaSO4. No
important as the initial rate of the SO2–CaCO3 reaction is slower discernible effect of CO2 is observed because the reaction between
than the initial rate of the SO2–CaO reaction (Liu et al., 2000). The the formed CaCO3 and SO2 is much faster than the transport of SO2
experimental results show that the CaO–SO2 reaction only con- to the unreacted CaO core, which must take place through the
tributes to the observed CaO conversion to sulfur species until the solid layer of reaction products. The results presented here clearly
particles are saturated with CO2. At a temperature of 534 1C, the show the pitfalls of predicting short reaction time outcomes solely
CaO–SO2 reaction will only compete with the reaction between from data gathered from long reaction time experiments.
CaO and CO2 within the first 0.35 s. At longer reaction times, the
dominating reaction mechanism for SO2 will be the reaction
between SO2 and CaCO3, which is also reflected in the levelling 4.5. Temperature
off of the reaction rate after 0.35 s at 534 1C.
Fig. 12 shows CaO conversion as a function of temperature
when the reaction gas contains CO2. Considering the Faxe FB
4.4. CO2 concentration results, there are no effects of increasing the temperature when
the reaction gas contains 20% CO2, while a small increase is
Fig. 11 shows the CaO conversion as a function of CO2 observed at the highest temperature for 10% CO2 in the reaction
concentration in the reaction gas. The CaO conversion decreases gas. In the case of 2xFaxe 850 (see Table 3), the temperature has a
when the CO2 concentration increases. It is also apparent that the negative effect on the CaO conversion when 25% CO2 is present in
effect of increasing the reaction time levels off as the CO2 the reaction gas. Observing no effect or even a negative effect of
concentration increases. The negative effect of increasing the CO2 temperature is different from the results obtained when conduct-
concentration is due to the carbonization reaction which, at CO2 ing similar experiments in atmospheric air, where nothing but a
partial pressures higher than the equilibrium pressure, competes positive effect of temperature was observed. The trends observed
with SO2 for any available CaO. When the CO2 partial pressure in Fig. 12 can be explained by the pore size distributions in the two
increases, the effect upon the CaO conversion seems to level off, CaO materials. It has been determined by Hg porosimetry (appa-
indicating that the reaction becomes independent of the CO2 ratus: Micromeritics autopore II 9220) that flash calcined lime-
concentration at high concentrations. That the CaO–CO2 reaction stone, like 2xFaxe 850, has a narrow secondary pore system with a
should be independent of the CO2 concentration at higher con- size distribution around 10 nm, which is significantly smaller than
centrations is in agreement with literature findings (Sun et al., the broad pore size distribution between 20 and 150 nm observed
2008; Kyaw et al., 1996; Bhatia and Perlmutter, 1983; Oakeson and for Faxe FB. Thus, blockage of the secondary pore system due to
Cutler, 1979; Dedman and Owen, 1962). the formation of CaCO3 in 2xFaxe 850, and thereby a reduction of
That increasing reaction time has less positive effect on CaO the reactive surface area, increases faster with temperature than
conversion at high CO2 partial pressure is due to saturation of the the reaction rates of SO2–CaO and SO2–CaCO3, here resulting in a
material with CO2, as also seen in Fig. 10. The results in Fig. 11 is negative temperature effect. In the literature (Hartman and Trnka,
additional evidence that the saturation rate is fast and indepen- 1993; Suyadal and Oguz, 1999), only a positive temperature effect
dent of CO2 partial pressure when it is high. was reported, which is most likely due to longer reaction times
and the more sintered CaO material used.

Fig. 12. CaO conversion as a function of reaction temperature. The experiments

Fig. 11. CaO conversion as a function of CO2 concentration in the reaction gas for were carried out in different CO2 concentrations, with a reaction time of 0.55 s and
partially calcined Faxe limestone, C SO2 ¼ 1000 ppm. an SO2 concentration of 1000 ppm.
176 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177

conditions that are similar to those present in industrial practice.

That is, however, a question of academic importance only, since
industrial processes utilizing short reaction times will nearly
always be carried out in a mode where mass transfer limitations
will be present, as in our experiments.
The experiments showed that a carbonating atmosphere inhi-
bits the conversion with respect to SO2, with the inhibiting effect
increasing with the CO2 content up to about 20% CO2. The
inhibition is due to formation of CaCO3 at the reaction surface.
