Desalination 338 (2014) 93105

Contents lists available at ScienceDirect

Desalination
journal homepage: www.elsevier.com/locate/desal

Inhibition of homogenous formation of calcium carbonate by poly

(acrylic acid). The effect of molar mass and end-group functionality
Ali A. Al-Hamzah a,b, Christopher P. East c, William O.S. Doherty c, Christopher M. Fellows a,
a
b
c

School of Science and Technology, The University of New England, Armidale, NSW 2351, Australia
Saline Water Desalination Research Institute, SWCC, P. O. Box 8328, Al Jubail 31951, Saudi Arabia
Sugar Research and Innovation, Centre for Tropical Crops and Biocommodities, Queensland University of Technology, Brisbane, Queensland 4001, Australia

H I G H L I G H T S

End-groups strongly affect the effectiveness of polymeric scale inhibitors of CaCO3.

Poly(acrylic acids) with end-groups of moderate hydrophobicity are most effective.
This is true for temperatures between 25 C and 100 C.
Effective scale inhibitors also stabilize less stable polymorphs of CaCO3.

a r t i c l e

i n f o

Article history:
Received 20 December 2013
Received in revised form 23 January 2014
Accepted 24 January 2014
Available online 22 February 2014
Keywords:
Crystallization
Calcium carbonate
Polyelectrolytes
Scale inhibitor
SEM (Scanning Electron Microscopy)
XRD (X-ray Diffraction)
Poly(acrylic acid)
ATRP (Atom Transfer Radical Polymerization)

a b s t r a c t
The ability of poly(acrylic acid) (PAA) with different end groups and molar masses prepared by Atom Transfer
Radical Polymerization (ATRP) to inhibit the formation of calcium carbonate scale at low and elevated temperatures was investigated. Inhibition of CaCO3 deposition was affected by the hydrophobicity of the end groups of
PAA, with the greatest inhibition seen for PAA with hydrophobic end groups of moderate size (610 carbons).
The morphologies of CaCO3 crystals were signicantly distorted in the presence of these PAAs. The smallest
morphological change was in the presence of PAA with long hydrophobic end groups (16 carbons) and the relative inhibition observed for all species were in the same order at 30 C and 100 C. As well as distorting morphologies, the scale inhibitors appeared to stabilize the less thermodynamically favorable polymorph, vaterite, to a
degree proportional to their ability to inhibit precipitation.
2014 Elsevier B.V. All rights reserved.

1. Introduction
Calcium carbonate is one of the most common scales found in both
thermal (e.g., multi-stage ash, MSF) and membrane (e.g., reverse osmosis) desalination processes. At lower temperatures, it is the main
component of alkaline scale, while at higher temperatures it is found
mixed with magnesium hydroxide in desalination plant scales [1]. The
exact temperature at which Mg(OH)2 deposition becomes competitive
with CaCO3 deposition will depend on the extent to which carbon dioxide is degassed from the brine [2].
In MSF desalination, CaCO3 generally appears above 45 C as a result
of the thermal decomposition of the bicarbonate ion (Re. (1)); increasing

Corresponding author. Tel.: +61 2 6773 2470.

E-mail address: cfellows@une.edu.au (C.M. Fellows).
http://dx.doi.org/10.1016/j.desal.2014.01.020
0011-9164 / 2014 Elsevier B.V. All rights reserved.

temperature pushes bicarbonate to carbonate by the entropicallyfavored reaction.

2

2HCO3 aq CO3 aq CO2 aq H2 Ol

Re:1

The precipitation of CaCO3 occurs when the ion product exceeds the
Ksp (Re. (2)). The concentration of Ca2+ and HCO
3 ions in standard seawater (Salinity = 35 g/kg) are 10.3 and 1.8 mM respectively, but the
will be much lower and sensitively dependent
concentration of CO2
3
on conditions [3].
2

Ca aq CO3 aq CaCO3 s

Re:2

Calcium carbonate can be found as an amorphous solid and in three

different crystalline forms, calcite (Ksp at 25 C = 3.3 109), aragonite
(Ksp at 25 C = 4.6 109) and vaterite (Ksp at 25 C = 1.2 108) [4,5]
listed in order of increasing solubility and decreasing thermodynamic

