SPE-200699-MS

Improved Scale Prediction for High Calcium Containing Produced Brine and
Sulfide Scales

Xin Wang, Guannan Deng, Saebom Ko, Alex Yi-Tsung Lu, Yue Zhao, Chong Dai, Samridhdi Paudyal, Bingjie
Ouyang, Sana Mateen, Amy T. Kan, and Mason B. Tomson, Rice University

Copyright 2020, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE International Oilfield Scale Conference and Exhibition originally scheduled to be held in Aberdeen, Scotland, UK,
24 - 25 Jun 2020. Due to COVID-19 the physical event was changed to a virtual event. The official proceedings were published online on 24 June 2020.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
In oil and gas industry, scaling prevention is one of the most important problems. While with more
aggressive drilling and exploitation, scale control for the unconventional scale under complex water
chemistry becomes more challenging. There are more chances to encountering with high temperature,
high pressure, high TDS and some unconventional scale conditions. The modeling of the sulfide scale
is notoriously difficult due to the extremely low solubility and complex water chemistry. Thus, the
thermodynamic data is rare for sulfide minerals. Metal-sulfide-bisulfide complexes bring a large uncertainty
for scale prediction. Another challenge is scale prediction in brine with high TDS, especially with high
calcium concentration. Thermodynamic data with common ions Ca2+ and SO42- is needed to improve
thermodynamic models. The objective of this paper is to extend our knowledge for these exotic scale
solubility predictions with both experimental studies and model validation. Some remaining questions in
Pitzer theory framework have been thoroughly reviewed and discussed to improve the scale prediction
for iron sulfide and high calcium condition. The newly derived models are able to predict the saturation
index (SI) within ±0.3 unit for iron sulfide and ±0.15 units for common sulfate scales, respectively. These
developed models have been incorporated into ScaleSoftPitzer for practical use in the oil and gas production.

Introduction
The risk of scale formation in produced water have increased concern for oil and gas industry costing
millions of dollars#8217; every year (Kan, et al., 2019). Scale formation normally occurs due to large
temperature or pressure changes, and the mixing of incompatible brines (Dai, et al., 2017). Since the major
inorganic components of most produced water are sodium chloride and calcium chloride, the scale prediction
in oil field commonly can be express as thermodynamic models to predict the solubility of conventional
scale minerals (e.g. halite, barite, calcite, gypsum, celestite, etc.) in aqueous systems containing: Na+, K+,
Mg2+, Ba2+, Sr2+, Cl−, HCO3−, CO32-, SO42-, H2O and CO2. Modeling of scale formation under such systems
had been discussed for many years with several different approaches (Helgeson 1974, Pitzer 1984, Moller
1989, Archer 1992, He 1992, Holmes 1997, Christov 2004, Li 2005, Li 2011, Dai 2015, Appelo 2015,
2 SPE-200699-MS

Dai 2017). Among these models, Pitzer theory is suggested to be one of the most popular and effective
models due to the capacity of dealing with the concentrated mixed electrolyte solutions over a wide range
of temperature and pressure. The Pitzer theory is based on parameterizing specific ion-ion interaction with
temperature, pressure and composition. Dai et al. (Dai 2017) have reported a comprehensive Pitzer model
to predict most common scale mineral and solution density for the produced water system. Based on this
well-developed Pitzer model, in this study we extend the model to cover ion sulfides and some extreme
brine composition such as high calcium concentration.
One key question remaining is the uncertainty of the solubility product constant (Ksp). Since the solubility
prediction is based on the saturation index (SI) defined by
accurate Ksp value serves as the basis to calculate SI. Unfortunately, the literature listed Ksp values of
sulfate mineral shows a large discrepancy, Barite, gypsum and anhydrite Ksp values could differ by as
much as 0.1, and 0.2, respectively, or greater (Blount 1977, Langmuir and Melchior
1985, Kharaka et al. 1989, Johnson et al. 1992 Appelo 2015). A similar question remains for the iron
sulfide mackinawite (FeSm) and due to the uncertainty of the secondary dissociation constant of hydrogen
sulfide, the solubility product constant is usually expressed as.
The solubility of mackinawite at room temperature has been investigated by many researchers, the reported
Ksp values vary from 10−2.94to 10−3.98, and only a few have addressed the temperature dependence (Naumov
1974, Luther 1996, Davison 1999, Benning 2000, Rikard 2006). It is risky to trust these literature values
without careful reexamination.
The second question is the complicated speciation of the active ions, especially for the iron sulfide
solubility prediction. The ferrous ion is very active and could undergo very complicated speciation with
sulfide ion under brine conditions, especially at pH 5-7. The major aqueous ferrous ion is not the free Fe2+,
but the iron-sulfide complex (Rickard 2007). Davison et al., 1999, found that the solubility of mackinawite
was partially explained by FeHS+, Lurther et al. 1996 suggested that Fe(HS)02,aq, or some polymeric such
as FexSx, could be found by scanning voltammetry. Rickard 2006 and our previous research (Wang, 2019)
suggested that a neutral species plays as a key component in determining the concentration of total
dissolved Fe(II) concentration. Thus, when developing an accurate model, an additional correction for these
stable species needs to be used. Also, for the sulfate minerals, the protonation of SO42- to HSO4−cannot be
ignored at high temperature conditions.
And the third question is the model deviation at high calcium and high sulfate conditions. This is attributed
to the uncertainty of the 2:2 ion interaction term, like and Due to their low
concentration in solution these interaction terms are too small to be determined under common experimental
conditions. In the produced water with extremely high calcium or sulfate concentrations, these uncertainties
result in large errors in solubility predictions.
The purpose of this study is to elucidate these problems and try to solve them with new experimental
data and more complete thermodynamic database. A reliable iron sulfide solubility prediction model, based
on Pitzer ion-interaction theory, has been developed with ion-complexation from pH 3 to 9 under a wide
range of brine composition with general prediction error is less than ± 0.3 SI. Newly derived Ca-SO4
interaction terms and mixture terms with a boundary of T<250°C, P<1400bar, Ca<1.3 m SO4<1.5 m are
determined by modeling with literature and laboratory-measured solubility data of gypsum, anhydrite and
barite. Regression gives a general prediction error of less than ±0.10 SI, with over 900 solubility data. These
newly derived improvements are included in ScaleSoftPitzer 2020.
SPE-200699-MS 3

