Page 1
1
FINAL TECHNICAL REPORT
March 1, 2013, through July 31, 2014
Project Title: ON-SITE PRODUCTION OF LIME TO REMOVE SO3 AND HCL
FROM ILLINOIS COAL COMBUSTION FLUE GAS
ICCI Project Number: 13/6A-1
Principal Investigator: Hong Lu, ISGS/UIUC
Project Manager: Debalina Dasgupta, ICCI
ABSTRACT
The project aimed at producing highly reactive calcium-based sorbents using a batch
reactor and a 5-10 lbs/hr sorbent activation process (SAP) unit at the Illinois State
Geological Survey (ISGS) for the removal of sulfur trioxide (SO3) and hydrochloric acid
(HCl) in Illinois coal combustion flue gases. More than twenty sorbents were prepared
and the reactivities of these measured using a thermogravimetric analyzer (TGA) and a
tubular reactor. SAP-limes performed better for HCl capture than commercial Trona and
quicklime but exhibited a lower capacity/reactivity for SO3 and sulfur dioxide (SO2)
removal. Partially calcined limestone had the maximum capacity for HCl removal at 70%
extent of calcination (EOC) and sorbents doped with calcium acetate hydrate
[(Ca(OAc)2·H2O)] had a greater than 30% higher capacity for HCl capture.
A techno-economic analysis for a SAP unit installed in a 550 MWe power plant burning
a high-Cl, high-S Illinois No. 6 coal was also conducted. The analysis assumed 70% EOC
during the sorbent production, 40% sorbent utilization during the acid gas removal, and a
90% HCl removal rate. The results revealed that the SAP technology is cost-effective for
on-site production of lime sorbents for HCl and SO3 removal. SAP’s overall cost of acid
gas removal is about 37% lower than a conventional dry sorbent injection (DSI) process.
Page 2
2
EXECUTIVE SUMMARY
The U.S. EPA’s final standards for mercury and other air toxics emissions from electric
generating units (EGUs), commonly referred to as the “Mercury and Air Toxics
Standards” (MATS) set emission limitations on hydrogen chloride (HCl) at 0.02 and
0.0004 lb HCl/MWh for coal-fired existing and new EGUs, respectively. Although HCl
in the flue gas can be efficiently captured in a flue gas desulfurization (WFGD) process
to meet the MATS’ requirement, there is no effective and economic approach to remove
the accumulated chlorides from the scrubber liquid. The application of WFGD facilities
in the utility industry to reduce sulfur dioxide (SO2) emission also leads to increasing
concerns on plume opacity caused by sulfur trioxide (SO3).
Dry sorbent injection (DSI) processes (Figure 1) using sorbents such as hydrated lime
and Trona are often employed to mitigate SO3, SO2, and HCl. The advantages of a DSI
process over a WFGD are: 1) lower capital and raw materials costs; 2) smaller
installation foot print; 3) ease of operation; and 4) more flexibility to fuel changes. A DSI
process upstream of the plant’s particulate control device (PCD), such as a baghouse, can
capture SO3 and HCl in the flue gas upstream of a WFDG. Sorbents may be injected at
different locations between the plant’s economizer and the PCD. Co-benefits of the
removal of HCl and SO3 include: 1) less equipment fouling and corrosion; 2) higher plant
efficiency by lowing air preheater operation temperature; and 3) higher efficiency for
mercury removal by activated carbon injection (ACI).
Figure 1. Block diagram of a SAP or DSI for acid gas removal in a typical power plant.
The Illinois State Geological Survey (ISGS) of the University of Illinois at Urbana-
Champaign, and Electric Power Research Institute (EPRI) have developed a patented
technology, sorbent activation process (SAP), for on-site production of sorbents to
removal pollutants in coal combustion flue gas. SAP can be used as a simple alternate to
DSI for pollutant mitigation. SAP technology allows for producing an acid gas sorbent at
a utility site and directly injecting the fresh sorbent into the flue gas upstream of the
plant’s PCD.
The project aimed at producing highly reactive calcium-based sorbents using a batch
reactor and the 5-10 lbs/hr SAP unit at the ISGS for the removal of SO3 and HCl in Illinoi
coal combustion flue gases. More than twenty sorbents were prepared and the reactivities
Page 3
3
of the sorbents were measured using a thermogravimetric analyzer (TGA) and a tubular
reactor. Performance and properties of the batch sorbents provided useful guidance to
prepare SAP sorbents. Sorbent characterizations included extent of calcination (EOC),
nitrogen adsorption for surface area, and particle size distribution. ISGS prepared lime
sorbents exhibited more than twice of the capacity of HCl capture in comparison to Trona
and quicklime at the same testing conditions, but exhibited a lower performance for SO3
and SO2 removal. The lab-synthesized sorbents had the best performance for HCl
removal when the EOC of the sorbents was at about 70%. Sorbents doped with calcium
acetate hydrate [Ca(OAc)2·H2O] had more than 30% higher capacity for HCl capture. A
combination of various sorbents may maximize HCl/SO3/SO2 removal by taking
advantage of the sorbents’ preferential adsorption on individual gases.
A techno-economic analysis study of SAP technology was performed for producing lime
sorbents for an on-site unit installed in a 550 MWe power plant burning a high-Cl, high-S
Illinois No. 6 coal. The analysis assumed producing a 70% EOC lime sorbent, 40%
sorbent utilization during the acid gas removal, and a 90% HCl removal rate. The results
from this techno-economic analysis revealed that SAP is a cost-effective technology for
producing lime sorbents for HCl and SO3 removal. SAP’s overall cost of acid gas
removal is about 37% lower than a conventional DSI process.
Page 4
4
OBJECTIVES
The project aimed at producing highly reactive calcium-based sorbents using a batch
reactor and the bench SAP unit at the ISGS for the removal of SO3 and HCl in Illinois
coal combustion flue gases. Specific objectives of the project include: 1) the
thermodynamic analysis of the reactions between selected sorbents and HCl/SO2/SO3 at
various process conditions; 2) the preparation of sorbents using a batch reactor and the
bench-scale SAP unit; 3) the evaluation of SAP sorbents using a TGA and a tubular
reactor in a typical Illinois coal flue gas; and 4) the techno-economic analysis of the on-
site sorbent production process installed in a 550 MWe power plant burning a high-S,
high-Cl Illinois coal.
The project tasks were as follows.
Task 1. Sorbent Thermodynamic Analysis A thermodynamic study was performed to identify operating conditions for achieving 90%
HCl removal. Corresponding equilibrium curves (partial pressure vs. temperature) were
developed for reactions between the sorbent and HCl, SO2, and SO3.
Task 2. Sorbent Preparation Using a Bench-Scale SAP
Over 20 sorbents were prepared using a batch reactor and the 5-10 lbs/hr bench-scale
SAP unit at the ISGS. Selected samples were characterized using various techniques such
as nitrogen adsorption for Brunauer-Emmett-Teller (BET) surface area and laser
scattering for particles size distribution.
Task 3. Sorbent Evaluation
The sorbents prepared in Task 2 were evaluated using a TGA and a tubular reactor with
acid gases in a typical Illinois coal flue gas. The temperature window to be examined in
the tests was based on the results obtained from the thermodynamic analysis in Task 1.
Task 4. Techno-Economic Analysis
A techno-economic analysis of the on-site sorbent production process for the removal of
SO3 and HCl in a 550 MWe power plant burning a high-S, high-Cl Illinois coal was
conducted. Capital and operating & maintenance (O&M) costs associated with the on-site
sorbent production and injection system were individually calculated.
