Journal of Membrane Science 600 (2020) 117859

Contents lists available at ScienceDirect

Journal of Membrane Science

journal homepage: http://www.elsevier.com/locate/memsci

Scaling mitigation in radio-frequency induction heated

membrane distillation
Arezou Anvari , Kartikeya M. Kekre , Avner Ronen *
Civil and Environmental Engineering Department, Temple University, Philadelphia, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Calcium sulfate (CaSO4) scaling is a limiting factor in membrane distillation as a result of concentration po­
Membrane distillation (MD) larization and elevated temperatures. We assessed the influence of radio-frequency magnetic fields, used as a
Calcium sulfate (CaSO4) scaling method of heating the membrane’s surface, as part of a membrane distillation (RF-MD) process. The influence of
Iron oxide functionalized carbon nanotubes
RF heating on CaSO4 scaling is addressed in terms of distillate flux and crystal formation, and the results are
(Fe-CNTs) membranes
Radio-frequency magnetic fields, induction
compared to conventional MD systems. The RF-MD system included a thermally conducting dual-layer mem­
heating (IH) brane containing a magnetic hydrophilic layer based on iron oxide-carbon nanotubes coated on a hydrophobic
membrane. Heating the solution was mainly done directly at the membrane-water interface, and scaling was
assessed at different feed flow velocities using feed solutions containing elevated concentrations of CaSO4 and
NaCl and at ionic composition simulating RO brine. Results show only sporadic small CaSO4 crystals on the
membrane surface following RF-MD, while high concentrations of small crystals were detected at the concentrate
stream exiting the filtration module. Furthermore, the RF-MD system was able to produce enhanced distillate
flux. The scaling mitigation mechanism is discussed in detail and hypothesized to be a result of the high-
frequency movement and collision of the ions in the solution.

1. Introduction of inorganic fouling (i.e., scaling) such as calcium sulfate (CaSO4) and
calcium carbonate (CaCO3) is considered a critical problem [7–9]. In
In the last decade, there is a growing interest in membrane distilla­ addition, scale layers on the membrane surface can intensify the nega­
tion (MD), a thermally driven membrane separation process, for pro­ tive influences of temperature polarization (TP) and CP, thus, reducing
ducing potable water from wastewater, seawater, and saline water [1,2]. the active surface of the membrane for water evaporation and conse­
In contrast to pressure-driven separation processes such as reverse quently decreasing distillate flux [10,11]. Water composition, flow ve­
osmosis (RO), which are impacted by high salt concentration (i.e. os­ locity, degree of supersaturation, and temperature have been shown to
motic pressure), MD systems are less sensitive to salt concentration as influence scale formation rate [12,13] in MD processes.
vapor pressure is less impacted by salinity. Pressure driven desalination, Controlling and removing CaSO4 scaling was shown difficult in
e.g., RO and NF, are restricted by osmotic pressure, leading to high comparison to CaCO3 [14]. While CaCO3 typically precipitates and ad­
required pressure, increased energy demand, and limiting phenomenon heres to heat exchangers and tubing prior to entering the filtration
as concentration polarization (CP) [3,4]. While MD has some advantages module, CaSO4 was shown to form suspended crystals that may reach
over pressure-driven separation when dealing with high salinity (i.e. the membrane surface and deposit [15] on it as scaling. Nghiem et al.
hypersaline) feed, membrane fouling is still considered a limiting phe­ [13] showed that membrane scaling occurred by CaSO4 was compara­
nomenon for MD application as it leads to a decrease in distillate flux tively higher than scaling by CaCO3 and led to distillate flux decline by
and changes in membrane’s surface properties. Fouling was shown to up to 92% following a 40 h experiment [13]. In addition, the elevated
increase membrane hydrophilicity and wettability, eventually leading to feed solution temperature, which is required for MD processes, was
a decrease in membrane selectivity [5,6]. As MD systems are typically shown to increase the rate of CaSO4 crystallization [16] (i.e., inverse
used to reclaim clean water from different degrees of saline solutions (i. solubility at T > 50 � C [14,17]). Up to now, multiple MD studies have
e., saline water to concentrated brine or produced water), accumulation been done on controlling and limiting membrane scaling [13,17–23],

* Corresponding author.
E-mail address: avner.ronen@temple.edu (A. Ronen).

