PROCEEDINGS, 43rd Workshop on Geothermal Reservoir Engineering

Stanford University, Stanford, California, February 12-14, 2018

SGP-TR-213

Scaling Assessment, Inhibition and Monitoring of Geothermal Wells

Arjen Boersma, Frank Vercauteren, Hartmut Fischer, Francesco Pizzocolo

High Tech Campus 25, 5656 AE, Eindhoven (the Netherlands)
francesco.pizzocolo@tno.nl

Keywords: Scaling, Inhibition, Monitoring, Geothermal Well, Scale Management, Experimental

ABSTRACT
The formation of inorganic, sparingly soluble salts from aqueous brines during geothermal energy production, is known as scale and it is
one of the major flow assurance problems. The damage caused by scale is one of the biggest challenges encountered by geothermal
industries. Scale deposition damages the downhole equipment and moreover decreases in flow rate leads to a loss in production. Scale
forms and deposits under supersaturated conditions, wherever the mixing of the incompatible types of water, formation water from the
bottom hole and the injected seawater, takes place. Or when temperature or pressure changes are severe enough to produce a
supersaturated solution. The deposited scale adheres on the surfaces of the producing well tubing and on parts of water handling
equipment, where it builds up in time and leads to problems in reservoirs, pumps, valves and topside facilities. The rapid increase of the
mineral deposits leads to inevitable damage of the equipment parts. As a consequence, suspension of production activities is necessary
for the recovery or replacement of damaged parts. The formation of scaling on the inner surface of casings depends on the scaling
tendency of the water and the surface chemistry and morphology of the casing. Many mitigation strategies exist, the most common are:
scale inhibitors and acids washes. The scaling tendency can only be lowered by a chemical intervention on the production fluids: reduce
the supersaturation of the salts in the water by inhibitors that form chemical complexes with the scaling ions. Preferably, no scale
inhibitors are used, and other ways of prevention are required. In this study we experimentally show how the amount of scaling that is
deposited on a substrate mainly depends on three parameters: the surface energy of the casing and the roughness and the bulk modulus
of its inner surface. A long term solution for scaling prevention is the implementation of a casing material that is intrinsically better
resistant against scaling. A new class of pipes that shows intrinsically improved scaling performances is glass fiber reinforced
composites. The inner surface of the pipes consist of the polymer matrix of epoxy or polyethylene and will have a different surface
energy than steel pipes.

1. INTRODUCTION
The formation of inorganic, sparingly soluble salts from aqueous brines during oil and gas production, and during geothermal energy
production, is known as scale and is one of the major flow assurance problems (University of Leeds). Scale forms and deposits under
supersaturated conditions, wherever the mixing of the incompatible types of water; formation water from the bottom hole and the
injected seawater, takes place. Or when temperature or pressure changes are severe enough to produce a supersaturated solution. The
deposited scale adheres on the surfaces of the producing well tubing and on parts of water handling equipment, where it builds up in
time and leads to problems in reservoirs, pumps, valves and topside facilities. Since the deposition/ adhesion takes place also in heat
transfer equipment such as boilers and heat exchangers, a decrease in the performance during the heat exchange as well as in flow rates
occurs. These problematic conditions are found not only in oil and gas fields but commonly also in geothermal plants where energy is
obtained from hot fluids. The rapid increase of the mineral deposits leads to inevitable damage of the equipment parts and to a decrease
of the production rate. As a consequence, suspension of production operations is necessary for the recovery or replacement of damaged
parts and to remove the clog generated by the scale deposition. The total costs of scale (for the oil and gas market) has been estimated at
USD 1.4 billion (Eroini, 2011).
Two common, most insoluble types of inorganic scale, are calcium carbonate (CaCO3) and barium sulphate (BaSO4). Both mineral
types are undesirable since they result in blockage of pipes and in many other flow assurance problems. A very well-known case of
scaling occurring near the wellbore is the formation of calcium carbonate when the pressure of the fluid drops and becomes lower than
the bubble point of CO2. Thus CO2 start degassing from the brine, resulting in pH increase and subsequently in an environment
saturated with respect to CaCO3. Scaling can also occur when incompatible types of water (e.g. seawater and formation water) are
mixed and forms insoluble salts (Mavredaki, 2014). Most of the research on scaling has been devoted to bulk phase scaling or on seeded
crystals. Only during the last few decades more attention has been given to the growth of scaling directly on surfaces. Currently, it has
been accepted that calcium carbonate reveals different growth kinetics in the bulk phase than the one dominating the deposition on a
surface (Zhang, 2012). The formation of scaling on the inner surface of casings depends on the scaling tendency (supersaturation) of the
water and the surface chemistry and morphology of the casing. Steel casings have been shown to be very sensitive for scaling, and many
mitigation strategies have been proposed: coatings, scale inhibitors, acids washes etc. The scaling tendency can only be lowered by a
chemical intervention of the production fluids: reduce the supersaturation of the salts in the water by inhibitors that form chemical
complexes with the scaling ions.
The mechanism in the formation of scaling can be divided in several steps:

1. Scaling salts in water become supersaturated, due to mixing temperature and/or pressure changes, etc.
2. Inorganic salts start to precipitate around small crystallization nuclei.

