Quantitative Evaluation of Calcium

Sulfate Precipitation Kinetics in the

Presence and Absence of Scale Inhibitors
Chunfang Fan, Amy T. Kan, SPE, Gongmin Fu, and Mason B. Tomson, SPE, Rice University Brine Chemistry
Consortium; and Dong Shen, BJ Chemical Services

Summary have been discussed in several papers (Liu and Nancollas 1975; He
In this study, the nucleation kinetics of calcium sulfate scale forma- et al. 1994a; Lioliou et al. 2006). The authors are not aware of simi-
tion in the presence and absence of scale inhibitors has been studied. lar nucleation kinetic studies at high temperatures. The objective of
The nucleation kinetics of calcium sulfates in 0–3.2 M NaCl solution this paper is to present (1) an experimental study of calcium sulfate
was measured from 0 to 200°C at various supersaturation conditions nucleation kinetics in the absence and presence of scale inhibitors
[saturation index (SI) = 0–1.3]. The phase behavior of these various over a wide range of temperatures (0.9–200°C), (2) calcium sulfate
calcium sulfate phases was monitored by X-ray diffraction (XRD). The nucleation kinetics and inhibition in the presence of methanol and
inhibition study was performed by evaluating the inhibition efficiency of MEG, and (3) a quantitative model to predict the nucleation kinet-
calcium sulfate precipitation at the same supersaturation and tempera- ics and minimum inhibitor need for calcium sulfates.
ture ranges as those of the nucleation study. Several polyphosphonate,
polymeric, and environmentally friendly inhibitors were found to be Quantitative Nucleation-Kinetics Model. Knowledge of nucleation
effective inhibitors. The study has shown that calcium sulfate scales kinetics is needed to predict scale formation and the correct use of inhibi-
are very difficult to inhibit at SI > 1. Calcium sulfate scale is inhibited tors in scale control. Rates of nucleation and crystal growth in the absence
most effectively by hexamethylenediamine tetra(methylene phosphonic of specific inhibitors are normally related to the mineral’s supersaturation
acid) (HDTMP) if the saturation index is less than approximately 1.0. ( )( )
state {e.g., SI, SI ( gypsum ) = log10 ⎡⎣ Ca 2+ ⋅ SO 24− ⋅ aH2 2O K spgypsum ⎤⎦;
In addition, a semiquantitative model to predict precipitation kinetics
of calcium sulfate as a function of temperature, pH, SI, and HDTMP
( )( )
SI ( anhydrite ) = log10 ⎡⎣ Ca 2+ ⋅ SO 24− K spanhydrite ⎤⎦}. The objective of
this research was to develop a corresponding set of equations for
concentration is established with experimental results from this study.
gypsum and anhydrate. In this study, the SI is calculated by the
The effect of methanol and methylene glycol (MEG) on calcium sulfate
ScaleSoftPitzer software (Kan et al. 2005). Previously, He et al. (1996,
precipitation kinetics and inhibition is also presented.
1999) developed a semiempirical quantitative predictive model to
predict the time for scale to form in the presence and absence of a
Introduction
certain concentration of inhibitors and to predict the minimum-inhibi-
The formation of calcium sulfates as scale deposits is undesirable in tor-concentration (MIC) need of a given well condition. The measured
many industrial processes, such as geothermal applications, oil and induction time in the absence of inhibitor (tind 0
) can be described by the
gas production, water and waste-water treatment, membrane filtra- following function, using barite as the example:
tion, and desalination (Cowan and Weintritt 1976). Calcium sulfate
dihydrate (gypsum) is commonly precipitated at low temperatures,
such as in membrane filtration, and calcium sulfate (anhydrite) is often (0
log tind )
, barite = 1.83 −
12.1
SI ( barite )
−
885.8
T
+
5460.3
SI ( barite ) ⋅ T
,
precipitated at elevated temperatures and pressures, such as during
steam stimulation. It has been observed that gypsum and other min- . . . . . . . . . . . . . . . . . . . . . . . . (1)
eral-salt solubilities can be affected severely in the presence of hydrate
where T is given in K. A similar equation has been developed for
inhibitors (e.g., methanol and glycols) (Kan et al. 2002, 2003). The inh
calcite. The influence of inhibitors on the nucleation time, tind , can be
formation of gypsum scale has been observed in a methanol-treated
related semiempirically in Eqs. 2 and 3 by assuming a separation of
offshore well located in the Gulf of Mexico. During steam stimula-
the effect of the inhibitor from that of the uninhibited mineral:
tion of a heavy-crude reservoir containing anhydrite, the formation
brine is heated to a very high temperature. Anhydrite scale becomes
problematic when the hot brine from steam stimulation is mixed with (inh
log tind )
, seconds = log10 tind
0
( )
, seconds + binh ⋅Cinh, . . . . . . . . . (2)
connate brine from a different zone at a lower temperature.
where Cinh is the concentration of inhibitor added, in mg/L, and binh
The nucleation induction period for calcium sulfate dihydrate
is an inhibitor-effectiveness term given in L/mg and is a function
nucleation has been measured in NaNO3 solutions (less than 0.48
of supersaturation (SI), temperature T ( in K), pH, and lattice-ion
M) at 25°C (Packter 1974) and in 0.5 M NaCl solutions at 25
molar ratio (R = [Ba2+]/[SO42−]). For barite and HDTMP as the scale
and 90°C (Liu and Nancollas 1973). These measurements were
inhibitor, binh can be expressed as
made over a wide range of calcium sulfate concentrations (up to
0.24 M), but the background electrolytes (nonlattice ions) either HDTMP
log binh = −0.006 − 1.43 ⋅ SI ( barite )
were fixed at a single concentration (Liu and Nancollas 1973) or
were varied as a function of the concentrations of the lattice ions + 968.6 / T + 0.102 ⋅ pH + 0.137 ⋅ log R. . . . . . . . (3)
(Packter 1974). He et al. (1994c, 1994d) have studied the primary The parameters in Eq. 3 are for HDTMP inhibition of barite.
nucleation kinetics and seeded growth of calcium sulfate dihydrate Similar equations for other scale-inhibitor interactions with barite
and barium sulfate in NaCl solutions to 6 M and 90°C. Limited and calcite are also available (He et al. 1997). For a given protec-
inhibition studies of gypsum in the presence of scale inhibitors tion time (tprotection) necessary to protect a well from scaling, Eq. 4,
derived from Eq. 2, can be used to predict the MIC (in mg/L) for
a given well and conditions:
Copyright © 2010 Society of Petroleum Engineers

This paper (SPE 121563) was accepted for presentation at the SPE International Symposium MIC =
( 0
log10 tprotection tind ), . . . . . . . . . . . . . . . . . . . . . . . . . . . (4)
on Oilfield Chemistry, The Woodlands, Texas, USA, 20–22 April 2009, and revised for binh
publication. Original manuscript received for review 23 February 2009. Revised manuscript
received for review 30 December 2009. Paper peer approved 19 January 2010. where binh is given in L/mg.