https://doi.org/10.1016/j.memsci.2020.117859
Received 5 November 2019; Received in revised form 14 January 2020; Accepted 17 January 2020
Available online 18 January 2020
0376-7388/© 2020 Elsevier B.V. All rights reserved.
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

leading to control and treatment pathways, including membrane surface concentrations of CaSO4 (1.8, 2.7, and 3.6 g/L) were used as the feed
modification, pre-treatment of the feed solution, and finally membrane solutions in all experiments. In addition, the simulated RO brine
descaling and cleaning through chemical treatment [24–28]. In addition included 35 g/L NaCl, 1.80 g/L CaSO4, 2.91 g/L MgSO4, 4.52 g/L MgCl2,
to these methods, recent studies evaluated the use of electromagnetic and 1.03 g/L KCl [15].
fields (e.g. microwave irradiation and magnetic fields) for heating the The saturation index (SI) of solutions was calculated based on eq. (1)
feed solutions in MD system while assessing their impact on inorganic [36]. Solutions with SI ¼ 0 are at equilibrium while the solutions with
scaling [29–34]. Despite a large number of studies on MD scaling, SI < 0 ​ and ​ SI > 0 are undersaturated and supersaturated, respectively.
limited work has been done on MD scaling of high saline water, such as s
found in RO brines or produced water. Recently, there is a growing SI ¼ logð o Þ (1)
s
understanding that brine discharge has significant environmental im­
pacts, and therefore, it should be treated and managed prior to release. where s is the solubility of CaSO4 at I (i.e., ionic strength), so is the
Since MD is suggested as a possible treatment approach, it is vital to solubility at I ¼ 0 (i.e., solubility product). As I is equal to the molality of
evaluate limiting phenomenon as scaling and derive methods to (NaCl þ 4 � molality of CaSO4), s is the solubility of CaSO4 at any
decrease their impact. particular concertation (see equation (4), section 3.6).
In our previous research, we demonstrated the use of radio-
frequency (RF) electromagnetic fields (i.e., RF induction heating (RF- 2.3. Composite Fe-CNTs/PTFE membrane
IH)) to directly heat the membrane’s surface in a vacuum membrane
distillation (VMD) module [35]. This approach proved to be signifi­ The preparation of iron oxide functionalized multiwall carbon
cantly efficient, leading to high distillate flux and low specific heating nanotubes (Fe-CNTs) has been described in our previous work [35].
energy, especially when treating high saline feed solutions, furthermore, Briefly, FeCl3 and FeSO4 with appropriate ratio were added to the DI
it does not require heating of the bulk feed solution prior to application. water followed by dispersion of CNTs by a horn sonicator (Fisher Sci­
The ‘self-heated’ membranes also resulted in a negligible TP in com­ entific FB705). The solution reached to a pH of 10 by adding 5 M NaOH
parison to an unmodified system. Here, we present a comprehensive solution and then stirring at 85 oC for 1 h. 1 M hydrochloric acid was
investigation of scaling and scaling mitigation behavior of a similar added to the solution to pH of 5 to acidify and wash iron oxide func­
system, based on RF-IH under high salinity feed conditions, specifically tionalized CNTs.
addressing CaSO4 scaling due to its severity. A composite dual-layer Membrane fabrication was according to previous studies using CNTs
hydrophilic-hydrophobic membrane was fabricated through chemical and PVA [37–39]. The CNTs were linked together with a matrix pro­
precipitation and spray coating. The hydrophilic layer contained vided by PVA through covalent bonding of carboxyl groups of the
multiwall carbon nanotubes (CNTs) functionalized with magnetic iron CNT–COOH and hydroxyl groups of PVA. The prepared Fe-CNTs powder
oxide nanoparticles (Fe3O4), and the hydrophobic layer was a com­ (2.3 mg/cm2) dispersed in the DI water and 0.46 mL/cm2 of 1 wt% PVA