1
Boersma, Vercauteren, Fischer and Pizzocolo

3. The nucleus can be the surface of the pipe or a inhomogeneity in the liquid.
4. If the nucleus forms in the liquid and does not adhere to the pipe, the salt crystal will grow, but does not form scaling in the
pipe.
5. If the nucleus is formed in the liquid, but is deposited onto the pipe, or the crystal is nucleated from the pipe wall, scale starts
to build up.
6. Once the initial crystals have started to grow on the wall, further scaling will occur.

To properly manage the scale related issues, the formation of scale must be prevented in the nucleation steps: this must not start at the
pipe inner surface. The nucleation stage of the scale can be influenced by the surface chemistry of the pipe. The most important
parameter to be assessed is the surface energy of the material on which the scale is deposited. But also the morphology and roughness
will play an important role.
A more long term solution would be the implementation of a casing material that is intrinsically better resistant against scaling. Thus has
bulk chemical and physical properties that prevent the deposition of inorganic salts. A new class of pipes that may show intrinsically
improved scaling performances is glass fiber reinforced composites. The inner surface of the pipes consist of the polymer matrix of
epoxy or polyethylene and will have a different surface energy than steel pipes.
Although the scaling of metal pipes has been a topic for research for many years, the scale formation of polymer pipes (Wang, 2005)
has not been extensively investigated yet. The majority of the literature describes the prevention of scaling by using scale inhibitors, or
the mechanical removal of scale. In this study we performed several scale deposition experiments, both on steel and polymeric materials
that can be used as wellbore casing.

2. METHOD

2.1 Theoretical background

The driving force for the inorganic scale to form is the saturation ratio (SR) or saturation index (SI). It is essential for reactive
crystallization to occur that the SR is greater than 1, in the mixed solution.

𝑎𝐶𝑎 ∙ 𝑎𝐶𝑂3
𝑆𝐼 = log(𝑆𝑅) = log ( ) (1)
𝐾𝑠𝑝

Where aca and aCO3 are the activities respectively of the Ca and CO3 ions and Ksp is the solubility product. The adopted solution of Ksp is
given by Larson (1942).
Since most scaling studies have only used thermodynamics and bulk kinetics to model scale formation, they have led to under or over
estimation of the scaling. For accurate influence of the surfaces on scaling, surface effects must be taken into account, such as
chemistry, pre-scaled surfaces and roughness. A study on the formation of scale on stainless steel revealed that at high SR values (>5)
bulk crystallization is favorable and surface scale is much less than when SR < 5 (Eroini, 2011).
The activity of calcium and carbonate depends on the concentration, the temperature and the pressure. For low concentrations, the
activity coefficient of ions (fi) is given by:

−0.51 ∙ 𝑧 2 ∙ √𝐼
log(𝑓𝑖 ) = (2)
1+ √𝐼

Where z is the charge of the ion and the ionic strength (I) of the solution is given by:
1
𝐼= ∙ ∑ 𝐶𝑖 ∙ 𝑧𝑖2 (3)
2

Where Ci is the molar concentration of the ith ion present in the solution and zi is the charge of the ions. The activity of each ion is now
given by:

𝑎𝑖𝑜𝑛 = 𝑓𝑖 ∙ 𝐶𝑖𝑜𝑛 (4)

Extensive research has shown scaling to be a two-stage process, with the first period identified as the induction period and the second
period known as the fouling period (Bargir, 2007). During the induction period a small amount of scale (inorganic particulate fouling)
accumulates on the surface without significantly affecting material performance. Although the quantity of scale formed is small, it is
enough to condition the surface and enable a thin layer of scale to form. It is this conditioned layer which is succeeded by the fouling
period, an overall decrease in the performance of the system. The induction period, while often neglected, offers much potential for
mitigating fouling. The formation of a conditioning layer during the induction period is a balance between the deposition and removal of
material at the interface between the solid and liquid. The deposition process is classically viewed as a heterogeneous nucleation
process, where foreign bodies or impurities act as nucleation sites. The energetics of heterogeneous nucleation is described as a
modification of homogeneous nucleation to account for the different interfaces and precipitation processes.
The interfacial tensions are composed of dispersive or Lifshitz-van der Waals, Lewis acid-base (or electron-acceptor/donor) and
electrical double layer force. In a changing pH environment, the charges on the surfaces can change (especially in the case of metals and
metal oxides), which leads to a change in surface tension due to the formation of electron/donor moieties. This will influence the
adhesion significantly. The surface energy is constructed of at least two contributions: dispersive and a-scalar (combined dipolar,
hydrogen bonding and induction forces) forces. For example metals have a significant a-scalar contribution, whereas fluorinated
polymers do not. An important parameter that steers the adhesion of scaling crystals to a surface is the work of adhesion (Mitchel, 2010
and van Krevelen, 1997). Taking subscript 1 for the scale, 2 for the substrate and 3 for the water, the work of adhesion can be written as:
2
 Boersma, Vercauteren, Fischer and Pizzocolo