CaCO3 reacts slower with SO2 than CaO. Furthermore, it has been
observed that CaCO3 causes pore blockage, resulting in serious
reduction of available reactive surface area. The experiments
further showed that increased surface area of the calcined lime-
stone had a positive effect on the CaO conversion. Higher reaction
temperatures had a clear positive effect in experiments conducted
in atmospheric air, while no effect or even a negative effect of
increasing temperature was observed at CO2 partial pressures of
Fig. 13. IR spectrum for Faxe 750 and the reaction product of Faxe 750 reacted at
different temperatures for 0.35 s in atmospheric air using an SO2 concentration of 10% or higher. This is explained by the reduction in reactive surface
1000 ppm. area, due to the formation of CaCO3, counteracting the positive
effect of increasing temperature on the SO2 reaction rate.
In the literature on SO2 emission from cement preheater
4.6. Reaction product towers, it has been stated that CaO recirculating in the preheater
can influence the SO2 emission. Based on the results presented
Fig. 13 shows the IR spectra recorded (apparatus: Perkin Elmer here, it is evident that CaO recirculating in the preheater does not
Spectrum model 2000) for Faxe 750 and the reaction product for contribute significantly to the SO2 removal in the top cyclones,
Faxe 750 reacted with SO2 at 515 1C and 590 1C. It is apparent that where SO2 formation takes place. This is so because, under
absorption peaks not found in the starting material are observed at carbonating conditions, any CaO will react instantaneously with
1100–1200 cm  1, 900–1050 cm  1 and 650 cm  1 for Faxe 750 CO2, forming CaCO3 in the upper preheater stages.
reacted at 590 1C. For Faxe 750 reacted at 515 1C only the peak at
900–1050 cm  1 is clear. Careful study of the figure reveals that a
weak peak at 1000–1200 cm  1 is also present for the reaction Acknowledgments
product at 515 1C. The peak found at 650 cm  1 for the reaction
product at 590 1C seems to be present only as a very weak peak at The work is part of the advanced technology platform ‘New
515 1C, if present at all. In general then, the peaks which differ from Cement Production Technology’ funded by the Danish National
the reaction starting material are weaker at 515 1C than at 590 1C. Advanced Technology Foundation, FLSmidth A/S, and Technical
Comparing the three peaks observed for Faxe 750 reacted at University of Denmark (Grant no. 2007-2-001).
590 1C, for instance, to the observations made by Ghardashkhani
and Cooper (1990), we find that the peaks at 650 cm  1 and 900– References
1050 cm  1 are due to the presence of CaSO3, whereas the peak at
1100–1200 cm  1 is due to the presence of CaSO4. Thus, at 590 1C, Allen, D., Hayhurst, A., 1996. Reaction between gaseous sulfur dioxide and solid
the SO2/CaO reaction product is a mixture of CaSO3 and CaSO4. At calcium oxide—mechanism and kinetics. J. Chem. Soc.—Faraday Trans. 92 (7),
1227–1238.
515 1C, the near disappearance of the peak at 1000–1200 cm  1, Alonso, M., Criado, Y., Abanades, J., Grasa, G., 2014. Undesired effects in the
combined with the presence of a peak at 900–1050 cm  1, indi- determination of co2 carrying capacities of cao during tg testing. Fuel 127 (0),
cates that the product is basically CaSO3 with traces of CaSO4. 52–61.
Bech, C., Gundtoft, L., 1998. Study of NOx, SOx, and CO mechanisms based on actual
Generally, peaks are weaker at 515 1C, accounting for the disap-
plant data. In: Cement Industry Technical Conference, 1998. 40th Conference
pearance of the peak at 650 cm  1. Record. 1998 IEEE/PCA, Piscataway, NJ, pp. 141–154.
If the intensities of the peaks are taken as an indication of the Bhatia, S.K., Perlmutter, D.D., 1983. Effect of the product layer on the kinetics of the
amount of a given product, the dominating product will be CaSO3 CO2-lime reaction. AIChE J. 29 (1), 79–86.
Borgwardt, R.H., 1970. Kinetics of the reaction of SO2 with calcined limestone.
in both cases. The reaction product at both temperatures is a Environ. Sci. Technol. 4 (1), 59–63.
mixture of CaSO3 and CaSO4. The main product being CaSO3 Borgwardt, R.H., Bruce, K.R., 1986. Effect of specific surface area on the reactivity of
corresponds well with the findings by other authors (Low et al., CaO with SO2. AIChE J. 32 (2), 239–246.
Borgwardt, R.H., Bruce, K.R., Blake, J., 1987. Investigation of product-layer diffusivity
1971; Ishizuka et al., 2000; Ghardashkhani and Cooper, 1990; Allen for CaO sulfation. Ind. Eng. Chem. Res. 26 (10), 1993–1998.
and Hayhurst, 1996; Li and Sadakata, 1999) using IR measurements Bruce, K., Gullett, B., Beach, L., 1989. Comparative SO2 reactivity of CaO derived from
to identify their reaction products. CaCO3 and Ca(OH)2. AIChE J. 35 (1), 37–41.
