Journal of the Air & Waste Management Association

ISSN: 1047-3289 (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/uawm18

Limestone Dissolution in Flue Gas Scrubbing:

Effect of Sulfite

Cynthia L. Gage & Gary T. Rochelle

To cite this article: Cynthia L. Gage & Gary T. Rochelle (1992) Limestone Dissolution in Flue
Gas Scrubbing: Effect of Sulfite, Journal of the Air & Waste Management Association, 42:7,
926-935, DOI: 10.1080/10473289.1992.10467043

To link to this article: https://doi.org/10.1080/10473289.1992.10467043

Published online: 06 Mar 2012.

Submit your article to this journal

Article views: 882

View related articles

Citing articles: 2 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=uawm20
ISSN 1047-3289 J. Air Waste Manage. Assoc. 42: 926-935

Limestone Dissolution in Flue Gas Scrubbing: Effect

ofSulfite

Cynthia L. Gage and Gary T. Rochelle

Department of Chemical Engineering
The University of Texas
Austin, Texas

Batch limestone dissolution experiments were carried out in a to predict measured dissolution rates at higher pH when
pti stat apparatus at 55°C with CO 2 sparging and dissolved CO2 is present. In these aqueous systems Toprac and
sulfite. Particle size distribution, utilization, sulfite in solution, Rochelle2 have shown that there is no effect of limestone
type. Other researchers have interpreted some experiments
limestone type, and the approach to calcite equilibrium were all
as having dissolution rates with surface kinetics contribu-
found to contribute to the limestone reactivity. In the absence of tions. The present work indicates that when sulfite is
sulfite, limestone dissolution was controlled solely by mass present in these aqueous systems, limestone dissolution
transfer. For a given stone under mass transfer control, film rates are inhibited. This effect has been successfully mod-
thickness was found to be independent of pH. The dissolution eled using surface kinetics combined with mass transfer
and a dependence on limestone type. Additional details and
rate in the presence of sulfite was controlled by a combined
applications of this work are given in Gage3 and Gage and
surface kinetics/mass transfer regime. SEM micrographs sup- Rochelle.4
ported this conclusion. A surface rate correlation was devel-
oped which accounted for observed inhibition by an inverse Previous Work
dependence on calcium sulfite concentration at the limestone
surface. While the form of the rate expression was applicable to
Mass Transfer Control
all stones, the surface rate constant was stone dependent. A
computer code which accounted for mass transfer with surface The dissolution rate of limestone particles in acid sys-
kinetics was tested against experimental observations of four tems has been previously treated as a mass transfer pro-
limestone types. Changes in pH and the concentrations of
cess.2'5"10 All of these systems were aqueous acid/CO2
systems which contained none of the other species present
calcium, carbonate, sulfite, sulfate, and adipic acid were accu- in FGD systems. The general conclusion was that for
rately modeled. systems with low partial pressures of CO2 and with pH less
than 5, the dissolution rate of limestone is diffusion con-
trolled. Additional researchers 1 ' 11 found that in systems of
Prediction of limestone dissolution is an important step in high partial pressures of CO2, the rate could be modeled as
determining scrubber performance of flue gas desulfuriza- a mass transfer process with the finite homogeneous reac-
tion (FGD) processes. Accurate predictions, in turn, require tion of aqueous CO2 to hydrogen ion and bicarbonate.
an understanding of the factors which affect the dissolution
process, especially those factors representative of scrubbing Surface Kinetics
systems. Most of the previous work on measurement of
limestone dissolution rates has been done in aqueous acid Only a few systems have been studied in which dissolu-
systems with varying partial pressures of CO2. The work tion was thought to be controlled by a combined regime of
has shown that at pH < 4 the dissolution rate is controlled mass transfer and surface kinetics. Most of these experi-
by mass transfer of H + to the surface. Chan and Rochelle1 ments 6 ' 8 ' 12 ' 13 were performed with a rotating disk appara-
used mass transfer with homogeneous chemical reactions tus using the results of Levich. Levich's development14
integrated the convective diffusion equations for a spinning
disk when a first order kinetic rate is applicable at the
surface. In this development the mass transfer contribution
Implications was easily separated from the overall rate.
Lund et al.6 studied limestone dissolution in HC1 solu-
Selection of a limestone source for slurry scrubbing tions at temperatures of - 15.6°C to 25°C using the rotating
should consider the specific reactivity of that source in disk apparatus. Only at very low temperatures did they find
the presence of dissolved sulfite. Among the 12 lime- dissolution to be controlled by combined mass transfer and
stones tested the reactivity varied an order of magni- surface kinetics. They calculated the surface concentra-
tude. Specific process options for slurry scrubbing, such tions and using a Langmuir model, postulated a rate
as forced oxidation, may have substantially better perfor-
mance because the limestone will be more reactive in the limiting step of the adsorption of H + onto the surface.
absence of dissolved sulfite. Computer models of lime- Rickard and Sjoberg13 used the rotating disk apparatus and
stone slurry scrubbing should incorporate the effect of found that limestone dissolution was controlled by com-
dissolved sulfite on the limestone dissolution rate. bined mass transfer and surface kinetics at pH 8.3 and