𝑊𝑎𝑑ℎ = 𝛾13 + 𝛾23 − 𝛾12 (5)

If this energy is large, then the scale prefers to be only in contact with water and the adhesion between scale and surface will be poor.
The lowest adhesion will be found if the interfacial tension between scale and surface is lower than the combined interfacial tension
between scale-liquid and surface-liquid. Since the surface tension of water and calcium carbonate are more or less fixed, the surface
energy between the substrate and scale crystals, 12, must be low to reduce adhesion between scale and surface.
The surface tensions as described above are only valid if the surface a completely smooth. Any roughness on the surface distorts the
surface energy, because of an increasing contact area (Good, 1998). Good introduced the ratio r:
𝑎
𝑟= (6)
𝐴

in which a is the real microscopic area and A the apparent area.

The apparent surface or interfacial energies between the rough surface and the scale, or the water increases with a factor r. The
roughness of a surface can be characterized by several different roughness parameters, such as R rms, (root mean squared) or Ra
(arithmetic average). The ratio r can be estimated using the roughness root mean squared by:

𝑅𝑟𝑚𝑠 2
𝑟= + 1 (7)
𝜆2

In which  is the distance between the peaks on the surface. So if the roughness increases, the apparent surface tension of the surface, 2
increases. And looking at the work of adhesions as was just described above, and increasing 2 results in a more positive work of
adhesion, so more scaling. For the assessment of the influence of the roughness on the surface tension, two parameters for the roughness
of the surface are needed.
Next to the influence of surface roughness on the interfacial tension of the scale and substrate, also the roughness of the substrate itself
plays a significant role (Bargir, 2007). Height distribution is consider to be a major factor during adhesion between rough surfaces.
When the scale is growing on a rough surface, the scaling is favored when the roughness is high. If, in that case, scale starts to grow, the
adhesion may be poor and can be easily washed away.

2.2 Scaling experiments

To quantify the relevance of the parameters that mostly influence the deposition of scale on the inner surface of the casing we created a
special experimental set-up. This was engineered as a result of the assessment of four literature tests: the quartz crystal microbalance
(QCM) experiment (Eroini, 2011, Mavredaki, 2014), the bulk jar test (Carpentier, 2014, 2015), the flow cell (Eroini, 2011) and the
dynamic flow rig (Mavredaki, 2014).
The QCM tests were performed to monitor the formation of scale on the steel surfaces and it was built into a flow cell.
In the bulk jar test scale cations and anions are heated up to 80°C and mixed together in a jar. The sample are immersed in the liquid and
rotated to obtain laminar or turbulent flow. The scaling tests were performed for 2 hours, and the amount of scaling is measured by
weighing the samples. Alternatively, only Calcium carbonate scale can be produced by using a simpler mixture at lower temperature of
50°C.
In the flow cell two components of brine are mixed in a jacked beaker at elevated temperatures. The mixed liquid is pumped through a
flow cell having windows to visually assess the growth of the scaling layer. A turbidity and pH meter monitor the change in liquid
properties. Only calcium carbonate is monitored.
Since we will be assessing the difference between surfaces, and not the scaling tendency of solutions, not a single value for the
saturation ratio (SR) is selected, but it is increased gradually during the test. A flow cell is more representative for down hole conditions,
but is more difficult to construct and only a single sample per experiment can be measured. Furthermore, high volumes of scaling brines
are required. The scaling must start after mixing, so sufficient induction time was allowed. This means that the saturation ratio (equation
1) at the point of injected of the second fluid into the first must not exceed a value of 20.
Finally, during the dynamic flow rig test the two brine solutions are mixed at elevated temperatures and forced through a stainless steel
capillary cell. The pressure built-up over this cell due to the precipitation of calcium carbonate is monitored in time. Various saturation
ratios were assessed between 54.8 and 2.5 for their scaling tendency. The induction time for scaling increases rapidly below a SR value
of 4.
The scaling experiment that was performed in this paper will result in an initial assessment of the scaling tendency of several surfaces
as a function of the coating properties, surface tension (), roughness (R), and modulus (E). The surface tension is varied by the
incorporation of hydrophobic and hydrophilic groups in the coatings. The roughness variations is introduced by the implementation of
nanotexturing on the metal substrate or on the coating surface by embossing. The modulus is varied by the use of different amounts of
crosslinker in the same formulations or by using a different substrate, such as polymer composite. Different coating formulations were
adopted to change the surface energy of the steel plates and are listed in Table 1.
Three steel substrates are used: Q-lab steel panels QD, R and S. These have different surface roughness. The surface modifiers and
coatings were applied from a very dilute solution to the steel substrates using a 10 μm doctor blade. Taking a volume fraction of 2 %,
this would yield a 200 nm coating layer on the substrates. The surface modifiers, solvents and curing protocol are listed in Table 1.
Three different epoxy formulations have been included, because this could give qualitative and quantitative indications about the
material properties that should be modified to improve the scaling behavior of glass fiber reinforced epoxy pipes.
The surface energies of all samples were measured using a Krüss DSA100 surface tension machine. The surface tension experiments
were done on coated QD panels, because these had a smallest distortion from surface roughness. A droplet of water is placed on the

