SPE 155127

A New Correlation to Predict the Stoichiometry of Mixed Scale: Iron-

Calcium Carbonate
H.A Alsaiari, SPE, Saudi Aramco, N. Zhang, S. Work, A. Kan, SPE, and M.B. Tomson, SPE, Rice University

Copyright 2012, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE International Conference and Exhibition on Oilfield Scale held in Aberdeen, UK, 30–31 May 2012.

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
Calcium carbonate and iron carbonate scales are widely observed in oil and gas production. Scale formation can be
useful for corrosion control; however, excessive scale buildup can lead to severe production loss. What is called calcite scale
in the field is almost always a solid solution of iron in calcite. Yet little attention has been paid to the precipitation of these
mixed calcium-iron carbonate scales. As a result, knowledge of the formation and inhibition of mixed calcium/iron scales is
very limited.
Normally, calcite scale formation is readily inhibited, whereas siderite inhibition is notoriously difficult. The solid-
solution transition from predominantly calcite to predominantly siderite properties is unknown. Besides, although the
solubility of mixed scale can differ by several orders of magnitudes from the solubility of its pure salts, scale prediction
models are normally developed based on the data from pure solids. Finally, the incorporation of iron into calcite solid
dramatically alters the kinetics of scale growth, as will be illustrated.
A series of experiments were performed to precipitate mixed iron-calcium carbonate (FexCa1-XCO3), ranging from
calcium-rich to iron-rich. The experiments were conducted at 7.3±0.2 pH in 0.5 M NaCl at 55 oC. The work was performed
with a new constant composition method, modified to handle oxygen sensitive ferrous carbonate scale and solid solutions.
Based upon the experimental results and a flux-based theoretical derivation, a new correlation in a form of a logistic
function has been developed to predict the composition of FexCa1-xCO3 as a function of the aqueous composition. The model
is an excellent representation for all of the experimental results, with R2 greater than 0.97. The correlation and methods
developed in this work can readily be adapted to other mixed scale systems. Laboratory results will be compared with field
observations and the consequences discussed.

Introduction
Understanding of the solid solution formation from the aqueous solutions is needed to predict and prevent solid
formation in various water treatment processes. Several studies have examined the thermodynamics of solid solution
formations in geochemical processes where equilibrium conditions may be assumed [Lippman, 1980 and 1982; Plummer
1987; Glynn and Reardon, 1990; Königsberger and Gamsjäger, 1990; Gamsjäger. et al., 2000; Pina and Putnis, 2002,
Astilleros, et al., 2003; Pina et al., 2004]. However, solid solution formation is rarely at equilibrium in industrial water
treatment processes. The key parameters needed for the prediction of solid solutions formation for industrial water treatment
processes are the stoichiometry of solid solution formation and kinetics at different aqueous compositions and conditions. If
these parameters were known, scientists and engineers would be able to project, with precision, the incorporation of ions
from waters into minerals. They would also help to infer the physicochemical conditions of natural crystallization
environments.
Typically, precipitation is studied by preparing a supersaturated solution and following the concentration of the lattice
ions with time. This method is known as free drift. Using free drift is particularly problematic to study solid solution
formation, because as the concentrations decrease the solid stoichiometry frequently changes. In addition, the amount of solid
solution formed is often very small as the reaction approaches equilibrium. Most of these restrictions have been overcome,
for the first time, by modifying the method of constant composition, introduced by Tomson and Nancollas, 1978. This
approach has the ability to hold the supersaturation of the aqueous solution at a constant value for long periods of time. At
very low supersaturation, macroscopic quantities of stochiometric and highly crystalline material can be formed and collected
2 SPE 155127

for analysis. It also solves the difficulty of determining the stoichiometry of precipitation when the experiment has
appreciable reduction in the initial concentrations of crystal lattice ions during the reaction.
This method also enables one one composition of solid solution to be studied at a time, at specific supersaturation
conditions. Using constant composition method enables researchers to study the influence of pH, ionic strength, seed type,
and degree of supersaturation with a precision hitherto unobtainable. This work was intended to; 1) develop an experimental
design to be workable for oxygen sensitive salts, and also to be applicable for studying the stoichiometry and kinetics of
crystal growth of solid solutions in addition to a single salt; and 2) develop a correlation to predict the stoichiometry of
iron/calcium carbonate solid solutions as a function of aqueous supersaturation with respect to the pure solid phases.

Experimental Procedure

Experimental Apparatus and Design

A picture and a schematic of experimental design are shown in Figures 1 and 2, respectively. The design has the
flexibility to deal with oxygen-sensitive ferrous ions. All fluid delivery lines (Peek Tubing) and experimental apparatus were
pre-flushed with argon gas several times before conducting the experiment. The titrant containers, titrant syringes, and titrant
lines were connected to the flushing gas cylinder and reaction cell by three-way valves to easily flush the lines and containers
in an oxygen-free environment. An oxygen gas absorber (SUPELCO, Model 503088) was connected to the outlet of the gas
cylinder as an additional measure to trap the trace concentration of oxygen in the gas cylinder. The precipitation reactions
were performed in a double-walled glass reaction vessel connected to a water bath for temperature control. The titrant
delivery lines were dipped in a separate double-walled glass vessel connected to a water bath to maintain the titrant solutions
to be at the same temperature as the reaction solution. A digital caliber was connected to a syringe pump and a PC to track
the amount of titrant-solution added during the precipitation processes.