94

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

stability. The crystal shape and morphology of calcium carbonate

precipitation is affected by factors such as temperature, pH, supersaturation, the ratio of [Ca2+]/[CO2
3 ]and the presence or absence of additive [6]. At high temperatures (T N 70 C), aragonite is favored, while
calcite is favored at low temperature (T b 30 C). At any temperature
all polymorphs eventually recrystallize to the thermodynamicallyfavored calcite [7]. Rhombohedral calcite is favored at a 1:1 [Ca2 +]/
2+
]/
[CO2
3 ] ratio, while scalenohedral calcite is favored when the [Ca
]
ratio
is
1.2
[7].
[CO2
3
To control scaling in desalination plants, several methods of control
have been adopted. The most important methods are acid treatment,
mechanical cleaning and the use of polymeric scale inhibitors. The primary method used historically to control scale formation in MSF desalination plants has been acid treatment. In this treatment, the pH of
seawater is maintained around 4.5 using acid, most often sulfuric acid
due to its low cost. Controlling the pH of treated seawater is crucial for
both inhibition of scale and prevention of corrosion. Acid reacts with
and HCO23 ions present in seawater yielding H2O and CO2
the CO2
3
and preventing formation of CaCO3 and Mg(OH)2 [8].
Scale inhibitors are chemical additives used to control the formation and/or deposition of scale. There are three common groups of polymeric scale inhibitors: polymers containing carboxylic acids such as
poly(acrylic acid) and poly(maleic acid); polymers containing phosphate groups, such as polyphosphates and polyphosphate esters; and
polymers containing sulfonate groups [9]. The attractive features of
these chemical additives include ease of handling, relatively low cost,
low dose rate, and ability to inhibit hard calcium sulfate scale formation
[10,11].
It has been suggested that scale inhibitors may operate by three
distinct mechanisms; by sequestration, dispersion, or adsorption [9].
As solubility is dened as the maximum concentration of dissolved
ions in equilibrium with solid phase at a xed temperature and background ionic composition, that maximum concentration of dissolved
ions can be reduced prior to nucleation by sequestration with scale
inhibitors.
Scale forms primarily heterogeneously, on interfaces (of bubbles,
plant surfaces, and particles of suspended matter), but this heterogeneous nucleation is still primarily in the bulk phase. If scale particles
formed in solution can be prevented from aggregating onto surfaces
and remain suspended in the nal brine, they will not contribute to scaling. By adsorbing to the surface of these particles and providing addition
electrostatic and/or steric stabilization, scale inhibitors can retard aggregation of these particles, effectively dispersing them in suspension until
they are ejected in the waste brine.
Differences in crystal form can arise from selective adsorption of
scale inhibitors on the points of crystal growth, causing differential reduction in the growth rate of different crystal planes. For example, in
the case of calcium carbonate, scale inhibitors can stabilize the vaterite
and prevent its transformation into calcite or aragonite [12]. The new
crystal morphologies may grow more slowly overall, may not aggregate
as effectively as the native crystal morphologies, or may form a deposit
that is more porous and more easily removed [13].
We have previously found in investigations of inhibition of the
related calcium oxalate (CaC2O4) scaling system (prevalent in sugar
manufacture) that the effectiveness of poly(acrylic acid) (PAA) scale inhibitors is sensitively dependent not only on dose and molar mass, but
also on the end-groups attached to the polymer [1416]. Reversibletermination radical polymerization methods, such as Reversible
AdditionFragmentation chain Transfer polymerization (RAFT) [17]
and Atom Transfer Radical Polymerization (ATRP) [18] provide a way
in which the molar mass and end-group functionality of PAA can
be controlled to a high degree of precision, enabling such structure
property relations to be conclusively established. For CaC2O4 scaling,
we found that most effective scale inhibition by PAA was obtained not
for hydrophilic end groups, or hydrophobic end-groups large enough
to generate signicant surface activity, but moderately hydrophobic

end-groups [14]. These polymers had the greatest impact on crystal

speciation and morphology [15].
The main aim of this paper is to determine if similar trends occur
in CaCO3 scaling as have been observed previously for CaC2O4. The
efciency of PAA of controlled molar mass with different end-groups
as inhibitors of CaCO3 crystallization in the bulk solution was determined at ambient and elevated temperatures, using conductivity and
turbidity measurements. We have also previously observed differing
impacts of this same set of PAAs on the decomposition of the HCO
3
ion which suggests that they should have different effects on CaCO3
scaling [19].

2. Experimental
2.1. Synthesis and characterization of poly(acrylic acid)
Poly(acrylic acid) with different end-groups (Table 2) and molar
masses were synthesized by Atom Transfer Radical Polymerization
(ATRP) of t-butyl acrylate and subsequent hydrolysis with triuoroacetic
acid, as described previously [15]. Molar masses of PAA were estimated
by 1H and NMR spectroscopy (Bruker-300) and Gel Permeation Chromatography (GPC Waters 1525 HPLC, Waters auto-sampler 712 WISP and
Waters 2414 RI detector).
The following PAAs were tested: carboxymethyl-1,1-dimethylPAA (CMM-PAA, Mn = 1500, 7600, 11,800); ethyl-isobutyrate-PAA,
(EIB-PAA Mn = 1700, 5100, 7200); cyclohexyl-isobutyrate-PAA (CIBPAA, Mn = 1700, 1900, 3500, 5100, 8400, 11,000, 13,200); n-hexylisobutyrate-PAA (HIB-PAA, Mn = 1400, 2000, 3600, 4200, 6700, 8900,
13,100); n-decyl-isobutyrate-PAA (DIB-PAA, Mn = 2400, 4500, 6200);
and n-hexadecyl-isobutyrate-PAA (HDIB-PAA, Mn = 1700, 4100, 9400,
7200) (see Supplementary material, Fig. S1).

2.2. Crystallization test conditions

Two stock solutions, 0.167 M CO23 as Na2CO3 (10,000 ppm) and
0.413 M Ca2+ as CaCl2 (16,500 ppm), were prepared. These solutions
were ltered and degassed using a 0.45 m Millipore solvent lter.
PAA solutions were prepared by dissolving 0.015 g of PAA in 20 mL
water (750 ppm) and were used after three days to ensure complete
dissolution of the polymer.
Tests were conducted at a pH of 9.2 under four sets of conditions of
varying temperatures and supersaturation levels (SL = Qsp/Ksp, where
Qsp is the solubility quotient), as outlined in Table 1. It should be
noted that the Ca2+ concentrations are much less than those expected
under typical thermal desalination conditions, and at a correspondingly
lower [Ca2+]/[CO2
3 ] ratio: thus while the supersaturation values are
comparable to thermal desalination conditions, they correspond to distinctly different environments for crystal growth [20].
Time-resolved measurements of crystal formation, and microscopic
examination of the crystals formed in these experiments was carried

Table 1
Crystallization test conditions with PAA.
Condition set

pH
T C
[Ca2+] ppm
[CO2
3 ] ppm
[Ca2+]/[CO2
3 ]
[PAA] ppm
9
Ksp 10
SL

9.2
25
66
100
1:1
1.50
4.95
556

60
66
100
1:1
1.50
2.80
983

100
36
30
1.8:1
0.50
1.69
277

100
66
100
1:1
6.70
1.69
1629

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

out to see if any changes in speciation and morphology analogous to

those seen in the calcium oxalate system could be observed for CaCO3.

95

(a)
610

Induction Time

2.2.2. Determination of conductivity at 60 C

Filtered (0.45 m cellulose acetate membrane) deionized water
(98.25 mL) was placed in a clean cell containing the conductivity
probe and thermometer. One drop of 0.1 M NaOH, Ca2+ as CaCl2 solution to give a nal concentration of 66 ppm and scale inhibitor solution
to give 1.5 ppm were added to the cell under stirring. Recording of conductivity started when the solution reached the target temperature and
20 s later CO23 as Na2CO3 solution to give a nal concentration of
100 ppm was added to the cell. Conductivity was measured by a conductivity probe and meter (Beta 81, CHK Engineering).