3
Boersma, Vercauteren, Fischer and Pizzocolo

surface and the contact angle between the droplet and the surface is measured. This is repeated with a droplet of diiodo methane. From
the contact angles, the dispersive and polar component of the surface energy can be calculated.

Type Name Solvent Curing process

Fluoroacrylate 2% in FC 770 Room temperature
Fluorinated
Fluorolink F10 2% in water/IPA Room temperature
Araldite (2:1) 2% in toluene 60 min at 80°C, 15 min at 150°C
Epoxy Araldite (2:5:1) 2% in toluene 60 min at 80°C, 15 min at 150°C
Araldite (2:1.5) 2% in toluene 60 min at 80°C, 15 min at 150°C
PDMS Sylgard 184 2% in hexane 2 hours at 80°C
Tetraethyl orthosilicate
2% ethanol 1 hour at 120°C
SiO2 (TEOS)
SmartCoat 1% ethanol Room temperature
TEOS/Decyl trimethoxy
Alkyl 2% ethanol 1 hour at 120°C
silane (DTMS) (1:1)

Table 1. Surface modifiers and solvents used in the scaling experiments.

The roughness of the substrates is measured using a Sensofar Optical Imaging Profiler. An area of approx. 500x500 μm2 is scanned
using the confocal microscope. Using the processing software, the most relevant roughness parameters are determined, such as the
average roughness (Ra), surface area increase, peak area (Sh) and dale area (Sd).
The scaling experiments are conducted as follows:

 The substrate is coated with the surface modifier

 Three polymer cylinders are glued on the substrate, two on top of the coated area, one on the uncoated steel (Figure 1)
 The cylinders are filled with 4 ml (0.1 M) CaCl2 solution
 A 0.2 M solution (4 ml) of NaHCO3 is added rapidly in the center of the cylinder. When the solutions are mixed in this way,
the CaCO3 will slowly starts to form after an induction period. This is needed to avoid the immediate crystallization of the
scale in the bulk.
 The scaling solution is left for 1 hours at room temperature and the scale will form on the substrate.
 Then, the cylinders are removed and the substrates are thoroughly washed, to remove the unreacted chemicals, and scaling
that is not formed on the substrates.
 After drying the scaling layers are characterized.
 The polymer tube on the untreated steel is used as reference measurement, and as a measure for the reproducibility of the
scaling experiments.
 All experiments yield a duplo for the scaling tendency.

Figure 1. Scaling experiment set-up of three glued polymer tubes on the substrates.
After all scaling experiments were performed, a series of 30 steel substrates were obtained, all having a layer of scale formed in the
polymer tubes. The polymer tubes were removed and the resulting scaling layers were characterized.

4
 Boersma, Vercauteren, Fischer and Pizzocolo

Figure 2. Large scaling experiments on 30 different substrate-surface modifier combinations. Left: overview. Center: samples
during scaling. Right: samples after scaling.
The characterization of the scale formation is a very important aspect of the current project. Normally, the weight changes are monitored
of samples that are introduced into scaling liquids. However, since we have only used small parts of the substrates for scaling, weight
measurements are not sufficient. Therefore, the specular light reflection from the surface was monitored using an Ocean Optics USB
4000 spectrophotometer at 650 nm. The reflected light from the non-treated steel surface was used as reference, and the absorbance of
the scaling layer was calculated according to:
𝐼
𝐴 = −𝑙𝑜𝑔 ( 𝑠𝑐𝑎𝑙𝑒) (8)
𝐼𝑠𝑡𝑒𝑒𝑙