mercial polytetrafluoroethylene (PTFE) base support used as it. We solution were sprayed layer by layer on a PTFE membrane. Spray
evaluate and analyze the impact of operating parameters and feed coating was done using an in-house built spray coater including an
concentration on distillate flux and scaling properties (i.e., crystal size airbrush (Master Dual-Action Siphon Feed Airbrush, Model S68) which
and quantity) and compare the RF-based approach (RF-VMD) to a was attached to XY plotter. As the membrane is hydrophobic, the hea­
conventional VMD system (C-VMD). Results demonstrate the significant ting/drying process was done using a heat gun (TCKLIFE, HGP73AC) to
influence of RF-IH on CaSO4 scaling mitigation, leading to an increase in immediately dry each sprayed layer. Then, the Fe-CNTs sprayed layer
distillate flux over time, even when treating hypersaline water. In was stabilized by a crosslinking solution (1.1 vol% Glutaraldehyde and
contrast, similar feed conditions resulted in significant distillate flux 0.55 vol% hydrochloric acid) and heated at 90 � C for 1 h. The prepared
decline over time in C-VMD systems as a result of membrane scaling. In membrane was dried and kept at room temperature.
addition, higher salt concentrations resulted in higher distillate flux,
contrary to C-VMD. To our knowledge, this phenomenon is shown for 2.4. Surface characterization
the first time in MD systems. Thus, making RF-VMD a promising
approach for treating high salt feeds with a high tendency for scaling. In order to image the membranes’ surface morphologies and detect
the formation of scaling on the membrane surface, scanning electron
2. Experimental microscopy (SEM, FEI Quanta 450FEG Scanning Electron Microscope),
and X-ray EDX (FEI Quanta 450FEG) were used, respectively.
2.1. Materials In addition, characterization of fabricated Fe-CNTs powder and
composite membranes was done by transmission electron microscopy
Flat sheet hydrophobic PTFE membranes (pore size 0.2 μm, thickness (TEM, FEI Quanta 450FEG), Fourier-transform infrared spectroscopy
150 μm, and water entry pressure > 45 psi) were purchased from Ster­ (FTIR, PerkinElmer Spectrum 100), and EDX of Fe-CNTs powder. SEM
litech (Sterlitech Inc. Kent, WA). Multi-Wall Carbon Nanotubes func­ images of the cross-section of Fe-CNTs/PTFE membranes have been
tionalized with Carboxylic groups via Plasma treatment (CNTs–COOH) done in our previous work. The results showed a stable Fe-CNTs layer
(outer diameter (30–50 nm), length (10–20μm), and purity > 95%) were with 45 wt. % iron oxide coated on the CNTs, high porosity of 85 � 4%,
purchased from CheapTubes (CheapTubes Inc., Brattleboro, VT). Poly­ and thickness of 30 μm on the PTFE membrane surface [35]. As dis­
vinyl alcohol (PVA) (MW 146–168 kDa), 50 wt% Glutaraldehyde solu­ cussed in our previous study [35], a 30 μm thermally conducting layer
tion, Hydrochloric acid (HCl) 37%, Iron (III) chloride hexahydrate was selected as it was mechanically stable while producing relatively
(FeCl3), Iron (II) sulfate hepatahydrate (FeSO4), Sodium chloride (NaCl), high distillate flux.
and calcium sulfate (CaSO4) were purchased from Sigma Aldrich (t.
Louis, MO). Potassium chloride (KCl) was purchased from MP Bio­ 2.5. Experimental design
medicals LLC (Solon, Ohio). Magnesium chloride (MgCl) and magne­
sium sulfate (MgSO4) were purchased from LabChem (Zelienople, PA The radio-frequency induction heated-vacuum membrane distilla­
16063). tion (RF-VMD), containing an induction heating system (UltraFlex SH-2
with a fixed frequency of 283 kHz), was previously described by Anvari
2.2. Feed solutions et al. (Fig. 1) [35]. The membrane module (effective area of 32 cm2) was
fed with an unheated solution (20 oC). In order to keep the temperature
Deionized (DI) water and a 35 g/L NaCl brine solution with different and feed concentration constant, the concentrate stream was not