570
550
530
510
490
470
450
0

500

1000

1500

2000

Time (s)

(b)
200
180
160

Absorbance (900nm)

2.2.1. Determination of turbidity and conductivity at 25 C

Filtered (0.45 m cellulose acetate membrane) deionized water
(246 mL) was placed in a 500 mL cleaned beaker containing the conductivity probe and meter (Beta 81, CHK Engineering). Under magnetic
stirring, one drop of 0.1 M NaOH, Ca2+ as CaCl2 solution to give a nal
concentration of 66 ppm and scale inhibitor solution to give a nal concentration of 1.5 ppm were added. The solution was pumped continuously through a 1 cm Quartz ow cell using a Gilson peristaltic pump
(Minipuls 2) with 4 mm silicon tubing and then returned to the main
mixture. CO23 as Na2CO3 solution was added to the beaker to give a
nal concentration of 100 ppm. A UVvis spectrophotometer (Unicam
spectrophotometer SP6-550) was used to measure the increase in turbidity with time by recording apparent absorption at 900 nm (where
absorbance is negligible). Recording of absorbance at 900 nm started
20 s after addition of CO2
3 solution. Due to the formation of air bubbles
at high temperatures which interfere with turbidity measurements, turbidity measurements were only recorded at 25 C (condition set 1).
Analog outputs from conductivity probe and spectrophotometer
were digitally converted using a Picolog A/D Converter 16 (16 Bit) and
Picolog recording software and data was acquired every 5 and 10 s.
After each experiment all equipment was ushed multiple times with
weak acid followed by R/O water.

Conductivity (S/cm)

590

140

Steady State

120
100
80
60
40
20
Induction Time

0
0

500

1000

1500

Time (s)
Fig. 1. Determination of induction time and steady state for three different experiments
(ne gray lines) and their average (thick gray line) for a solution containing Ca2 +
and CO2
ions and 1.5 ppm HDIB-PAA (Mn = 17,200) under condition set 1 (SL = 556,
3
T = 25 C) by conductivity (a) and turbidity (b).

2.3. Steady state and induction time

The steady state (SS) is dened as the conductivity of the system
when it is under equilibrium, that is, the measured conductivity when
there is no more precipitation and conductivity value remain stable.
Inhibition efciency supersaturation level (% IE) was determined by
applying the following equation (Eq. (1))


% IE


SSC Blank
 100
C C Blank

where SS is the steady state value of conductivity, CBlank is the equilibrium conductivity value of the system with no PAA under the given
conditions, and C0 is the conductivity value before any CaCO3 scale formation under the given conditions. For example % IE of PAA at without
any CaCO3 scale formation = 100% and IE of blank = 0%.
Induction time (IT) is a measure of the time before any effects of
crystal formation on macroscopic properties of the reaction mixture
can be observed. It was measured by tting a straight line to the decreasing part of the conductivity curve (Fig. 1a) or increasing part of
the turbidity curve (Fig. 1b) and a straight line to the initial invariant
part of the curve. The induction time is then measured as the intersection of these two lines, as illustrated in Fig. 1.

2.4. Scanning Electron Microscopy (SEM), X-ray Diffraction (XRD) and

Fourier Transform Infrared Spectroscopy (FTIR)
Crystals of CaCO3 in the presence and absence of PAA (Mn 2000)
under condition set 4 were collected after 1000 s and ltered through
0.45 m pore size cellulose acetate lter paper and then characterized
by SEM, ATR-FTIR and XRD. Scale samples were gold-coated and SEM
images obtained using a FEI Quanta 200 Environmental SEM at an accelerating voltage of 15 kV. XRD was carried out using a Rigaku diffraction
camera with an X-ray generator with Cu K radiation of wavelength
1.5418 at the X-ray Analysis Facility, Queensland University of Technology, Brisbane, Australia. FTIR was carried out using a Varian 660-IR
Spectrometer.

3. Results
The conductivity and turbidity measurement curves of solutions
containing Ca2 + and CO23 ions and PAA with different end groups
and molar masses under the conditions described in Section 2 are detailed below. First, results obtained at 25 C (condition set 1), then 60
C (condition set 2) and then 100 C at relatively low (condition set 3)
and relatively high (condition set 4) supersaturation values. Following

96

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

Table 2
Induction times (IT) and % inhibition efciency (% IE) of PAA for CaCO3 formation under condition set 1. Unclear indicates that no clear transition corresponding to an induction time could
be observed. (1) Group one, no precipitation and no signicant change in conductivity. (2) Group two, minor precipitation with long IT. (3) Group three, more signicant precipitation and
shorter IT.

End group
terminated
PAA

CMM

EIB

CIB

HIB

DIB

Mn

Blank

SS
conductivity
(S/cm)

IT
turbidity

% IE

419

0
100

IT conductivity

Group

225

400

>5000

>5000

1457

1.5

592

7633

1.5

582

94.2

Unclear

Unclear

11773

1.6

576

90.5

Unclear

Unclear

1669

1.3

592

>5000

>5000

5065

1.3

481

35.8

796

+110
225

866

+208
332

7180

1.3

514

54.9

460

+81
185

610

+212
23

1689

1.4

592

100

>5000

>5000

3518

1.2

592

100

>5000

>5000

5088

1.4

592

84.4

Unclear

Unclear

8400

1.2

560

81.5

Unclear

Unclear

9954

1.3

497

45.1

510

+50
75

100

749

+204
25

790

+204
25

433

+146
50

10988

1.2

505

49.7

620

+86
160

13209

1.1

456

21.4

318

+33
23

1403

592

100

>5000

>5000

1981

1.2

592

100

>5000

>5000

3563

1.2

592

100

>5000

>5000

4224

1.2

592

100

>5000

>5000

6723

1.2

592

100

>5000

>5000

8928

1.1

511

53.2

829

+70
99

1190

+15
209

13094

1.1

461

24.3

314

+30
13

421

+103
13

2422

1.3

546

73.4

Unclear

Unclear

4472

1.3

530

64.2

Unclear

Unclear

6203

1.5

503

48.6

Unclear

Unclear

1687

491

41.6

706

4135

497

45.1

1095

9391

489

40.5

609

17167

470

29.5

507

+6
16

959

+266
239

1150

+238
365

+40
27

827

+166
15

+16
35

676

+104
139

+92
397

HDIB

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

100

97

3.1. Inhibition of CaCO3 crystallization at room temperature (condition set 1)

90
80
70

% IE

60
50
40
30
20
10
0
0

5000

10000

15000

20000

Mn
Fig. 2. Inhibition efciency (% IE) of CaCO3 formation by PAA with different molar masses
and end groups ( CMM, EIB, CIB, HIB, DIB and HDIB) under condition set
1 (SL = 556, T = 25 C).