Copyright 1992—Air & Waste Management Association

926 J. Air Waste Manage. Assoc.

25°C. Additionally their surface rate constant was depen-
dent on the stone type.
Compton and Daly12 measured the dissolution rate of Notation
Iceland Spar at 25°C with varying partial pressures of CO2
using a rotating disk. They varied pH between 3 and 6.2 and B = agitation enhancement factor, cm" 1
found surface kinetics contributions in buffered systems at AC = concentration driving force, gmol/1
conditions close to equilibrium to calcite (pH 6.11 with CCS = relative saturation of calcium carbonate
0.091 or 0.167 atm CO2). Bjerle and Rochelle,15 using a pH CSS = relative saturation of calcium sulfite
d = diameter of a spherical particle, cm
stat apparatus, also noted surface kinetics limitations for dj = diameter of a particle from size fraction i at a
plane surfaces of limestones at conditions near equilibrium given time, cm
(pH 4.8 with CO2 sparging). doi = initial diameter of a spherical particle from size
fraction i, cm
Inhibitors D = diffusivity, cm2/sec
fi = volume fraction remaining from size fraction i
Several species are known to adsorb or precipitate onto F = total fraction remaining
calcite and may therefore inhibit dissolution rates. Calcite kc = surface kinetic rate constant, gmol5/^2-cm/liter5^2-
dissolution is inhibited by a number of dissolved metals, sec
including zinc, copper, iron, manganese, magnesium, and Kg, = solubility constant for calcium carbonate, gmol2/l2
scandium.16"22 Phosphate is also known to inhibit lime- mM = millimolar
Np = number of particles
stone dissolution.23"26 Several investigators27"29 report Npi = number of particles from size fraction i
that Al + 3 and F~ inhibit limestone dissolution at conditions t = time, sec
typical of flue gas desulfurization. Vtot = total initial volume of solids, cm3
In another FGD application, Chan and Rochelle1 quanti- Voi = initial volume of a particle from size fraction i,
fied inhibition by sulfite in concentrations greater than 1 cm3
mM at conditions far from equilibrium to calcite. Measured
rates for reagent calcite were significantly less than values Greek letters
predicted by a mass transfer model which accounted for the 8 = film thickness, cm
buffering effect of sulfite/bisulfite. At some conditions rates (J>i = volume fraction of particles between size i and i +
measured in sulfite were less than the zero sulfite rates. 1
They successfully modeled the inhibition as a shift in p = molar density of calcite, gmol/cm3
surface equilibrium from the calcite solubility product to a
mixed calcium sulfite/carbonate solid solution. Subscripts
Jarvis et al. 29 studied FGD solution effects on ten dif- b =bulk
ferent limestone types. Solution species which were tested eq = equilibrium
included pH, Ca + + , Mg + + , Na + , SOj, SO=, COJ, F e + + + , 0 = initial
Al + + + , F~, and adipic acid. Of the species studied, magne- p = pseudo
sium and adipic acid were found to be minor inhibitors s = surface
while sulfite and aluminum fluoride complexes were noted
to be significant inhibitors to the limestone dissolution rate.
The magnitude of sulfite inhibition was found to be depen-
dent on limestone type. Another important conclusion of
their work was that the CO2 hydrolysis reaction was Unfortunately this equation cannot be used directly since
insignificant in FGD systems where other buffer species the total area at any time depends on the rate of dissolution,
such as sulfite are present. the total elapsed time, and the number of particles for each
The present work addresses the effect of sulfite on size. Thus a method must be developed to determine the
limestone dissolution. Reagent limestone is used at condi- area from the measured dissolution.
tions near equilibrium to calcite. Dissolution rates are Polydispersed stones have been previously modeled by
modeled as mass transfer with the calcite surface concentra- Chan and Rochelle1 and by Toprac and Rochelle.2 In both
tions controlled by a surface kinetics rate involving sulfite studies the total rate of dissolution was expressed using the
at the surface. The effect of sulfite on different limestones is initial particle size distribution and the slope of the total
also studied by applying the model to the work of Jarvis et fraction remaining versus time curve. Using a similar
al. 29 and Chan and Rochelle.1 approach here, Equation 1 can be integrated for a single
particle of initial size dOi to give:
Model

This predictive model for limestone dissolution uses film t = doi (3)
theory to describe the mass transfer. For spherical particles
the dissolution flux is given by Equation 1. The parameter tat is found by iteration using Equation 4
p dV -DAC until the desired utilization (or total fraction remaining) is
T^lllY (1) reached.
TTd2NP dt
Since the area at any time is well defined for monodis- U =1- F =1 - (4)
persed limestones, Equation 1 can be used directly to
calculate either 8 or AC from experimental data when the The number of particles for any initial volume, Voi, can
other variable and the rate are known. The film thickness, also be written in terms of the particle size distribution by
8, is a function of the particle size. Equation 5.
For polydispersed stones the equivalent flux using film
theory can be calculated using an "average" film thickness,
8 ave as shown in Equation 2. NPi = (5)
PVtot dF -DAC
Flux = (2) Once (at is found for a given utilization using Equation 4,
i dt Equations 5 and 3 can be used to find the area available

July 1992 Volume 42, No. 7 927

from each size fraction. These areas are then added to give concentrations. This model is subject to several flux con-
the total area needed by Equation 2. straints. Since sulfite and sulfate are not consumed or
Since prediction of limestone dissolution rates requires produced at the limestone surface, the net flux of each of
knowledge of the surface concentrations, a chemical mass these species must be zero, as shown in Equations 12 and
transfer model has been written to predict DAC. This model 13.
is a mass transfer extension to the Bechtel Modified Radian
Equilibrium (BMREQ) program.30 BMREQ is a program DHSo3([HSO3-]s - [HSO3-]b)
which solves the set of equilibrium and mass balance
expressions applicable to an FGD system. The thermody- + DSO3([SO3=]S - [SO3=]b) = 0 (12)
namic database of this model comes from the work of DHSO4([HSO4-]S - [HSO4-]b)
Lowell et al. 31 who amassed (or measured) the temperature-
dependent parameters for the equilibrium constants of 41 + DSO4([SO4=]S - [SO4=]b) = 0 (13)
chemical reactions. Lowell and his coworkers also assem-
bled the temperature dependent parameters needed for the In these equations and all future expressions, the
extended Debye-Huckel expression which predicts activity pseudoconcentration subscript has been dropped in order to
coefficients. The BMREQ database was used as written simplify notation. All future concentrations will imply
with the exception of the temperature dependent parame- pseudoconcentrations unless otherwise noted.
ters for CaHCO3+, CaCO3°, and calcite solubility. These Since FGD limestone dissolves in a high ionic environ-
parameters were updated to the values given by Plummer ment, the potential gradient at the surface is very small.
and Busenberg32 as given in Equations 6 through 8. The mass transfer model neglects this gradient and re-
Additionally, the model has been modified to assure charge quires the charge flux to be zero between the surface and
balance in the bulk solution. bulk (Equation 14).
^ = 1209.120 + 0.31294T 2DCa([Ca++]s - [Ca++]b) + DHco3([HCO3-]s - [HCO3-]b)

34765.05 + 2DCo3([CO3=]s - [CO3-]b) + DHSo3([HSO3-]s

-478.782 log T (6)
- [HSO3-]b) + 2DSO3([SO3=]S - [SO3=]b)
= -1228.732 - 0.299444T + DHSo4([HSO4-]s - [HSO4-]b) + 2DSo4([SO4=]s
35512.75 - [SO4=]b)+ DH([H+]S - [H+]b) + DOH([OH-]S
+ 485.818 log T (7)
- [OH"]b) - 0 (14)
logK,. = -171.9065 - 0.077993T ++ +
If additional cations, such as Mg or Na , or if there are
2839.319 additional anions, such as adipate, these are also included
+ 71.595 log T (8) in the charge flux expression. Note that with a zero charge
flux, electronegativity at the surface is not explicitly satis-
BMREQ is used to calculate activity coefficients, concen- fied.
trations, and equilibrium constants of the 41 FGD species. The last flux constraint which must be satisfied is that
Then the mass transfer extension is called to calculate
limestone surface species and reactivity. The model as- the total calcium flux must equal the total carbonate flux
sumes equilibrium in the bulk and at the surface of the (Equation 15).
limestone dependent on the applicable boundary condition. DCa([Ca++]s - [Ca++]b) = DHco3([HCO3-]s - [HCO3-]b)
Diffusivities used in the model were those values given by
Tseng33 and Chan and Rochelle1 and are corrected for + DCo3([CO3=]s - [CO3=]b) (15)
temperature. The number of variables is reduced by de-
fining pseudoconcentrations, pseudodiffusivities, and pseu- The boundary condition which is satisfied at the lime-
doequilibrium constants. stone surface depends on the species present in the bulk
Pseudoconcentrations are calculated from the equilib- solution. In the absence of sulfite species, the boundary
rium concentrations found by the rigorous model. A condition is given by Equation 16, the solubility product of
pseudoconcentration is defined as the sum of the ion pairs calcium carbonate.
with their parent species. For example, the pseudoconcen- [Ca ++ ] S [C(V] S = Ksp(calcite) (16)
tration of bicarbonate is given by Equation 9.
It will be shown in the results section that in the presence
[HCO3-]p = [HCO3-] + [CaHCO3+] of sulfite, limestone dissolves under a combined regime
+ [MgHCO3+] + [NaHCOgo] (9) where resistances from both mass transfer and surface
kinetics affect the overall rate. The form of the surface
Pseudiffusivities and pseudoequilibrium constants are kinetics rate expression is written in terms of activities and
found from the pseudoconcentrations. The bicarbonate is given by Equation 17.
pseudodiffusivity is given by Equation 10.
5
(CaCO°3eq - CaCO 3s )
Flux = (17)
D, (10) CaSO 3s - CaCO 3s
Therefore, when sulfite is present, the boundary condi-
Here the summation over i is for the same species as tion is the continuity of flux at the surface. In this case the
given in Equation 9. The pseudoequilibrium constant for calcium flux from the surface must equal the flux deter-
bicarbonate is determined from the bulk pseudoconcentra- mined by the surface kinetic rate of Equation 17. In the
tions. This constant is given by Equation 11.
limit of zero sulfite, the surface rate equation gives a
[H+]p[CO3=]F boundary condition of CaCO3°s '-> CaCO3°eq.
K,•HCOo- - (11) In the model the bisection technique is used to iterate on
[HCO3-]p
the surface concentrations and fluxes by assuming hydro-
Once the bulk pseudoconditions are calculated, the mass gen ion concentration at the surface, calculating equilib-
transfer model uses film theory to converge on the surface rium species, and testing against the boundary condition.