Experimental Design
Iron sulfide solubility measurement
For iron sulfide solubility study, the detailed apparatus design and experiment procedure can be found in
our previous study (Wang, 2019). Experiments were performed in a strictly anoxic plug-flow reactor as
shown in Figure 1. The iron (II) and sulfide (-II) solutions were prepared separately, and a third bottle was
used for dilution and pH control. All solutions were sparged with argon gas for a least 2 hours to control the
oxygen level to below 10 ppb. After deoxygenation, the solutions were taken by gas-tight syringe pumps,
heated in heating loop by water bath and mixed right before the reactor. The reactor was a glass column
with two pieces of carbon steel C1018 coupons. For each experiment, the mixed solution was pumped
through the reactor for 20 hours to allow the iron sulfide scale fully coating on the coupon surface. The
ferrous ion concentration was checked once per hour until the system reaches equilibrium and the solubility
measurements were then determined. A 0.45 µm cellulose acetate filter was directly linked at the outlet of
the reactor to remove suspended FeS particles. The ferrous ion and sulfide concentrations were measured
by colorimetric methods, and effluent pH was measured in a closed anoxic vial within the water bath,
to keep the same condition as within the reactor. After each experiment, Ar gas was flowed through the
reactor for at least 10 hours to dry the reactor and coupons. FeS on coupons surface was analyzed by X-
ray diffraction (XRD) and/or scanning electron microscopy energy dispersive spectroscopy (SEM/EDS) for
crystal morphology. Typical experimental condition is listed in Table 1.

Figure 1—Anoxic plug-flow reactor for iron sulfide solubility study

Table 1—Typical reaction conditions.

Parameters Conditions

Temperature 25°C/ 70°C/ 90°C

pH 6.75/ 6.35/ 6.00 ± 0.10
Reaction time 8 − 168 hr
Fe(II) concentration 0.50 − 1.8 × 10−4 M (3 − 10 ppm Fe2+)
S(-II) concentration 0.15 − 3.0 × 10−3 M (5 − 98 ppm S2-)
NaCl 0.15 − 4.30 M
PIPES / MES buffer 0.010 M
CaCl2·2H2O 0.025 M (1000 ppm Ca2+)
4 SPE-200699-MS

Anhydrite solubility measurement

The experimental procedure follows a traditional hydrothermal reactor approach. A specially designed 500
ml autoclave was used to contain a 260 ml borosilicate glass reactor, where about 2 grams of anhydrite
powder stayed inside the reactor to equilibrate with the about 200 g background solution with high
concentration of calcium. After more than 2 days of shut-in, samples were withdrawn and filtered to analyze
both the dissolved Ca2+ and SO42- concentrations. A small amount of water vapor existed in the headspace
of the 150 ml reactor. The mass (<0.4 g at T<220°C) was corrected for in the solubility calculation. When
using ICP-OES to measure low concentrations of SO42- (< 5 mg/L) in high TDS NaCl-CaCl2 (up to 1 m
ionic strength) brine, it was important to optimize the ICP-OES method to gain better signal/concentration
slope and it was important to matrix match between measured samples and standards. More details of these
experimental procedures can be viewed in Deng, et al., 2018.
Some preliminary experiments were used to measure anhydrite solubility in both pure water and in 4 M
NaCl solutions; these results agreed with literature results very well, verifying both solubility equilibrium
and analytical procedure. Anhydrite solubility from 120°C to 210°C was then measured in CaCl2 solution
with calcium concentration ranging from 0.1 m to 1.3 m, as well as NaCl-CaCl2 solution with similar [Ca]
range, but at constant ionic strength of 4 m.