Induction times were determined by conductivity and turbidity

measurements in the absence and presence of PAAs under condition
set 1. No statistically signicant difference between the IT values
determined by conductivity and turbidity measurements for CaCO3 formation at room temperature. However, at short times it can be seen that
systematically longer IT values are reported using conductivity rather
than turbidity, suggesting that the initially formed colloidal particles
scatter considerable light while their effect on conductance is still
marginal compared to the starting conditions of the solution (see Supplementary material, Fig. S2). The IT values measured by the two
methods thus correspond to different physical states of the system
and are not directly comparable.
For condition set 1, CBlank = 419 S/cm and C0 = 592 S/cm. The inhibition efciency, conductivity and turbidity measurements of PAA
under condition set 1 allowed the PAA to be divided into three groups
(Table 2, Fig. 2).
Group One For the lowest molar mass of PAA (Mn ~ 2000) with
hydrophilic end group and short and medium hydrophobic end groups, such as carboxymethyl-1,1-dimethyl
(CMM), ethyl isobutyrate (EIB), hexyl isobutyrate (HIB)
and cyclohexyl isobutyrate (CIB) terminated-PAAs, excellent % IE were obtained. No precipitation was observed

the presentation of these results, the results of characterizations carried

out on crystals formed at 100 C (condition set 4) SEM, FTIR, and XRD
will be given.

Table 3
Induction times (IT) and % inhibition efciency (% IE) of PAA for CaCO3 formation under condition sets 2, 3 and 4.

Mn

SS
(S/cm)

IT
conductivity

453

175

End groups
terminated
PAA
Blank

% IE

Condition
set 2

2106

1.3

588

217

+166
80

68.3

7633

1.5

521

110

+12
5

34.4

1669

1.3

573

202

+58
61

61.1

7180

1.3

494

128

+42
16

21.1

1689

1.4

600

304

+126
121

74.4

5088

1.3

521

158

+41
28

34.9

589

250

+12
90

68.8

Condition
set 3

67.8

Group

Condition
set 4

65.9

CMM
2

84.3

75.3

EIB
2

84.3

93.2

CIB

1403

100

100

HIB
3563

1.3

522

176

+34
49

50.3

2422

1.3

503

163

25.5

25.5

4472

1.3

504

130

+26
9

25.6

1687

506

182

+68
17

26.9

4135

484

188

+66
29

15.8

21.6

18.2

DIB
2

12.7

45.5

HDIB
2

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

and the turbidity and conductivity changed very slowly,

therefore no induction time could be reported (Table 2).
Group Two For decyl isobutyrate (DIB) terminated-PAA, HIB-PAA and
CIB-PAA of moderate molar mass (Mn ~ 6000), % IE were
generally good, and little precipitation was observed,
with high induction times.
Group Three For the highest molar mass of PAA (Mn N 9000) for all end
groups except CMM, and for all molar masses for the longest end group hexadecyl isobutyrate (HDIB), % IE is less
than 50%, with distinct induction times.

1403 (HIB)

850

Cond.uctivity (S/cm)

98

1669 (EIB)
1689 (CIB)
2106 (CMM)

800

2422 (DIB)
1687 (HDIB)

750
Blank

700

650

3.2. Inhibition of CaCO3 crystallization at elevated temperatures (condition

sets 2, 3 and 4)

600
0

200

400

600

800

1000

Time (s)
Since CaCO3 is the predominant alkaline scale in MSF plants at temperatures less than 100 C, the % IE was determined at 60 and 100 C.
Under condition set 2, conductivity measurements were used to determine the % IE of PAA with different end groups and molar mass to inhibit
CaCO3 crystallization in bulk solution for the polymers which performed
very well (IE = 100%) under condition set 1.

3.2.1. Inhibition of CaCO3 crystallization at 60 C

Under condition set 2, conductivity measurements was used to determine the % IE and induction time of PAA with different end groups
and molar mass in the range of (14009000) to inhibit CaCO3 formation
in bulk solution (Table 3, Fig. 3).
For condition set 2, CBlank = 453 S/cm and C0 = 650 S/cm. The % IE
and induction time of PAA with different end groups and molar mass
allow the PAA to be divided into two groups.
Group One Low molar mass PAA with hydrophilic end group (CMM)
and short and medium hydrophobic end groups, such as
EIB-PAA (Mn = 1669), HIB-PAA (Mn = 1403) and CIBPAA (Mn = 1689) showed good inhibition efciency
(% IE N 50) and high induction times.
Group Two PAA with molar mass more than 4000 for hydrophilic end
group (CMM) and short and medium hydrophobic end
groups (EIB, CIB and HIB), and low molar mass PAA with
long hydrophobic end groups (DIB and HDIB), which
showed low inhibition efciency (% IE b 30%) and low induction times.

Fig. 4. Conductivity measurements of solutions containing Ca2+ and CO2

ions and 0.50
3
ppm PAA (Mn 2000) with different end groups under condition set 3 (SL = 277, T =
100 C) with end groups: HIB, CIB, EIB, CMM, DIB, HDIB and Blank.