December 2010 SPE Journal 983

 TABLE 1—SCALE INHIBITORS USED IN THIS STUDY
Abbreviated Name Active Ingredient

HDTMP Hexamethylenediamine tetra(methylene phosphonic acid)

BHPMP Bishexamethylenediamine penta(methylene phosphonic acid)
DTPMP Diethylenetriamine penta(methylene phoshpohnic acid)
NTMP Nitrilo tris(methylenephosphonic acid)
PPCA3800 Phosphinopolycarboxylate MW 3,800
PPCA1900 Phosphinopolycarboxylate MW 1,900
MAT 1000 Maleic acid terpolymer MW 1,000
MAC 2000 Maleic acid copolymer MW 2,000
SCC1500 Sulfonate carboxylate copolymer MW 1,500
PVS15000 Polyvinylsulfonate MW 15,000
CMI 15 Carboxymethyl inulin, degree of substitution = 1.5%
CMI 25 Carboxymethyl inulin, degree of substitution = 2.5%

TABLE 2—RANGES OF TESTING PARAMETERS

FOR CALCIUM SULFATE NUCLEATION-KINETICS STUDY
2+ 2−
Ca (M) SO4 (M) Temperature (°C) NaCl (M) SI Inhibitor Concentration (mg/L)

Low-Temperature Calcium Sulfate Nucleation-Kinetics Experiments

0.012–0.4 0.012–0.4 0.9–90 0.0–3.2 0.2–1.3 (gypsum) 0–5

High-Temperature Calcium Sulfate Nucleation-Kinetics Experiments

0.0–0.076 0.0–0.076 100–200 0.1–1.0 −0.09–1.3 (anhydrite) 0–10

Materials and Methods stock solutions were added to the anionic solutions to prepare inhibi-
Materials. All chemicals used in this study were reagent grade tor concentrations from 0.2 to 10 mg/L.
except for the scale inhibitors, which were technical grade. Solutions
containing the lattice cations and anions were prepared with 0–3.2 Calcium Sulfate Nucleation and Inhibition at 0.9–90°C. The
M NaCl as a background electrolyte. For the nucleation-inhibition experimental protocol used in this study is similar to the procedure
experiments, buffering solutions with either 1 mM MES at 6.0 pH published by He et al. (1994a; 1994b). The range of experimental
[2-(N-Morpholino)ethanesulfonic acid, pKa = 6.21 at 25°C] or 5 mM parameters tested is listed in Table 2. The cationic and anionic
PIPES at pH 6.7 (1,4 piperazinediethanesulfonic acid, pKa = 7.1 at solutions were maintained at the desired temperature before mix-
25°C) were also added to the brine solution. For some experiments, ing. Equal volumes of cationic and anionic solutions were mixed in
the cationic and anionic solutions were adjusted to pH 3.7–8.7 with a vial in a jacketed glass reactor and kept at a constant temperature
NaOH or HCl to test the efficiency of scale inhibition at various solu- by circulating water from a constant-temperature water batch.
tion pH values. For most experiments, equal molar concentrations The solution was stirred by a Teflon-coated magnetic stirring bar.
of calcium and sulfate were used. Lattice cation/anion ratios of 2, 5, Either the solution turbidity or calcium (Ca) concentration was
and 10 were also tested. Commercial-grade scale inhibitors used in monitored visually using a laser light (wavelength = 532 nm) or
0 inh
this study are listed in Table 1. Scale-inhibitor stock solutions were by a Ca electrode. The induction period (tind or tind ) is defined as
prepared and neutralized to pH 6. Small volumes of scale-inhibitor the time between the generation of a supersaturated state by mixing
the two solutions and the first observed change in Ca concentration
or turbidity monitored by a drop in the Ca-electrode signal (by
0.5 mV) or by visually observable laser-light scattering. A drop
SI = 0.81 SI = 0.72 of 0.5 mV in the Ca-electrode signal is equivalent to a change of
4% in Ca concentration (see Fig. 1). The drawback of using a Ca
SI = 0.59 SI = 0.42 electrode is that (1) its response is temperature sensitive and (2) it
can be used only at less than 40°C.
72 Calcium Sulfate Nucleation and Inhibition in the Presence of
70 3 M NaCl, 0.2 M SO4, 25°C
Response (mV)

Hydrate Inhibitors. The effect of methanol and MEG on calcium

Ca Electrode

68
66 sulfate nucleation kinetics and inhibition has been tested at 25°C
64 using the same batch reactor as discussed earlier. The supersatu-
62 rated calcium sulfate brine was prepared by mixing the cationic
60 solution containing 140 mM Ca, 1 M NaCl (pH 6.7), 140 mM
58 sulfate, 1 M NaCl (pH 6.7), and different volume fractions of
56
54
methanol and MEG. For some experiments, 0.5–1 mg/L HDTMP
0 2,000 4,000 was also added to the mixture as a scale inhibitor. Methanol and
Time, seconds MEG increase gypsum SI, which can be calculated using the
ScaleSoftPitzer software. Because methanol increases gypsum
Fig. 1—Ca electrode responses of four low-temperature cal- supersaturation substantially and induces rapid nucleation, a sec-
cium sulfate nucleation-kinetics studies at four gypsum SI ond set of experiments was run at lower supersaturation with a
values, 0.42, 0.59, 0.72, and 0.81. brine prepared by mixing a cationic solution of 200 mM Ca and 1