Cembureau, 2000. “Best Available Techniques” for the Cement Industry.
Coppola, A., Montagnaro, F., Salatino, P., Scala, F., 2012. Fluidized bed calcium
looping: the effect of SO2 on sorbent attrition and CO2 capture capacity. Chem.
5. Conclusion Eng. J. 207 and 208 (0), 445–449.
Dedman, A., Owen, A., 1962. Calcium cyanamide synthesis. Part 4. The reaction CaO.
CO2 ¼ CaCO3. Trans. Faraday Soc. 58 (478), 2027.
This paper focuses on the CaO/SO2 reaction at reaction times Fogler, H.S., 2006. Elements of Chemical Reaction Engineering. Prentice-Hall, Upper
shorter than 1 s. It was shown that the results observed were Saddle River, NJ.
Ghardashkhani, S., Cooper, D., 1990. A thermogravimetric study of the reaction
strongly influenced by external and internal mass transfer limita-
between sulfur-dioxide and calcium-oxide. Thermochim. Acta 161 (2), 327–337.
tions. When the intrinsic rates of reactions in the initial phase are Grasa, G., Alonso, M., Abanades, J., 2008. Sulfation of CaO particles in a carbonation/
very high to extremely high, transport through either gas phase or calcination loop to capture CO2. Ind. Eng. Chem. Res. 47, 1630–1635.
solid phase becomes the controlling mechanism governing the Gullett, B., Bruce, K., 1987. Pore distribution changes of calcium-based sorbents
reacting with sulfur-dioxide. AIChE J. 33 (10), 1719–1726.
rate obtained. It will be difficult to establish experimental condi- Hansen, J.P., 2003. SO2 Emissions from Cement Production (Ph.D. thesis). Technical
tions that allow the determination of intrinsic reaction rates under University of Denmark.
 M.H. Rasmussen et al. / Chemical Engineering Science 134 (2015) 169–177 177

Hartman, M., Trnka, O., 1993. Reactions between calcium-oxide and flue-gas Miller, S., Hansen, J., 2004. Methods for reducing SO2 emissions. In: Cement
containing sulfur-dioxide at lower temperatures. AIChE J. 39 (4), 615–624. Industry Technical Conference, 2004, IEEE-IAS/PCA, Piscataway, NJ, pp. 79–92.
Hu, G., Dam-Johansen, K., Wedel, S., 2008b. Oriented nucleation and growth of Milne, C., Silcox, G., Pershing, D., Kirchgessner, D., 1990. High-temperature, short-
anhydrite during direct sulfation of limestone. Cryst. Growth Des. 8 (4), time sulfation of calcium-based sorbents. 2. Experimental-data and theoretical-
1181–1185. model predictions. Ind. Eng. Chem. Res. 29 (11), 2201–2214.
Hu, G., Dam-Johansen, K., Wedel, S., Hansen, J.P., 2007a. Enhancement of the direct Munoz-Guillena, M., Linares-Solano, A., Salinas-Martinez de Lecea, C., 1994a. A
sulfation of limestone by alkali metal salts, calcium chloride, and hydrogen study of CaO–SO2 interaction. Appl. Surf. Sci. 81 (4), 409–415.
chloride. Ind. Eng. Chem. Res. 46 (16), 5295–5303. Munoz-Guillena, M., Linares-Solano, A., Salinas-Martinez de Lecea, C., 1994b. A
Hu, G., Dam-Johansen, K., Wedel, S., Hansen, J.P., 2007b. Direct sulfation of study of CaO–SO2 interaction in the presence of O2. Appl. Surf. Sci. 81 (4),
limestone. AIChEt J. 53 (4), 948–960. 417–425.
Hu, G., Dam-Johansen, K., Wedel, S., Peter Hansen, J., 2006. Review of the direct Mura, G., Lallai, A., Olla, P., 1991. On the kinetics of dry desulfurization with calcium
sulfation reaction of limestone. Prog. Energy Combust. Sci. 32 (4), 386–407. oxide. Chem. Eng. J. 46 (3), 119–128.
Hu, G., Dam-Johansen, K., Wedel, S., Shang, L., Hansen, J.P., 2008a. Initial kinetics of Oakeson, W., Cutler, I., 1979. Effect of CO2 pressure on the reaction with CaO. J. Am.
the direct sulfation of limestone. AIChE J. 54 (10), 2663. Ceram. Soc. 62 (11–1), 556–558.