928 J. Air Waste Manage. Assoc.

Convergence is reached when the appropriate boundary pH Stat Apparatus
condition and flux expressions are satisfied.
Titrant
Experimental Apparatus and Procedure Demand Module
To
Limestone dissolution rates were measured in a batch Recorder
mode using a pH stat apparatus consisting of a Fisher
Model 360 electrometer and Model 395 Burette Dispenser. Electrometer
The basic stoichiometry used in this apparatus is given in
Equation 18.
CaCO3 + 2HC1 -» Ca ++ + H2O + CO2 + 2C1" (18) CO,
HCI
Experiments were performed in 500 ml of solution pH Probe Tilrant
contained in a jacketed 700 ml reactor. The solution
temperature was controlled at 55°C ± 0.5 (Figure 1).
Presaturated CO2 was sparged through the reactor solution
during the experiments at a rate of 315 cc/min. All experi-
ments were performed in a background solution of 0.1 M i cm
i
CaCl2 to provide constant ionic strength and to reduce
changes in solution concentration as the limestone dis-
solved. The acid titrant was 1.00 N HC1. Most experiments
used reagent grade limestone whose particle size distribu-
tion was measured with Model TAII Coulter Counter cm|4-
->|5cm|
Recirculating
(Table I). Experimental pH varied from 4.8 to 5.3.
L
=*«-Water Bath
Corning rugged bulb combination electrodes were used to
measure pH. These electrodes were reconditioned every two Figure 1. Experimental apparatus.
weeks by replacing the reference electrode solution and
treating the electrode for one hour in warm dilute acid
solution. This procedure reduced electrode errors by pre- Experimental Methods of Previous Work
venting clogging of the reference junction arid by replacing
adsorbed Na + with H + on the pH electrode tip. Electrodes The rate model was also applied to work done by Chan
were calibrated at 56°C before each experiment using a pH and Rochelle1 and by Jarvis et al. 29 Using work done by
5.06 buffer (0.01 N adipic acid-sodium hydroxide in 0.1 M other researchers allows the model to be tested against a
CaCl2). wide range of conditions representative of scrubbers and
Sulfate is a typical species present in scrubbing systems hold tanks. Table I lists the characteristics of the four
and is a known inhibitor for calcium sulfite crystallization. limestones analyzed by the rate model.
Therefore sodium sulfate was added to the solution to give The work performed by Chan and Rochelle1 was done
sulfate concentrations ranging from 0 to 12 mM. Calcium using reagent CaCO3 in a batch pH stat apparatus with N2
sulfite saturations (CSS) ranged from 0 to 2.5 by addition of sparging. The bulk of the work was done at 25°C, although
0.5 M sodium sulfite solution. Sulfite concentrations were one series was performed at 55°C and pH 5. Rates for these
monitored during the experiments by collecting and fil- experiments were measured at 50 percent dissolution.
tering approximately 2 mis of solution. This was added to a These reagent experiments combine well with the present
preweighed aliquot of 0.01 N iodine and reweighed to work by addressing the effect of sulfite in a very low
determine actual sample size. The sulfite concentration was saturation of CaCO3 and by providing extensive data at
then calculated by back titration with 0.01 N sodium 25°C.
thiosulfate. Calcium sulfite saturation was calculated using The experimental work done by Jarvis et al. 29 involved
the solubility product measured by Tseng.33 This value is the dissolution of industrial limestones screened to 325-
2.474 x 10- 7 M 2 at55°C. 400 mesh. The pH-stat apparatus used was a 2 liter sealed
At the start of the experiment sodium sulfate was added reactor operating in a batch mode for the solids and
to 500 ml of CaCl2 solution and then the mixture was
warmed to temperature. CO2 was sparged through the
solution for approximately thirty minutes to presaturate Table I. Limestone characteristics.
the solution. Once equilibrium was obtained sodium sulfite
was added to give the desired concentration. This concentra- Reagent Reagent
tion was checked by iodometric titration to confirm the 1 2
starting value. An experimental run was initiated by adding Fisher Fisher HSTC Monteagle
Source (1988) (1981) Beachville, On Tiftonia, Tn
1 g of limestone. During the experiment pH was then held
constant by automatic addition of titrant. During a run Mg(Wt.%) 0.01 0.2 0.6
electrode millivolt signals and the titrant acid addition rate BET Area
were continuously monitored on a strip chart recorder. The (nvVg)* 0.22 0.27 2.23
cumulative dissolution of CaCO3 was found directly from PSD (p,m)**
the titrant rate using the stoichiometry of Equation 19. For % < 6.35 0.0 1.6 0.0 0.0
experiments lasting longer than 30 minutes, the sulfite % < 12.7 0.9 69.6 0.0 0.2
% < 25.4 13.5 99.8 1.1 2.0
concentration was checked after one half hour and then % < 50.8 95.3 100 95.1 89.1
after each additional hour. When there was a drop in sulfite % < 80.6 99.0 100 99.3 99.6
concentration because of oxidation to sulfate, additional Mean
sulfite was added as Na2SO3/NaHSO3 at the pH of the diameter 32.4 10.3 41.5 41.5
experiment. As a result sulfite concentrations averaged
* Measured with krypton to improve accuracy.
within 8 percent of the starting value throughout the ** Particle size distribution (PSD) shows cumulative volume per-
duration of the experiment. At the end of a run, the sulfite cent less than the specified diameter. Mean diameter based on the
concentration was checked again and the solution was volume percent PSD for the two reagent stones. Industrial stones
filtered to collect the solids for SEM or BET analyses. were prescreened to give the listed sizes.