984 December 2010 SPE Journal

 Pressure anionic solutions contain 1 M NaCl, 1 mM MES (pH 6.02), and
various concentrations of Ca2+ and SO42−. A third inhibitor stock
Pump A
solution was prepared, containing the same composition as the
anionic solution and 120 mg/L inhibitor. The flow rate of the third
Pump B
pump was set at 0.017, 0.083, and 0.17 mL/min so that the final
Pump C Backpressure inhibitor concentrations would be 1.0, 5.0, and 10 mg/L. To
Oil bath Ca
Ca2+ prevent inhibitor degradation at high temperatures, all solutions
were degassed by vacuum for 10 minutes and purged with argon
SO42+
before use. The experiments were run by pumping the anionic and
cationic solutions at a combined flow rate of 2 mL/min through
Cooling bath the flow loop (i.e., a residence time of 64.5 seconds in the reac-
tion coil). Before each experiment, the reaction coil was heated
to 100°C as a solution containing 10 mg/L inhibitor was pumped.
Fig. 2—Apparatus for the high-temperature calcium sulfate Each solution was pumped for 10 minutes into the reaction coil to
nucleation-kinetics study. All surfaces contacting the fluids monitor whether scale formed. If scale did not form, the inhibitor
are made out of either Hastelloy C or PEEK material to avoid concentration was reduced to 5 mg/L, the scaling was monitored
corrosion at high temperature. for another 10 minutes, and the process was repeated at 1-mg/L
inhibitor concentration. If no scale was formed at this temperature,
the temperature was increased by 25°C, and the inhibitor evalu-
ation was repeated from 10 to 1 mg/L. These experiments were
M NaCl (pH 6.7) with an anionic solution of 20 mM sulfate 1 M repeated with a 25°C increment up to 200°C. Similar experiments
NaCl (pH 6.7), and different volume fractions of methanol. were repeated with 0.01 to 0.05 M Ca2+ and SO42− concentrations.
Anhydrite nucleation inhibitions were monitored at 0 to 10 mg/L
Calcium Sulfate Nucleation and Inhibition Experiments at HDTMP concentrations.
100–200°C. The apparatus consists of two high-pressure syringe
pumps and a flow loop with automatic data recording for dif- XRD Analysis. Four samples of low-temperature calcium sulfate
ferential pressure and Ca-electrode responses (Fig. 2). An oil precipitates were prepared by mixing an equal volume of 0.2 M
bath is used to heat the flow loop. The flow loop consists of two CaCl2·2H2O and 0.2 M Na2SO4 solutions in deionized water at 4,
165-cm-long coils of 0.107-cm-inner-diameter (ID) Hastelloy C 25, 40, 70, and 90°C in the same batch reactor described earlier.
tubings, with approximately 90 cm of the tubing submerged in oil Solid precipitates were filtered and dried at room temperature.
for heating the Ca2+ and SO42− solutions, respectively. The two solu- One high-ionic-strength precipitate was prepared by mixing equal
tions are injected into the flow loop by two high-pressure syringe volumes of CaCl2·2H2O (1 M) and Na2SO4 (1 M) in 3.2 M NaCl
pumps with a flow-rate accuracy of ±0.5%. The heated solutions solutions at 70°C in a batch reactor. The precipitate was separated
are combined through a T into a 240- or 426.7-cm-long coil, 0.107- from supernatant by centrifugation at 1,000 rev/min for 10 min-
cm ID Hastelloy C tubing. A third syringe pump containing the utes and resuspended in deionized water to remove excess salt.
inhibitor to be tested is introduced into the anion stream through This procedure was repeated six times to remove NaCl from the
the T. Mixing takes place at the same temperature and pressure. precipitate. The precipitate was then dried at 60°C for 10 hours.
All high-temperature surfaces (e.g., of tubing, valves, mixers) One high-temperature precipitate was synthesized in the flow loop
are constructed of Hastelloy C for corrosion resistance at 200°C where the reaction coil was replaced with a 1/8-in.-ID, 110-cm-
and ease of cleaning. After the reaction coil, the fluid is cooled long stainless-steel coil at 200°C using a solution of Na/Ca/Mg/K/
to room temperature by means of a 127-cm-long, 0.05-cm-ID Cl/SO4 matrix at a Ca/SO4 ratio of 2 and anhydrite SI of 1.2. The
polyether ether ketone (PEEK) tubing and a backpressure valve solid was formed at 10 to 25 cm from the entrance to the flow loop,
at 250 psia, a calcium-ion-selective electrode, and a fraction col- indicating rapid precipitation. The precipitate was separated from
lector. A differential-pressure transducer is connected to the front the column at the end of precipitation experiment and analyzed by
and to the end of the flow loop to monitor the differential pres- XRD. Powder XRD analysis was conducted with a Rigaku D/Max
sure of the reaction coil. Data acquisition is accomplished with a Ultima II configured with a vertical theta/theta goniometer with a
digital multimeter connected to a computer for data logging. The Cu Kα source at 50 kV and 40 mA.
reaction-coil volume is 2.15 mL for the 240-cm tubing and 3.84
mL for the 426.7-cm tubing. The heat-transfer rate in this tubing Results and Discussion
has been measured to be 25°C/s at the flow rates needed, which is Calcium Sulfate Nucleation Kinetics in the Absence of Scale
equivalent to approximately 7 seconds for the solution to be heated Inhibitors at Low Temperatures (0–90°C). Previously, He et al.
from 25 to 200°C. Effluent samples are occasionally analyzed by (1994c) conducted an extensive study of gypsum nucleation kinet-
inductively coupled plasma (ICP) for Ca, sulfur, and phosphorus ics between 25 and 90°C and 0 and 6 M ionic strength. The
concentrations. current study extended that experimental range to include lower
In the nucleation experiments, a cationic solution and an temperature and ionic-strength data. On the basis of the classical
anionic solution are prepared at 1 M NaCl and various Ca2+ and homogeneous nucleation theory, the induction time is related to the
SO42− concentrations so that the mixed solution is supersaturated inverse of SI to the second power and the slope of the correlation
with anhydrite (0–1.3 SI) at 100–200°C. The solutions are pumped in the region of homogeneous nucleation is proportional to the
at 0.1- to 3-mL/min flow rates to determine when a precipitate interfacial surface tension by
forms. Differential pressure is a good confirmation for scaling,
but it is not sensitive enough to show when precipitation starts. B
Change in Ca concentration, either with ICP analysis or by Ca
0
log tind = − C , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
SI 2
electrode, appears to be a better indicator for the onset of scale
formation. A series of experiments was performed to establish the where C is a constant and B is the slope and is a function of tem-
relationship of precipitation kinetics as a function of temperature perature and interfacial surface tension:
and SI, and the range of temperature and supersaturation states
are listed in Table 2.
 3Vm2 N A f
Systematic inhibitor-evaluation studies were performed with B= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (6)
HDTMP plus four other inhibitors. The inhibitor tests were ( 2.3RT )3
performed by evaluating the inhibition efficiency of anhydrite
precipitation at supersaturation of 0–2.0 SI, 100–200°C, and 1-, where R is the gas constant, T is the absolute temperature, NA is
5-, and 10-mg/L active inhibitor concentrations. The cationic and Avogadro’s number, Vm is the molar volume of the crystal (74.7

December 2010 SPE Journal 985

 4 5
3.5

log t0, seconds

log t0, seconds

3 4
2.5 Slope = 0.76 3
2
1.5 2 Slope = 0.77
4°C
1 1
0.5
0
0 0 5 10 15 20
0 5 10
1/SI2 1/SI2
(a) (b)

5
5

log t0, seconds

4
log t0, seconds

4
3
3
2
2
1 Slope = 0.70 1 Slope = 0.56
25°C 40°C
0 0
0 5 10 15 0 5 10
1/SI2 1/SI2
(c) (d)

Fig. 3—Logarithm induction periods, log(t 0ind), as a function of the inverse of SI to the second power for calcium sulfate nucleation
at (a) 4°C, (b) 15°C, (c) 25°C, and (d) 40°C at 0.6–4.0 M ionic strength. Also shown are the regions dominated by homogeneous
and heterogeneous nucleation. The points represent experimental data from this study. Note that the first slope in each figure
has been labeled and used to calculate the interfacial surface tensions.

cm3/mol for gypsum), f is the correction factor for heterogene- temperature on the induction period and to model conditions that
ous nucleation [f ≈ 0.01 (Söhnel and Mullin 1988)], and  is a might occur in undersea pipelines.
geometric shape factor [ = 50 for rectangular parallelepipeds An empirical relationship was suggested for the effect of tem-
0
(Nielsen 1964)]. Plots of log tind ( )
vs. 1/SI2 for data from 4 to 40°C perature on the induction period, which is similar to the Arrhenius
are shown in Fig. 3. Similar to the data of He et al. (1994c) and equation for the temperature dependence of the rate constant (He
Söhnel and Mullin (1988), the data can be differentiated into two et al. 1994c):
regions. Homogeneous nucleation is dominant at high supersatura-
tions, and heterogeneous nucleation is dominant in the low-super- 0
log 1 tind ( )
= a − E 2.3 RT , . . . . . . . . . . . . . . . . . . . . . . . . . . . (7)
saturation range. The interfacial surface tensions estimated from
the slope of Fig. 3 are 40.6, 42.6, 42.6, and 41.6 mJ/m2 at 4, 15, where E is the activation energy for nucleation and a is a constant.
25, and 40°C, respectively. These values are within experimental 0
In Fig. 4, log 1 tind ( )
-vs.-1/T data from He et al. (1994c) (25– 90°C,
error for gypsum when comparing with those reported by He 3 M NaCl) and from this study (4–25ºC, 1 M NaCl) are plotted.
et al. (1994c) (39.3–42.6 mJ/m2 at 25–50°C). Note that the data in Both sets of data are linearly related, and the slope corresponds to
Fig. 3 represent nucleation-kinetics data of varying ionic strength an activation energy of 48.2 kJ/mol. These results indicate that the
(0.6–4 M ionic strength), whereas the data in He et al. (1994c) two sets of data are internally consistent. These data are used to
were measured at constant ionic strength. This work extended the develop a quantitative model, which will be discussed later.
temperature range of those data to 0.9°C to evaluate the effect of
XRD Analysis of Calcium Sulfate Precipitates Prepared in 0
and 3.2 M NaCl Solutions at 4–90°C. The precipitate formed
between 4 and 90°C in pure water has been identified as gypsum,
SI=0.55, this study SI=0.55, He et al. according to XRD analyses (Fig. 5). This is consistent with obser-
vations of He et al. (1994c). However, the precipitate prepared at
0 70°C in 3.2 M NaCl was identified as hemihydrate, indicating that
the nucleation studies mentioned earlier may represent the nucle-
–0.5 ation kinetics of a mixed-solid phase of gypsum and hemihydrate
log(1/t0), seconds