Iisa, K., Tullin, C., Hupa, M., 1991. Simultaneous sulfation and recarbonation of Oneill, E., Keairns, D., Kittle, W., 1976. Thermogravimetric study of sulfation of
calcined limestone under PFBC conditions. In: 11th International Conference on limestone and dolomite, effect of calcination conditions. Thermochim. Acta 14
Fluidized Bed Combustion, ASME, Brussels, Belgium 1991, pp. 83–90. (1–2), 209–220.
Illerup, J., Dam-Johansen, K., Lunden, K., 1993. High-temperature reaction between PCA, 2011. Innovations in Portland Cement Manufacturing. Potland Cement
sulfur dioxide and limestone-VI. the influence of high pressure. Chem. Eng. Sci. Association.
48 (11), 2151–2157. Qin, Z., 1995. Thermogravimetric studies of the reaction of CaO with SO2. J. Thermal
Ishizuka, T., Kabashima, H., Yamaguchi, T., Tanabe, K., Hattori, H., 2000. Initial step Anal. 45 (1), 211–219.
of flue gas desulfurization—an IR study of the reaction of SO2 with NOx on CaO. Ryu, H., Grace, J., Lim, C., 2006. Simultaneous CO2/SO2 capture characteristics of
Environ. Sci. Technol. 34 (13), 2799–2803. three limestones in a fluidized-bed reactor. Energy Fuels 20, 1621–1628.
Klotz, B., 1997. New developments in precalciners and preheaters. In: Cement Salmento, J., Shenk, R., 2004. Accurately predicting cement plant emissions. In:
Industry Technical Conference. 1997. XXXIX Conference Record, IEEE/PCA, IEEE-IAS/PCA 2004 Cement Industry Technical Conference (IEEE Cat.
Piscataway, NJ, pp. 255–280. No04CH37518), pp. 333–343.
Kyaw, K., Kanamori, M., Matsuda, H., Hasatani, M., 1996. Study of carbonation Shi, L., Xu, X., 2001. Study of the effect of fly ash on desulfurization by lime. Fuel 80
reactions of Ca–Mg oxides for high temperature energy storage and heat (13), 1969–1973.
transformation. J. Chem. Eng. Jpn. 29 (1), 112. Shih, S.-M., Hung, J.-T., Wang, T.-Y., Lin, R.-B., 2004. Kinetics of the reaction of sulfur
Li, Y., Buchi, S., Grace, J., Lim, C., 2005. SO2 removal and CO2 capture by limestone dioxide with calcium oxide powder. J. Chin. Inst. Chem. Eng. 35 (4), 447–454.
resulting from calcination/sulfation/carbonation cycles. Energy Fuels 19, Simons, G.A., Parker, T.E., Morency, J.R., 1988. The oxygen reaction order of SO2 with
1927–1934. CaO. Combust. Flame 74 (1), 107–110.
Li, Y., Sadakata, M., 1999. Study of gypsum formation for appropriate dry Sun, P., Grace, J.R., Lim, C.J., Anthony, E.J., 2008. Determination of intrinsic rate
desulfurization process of flue gas. Fuel 78 (9), 1089–1095. constants of the CaO–CO2 reaction. Chem. Eng. Sci. 63 (1), 47–56.
Liu, H., Katagiri, S., Kaneko, U., Okazaki, K., 2000. Sulfation behavior of limestone Suyadal, Y., Oguz, H., 1999. Dry desulfurization of simulated flue gas in a fluidized-
under high CO2 concentration in O2/CO2 coal combustion. Fuel 79 (8), 945–953. bed reactor for a broad range of SO2 concentration and temperature: a
Low, M., Goodsel, A., Takezawa, N., 1971. Reactions of gaseous pollutants with comparison of models. Ind. Eng. Chem. Res. 38 (8), 2932–2939.
solids. i. Infrared study of sorption of SO2 on CaO. Environ. Sci. Technol. 5 (12), Tullin, C., Ljungstrom, E., 1989. Reaction between calcium-carbonate and sulfur-
1191. dioxide. Energy Fuels 3 (3), 284–287.
Manovic, V., Anthony, E., 2010. Competition of sulphation and carbonation reac- Worrell, E., Price, L., Martin, N., Hendriks, C., Meida, L.O., 2001. Carbon dioxide
tions during looping cycles for CO2 capture by CaO-based sorbents. J. Phys. emissions from the global cement industry. Annu. Rev. Environ. Resour. 26, 303.
Chem. Al. 114, 3997–4002. Yang, R.T., Shen, M.-S., Steinberg, M., 1978. Fluidized-bed combustion of coal with
Marsh, D., Ulrichson, D., 1985. Rate and diffusional study of the reaction of calcium- lime additives: catalytic sulfation of lime with iron compounds and coal ash.
oxide with sulfur-dioxide. Chem. Eng. Sci. 40 (3), 423–433. Environ. Sci. Technol. 12 (8), 915–918.