Page 5
5
INTRODUCTION AND BACKGROUND
HCl is a hazardous atmospheric pollutant which has adverse effects on both human health
and the wider environment. The U.S. EPA restricts HCl emission to 0.02 and 0.0004 lb
HCl/MWh for existing and new EGUs respectively. [1] In Illinois and other Midwest
bituminous coals, the chlorine content can be as high as 0.54 wt% [2], well above 0.1 wt%
average chlorine content of all U.S. coals. Although wet scrubbers are primarily used for
the removal of SO2, they can effectively capture HCl (>95%) to satisfy the EPA’s
regulations on HCl emissions. However, the removal of HCl via a WFGD leads to
substantial accumulation of chlorides in the scrubber liquid, causing severe equipment
corrosion issues and extra cost to periodically purge the system with low chloride water.
A cost-effective approach to remove HCl upstream of a scrubber is necessary for abiding
by the EPA regulations and minimizing corrosion to the system.
The installation of a NOx removal unit and a WFGD allows a power plant to burn high-S,
high-Cl Illinois coal. The SO3 can only be partially captured at a particulate control
device and a wet scrubber. SO3 in the flue gas may form sub-micro aerosols, causing
public concerns on a visible plume. SO3 mitigation also improves mercury removal by
use of activated carbon injection [3] with a subsequent increase of energy efficiency as a
result of a lower air preheater operation temperature.
DSI is a common technology for the removal of SOx and HCl. (Figure 1) Compared to a
scrubbing system, this technology features 1) lower capital cost; 2) smaller foot print; 3)
user friendly operation; 4) lower raw materials cost; and 5) fuel flexibility. DSI is mostly
used as an alternative to SO2 removal technology for low sulfur coal combustion flue
gases. In a power plant application, the SAP can be used as an alternative to a DSI
system, combining on-site sorbent production from raw materials and sorbent injection
for pollutant mitigation.
In this project, we used a batch reactor and a bench scale SAP unit to produce highly
reactive lime sorbents and evaluated performance of these sorbents for the removal of
acid gases from flue gas at a desired temperature window selected through
thermodynamic analysis. The concentrations of acid gases are typical of a flue gas
generated from burning Illinois basin coal. A techno-economic analysis was conducted
for an on-site sorbent production process for acid gas removal installed in a 550 MWe
power plant burning a high-S, high-Cl Illinois No. 6 coal.
EXPERIMENTAL PROEDURES
Sorbent Preparation
More than twenty sorbents were prepared using a batch reactor (Tables 1 and 2) and the
bench scale SAP unit (Table 3). The performance and properties of the batch reactor-
made sorbents provided a guideline to the sorbent preparation using the SAP, a more
complicated unit than the batch reactor. A typical procedure using the batch reactor is as
follows:
Page 6
6
About one gram of the limestone was evenly placed on a 1.5”x2” flat metal sample boat.
The sample boat was inserted in the cold section of a horizontal quartz tube (2-in ID)
which was preheated to the reaction temperature by a Lindberg furnace. A metal boat was
chosen to facilitate heat transfer during the calcination. The quartz tube was purged with
N2 for 10 min before the sample boat was pushed to the hot zone of the furnace and kept
at this location for a predetermined time. The temperature was controlled and monitored
by a K-type thermocouple above the sample boat. A N2 flow of 250 sccm purged the
reactor during the entire process. At the end of the reaction time, the sample boat was
brought back to the cool section of the reactor for a few minutes and stored in a container
with desiccant. During this procedure, limestone was partially decomposed to CaO
(CaCO3 CaO + CO2). For sorbents with dopants, the dopants (calcium acetate hydrate
sodium acetate , sodium chloride, sodium carbonate) were added onto CaO/CaCO3
sorbent via the following approach: 1) Two grams of limestone and calculated amount of
dopants were added into 50 ml DI water and the slurry was vigorously stirred for 10
minutes; 2) The slurry was evaporated to a paste in the air and then dried in an oven at
105 °C overnight; 3) The solids were calcined at 900 °C in the presence of nitrogen.
Products obtained were sealed in bottles filled with nitrogen. The batch sorbents have
been summarized in Tables 1 and 2. Note that the BET surface areas in the tables are
based on the total sorbent weight, not the CaO content of the sorbents.
Table 1. Preparation conditions and properties of the batch sorbents.
Sample Temperature Residence
time EOC
BET surface
area Note
°C min % m2/g
IF001 1000 1.5 31 3.6 LS-AC3
IF002 1000 3 92 7.6 LS-AC3
IF003 900 3 45 4.9 LS-AC3
IF004 900 5 91 8.6 LS-AC3
IF005 800 3 11 2.4 LS-AC3
IF006 800 5 20 3.2 LS-AC3
IS020 900 5 86 10.7 LS-X8
IS021 900 10 98 9.2 LS-X8
IS022 900 15 97 10.7 LS-X8
Table 2. Preparation conditions and properties of the batch sorbents with dopants.
Sample Temperature Residence
time EOC Note
°C min %
IS008 900 3 43 2wt% NaCl/LS-AC3
IS013 900 3 23 2wt% Ca(OAc)2 /LS-AC3
IS014 900 3 66 2wt% NaOAc/LS-AC3
IS018 900 3 42 2wt% Na2CO3/LS-AC3
IS019 900 3 49 2wt% Na2SO3/LS-AC3
Page 7
7
The bench-scale SAP (Figure 2) is an L-shaped entrained-flow reactor with an inner
diameter of 6 inch. It has several ports for either monitoring the temperatures or injecting
limestone particles. A Krom Schroder BIC-65 burner, a pre-mix burner with a maximum
capacity of 70 kW is located about 22-inch upstream of the injection port and burnt
propane as the main heat source in this unit. After preheating the SAP, limestone was
injected into the reactor via a volumetric screw feeder that aided by a purge gas. Inside
the SAP, limestone was partially decomposed to CaO. Some of the products are expected
to be re-carbonated (CaO + CO2 CaCO3) in the tail gas before it can be collected at the
effluent of the SAP through a cyclone. Doped sorbents were produced by feeding
premixed limestone and dopant. Detailed preparation conditions and properties of the
SAP-made sorbents are given in Tables 3 and 4.
Figure 2. Front look of SAP and temperature port locations.
Table 3. Preparation conditions and properties of the SAP sorbents.
Sample
SAP operating temperatures Sorbent properties
Note T1 T2 T3 EOC
BET surface
area
°C °C °C % m2/g
IS010 677 581 464 13 2.0 LS-AC3
R11 980 768 629 61 5.0 LS-X8
R17 1044 813 648 67 7.3 LS-X8
R13 1110 886 727 81 5.9 LS-X8
R15 1146 944 773 86 4.1 LS-X8
R21 1137 914 752 86 4.0 2 wt.% Ca(OAc)2
/LS-X8
R31 1174 970 793 88 3.2 2 wt.% Na2SO3
/LS-X8
M11 936 681 526 67 4.2 LS-M6
M12 829 610 493 51 3.3 LS-M6
Page 8
8
Table 4. Median particle size of raw limestone and SAP sorbents.
Sample Median particle size
µm
LS-X8 18
R11 20
R17 21
R13 23
R15 24
R21 21
R31 24
LS-M6 9
M12 12
Sorbent Characterizations
Selected samples were subjected to various characterizations. BET surface area was
measured by N2 adsorption and desorption isotherms at -196 °C using a Monosorb
surface area analyzer (Quantachrome Inc.). Volume-based particle size distribution
(PSD) was measured using a laser scattering particle size analyzer (Horiba Instruments
Inc., LA-950). The instrument can measure particles ranging from 0.3 nm to 30 mm. The
sorbents were also subjected to full calcination using a TGA (Thermo Scientific
Versatherm) to determine the EOC of a sorbent.