–1 (Fig. 5). Adding salt to the solution reduces the water activity and
favors dehydration. This is consistent with previous studies show-
–1.5 ing that gypsum can undergo phase transition under conditions that
–2 favor dehydration, such as high temperature and total dissolved
solids (Hardie 1967).
–2.5
–3 Calcium Sulfate Nucleation Kinetics in the Absence of Scale
–3.5 Inhibitors at High Temperatures (100–200°C). The typical
0.0025 0.003 0.0035 results of the nucleation at 178°C where the solutions contained 1
1/T, K
M NaCl, 36 mM Ca2+, 1 M NaCl, and 36 mM SO42−, respectively,
are shown in Fig. 6a. The SI with respect to anhydrite is 0.78 at
Fig. 4—The temperature dependence of the inverse of the 178°C. The solutions were pumped into the flow loop at 3 mL/min
induction period, log(t 0ind), for gypsum at a constant SI of 0.55 each for 60 minutes (total flow of 6 mL/min), and no precipitation
± 0.05. The brines contain 1 M NaCl (this study) and 3 M NaCl was observed. The pump flow rate was reduced to 2 mL/min (total
(He et al. 1994c). flow = 4 mL/min), and precipitation was observed as shown in the

986 December 2010 SPE Journal

 Gympsum Ppt at 4°C Gympsum Ppt at 70°C
4,000 4,500
4,000
3,000 3,500

Intensity

Intensity
3,000
2,000 2,500
2,000
1,000 1,500
1,000
0 500
0
0 50 0 20 40 60 80
2θ 2θ
(a) (b)

Gympsum Test at 90°C Hemihydrate Ppt 3MNaCl, 70°C

8,000 3,000
2,500
6,000
Intensity

Intensity
2,000
4,000 1,500
1,000
2,000 500
0 0
0 20 40 60 80 0 20 40 60 80
2θ 2θ
(c) (d)

Fig. 5—Plots of XRD spectrum of calcium sulfate precipitates prepared in (a) deionized water at 4°C, (b) deionized water at 70°C,
(c) deionized water at 90°C, and (d) 3 M NaCl solution at 70°C. The spectra in Figs. 5a through 5c are identified as gypsum, and
the spectrum in Fig. 5d is identified as hemihydrate. The thin lines are the corresponding spectra of gypsum and hemihydrate
from the spectrum library.

reduction of Ca2+ and SO42− concentrations. The flow loop eventu- 1 M NaCl, and 40 mM SO42−, respectively, is shown in Fig. 7. The
ally plugged with scale, as indicated by pressure buildup (Fig. 6b). SI with respect to anhydrite is 0.49 at 148°C. The solution was
At 4- and 6-mL/min flow rates, the travel time is approximately 32 pumped into the flow loop at a total flow rate of 2 mL/min for 111
and 21 seconds in the reaction coil, respectively. A similar study minutes and at 0.2 mL/min for 280 minutes; no precipitation and
at 148°C where the solutions contained 1 M NaCl, 40 mM Ca2+, no pressure buildup were observed. The solution travels through

Ca (mg/L) SO4 (mg/L) CaSO4 ppt (mg)

6 mL/min 4 mL/min 6 mL/min 4 mL/min

2,000 16 120
[Ca2+][CaSO4], mg/L

14
Pressure, psi

100
CaSO4 ppt, mg

1,500 12 80
10 60
1,000 8 40
6 20
500 4 0
2 0 50 100 150
0 0
Run Time, minutes
0 200 400 600
Volume, mL
(a) (b)

Fig. 6—Plots of (a) effluent Ca2+ and SO42− concentrations and corresponding accumulative anhydrite mass precipitated after a
supersaturated brine (anhydrite SI = 0.78 at 178°C) flowing through the high-temperature apparatus; (b) pressure vs. run time for
experiment in Fig. 6a. In these experiments, the initial Ca2+ and SO42− concentrations are 696 and 1734 mg/L, respectively.

Ca (mg/L) SO4 (mg/L)

2 mL/min 0.2 mL/min
2,500
2,000
mg/L

1,500
[Ca2+],

1,000
500
0
0 50 100 150 200 250 300
Volume, mL

Fig. 7—Effluent Ca2+ and SO42− concentrations after a supersaturated brine (anhydrite SI = 0.49 at 148°C) flowing through the high-
temperature apparatus. In these experiments, the initial Ca2+ and SO42− concentrations are 787 and 2017 mg/L, respectively.