Iscale is the light intensity of reflected light from the scaling layer and Isteel is the reflected light from the steel substrate. First, the light
reflection of the scale layer that was formed was measured, then the scale was removed by the application and removal of an adhesive
tape. The forces generated in this removal were max 2.5 N/cm, but could also be close to 0 N/cm, when the scale was not attached to the
substrate. When the scale was not firmly attached to the substrate, it came off and a cleaner surface was left behind. This was observed
in a lower value of the absorbance. This experiment resulted in information about the adhesion between scale and substrate.
The scaling experiments on the glass fiber reinforced (GFR) epoxy tubing was done in a similar way as the steel samples. Some parts of
the inside of the epoxy tubing were coated with the surface modifiers as described above. The same scaling experiments were done
using the polymer cylinders glued to the inside of the tube. Some of the polymer cylinders glued on the inside of the GFR epoxy tube
are shown in Figure 3. Each cylinder was filled with the scaling liquid and left for 1 hour, after which the liquid was removed and the
ring was washed. Then, a new position in the tube was treated. After drying of the samples, similar light reflection experiments were
performed as were done for the steel substrates.

Figure 3. Scaling experiments on the inside of the Glass Fiber Reinforced epoxy tube.

5
Boersma, Vercauteren, Fischer and Pizzocolo

3. RESULTS & DISCUSSION

The first series of experiments is conducted at 23 °C. The beaker is filled with a CaCl2 solution of 0.05M, the CO2 concentration is kept
atmospheric (~300 ppm). A solution of 0.002 M NaHCO3 is added slowly to the CaCl2 solution. Using the equations above, a mixture of
0.05M CaCl2 and 0.002M NaHCO3 (1:1) has a SR value of 17. At the point of injection, a high NaHCO3 concentration exist after which
the NaHCO3 is diluted into the rest of the liquid. However, due to mixing and dilution of the NaHCO3 and CaCl2 liquids the SR value
will never exceed the 21. At the end of the experiment all the liquid has a SR of 17, although the reaction of the calcium with the
carbonate will reduce the SR. During the experiment the HCO3- is converted to CO32-. The concentration of CO32- is low compared to
the concentrations of Ca2+ and HCO3-. In the initial reaction of Ca2+ with CO32- the consumed carbonate is replenished by the
bicarbonate, and the concentration of carbonate remains rather constant: the SR value decreases only slightly. Once the amount of
reacted calcium is equal to half of the added bicarbonate, the concentration of carbonate decreases orders of magnitude and the SR
becomes almost zero.
Assuming that serious scaling will occur at SR values above 5, when adding NaHCO3 to Ca, scaling will start only when half the
(0.002M) NaHCO3 is added to the (0.05M) calcium solution. On the other hand, when the calcium solution is added to the bicarbonate
solution, the SR value is initially above 20 and stays this high almost to the end of the addition (Figure 4). This is caused by the fact that
the ionic strength increases upon addition of the CaCl2, and without calcium the carbonate concentration is substantially higher, thus
leading to a higher SR value.

Figure 4. Saturation ratio development of calcium and bicarbonate upon the addition of one to the other. [NaHCO3] = 0.002 M,
[CaCl2] = 0.05 M.
For the heterogeneous generation of scale on the substrates, it is beneficial that the scale forms slowly, and does not precipitates
immediately after adding one of the components. Therefore, adding the NaHCO3 to the CaCl2 will produce a slower scale that will
preferably adhere to the samples. This was proven by an initial experiment using these two chemicals. However, it was also observed
that the concentrations of chemicals can be higher, and further scaling experiments will be conducted using 0.1 M CaCl2 and 0.2 M
NaHCO3. This difference between experiments and theory is caused by the nature of the NaHCO3 solution, which changes in time,
because of the conversion of HCO3- to CO3-2.

The surface energies of all samples were measured using a Krüss DSA100 surface tension machine. From the contact angles, the
dispersive and polar component of the surface energy can be calculated. The values for these energies are listed in Table 2.

Substrate Surface modifier total [mN/m] dispersive [mN/m] polar [mN/m]

Fluoroacrylate 8.9 ± 0.66 8.5 ± 0.58 0.4 ± 0.08
Fluorolink F10 14.1 ± 2.2 10.6 ± 1.48 3.5 ± 0.72
Araldite (2:1) 32.2 ± 2.26 29.3 ± 0.68 2.8 ± 1.58
Araldite (2:5:1) 29.4 ± 1.49 25.1 ± 0.74 4.2 ± 0.74
QD Araldite (2:1.5) 34.6 ± 8.28 29.7 ± 6.43 4.9 ± 1.85
Sylgard 184 23 ± 0.78 22.1 ± 0.66 0.8 ± 0.12
TEOS 29.7 ± 1.34 27.7 ± 0.86 2 ± 0.48
SmartCoat 5.3 ± 0.38 5.2 ± 0.3 0.1 ± 0.08
TEOS/DTMS (1:1) 21.1 ± 1.86 20 ± 1.58 1 ± 0.28
6
 Boersma, Vercauteren, Fischer and Pizzocolo