Sorbent Evaluation
The performance of the sorbents for acid gas adsorption was evaluated using the TGA
and a tubular reactor. TAG is a common and effective tool to evaluate the performance of
lime sorbents for acid gases (HCl and/or SO2) [4-6] and CO2 capture [7-9] despite that
the reaction conditions including reaction time and gas-solid contact patterns are much
different in the TGA from a real applications. A typical procedure using the TGA for the
sorbent evaluation was as follows: 1) a ~4 mg of sample was heated from room
temperature to reaction temperature in N2 at 10 °C/min; 2) the sample was kept at
reaction temperature for 10 min; 3) N2 was switched to reacting gas such as 600 ppm
HCl, 3000 ppm SO2, or 105 ppm SO3 (balance N2) at 400 °C for 90 min unless otherwise
stated.
Performance of a selected sorbent and a commercial sorbent were evaluated using the
tubular reactor (1/4-in ID) with typical concentrations of acid gases in an Illinois coal
burning flue gas: 300 ppm HCl, 2500 ppm SO2, and 50 ppm SO3. The system included a
gas mixing section and a tubular reactor section. The gas mixing section included gas
cylinders and mass flow controllers (MFC) to generate a desired gas stream. SO3 was fed
into the gas line using a Perkin Elmer syringe pump. The gas line was heated with a
heating tape to prevent the SO3 from condensing. The reactor section included a tubular
reactor, a furnace, and a temperature controller (Omega CN7500). Gas breakthrough
Page 9
9
curves were obtained via monitoring gas concentrations in the effluent using a mass
spectrometry technique-based residue gas analyzer (MKS Cirrus 2). To enhance the
utilization of the sorbent, inert silica sand (210 to 297 µm) was used to disperse sorbent
particles at a ratio of 20 g-silica/g-sorbent. Roughly, the residence time of the reaction
gas within the sorbent bed was extended for 20 times. Moreover, channeling effect was
also effectively reduced, improving better gas-particle contact effect.
RESULTS AND DISCUSSION
Thermodynamic Analysis
Thermodynamic analysis methodology
The thermodynamic analysis study was focused on reactions between CaO and acid
gases. Dominant products of a reaction system at equilibrium were identified by
minimizing the change of Gibbs free energy of the reaction system according to Eq. 1:
ΔG = ∑ (µio + RTlnϕiPi) + ∑ (µj
o + RTlnaj) (Eq. 1)
where ΔG is the change of Gibbs free energy, i is a gas component, j is a pure condensed-
phase component, n is the number of moles, µo is the standard chemical potential, ϕ is the
fugacity coefficient, P is pressure, T is temperature, R is the gas constant, and a is the
activity of the reaction species. The equilibrium constant K of a reaction was calculated
according to Eq. 2:
K = exp[–ΔG0/(RT)] (Eq. 2)
Equilibrium pressures of a reaction at various temperatures were calculated using Eq. 3:
Kp = ∏ Pia (Eq. 3)
where Kp is the equilibrium constant of a reaction and a is the stoichiometric number of
gas component i.
Adsorption reaction
The dominate reactions between a CaO sorbent and acid gases identified from the
thermodynamic analysis are as follows (100-900 °C):
CaOHCl (s) = CaO (s) + HCl (g) (Rx. 1)
CaCl2 (s) + H2O (g) = CaO (s) + 2HCl (g) (Rx. 2)
CaSO3 (s) = CaO (s) + SO2 (g) (Rx. 3)
CaSO4 (s) = CaO (s) + SO3 (g) (Rx. 4)
The equilibrium pressure as a function of temperature obtained via Eq. 3 is given in
Figure 3. Above each equilibrium curve, the adsorption reactions occur and the acid gas
is captured by the sorbent. Below each equilibrium curve, the desorption reactions
(reverse reaction) proceed.
Temperature window
In an ideal operation window, a sorbent reacts with both HCl and SO3 but not SO2 since
SO2 can be captured by a usually less expensive approach of flue gas desulfurization
process. When the temperature is lower than 790 °C, the tendency to react with the acid
Page 10
10
gases by CaO is HCl<SO2<SO3 thermodynamically as shown in Figure 3. At
temperatures above 790 °C, forward reactions of Rx. 1 and 3 and reverse reactions of Rx.
2 and 4 occur, leading to HCl and SO3 but not SO2 capture by the sorbents.
Figure 3. Equilibrium pressures of acid gases using a CaO sorbent.
Figure 4. Adsorption temperature window for a 90% HCl removal from a 600 ppm HCl
stream using a CaO sorbent. Left: Rx. 1 mechanism; Right: Rx. 2 mechanism.
To capture 90% of HCl from a flue gas containing 600 ppm HCl, the thermodynamic
analysis suggests that the adsorption operation should occur below 510 °C (Figure 4).
Selectivity
The selectivity of HCl or SO3 over SO2 by a lime sorbent is defined as the ratio of
equilibrium pressure of HCl (PHCl/PSO2) or SO3 (P SO3/PSO2) over SO2. The lower the ratio
is, the better the sorbent adsorbs HCl or SO3. Figure 5 exhibits that regardless of which
HCl adsorption mechanism dominates, a higher reaction temperature leads to a lower
ratio of the equilibrium pressures of HCl over SO2, suggesting that the preferential
capture of HCl over SO2. With decreasing reaction temperature, SO3 adsorption is
preferred over SO2 (Figure 6). In this project, major tests of the sorbents were run at
400 °C to keep high selectivity of HCl over SO2.
200 400 600 8001E-14
1E-11
1E-8
1E-5
0.01
790 C
Ac
id g
as
eq
uilib
riu
m p
res
su
re (
ba
r)
T (C)
CaOHCl=CaO+HCl
CaCl2+H2O=CaO+2HCl
CaSO3=CaO+SO2
CaSO4=CaO+SO3
200 400 600 8001E-14
1E-11
1E-8
1E-5
0.01600 ppm
Desorption
HC
l e
qu
ilib
riu
m p
res
su
re (
ba
r)
T (C)
CaOHCl=CaO+HCl
Adsorption
60 ppm
520 C
200 400 600 8001E-14
1E-11
1E-8
1E-5
0.01600 ppm
Desorption
HC
l e
qu
ilib
riu
m p
res
su
re (
ba
r)
T (C)
CaCl2+H2O=CaO+2HCl
Adsorption
60 ppm
510 C
Page 11
11
Figure 5. Adsorption selectivity of HCl over SO2 using a CaO sorbent. Left: Rx. 1
mechanism; Right: Rx. 2 mechanism.
Figure 6. Selective adsorption of SO3 over SO2 using a CaO sorbent.
Sorbent Preparation and Characterizations
Particle size
Particle size measurement shows that the median particle size of limestone AC-3 (IF000)
was 4.0 µm and 90% (volume) of the particles were smaller than 6.7 µm (Figure 7).
Small sizes are beneficial for fast calcination because shorter time exposure to high
temperature is required during calcination. High temperature calcination is known to lead
to particle sintering and performance deterioration. The results in Figure 7 show that the
produced sorbents had larger size than the raw limestone (IF000). There are two possible
explanations: particle sintering due to high temperature and expanding due to the fast
release of CO2 during the calcination. The median size of sample IF002 (9.1 µm) was
about 60% larger than that of sample IF004 (5.7 µm) even though these samples had
comparable EOCs. The larger size of the former sample was probably caused by sintering
at the higher preparation temperature.
200 400 600 8001
100
10000
1000000
1E8
PH
Cl/P
SO
2
T (C)
2CaO+HCl+SO2
200 400 600 8000.1
10
1000
100000
1E7
1E9
1E11
PH
Cl/P
SO
2
T (C)
2CaO+2HCl+SO2
200 400 600 800
1E-17
1E-14
1E-11
1E-8
PS
O3
/PS
O2
T (C)
2CaO+SO3+SO2
Page 12
12
Figure 7. Particle size distributions of the limestone (IF000) and sorbents. Left: Size
distribution vs. diameter; Right: Accumulated size distribution vs. diameter.