December 2010 SPE Journal 987

 3
This study Field scale sample
2.5
log t, seconds
600
2
500
1.5

Intensity
400
1 Slope = 0.79 300
0.5
200
0
100
1 2 3 4 5 6 7
0
1/Sl2
0 50 100
Fig. 8—Plot of logarithm of anhydrite precipitation time (sec- 2θ
onds) vs. the inverse of anhydrite SI values squared at 178°C.
Fig. 9—XRD spectrum of the laboratory-prepared calcium sul-
fate precipitates and a field scale sample. The XRD spectrum
of both samples has been identified as anhydrite.
the reaction coil in 64 and 645 seconds at 2- and 0.2-mL/min flow
rates, respectively. Similar systematic experiments have been con-
ducted at 118–200°C and 0.35–0.80 SI in 1 M NaCl solution. The induction time of a more-supersaturated calcium sulfate solution
logarithm of the measured precipitation time (in seconds) is also (gypsum SI = 1.06 at 40°C) vs. HDTMP concentration is plotted.
inversely related to the SI to the second power at 178°C (Fig. 8). As shown in Fig. 11c, much higher HDTMP concentration is
The slope of the homogeneous nucleation corresponds to an inter- needed to inhibit this brine at gypsum SI = 1.06. He et al. (1996)
facial surface tension of 92.9 mJ/mol at 178°C [molar volume of have shown that phosphonate inhibitors can inhibit barite precipita-
anhydrite = 45.9 cm3/mol (Azam 2007)], which is almost twice as tion at a supersaturation of SI = 2.5–3 at moderate concentrations.
high as that for gypsum and is similar to that for barite (71–140 Calcium sulfates are considerably more soluble than barite, and,
mJ/mol) reported in the literature. therefore, it is not surprising that conventional scale inhibitors do
not work as well for calcium sulfate scales, when compared at
XRD Analysis of Calcium Sulfate Precipitates Prepared at the same SI values. Fig. 11c also suggests that even with a higher
200°C and 1 M NaCl Solution. The precipitates prepared in this HDTMP concentration, the increase of inhibition time is shorter
study and a field scale sample from an oil well that has been stimu- compared with the one at lower SI values. These evidently indicate
lated by high-pressure steam are plotted in Fig. 9. Both samples that, although HDTMP is the most effective inhibitor to inhibit
have similar XRD spectra and are identified as anhydrite. calcium sulfate scales, a better inhibitor is needed, especially for
brines with very high SI values and at high temperatures.
Calcium Sulfate Inhibition by Scale Inhibitors at 25–200°C. Fig. 12 shows typical experimental results of a calcium sulfate
The induction time of calcium sulfate precipitation at 25–70°C at inhibition at high temperatures. The solution is similar to that
a constant supersaturation state (gypsum SI = 0.63) and in the pres- used in Fig. 7, except that the solution also contained 1–10 mg/L
ence of 1 mg/L of 13 different scale inhibitors is plotted in Fig. 10a. HDTMP. The anhydrite saturation indices varied from −0.09 to
The two horizontal lines are the markers for 30 minutes and 2 1.00 at 100–200°C, in 25°C increments. The effluent Ca2+, SO42−,
hours. Only a few scale inhibitors [HDTMP, BHPMP, PPCA3800, and phosphonate concentrations vs. solution volume are plotted
MAC 2000, and CMI 25 (see Table 1)] are sufficiently effective in Fig. 12. Anhydrite-scale formation was completely inhibited by
to inhibit calcium sulfate precipitate at 25°C for more than 30 HDTMP up to 175°C (SI = 0.79), whereas precipitate was observed
minutes. At 40°C, only HDTMP is effective at inhibiting calcium at similar conditions in the absence of HDTMP (see Fig. 7).
sulfate precipitation for more than 30 minutes; and, at 70°C, no Similarly, HDTMP was able to inhibit anhydrite precipitation
inhibitor at 1-mg/L concentration is effective at inhibiting calcium at 200°C (SI = 1.00) when the injection inhibitor concentration
sulfate precipitation from this supersaturated solution (gypsum was 5 or 10 mg/L. However, 1 mg/L HDTMP was not able to
SI = 0.63) for more than 30 minutes. With a less-supersaturated inhibit anhydrite precipitation at 200°C (SI = 1.00). Note that the
solution (gypsum SI = 0.43 at 70°C), most scale inhibitors are measured phosphonate concentrations are similar to the injection
effective at 1- to 5-mg/L concentrations (Fig. 10b). Similarly, the concentrations at < 150°C. At temperatures greater than 150°C, the
inhibition of calcium sulfate precipitation with HDTMP, BHPMP, effluent phosphonate concentrations were significantly less than
PPCA3800, and MAC2000 at 175 and 200°C is compared in the injection concentrations of 5 and 10 mg/L, indicating that a por-
Table 3. The solution contains 20 mM calcium and sulfate in 1 M tion of phosphonate precipitated with Ca2+ in solution at such high
NaCl solution and is supersaturated with respect to anhydrite (0.75 temperature to form Ca-phosphonate pseudoscale. In a separate
SI at 175°C and 1 SI at 200°C). It was found that all four inhibitors experiment, the thermal stability of HDTMP is evaluated. Free-
were able to inhibit anhydrite precipitation at 0.75 SI and 175°C, phosphate concentration was measured to determine if HDTMP
but only the two pentaphosphonates (HDTMP and BHPMP) were degraded at high temperatures. Approximately 8% phosphorus in
able to inhibit anhydrite precipitation at 1 SI and 200°C. HDTMP stock solution is free phosphate. Approximately 14.7 and
These results indicate that HDTMP is the most effective inhibi- 15% of the total phosphorus in effluent solutions of the 178 and
tor to inhibit calcium sulfate scales, and this is consistent with the 200°C experiments, respectively, is free phosphate, indicating that
observations of Liu and Nancollas (1975) and He et al. (1994a). only a small portion (approximately 7%) of HDTMP is degraded at
Additional studies were conducted with HDTMP. Fig. 11 shows high temperatures, under the conditions of these experiments.
the inhibition of calcium sulfate precipitation at low temperatures The temperature vs. Ca concentrations that represent the range
as a function of HDTMP concentration, SI, and pH. In Fig. 11a, a of conditions tested in this study are plotted in Fig. 13. The black
( HDTMP
linear relationship of the log tind )
vs. HDTMP concentrations squares and lines are the experimental data points and the anhydrite
SI isopleths, respectively. The anhydrite SI is labeled on the graph,
is observed. These data support the functional form of Eq. 2 (i.e.,
the logarithm of induction time is linearly related to the inhibitor adjacent to the black lines. A total of 35 data points were evalu-
concentration, and the slope is binh). The data in Fig. 11b show the ated, representing the anhydrite SI range of −0.61 to 2. Anhydrite
relationship of inhibition efficiency with pH. The inhibitor is more scale did not form at 0.48 ± 0.03 SI, or lower. Between 0.48 ±
effective at higher pH. This is because the inhibition efficiency 0.03 and 0.73 ± 0.03 SI, anhydrite scale can be controlled with 1
is related to the solution chemistry of inhibitors, and similar pH mg/L HDTMP. Between 0.73 ± 0.03 and 0.99 ± 0.03 SI, anhydrite
dependence has also been observed for barite and calcite inhibition scale can be controlled with 5 mg/L HDTMP. At SI values greater
(He et al. 1996, 1999; Tomson et al. 1994, 2003). In Fig. 11c, the than approximately 1, the anhydrite scale cannot be controlled even

988 December 2010 SPE Journal

 0.63 Sl, 1 ppm inhibitor 25°C 40°C 70°C

100,000

Induction Time, seconds

10,000

1,000

100

10

(a)

5 ppm 1ppm
70°C, 1.4M l, 0.43 SI

100,000
Induction Time, seconds

10,000

1,000

100

10

(b)

Fig. 10—Comparison of induction times of calcium sulfate precipitation in the presence of different scale inhibitors. (a) Induction
times of calcium sulfate precipitation at constant supersaturation state of gypsum SI = 0.63 and three different temperatures,
25, 40, and 70°C. Three different brines containing 1 M NaCl, 1 mg/L of different scale inhibitors, and varied calcium and sulfate
concentrations that correspond to gypsum SI = 0.63 at 24, 40, and 70°C were used in this study. (b) Induction time of calcium
sulfate precipitation at 70°C in a supersaturated brine of gypsum SI = 0.43 and inhibitor concentrations of 1 and 5 mg/L. The
horizontal lines are 30-minute and 2-hour time lines.

with 10 mg/L HDTMP, possibly because of inhibitor precipitation. developed an equation for nucleation time as a function of tem-
At high temperatures, anhydrite scale may be difficult to inhibit. perature and SI by using the functional form previously defined for
Currently, we have observed that anhydrite scale can be inhibited a barite-nucleation-kinetics model (He et al. 1996). The results are
by HDTMP if the SI is below approximately 1.0. Anhydrite scale listed in Table 4. The low-temperature data were fitted using a non-
is difficult to inhibit if the SI is greater than approximately 1.0, linear least-squares method to yield Eq. 8 by SigmaPlot. Because
partly because phosphonate inhibitors may not be soluble enough the temperature dependence of anhydrite nucleation is not well
in the presence of divalent metal ions at high temperatures. established, at this point, the third term in Eq. 9 is adopted from
barite nucleation. This may be justifiable because the temperature
Modeling Calcium Sulfate Nucleation Kinetics in the Presence dependence of nucleation kinetics may be similar between anhy-
and Absence of Scale Inhibitor. From the preceding data, we drite and barite with similar interfacial tension. Similarly, from

TABLE 3—INHIBITION OF CALCIUM SULFATE PRECIPITATION

BY FIVE DIFFERENT INHIBITORS AT 175 AND 200°C
Temperature (°C) 175 175 175 200 200 200
SI (anhydrite) 0.75 0.75 0.75 1.00 1.00 1.00
Inhibitor (mg/L) 10 5 1 10 5 1
HDTMP No ppt No ppt No ppt No ppt No ppt ppt
BHPMP No ppt No ppt No ppt No ppt No ppt ppt
PPCA3800 No ppt No ppt No ppt ppt ppt ppt
MAC2000 No ppt No ppt No ppt ppt ppt ppt
SCC1500 No ppt No ppt No ppt ppt ppt ppt

December 2010 SPE Journal 989

 70°C, 0.43 Sl
70°C, pH 6.7
pH 8.7 pH 6.7 pH 5.7
SI=0.43 Sl=0.63 Sl=0.83
5.0 pH 4.7 pH 3.75
5.0
4.0 4.5
log tinh, seconds

log tinh, seconds

4.0
3.0 3.5
3.0
2.0 2.5
2.0
1.0
1.5
0.0 1.0
0 0.5 1 1.5 0 0.2 0.4 0.6 0.8
[HDTMP], mg/L
(a) (b) [HDTMP], mg/L

40°C, 1.06 Sl, pH 6.7

4.5
log tinh, seconds

4.0
3.5
3.0
2.5
2.0
1.5
1.0 pH = 6.7
0.5
0.0
0 5 10 15
[HDTMP], mg/L
(c)
inh
Fig. 11—Plots of log(t ind ) vs. the HDTMP concentrations at (a) three different supersaturation states for gypsum SI = 0.43, 0.63,
and 0.83, each at 70°C (the solutions were buffered at pH 6.7); (b) a constant supersaturation state of SI = 0.43, at 70°C and with
five different pHs; and (c) constant supersaturation, SI = 1.06, 40°C. The solution of Plots a and b contain 1 M NaCl. The solution
of Plot c contains 0.1 m CaCl2·2H2O and 0.2 m Na2SO4 without additional NaCl.