QD None 36.7 ± 1.7 29.9 ± 1.0 6.8 ± 0.7

R None 37.7 ± 1.3 30.6 ± 0.8 7.1 ± 0.6
S None 32.7 ± 1.4 28.4 ± 1.1 4.4 ± 0.3
Epoxy plate None 46.9 ± 4.0 38 ± 2.4 8.9 ±1.7
CaCO3 None 39.7 36.0 3.7
Water 52.2 19.9 72.1

Table 2. Surface energies for the substrates and surface modifiers.

The roughness of the substrates is measured using a Sensofar Optical Imaging Profiler. Typical images of the four surfaces are shown in
Figure 5 for the three steel panels R, QD and S, and the GFR epoxy substrate. The R-panel is dull finished, and has a more scattered
surfaces, whereas QD and S are brushed and polished and show a striped pattern. The epoxy surface appears to be much smoother than
the steel plates.

Figure 5. Surface images of the 3 steel substrates panel-R, panel-QD, panel-S and one epoxy substrate.

The values for the roughness parameters are listed in Table 3.

Sample Finish Roughness Sa [m] Area of peaks, r1 Area increase, r2

Q-la Qd Smooth 0.38 0.81 1.01
Q-lab R Dull 1.18 0.49 1.07
Q-lab S Ground 1.02 0.78 1.35
GFR epoxy No 0.044 0.93 1.01

Table 3. Types and surface properties of steel Q-panel substrates and glass fiber reinforced epoxy. r1 = Sh/(Sd+Sh) (Sh = mean
hill area; Sd = mean dale area); r2 =1 + Sdr (=Developed interfacial area ratio).
7
Boersma, Vercauteren, Fischer and Pizzocolo

A set of 30 steel substrates were obtained from the scale experiments. After removing the polymer tubes the scale layers were
characterized (Figure 6). The specular light reflection from the surface was measured. The calculated absorbances for the 3 steel panels
are listed in tables 4, 5 and 6.

Figure 6. Light reflection measurements on the scaled samples.

Surface modifier A_Scale A_Tape abs

Bare 0.947 0.947 0.000
Sylgard 184 0.269 0.174 0.095
SmartCoat 0.078 0.021 0.057
TEOS/DTMS 0.130 0.097 0.033
TEOS 0.383 0.291 0.092
Fluorolink F10 0.049 0.012 0.037
Fluorocrylate 0.370 0.028 0.343
Araldite (2:1) 0.871 0.812 0.059
Araldite (2:1:5) 0.938 0.935 0.003
Araldite (2.5:1) 0.758 0.737 0.021

Table 4. Q-Panel R: light absorption of the scale layer formed on the treated substrates, before (A_Scale) and after removal by
adhesive tape (A_Tape).

Surface modifier A_Scale A_Tape abs

Bare 1.092 1.092 0.000
Sylgard 184 0.471 0.064 0.407
SmartCoat 0.346 0.097 0.249
TEOS/DTMS 0.448 0.453 -0.005
TEOS 0.487 0.502 -0.015
Fluorolink F10 0.348 0.263 0.086
Fluorocrylate 0.368 0.158 0.209
Araldite (2:1) 1.041 0.979 0.063
Araldite (2:1:5) 0.917 0.876 0.041
Araldite (2.5:1) 0.916 0.880 0.036

Table 5. Q-Panel QD: light absorption of the scale layer formed on the treated substrates, before and after removal by adhesive
tape.

8
 Boersma, Vercauteren, Fischer and Pizzocolo

Surface modifier A_Scale A_Tape abs

Bare 0.553 0.553 0.000
Sylgard 184 0.314 0.290 0.023
SmartCoat 0.087 0.083 0.004
TEOS/DTMS 0.112 0.114 -0.003
TEOS 0.204 0.129 0.074
Fluorolink F10 0.058 0.057 0.001
Fluorocrylate 0.215 0.083 0.133
Araldite (2:1) 0.504 0.437 0.067
Araldite (2:1:5) 0.567 0.497 0.071
Araldite (2.5:1) 0.409 0.333 0.076

Table 6. Q-Panel S: light absorption of the scale layer formed on the treated substrates, before and after removal by adhesive
tape

When comparing the bare, untreated steel panels, the S-panel seems to give the least scaling. This is also obvious in the microscope
images of the three scaling layers (Figure 7). The amount of crystals (measured as a surface coverage) attached to the S surface is much
smaller than the R and QD surfaces. And it appears that the crystal size on the QD panel is smaller.