The sizes of the SAP sorbents exhibited the similar trend: particles were bigger when the
sorbents were prepared at higher temperature (Figure 8). Median particle sizes of the
sorbents from LS-X8 without dopants increased linearly with the preparation temperature
(Figure 8). The median particle size of the sorbent R15 (prepared at 944 °C, T2) was 24
µm, which is 25% greater than that of the sorbent R11 (prepared at 768 °C, T2). The
median size of sorbent R15 was also 33% greater than that of the raw limestone at 18 µm.
Particles tend to sinter when temperature is higher than their Tammann temperature,
which is estimated as 0.52 times of the particles’ melting point temperature (K). The
higher temperature caused sintering which led to larger particle sizes.
Figure 8. Median particle size and EOC of the SAP-sorbents (R11, R17, R13, and R15)
made at different temperatures with the same limestone.
Dopants also affected particle sizes. Sorbents R15, R21, and R31 were prepared at similar
SAP conditions but with different dopants. The sorbents had similar EOCs (86% - 88%).
However, they showed different size distributions and their median sizes ranged from 21
to 24 µm (Table 4 and Figure 9). Median size of sorbent R21 was one eighth smaller than
1 100
5
10
15
20
q (
%)
Diameter (m)
IF000
IF001
IF002
IF003
IF004
IF005
IF006
1 100
20
40
60
80
100
Un
ders
ize (
%)
Diameter (m)
IF000
IF001
IF002
IF003
IF004
IF005
IF006
50%
60%
70%
80%
90%
100%
17
19
21
23
25
700 800 900 1000
EOC
Med
ian
Siz
e (µ
m)
Temperature (°C)
Page 13
13
those of sorbents R15 and R31. The smaller particles of sorbent R21 might be caused by
the interactive effect between dopant Ca(OAc)2 and the lime particles when the dopant
decomposed during its multiple-step decomposition to CaO.
EOC
EOC is an index of the degree that a raw material is calcined. During the sorbent
preparation, the limestone was partially decomposed to CaO. The sorbents were a
mixture of un-calcined limestone (CaCO3) and freshly produced CaO. To determine the
EOC, the sorbents were subjected to a full calcination using the TGA. The EOC was
calculated from the weight loss during the fully calcination using the following equation:
EOC = (1-wL/wL0)/(1-wL) (Eq. 4)
where wL and wL0 are weight losses during the fully calcination of the sorbent and the
raw limestone, respectively, measured in the TGA calcination. Higher temperature and
longer residence time during the sorbent preparation resulted in higher EOC (Tables 1
and 3). At higher temperature, the limestone had faster kinetics and stronger tendency to
decompose, resulting in the greater EOC. For examples, the batch samples IF002 and
IF004 had EOCs of 92% and 91%, respectively, much greater than those of the samples
IF005 and IF006 (Table 1). The only difference among the four samples is that the former
two were prepared at higher temperatures. Figure 8 demonstrates that the EOC of the
SAP sorbents increased proportionally with the sorbent preparation temperature.
Figure 9. Particle size distributions of SAP-sorbents with different dopants prepared
from LS-X8.
BET surface area
Limestone partially decomposed to CaO during the preparation, releasing CO2 and
producing porous structure that increased the surface area of the sorbents. Surface area of
six batch samples (IF001-IF006) was plot against their EOC. The surface area of the
sorbents increased linearly with the increasing EOC (Figure 10).
1 10 100 1000
0
4
8
12
16
q (
%)
Diameter (m)
R15
R21
R31
Page 14
14
Figure 10. BET surface area vs. EOC (Batch sorbents from LS-AC3).
Sorbent Evaluation
Sorbent reactivity for HCl adsorption using the TGA
Thermodynamic analysis discussed previously implies that the adsorption should be
carried out at temperatures lower than 510 °C to achieve 90% HCl removal by CaO in a
600 ppm HCl stream and higher temperature is beneficial to selectivity of HCl over SO2.
HCl adsorption capacities of the batch sorbents without dopants are summarized in
Figure 11. IF003 exhibited the largest capacity while IF005 and IF006 showed much
smaller capacities than the other sorbents. The smaller capacities of the latter samples
might be due to their limited surface area accessible by HCl molecules. Although IF002
and IF004 possessed much greater surface area than samples IF001 and IF003, the latter
demonstrated better performance. During sorbent preparation, IF001 and IF003 were
exposed to high temperatures at a shorter period than IF002 and IF004 at 1000 and
900 °C. A greater surface area of a sorbent was due to the formation of porous structures
during CaCO3 calcination. The smaller surface area of samples IF001 and IF003 was due
to the smaller degree of calcination. Limestone decomposition happens gradually from
exterior to interior according to shrinking core model. Partial calcination of IF001 and
IF003 should occur majorly at the exterior of the particles. Therefore, their surface area
might be more concentrated on the exterior part of the particles. IF002 and IF004
developed with greater surface area due to their longer calcination time which calcined
over 90% of the precursor particles. However, during the diffusion controlled reaction,
HCl molecules were consumed up at the exterior before they were able to reach the
surface CaO deep inside the particles during the 90 min reaction period. This indicates
that only partial of the available surface CaO was utilized for HCl adsorption. Since
residence time of sorbent particles in a duct DSI process is only a few seconds, it is
reasonable to predict that only a limited portion of the surface area will have the
opportunity to react with an acid gas. Therefore, it may not be necessary to fully calcine
limestone to produce a high surface area product. Suitable sorbents that are partially
calcined may achieve comparable performance to a fully calcined one but at lower cost.
0 20 40 60 80 100
2
4
6
8
BE
T s
urf
ace a
rea (
m2/g
)
EOC (%)
Page 15
15
Figure 12 exhibits a more clear view on how the adsorption behavior of IF004 and IF006
evolved with respect to different surface areas. In the initial 10 min, IF004 (8.6 m2/g) and
IF006 (3.2 m2/g) had comparable adsorption rates, indicating that HCl was reacting with
CaO on the most exterior part of the particles. After that, the reaction rate of IF006
gradually decreased and became negligible after about 120 min, indicating HCl diffused
through the CaCl2/CaOHCl product layer. IF004 maintained a faster rate during the entire
900 min adsorption period due to its greater surface area. The somehow slower reaction
rate at later stage could be attributed to larger diffusion resistance of HCl into the interior
pores of the sorbent particles.
Sorbents from different types of limestone showed different performance for HCl
adsorption. Figure 13 showed the performance of two batch sorbents from two types of
limestone at the same conditions. IS020 from LS-X8 exhibited high HCl capture capacity
of about 0.20 g-HCl/g-sorbent while IF004 from LS-AC3 only 0.11 g-HCl/g-sorbent.
Figure 14 presented the performance of two sorbents from two types of limestone made
0 30 60 900.00
0.03
0.06
0.09
0.12
1 2
0.000
0.001
Ca
pa
cit
y (
g-H
Cl/
g-s
orb
en
t)
Time (min)
IF 001
IF 002
IF 003
IF 004
IF 005
IF 006
Cap
acit
y (
g-H
Cl/
g-s
orb
en
t)
Time (min)
IF 001
IF 002
IF 003
IF 004
IF 005
IF 006
0 200 400 600 8000.0
0.1
0.2
0.3
0.4
0.5
0 10 20 30 40 50 60 70 80
0.00
0.01
0.02
0.03
0.04
Cap
acit
y (
g-H
Cl/
g-s
orb
en
t)
Time (min)
IF 004
IF 006
Ca
pa
cit
y (
g-H
Cl/
g-s
orb
en
t)
Time (min)
IF 004
IF 006
0 20 40 60 80
0.0
0.1
0.2
Ca
pa
cit
y (
g-H
Cl/g
-so
rbe
nt)
Time (min)
IS-020
IF-004
0 20 40 60 80
0.00
0.02
0.04
0.06
0.08
0.10
Ca
pa
cit
y (
g-H
Cl/g
-so
rbe
nt)
Time (min)
R11
M11
Figure 12. HCl adsorption over
extended time at 400 °C. Figure 11. HCl adsorption of
various sorbents.