the HDTMP inhibition data, we derived the equation for binh using value should be very similar for HDTMP/anhydrite and HDTMP/
the functional form previously defined for HDTMP in the barite barite systems. Certainly, more data are needed to fine tune this cal-
nucleation-inhibition model (He et al. 1996) for both gypsum and cium sulfate nucleation-kinetics model. The model in Table 4 can
anhydrite. To date, there are not enough data to warrant a complete serve as a first-order estimate of the precipitation rate and inhibitor
curve fit of b terms with the high-temperature data. The T, pH, need for a given system. As will be discussed in the following, this
and R terms in Eq. 11 are adopted from that for HDTMP/barite. model provides a reasonable estimate of gypsum precipitation and
Because these three terms are related to HDTMP properties, the inhibition in the presence of low concentrations of methanol and

Ca (mg/L) SO4 (mg/L) HDTMP (mg/L)

100°C 120°C 150°C 175°C 200°C

2,500 10
[Ca ] and [SO4 ], mg/L

2+
[Ca ]
8
[HDTMP], mg/L

2,000
2–

1,500 6
2-
1,000 [SO4 ] 4
2+

500 2

0 0
0 50 100 150 200 250 300 350
Volume, mL

Fig. 12—Plot of calcium sulfate and HDTMP concentrations vs. the volume of the calcium sulfate supersaturated solution used in
the high-temperature nucleation-test apparatus. The calcium and sulfate concentrations of the feed solution are shown as dotted
horizontal lines ([Ca] = 775 mg/L; [SO4] = 1953 mg/L). The anhydrite saturation indices were −0.09, 0.21, 0.50, 0.79, and 1.00 at
100, 125, 150, 175, and 200°C, respectively. The experiment was run continuously at five temperatures (separated by the dashed
lines) from 100 to 200°C, at 25°C increments. Note that at each temperature condition, initial HDTMP concentrations of 1.0, 5.0,
and 10 mg/L were tested for 10 minutes. The HDTMP concentrations (dotted line) are the measured HDTMP concentration in the
effluents. HDTMP concentrations in the effluents are significantly lower than that in the feed solution at 175 and 200°C.

990 December 2010 SPE Journal

 Constant Anhydrite Sl Isopleths

Temperature, °C

[Ca2+], mg/L

Fig. 13—High-temperature calcium sulfate scaling and inhibition isopleths, where the lines connecting the squares are constant
SI isopleths. Experiments have been tested at the condition represented by the squares. In each experiment, mCa2+ = mSO24− . The
numbers at the top of the isopleths represent the SI values of each line connecting the squares. Four regions have been identified
on the basis of the SI isopleths: (1) no calcium sulfate scale forms; (2) 1 ppm HDTMP is needed; (3) 5 ppm HDTMP is needed;
and (4) scale cannot be controlled at 10-ppm-HDTMP level. The residence time in the flow loop is 1–10 minutes.

MEG. Also, it has been compared to the anhydrite-scale-inhibition (MIC is calculated from Eq. 4 and Eqs. 8 through 11, assuming a
problem of an actual steam-stimulation project, and the inhibitor protection time of 2 hours).
need predicted from these equations is quite reasonable compared
with that observed in the field. Calcium Sulfate Nucleation Kinetics and Scale Inhibition in the
In Fig. 14, the calculated induction time (based on Eq. 8) vs. Presence and Absence of Hydrate Inhibitors and Scale Inhibi-
temperature at SI values of 0.2 to 1.0 is plotted. The horizontal tors. Fig. 16 shows the induction time of calcium sulfate scale in
dotted line is the 2-hours time line. From Fig. 14, the top region the presence of methanol at two supersaturation conditions. In the
is the region where inhibitor is not needed. At the temperature and first supersaturation condition (Fig. 16a), gypsum SI is 0.39 in the
SI conditions in the middle region, low inhibitor concentration is absence of methanol. When the methanol concentration increases,
needed to inhibit calcium sulfate scale. At the temperature and SI methanol significantly reduces the gypsum solubility and, there-
conditions of the bottom region, the scale is difficult to inhibit, fore, gypsum SI increases dramatically to a maximum of 1.8 at
except with high concentrations of a pentaphosphonate or a tetra- 50% methanol concentration. The corresponding induction time
phosphonate, such as HDTMP. At 4–200ºC, calcium sulfate scale drops significantly in the presence of methanol. The calculated
becomes very difficult to inhibit when either gypsum or anhydrite induction time based on Eq. 8, which agrees well with the mea-
SI is greater than 1.0. This relationship is illustrated in Fig. 15, sured induction time, is also plotted in Fig. 16. Also plotted is the
where MIC vs. SI values are plotted at four different temperatures calculated MIC for this brine based on Eqs. 4, 8, and 10 and 2

TABLE 4—PROPOSED NUCLEATION KINETIC MODEL

FOR CALCIUM SULFATE PRECIPITATION AND INHIBITION FROM 0 TO 200°C
Calcium-Sulfate Nucleation in the Absence of Inhibitors Equation Number

At Low Temperature (0–100°C, Gypsum SI > 0.1)

0.221(± 0.079 ) 2171.2 (±93.3 ) 1.715 (±0.196 )
( )
log tind
0
=−6.297 (±0.340 )−
SI
+
T (K) ⋅SI (
0.285 ±0.025 )
+
1 +I 0.5
, r = 0.96 8

At High Temperature (100–200°C, Anhydrite SI > 0.1)

2.830 ( ±1.189 ) 885.8 1766.3 ( ±578.9 )
log tind
0
( )
= 2.148 (±0.361)−
SI
−
T (K )
+
SI ⋅T (K )
, r = 0.92 9

HDTMP/Calcium-Sulfate-Inhibition Efficiency Function (binh)

At Low Temperature (0–100°C, Gypsum SI > 0.1)

241.8 (±25.0 ) Ca2+
log binh
HDTMP
(L/mg ) = − 1.48 (±0.24 )− T (K ) ⋅SI
+ 0.091(±0.025 ) ⋅pH + 0.18 (±0.19 )⋅log , r = 0.90 10
SO24−

At High Temperature (100–200°C, Anhydrite SI > 0.1)

968.6 Ca2+
log binh
HDTMP
(L/mg )=− 1.98 (±0.51 )−1.55 (±⋅0.57 ) SI+ T (K )
+ 0.10⋅ pH +0.14 ⋅ log , r = 0.89 11
SO42−

December 2010 SPE Journal 991

 1×107

1×106

Induction Time, seconds

1×105 No inhibitor needed

2 hours line
1×104

1×103 Inhibitor needed

1×102

10

Difficult to Inhibit
1
0 20 40 60 80
Temperature, °C

Fig. 14—Plots of calculated calcium sulfate induction time based on Eq. 8 vs. temperature and gypsum SI values. The top region
represents the gypsum SI and temperature range in which no inhibitor is needed for scale inhibition of 2 hours. The middle region
represents the region of gypsum SI and temperature conditions where calcium sulfate scale can be inhibited easily by several
inhibitors. The bottom region represents the gypsum SI and temperature conditions under which inhibition is difficult with any
inhibitors, except high concentration of HDTMP.