July 1992 Volume 42, No. 7 929

Table II. Rate constants and film thicknesses to 50°C for the Experimental Results
experimental limestones."
Results from all of the limestones are summarized in
8(xlO4) BET area Mg
Stone 13 A
cm m 2 /g wt% Table II. Detailed results from the four major limestones
kc(xlO )
are shown in Tables III-VI. The tables include the bulk
Reagentt 24.07 3.4 0.227 0.01 solution concentrations, the measured dissolution rates,
Reagent** 2.44 7.3 — — the surface activities which satisfy the measured rates, and
HSTC 5.93 9.3 0.267 0.20 the rates predicted by the combined mass transfer/surface
Monteagle 14.06 7.1 2.226 0.60 kinetics model. The measured fluxes were calculated from
Maysville 17.50 9.0 1.174 1.7 the measured absolute rates with spherical areas based on
Arapahoe 5.53 7.5 0.305 0.12 the particle size distributions. The surface activities were
Iceland Spar 2.18 7.3 1.052 0.04
Edwards 4.63 9.7 0.837 0.10 calculated from the mass transfer model by iterating on the
Limestone B 12.71 8.1 1.062 0.10 surface concentrations to satisfy the measured fluxes. The
Limestone C 5.74 8.1 1.188 0.0 predicted fluxes were calculated from the mass transfer
Limestone D 8.93 9.3 1.502 3.1 model using the surface kinetics rate expression as the
Limestone A 4.70 8.1 1.946 0.19 boundary condition. In analyzing the results of the batch
* Film thickness from measuredfluxesat pH 5.5, CCS = 0.026, and limestone dissolution experiments the variables of interest
CSS = 0.0. Rate constants from measured fluxes at pH 5.5, CCS = relative to FGD systems are pH, CSS (calcium sulfite
0.026, and CSS = 1.6. saturation), and CCS (calcium carbonate saturation). For
A
Rate constant has units of M5/2-cm/sec. this work presented in Table III, pH varied from 4.8 to 5.3,
t For reagent limestone T = 55°C, pH = 5.3, CCS = 0.26, and CSS varied from 0 to 2.8, and CCS varied from 0.02 to 0.26.
CSS = 2.0. Within these ranges, the measured flux for reagent CaCO3
** For this reagent T = 25°C, pH = 5.0, CCS = 0.0006, and CSS = varied from 9.32 x 1 0 ~ 1 0 t o l . 5 3 x 10~ 8 gmole per cm 2 -sec.
2.3. Saturation to gypsum (calcium sulfate) was 0.9 and the
total area at 50 percent dissolution was 2820 cm 2 .
For the industrial stones 2 9 presented in Tables IV and V,
flow-through for the liquid. The bulk of their work was pH varied from 4.5 to 6.1, CSS varied from 0 to 1.9, and CSS
done at 50°C on two stones by varying CaCC>3 saturation at varied from 0.01 to 0.25. The measured fluxes ranged from
two fixed CaSO 3 saturations. For most experiments carbon- 2.9 x l O " 1 0 to 1.38 x 10~ 8 gmole per cm 2 -sec for HSTC
ate was fixed at 2 mM. Dissolution rates of eight other limestone and from 5.1 x 10~ 1 0 to 2.22 x 10~ 8 gmole per
stones were measured at pH 5.5 in the absence and
presence of sulfite. In this work, initial rates were reported.
These experiments test the applicability of the combined
Table IV. Results for HSTC limestone.
mass transfer/surface kinetics regime on industrial lime-
stones. Flux
Surface solution (gmol/cm2/sec)"
Bulk solution (so.) (M)t Pro
i I tJ~
Table III. Results for reagent limestone experiments at 55°C, [SO3=] Other aCaSO3° aCaCO3° jviea- dictedA
sured9 ) (xlO9)
0.01M Na2SO4 and CO2 sparging. pH (xlO 3 ) (xlO 3 ) (xlO3) (xlO7) (xlO

Bulk solution Surface solution Flux 4.5 0.0 Base* 70.7 7.60 7.58
(M) (M)t (gmol/cm2-sec)* 4.5 30 Base 1.63 0.61 13.77 14.32
5.0 6.0 Base 70.7 2.87 2.81
[SO3=] aCaSO3° aCaCO3° Measured PredictedA 5.0 8.5 Base 1.25 1.22 6.66 7.81
pH (xlO3) (xlO3) (xlO7) (xlO 9 ) (xlO9) 5.0 17 Base 2.04 1.04 7.26 7.23
5.5 0.0 Base 70.7 1.09 1.06
4.8 0.0 74.2 12.42 11.62 5.5 0.75 Base 0.36 6.87 2.61 3.52
4.8 0.6 0.27 12.3 14.43 15.40 5.5 3 Base 0.98 4.39 3.38 3.42
4.8 1.0 0.36 9.93 15.27 16.01 5.5 6 Base 1.63 3.23 2.55 2.75
4.8 4.9 0.93 4.72 15.22 15.33 5.5 6 Base 1.66 3.40 2.90 2.75
4.8 9.7 1.50 3.70 13.64 12.81 5.8 0.0 Base — 70.7 0.34 0.59
4.8 14.7 1.99 3.05 10.13 10.89 5.8 2 Base 0.89 10.3 1.64 1.58
4.8 14.7 2.05 3.25 12.34 10.89 5.8 3.75 Base 1.49 8.28 1.04 1.14
4.8 20.0 2.57 2.88 9.97 9.40 6.1 0.0 Base — 70.7 0.27 0.31
5.0 0.0 74.2 5.59 6.23 6.1 1.25 Base 0.69 24.1 0.47 0.69
5.0 5.0 1.04 7.47 5.82 6.47 6.1 2.5 Base 1.33 22.5 0.41 0.42

5.1 0.0 •
74.2 - 4.77 5.00 5.0 17 8CO3= 1.83 1.66 5.30 5.14
5.1 2.0 0.57 13.6 5.99 6.32 5.5 6 8CO3= 1.53 0.98 1.63 1.03
5.5 6 100Mg ++ 0.76 3.32 3.62 3.72
5.2 0.0 74.2 3.74 4.13 5.5 3 0SO4= 1.09 5.42 4.02 3.13
5.2 1.0 0.43 28.5 7.35 6.64 5.5 6 0SO4= 1.74 3.67 2.70 2.46
5.2 3.0 0.92 18.6 5.12 4.69 5.5 3 6 Adipic 0.74 3.27 4.07 5.19
5.2 3.0 0.93 18.7 5.27 4.69 5.5 6 6 Adipic 1.47 3.23 4.25 3.57
5.2 4.8 1.33 16.1 2.58 2.84 5.7 1.5 ** 0.94 28.5 0.56 0.46
5.7 2 1.24 27.7 0.50 0.40
5.3 0.0 74.2 2.80 2.96 5.8 15 5Ca + + 1.36 1.76 8.44 7.30
5.3 1.0 0.41 34.0 4.23 3.98 5.8 30 5Ca + + 1.90 1.15 7.36 7.16
5.3 2.3 0.78 26.8 2.60 2.65 6.1 2.5 100Mg+ + 0.88 18.7 0.31 0.73
5.3 5.6 1.76 24.2 1.30 1.33
5.3 6.1 1.90 23.9 0.93 1.25 t Activities calculated using the measured fluxes and 8 = 9.3 (jum.
w
5.3 6.3 1.97 24.0 1.09 1.21 Measured fluxes are initial rates with 510 cm2 total area.
A
Fluxes predicted using kc = 5.801 x 10~13 M 5/2 cm/s and 8 = 9.3
t Activities calculated using the measured fluxes and 8 = 3.4 |xm. |xm.
* Measured fluxes are at 50% dissolution with 2820 cm2 total area. * Base conditions: 50°C, 0.04 M Ca ++ , 0.015 M Mg ++ , 0.027 M
A
Fluxes predicted using k,, = 2.317 x 10" 12 M 5/2 cm/s and 8 = 3.4 Na+, 0.015 M SO4=, 0.002 m CO3=, and 0.0962 M Cl~.
** 144 mM Ca ++ and 3 mM CO3=.