Reaction Solution Preparation

Reagent grade chemicals and deionized water were used to prepare the reaction and titrant solutions. Magnetite powder
(Aldrich grade) with a particle size less than 5 micron was used as seed materials. In a typical experiment, nitrogen gas with
1% CO2 was sparged through 600 ml of 0.5 M NaCl solution in water-jacked Pyrex cell to exclude oxygen. 1.0 % CO2 gas
was used to stabilize the concentration of total carbonate. The sparging process was maintained throughout the entire time of
each experiment. The oxygen content was measured at the beginning and at the end of each experiment by a colorimetric
analysis technique. (O2 content was less than 5 ppb). The solution in the reaction cell was supersaturated with respect to
calcium carbonate and iron carbonate by addition of sodium bicarbonate salt (Fisher grade), calcium chloride dihydrate salt
(EMD grade), and ferrous chloride 4-hydrate salt (J.T. Bakker grade). The salts were added to the reaction cell after one hour
of the sparging. The amount of salts added was based upon the level of supersaturation required for precipitating a specific
composition of FeXCa1-XCO3 solid solution. A known amount of de-aerated seed crystals was added to the reaction solution to
induce the precipitation.

Figure 1: Picture of the experimental design and its basic components.

SPE 155127 3

Figure 2: Schematic of the experimental design of constant composition method: 1) magnetic stirrer; 2)water bath; 3)pH
control; 4) reaction cell; 5) syringe pump; 6) CO2 gas; 7) argon gas; 8&9) oxygen scavenger; 10) Titrant-1 container; 11)
Titrant-2 container; 12) glass syringe with caliper; 13) sampling point; and 14) computer recorder.

Titrant Preparation
Two titrants, each in a separate container, were used to maintain the experiment conditions at constant thermodynamic
driving force (constant pH, ionic strength, and initial concentrations). Titrant-1 contained the cations (calcium and ferrous
ions) while titrant-2 contained anions (bicarbonate and hydroxide ions). Titrant-1 was prepared by adding calcium chloride
dihydrate (EMD, assay 99.0-105%) and sodium chloride (SIGMA-ALDRICH, assay 99.0%) salts to DI water. The solution
was sparged with argon gas (ultra high purity) for about two hours. Ferrous chloride 4-hydrate salt (J.T.BAKER, assay
101.5%) was added into the titrant container afterward to avoid the oxidation of ferrous ions. Similarily, titrant-2 was
prepared by adding sodium hydroxide solution (TRISTAR, 1.000±0.002N), sodium chloride salt (SIGMA-ALDRICH, assay
99%), and sodium bicarbonate salt (Fisher grade), to DI water and, then, the mixture was deoxygenated by argon gas. At the
end of the de-oxygenation processes the titrant containers were maintained at a little positive pressure to prevent
contamination with air. Samples were collected from each titrant for analysis. The equations below were developed to
calculate the total concentrations of cations and anions in titrants. These equations are similar to what has been reported by
Zhang and Nancollas for a pure solid phase, [Zhang and Nancollas, 1992], but modified to include both Ca2+ and Fe2+. For
the reaction:
xFe2+ + (1-x)Ca2+ + HñCO3ñ-2 = Ca1-xFexCO3(s) + ñH+ Eq. 1

where: 0 ≤ x ≤ 1, and 0 ≤ ñ ≤ 2

Titrant-1:
[CaCl2]Titr = 2[CaCl2]Sol + [CaCl2]Ppt Eq. 2
[FeCl2]Titr = 2[FeCl2]Sol + [FeCl2]Ppt Eq. 3
[FeCl2]Ppt = [x/(1-x)][CaCl2]Ppt Eq. 4
[NaCl]Titr = 2[NaCl]Sol – 2CPpt Eq. 5
CPpt = [CaCl2]Ppt + [FeCl2]Ppt Eq. 6

Titrant-2:
[NaHCO3]Titr = 2[NaHCO3]Sol + [NaHCO3]Ppt Eq. 7
[NaHCO3]Ppt = CPpt Eq. 8
[NaOH]Titr = ñ CPpt Eq. 9
4 SPE 155127

Where,
The subscripts Titr, Sol, and Ppt refer to the titrant solution, reaction solution, and the precipitate concentration, in mol/L,
respectively. CPpt in Eq. 6 is the concentration of the precipitate; any value can be selected for Cppt. To have enough volume
of titrant addition during the course of the experiment, it is recommended to select small concentration for CPpt (≤ 0.1 mol/L)
when the precipitation rate is slow and large concentration (> 0.1 mol/L) when the precipitation rate is high. ñ is the average
number of protons bound to a carbonate ion which can be expressed as shown in the following equation:

[HCO3 ]  2[H 2 CO3 ]

ñ Eq. 10
C

Where,
C is the concentration of total carbonate. The value of ñ is calculated from the ionization constant of carbonates, pH, and total
carbonate concentration, or can be read from an alkalinity titration. Plot of ñ as a function of pH at 55oC is shown in Figure
3.