25°C 90°C 150°C 200°C

10 35
9
30
8
7 25
MIC, mg/L
MIC, mg/L

6
20
5
4 15
3 10
2
5
1
0 0
0 0.5 1 0 0.5 1
Gypsum Sl Anhydrite Sl
(a) (b)

Fig. 15—Plots of MIC (HDTMP) calculated from Eq. 4 with parameters estimated from Eqs. 8 through 11 for gypsum and anhydrite
SI values at 25, 90, 150, and 200°C. The protection time is 2 hours, and pH of the solutions is 6.

log t0 log tinh

log t0 calc log t0 MIC Gypsum Sl
6.0 2.5
Sl MIC
4.0
3.5 8.0 5.0 2.0
log t0, log tinh, seconds

3.0 7.0
Sl, MIC, mg/L

4.0 1.5
log t0, seconds

Sl, MIC, mg/L

6.0
2.5
5.0 1.0
2.0 3.0
4.0
1.5
3.0 2.0 0.5
1.0
2.0
0.5 1.0 0.0
1.0
0.0 0.0
0 20 40 60 0.0 –0.5
0 20 40 60 80
[MeOH], v/v % [MeOH], v/v %
(a) (b)

Fig. 16—Plots of calcium sulfate induction time, gypsum SI, and MIC vs. methanol concentrations. (a) The supersaturated brine
containing 70 mM Ca and SO4 in 1 M NaCl is mixed with various volume fractions of methanol at 25°C. (b) The supersaturated
brine containing 100 mM Ca and 10 mM sulfate in 1 M NaCl is mixed with various volume fractions of methanol at 25°C and in
the presence and absence of 1 mg/L HDTMP. The changes of gypsum SI and MIC values because of methanol are calculated by
the ScaleSoftPitzer program.

992 December 2010 SPE Journal

 logt0 logtinh calc logtinh ( )inh
time, log tind , by approximately one order of magnitude, suggest-
calc logt0 Sl MIC ing more experimental data and modeling are needed, especially
for the effect of MEG and methanol on calcium sulfate nucleation
7.0 0.70 and inhibition.
0.60
logt0, logtinh, seconds
6.0
0.50
Conclusions

Sl, MIC, mg/L

5.0
The following conclusions can be made from this work:
0.40
4.0 • Common scale inhibitors show a narrow calcium sulfate SI range
0.30 (0.3–1 SI) of effective threshold inhibition. It is difficult to inhibit
3.0 calcium sulfate scale if the brine is supersaturated at SI > 1 for
0.20
2.0 gypsum or anhydrite.
0.10 • The most effective calcium sulfate inhibitor is HDTMP. Other
1.0 0.00 effective inhibitors include pentaphosphonates, phosphinopoly-
−0.10
carboxylates, maleic acid terpolymer, and carboxymethyl inulin,
0.0
0 20 40 60 80 10 in decreasing efficiency.
MEG, v/v % • A semiempirical model has been developed for prediction of both
gypsum and anhydrite nucleation-kinetics rates. This model can
Fig. 17—Plots of calcium sulfate induction time, SI, and MIC be used as a first-order estimate of the kinetics of precipitation.
vs. MEG concentrations. The supersaturated brine containing • Calcium sulfate nucleation kinetics is greatly accelerated in the pres-
70 mM Ca and SO4 in 1 M NaCl is mixed with various volume ence of methanol and, to a lesser extent, in the presence of MEG.
fractions of MEG at 25°C and in the presence and absence of • In the presence of methanol or MEG, more scale inhibitors are
0.5 mg/L HDTMP. The SI and MIC values are calculated by the needed to inhibit calcium sulfate precipitation. The amount of
ScaleSoftPitzer program.
inhibitor needed and the nucleation kinetics can be modeled
with the proposed model at < 50% methanol or MEG concen-
trations.
hours protective time. As shown, as little as 0.02 mg/L HDTMP • More research is needed to study calcium sulfate nucleation
is needed in the absence of methanol and as much as 8.4 mg/L kinetics at high temperatures (100–200°C) and over a wider
HDTMP is needed in the presence of 50% methanol. range of conditions. Research to identify more-effective calcium
The methanol effect was also tested using brine that is near sulfate scale inhibitors is also needed.
saturation with respect to gypsum in the absence of methanol
(Fig. 16b). In Fig. 16b, the induction times, log tind 0 inh
or tind , were ( ) Nomenclature
measured in the presence and absence of 1 mg/L HDTMP as scale
inhibitor. Also plotted is the calculated MIC for gypsum scale a = constant
aL− = activity of anion
based on Eqs. 4, 8, and 10 and 2 hours protective time. Adding
aMe + = activity of cation
methanol to the brine affects gypsum supersaturation and causes
simultaneous reduction in the induction time, log tind 0
. HDTMP ( ) binh = inhibition-efficiency factor, a specific parameter for a
concentration (1 mg/L) is higher than MIC at 25% methanol specific scale inhibitor
concentration and lower than the calculated MIC (1.39 mg/L) at B = the slope of the plot of log(tind
0
) vs. 1/SI2
37.5% methanol concentration. Indeed, the observed log tind inh
is ( ) C = constant
0
( )
much greater than log tind at < 37.5% methanol concentration and Cinh = scale-inhibitor concentration, mg/L
inh
log tind and log tind0
( ) ( )
are similar at 50% methanol concentration. E = activation energy for nucleation, J/mol
f = correction factor for heterogeneous nucleation [ f ≈
The calculated MIC at 75% methanol is less than 1 mg/L. There-
fore, one would expect to observe a nucleation time greater than 2 0.01(Söhnel and Mullin 1988)].
hours at 75% methanol. However, this was not observed, indicat- Ksp = conditional solubility product
ing that HDTMP did not inhibit calcium sulfate precipitation as NA = Avogadro number
our model would predict in a 75% methanol solution. The model R = gas constant
appears to predict the induction time, log tind 0
, reasonably well at ( ) (
SI = saturation index = log10 aMe + ⋅ aL− K sp)
low methanol concentration (see Fig. 16a), and the data deviates T = temperature, K
from predicted log tind 0 inh
, tind ( )
at > 50% methanol. Note that the solid
0
tind = calcium sulfate nucleation-induction time in the absence
formed at high methanol concentration in fact may be hemihydrate inh
of scale inhibitor
or anhydrite (Sandengen 2006), and, therefore, Eq. 8 may not be tind = calcium sulfate nucleation-induction time in the presence
applicable in this case. Previously, it has been observed that addi- of scale inhibitor
tional functionality is needed to model the nucleation kinetics of tprotection = time necessary to protect a well from scaling, seconds
barite in the presence of methanol (Tomson et al. 2005; Fan et al. Vm = the molar volume of the crystal (74.7 cm3/mol for gypsum
2009). Certainly, more data are needed to extend this calcium and 45.9 cm3/mol for anhydrite)
sulfate nucleation-inhibition model to the cosolvent system.  = the geometric shape factor [ = 50 for rectangular paral-
Fig. 17 shows the induction time of calcium sulfate precipitates, lelepiped (Nielsen 1964)]
0
log tind inh
, tind ( )
, in the presence of MEG and HDTMP (0.5 mg/L).
The brine composition is identical to that of Fig. 16a. Because the Acknowledgments
effect of MEG on gypsum solubility is less than that with metha-
nol, the changes in gypsum SI and induction time, log tind 0
, are ( ) This work was financially supported by a consortium of compa-
nies, including Baker-Petrolite, BJ Chemical Services, BP, Cham-
more moderate as compared with those in Fig. 16a. Also included pion Technologies, Chevron, Clariant, ConocoPhillips, Hallibur-
in Fig. 17 are the calculated MIC and calculated log tind 0 inh
, tind , ( ) ton, Kemira, Marathon Oil, M-I SWACO, Nalco, Occidental Oil
based on Eqs. 4, 8, and 10. The maximum MIC is predicted to be and Gas, Petrobras, Saudi Aramco, Shell, Statoil, and Total, and
0.34 mg/L. Therefore, HDTMP significantly slows down calcium the National Science Foundation through the Center for Biologi-
sulfate kinetics and increases nucleation time by one to two orders cal and Environmental Nanotechnology [EEC-0118007], US EPA
of magnitude when comparing to the case without scale inhibitor. ORD/NCER/STAR nanotechnology program, China-U.S. Center
Again, we observe fair agreement between the calculated and for Environmental Remediation and Sustainable Development, and
observed log tind 0
( )
, but the equations overestimate the induction the Advanced Energy Consortium.