930 J. Air Waste Manage. Assoc.

Table V. Results for monteagle limestone.
TTlnv Experiments at 25°C and N2 sparging.
F1UX
Surface solution (gmol/cm2-sec)w Bulk solution Surface solution 1Flux
BUIK solution uvu (M)t Plea- Pre- (M) (M)t (gmol/cm*-secr
[SO3=] Other aCaSO3° aCaCO3° sured dicted'^ [SO3=] aCaSO 3 ° aCaCO3° Measured PredictedA
pH (xlO 3 ) (xlO3) 3
(xlO ) (xlO ) (xlO ) (xlO 9 )
7 9
PH (xlO 3 ) (xlO3) (xlO 7 ) (xlO 9 ) (xlO 9 )
4.5 0.0 Base* 70.7 9.91 9.92 4.5 0.0 55.5 4.87 4.87
4.5 30 Base 1.84 0.84 22.21 22.80 4.5 0.9 0.43 1.43 7.02 7.18
5.0 0.0 Base — 70.7 3.95 3.68 4.5 3.0 0.75 0.70 7.55 7.96
5.0 8.5 Base 1.49 1.96 12.92 12.84 4.5 5.8 1.22 0.69 9.08 8.14
5.0 17 Base 2.34 1.60 14.05 12.16 4.5 8.7 1.45 0.53 8.83 8.08
5.5 0.0 Base 70.7 1.31 1.39 4.5 27.0 2.90 0.34 9.07 7.33
5.5 0.75 Base 0.44 10.0 4.8Q 5.62 4.5 38.0 3.41 0.22 7.28 6.97
5.5 3 Base 1.21 7.07 8.25 6.09
5.5 3 Base 1.03 4.88 5.22 6.09 5.0 0.0 — 55.5 1.36 1.64
5.5 6 Base 1.67 3.47 4.00 5.27 5.0 0.9 0.50 0.96 3.39 4.48
5.5 6 Base 1.63 3.23 3.32 5.27 5.0 3.2 1.19 0.74 4.41 4.82
5.5 6 Base 1.67 3.50 3.97 5.27 5.0 6.3 1.89 0.56 4.41 4.52
5.8 0.0 Base 70.7 8.75 7.66 5.0 13.2 3.19 0.31 3.13 3.86
5.8 2 Base 1.02 14.0 4.39 3.09 5.0 20.0 4.34 0.16 1.56 3.43
5.8 3.75 Base 1.58 9.79 2.93 2.42
0.0 Base 70.7 0.48 0.39 5.5 0.0 55.5 0.74 0.68
6.1
1.25 Base 0.75 29.5 1.69 1.42 5.5 1.0 0.78 2.03 2.60 2.86
6.6
2.5 Base 1.36 23.9 1.05 0.96 5.5 1.8 1.35 2.81 3.97 2.85
6.1 2.6 2.21
5.5 1.73 3.89 2.70
5.0 17 8CO3= 1.96 1.90 8.58 9.28 5.5 5.8 2.85 0.33 0.85 2.19
5.5 6 100Mg ++ 1.28 3.74 6.25 6.99
5.5 3 0 SO4= 1.17 6.42 6.59 5.65 5.75 0.0 55.5 0.54 0.49
5.8 2 6 Adipic 0.79 9.48 4.21 4.50 5.75 1.0 0.82 2.02 1.69 2.14
1.45 4.40 5.75 1.7 1.30 1.90 1.93 2.07
5.8 3.75 6 Adipic 9.16 3.04 3.5 2.32 0.97 1.74
5.8 15 5Ca ++
1.57 2.56 18.20 13.57 5.75 1.23
5.8 30 5Ca + + 1.96 1.32 12.99 13.64 6.0 0.0 55.5 0.36 0.35
6.1 2.5 100Mg+ + 0.90 20.2 1.09 1.59 6.0 1.0 0.85 2.02 0.90 1.51
t Activities calculated using measured fluxes and 8 = 7.1 ixm. 6.0 1.7 1.10 1.06 0.46 1.47
"A Measured fluxes are initial rates with 51012cm25/2
total area. 6.0 3.1 2.32 0.83 0.34 1.19
Fluxes predicted using kc = 1.408 >c 10" M cm/s and 8 = 7.1 5.0* 0.0 74.2 3.94 3.94
ixm. 5.0 0.7 0.33 2.26 5.58 7.17
* Base conditions: 50°C, 0.04
i M Ca + + , 0.015 M M g + + , 0.027 M 9£.0 ft n QQ O OP*
71.01VI
+ O.XJ
Na , 0.015 M SO4=, 0.002 M CO3=, and 0.0962 M Cl". 1.58 1.80 8.88
5.0 5.8 6.62
5.0 11.1 2.39 0.87 5.58 5.60
t Activities calculated using measured fluxes and 8 = 7.3 ixm.
cm2-sec for Monteagle limestone. Gypsum saturation was Measured fluxes are at 50% dissolved with 8300 cm total area.
A
1.1 and the initial total area was 510 cm2. Fluxes predicted using kg = 2.436 x 10~ 13 M 5/2 cm/s and 8 = 7.3
fim.
For the work of Chan and Rochelle1 presented in Table * This series performed at 55°C. Activities at 55°C calculated using
VI, pH varied from 4.5 to 6.0, CSS varied from zero to 7.2, 8 = 6.3 |xm. Fluxes at 55°C predicted using kg = 4.336 x 10~12M5/2
and CCS was always less than 0.01. Within these ranges, cm/s and 8 = 6.3 |xm.
the measured flux for reagent CaCO3 varied from 3.4 x
10~ 10 to 9.08 x 1O~9 gmole per cm2 sec. Gypsum saturation
Discussion
was zero and total area at 50 percent was 8300 cm2.
For reagent limestones which are polydispersed (this SEM micrographs were taken of many of the partially
work and Chan and Rochellel), the measured fluxes were dissolved limestones. Figures 2-4 show some of the typical
taken at the time required for the cumulative dissolution to features observed. All micrographs include a reference
reach 50 percent. The flux was calculated using Equations length of 10 microns. Figure 2 is a view of the undissolved
2—5. For the industrial stones which were screened, the reagent calcite. The crystals are fused blocky crystallites
measured fluxes were calculated from the measured initial typical of precipitated calcite. Figure 3 is a micrograph of
rates using Equation 1 and assuming a monodispersed
stone of 41.5 microns in spherical diameter.
The measured fluxes from the zero sulfite cases were
used with the mass transfer model to calculate apparent
film thicknesses. In the absence of sulfite, limestone dissolu-
tion is controlled by mass transfer, thus the boundary
condition in the mass transfer model is given by Equation
16. This condition along with the bulk solution composi-
tions specifies the applicable driving forces. When these
calculated forces are used with the measured rates, Equa-
tion 1 or 2 can be solved to give an estimate of the
hypothetical film thickness around the particles. Once the
film thickness was estimated, it was used with the sulfite
experimental rates to calculate the surface activities shown
in the Tables. The form of the kinetic rate expression was
found by plotting surface concentrations against experimen-
tal rates. The surface rate constants and film thicknesses
for all limestones are shown in Table II. Figure 2. Reagent 1 calcite-raw sample.