HnCO3n-2 at 55 oC & IS=0.5 M NaCl

2
Number of proton bounded,

1.8
1.6
1.4
1.2
1
n

0.8
0.6
0.4
0.2
0
2 3 4 5 6 7 8 9 10 11 12
pH

Figure 3: Average number of protons bound per carbonate ion as function of solution pH at IS = 0.5 M and 55oC. Total
carbonate was calculated by mass balance equation. The activity coefficient of the proton was calculated by using Davies
equation. The ionization constants of carbonate, Ka1 and Ka2, at 55OC were obtained from SSP Software (pK1 = 6.286 and
pK2 = 10.157)

Precipitation Technique at Constant Composition Method

pH dropped as precipitation occurred according to the following reaction

( ) ̃
̃ ( ) ̃ Eq. 11

The output of the pH (Orion pH electrode) in the controller (EUTECH-alpha-pH2000w) was constantly compared with a
preset value. The difference in electric potential triggered a motor of a syringe pump (Harvard Apparatus-
Infusion/Withdrawal Pump) to deliver titrants from two syringes (SGE- 100 ml class syringe) into the reactor cell thereby
maintaining a constant thermodynamic driving force. The processes repeated each time pH dropped. The amount of titrants
added into the reaction solution with time during the course of the experiment was tracked by a digital recording caliber
(WIN-2 RS232 INTERFACE) connected to a PC. Samples from the reaction solution were collected each 10 minutes for
analysis.

Experiment Tactic and Ideal Experiment

The tactic used in this work was based upon systemmatic trial and error. A series of experiments was conducted, each
at different initial concentrations of lattice ions at 55oC. The concentrations of calcium and bicarbonate ions were selected to
be within the precipitation zone of calcium carbonate, and also to reflect the concentrations that commonly exist in the
SPE 155127 5

natural waters and industrial brines. On the other hand, ferrous iron concentrations were designed to be within the
precipitation zone of iron carbonate and outside the precipitation region of ferrous hydroxide ([Fe2+] < 10-2.4 M, and pH ≤
7.5). Visual MINTEQ package and SSP software were used to diagnose the precipitation regions.
The concentrations of lattice ions were varied in each experiment by assuming an arbitrary value for x in the chemical
formula, FexCa1-xCO3, then the reaction and titrant solutions were prepared as described in Eqs. 2-10. The concentrations of
calcium and ferrous ions were then monitored during the course of the precipitation experiment. The first experiment was
usually considered as non-ideal due to the change in the concentrations of lattice ions. The level of change in the
concentrations corresponded to how much the assumed value “x” was far from the correct value. In the subsequent
experiment, the value of x was projected by performing a mass balance on the concentrations of iron and calcium ions in the
preceding experiment. The titrant concentrations were then adjusted accordingly. The procedures were repeated for several
iterations until no appreciable change in the concentrations of iron, calcium, and bicarbonate was observed in the reaction
solution. Two to three experiments were generally needed to reach constant composition versus time. The acceptable change
in lattice ion concentrations was within 2% for calcium and iron ions, and 5% for bicarbonate. At this point, the constant
composition experiment was considered successful in which the rate of titrant addition is characteristic of growth rate and
Fe/Ca ratio is characteristic of stoichiometry of the solid solution formed. Plots of the concentrations of constituent ions and
amount of titrant added, for ideal experiments, are shown in the discussion section.

Analyses and Sampling Procedures

The concentrations of iron, calcium and bicarbonate in the reaction solution were measured periodically by collecting
samples from the reactor cell. Before collecting the sample, the sampling line was flushed with 10 milliliters of the reaction
solution. Five milliliters of the reaction solution were withdrawn by a plastic syringe. The sample was filtered through a 0.45
m filter to remove suspended solids. One half (0.5) gram of the filtered samples was then transferred into a vial and diluted
with 10 grams of 2% HNO3 to be within the measurable range. Calcium and iron concentrations were measured at the end of
the experiment using an inductive coupled plasma atomic emission spectrophotometer (ICP-AES). Total alkalinity was
measured immediately after sampling to minimize further precipitation of carbonate salts. The alkalinity measurement was
performed by transferring 4 ml of the filtered sample to a small glass tube already containing bromocresol green-methyl red
indicator powder (HACH). Then, the sample was titrated with 0.05 N HCl till the end-point. At the end point, the sample
color turned from blue to red. The total alkalinity was calculated by using the equation below:
[ HCl ]  Vol HCl
Alk  Eq. 12
Vol sample

Where,
Alk : total alkalinity in mol/L
[HCl] : concentration of hydrochloric acid in mol/L
VolHCl : added volumes of HCl in liter
Volsample : volumes of sample in liter.

At the end of precipitation experiments, aqueous solution was rapidly removed from the reaction cell and the leftover crystals
were dried in-situ inside the reaction cell with ultra pure argon gas at 60 oC and 1 atm. The crystals were left for couple of
days, and then analyzed using X-ray diffraction (XRD), a scanning electron microscope, SEM (FEI Quanta 400 ESEM FEG),
and Energy-dispersive X-ray spectroscopy (EDX) systems.