Figure 7. Microscope images of the 3 bare Q-panels, R, QD and S. The calcium carbonate crystals can be seen as the black dots.
Although the surface roughness is almost the same for the Q-panels R and S, the peak density for S is much higher, meaning sharper and
more peaks on the surface. So, the roughest surface shows to smallest amount of scaling. The smoothest surface shows the most scaling.
An example of the influence of the surface modifier on the scaling tendency is shown in Figure 8. The dark areas are filled with calcium
carbonate particles, the light area is reflected light. Three coatings inhibit scale formation, a third (PFTES - Perfluorooctyl
triethoxysilane) not.
The conventional way of processing scaling data is by plotting the amount of scaling against the surface energy of the substrate. This is
done for the data collected before (A_Scale) and after (A_Tape) removal of all steel substrates using adhesive tape (Figure 9). There is a
clear correlation between surface energy and amount of scaling for each Q-Panel. However, each steel panel result in a different amount
of scaling. This means that the roughness of the substrate still has a significant influence on the scaling tendency.

9
Boersma, Vercauteren, Fischer and Pizzocolo

Figure 8. Scaling on coated steel surfaces for 4 different coatings: the PFTES was used in experiments on glass and does not
show scale inhibition on steel.

Figure 9. Scaling data (A_scale and A_Tape) plotted against the total surface energy of the substrates.

10
 Boersma, Vercauteren, Fischer and Pizzocolo

The scaling behavior depends on the difference between surface properties of the scaling material and the surface. The work of adhesion
can be used to calculate the energy needed to separate a scale crystal from the surface (equation 5). The interfacial energies between the
substrate, scale and water can be calculated using the surface energies of the individual components in air. Furthermore, the work of
adhesion between scale and a rough surface must be modified because of the differences in contact area:

𝑊𝑎𝑑ℎ = 𝛾13 + 𝛾23 ∙ 𝑟2 − 𝛾12 ∙ 𝑟1 (9)

Where r2 is the increase in surface area due to the substrate roughness. When the calcium carbonate crystal is released from the surface,
water will contact a larger surface because of the roughness. So, r2 > 1. r1 is the area of contact between the crystals and the surface.
This area is smaller than the total area, because the crystals can only adhere to the surface by contact with the peaks of the surface
roughness. Consequentially, r1 < 1. We now see that all substrates can be described by a single scaling relation, that depends on the
surface energies and some roughness parameters.

Figure 10. Effective scaling tendency (A_Tape) against the work of adhesion of calcium carbonate on the surfaces, corrected for
the surface roughness.
The scaling experiments on the glass fiber reinforced epoxy tubing were done in a similar way as the steel samples. Collecting the
results it has been found that, surprisingly, the scaling on the Smartcoat coating on GFR epoxy is larger than the same coating on the
steel substrates. The scaling on the Fluoroacrylate coating is lower than on the bare epoxy ring, but can be removed easily by the
adhesive tape.
The reason that the three epoxy coatings (A2:1, A2.5:1 and A2:1.5) were applied on the steel and on the epoxy ring was the assessment
of the influence of scaling on a modified epoxy material. A2:1 is a commercial epoxy, A2.5:1 is made from the same epoxy components
(epoxide and amine), but having a larger amount of epoxide groups. A2:1.5 has a higher amount of amine groups, thus changing the
surface properties. The surface having the higher amount of epoxide groups (A2.5:1) initially shows more scaling, but can be removed
easily. On the other hand, higher amounts of amine (A2:1.5) initially has lower scaling tendency, which cannot all be removed.
When comparing the scaling on the epoxy coatings on the tube and the steel plates, the scaling on the tube is much lower than on the
steel plates. This is not caused by the difference in surface roughness, because the smoother surface of the epoxy would result in a
higher scaling tendency. Another important difference is the hardness or modulus of the substrates. The coated steel substrates are
assumed to have the same modulus as the bare steel, because the thickness of the coatings is very small (~200 nm). If the thickness of
the coating increases, than the surface modulus decreases, because it becomes like a pure polymer. This was shown by a scaling test on
a thick layer of PDMS, where the light absorption of the scale layer was much lower than that of the thin PDMS layer on the steel.
The adhesion force between scale and substrate is determined by the work of adhesion, and by the surface bulk modulus. The bulk
modulus is the most relevant parameter, because the scale crystals are penetrating into the material. The Poisson’s ratio has to be taken
into account as well. The bulk modulus (K) can be calculated from the tensile modulus (E) by:
𝐸
𝐾= (10)
3 ∙ (1− 𝑣 2 )

Where v is the Poisson’s ratio. Typical value of the bulk modulus of steel is 160 GPa, and 50 GPa for epoxy. This means that the work
of adhesion Wadh must be corrected by the square root of the ration between those bulk moduli. This leads to higher work adhesion of the
epoxy samples. Finally, the final relation between scaling, surface energy, surface roughness and surface bulk modulus now becomes:
𝐾 𝐾
𝐴 𝑇𝑎𝑝𝑒 ~ 𝐹 ~ ~ (11)
𝑊𝑎𝑑ℎ 𝛾13 + 𝛾23 ∙ 𝑟2 − 𝛾12 ∙ 𝑟1

Where F is the pull-off force (as intended by Chaudhury, 2005).