3.2.2. Inhibition of CaCO3 crystallization at 100 C (condition sets 3 and 4)

at 100 C were greater than
As the conductivities of Ca2+ and CO2
3
at lower temperatures, causing the total conductivity to exceed the scale
of the conductivity meter used, the system of CaCO3 crystallization recording was changed by increasing the cell constant. Considering the
decrease in % IE and induction time of PAA with molar mass more
than 2000 under condition set 2 (60 C), only the lowest molar mass
PAA with different end groups were chosen for investigation under condition sets 3 and 4 (Table 3). The concentration of PAA was increased to
6.7 ppm under condition set 4 to make the trends in % IE and induction
time of PAA with different end groups more evident. While condition
set 4 maintained the same equimolar Ca2 +:CO23 ratio as condition
sets 1 and 2, condition set 3 employed an excess of Ca2+.
Conductivity results for condition sets 3 and 4 are summarized in
Figs. 4 and 5 respectively.
It can clearly be seen that at higher temperatures the difference between PAA bearing different end-groups is reduced, but not to any dramatic extent, and that the relative effectiveness of the different PAA is
retained under the two sets of conditions; the only difference in the reversal in order of HDIB-PAA and DIB-PAA in order of effectiveness at the
higher SL.

100
90
1403 (HIB)

80

980

1689 (CIB)

Conductivity(S/cm)

70

% IE

60
50
40
30
20

1669 (EIB)
2106 (CMM)

940

1687 (HDIB)
2422 (DIB)

900
Blank

860

10
0
0

2000

4000

6000

8000

10000

Mn
Fig. 3. Inhibition efciency (% IE) of CaCO3 formation by PAA with different end groups
and molar masses ( CMM, EIB, CIB, HIB, DIB and HDIB) under condition set
2 (SL = 983, T = 60 C).

820
0

100

200

300

400

500

600

Time (s)
Fig. 5. Conductivity measurements of solutions containing Ca2+ and CO2
ions and 6.7
3
ppm PAA (Mn 2000) with different end groups under condition set 4 (SL = 1629, T =
100 C) with end groups: HIB, CIB, EIB, CMM, DIB, HDIB and Blank.

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

99

Fig. 6. SEM micrograph magnication (500) of rod-like CaCO3 crystals, identied by XRD
as largely aragonite, in the absence of PAA under condition set 4 (SL = 1629, T = 100 C).

The similarity between the results under condition set 3 and condition set 4 requires explanation. At the same supersaturation level, less
scale control would be expected at 0.5 ppm than 6.7 ppm; but very similar results are seen in both cases. However, the concentration of the
scale-forming ions is considerably lower in condition set 3 and the
ratio of SL values in the two systems (5.9) is roughly comparable to
the difference in PAA concentrations (13.4).

3.3. SEM, XRD and FTIR results

In the absence of PAA, CaCO3 crystals occurred primarily as a mixture
of calcite and aragonite in a rod-like morphology as shown by SEM
(Fig. 6) with traces of rhombohedral calcite (Fig. 11(a)) and hexagonal

Fig. 8. XRD of CaCO3 crystals after subtraction of background under condition set 4 (SL =
1629, T = 100 C). A: aragonite, C: calcite, V: vaterite. (a) in the absence of PAA (Blank);
(b) in the presence of CMM-PAA (Mn = 2100); (c) in the presence of DIB-PAA (Mn =
2400); (d) in the presence of HDIB-PAA (Mn = 1700).

Fig. 7. FTIR spectra of CaCO3 crystals prepared under condition set 4 (SL = 1629, T =
100 C). A: aragonite, C: calcite, V: vaterite. (a) in the absence of PAA; (b) in the presence
of DIB-PAA (Mn = 2400).

orettes of vaterite (Fig. 12(a)). FTIR results showed two peaks at

711 cm1 indicating calcite polymorph and at 1082 cm1 indicating
aragonite formation (Fig. 7(a)) [21], while the XRD of these crystals suggest that they are largely aragonite (Fig. 8(a)).
After background correction, XRD results showed that the CaCO3
present was a mixture of calcite and aragonite (Fig. 8). Those results

100

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

Fig. 9. SEM micrographs ( 500) of CaCO3 crystals in the presence of PAA (Mn 2000) with different end groups under condition set 4 (SL = 1629, T = 100 C). (a) CMM-PAA;
(b) CIB-PAA; (c) HIB-PAA; (d) DIB-PAA; (e) HDIB-PAA.

may indicate that the rod-like morphology consisted originally of aragonite polymorph as was observed by Yang et al. [22].
In the presence of low molar mass PAA with different end groups,
two points can be observed from the SEM micrographs ( 500) of

CaCO3 crystals. First is the signicant reduction in the population of

CaCO3 crystals in order of HIB N CIB N CMM N HDIB N DIB which is compatible with the conductivity measurements under condition set 4
(Fig. 9). Second is the distortion in crystals of the different polymorphs

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

101

Fig. 10. SEM micrographs (2000) of aragonite rod-like CaCO3 morphology under condition set 4 (SL = 1629, T = 100 C) in the absence (a) and in the presence of PAA (Mn ~ 2000) with
different end groups; (b) CMM-PAA; (c) HIB-PAA; (d) DIB-PAA; (e) HDIB-PAA.

of aragonite (Fig. 10), calcite (Fig. 11), and vaterite (Fig. 12) in the same
order. This is presumably due to adsorption of PAA with different end
groups to the active faces of growing nuclei of CaCO3 [23].
The results of SEM (Fig. 9) and XRD (Fig. 8(b)) for PAA with a hydrophilic end group (CMM) showed the CaCO3 crystals to be mostly calcite
of rhombohedral morphology, with distorted edges. The hydrophilicity
of the CMM group may make the effect of PAA on the growth stage of
calcium carbonate precipitation more pronounced than at the nucleation stage, and it appears to be much less selective in its effect, in
that all crystals formed in the presence of CMM-PAA are signicantly

distorted, while the rod-like crystals formed in the presence of DIBPAA and HDIB-PAA show little distortion (Fig. 10).
In contrast to PAA-CMM, the SEM of crystals grown in the presence
of PAA with long hydrophobic end groups (DIB and HDIB) shows crystals of CaCO3 were a mixture of rod-like aragonite, vaterite orettes
and a few single crystals of rhombohedral calcite (Figs. 912). The formation of vaterite ower polymorph may be due to the adsorption of
PAA on active vaterite surfaces, preventing the transformation of
vaterite into aragonite or calcite. The distortion of different polymorphs
by DIB-PAA and HDIB-PAA was the least in comparison to other