Figure 13. HCl adsorption capacity of
sorbents made at the same conditions
from different types of limestone: LS-
X8 (IS020) and LS-AC3 (IF004).
Figure 14. HCl adsorption capacity of
sorbents made at the similar conditions
from different types of limestone: LS-
X8 (R11) and LS-M6 (M11).
Page 16
16
at similar conditions from the SAP. M11 prepared from LS-M6 had a capacity of 0.088
g-HCl/g-sorbent, much greater than that of R11 despite that the latter had a greater
surface area and similar EOC. The smaller particle sizes of LS-M6-based sorbents may
have contributed to the better performance.
Sorbents with different dopants shows different performance. HCl adsorption reactivity
of the batch sorbents with dopants is summarized in Figure 15. IS014 with dopant
NaOAc exhibited the highest capacity, about 0.097 g-HCl/g-sorbent in the testing,
followed closely by IS013 with dopant Ca(OAc)2 at 0.075 g-HCl/g-sorbent. Sorbents
doped with Na2SO3 (IS019), NaCl (IS015), and Na2CO3 (IS018) had capacities between
0.03 and 0.55 g-HCl/g-sorbent. It is interesting to note that both IS014 and IS013 were
doped with acetate salts. Compared to the sorbents IS014 - IS019, IS013 had the lowest
EOC (23%). The multiple decomposition steps of calcium acetate hydrate
Ca(CH3COO)2·H2O might have produced porous structure that was beneficial to the gas-
solid reaction [10].
Figure 15. HCl adsorption capacity of sorbents with dopants made from the batch
(Left) and SAP (Right)
In Figure 15, the performance of SAP sorbent (R15) without dopant is compared with the
sorbents (R21 and R31) with dopants that show good performance from batch sorbents.
Doped with Ca(OAc)2 and Na2SO3, R21 and R31 exhibited 31% and 36%, respectively,
higher capacities than R15 for HCl adsorption. Note that all three sorbents were prepared
at similar SAP conditions but R15 did not have any additives. The results demonstrate
that dopants Ca(OAc)2 and Na2SO3 effectively enhanced performance of lime sorbents
for HCl capture. For comparison, the performance of commercial Trona (EnProveTR,
Natronx) for HCl adsorption was also included. The dopant enhanced SAP sorbents
exhibited much better performance than the commercial Trona at the test conditions
either at the initial reaction period or the entire extended reaction stage, indicating the
advantage of SAP sorbents over the Trona for HCl mitigation.
An interesting observation is that a maximum exists when sorbent capacity for HCl
removal vs. EOC is plotted (Figure 16). At higher preparation temperature, the sorbent
had higher EOC as shown in Figure 8, which means that the sorbent possessed greater
percentage of CaO. However, the capacity of the sorbent for capturing HCl did not
always increase with the EOC. There was a maximum capacity for HCl capture at about
70% EOC for both the batch and SAP sorbents. When the EOC was greater than about
0 20 40 60 80
0.00
0.02
0.04
0.06
0.08
0.10
Ca
pa
cit
y (
g-H
Cl/g
-so
rbe
nt)
Time (min)
IS-008
IS-013
IS-014
IS-018
IS-019
0 20 40 60 80
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Ca
pa
cit
y (
g-H
Cl/
g-s
orb
en
t)
Time (min)
R15
R21
R31
Trona
Page 17
17
70%, the capacity started to decrease with the EOC. Sorbent prepared at higher
temperature or longer residence would sinter more as discussed above. Despite a sorbent
may have a greater EOC (more CaO per gram sorbent), the amount of active CaO sites
accessible by the gaseous HCl may be less than other sorbents at smaller EOCs. The
existing of a maximum capacity is useful to determine optimal conditions for producing
sorbent using the SAP. Energy can be saved by partially calcining limestone while
achieving better performance for acid gas removal. Assuming limestone are sphere
particles, at 70% EOC, only the exterior 1/3 (by diameter) of the limestone particles was
calcined. (Figure 16)
Figure 16. HCl adsorption capacity vs. EOC (Left, batch sorbents IF001-006; Right ,
SAP sorbents R11, 17, 13, and 15).
The HCl capacities of a SAP sorbent (R17) and the commercial Trona were compared at
different testing temperatures (Figure 17). R17 had capacities of 0.014, 0.072, and 0.127
g-HCl/g-sorbent at 200, 300, and 400 °C. The Trona showed very limited capacity for
HCl adsorption at all of the testing temperatures (<0.02 g-HCl/g-sorbent), demonstrating
that the SAP lime sorbents are better than the Trona for HCl adsorption, especially when
the temperature is at 300 °C or higher.
Figure 17. HCl adsorption capacity of a SAP sorbent (R17, Left) and the commercial
Trona (Right) at 200, 300, and 400 °C.
0
0.04
0.08
0.12
10% 30% 50% 70% 90% 110%
Cap
acit
y (g
-HC
l/g-
sorb
en
t)
EOC
0.00
0.04
0.08
0.12
50% 60% 70% 80% 90%
Ca
pa
cit
y (
g-H
Cl/
g-s
orb
en
t)
EOC
0 30 60 900.00
0.03
0.06
0.09
0.12
Ca
pa
cit
y (
g-H
Cl/
g-s
orb
en
t)
Time (min)
400 C
300 C
200 C
0 20 40 60 80
0.00
0.01
0.02
0.03
0.04
0.05
Ca
pa
cit
y (
g-H
Cl/g
-so
rbe
nt)
Time (min)
400 oC
300 oC
200 oC
Limestone particles calcined
(orange) at 70% EOC
Page 18
18
Sorbent reactivity for SO2 adsorption using the TGA
Selected sorbents were tested for SO2 capture. The sorbents IS020 - IS022 showed
capacities between 0.065 and 0.070 g-SO2/g-sorbent (Figure 18). All of the three sorbents
displayed two distinguished reaction states: a fast reaction in the initial 15 minutes
followed by a much slower one. The Trona, however, exhibited a capacity of 0.106 g-
SO2/g-sorbent at the same conditions and it did not show a slow reaction stage at the
testing conditions (Figure 20). Even at 200 and 300 °C, the Trona captured 0.05 and 0.06
g-SO2/g-sorbent.
The SAP sorbents M11 and M12 also exhibited low capacity for SO2 capture (Figure 20)
while had high capacity for HCl adsorption (Figure 14). Note that the Trona had an SO2
capacity of 0.106 g/g-sorbent at 400 °C (Figure 19), an order greater than those of the
SAP sorbents, and an HCl capacity of 0.016 g/g-sorbent (Figure 17), an order smaller
than those of the SAP sorbents. These suggest that the SAP lime sorbents be effective for
HCl adsorption but not SO2.