Results and Discussion

Iron sulfide speciation and solubility modeling
According to our previous research, the freshly precipitated iron sulfide was always identified to be
mackinawite under experimental conditions. No phase transformation was observed during the experiment
period (Wang, 2019). This result is in agreement with many previous researchers (Liu, 2017; Benning,
2000; Rickard, 2006). Thus, in this paper, the iron sulfide discussed below is referring to mackinawite. The
solubility data of mackinawite from literature were also reviewed and selected. The data sources and their
experimental conditions are summarized in Table 2.

Table 2—Summary of data sources in iron sulfide solubility modeling

Reference T range(°C) pH IS (mol/ kg H2O) No. of data

This study (Wang 2019) 23 − 90 5.86 − 6.93 0.22-4.27 30

Liu, et al. 2017 23 − 125 4.22 − 5.09 0.01-5.03 50
Rickard 2006 23 3.16 − 9.66 0.10-0.51 78

As discussed in the introduction, the complicated speciation of the aqueous Fe-S system presents
considerable difficulty for the solubility modeling. Our previous study suggested that a neutral species
may have an important contribution to the concentration of total aqueous ferrous ion, Fe(II). The aqueous
complex FeS is an uncharged neutral species with undetermined total stoichiometry that might range from
FeS and Fe2S2 to Fe150S150 with Fe to S ratio of 1 (Theberge 1997), with size smaller than 2 nm and could
remain stable in water. The stability constant and the concentration to represent these species, when at
equilibrium with mackinawite solid, are shown in equation (1) and (2).

(1)

(2)
SPE-200699-MS 5

Notice here the concentration of this neutral species would be a constant at a given temperature,
pressure and ionic strength condition, but independent from the total [Fe(II)] and total [S(II)] concentrations
and pH. This is conventionally referred to as the intrinsic solubility of the mineral (Dyrssen 1988, Stuart
2005, Rickard 2007). Also, since is a constant, the percentage of this neutral species in total
aqueous Fe(II) would be more important at higher pH and higher sulfide conditions. To address the solubility
prediction of mackinawite, equation (3) has been applied for speciation and the molality of ferrous ion and
bisulfide was calculated by speciation equation (4) and (5). The activity coefficients were calculated using
Pitzer theory. Ka1,H2S is the dissociation constants of hydrogen sulfide, the temperature dependence of these
parameters were taken from Suleimenov and Seward, 1997.
(3)

(4)

(5)

The solubility product of mackinawite (Ksp,FeSm) from different literature is plotted in Figure 2.
The mackinawite Ksp value could differ by as much as even at room temperature
conditions. The difference in experimentally measured solubility constants appears to be small Berner 1967
Morse 1987 Benning 2000
Rickard 2006 (pH dependent log10Ksp=−3.5). While the PhreeQC reports Ksp= −4.65 (Parkhurst 2013).
Only Naumov 1971 and Benning 2000 reported the temperature dependence of mackinawite, while in
Naumov 1971 the Ksp of mackinawite was determined by characterization of thermodynamic properties
of Fe2+, HS− and FeS solid, the parameter estimation involved several steps of regression, such
that the error propagation might be large. Usually the determination of Ksp should be mostly independent
of the activity coefficient model such as Pitzer model, because the calculation of activity coefficient itself
carries error. Since the and is from the experimental measurement, which would be
used to calculate the Ksp value in equation the speciation of
ferrous ion would influence the Ksp calculation. Thus, Ksp of mackinawite is also fitted in this study to
improve the prediction model.
6 SPE-200699-MS

Figure 2—Discrepancy of Kspof mackinawite from different sources.

The saturation index of iron sulfide (SIFeS), as shown in equation (6), is a function of solubility product
(Ksp), activity coefficients (γ) of each compound and pH (aH+). Theoretically, if the solutions have reached
equilibrium with the solids, then the SI should be 0. The sum of SI2 values for each solubility measurement
are minimized to fit thetemperature dependent parameters of stability constant and Pitzer coefficients.
The particle swamp algorithm inMATLAB was used to optimize the objective function. The temperature
dependence of KspFesm and were fitted as shown in equation (7) and equation (8), respectively.
Our calculated mackinawite Ksp is also plotted in Figure 2, compared with other experimental values, the
difference is within 0.15 logKsp under room temperature conditions. This deviation might be due to the
speciation of Fe(II). And the temperature dependence of the Kspis closer to Naumov 1977, with largest error
of -0.35 logKsp under high temperature conditions.

(6)

(7)

(8)

Using the fitting result of stability constant, the calculated SI values for mackinawite solubility
under various conditions are plotted in Figure 3. The mean SI value for the 158 data points is -0.0003
with a standard error of estimate of the true SI equal to ±0.015 SI units, showing good
prediction from 23 to 125°C, 0 to 5 M ionic strength and 3 (or lower) to 9 pH, S(-II) to Fe(II) molar ratio 0.01
to 10,000. SI values represent the predicted scaling and corrosion risks of iron sulfide under each condition.
These results imply good applicability of this model in iron sulfide scale prediction and control with a wide
range of brine composition in different industrial processes.
SPE-200699-MS 7

Figure 3—Calculated SI values of mackinawite, with T= 23 − 125°C, 0 − 5 M NaCl, pH = 3.16 − 9.66.