2
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

recycled back into the feed tank following MD but was used as a single which are crucial properties of self-heated membranes as they signifi­
pass system (except in section 3.6). In all experiments, the feed solu­ cantly impact on the distillate flux, were changed by spray duration and
tion’s temperature was continuously monitored and controlled using a optimized based on robustness, mechanical stability, and heating rate of
cooling bath. For section 3.6 experiments (concertation factor experi­ Fe-CNTs–PVA layer [35]. As discussed in our previous work [35], the
ments), temperature was kept constant while the salt concentration hydrophilicity of the coated layer was shown to enhance the distillate
increased. In addition, the feed solutions were continuously mixed using flux by providing an improved contact of water with the heated layer,
magnetic stirrer to prevent the precipitation of CaSO4 in the feed solu­ leading to higher vapor pressure in the vicinity of the membrane surface.
tion. The distillate was collected by vacuum at a constant pressure (20 Moreover, the PTFE membranes’ hydrophobicity, porosity, and pore size
kPa) and the condensed distillate was weighed continuously by using an were relatively unchanged after the spray coating [35].
electronic balance. The normalized flux (N.F) was measured based on
the ratio of distillate flux to the initial distillate flux of DI water (J/Jo) 3.2. Distillate flux under scaling conditions
and presented as a measure of CaSO4 scaling. Distillate flux was calcu­
lated using equation (2). In order to characterize the impact of RF-IH on CaSO4 scaling, C-
M VMD and RF-VMD systems were fed with synthetic solutions designed to
J¼ (2) simulate seawater salinity (35 g/L NaCl and 1.8 g/L of CaSO4). CaSO4
A:Δt
concentration was selected to be close or above the saturation limit of
where J (kg/m2.hr) is the distillate flux, M (Kg) is the mass of collected CaSO4 according to specific temperatures. At 25� C and 65 � C, the
permeate, A(m2) is the effective membrane area, and Δt ​ ðhrÞ is the maximum saturation concentrations are 2.4 g/L and 1.5 g/L, respec­
process time [40]. Each experiment was replicated at least three times tively [41,42].
and error bars represent the standard deviation. In addition, results were Scale development of both systems was evaluated according to
analyzed to determine statistical significance (p < 0.01). operational parameters of the MD system, i.e. flux changes over time.
The salinity of condensed and concentrated solutions was continu­ Both systems had similar inlet conditions based on our previous work
ously monitored by using electrical conductivity (EC, Thermo scientific- [35], i.e., flow velocity of 2.33 cm/min and a vacuum pressure of 20
Orion Star A329). In addition, CaSO4 crystals’ number in the concen­ kPa, but different feed temperatures of 65oC for C-VMD and 20 oC for the
trate solutions was measured by imaging a 1mL of solution using a mi­ RF-VMD. While distillate flux changed during our experiments, no
croscope (Zeiss Axio Examiner Z1) and analyzing the images by image decrease in rejection was observed in both systems (NaCl
processing software (ImageJ). rejection>99%). Although the deposited inorganic salt layer was shown
to decrease the hydrophobicity of the PTFE membrane and cause a
3. Results and discussion decrease in rejection [6], the modified dual hydrophilic/hydrophobic
membrane structure was previously shown to reduce the influence of
3.1. Fabricated Fe-CNTs/PTFE membranes scaling on membrane rejection [43].
Distillate flux changes are addressed as normalized flux (N.F) values
As shown in our previous work [35], single or clusters of iron oxide and are presented as a function of time (Fig. 3). The C-VMD system (J0 ¼
nanoparticles with a 10–20 nm diameter were coated on Multiwall CNTs 0.6 � 0.3) resulted in a 50% reduction in N.F over 7 h of experiment, the
(Fig. 2a-enlarged image). IR spectrum analysis confirmed the Fe–O reduction is attributed to the precipitation of CaSO4 crystals on the
bonds of iron oxide at 525 and 545 cm 1 (Fig. 2b). Surface morphology membrane surface due to CP. Scaling occurs as calcium and sulfate ions
and electrical resistance analyses confirmed the stability and high are rejected by the membrane, leading to their supersaturation and
electrical conductivity of CNTs after modification by Fe nanoparticles precipitation on the membrane surface. In addition to CP, the elevated
[35]. Layer-by-layer spray coating was used to deposit the iron temperature was shown to decrease the solubility of CaSO4 due to its
oxide-functionalized carbon nanotubes (Fe-CNTs) and polyvinyl alcohol exothermic formation energy [44]. In all experiments, the average so­
(PVA) on a PTFE membrane, leading to a stable 30μm thick lution temperature was 65 oC, leading to supersaturation and resulting
Fe-CNTs–PVA thermally conducting and magnetic layer (Fig. 2). The in CaSO4 crystals formation on the membrane surface. While scaling led
coated layer was highly hydrophilic (contact angle of 19.5 � 1.5 0) and to the N.F decline in the C-VMD, an opposite behavior was seen in the
contained a 45 wt. % loading of iron oxide (according to SEM-EDX RF-VMD system (J0 ¼ 3 � 0.2), where N.F values increased by up to
analysis [35]). The thickness and mass loading of the Fe-CNTs layer, 115% and stabilized within the first 2 h of heating as presented in our
previous work [35]. The proposed explanation for the RF-MD scaling

Fig. 1. Schematic diagram of the RF-VMD system.