0 20 40 60 80
0.00
0.02
0.04
0.06
0.08
0.10
Cap
acit
y (
g-S
O2/g
-so
rben
t)
Time (min)
IS-020
IS-021
IS-022
0 20 40 60 80
0.00
0.03
0.06
0.09
0.12
Ca
pa
cit
y (
g-S
O2/g
-so
rbe
nt)
Time (min)
Trona 300C-200C
Trona 300C-300C
Trona 300C-400C
0 20 40 60 80
0.000
0.004
0.008
0.012
0.016
Ca
pa
cit
y (
g-S
O2/g
-so
rbe
nt)
Time (min)
M12
M11
0 30 60 900.00
0.03
0.06
0.09
0.12
Ca
pa
cit
y (
g/g
-so
rbe
nt)
Time (min)
R17 HCl
R17 SO2
Quicklime HCl
Quicklime SO2
Figure 19. SO2 adsorption capacity
of the Trona at 200, 300, and 400 °C
Figure 18. SO2 adsorption capacity
of sorbents made from LS-X8.
Figure 20. SO2 adsorption capacity of
SAP sorbents from LS-M6.
Figure 21. Comparison of SAP lime
sorbent (R17) with a commercial
quicklime for SO2 and HCl adsorption
Page 19
19
In Figure 21, a SAP sorbent (R17) without any dopant is compared with commercial
quicklime. While the SAP sorbent exhibited faster reaction for HCl adsorption than the
quicklime, it was slower for SO2 removal. This reaction behavior is wanted since the lime
sorbent was expected to capture HCl, but not SO2, which can be removed by a usually
less expensive approach, a wet scrubber. When both HCl and SO2 are required to be
removed from a gas stream, a combined utilization of the SAP sorbents and a commercial
Trona or quicklime is expected to have a better performance.
Dopants also showed effect on the SO2 adsorption. SO2 adsorption capacities of the batch
sorbents with various dopants are summarized in Figure 22. Sorbent with dopant NaOAc
(IS014) captured 0.038 g-SO2/g-sorbent. Sorbents with dopants of Ca(OAc)2 (IS013),
Na2CO3 (IS018), or Na2SO3 (IS019) captured less than 0.02 g-SO2/g-sorbent. Sorbent
IS013 with dopant Ca(OAc)2 captured 0.075 g-HCl/g-sorbent (Figure 15) while only
moderate SO2, making it a promising candidate to selectively remove HCl over SO2.
Sorbents with dopants of Ca(OAc)2 and Na2SO3 were prepared using the SAP. The
performance of the sorbents was compared with the commercial Trona and the sorbent
R15 which was prepared at similar conditions using the SAP but no dopant (Figure 23).
R31 doped with Na2SO3 had a capacity of 0.016 g-SO2/g-sorbent. R21 with Ca(OAc)2
and R15 without any dopants had capacities of 0.010 and 0.008 g-SO2/g-sorbent,
respectively. Although R21 and R31 had 31% and 36% higher capacities than R15 for
HCl adsorption, they had 25% and 100%, respectively, higher capacities than R15 for
SO2 adsorption. This suggests that dopant Ca(OAc)2 not only enhanced the sorbent’s
capability to capture HCl but also enhanced its selectivity on HCl over SO2. The
enhancement of HCl adsorption over SO2 by R21 (31% vs. 25%) might due to the
structure change caused by Ca(OAc)2 that promotes diffusion of reaction gas to the active
CaO sites. The enhancement SO2 adsorption over HCl by R31 (100% vs. 36%) with
Na2SO3 might due to seeding effect of SO2-
that is preferential to SO2 adsorption,
resulting in a significant enhancement for SO2 adsorption.
0 20 40 60 80
0.00
0.01
0.02
0.03
0.04
0.05
Ca
pa
cit
y (
g-S
O2/g
-so
rbe
nt)
Time (min)
IS-013
IS-014
IS-018
IS-019
0 20 40 60 80
0.000
0.004
0.008
0.012
0.016
0 20 40 60 80
0.000
0.004
0.008
0.012
0.016
Ca
pa
cit
y (
g-S
O2/g
-so
rbe
nt)
Time (min)
R15
R21
R31
0 20 40 60 800.00
0.03
0.06
0.09
0.12
Ca
pacit
y (g
-SO
2/g
-so
rben
t)
Time (min)
Trona
Ca
pa
cit
y (
g-S
O2/g
-so
rbe
nt)
Time (min)
Figure 22. SO2 adsorption capacity of
sorbents with dopants.
Figure 23. SO2 adsorption capacities of
sorbents with dopants: R21 with 2 wt.%
Ca(OAc)2 and R31 with 2 wt.% Na2SO3.
Page 20
20
Sorbent reactivity for SO3 adsorption using the TGA
A SAP sorbent (IS010) and the Trona were compared for SO3 adsorption. After one hour
of reaction, IS010 showed a capacity of 0.0017 g-SO3/g-sorbent at 400 °C while the
Trona sorbent exhibited a capacity of 0.005 g-SO3/g-sorbent at the same conditions
(Figure 24). Noting that the SAP sorbents showed much better HCl capacity, a combined
utilization of SAP sorbents and a Trona is expected to have a better performance when
both HCl and SO3 are required to be removed from a gas stream.
Figure 24. SO3 adsorption capacity of SAP and Trona sorbents at 400 °C.
Sorbent reactivity for acid gas adsorption using the tubular reactor
A SAP sorbent (R17) and a commercial quicklime sorbent were evaluated with acid
gases typical in an Illinois basin coal combustion flue gas (300 ppm HCl, 2500 ppm SO2,
and 50 ppm SO3 in N2) using the tubular reactor. Though no HCl and SO3 breakthrough
was observed during the 100 min tests, SO2 in the effluent almost reached a plateau
(Figure 25), indicating the sorbents favored reactions with HCl and SO3. An interesting
observation is that the amount of SO2 captured by R17 was about half of that by the
quicklime in both the TGA and tubular reactor tests (Table 5).
Figure 25. Evolution of HCl and SO2 in the effluent of the tubular reactor using acid
gases typical in an Illinois coal combustion flue gas.
0 20 40 60 80
0.00
0.01
0.02
0.03
0.04
0.05
Ca
pa
cit
y (
g/g
-so
rbe
nt)
Time (min)
IS-010 HCl
IS-010 SO3
Trona HCl
Trona SO3
Page 21
21
Table 5. Sorbent utilization and weight change for HCl and SO2 adsorption tested with
the TGA and tubular reactor.
SO2 HCl
Sorbent utilization using
the tubular reactor
Capacity using
the TGA, g/g
Capacity using the
TGA, g/g (at 70 min)
R17 17% 0.028 0.114
Quicklime 30% 0.063 0.067
Techno-Economic Analysis of SAP for Acid Gas Removal
A techno-economic analysis was conducted for a conceptual on-site SAP process to
remove acid gases in the flue gas of a 550 MWe power plant burning high-Cl, high-S
Illinois No. 6 coal. The acid gas removal cost was compared with that of a traditional DSI
process. Sensitivities of the cost of the SAP to sorbent utilization rate and limestone price
were also addressed.
Process description and assumptions
A schematic diagram of a power plant using either a SAP or DSI system for acid gas
mitigation is presented in Figure 1. Sorbent powders can be injected using either the SAP
or DSI at different locations between the furnace boiler and the inlet of the plant’s
particulate control device. In the SAP technology, lime sorbents are produced on-site and
injected directly into the plant’s flue gas. The process takes advantage of on-site coal and
limestone supplies, handling facilities, and energy sources. Pulverized coal and limestone
are separately injected into the combustion and sorbent activation zones of the SAP. In
the SAP, micronized (<45 microns) limestone particles are thermally decomposed to lime
in 1 to 3 seconds. The freshly produced lime sorbent particles are injected directly into
the flue gas. Essentially, a SAP unit is a combination of a limestone calciner unit and a
DSI unit (Figure 26). Thus, the cost of the SAP system was estimated to be a combined
cost of the two units.