102

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

Fig. 11. SEM micrographs (30,000) of calcite rhombohedral morphology under condition set 4 (SL = 1629, T = 100 C) in the absence (a) and in the presence of PAA (Mn ~ 2000) with
different end groups; (b) CMM-PAA; (c) HIB-PAA; (d) DIB-PAA; (e) HDIB-PAA.

hydrophobic end groups of PAA. Moreover, the rod-like crystals were

shorter (~ 28 and 29 m in the presence of PAA with HDIB and DIB respectively) than the same morphologies in the absence of PAA (~ 63
m) (Fig. 10).
Due to the control of CaCO3 formation by HIB-PAA and CIB-PAA, no
peaks could be observed by FTIR or XRD. However, there were traces
of single crystals in different habits with rod-like crystals (14 m),
rhombohedral calcite and vaterite all detected by SEM. Those images
illustrate the highest distortion in different forms, which may be due
to the mid-length hydrophobic end groups discouraging PAA chains
from desorbing from the nuclei of CaCO3 as fast as PAA with hydrophilic

end groups, delaying the growth of the CaCO3 nuclei. Alternatively, the
hydrophobic end-groups may lead to preferential adsorption at the
edges between more and less charged surfaces, giving a higher degree
of distortion per amount of polymer adsorbed.
3.4. Discussion
The target molar mass of scale inhibitors for good control of scale formation is 10003000 [9]. In the results presented here for CaCO3 scaling,
PAAs in that molar mass range were generally most effective in inhibition of CaCO3 crystallization. The results for all conditions showed that

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

103

Fig. 12. SEM micrographs (10,000) of vaterite morphology under condition set 4 (SL = 1629, T = 100 C) in the absence (a) and in the presence of PAA (Mn ~ 2000) with different end
groups; (b) CMM-PAA; (c) CIB-PAA; (d) HIB-PAA; (e) DIB-PAA; (f) HDIB-PAA.

the nature of the end groups terminating PAA played a very important
role, as previously reported for calcium oxalate scaling. In general, the
two measurements of effectiveness (induction time, IT, which relates
to the nucleation stage of scale formation and inhibition efciency,
% IE, which relates to the growth stage of scale crystals) tracked each
other closely, with an additive giving a good result by one measure
also giving a good result with the other.
Induction time was affected by the hydrophobicity of end groups,
with the hydrophobic end groups having a usually longer induction
time than the hydrophilic end groups, as previously found in the calcium oxalate system [14]. This may possibly be because the hydrophobic
end groups discourage the PAA chains from desorbing from the nuclei of
CaCO3 as fast as PAA with hydrophilic end groups, delaying the growth

of the CaCO3 nuclei. However, IT measurements showed a higher variability and a reduced difference between PAA than measurements of %
IE, consistent with the greater variability in the nucleation stage of
scaling.
The % IE of CaCO3 precipitation was affected strongly by the size of
the end groups. PAAs with short end groups with both hydrophilic
(CMM) and hydrophobic (EIB) and middle hydrophobic end groups
(CIB and HIB) of PAA were found to have excellent inhibition efciency
(100%) at room temperature and are more efcient than PAA with long
hydrophobic end groups (DIB and HDIB) at all temperatures investigated (25100 C). At high temperature, the inhibition efciency of PAA
with hydrophobic end groups of moderate size (CIB and HIB) remained
superior. The inhibition efciency and induction time of PAAs decreased

104

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

Table 4
The relationship between the nature and length end groups of PAA and the distortion and speciation of CaCO3 polymorphs under condition set 4.
End-group

% IE

Distortion

Average rod
length (m)

Average vaterite
diameter (m)

Predominant polymorph

None
CMM
HIB/CIB
DIB
HDIB

0
66
100/94
20
45

None
Moderate
High
Low
Low

63
19
14
29
28

20
9.5
9
13
21

Rod-like
Rhombohedral calcite
Mixture of rod-like, vaterite and
rhombohedral calcite

with the increasing temperature, which is reasonable since at higher

temperatures, the solubility product of CaCO3 will decrease to give a
greater thermodynamic driving force for crystallization. Therefore,
more rapid and less reversible adsorption of PAA to delay the crystal
growth of CaCO3 will be more required than at low temperature.
It has been found for calcium oxalate that the best performance of
PAA inhibitors occurs at an equimolar ratio of cations to anions [24],
while the similarity between results obtained in this work at very different inhibitor concentrations (0.5 and 6.7 ppm) at different Ca2+:CO2
3
ratios (1.8 and 1.0) may suggest that in the calcium carbonate scaling
system an excess of cations is preferable. However, the concentrations
of scale-forming ions were also different in the two systems, making it
impossible to draw any conclusion.
Different charge densities readily develop on different faces of growing crystals, with molecular dynamics being an ideal tool to estimate
these values [25]. Based on our previous results on calcium oxalate
where marked changes in crystal morphology were observed [15], we
advanced the hypothesis of edge activity [26], where end-group modied polyelectrolytes might partition selectively to edges between
growing interfaces with different charge densities specically, to the
edge between a highly negatively-charged and a near neutral face
as a mechanism to achieve greater effectiveness than indiscriminate adsorption of scale inhibitor to crystallites. In order to investigate this further, SEM and XRD measurements were carried out on the crystal
products formed under condition set 4.
The formation of different polymorphs of CaCO3 crystals precipitated
under condition set 4 ([Ca2+]/[CO2
3 ] = 1, SL = 1629 and T = 100 C)
can be explained by a dissolutionrecrystallization mechanism, where
the least thermodynamically stable phase is initially formed, and
dissolves and reprecipitates in the absence of scale inhibitor to a more
stable phase.
We suggest that the less thermodynamically stable vaterite phase is
formed rst, consistent with Ostwald's rule of stages which hold that
the least thermodynamically stable phase is usually kinetically preferred [27]. All scale inhibitors can bind to this phase to some extent,
retarding its dissolution and leading to distorted crystals. Aragonite
and calcite then both form from these vaterite crystals by a process of
dissolution and recrystallization, with the crystals formed being considerably more distorted and showing a higher degree of twinning then in
the absence of scale inhibitor.
The nature and length of end groups of PAA play a signicant role in
the distortion of different polymorphs of calcium carbonate. The highest
distortion in those polymorphs was in the presence of PAA with end
groups of moderate hydrophobicity (HIB and CIB) while the lowest distortion was in the presence of PAA with long end groups (DIB and HDIB)
(Figs. 1012). This distortion is likely to be dependent both on the selectivity of PAA with different end groups to adsorb on different active
faces of CaCO3 crystals and its likely rate of adsorption/desorption to/
from those active faces [28]. It is to be expected that PAA will adhere
most readily to the most positively charged faces of the growing crystals. Hydrophobic end-groups would be expected to prefer relatively
uncharged surfaces, possibly directing PAA toward edges; while longer
hydrophobic groups may self-assemble, leading to unselective PAA adsorption and retarding desorption.
In Table 4 it can be seen that effective scale inhibitors also gave
shorter rod-like crystals and smaller vaterite orettes, suggesting