Furthermore, more than 78% of the SI values are within the range of [-0.2, 0.2], which indicated the
relative errors of solubility predictions are within 25.0%. Only 6 data points have SI value out of the [-0.4,
0.4] range, such deviation might be due to random measurement errors or the limitation of measurement
and slight oxidation or poor solid separation could easily lead to error estimation of Fe(II). Also, due to
the complexity of Fe(II)-S(-II) chemistry, other iron sulfide complexes like FeHS+ and FeSn have also been
reported. These iron sulfide complexes could also result in more deviation. More research is needed to
validate the stability constant of different species and the solubility data over wider temperature range.
Compared with other commercially available software like PhreeQC and our previous software SSP
2019, as shown in Figure 4, this model shows a substantial improvement in the iron sulfide solubility
prediction, with a wide range of brine and physiochemical conditions, which could cover most field
conditions. This model has been included in our ScaleSoftPitzer 2020.
8 SPE-200699-MS

Figure 4—Comparison of the developed model in SSP2020 with SSP 2019 and PhreeQC

Solubility modeling for sulfate scale under high calcium condition

Compared with iron sulfide, sulfate scales (e.g, barite, gypsum, anhydrate, celestite et al.) have been
discussed repetitively for many years in the literature using the Pitzer equation framework. It is natural to
think that the sulfate mineral thermodynamics has become well studied over the years, but as mentioned in
the introduction, many other problems also exist and are neglected in literature models.
One key question is the uncertainty of solubility product constant (Ksp) of sulfate minerals. As shown
in Figure 5, the literature listed Ksp values of sulfate mineral show a large discrepancy, even from widely
trusted models, such as Solmineq.88, SUPCRT.92, PHREEQC. A significant source of uncertainty of Ksp
values is from the uncertainty of the chemical potentials of M2+ and ions at infinite dilution, especially
at high temperature. Two approaches are generally applied to calculate Ksp values,
a) Gibbs free energy approach,

b) Solubility approach,

Using barite as an example, in the Gibbs free energy approach, and are determined by

characterization of thermodynamic molar properties of BaCl2, Na2So4 and

NaCl solutions. This approach is generally taken in SUPCRT.92 regression process, using mainly data of
standard volumetric properties (standard partial molar volume/compressibility/expansibility), along with the
Gibbs-Duhem equation and various Maxwell relations, to estimate other standard thermodynamic properties
of ionic species (Helgeson 1976). Comparing to the free energy approach, the solubility approach is more
directly linked to experimental data. This is the method applied by C.W Blount (Blount 1977), using his
measured solubility data at high temperature and pressure. Blount#8217;s results agree with Ksp from
SUPCRT.92, Solmineq.88 below 200°C (Figure 5, up), however, PHREEQC#8217;s result deviates from
other sources significantly: at 25°C, it is widely agreed that (Langmuir 1985, Jiang
1992), while PHREEQC gives (Appelo 2015).
SPE-200699-MS 9

Figure 5—The discrepancy of Ksp of barite, gypsum and anhydrite at vapor pressure from different
sources. "Solmineq.88" from Kharaka et al. 1989, "SUPERCRT.92" from the software SUPERCRT.92,
originated from Johnson, Oelkers and Helgeson. 1992. "PHREEQC 3.0" from Appelo 2015.

When using the solubility approach to calculate Ksp, the speciation of SO42- to HSO4− is often assumed
negligible, but in fact should not be neglected at high temperature. The equilibrium can be expressed as,
(9)
10 SPE-200699-MS

(10)

The association constant in equation (9) was studied at high temperature by two independent
researchers (Oscarson, Izatt, et al. 1988; Dickson, Wesolowski, et al. 1990), their results of association
constant agreed with each other very well. The degree of association in BaSO4- H2O system,
, was calculated using solubility data of barite in pure water from Blount 1977,
with the activity coefficient calculated by the Pitzer model. The result is plotted in Figure 6. In a BaSO4-
H2O system, is almost negligible below 150°C, but increase to 5% at 200°C and further to about 20%
at 250°C, speciation of therefore should be considered above 175°C in the sulfate chemistry. The
speciation to would decrease with ionic strength. This means may become negligible even
at T=200°C in typical brine conditions with high ionic strength. However, when calculating Ksp of barite
based on dilute BaSO4-H2O system, should be considered.

Figure 6—Calculated degree of association in the Barite-H2O system with temperature

at various pressure. The total sulfate concentration is taken from C.W. Blount 1977.