Results and Discussion

Formation of aragonite
Calcium carbonate occurs in several polymorphic forms. Calcite and aragonite are the common types of calcium
carbonate that exist in nature and in industrial processes. The values of the thermodynamic solubility product constants of
aragonite and calcite at 55 oC are 10-8.44 and 10-8.71, respectively, [SSP, Tomson and Kan, Rice University]. The solubility
product constant of aragonite is about 0.27 SI units higher than that of calcite at 55°C. Although the difference in Ksp is not
significant, but it is probably worthy to mention that aragonite was the predominant calcium carbonate solid formed along
with siderite during the precipitation processes. Surprisingly, even at approximately 1.4 % iron content in the solid solution,
aragonite was formed. This revealed that the metastable calcium carbonate (aragonite) is probably the preferable form in the
presence of iron ions. These results were consistent with the results previously reported by Herzong, et al., (1989) and
Langerak, et al., (1999). They reported that ferrous iron inhibited calcite growth but not aragonite growth. The inhibition
effect on calcite formation probably results from the presence of ferrous iron, not in solid solution, but in the aqueous
solution. Nevertheless, to the best knowledge of the author, no reports are available in the literature showing the conditions at
which calcite forms in the presence of iron.
6 SPE 155127

Some of the peaks of XRD analysis of one of the iron-rich-solid sample (FexCa1-xCO3, x1) are in a good agreement with the
beaks reported by Johnson for pure siderite sample, [Johnson 1990]. However, the vast majority of the peaks was in a good
agreement with magnetite peaks. This was probably due to the pre-existence of magnetite as crystal seeds and also probably
due to the spontaneous oxidation of iron carbonate during drying processes. It had been observed that, during the drying
processes, the color of the solid gradually changed from white to green and then, eventually, either to brown or black based
upon the content of iron on the precipitated solid solution. This observation is consistent with what has been reported in the
literature. According to Chai and Navrotsky (1994), the final product of siderite decomposition is hematite (Fe2O3) in an
oxidizing atmosphere, magnetite (Fe3O4) in a CO2 atmosphere and in vacuum, and magnetite and wüstite (FeO) in an inert
atmosphere, [Chai and Navrotsky, 1994]. Magnetite is commonly observed in tubing in gas and oil wells.

Composition of Iron/Calcium-Carbonate Solid Solutions

A series of iron-calcium carbonates solid solutions, ranging from calcium-rich to iron-rich, was forced to precipitate by
changing the initial supersaturation of pure phases in reaction solution. The experimental conditions (initial concentrations of
lattice ions, composition of pure phases in the solid, saturation indexes of pure phases, pH and temperature) are summarized
in Table 1. Each constant composition experiment had different initial aqueous concentrations of calcium and iron, and
calcium to iron molar ratio. The difference in supersaturation with respect to pure iron carbonate and calcium carbonate
increased gradually from Experiemnt-1 to Experiment-9. The supersaturation was expressed by saturation index (SI) with
respect to pure phases. The saturation index of the pure iron carbonate is defined as:
aFe 2 aCO 2
SI FeCO3  log 3
Eq. 13
K SP,FeCO3 (T , P)

And the saturation index of the pure calcium carbonate is defined as

aCa2 aCO 2
SI CaCO3  log 3
Eq. 14
K SP,CaCO3 (T , P)

where,
aFe2+ , aCa2+, and aCO3(2-) are the aqueous activities of ferrous ions, calcium ions, and carbonate ions, respectively. Ksp,FeCO3, and
Ksp,CaCO3, are the thermodynamic solubility products constant of FeCO3 and CaCO3 , respectively. The ion activities and SI
were calculated using SSP software, developed by Tomson and Kan, Rice University.

Figure 4 shows the trends of the ion concentrations of calcium, iron and alkalinity with time for one of the experiments,
(Fe0.37Ca0.63CO3). The concentrations of the constituent ions were held constant during the course of the experiment. The pH,
the actual amount of titrant added, and the corrected amount of titrant added are shown in Figure 5. The corrected amount of
titrant added in Figure 5 represents the amount of titrant added per constant effective surface area of seed material. This is
because the effective surface area available for crystal growth increases as the crystals grow. As a result, this leads to a more
rapid addition of titrant. According to Zhang and Nancollas (1992), Eq.15 was used to calculate the corrected amount of
titrant added for a constant effective surface area as a function of actual amount of titrant added, [Zhang and Nancollas;
1992].
1 p
 MC pptV 
1   1
 m0  Eq. 15
Vc 
 MC ppt 
(1  p ) 
 m0 

Vc : the corrected amount of titrant for a constant effective surface area, in liter.
M : the molecular weight of the precipitated solid solution in gram per mole.
V : the actual amount of titrant added in liter.
m0: the mass of crystal, in grams, at t = 0 .
Cppt : the concentration of the precipitated solid solution, mol/L.
p : the parameter of growth type, p = 2/3 for isotropic three dimensional growth, [Zhang and Nancollas; 1992].
SPE 155127 7

1600 240
Ca Alk Fe

& Alk, ppm

1200 180

Fe2+, ppm
2+
800 120

400 60
Ca

0 0
0 20 40 60 80 100 120 140 160 180
Time, min
Figure 4: Iron, calcium, and bicarbonate concentrations in the precipitation experiment of Fe0.37Ca0.63CO3 solid solution. (♦)
aqueous calcium concentration in mg/l. (•) aqueous iron concentration in mg/l. (■) Alkalinity concentration in mg/l. The first
20 minutes of the experiments was ignored for precipitation stabilization.