A 550 MW net output subcritical unit (Case 9 in Reference [11]) was selected for the
analysis. Selected properties of an Illinois No. 6 coal used in this case are listed in Table
6. In Case 9, Illinois No. 6 coal and primary air are introduced into the boiler through the
wall-fired burners and the flue gas exits the boiler through the selective catalytic reactor
and is cooled to 169 °C in the air preheater before passing through a baghouse for
particulate removal. Mass balance, flow rates, and assumptions of the sorbent
performance in the SAP are given in Tables 7 and 8. Because no SO3 concentration is
reported in case 9 in [11], it was assumed that 98.5% and 1.5% of the total sulfur in the
coal was converted to SO2 and SO3, respectively, generating a SO2 concentration of 2,102
ppm. The amounts of the limestone and coal feed rates were calculated using the process
conditions and assumptions listed in Tables 6 to 9.
Page 22
22
Figure 26. SAP (left) and traditional DSI (right) for acid pollutant control.
Table 6. Properties of Illinois No. 6 coal. [11]
Item As Received Dry
HHV, kJ/kg 27,113 30,506
Chlorine, wt.% 0.29 0.33
Sulfur, wt.% 2.51 2.82
Ash 9.70% 10.91%
Table 7. Mass balance and flow rate. [11]
Item Value Unit
Coal feed rate 198,391 kg-coal/hr
Volumetric flue gas rate before the baghouse 72,904 kg-mol/hr
Cl flow rate before the baghouse 575 kg-Cl/hr
S flow rate before the baghouse 4,980 kg-S/hr
HCl concentration before the baghouse 222 ppmv
SO2 concentration before the baghouse 2,102 ppmv
SO3 concentration before the baghouse 32 ppmv
Fly ash rate 15,390 kg-ash/hr
Table 8. Mass balance and assumptions for SAP.
Item Value Unit
Sorbent utilization by molar 40%
HCl removal rate 90%
SO3 and SO2 removal rate 40%
Limestone feed rate 17,385 kg-LS/hr
Limestone calcination rate by the SAP 70%
Total sorbent produced by the SAP 12,030 kg-sorbent/hr
Limestone silo
Limestone convey
Limestone feeder
Limestone blower
To sorbent injection point
Coal
Sorbent silo
Sorbent convey
Sorbent feeder
Sorbent blower
To sorbent injection point
Air
Page 23
23
Table 9. Mass balance and assumptions of the caliner unit of the SAP.
Item Value Unit
Heat loss 10%
Coal feed rate 1,454 kg-coal/hr
Air coal ratio 0.347 kg-mol-air/kg-coal
Cost methodology and assumptions
Cost is categorized as total project cost (TPC), operating and maintenance cost (O&M),
and equivalent annual cost (EAC). The cost analysis of the DSI unit of the SAP system
was based on the IPM model for a DSI process [12]. The cost analysis of the calciner unit
of the SAP system without an injection system, was based on a pulverized coal-fired
boiler model [13]. Basic parameters and assumptions for the cost calculation are given in
Tables 10 and 11.
Table 10. Cost calculation assumptions.
Variable Label Value Unit Note
Net Output A 550 MWe
Retrofit factor B 1
Limestone feed rate M 17.385 tonne/hr
Limestone cost N 30 $/tonne
Coal cost 52 $/tonne
Sorbent waste rate P 16.8 tonne/hr Based on the final products of
CaSO3, CaSO4, and CaCl2
Aux power Q 0.63 % M*20/A
Fly ash waste rate R 15.4 tonne/hr
Waste disposal cost S 30 $/tonne
Aux power cost T 0.06 $/kWh
Operating labor rate U 60 $/hr Labor costs including all
benefits
Effluent temperature 400 °C Including both sorbent and flue
gas exiting the furnace
TPC cost of the DSI unit of the SAP is a sum of the capital, engineering, and construction
cost (CECC) as well as financing expenditures. Details of these items are presented in
Table 11.
Page 24
24
Table 11. Total project cost of the DSI unit of the SAP.
Item Label Value Unit Note
BM 16,911,927 $ 7,516,000*(B)*(M)0.284
BM 31 $/kW
Engineering and construction
management costs
A1 845,596 $ 5% of BM
Labor adjustment for 6 x 10
hour shift etc.
A2 845,596 $ 5% of BM
Contractor profit and fees A3 845,596 $ 5% of BM
CECC 19,448,716 $ BM+A1+A2+A3
CECC 35.4 $/kW
Owner’s home office costs
(owner’s engineering,
management, and procurement)
B1 972,436 $ 5% of CECC
Allowance for funds used during
construction
B2 0 $ 0% of the CECC as the
projects are to be
completed <1 year
TPC 20,421,152 $ CECC+B1+B2
TPC 37.1 $/kW
O&M cost of the DSI unit of the SAP includes the fixed (FO&M) and variable (VO&M)
components. The FO&M is further subdivided into the costs associated with additional
operations (FO&MO), maintenance labor and materials (FO&MM), and administrative
and support labor (FO&MA). The VO&M is composed of the reagent cost (VO&MR),
which is a function of the sorbent capacity and acid gas concentrations, waste disposal
cost (VO&MW), and additional auxiliary power required (VO&MP). The disposal cost
accounts for both the sorbent and the fly ash waste. Summary of the O&M calculations is
given in Table 12.
Table 12. O&M cost of the DSI unit of the SAP.
Variable Value Unit Note
FO&MO 0.227 $/(kW yr) (1 additional operator)*2080*U/(A*1000)
FO&MM 0.307 $/(kW yr) BM*0.01/(A*B*1000)
FO&MA 0.010 $/(kW yr) 0.03*(FO&MO+0.4*FO&MM)
Total FO&M 0.54 $/(kW yr) FO&MO+FO&MM+FO&MA
Total FO&M 299,693 $/yr
VO&MR 0.49 $/MWh M*R/A
VO&MW 2.55 $/MWh (N+P)*S/A
VO&MP 0.38 $/MWh Q*T*10
Total VO&M 3.42 $/MWh VO&MR+VO&MW+VO&MP
Total VO&M 16,478,358 $/yr
Total O&M 16,778,050 $/yr
Page 25
25
TPC and O&M costs of the limestone calciner unit of the SAP are summarized in Table
13. The total O&M cost is overestimated because the SAP unit does not have water and
steam tubes that are included in the boiler model.
Table 13. Cost evaluation for the limestone calciner unit of the SAP.
Item Label Value Unit Note
Equipment cost C1 1,672,679 $ 1979 dollars
Direct installation cost C2 664,390 $ 1979 dollars
indirect installation cost C3 550,217 $ 1979 dollars
TPC 2,887,286 $ C1+C2+C3
TPC 8,672,005 $ 2010 dollars
TPC 15.8 $/kW
Total VO&M D1 100,698 $/yr 1979 dollars, excluding
coal cost
Total FO&M D2 539,087 $/yr 1979 dollars
Total O&M 639,785 $/yr D1+D2, excluding coal
cost
Total O&M D0 1,921,603 $/yr 2010 dollars, excluding
coal
Coal E0 662,157 $/yr
Total O&M 2,583,760 $/yr D0+E0, 2010 dollars
The total project and O&M cost of separate units of the SAP is summarized in Table 14.
The TPC and O&M costs of the DSI unit are 235% and 649% more than their
counterparts of the calciner unit. A much higher cost of O&M in the DSI unit is
associated with greater amounts of limestone handling and processing than that of coal in
the calciner unit (17,385 vs. 1,454 kg/hr).
Table 14. Cost summary of the SAP.