With the speciation of HSO4−is considered, it is possible to calculate barite Ksp based on solubility data.
The data sources are critically selected and listed in Table 4, these data are selected because of their low
ionic strength, and therefore the calculation of activity coefficient has high certainty. Two extra data sources
are included compared to data used by Blount 1977. The Jiang 1996 data gives an accurate Ksp prediction
from 0°C to 80°C, updating data from Templeton 1960. The Shi et al. 2012 data pertains to barite in 0.1m
NaCl, it is the only measurement at 0°C and high pressure. Also, at the highest pressure of 1400bar, Blount’s
experimental solubility data only covers temperature around 155°C. So at this pressure we use smoothed
data given by Blount. There is some uncertainty in this data due to extrapolation, but this is the only data
source and it would give a reasonable prediction boundary to the Ksp model. The calculated barite Ksp is
shown in equation (11), the overall goodness of fit for these solubility data is evaluated by the root-mean-
square error (RMSE) of saturation index (SI), which is listed also in Table 4. A robust regression technique
was applied, where a few data with large deviation was given less weight.
SPE-200699-MS 11

Table 3—Data to calculate Barite Ksp from 0°C to 250°C, <1400 bar.

Data Source T/P range and composition No. RMSE of SI

Templeton 1960 25°C to 90°C, 1bar, Barite in H2O 6 0.05

Jiang 1996 0°C to 80°C, 1bar, Barite in 0.0002 60 0.02
m to 0.0079 m Na2SO4 solution
Blount 1977, experimental 22°C to 249°C, 38 0.05
<1400bar, Barite in H2O
Shi et al. 2012 0°C to 100°C, <1517 bar, 12 0.06
Barite in 0.1m NaCl solution
Blount 1977, Smoothed 60°C to 250°C, 9 0.05
1400bar, Barite in H2O

(11)

Here P<1400bar and 273K<T<523K.

The fitted Ksp values of barite is compared with values given by SUPCRT.92, results plotted in Figure 7.
At vapor pressure and 550 bar, the difference is within 0.05 logKsp, at 1000 bar the maximum difference is
0.15 logKsp. The difference between ours and those reported by Blount 1977 at temperature below 200°C
is also very small (<0.05 logKsp); however our results at T>225°C is considerably lower, e.g., at 250°C,
vapor pressure, our calculated barite logKsp is -10.95 compared to Blount’s -10.80, due to the correction of
speciation at high temperature.

Figure 7—Regression results of barite Ksp, and its comparison with SUPCRT.92

Unlike barite, the Ksp of gypsum and anhydrite cannot be determined unambiguously with solely
binary solubility data, because of their relatively large solubility. The mean activity coefficient needs to be
determined first. For gypsum we take directly the Ksp value from Solmineq.88, since the limiting properties
of both and might be more accurately characterized at low temperature (<100°C). The above
limiting properties reported in literature might involve larger uncertainties T>100°C, which leads to a
relatively large discrepancy of anhydrite Ksp value (Figure 5, bottom). We examine the Ksp of anhydrite
using isocoulombic principle (Langmuir 1985), assuming the isocoulombic reaction in Equation 12 has a
constant heat capacity change of reaction , then use barite Ksp determined previously to indirectly
12 SPE-200699-MS

calculate anhydrite Ksp (equation 13). Our calculated results are surprisingly similar to Ksp anhydrite given
by Solmineq.88, though showed small deviation above 200°C at vapor pressure (Figure 8), which renders
us to speculate that the values given in SOLMINEQ.88 also follows the isocoulombic approach.
(12)

(13)

Figure 8—Ksp anhydrite with temperature and pressure, compare values

calculated by isocoulombic principle in this study and those given in Solmineq.88.

Parameterization of Ca:SO4 interaction in Pitzer model. Characterization of the 2:2 cation-anion

interaction with temperature and pressure has been a long-standing difficult problem. The interaction
between had been characterized by Phutela and Pitzer 1986 and later corrected by Archer and Rard
1998 with regression of the thermodynamic data of the MgSO4-H2O system. However, this is not possible
for , in the CaSO4-H2O system, according to the Pitzer model,

(14)

Because of its much lower solubility, thus low value of m and I in equation 14, only , which
is a large negative value, can be estimated accurately using solubility data in the CaSO4-H2O system.
Contribution of term is small, contribution from and terms are less or close to
experimental uncertainty, their contributions to the mean activity coefficient are adding a small number to
a large number, therefore regression of with m would not give good estimation of these terms.
To estimate and it is necessary to raise their contribution by increasing the ion
concentration. This can only be realized in the common-ion Ca system or common ion SO4 system.
SPE-200699-MS 13

Common ion solubility data, CaSO4 solubility in high concentration of either CaCl2 or Na2SO4, are needed
to bettercharacterize the interaction of . This characterization is also practically important, because
most brines show high calcium low sulfate properties. The calcium concentration could be as high as 0.5 M,
or higher. For the other two interactions, and practically their concentration is many orders
of magnitude lower compared to it is even more difficult to characterize them, but on the other
hand, there is no practical need to characterize them because their contributions to the activity coefficient
are negligible.
The common ion Ca2+ or common ion SO42- data sources used in regression are listed in
Table 4. In the regression process in the Pitzer model, besides the Ca-SO4 interaction parameters
another two mixture terms, and are also
estimated. Optimization of the T/P coefficients of these terms takes a matrix approach, the RMSE for
saturation index (SI) is roughly 0.05, which is equivalent to roughly ±6% uncertainty of relative solubility.
Figures 9 and 10 show the plotted solubility of calcium sulfates in Na2SO4 solution and CaCl2 solution,
respectively. The behavior of calcium sulfates solubility in these two common- ion solutions are different,
while both show a "v" shape profile at [Ca] or [SO4] = 0.05m, even higher [Na2SO4] will increase the
solubility, while even higher [CaCl2] will decrease the solubility at T=85°C.