12 20
V_actual
V_actual, (ml) & pH

16

V_corrected, (ml)
9
pH 12
6
8
V_corrected
3
4

0 0
0 20 40 60 80 100 120 140 160 180
Time ,min

Figure 5: (◊) actual volume of titrant added in milliliters. (о)The corrected volume of titrant added due to change in surface
area of the crystal seeds in the precipitation experiment of Fe0.37Ca0.63CO3 . (Δ) pH reading.
8 SPE 155127

Table 1: Summary of the experimental conditions of constant composition experiments

T,oC IS Ca2 Fe2 Alk aCa2 aFe 2 aCa2 ,eq aFe 2 ,eq pH,
+ +
EXP , , , , , ±0.0 SIaragonit SIsiderit Seed type XFeCO

# ±0.0 M mg/ mg/ mg/ mM mM mM mM 2 a a b

e e 3
1 55 0.5 2783 25.1 213. 7.1 1.04 1.12 Fe0.01Ca0.99C 0.012
11.38 0.064 3.570 0.017
2 l 0l l0 O3
2 55 0.5 2304 136. 212. 7.37 1.03 1.93 magnetite 0.025
9.83 0.355 1.029 0.005
0 0
3 55 0.5 1511 177. 219. 7.29 0.88 2.08 0.151
6.89 0.501 1.488 0.007 magnetite
6 0
4 55 0.5 1093 178. 181. 7.28 0.60 1.94 0.367
5.19 0.517 1.558 0.007 magnetite
3 4
5 55 0.5 957.7 153. 284. 7.42 0.90 2.23 0.442
4.68 0.448 0.817 0.004 magnetite
7 4
6 55 0.5 976.7 155. 155. 7.19 0.40 1.73 0.528
4.68 0.459 2.358 0.011 magnetite
9 8
7 55 0.5 532.6 169. 170. 7.27 0.22 1.84 0.822
2.65 0.518 1.631 0.008 magnetite
9 6
8 55 0.5 361.7 168. 188. 7.34 0.14 1.93 0.933
1.86 0.527 1.182 0.006 magnetite
4 9
9 55 0.5 260.9 88.3 289. 7.4 0.35 2.01 0.99
1.37 0.287 0.896 0.004 magnetite
1 9
a. Saturation indexes were calculated by SSP-Software.
b. Molar fraction of ferrous iron in solid solution.

Theoretical Derivation of Solid Solution Stoichiometry as a Function of Aqueous Solution Composition

In a given aqueous solution containing seed crystals and supersaturated with more than one solid phase, the rate of
growth of a solid solution, e.g. (FexCa1-xCO3), based on kinetics and reaction probability can be expressed by the following
relation:
Rate = {driving force}{probability of reaction of ions in solution} Eq. 16
For pure calcium carbonate solid (CaCO3) the rate equation can be expressed as:
 aCa2 
Rate CaCO3  kCaCO3 (aCa2  aCa2 ,eq ) m    Eq. 17
 a 2  a 2 
 Ca Fe 

where,

RateCaCO3 : rate of precipitation of calcium carbonate in mol L-1 s-1

aCa2 : activity of calcium at time zero
aCa2 ,eq : activity of calcium at equilibrium
m : order of rate equation; m = 1.0 in this work

aCa2
: probability that Ca2+ will be added to the solid mixture.
aCa2  aFe 2

Similarly, for iron carbonate solid (FeCO3), the rate equation can be expressed as:

 a Fe 2 
Rate FeCO3  k FeCO3 (aFe 2  aFe 2 ,eq ) m    Eq. 18
 a 2  a 2 
 Ca Fe 
SPE 155127 9

At constant composition precipitation of FexCa1-xCO3 , the rate of Fe2+ addition is related to the rate of Ca2+ addition via fixed
parameter, ζ, therefore:

Rate FeCO3
 Eq. 19
Rate CaCO3

RTotal  Rate FeCO3  RateCaCO3 Eq. 20

Where: RTot, is the total rate of titrant addition to maintain a constant composition of solid solution FexCa1-xCO3.

RateFeCO3 = x RTot Eq. 21

RateCaCO3 = (1-x) RTot Eq. 22

Rearranging Eqs. 21 and 22 leads to:

Rate FeCO3 Rate CaCO3

 Eq. 23
x 1 x
where, x is the molar composition of FeCO3 in the precipitated solid solution, FexCa1-xCO3.