TPC, $ TPC, $/kW Total O&M, $/yr
DSI unit 20,421,152 37.1 16,778,050
Calciner unit 8,672,005 15.8 2,583,760
SAP (DSI and calciner units) 29,093,156 52.9 19,361,810
Cost comparison between the SAP and DSI
The cost of a DSI system applied to Case 9 using commercial quicklime was analyzed
using the same assumptions and methodology for the DSI unit of the SAP discussed
above. The EACs for SAP and DSI systems at different capital cost and equipment life
span scenarios are presented in Table 15. In both the SAP and DSI cases, O&M costs are
5 to 20 times higher than the annualized project costs. Assuming the same sorbent
utilization and acid gas removal rates for either a SAP or a conventional DSI unit, the
Page 26
26
EACs of the SAP are 36% to 38% lower than the conventional DSI process in the three
scenarios considered in this study. The primarily reason for the lower cost of the SAP is
attributed to cost of producing a lime sorbent in SAP than purchasing it separately for use
with the conventional DSI ($105 vs. $30). In a commercial limestone calcination process,
quicklime is prepared in a rotary kiln operating at 1100 °C and 2-4 hours calcination time.
In contrast, a SAP sorbent is prepared in 1-3 second heat treatment time. These results
demonstrate that SAP is cost-effective technology for acid gas mitigation.
Table 15. Comparison of the costs using the SAP and DSI for the acid gas removal.
O&M, $/yr
Project cost
TPC, $ Interest
of capital
Lifetime of
equipment, yr
Annualized
project cost, $/yr
EAC, $/yr
SAP 19,361,810 29,093,156 5% 15 2,802,901 22,164,712
DSI 34,023,683 17,320,707 5% 15 1,668,717 35,692,399
SAP 19,361,810 29,093,156 10% 15 3,824,987 23,186,798
DSI 34,023,683 17,320,707 10% 15 2,277,219 36,300,902
SAP 19,361,810 29,093,156 10% 20 3,417,271 22,779,082
DSI 34,023,683 17,320,707 10% 20 2,034,484 36,058,167
A sensitivity analysis was also performed to determine the impacts of the sorbent
utilization and price of limestone on the O&M cost of the SAP (Figure 27). The total
O&M cost initially decreases with the increasing calcium utilization of the sorbent and
gradually levels off at about 50% calcium utilization, assuming the same acid gas
removal rates in Table 8. On the other hand, the total O&M increases linearly with the
cost of limestone. Since the transportation fee accounts for a large fraction of the cost of
the limestone, the results indicate that the cost of SAP sorbents is sensitive to the distance
between a quarry and a utility site.
Figure 27. Sensitivity of the total O&M cost to the sorbent utilization and limestone price.
15.0
20.0
25.0
30.0
0% 20% 40% 60% 80%
Tota
l O&
M, M
illio
n $
Sorbent utilization (molar, %)
10.0
15.0
20.0
25.0
30.0
0 20 40 60 80
Tota
l O&
M, M
illio
n $
Limestone Price ($)
Page 27
27
CONCLUSIONS AND RECOMMENDATIONS
Conclusions
The ISGS prepared lime sorbents showed better capacity for HCl capture, more
than twice of that by commercial Trona and quicklime but exhibited worse
performance for SO3 and SO2 removal.
Partially calcined limestone had the maximum capacity for HCl removal at 70%
extent of calcination (EOC).
The sorbents doped with calcium acetate hydrate [Ca(OAc)2·H2O] had more than
30% higher capacity for HCl capture.
Combination of various sorbents may optimize HCl/SO3/SO2 removal.
Batch sorbents provided useful guidance for producing desired sorbents by
sorbent activation process (SAP).
The SAP technology is cost-effective for on-site production of lime sorbents for
HCl and SO3 removal. SAP’s overall cost of acid gas removal is about 37% lower
than a conventional dry sorbent injection (DSI) process.
Recommendations for Future Work
Recommended future work incudes preparing 100 grams of two sorbents identified in this
project using the SAP and testing the sorbents with an entrained duct reactor with
combustion gases in which the sorbent particles have a residence time similar to that in
power plant applications. This will help in comparing to the results obtained using the
TGA and tubular reactor, and generate more realistic data to advance the SAP technology
towards slipstream testing and eventual commercialization of the technology for acid gas
capture applications.
REFERENCES
1. U.S.EPA, 40 CFR Parts 60 and 63, National Emission Standards for Hazardous
Air Pollutants. February 16, 2012.
2. Ruch, R., H. Gluskoter, and N. Shimp, Occurrence and Distrubution of
Potentially Volatile Trace Elements in Coal, in Enviromental Geology Notes.
August 1974.
3. Presto, A.A. and E.J. Granite, Impact of Sulfur Oxides on Mercury Capture by
Activated Carbon. Environmental Science & Technology, 2007. 41(18): p. 6579-
6584.
4. Partanen, J., et al., Absorption of HCl by Limestone in Hot Flue Gases. Part I: the
Effects of Temperature, Gas Atmosphere and Absorbent Quality. Fuel, 2005.
84(12–13): p. 1664-1673.
5. Chin, T., et al., Hydrated Lime Reaction with HCl under Simulated Flue Gas
Conditions. Industrial & Engineering Chemistry Research, 2005. 44(10): p. 3742-
3748.
6. Yan, R., et al., Kinetic Study of Hydrated Lime Reaction with HCl. Environmental
Science & Technology, 2003. 37(11): p. 2556-2562.
Page 28
28
7. Abanades, J.C. and D. Alvarez, Conversion Limits in the Reaction of CO2 with
Lime. Energy & Fuels, 2003. 17(2): p. 308-315.
8. Blamey, J., et al., Reactivation of CaO-Based Sorbents for CO2 Capture:
Mechanism for the Carbonation of Ca(OH)2. Industrial & Engineering Chemistry
Research, 2011. 50(17): p. 10329-10334.
9. Lu, Y., Evaluation of Dry Sorbent Technology for Pre-combustion CO2 Capture
(phase II). 2013, Illinois Clean Coal Institute.
10. Lu, H., A. Khan, and P.G. Smirniotis, Relationship between Structural Properties
and CO2 Capture Performance of CaO−Based Sorbents Obtained from Different
Organometallic Precursors. Industrial & Engineering Chemistry Research, 2008.
47(16): p. 6216-6220.
11. Cost and Performance Baseline for Fossil Energy Plants Volume 1. 2013,
DoE/NETL.
12. Documentation Supplement for EPA Base Case v.4.10_FTransport – Updates for
Final Transport Rule. June, 2011, U.S. EPA.
13. Captial and Operating Costs for Industrial Boilers. June, 1979, U.S. EPA.
Page 29
29
DISCLAIMER STATEMENT
This report was prepared by Hong Lu of the Illinois State Geological Survey at the
University of Illinois at Urbana-Champaign (ISGS-UIUC), with support, in part, by
grants made possible by the Illinois Department of Commerce and Economic
Opportunity through the Office of Coal Development and the Illinois Clean Coal
Institute. Neither Hong Lu, ISGS-UIUC, nor any of its subcontractors, nor the Illinois
Department of Commerce and Economic Opportunity, Office of Coal Development, the
Illinois Clean Coal Institute, nor any person acting on behalf of either:
(A) Makes any warranty of representation, express or implied, with respect to the
accuracy, completeness, or usefulness of the information contained in this report,
or that the use of any information, apparatus, method, or process disclosed in this
report may not infringe privately-owned rights; or
(B) Assumes any liabilities with respect to the use of, or for damages resulting from
the use of, any information, apparatus, method or process disclosed in this report.
Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute or imply its
endorsement, recommendation, or favoring; nor do the views and opinions of authors
expressed herein necessarily state or reflect those of the Illinois Department of
Commerce and Economic Opportunity, Office of Coal Development, or the Illinois Clean
Coal Institute.
Notice to Journalists and Publishers: If you borrow information from any part of this
report, you must include a statement about the state of Illinois’ support of the project.