again that they have a strong impact on growth as well as nucleation. Although the inhibition efciency of PAA with DIB and HDIB end groups
was the lowest (18.2 17 and 45.5 5.7 respectively), the action of
these polymers as scale inhibitors can clearly be seen in the reduction
of rod-like population (comparison with the blank) and the stabilization
of vaterite metastable polymorph (Fig. 9(d, e)). The correlation between
% IE and amount of rod-like polymorph formed can be observed in these
images.
Finally, in contrast to the action of these same scale inhibitors previously reported on the calcium oxalate system [15], all changes in morphology observed in the calcium carbonate system tended to a loss of
denition in the structures. This suggests a relatively indiscriminate adsorption of polymer to all surface and/or edges, which is not surprising
considering the high (6.7 ppm) concentration of inhibitor employed in
condition set 4. Work is currently underway on the behavior of PAA at
lower concentrations and over a range of ion ratios.
4. Conclusion
The inhibition efciency of PAA with different end groups and molar
masses as scale inhibitors to prevent CaCO3 formation in bulk solution
was studied at temperatures ranging between 25 and 100 C using conductivity and turbidity. The results showed that both molar mass and
the nature of the end group of PAA affect the inhibition of CaCO3 scale.
Low molar mass PAA with short and moderately-sized end groups had
relatively high inhibition efciency and long induction times under all
conditions investigated, while all high molar mass PAAs had poor
inhibition efciency of CaCO3. Low molar mass PAA with long hydrophobic end groups had poor inhibition efciency of CaCO3 scale. At
room temperature, the lowest molar mass of PAA with hydrophilic
end group showed good efciency in the inhibition of CaCO3 scale
making it suitable for use as a scale inhibitor in RO desalination.
However, with increasing temperature, the lowest molar mass of PAA
with different middle hydrophobic end groups gave a better inhibition
efciency and induction time than PAA with hydrophilic end groups.
These results suggest that the lowest molar mass PAA with end groups
of moderate hydrophobicity are more suitable as scale inhibitors in MSF
desalination.
Effectiveness of the inhibitors declined with increasing temperature,
but the relative effectiveness of the inhibitors according to end-group
remained essentially the same. At 100 C, 0.5 ppm PAA applied to a
system with SL = 277 gave approximately the same inhibition efciency (% IE) as 6.7 ppm applied at SL = 1629, with the order of efciency of
the different end-groups being almost identical.
The nature and length of end groups of PAA had a major effect on
the morphologies of CaCO3 obtained at 100 C with 6.7 ppm. The highest
distortion in the CaCO3 polymorphs was for PAA with mid-hydrophobic
end groups. However, the lowest distortion in different CaCO3 polymorphs was for PAA with long-hydrophobic end groups. These results
may due to its rate of adsorption/desorption on the active faces of
CaCO3 polymorphs. The conductivity and morphology results are compatible, where the % IE of CaCO3 formation and the distortion of its morphologies have the same order of HIB N CIB N CMM N HDIB N DIB. As
previously observed with the calcium oxalate system, the PAA appears
to be stabilizing the initially-formed kinetically-favored product so it
is not transformed to the thermodynamic product, though less

A.A. Al-Hamzah et al. / Desalination 338 (2014) 93105

comprehensively. The crystals formed are more ill-dened than observed in the calcium oxalate system, which may be related to the
higher concentration of polymer employed, a smaller difference in
charge density between crystal growth faces, or some other unidentied feature of the system.

[7]

[8]
[9]

List of acronyms
ATRP
Atom Transfer Radical Polymerization
CMM-PAA carboxymethyl-1,1-dimethyl-terminated poly(acrylic acid)
EIB-PAA ethyl-isobutyrate-terminated poly(acrylic acid)
CIB-PAA cyclohexyl-isobutyrate-terminated poly(acrylic acid)
DIB-PAA n-decyl isobutyrate-terminated poly(acrylic acid)
FTIR
Fourier Transform Infrared spectroscopy
HDIB-PAA n-hexadecyl isobutyrate-terminated poly(acrylic acid)
HIB-PAA n-hexyl isobutyrate-terminated poly(acrylic acid)
IT
induction time
MSF
multi-stage ash thermal desalination
PAA
poly(acrylic acid)
RAFT
Reversible AdditionFragmentation chain Transfer
polymerization
SEM
Scanning Electron Microscopy
SL
supersaturation level
SS
steady state value of a parameter
XRD
X-ray diffraction
% IE
Percent inhibition efciency