Table 4—Source of common ion solubility data used to characterize Ca: SO4 interaction

Mineral T/K P/bar system-H2O No. Reference Common ion

Anh 115 2 Ca-SO4-Na-Cl 78 Deng 2018 Ca<1.33m

230 25
Anh 50 1 Ca-SO4-Cl 12 Li & Demopoulos Ca<1.5m
80 1 2005
Anh 150 5 Ca-SO4-Na 20 Freyer & SO4<1.12m
200 15 Voigt 2004
Anh 150 5 Ca-SO4-K 7 Freyer & SO4<1.33m
200 15 Voigt 2004
Gyp/Anh 25 1 Ca-SO4-Na 120 Block & SO4<1.5m
85 1 Ca-SO4-Na-Cl Waters 1968

Gyp 25 1 Ca-SO4-Cl 60 Li & Demopoulos Ca<2.06m

80 1 2005
Gyp 25 1 Ca-SO4-Na-Cl 16 Tanji 1969 Ca<0.05 SO4<0.05
Ca-SO4-Mg-Cl
Ca-SO4-Cl
Ca-SO4-Na
Bar 25 35 Ba-SO4-Na- 16 Shi, et al. 2012 Ca<0.55m
250 1388 K-Mg-Ca-Cl
14 SPE-200699-MS

Figure 9—Predicted solubility of gypsum (upper curves) and anhydrite (lower curves)
in Na2SO4 solution (<1 m) at various temperature, compared with experimental data.

Figure 10—Predicted solubility of gypsum (upper curves) and anhydrite (lower curves)
in CaCl2 solution (<1 m) at various temperature, compared with experimental data.

Conclusions
Motivated by the practical needs in the oil and gas industry, this study focuses on improving the Pitzer
model to accurately predict the solubility of the unconventional sulfide scale and the solubility of sulfate
scale under high calcium and high sulfate conditions. These improvements could help the scale prediction
in oil and gas industry with more aggressive and extreme conditions.
SPE-200699-MS 15

With a speciation approach on the iron sulfide system, a reliable iron sulfide solubility prediction model
has been developed, under a wide range of brine compositions (temperature 23 to 125°C, 0 to 5 M ionic
strength and pH 3 (or lower) to 9, S(-II) to Fe(II) molar ratio 0.01 to 10,000) with general prediction error is
less than ± 0.3 SI. While with corrected Ksp value for sulfate scale and the newly derived Ca-SO4interaction
term, the improved solubility model of gypsum, anhydrite and barite have extended the prediction boundary
to T<250°C, P<1400bar, Ca<1.3m, SO4<1.5m with prediction error of less than ±0.15 SI. These newly
derived improvements are included in ScaleSoftPitzer2020.
There are still deviations remaining due to not enough data for fitting the virial coefficients in the mixed
electrode under some extreme conditions, especially for the iron sulfide system. More experiments to study
the solubility of these unconventional scale under wide ranges of different condition would be required to
further improve the model predictions. In summary, this study improved our prediction for these exotic scale
solubility predictions with both experimental studies and model validation. The newly developed model
extends the valid ranges of temperature, pressure and chemical composition in produced brine to better fit
the more complicated field conditions under extreme conditions.

Acknowledgement
This work was financially supported by Brine Chemistry Consortium companies of Rice University,
including Aegis, Apache, BHGE, BWA, Chevron, ConocoPhillips, Coastal Chemical, EOG Resources,
ExxonMobil, Flotek Industries, Halliburton, Hess, Italmatch, JACAM, Kemira, Kinder Morgan, Nalco,
Oasis, Occidental Oil and Gas, Range Resources, RSI, Saudi Aramco, Schlumberger, Shell, SNF, Statoil,
Suez, Total and the NSF Nanosystems Engineering Research Center for Nanotechnology-Enabled Water
Treatment (ERC-1449500).