11
Boersma, Vercauteren, Fischer and Pizzocolo

Figure 11. Comparison of the scaling on the three steel substrates and the glass fiber reinforced epoxy substrates.
No significant amount of scale is left on the epoxy tube. This is consistent with the findings on the steel panels.
Analyzing these results, it seems to be a transition area between low and high amounts of scale development. When the value
𝐾
is smaller than 6 GPa/(mN/m), low amounts of scale will occur. When this value becomes higher than 6 GPa/(mN/m),
𝛾13 + 𝛾23 ∙ 𝑟2 − 𝛾12 ∙ 𝑟1
scale starts to develop until a value of 10 GPa/(mN/m), where the whole surface is covered with a scale layer. Although this relation
must be furtherly validated for more substrate-scale systems, it is a very promising approach to be used to assess the scaling tendency of
various types of wellbore systems.

4. CONCLUSIONS
In our study we show that the amount of scaling that is deposited on a substrate depends on three parameters: surface energy of the
casing, roughness of the surface and the bulk modulus of the surface. These parameters were used for the development of a single
parameter that can describe the scaling tendency of a surface. A more long term solution for scaling prevention is the implementation of
a casing material that is intrinsically better resistant against scaling. A new class of pipes that may show intrinsically improved scaling
performances is glass fiber reinforced composites. The inner surface of the pipes consist of the polymer matrix of epoxy or polyethylene
and will have a different surface energy than steel pipes. Several sets of experiments were done on steel and polymer matrix casings
under static scaling conditions, where a mixture of CaCl2 and NaHCO3 was placed on top of the surface treated substrates for certain
periods of time. The scale that formed was a combination of heterogeneous nucleated on the surface and homogeneous nucleated in the
bulk solution. The latter scale was removed by rinsing the samples and an adhesive tape after drying to only assess the scale that was
grown on the surface. Moreover, we have shown that the measurement of the light reflection from the scaled surface is a good way of
assessing the amount of scaling. With the results collected in this study, it is possible: predict the scaling tendency under specific
conditions, measure the scale to validate the predictions and assist the development of material for pipelines to lower scaling risks.

REFERENCES
Bargir, S.M., Nanoscale investigation of surface phenomena in the water treatment industry using the atomic force microscope, PhD
Thesis, Cranfield University, 2007.
Carpentier, T.V.J., et al., Evaluation of anti-fouling Surfaces for Prevention of Mineral Scaling in Sub-surface Safety Valves, SPE-
169750-MS, (2014).
Carpentier, T.V.J., et al., Liquid Infused Porous Surfaces for Mineral Fouling Mitigation, J. Coll. Intern. Sci., 444, (2015), 81-86.
Chaudhury, M.K., et al., The influence of elastic modulus and thickness on the release of the soft-fouling green alga Ulva linza from
poly(dimethylsiloxane) model network, Biofouling, 21(1), (2005), 41-48.
Eroini, V., Kinetic Study of Calcium Carbonate Formation and Inhibition by using an in-situ Flow Cell, PhD Thesis, University of
Leeds, (2011).
Good, R.J., et al., A new approach for determining roughness by means of contact angles in solids, Mittal. Festchrijft, (1998), 181-197.
Larson, T.E., Buswell, A.M., Calcium Carbonate Saturation Index and Alkalinity Interpretations, J. Am. Water Works Ass., 34 (11),
(1942), 1667-1679.
Mavredaki, E., Neville, A., Prediction and Evaluation of Calcium Carbonate Deposition at Surfaces, SPE-169796-MS, (2014).
Mitchell, O.E., et al., A model of carbon dioxide dissolution and mineral carbonation kinetics, Proc. R. Soc. A, 466, (2010), 1265-1290.
University of Leeds, www.institutes.engineering.leeds.ac.uk/functional-surfaces/
van Krevelen, D.W., Properties of polymers, Elsevier (Amsterdam), (1997), p239.
12
 Boersma, Vercauteren, Fischer and Pizzocolo

Wang, Z., et al., How and why does scale stick – can the surface be engineered to decrease scale formation and adhesion, SPE 94993,
(2005).
Zhang, W., et al, Preparation of Calcium Carbonate Superfine Powder by Calcium Carbide Residue, Energy Proceedings, 17, (2012),
1635-1640.

13