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.desal.2014.01.020.
References
[1] A.M. Shams El Din, M.E. El-Dahshan, R.A. Mohammed, Inhibition of the thermal decomposition of HCO
3 : a novel approach to the problem of alkaline scale formation
in seawater desalination plants, Desalination 142 (2002) 151159.
[2] F. Rahman, Z. Amjad, Scale formation and control in thermal desalination systems,
The Science and Technology of Industrial Water Treatment, CRC Press, Boca Raton,
2010, pp. 271296.
[3] K. Al-Anezi, N. Hilal, Effect of carbon dioxide in seawater on desalination: a comprehensive review, Sep. Purif. Rev. 35 (2006) 223247.
[4] G. Wolf, E. Knigsberger, H.G. Schmidt, L.-C. Knigsberger, H. Gamsjger, Thermodynamic aspects of the vateritecalcite phase transition, J. Therm. Anal. Calorim. 60
(2000) 463472.
[5] D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm, S.M. Bailey, I. Halow, K.L.
Churney, R.L. Nuttall, Selected values of thermodynamic constants, in: R.C. Weast,
D.R. Lide, M.J. Astle, W.H. Beyer (Eds.), Handbook of Chemistry and Physics, 70th
ed., CRC Press, Boca Raton, Florida, 1989.
[6] O. Cizer, K. Balen, J. Elsen, D. Gemert, Crystal morphology of precipitated calcite crystals from accelerated carbonation of lime binders, ACEME08, 2nd International

[10]

[11]
[12]

[13]

[14]

[15]

[16]

[17]
[18]

[19]

[20]
[21]

[22]
[23]

[24]
[25]

[26]

[27]
[28]

105

Conference on Accelerated Carbonation for Environmental and Materials Engineering, Rome, 2008.
W. Jung, S. Kang, C. Kim, C. Choi, Particle morphology of calcium carbonate precipitated by gasliquid reaction in a CouetteTaylor reactor, Chem. Eng. Sci. 55 (2000)
733747.
D. Hasson, R. Semiat, Scale control in saline and wastewater desalination, Isr. J.
Chem. 46 (2006) 97104.
E. Senogles, W.O.S. Doherty, O.L. Crees, Scale inhibitors, polymeric, in: J. Kroschwitz
(Ed.), Encyclopedia of Polymer Science and Technology, 2nd ed., Wiley-Interscience,
Boca Raton, 1996.
C. Gabrielli, M. Keddam, H. Perrot, A. Khalil, R. Rosset, M. Zidoune, Characterization
of the efciency of the antiscale treatment of water. Part I: chemical processes,
J. Appl. Electrochem. 26 (1996) 11251132.
O.A. Hamed, H.A. Al-Otaibi, Prospects of operation of MSF desalination plants at high
TBT and low antiscalant dosing rate, Desalination 256 (2010) 181189.
Z. Amjad, R.W. Zuhl, Effect of heat treatment on the performance of deposit control
polymers as calcium carbonate inhibitors, Corrosion 2007 NACE International,
Nashville2007.
O.A. Hamed, M.A.K. Al-So, G.M. Mustafa, A.G. Dalvi, The performance of different anti-scalants in multi-stage ash distillers, Desalination 123 (1999)
185194.
A.D. Wallace, A. Al-Hamzah, C.P. East, W.O.S. Doherty, C.M. Fellows, Effect of poly
(acrylic acid) end-group functionality on inhibition of calcium oxalate crystal
growth, J. Appl. Polym. Sci. 116 (2010) 11651171.
C.P. East, A.D. Wallace, A. Al-Hamzah, W.O.S. Doherty, C.M. Fellows, Effect of poly
(acrylic acid) molecular mass and end-group functionality on calcium oxalate crystal morphology and growth, J. Appl. Polym. Sci. 115 (2010) 21272135.
W.O.S. Doherty, C.M. Fellows, S. Gorjian, E. Senogles, W.H. Cheung, Inhibition of calcium oxalate monohydrate by poly(acrylic acids) with different end-groups, J. Appl.
Polym. Sci. 91 (2004) 20352041.
M. Siauw, B.S. Hawkett, S. Perrier, Short chain amphiphilic diblock co-oligomers via
RAFT polymerization, J. Polym. Sci. A Polym. Chem. 50 (2012) 187198.
H. Minami, A. Tanaka, Y. Kagawa, M. Okubo, Preparation of poly(acrylic
acid)-b-polystyrene by two-step atom transfer radical polymerization in supercritical carbon dioxide, J. Polym. Sci. A Polym. Chem. 50 (2012) 25782584.
A. Al-Hamzah, C.M. Fellows, Apparent inhibition of thermal decomposition of
hydrogencarbonate ion by poly(acrylic acid). The effect of molar mass and endgroup functionality, Desalination 332 (2014) 3343.
Y. Wang, University of New South Wales (Sydney), 2005.
K. Naka, D. Keum, Y. Tanaka, Y. Chujo, Control of crystal polymorphs by a latent
inductor: crystallization of calcium carbonate in conjunction with in situ radical
polymerization of sodium acrylate in aqueous solution, Chem. Commun. (2000)
15371538.
D. Yang, L. Qi, J. Ma, Well-dened star-shaped calcite crystals found in agarose gels,
Chem. Commun. (2003) 11801181.
E. Hadicke, J. Rieger, I. Ursula Rau, D. Boeckh, Molecular dynamics simulations of the
incrustation inhibition by polymeric additives, Phys. Chem. Chem. Phys. 1 (1999)
38913898.
W. Al-Thubaiti and C. M. Fellows, unpublished work, 2013.
J. Rieger, E. Hadicke, I.U. Rau, D. Boeckh, A rational approach to the mechanisms of
incrustation inhibition by polymeric additives, Tenside Surfactants Deterg. 34
(1997) 430435.
C.M. Fellows, A. Al-Hamzah, C.P. East, W.O.S. Doherty, E.J. Smith, Edge-active polymers for tailoring crystal morphology, 12th Pacic Polymer Conference, Jeju,
Korea, 2011.
T. Threlfall, Structural and thermodynamic explanations of Ostwald's rule, Org. Process Res. Dev. 7 (2003) 10171027.
A.S. Schenk, I. Zlotnikov, B. Pokroy, N. Gierlinger, A. Masic, P. Zaslansky, A.N. Fitch, O.
Paris, T.H. Metzger, H. Coelfen, P. Fratzl, B. Aichmayer, Hierarchical calcite crystals
with occlusions of a simple polyelectrolyte mimic complex biomineral structures,
Adv. Funct. Mater. 22 (2012) 46684676.