Rearranging Eq.23 and substituting Eq.19 leads to:

Rate FeCO3 x
  Eq. 24
RateCaCO3 1 x

Rearranging Eq.24 and substituting for ζ:

1
x Eq. 25
1
1

1
x Eq. 26
Rate CaCO3
1
Rate FeCO3

Eq. 26 is a form of logistic function. Substituting Eqs.17 and 18 in Eq.26 leads to:

1
xFeCO 3  Eq. 27
(aCa2  aCa2 ,eq )  aCa2
1  ks
(a Fe 2  a Fe 2 ,eq )  a Fe 2

where,
xFeCO3 : molar composition of siderite in the precipitated solid solution of FexCa1-xCO3
k
ks: the ratio of rate constant of pure phases; ks  CaCO3
k FeCO 3
a Fe 2 : activity of iron in solution at initial condition
a Fe 2 ,eq : activity of iron in solution at equilibrium
aCa2 : activity of calcium in solution at initial condition
aCa2 ,eq : activity of calcium in solution at equilibrium
10 SPE 155127

The activity at equilibrium can be calculated from solution pH and thermodynamic solubility product. It is assumed that the
thermodynamic solubility product constant, Ksp , of pure phases can be used also to calculate the equilibrium concentrations
of the constituent ions of the solid solution end-members. The typical expression of thermodynamic solubility product is:
K sp(T ,P )  acation,eq  aanion,eq Eq. 28
K SP(T ,P )
acation,eq  Eq. 29
aanion,eq
For carbonate solid solution, applying mass balance for total carbonate; Eq.29 can be expressed as:

K SP .aH2 
acation,eq  Eq. 30
K a1  K a 2  K H  PCO 2 gas
CO 2

Where,
Ksp : thermodynamic solubility product in molal unit, m2
a H  : activity of hydrogen, m
Ka1 : first ionization constant of H2CO3, m
Ka2 : second ionization constant of H2CO3, m
KH : Henry’s constant, m.atm-1
PCO2 : partial pressure of CO2, atm
 gas
CO 2
: Fugacity coefficient of CO2, unitless.

Validating Theoretical Model with The Experimental Results of FexCa1-xCO3:

To validate the model in Eq.27 the experimental results of solid solution composition at 55oC were plotted in
(aCa2  aCa,eq )  aCa2
Y-axis versus in X-axis. Figure 6 shows the experimental results and the trend produced from the
(a Fe 2  a Fe ,eq )  a Fe 2
theoretical model in Eq.27. The trend of the theoretical model is in a good agreement with the experimental results with
regression coefficient (R2) greater than 0.97. The value of ks, in Eq.27(4.3.15) was obtained by non-linear least square curve
fitting. ks was found to be 0.019. ks does not have units because it represents the ratio of the rate constants of end members,
kCaCO3
ks  . Macro SolvStat.xls package was used to obtain the regression statistics of the theoretical model, [Billo, 2001].
k FeCO 3
The value of ks, the standard deviation of ks, and the regression coefficient were shown in Table 2. More research has to be
done to address the the value of ks of FexCa1-xCO3 solid solution.

1.2
1 Experimental data
Correlation
0.8
XFeCO3

0.6
0.4
0.2
0
1 10 100 1000 10000
[(aCa - aCa,eq)aCa ]/[(aFe -aFe,eq)aFe ]

(aCa2  aCa,eq )  aCa2

Figure 6: Molar composition of solid solution, FexCa1-xCO3 versus at 55 oC Circles (•) represent
(a Fe 2  a Fe ,eq )  a Fe 2
actual experimental data. Line, ‫ــــــ‬, represents the theoretical model in Eq.27.
SPE 155127 11

Table 2: The correlation coefficients, standard deviation of the coefficients, and regression coefficient of FexCa1-
xCO3 system.

Coefficient Value Standard Deviation

ks 0.0196 0.0025

R2 0.97 -

The plot of solid solution composition, Y-axis, as a function of the difference in the aqueous solution supersaturation
with respect to pure iron carbonate and calcium carbonate, X-axis, is shown in Figure 7. As shown in Figure 7, at ∆SI = 0,
(∆SI = SIFeCO3 – SICaCO3, Eqs. 13 and 14) the crystallized solid solution was calcium-rich with probably trace amount of iron
carbonate (FeCO3 < 2%). In contrast, at ∆SI ≥ 1.5, the crystallized solid solution was iron-rich. In order to crystallize an equi-
molar solid solution of iron/calcium carbonate (Fe0.5Ca0.5CO3) at 55oC, 1 atm, and 0.5 M electrolyte solution (NaCl), the
aqueous solution should be supersaturated with respect to siderite approximately thirteen times more than the supersaturation
of aragonite (SIsid-SIcal ≈ 1.1). For example, if calcium concentration in water is 1000 mg/L, the concentration of iron should
be ~82 mg/L to allow an equi-molar of iron-calcium carbonate solid solution to precipitate.
The formation of calcium-rich solid solution at ΔSI = 0 reveals that calcium carbonate was kinetically favored over
iron carbonate in spite of its higher solubility. Also, it probably indicates that the presence of calcium ions impairs the
precipitation rate of iron carbonate. This observation is consistent with what has been observed in the previous work of batch
and continuous feeding experiments [ Alsaiari et al. 2008]. It was observed that the precipitation rate of iron carbonate was
markedly impacted when calcium ions were present at certain concentration in aqueous solution, [Ca2+] > 350 mg/L. This
result probably explains the common field observation wherein the deposit materials on the pipe are rich with calcium
carbonate and has low content of iron carbonate (FeCO3 < 10%) [Naser-Eldin, et al., 2002]. This phenomenon is also in a
good agreement with what has been noticed by Putnis et al. with respect to barium sulfate (barite; BaSO 4) and strontium
sulfate (celestite; SrSO4), [Putnis, et al., 1992]. Barium sulfate has lower solubility as iron carbonate, (Ksp=10-9.97 at 25oC,
;Stumm & Morgan) than strontium sulfate(Ksp=10-6.63 at 25oC, ;Stumm & Morgan). Putnis et al., found that the threshold
supersaturation value was around 10,000 for barium sulfate and 100 for strontium sulfate. Therefore, one can conclude that at
low ΔSI, the pure phase having higher solubility precipitates first and probably constitutes the core of the solid solution. This
phenomenon might be attributed to the difference in the characteristic water loss rate constant of the free ions in solution
and/or probably due to the difference in the interfacial free energy of the pure endmembers (inversely proportional to logKsp)
which might result in different values of the energy barriers.