July 1992 Volume 42, No. 7 931

 developed by Toprac and Rochelle.2 Following their theory,
the film thickness was derived and is given by Equation 19.
d
8 =
2 + 1.612Bd (19)
In this equation B is a mass transfer enhancement factor
accounting for the effects of agitation and the contribution
of the CO2 hydrolysis reaction to limestone dissolution. In
the absence of CO2, Toprac found B to have a value of 400
cm" 1 . Under CO2 sparging, B varied from 260 to 880 cm" 1
as pH increased from 4 to 5. B increases because the
hydrogen ion concentration decreases, the contribution of
the CO2 hydrolysis becomes more important. Using B =
400 cm" 1 on the limestones of this study, the calculated 8's
ranged from 7 to 9 jxms. If B = 880 cm" 1 ,8 varied from 4 to
Figure 3. Dissolved reagent 1—Mass transfer regime. Exp. 6 |xms. These values are in good agreement with the
67-pH 5.3,0.1 M CaCl2,0.85 atm CO2,55°G, 55% dissolved. measured film thicknesses.
In this study 8 varied between the four limestones. Stone
to stone variations could occur for several reasons. For
reagent limestones, differences in particle size distribution
the reagent limestone at 55 percent dissolution in the will cause 8 ave to vary. Also, as discussed previously, CO2
absence of sulfite. In this environment dissolution is con- sparging will be expected to cause a change in 8 since the
trolled by mass transfer^ and the expected effect on the model does not explicitly account for the CO2 contribution
crystal shape is to round off the corners and edges. This to the rate. The work of Chan was performed in the absence
result is clearly seen in the micrograph. Figure 4 is a view of of GO2, and therefore the measured 8 would be larger than
reagent limestone at 55 percent dissolution in the presence the present work which used CO2 sparging.
of 6.1 mM SOg1 at pH 5.3. From this micrograph it is clear
that a different dissolution phenomena is occurring when For the industrial stones which have the same diameter,
sulfite is present. Spikes and nodules were typical features the differences occur because of differences in the surface
seen in all the sulfite experiments at all pH values. The roughness. These stones have been screened to give the
industrial limestones also produced these unusual surface same spherical size, yet BET measurements show that the
features. It appears that some of the surfaces have been Monteagle stone has ten times the surface area of the
blocked from dissolving while opposing faces are eaten away HSTC stone (Table II).
towards the center. This observation of fast and slow Sulfite Experiments
dissolving sites supports the conclusion that in the presence
of sulfite, limestone dissolution is controlled by surface Figure 5 shows measured and predicted rates for the
reaction kinetics. three pH values studied in this work. The data show that at
low concentrations of sulfite and for a given pH, the
dissolution rate is enhanced over the rate observed in the
Zero Sulfite Experiments absence of sulfite. The experiments also indicated that at
higher concentrations of sulfite, the rates become inhibited
If mass transfer controls the dissolution rate the esti- and in some cases are even lower than the zero sulfite rates.
mated film thickness should not depend on pH. This was These phenomena concur with the observations made by
found to be true, since for a given limestone, a single value Chan and Rochelle.1 The ideal rate for pH 4.8 shown on
of 8 predicted the measured rates to within 15 percent at all Figure 5 is the rate calculated by assuming that mass
pH values. The single exception was the HSTC case at pH transfer alone controls the dissolution. In this situation
5.8, an experiment which Jarvis 29 identified as an outlier. sulfite acts only as a buffer, and the limestone surface is
The measured values of 8 ranged from 3.4 to 9.7 n-m. saturated to calcite. It is clear from Figure 5 that mass
Using stagnant film theory, predicted values of 8 for the transfer alone can not predict the observed rates because it
four limestones range from 14 to 21 |xm. These values were does not account for the inhibiting effect of sulfite. The
from two to five times the measured values. Film thickness predicted rates shown for all three values of pH are those
can also be predicted using the mass transfer approach values calculated by the combined rate model including
mass transfer and surface kinetics with the rate constants
and film thicknesses listed in Table II. The combined model
does a good job of predicting the observed rates at all values
of pH and sulfite concentrations. It accounts for both the
buffering and inhibiting effects of sulfite.
Figure 6 shows a comparison of the measured fluxes and
the model predictions for two industrial stones. In general
the complete model predicts the measured rates within 25
percent over two orders of magnitude in the rate. This
shows that the model can account for changes in pH and
concentrations of sulfite, sulfate, carbonate, calcium, and
magnesium. However, the model does not do well for high
magnesium at high pH. Magnesium is a known inhibitor of
calcite dissolution.21'34 In the present form, the surface
kinetics model can not account for this inhibition. Addition-
ally some variation is expected between the model and the
measured rates because the sulfite concentrations used in
the model were only the nominal values because the actual
Figure 4. Dissolved reagent 1—surface kinetics regime. measured concentrations were not reported.
Exp. 77—pH 5.3, 0.1 M CaCI2, 0.85 atm CO 2 , 55°C, 6.1 mM The combined model was tested against the rates mea-
SO 3 ,55% dissolved. sured in sulfite solutions by Chan and Rochelle.1 Table VI

932 J. Air Waste Manage. Assoc.

 Bulk Relative Saturation of CaS03
Figure 5. Effect of sulfite on fluxes of reagent 1 limestone in 0.1 M CaCI2 and 10 mM sulfate
at 55°C curves predicted by combined rate model.

gives the comparison between the measured and predicted saturations. Crystallization onto limestone would not be
rates. This table shows poor agreement for the pH 6 expected in scrubber systems where sulfite seed crystals
experiments where the presence of small amounts of carbon- exist to provide the surface area needed for crystallization.
ate could significantly affect the dissolution rate. The In the present work it was not possible to exceed a calcium
concentration of this species was not monitored during the sulfite saturation of 3 without significant crystallization of
experiments, and in modeling these experiments the CO2 the calcium sulfite from the clear solution.
partial pressure was set to 0.001 atm. If sparging were not
adequate, this lack of carbonate concentration accuracy Surface Kinetics Rate Expression
would be magnified at high pH resulting in overprediction
of the rate. The combined model calculates rates by converging on
Inadequacies in the model can be found by plotting the the surface concentrations which satisfy both the mass
ratio of the predicted flux to the measured flux versus the transfer expression and the surface kinetics rate for a given
independent variables. A valid model is indicated by ran- film thickness and kinetics constant, kc. The continuity of
dom distribution of the data about one. Any trend in the flux at the surface is the boundary condition with the
data would indicate a poor model. Figure 7 shows a surface kinetics rate as shown in Equation 17. This expres-
graphical comparison for the predicted to measured flux sion shows that sulfite inhibits the dissolution by forming
ratio against the calcium sulfite saturation for both reagent CaSO3 which may be adsorbing onto the surface.
limestones. The Chan and Rochelle data sets at pH 6 were The inclusion of this rate expression into the mass
not included in this figure since there were recognized transfer model means that the limestone dissolution rate
problems in modeling these data. For CSS < 3 a majority of can be controlled by mass transfer, surface kinetics, or a
the cases are predicted to within 25 percent of the measured mixture thereof. Equation 17 shows that surface kinetics
rate. However, for the reagent limestone of Chan and resistance will be significant whenever sulfite is present in
Rochelle there is poor agreement at the high sulfite, high solution. When the sulfite concentration is zero, the kinetic
pH data where the saturation of CaSO^ is greater than 3. It rate of Equation 17 becomes infinitely fast, and the surface
could be that in these experiments calcium sulfite was concentrations approach saturation to calcite. When this
crystallizing on the limestone and blinding it completely happens, limestone dissolution is controlled by the finite
rather than causing the milder adsorption seen at lower mass transfer rate. Mass transfer will also control the