References
Appelo C.A.J. (2015). Principles, caveats and improvements in databases for calculating hydrogeochemical reactions in
saline waters from 0 to 200°C and 1 to 1000atm. Applied Geochemistry, 55, 62–71.
Archer D.G. and Rard J.A. (1998). Isopiestic investigation of the osmotic and activity coefficients of aqueous MgSO4
and the solubility of MgSO4-7H20(cr) at 298.15 K: thermodynamic properties of the MgSO4 + H20 System to 440K.
J. Chem. Eng. Data, 43 (5), 791–806.
Benning, L. G., Wilkin, R. T., and Barnes, H. L. (2000). Reaction pathways in the Fe-S system below 100 C. Chemical
Geology, 167(1-2), 25–51.
Block J, Waters OB. 1968. Calcium sulfate-sodium sulfate-sodium chloride water system at 25. deg. to 100. deg. Journal
of chemical & engineering data., 1968, 13 (3), p.336
Blount, C.W. (1977). Barite Solubilities and Thermodynamic Quantities up to 300 degrees C and 1400 bars. Am Mineral,
62 (9-10), 942–957.
Christov, C., & Moller, N. (2004). Chemical equilibrium model of solution behavior and solubility in the H-Na-K-OH-C1-
HSO4-SO4-H20 system to high concentration and temperature. Geochimica et Cosmochimica Ada, 68 (6), 1309–1331.
Dai, Z., Kan, A., Zhang, F., et al (2015). A Thermodynamic Model for the Solubility Prediction of Barite, Calcite, Gypsum,
and Anhydrite, and the Association Constant Estimation of CaSO4 (0) Ion Pair up to 250 C and 22000 psi. Journal
of Chemical & Engineering Data, 60 (3), 766–774.
Dai, Z., Kan, A. T., Shi, W., et al (2017). Solubility measurements and predictions of gypsum, anhydrite, and calcite
over wide ranges of temperature, pressure, and ionic strength with mixed electrolytes. Rock Mechanics and Rock
Engineering, 50 (2), 327–339.
Davison, W., Phillips, N., et al (1999). Soluble iron sulfide species in natural waters: reappraisal of their stoichiometry
and stability constants. Aquatic Sciences, 61 (1), 23–43.
Deng, G., Kan, A. T., Dai, Z., et al. (2018, June). Impact of High Calcium Concentration on Sulfate Scale Prediction at
High Temperature from 120° C to 220° C. In SPE International Oilfield Scale Conference and Exhibition. Society of
Petroleum Engineers.
Dickson A.G., Wesolowski D.J., et al (1990). Dissociation constant of bisulfate ion in aqueous sodium chloride solutions
to 250°C. J Phys. Chem, 94.
Dyrssen, D. (1988). Sulfide complexation in surface seawater. Marine Chemistry, 24 (2), 143–153.
16 SPE-200699-MS