1.2

0.8
XFeCO3

0.6

0.4

0.2

0
-0.5 0 0.5 1 1.5 2
ΔSI = SIFeCO3-SICaCO3

Figure 7: Molar composition of solid solution, FexCa1-xCO3, as a function of the difference in the supersaturation with
respect to pure end members at 55 oC .

SEM images and EDX analysis:

The SEM images provide information about the spatial distribution of the crystals and knowing whether the
nucleation and crystal growth of FexCa1-xCO3 occurred on the bulk aqueous solution or on the surface of the crystal seeds. An
increase in the crystal size of the seed material was observed in all the SEM images taken for the solid at the end of the
12 SPE 155127

experiment. This observation revealed that the growth of iron-calcium carbonate crystals most likely occurred only on the
surface of seed materials. SEM images of the surface of magnetite seeds before and after the precipitation experiments, at X
= 0.37 and 1.0 in the formula FexCa1-xCO3, were shown in Figure 8. SEM images of the crystal growth on surfaces of two
other different types of seed material: zero iron powder (Aldrich grade), and experimentally pre-prepared Fe0.01Ca0.99CO3
crystals, are shown in Figures 9 and 10. The grow and alteration of the spherical shape of zero iron due to precipitation, and
the increase in the crystal size of the Fe0.01Ca0.99CO3 supported the assumption that the precipitation of FexCa1-xCO3 solid
solutions occurred on the surface of seed materials.
Further, EDX analysis was conducted and the analysis results of three different surfaces: Fe0.01Ca0.99CO3,
Fe>0.99Ca<0.01CO3, pure zero iron powder, were shown as an example. The analysis of EDX on the surface of calcium-rich
solid solution, Fe0.01Ca0.99CO3, displayed a large peak of calcium atoms as shown in Figure 11. The large peaks of calcium
atoms and the small peaks of iron atoms were proportional to the high content of calcium atom and low content of iron atom
in the analyzed surface of the solid solution. In contrast with Fe0.01Ca0.99CO3 , a large peak of iron was observed for iron-rich
solid solution (Fe>0.99Ca<0.01CO3) and for pure zero iron samples, Figures 12 and 13. Similarly, the sizes of the peaks were
proportional to the content of iron and calcium atoms in these samples. The quality of EDX analysis was not good enough to
obtain further information on Fe/Ca ratio in iron-calcium carbonate solid solutions surfaces.

a b c

Figure 8: SEM images of the crystal before and after the experiment. (a) SEM image of pure magnetite used as crystal seeds
before the experiment. (b) SEM image after the precipitation experiments of Fe0.37Ca0.63CO3. (c) SEM image of after the
precipitation experiments of Fe1.0Ca0.0CO3.

a b

Figure 9: SEM images of the crystal before and after the experiment. (a)SEM image of pure zero iron powder used as crystal
seeds before the experiment. (b) SEM image after the precipitation experiments of Fe1.0Ca0.0CO3.
SPE 155127 13

a b

Figure 10: SEM images of the crystal before and after the experiment. (a)SEM image of pure zero iron used as crystal seeds
before the experiment. (b) SEM image after the precipitation experiments of Fe1.0Ca0.0CO3.

Figure 11: EDAX analysis of a sample of calcium-rich solid solution (Fe0.01Ca0.99CO3). Large peak of calcium and small
peak of iron reflect the content of iron and calcium in the precipitate.

Figure 12: EDAX analysis of a sample of iron-rich solid solution (Fe1.0Ca0.0CO3). Large peak of iron and small peak of
calcium reflect the content of iron and calcium in the precipitate.
14 SPE 155127

Figure 13: EDAX analysis of a sample of zero iron powder. Large peak of iron reflects the content of iron. The very small
peaks are probably due to the impurity in the sample of zero iron powder.

Conclusion

 A series of experiments were performed successfully to precipitate mixed iron-calcium carbonate (FexCa1-XCO3),
ranging from calcium-rich to iron-rich.
 The work was performed with a new constant composition method, modified to handle oxygen sensitive ferrous
carbonate scale and solid solutions.
 Based upon the experimental results and a flux-based theoretical derivation a new correlation in a form of a logistic
function has been developed to predict the composition of FexCa1-xCO3 as a function of the aqueous composition.

 The model is an excellent representation for all of the experimental results, with R2 greater than 0.97.
 The correlation and methods developed in this work can readily be adapted to other mixed scale systems.