10"

o HSTC

• Monteagle

£5 10-
2 O

10"
1 0" 1 0"

Measured Flux (gmol/cm2-sec)

Figure 6. Parity plot for HSTC and monteagle limestones, predicted values from the
combined rate model.

July 1992 Volume 42, No. 7 933

 Reagent 1

Chan & Rochelle

2-

0 1 2 3 4 5 6 7 8

Bulk Saturation of CaSO3

Figure 7. Test of combined model as a function of CaSO3 saturation for reagent limestones.

dissolution rate when the surface concentration of the imately the same slope. Therefore, the scandium work
calcium carbonate ion pair approaches its equilibrium could be expressed in a rate form similar to the one written
value. In this case the surface rate of Equation 17 becomes for the sulfite work, and the analysis that scandium ad-
zero. For the extreme case of a zero mass transfer rate, the sorbed onto the limestone could be extended to sulfite
limestone surface species equal their bulk values, and the adsorption onto limestone.
dissolution is controlled entirely by the surface kinetic rate. The Nestaas and present models can also be compared to
The surface kinetics rate applicable in sulfite solutions is the previous reagent model of Chan and Rochelle.1 In their
different from the kinetics form found by other researchers work the rate of dissolution was also given by a driving force
using simple CC^/water systems. A direct comparison between pseudoequilibrium and the bulk conditions. How-
between expressions is difficult because of the presence of ever, because their work was done with nitrogen sparging,
sulfite as an inhibitor. Most of the previous work which the relative saturation to calcium carbonate was approxi-
found a surface kinetics contribution had been done in mately zero at all pH values. Thus when compared to
CO2 /water systems6-8'13 or in systems where a non- Nestaas model, their work would become a single line if the
inhibiting buffer Was present. 12 In all of these cases, the rate were plotted against the surface concentration. How-
Levich development was used for the experimental analy- ever, since they never exceeded pseudoequilibrium in the
ses. This development has a built in assumption of a first bulk solution, they did not completely test the theoretical
order surface kinetics rate expression. The sulfite rate applicability of a pseudoequilibrium constant. The present
expression was regressed from the measured data with no work which studied different bulk relative saturations to
assumptions as to its potential form, The resulting equa- calcium carbonate indicates that the value of the pseudoequi-
tion shows a half-order dependence rather than first order. librium constant would be dependent on the bulk satura-
A common result, though, between the previous and present tion of calcium carbonate and sulfite. Thus a significant
work is that the rate constant in the surface reaction is advantage to the present model over these pseudoequilib-
stone dependent. rium models is that Equation 17 will work with all bulk
The dissolution rate form in sulfite also differs from the saturations of calcium carbonate, whereas the pseudoequi-
forms of other inhibited systems which contained metal librium models require a different pseudoconstant which
ions 20 ' 21 or scandium ions.22 The rate form empirically must be measured experimentally for each different bulk
determined by Nestaas and Terjesen22 is shown in Equa- saturation.
tion 20. As shown in Table III the surface rate constants varied
over an order of magnitude between the 12 limestones. This
d[Ca(HCO3)2] constant, however, could not be correlated to Mg content or
= &'([Ca(HCO3)2]eq
~dt to BET surface area. The activation energy of the constant
was determined for two limestones. Using the constants for
- [Ca(HCO3)2]) - fc"[ (20)
Reagent limestone at 55°C in this work and 25°C from the
This rate form leads to an equivalent development of a work of Chan, E a was calculated to be 14.7 kcal/gmol. Over
driving force determined by a pseudoequilibrium constant the same temperature range, E a for HSTC was 11.9 kcal/
(less than thermodynamic equilibrium) and the bulk concen- gmol. Interpretation of these values is complicated by the
trations. The pseudoequilibrium conclusion had been de- fact that the elementary steps of the adsorption of sulfite
rived from the earlier work. 20 ' 21 and reaction with calcite are unknown, and how the
Although there are significant differences between the temperature dependencies of these elementary steps might
Nestaas rate form and the present rate expression, sulfite affect the surface rate constant is not understood at this
has effects similar to those of scandium. A plot of the time.
dissolution rate against the surface concentration of scan-
dium yielded a family of lines representing different bulk
saturations of calcium bicarbonate, but each line producing Conclusions
approximately the same slope. If the reagent sulfite data is
plotted in a similar manner with the rate against the In the absence of sulfite, limestone dissolution was
surface concentration of CaSO3, a family of lines results. controlled solely by mass transfer. For a given stone under
Each line represents a different pH (or equivalently, a mass transfer control, film thickness was found to be
different bulk saturation of calcium carbonate) with approx- independent of pH.