December 2010 SPE Journal 993

References Lioliou, M.G., Paraskeva C.A., Koutsoukos P.G., and Payatakes, A.C.
Azam, S. 2007. Study on the geological and engineering aspects of anhy- 2006. Calcium sulfate precipitation in the presence of water-soluble
drite/gypsum transition in the Arabian Gulf coastal deposits. Bulletin polymers. Journal of Colloid and Interface Science 303 (1): 164–170.
of Engineering Geology and the Environment 66 (2): 177–185. doi: doi: 10.1016/[Link].2006.07.054.
10.1007/s10064-006-0053-2. Liu, S.T. and Nancollas, G.H. 1973. Linear crystallization and induction-
Cowan, J.C. and Weintritt, D.J. 1976. Water-Formed Scale Deposits. Hou- period studies of the growth of calcium sulphate dihydrate crystals.
ston, Texas: Gulf Publishing Co. Talanta 20 (2): 211–216. doi: 10.1016/0039-9140(73)80268-1.
Fan, C., Kan, A.T., and Tomson, M.B. 2009. Barite nucleation and inhibi- Liu, S.T. and Nancollas, G.H. 1975. A kinetic and morphological study
tion at 0-200°C, with and without hydrate inhibitors. Paper SPE 121559 of the seeded growth of calcium sulfate dihydrate in the presence of
presented at the SPE International Symposium on Oilfield Chemistry, additives. Journal of Colloid and Interface Science 52 (3): 593–601.
The Woodlands, Texas, USA, 20–22 April. doi: 10.2118/121559-MS. doi: 10.1016/0021-9797(75)90285-4.
Hardie, L.A. 1967. The Gypsum-Anhydrite Equilibrium at One Atmosphere Nielsen, A.E. 1964. Kinetics of Precipitation, Vol. 18. New York: Interna-
Pressure. The American Mineralogist 52 (January–February): 171–200. tional Series of Mongraphs on Analytical Chemistry, Pergamon Press.
He, S., Oddo, J.E., and Tomson, M.B. 1994a. The Inhibition of Calcium and Packter, A. 1974. The precipitation of calcium sulphate dihydrate from
Barium Sulfate Nucleation and the Relationship to Scale Control. Pre- aqueous solution: Induction periods, crystal numbers and final size.
sented at the NACE International CORROSION 94 49th Annual Con- Journal of Crystal Growth 21 (2): 191–194. doi: 10.1016/0022-
ference and Exposition, Baltimore, Maryland, 28 February–4 March. 0248(74)90004-9.
He, S., Oddo, J.E., and Tomson, M.B. 1994b. The inhibition of gypsum Sandengen, K. 2006. Prediction of mineral scale formation in wet gas
and barite nucleation in NaCl brines at temperatures from 25 to condensate pipelines and in MEG (mono ethylene glycol) regeneration
90°C. Applied Geochemistry 9 (5): 561–567. doi: 10.1016/0883- plants. PhD thesis, Department of Materials Science and Engineering,
2927(94)90018-3. NTNU, Trondheim, Norway.
He, S., Oddo, J.E., and Tomson, M.B. 1994c. The Nucleation Kinetics Söhnel, O. and Mullin, J.W. 1988. Interpretation of crystallization induction
of Calcium Sulfate Dihydrate in NaCl Solutions up to 6 m and 90°C. periods. Journal of Colloid and Interface Science 123 (1): 43–50. doi:
Journal of Colloid and Interface Science 162 (2): 297. doi: 10.1006/ 10.1016/0021-9797(88)90219-6.
jcis.1994.1042. Tomson, M.B., Fu G., Watson M.A., and Kan, A.T. 2003. Mechanisms of
He, S., Oddo, J.E., and Tomson, M.B. 1994d. The Seeded Growth of Mineral Scale Inhibition. SPE J. 18 (3): 192–199. SPE-84958-PA. doi:
Calcium Sulfate Dihydrate Crystals in NaCl Solutions up to 6 m and 10.2118/84958-PA.
90°C. Journal of Colloid and Interface Science 163 (2): 372–378. doi: Tomson, M.B., Kan, A.T., and Fu, G. 2005. Inhibition Of Barite Scale In
10.1006/jcis.1994.1116. The Presence Of Hydrate Inhibitors. SPE J. 10 (3): 256. SPE-87437-
He, S., Oddo, J.E., and Tomson, M.B. 1995. The Nucleation Kinetics of PA. doi: 10.2118/87437-PA.
Barium Sulfate in NaCl Solutions up to 6 m and 90°C. Journal of Colloid Tomson, M.B., Kan, A.T., and Oddo, J.E. 1994. Acid/Base and Metal
and Interface Science 174 (2): 319–326. doi: 10.1006/jcis.1995.1397. Complex Solution Chemistry of the Polyphosphonate DTPMP versus
He, S.L., Kan, A.T., and Tomson, M.B. 1996. Mathematical Inhibitor Temperature and Ionic Strength. Langmuir 10 (5): 1442–1449. doi:
Model for Barium Sulfate Scale Control. Langmuir 12 (7): 1901–1905. 10.1021/la00017a021.
doi: 10.1021/la950876x.
He, S.L., Kan, A.T., and Tomson, M.B. 1999. Inhibition of calcium carbon-
ate precipitation in NaCl brines from 25 to 90°C. Applied Geochemistry Chunfang Fan has been a research scientist in the Department
14 (1): 17–25. doi: 10.1016/S0883-2927(98)00033-X. of Civil and Environmental Engineering at Rice University since
2008. He holds a PhD degree in geochemistry from George
He, S.L., Kan, A.T., Tomson, M.B., and Oddo, J.E. 1997. A New Interactive
Washington University. Amy Kan is a senior research scientist at
Software for Scale Prediction, Control, and Management. Paper SPE the Department of Civil and Environmental Engineering, Rice
38801 presented at the SPE Annual Technical Conference and Exhibi- University. She holds a PhD degree from Cornell University. Kan
tion, San Antonio, Texas, USA, 5–8 October. doi: 10.2118/38801-MS. serves as an associate editor for SPE J. and as a member of
Kan, A.T., Fu, G., and Tomson, M.B. 2002. Effect of Methanol on Carbonate the executive committee on SPE International Oilfield Scale
Equilibrium and Calcite Solubility in a Gas/Methanol/Water/Salt Mixed Conference. Gongmin Fu has been a research scientist at
System. Langmuir 18 (25): 9713–9725. doi: 10.1021/la025620n. the Department of Civil and Environmental Engineering, Rice
Kan, A.T., Fu, G., and Tomson, M.B. 2003. Effect of Methanol and Eth- University, since 1992. Fu holds a PhD degree in environmental
ylene Glycol on Sulfates and Halite Scale Formation. Ind. Eng. Chem. chemistry from Drexel University. Mason B. Tomson has been
a professor of civil and environmental engineering at Rice
Res. 42 (11): 2399–2408. doi: 10.1021/ie020724e.
University since 1987. He holds a PhD degree in physical chem-
Kan, A.T., Wu, X., Fu, G., and Tomson, M.B. 2005. Validation of Scale istry from Oklahoma State University. Dong Shen is a chemist
Prediction Algorithms at Oilfield Conditions. Paper SPE 93264 pre- at BJ Chemical Services. He holds a PhD degree in physical
sented at the SPE International Symposium on Oilfield Chemistry, The chemistry from the Institute of Chemistry, Chinese Academy of
Woodlands, Texas, USA, 2–4 February. doi: 10.2118/93264-MS. Sciences.

994 December 2010 SPE Journal