Freyer, D. Voigt W. (2004). The measurement of sulfate mineral solubilities in the Na-K-Ca-C1-SO4-H20 system at
temperatures of 100, 150 and 200°C. Geochimica et Cosmochimica Acta, 68 (2), 307–318.
Greenberg, J.P. and Moller, N. (1989). The Prediction of Mineral Solubilities in Natural-Waters − a Chemical-Equilibrium
Model for the Na-K-Ca-C1-SO4-H20 System to High-Concentration from 0 oC to 250 oC. Geochim Cosmochim Ac,
53 (10), 2503–2518.
He, S. (1992). The carbonic acid system and solubility of calcium carbonate and sulfate in aqueous solutions over a wide
range of solution composition, temperature and pressure (Doctoral dissertation, Texas A & M University).
Helgeson, H. C. and Kirkham, D. H. (1974). Theoretical prediction of the thermodynamic behavior of aqueous electrolytes
at high pressures and temperatures; I, Summary of the thermodynamic/electrostatic properties of the solvent. American
Journal of Science, 274 (10), 1089–1198.
Holmes, H. F., J. M. Simonson and R. E. Mesmer (1997). "Aqueous solutions of the alkaline earth metal chlorides.
Corrected constants for the ion-interaction model (vol 26, pg 271, 1994)." The Journal of Chemical Thermodynamics,
29 (11), 1363–1373.
Jiang C. (1996). Solubility and solubility constant of barium sulfate in aqueous sodium sulfate solutions between 0 and
80°C. Journal of Solution Chemistry, 25, 105–111.
Johnson, J.W., Oelkers, E.H. and Helgeson, H.C. (1992). SUPCRT92: A software package for calculating the standard
molal thermodynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bar and 0 to 1000
C. Computers & Geosciences, 18 (7), 899–947.
Kan, A. T., Dai, J. Z., Deng, G., et al (2019). Recent Advances in Scale Prediction: Approach and Limitations. SPE Journal.
Kharaka, Y.K., Gunter, W.D., Aggarwal, P.K., et al. (1989). SOLMINEQ. 88: A computer program for geochemical
modeling of water-rock interactions. Department of the Interior, US Geological Survey.
Langmuir D. and Melchior D. (1985). The geochemistry of Ca, Sr, Ba and Ra sulfates in some deep brines from the Palo
Duro Basin, Texas. Geochimica et Cosmochimica Acta, 44 (11), 2423–2432.
Li Z. and Demopoulos G. (2005). Solubility of CaSO4 phases in aqueous HC1 + CaCl2 solutions from 283 K to 353 K.
1 Chem. Eng. Data. 50, 1971–1982.
Li, J., and Duan, Z. (2011). A thermodynamic model for the prediction of phase equilibria and speciation in the H20-0O2-
NaC1-CaCO3-Ca504 system from 0 to 250 C, 1 to 1000 bar with NaC1 concentrations up to halite saturation.
Geochimica et Cosmochimica Acta, 75 (15), 4351–4376.
Liu, Y., Zhang, Z., Bhandari, N., Dai, Z., et al(2017). New approach to study iron sulfide precipitation kinetics, solubility,
and phase transformation. Industrial & Engineering Chemistry Research, 56 (31), 9016–9027.
Luther, G. W., Rickard, D. T., Theberge, S., et al (1996). Determination of metal (bi) sulfide stability constants of Mn2+,
Fe2+, Co2+, Ni2+, Cu2+, and Zn2+ by voltammetric methods. Environmental Science & Technology, 30 (2), 671–679.
Moller, N. (1988). The Prediction of Mineral Solubilities in Natural-Waters − a Chemical-Equilibrium Model for the Na-
Ca-C1-SO4-H20 System, to High-Temperature and Concentration. Geochimca et Cosmochim Acta, 52 (4), 821–837.
Naumov, G. B., Ryzhenko, B. N., and Khodakovsky, I. L. (1974). Handbook of thermodynamic data. Transl. into
ENGLISH of the pub!." Handbook of Thermodyanic Data" Moscow, Atomizdat., 1971.
Oakes C.S., Pitzer K.S. and Michael S.S. (1998). The system NaCl-CaC12-H20: part 3. Heats of dilution and mixing at
373 to 573 K and 21.5 MPa using a new high-temperature, flow-through calorimeter. Geochimica et Cosmochimica
Acta, 62 (7), 1133–1146.
Oscarson JL, Izatt RM, Brown PR, Pawlak Z, Gillespie SE, Christensen JJ (1988). Thermodynamic quantities for the
interaction of S042- with H+ and Na+ in aqueous solution from 150 to 320°C. J Solution Chem,17(9),841–863.
Pabalan, R.T. and Pitzer, K.S. (1988). Heat capacity and other thermodynamic properties of Na2SO4 (aq) in hydrothermal
solutions and the solubilities of sodium sulfate minerals in the system Na-C1-504-0H-H20 to 300 C. Geochimca et
Cosmochim Acta, 52 (10), 2393–2404.
Parkhurst, D. L., and Appelo, C. A. J. (2013). Description of input and examples for PHREEQC version 3: a computer
program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations (No. 6-A43).
US Geological Survey.
Phutela, R. C. and K. S. Pitzer (1986). "Densities and apparent molar volumes of aqueous magnesium sulfate and sodium
sulfate to 473 K and 100 bar." Journal of Chemical & Engineering Data, 31 (3), 320–327.
Pitzer, K. S. and G. Mayorga (1974). "Thermodynamics of electrolytes. III. Activity and osmotic coefficients for 2-2
electrolytes." Journal of Solution Chemistry, 3 (7), 539–546.
Pitzer, K. S., J. C. Peiper and R. H. Busey (1984). "Thermodynamic properties of aqueous sodium-chloride solutions."
Journal of Physical and Chemical Reference Data,13(1), 1102.
Rickard, D. (2006) "The solubility of FeS." Geochimica et Cosmochimica Acta, 70 (23), 5779–5789.
Rickard, D. and Luther, G. W. (2007). Chemistry of iron sulfides. Chemical reviews, 107 (2), 514–562.
Rogers, P. and Pitzer, K.S. 1981. High-temperature thermodynamic properties of aqueous sodium sulfate solutions. The
Journal of Physical Chemistry, 85 (20), 2886–2895.
SPE-200699-MS 17

Rowland, D., Königsberger, E., Hefter, G., et al. (2015). Aqueous electrolyte solution modelling: Some limitations of the
Pitzer equations. Applied Geochemistry, 55, 170–183.
Seewald, J. S., and Seyfried, W. E. (1991). Experimental determination of portlandite solubility in H20 and acetate
solutions at 100-350 °C and 500 bars: Constraints on calcium hydroxide and calcium acetate complex stability.
Geochimica et Cosmochimica Ada, 55 (3), 659–669.
Shi, W., Kan, A.T., Fan, C.F., et al. 2012. Solubility of Barite up to 250 degrees C and 1500 bar in up to 6 m NaC1 Solution.
Ind Eng Chem Res 51 (7), 3119–3128.
Stuart, M., and Box, K. (2005). Chasing equilibrium: measuring the intrinsic solubility of weak acids and bases. Analytical
chemistry, 77 (4), 983–990.
Suleimenov, O. M. and Seward, T. M. (1997). A spectrophotometric study of hydrogen sulphide ionisation in aqueous
solutions to 350 C. Geochimica et Cosmochimica Acta, 6/(24), 5187–5198.
Tanji. K.K. 1969. Solubility of gypsum in aqueous electrolytes as affected by ion association and ionic strengths up to
0.15M and 25°C. Environmental Science & Technology, 3 (7), 656.
Theberge, S. M., and Luther III, G. W. (1997). Determination of the electrochemical properties of a soluble aqueous FeS
species present in sulfidic solutions. Aquatic Geochemistry, 3 (3), 191–211.
Wang, X., Ko, S., Liu, Y., et al(2019, March). Kinetics and Thermodynamics of Iron Sulfide, Precipitation, Deposition
and Control. In SPE International Conference on Oilfield Chemistry. Society of Petroleum Engineers.