934 J. Air Waste Manage. Assoc.

 The presence of sulfite in solution was found to have a 13. Rickard, D. T.; Sjoberg, E. L. "Mixed kinetic control of calcite
dual effect on the dissolution of limestone. The sulfite/ dissolution rates," Am. J. Sci. 183:815 (1983).
bisulfite pair acted as a buffer at the limestone interface by 14. Levich, V. G. Physicochemical Hydrodynamics, Prentice-Hall
Inc., Englewood Cliffs, 1962.
providing hydrogen ion consumed during the dissolution 15. Bjerle, I.; Rochelle, G. T. "Limestone dissolution from a plane
reaction. But sulfite also inhibited dissolution by adsorbing surface," Chem. Eng. Sci. 39:183 (1984).
on the limestone as calcium sulfite. Because dissolution was 16. Kitano, Y.; Kanamori, N.; Yoshioka, S. "Adsorption of zinc and
enhanced, the rate in the presence of sulfite was controlled copper ions on calcite and aragonite and its influence on the
transformation of aragonite to calcite," Geochim. J. 10: 175
by a combined surface kinetics/mass transfer regime. The (1976).
sulfite inhibition effect was accurately modeled in the 17. Clarke, E. T.; Loeppert, R. H.; Ehrman, J. M. "Crystallization
surface kinetics expression by an inverse dependence on of iron oxides on calcite surfaces in static systems," Clays and
calcium sulfite concentration at the limestone surface. In ClayMin. 33:152 (1985).
18. Barton, P.; Vatanatham, T. "Kinetics of limestone neutraliza-
modeling the dissolution rates, continuity of flux at the tion of acid waters," Env. Sci. Tech. 10:262 (1976).
surface of the limestone was satisfied using the surface 19. Rochelle, G. T.; Chan, P. K.; Toprac, A. J. "Limestone
kinetics equation. Dissolution in Flue Gas Desulfurization Processes," U.S. EPA,
The surface rate expression did a good job of predicting publication PB83-252833,1983.
20. Erga, O.; Terjesen, S. G. "Kinetics of the heterogenous
the dissolution of limestone for a variety of stones and over reaction of calcium bicarbonate formation, with special refer-
a range of solutions compositions. Changes in the pH and ence to copper ion inhibition," Ada Chem. Scand. 10: 872
the concentrations of calcium, carbonate, sulfite, sulfate, (1956).
and adipic acid were accurately predicted. While the form of 21. Terjesen, S. F.; Erga, Cv, Thorsen, G.; Ve, A. "II. Phase
boundary processes as the rate determining steps in reactions
the rate expression was applicable to all stones, the rate between solids and liquids," Chem. Eng. Sci. 14: 277 (1961).
constant was stone dependent. This constant could not be 22. Nestaas, I.; Terjesen, S. G. "The inhibiting effect of scandium
correlated to Mg content or to BET surface area, however ions upon the dissolution of calcium carbonate," Ada Chem.
Ea was estimated to be on the order of 10 kcal/gmol. Scan. 23: 2519 (1969).
23. Freeman, J. S.; Rowell, D. L. "The adsorption and precipita-
When compared to other empirical rate models deter- tion of phosphate onto calcite," J. Soil Sci. 32: 75 (1981).
mined in inhibited systems, the present model was found to 24. House, W. A.; Donaldson, L. "Adsorption and coprecipitation
be superior since it did not require the experimental of phosphate on calcite," J. Coll. Int. Sci. 112:309 (1986).
determination of solution dependent pseudoequilibrium 25. Giannimaras, E. F.; Koutsoukos, P. G. "The crystallization of
calcite in the presence of orthophosphate," J. Coll. Int. Sci.
constants. 116:423 (1987).
26. Morse, J. W. "Dissolution kinetics of calcium carbonate in sea
Acknowledgments water. V. Effects of natural inhibitors and the position of the
chemical lysocline," Am. J. Sci. 274:638 (1974).
27. Mori, T.; Matsuda, S.; Nakajima, F.; Nishimura, T.; Arikawa,
The authors wish to thank Radian Corporation and, in Y. "Effect of Al 3+ and F~ on desulfurization reaction in the
particular, Jim Jarvis, for their cooperation in providing limestone slurry scrubbing process," Ind. Eng. Chem. Process
experimental data. The authors also thank the Electric Dev. 20:144 (1981).
Power Research Institute for its partial funding through 28. Farmer, R. W.; Jarvis, J. B.; Moser, R. "Effects of Aluminum;
Fluoride Chemistry on Wet Limestone Flue Gas Desul-
the Radian Corporation. furization," paper presented at Spring National AIChE Meet-
ing, Houston, 1987.
References 29. Jarvis, J. B.; Meserole, F. B.; Selm, T. J.; Rochelle, G. T.; Gage,
C. L.; Moser, R. E. "Development of a Predictive Model for
Limestone Dissolution in Wet FGD Systems," paper presented
1. Chan, P. K.; Rochelle, G. T. "Limestone dissolution—effects of at EPA/EPRI Combined FGD and Dry SO2 Control Sympo-
pH, CO2, and buffers modeled by mass transfer," ACS Symp. sium, St. Louis, 1988.
Ser. 188: 75 (1982). 30. Epstein, M. "EPA Alkali Scrubbing Test Facility: Summary of
2. Toprac, A. J.; Rochelle, G. T. "Limestone dissolution in stack Testing Through October 1974," U.S. EPA 600/7-7-105
gas desulfurization," Env. Prog. 1: 52 (1982). (PB274544), 1977.
3. Gage, C. L. Limestone Dissolution in Modeling of Slurry 31. Lowell, P. S.; Ottmers, D. M.; Schwitzgebel, K; Strange, T. I.;
Scrubbing for Flue Gas Desulfurization, Ph.D. Dissertation, Deberry, D. W. "A Theoretical Description of the Limestone
December 1989. Injection-Wet Scrubbing Process," U.S. EPA, APID 1287, PB
4. Gage, C. L.; Rochelle, G. T. "Modelling SO2 Removal in Slurry 1931-029,1970.
Scrubbing as a Function of Limestone Type and Grind," in 32. Plummer, L. N.; Busenberg, E. "The solubilities of calcite,
Proceedings: 1990 SO2 Control Symposium, EPRI Report aragonite, and vaterite in CO2-H2O Solutions between 0 and
GS-6963, Vol. 3, p. P-143 to P-164, September 1990. 90°C, and an evaluation of the aqueous model for the System
5. Berner, R. A.; Morse, J. W. "Dissolution of calcium carbonate CaCO3-CO2-H2O," Geochim. Cosmochim. Ada, 46: 1011
in sea water-IV. Theory of calcite dissolution," Am. J. Sci. (1982).
274:108 (1974). 33. Tseng, P. C; Rochelle, G. T. "Dissolution rate of calcium
6. Lund, K.; Fogler, H. S.; McCune, C. C; Ault, J. W. "Acidization- sulfite hemihydrate in flue gas desulfurization processes,"
II. The dissolution of calcite in hydrochloric acid," Chem. Eng. Env. Prog. 5: 34 (1986).
Sci. 30: 825 (1975). 34. Laslo, D.; Chang, J.C.S. "Pilot Plant Tests on the Effects of
7. King, C. V.; Liu, C. L. "The rate of solution of marble in dilute Dissolved Salts on Lime/Limestone FGD Chemistry," paper
acid," Am. Chem. Soc. J. 55:1928 (1933). presented at EPA/EPRI Flue Gas Desulfurization Sympo-
8. Sjoberg, E. L.; Rickard, D. T. "Temperature dependence of sium, New Orleans, 1983.
calcite dissolution kinetics between 1 and 62°C at pH 2.7 to 8.4
in aqueous solutions," Geochim. Cosmochim. Ada, 48: 485
(1984).
9. Sjoberg, E. L.; Rickard, D. T. "The effect of added dissolved
calcium on calcite dissolution kinetics in aqueous solutions at At the time of writing, C. L. Gage and G. T. Rochelle were
25°C," °C," Chem. Geo. 49:405 (1985). with the Department of Chemical Engineering, the Univer-
10. Uchida, S.; Wen, C. Y.; Michael, W. J. "Dissolution rate of
limestone into acid solution," J. Chin. Inst. Chem. Eng. 5: 111 sity of Texas at Austin, Austin, TX 78712. C. L. Gage is
(1974). presently with the U. S. EPA, Research Triangle Park, NC.
11. Wallin, M.; Bjerle, I. "A mass transfer model for limestone This manuscript was submitted for peer review on August
dissolution from a rotating cylinder," Chem. Eng. Sci. 44: 61 7,1991. The revised manuscript was received on March 23,
(1989). 1992.
12. Compton, R. G.; Daly, P. J. "The dissolution kinetics of Iceland
spar single crystals," J. Coll. Int. Sci. 101:159 (1984).

July 1992 Volume 42, No. 7 935