Acknowledgement

This work was financially supported by Saudi Aramco, the Brine Chemistry Consortium companies of Rice University,
including Baker Hughes, BP, Champion Technologies, Chevron, Clariant, ConocoPhillips, Dow, Halliburton, Hess, Kemira,
Kinder Morgan, Marathon Oil, Multi-Chem, Nalco, Occidental, Petrobras, Saudia Aramco, Schlumberger, Shell, Shengli
Engineering & Consulting Co., Statoil, Southwestern Energy and Total, China-U.S. Center for Environmental Remediation
and Sustainable Development.

References

1. Alsaiari,H.A., Yean,S., Tomson, M.B., and Kan, A.,” Iron (II)-Calcium Carbonate: Precipitation Interaction. Paper SPE 114064
presented at the 2008 SPE International Oilfield Scale Conference, Aberdeen, UK, 28-29 May. DOI:10.2118/114064-MS.
2. Alsaiari,H.A., Tomson, M.B., and Kan, A.,” Effect of Calcium and Iron (II) Ions on the Precipitation of Calcium Carbonate and
Ferrous Carbonate. SPE Journal, SPE-121553, 2009.
3. Astilleros, J.M., Pina, C.M., Fernandezdiaz, L., and A. Putnis, A., “Supersaturation functions in binary solid solution–aqueous
solution systems”, Geochimica et Cosmochimica Acta, Vol. 67, No. 9, pp. 1601–1608, 2003.
4. Billo, E.J.,:” Excel for chemists: A comprehensive guide”. Second Edition, 2001.
5. Chai, L., and Navrotsky, A., “Enthalpy of formation of siderite and its application in phase equilibrium calculation.” American
Mineralogist, Vol.79, page 921-929, 1994.
6. Gamsjager, H., Konigsberger, E., and Preis, W.,:” Lippmann diagrams: Theory and application to carbonate systems”. Aquatic
Geochemistry, 6: 119-132, 2000.
7. Glynn, P. D., Reardon, E.J.,:” Solid-solution aqueous-solution equilibria: thermodynamic theory and representation. Am J Sci
278: 164-201, 1990.
8. Herzong, R.E., hi, Q., Patil, J.N., and Katz, J.L.:” Magnetic Water Treatment: The effect of Iron on Ccalcium Carbonate
Nucleation and Growth”. Langmuir 1989, 5, 861-867.
9. Johnson, M.L.:” Ferrous Carbonate Precipitation Kinetics- A Temperature Ramped Approach.” Thesis, Rice University, 1990.
10. Konigsberger, E, and Gamsjager, H., “ Solid-solute phase equilibria in aqueous solution III. A new application of old chemical
potentiometer” Marine Chemistry, Vol.30, pp.317-327, 1990.
11. Langerak, E.P. Beekmans, M.M. Beun, J.J. Hameler, H.V. and Lettinga, G.:” Influence of Phodphate and Iron on the Extent of
Calcium Carbonate Precipitation during anaerobic Digestion”. Chem Technol Bihechnol 74:1030-1036, 1999.
SPE 155127 15

12. Lippmann, F.,:” Phase diagram depicting aqueous solubility of binary mineral systems”. N.Jb. Miner. Abh. 139, 1-25, 1980.
13. Lippmann, F.,: “Stable and metastable solubility diagram for the system CaCO3-MgCO3-H2O at ordinary temperature”. Bull.
Mineral. 105, 273-279, 1982.
14. Nasr-El-Din, H.A. , Lynn, J.D., Hashem, M.K. , and Bitar, G.“Field Application of A Novel Emulsified Scale Inhibitor System to
Mitigate Calcium Carbonate Scale in A Low Temperature, Low Pressure Sandstone Reservoir in Saudi Arabia” SPE-77767, SPE
Annual Technical Conference and Exhibition held in San Antonio, Texas, 29 September–2 October 2002.
15. Pina, C.M., and Putnis, A., “The kinetics of nucleation of solid solutions from aqueous solutions: A new model for calculating
non-equilibrium distribution coefficients”, Geochimica et Cosmochimica Acta, Vol. 66, No. 2, pp. 185–192, 2002.
16. Pina, C.M., Putnis, A., and Astilleros, J.M., “The growth mechanisms of solid solutions crystallizing from aqueous solutions”,
Chemical Geology 204, page 145– 161, 2004.
17. Plummer, L.N. and Busenberg, E.:” Thermodynamics of aragonite-strontianite solid solutions: Results from stoichiometric
solubility at 25 and 76°C “. Geochimica el Cosmochimica Acta Vol51: 1393-1411, 1987.
18. Putnis, A., Fernandez-Diaz, L., and Prieto, M., “ Experimentally produced oscillatory zoning in the (Ba,Sr)SO4 solid solution”,
Nature, Vol.358, 1992.
19. Tomson, M.B., and Nancollas, G.H., :” Mineralization Kinetics: A Constant Composition Approach”.Science 2, June 1978: Vol.
200 no. 4345 pp. 1059-1060.
20. Zhang, J., and Nancollas, G.H., “Influence of calcium/sulfate molar ratio on the growth rate of calcium sulfate dehydrate at
constant supersaturation”, Journal of Crystal Growth, Vol118, Issues 3-4, P-287-294, 1992.