3
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

Fig. 2. SEM cross-section of Fe-CNTs/PTFE membrane, with enlarged TEM of Fe-CNTs layer; (b) FTIR analysis of Fe-CNTs showing Fe–O bonds of iron oxide at 525
and 545 cm 1.

factors should be considered: (1) the influence of the high-frequency

electromagnetic field on Caþ2 and SO-2 4 ions collision, leading to crys­
tallization in the bulk solution; (2) the influence of high frequency
magnetic field on the mobility of Naþ and Cl- ions, leading to damage of
intermolecular bonding of CaSO4 crystals in addition to the damage of
intermolecular bonding of water molecules. Ions exposed to an elec­
tromagnetic field were shown to be influenced by Lorentz force, which is
perpendicular to the magnetic field and enhances linearly according to
the ion’s charge [46,47]. This leads to movement and fluctuation of ions
and to enhance their collisions, resulting in improved crystal nucleation
and the formation of small suspended CaSO4 crystals in the bulk solution
[29,30,45,47–49]. As permeate drag force is influenced by size and mass
[50], the suspended small crystals tend to remain in the solution rather
than reach the membrane surface. Furthermore, Silva and et al. [51]
showed that Ca2þ ions were more hydrated in a solution exposed to a
magnetic field, resulting in the formation of small insoluble crystals
because of the weak adhesion between cations and anions. Small sus­
Fig. 3. N.Fs of C-VMD and RF-VMD systems as a function of time at feed ve­ pended crystals can be more easily removed from the surface of the
locity of 2.33 cm/min and vacuum pressure of 20 kPa for the distillation of 35 membrane by shear as a result of the cross-flow. The formation of
g/L NaCl with 1.8 g/L CaSO4 solution and feed inlet temperatures of 65 oC and crystals with small size under magnetic fields was addressed in several
20 oC, respectively.
studies considering magnetic fields as a method to increase crystalliza­
tion of inorganic salts in the bulk solution [29,30,45,52,53]. In order to
mitigation mechanism is presented in the next section (section 3.3). prove the influence of RF-IH on the formation of crystals in our system’s
EDX and SEM images of the membranes’ surface following distilla­ bulk solution, the concentrate stream was collected and evaluated for
tion (feed solution of 1.8 g/L CaSO4 and 35 g/l NaCl) are presented in the average crystal number and size (Fig. 5a–d). Several mL from the
Fig. 4. SEM image of the C-VMD shows the presence of cubical needle- concentrate were imaged using light microscopy and images were
like shaped crystals, which are hypothesized as the reason for N.F analyzed by an image processing software. As theorized, a higher
decline (Fig. 4a). EDX analysis (Fig. 4c) indicates the presence of cal­ number of suspended crystals were detected following the RF-VMD (13,
cium, oxygen, and sulfate elements, which are the main components of 620 � 1,853 crystals/mm2), while for C-VMD, a significant lower con­
CaSO4 crystals. In contrast, no apparent crystals were detected at the centration of crystals was measured (i.e., 6 times less, 2,380 � 493
surface of the membrane following the RF-VMD processes (Fig. 4 b, d). crystals/mm2). In both cases, suspended crystals had an average size of
EDX data also confirms the presence of iron nanoparticles embedded in 5–10 μm, therefore, we assume larger crystals deposited on the mem­
the CNTs matrix composing the membrane. The data obtained by EDX brane surface.
and SEM images are consistent with the distillate flux results, indicating In addition to the influence of the high-frequency magnetic field on
that the RF-MD distillation process may result in limited or complete calcium and sulfate ions, the mobility of other ions may also contribute
mitigation of CaSO4 scaling. to scaling mitigation. High-frequency magnetic fields were shown to
increase the mobility of Naþ and Cl- ions [54]. High ion mobility and
3.3. Proposed scaling mitigation mechanism magnetic fields were shown to decrease CaCO3 scaling [46]. We assume
a similar phenomenon occurs in our system, where the RF-IH leads to
Previous studies have described the influence of magnetic fields on breaking of crystals’ structure and prevents the formation and deposi­
ions and inorganic salt crystallization along with their fouling behavior tion of CaSO4 scaling.
by several mechanisms including an increase of nucleation and crystal As a result of both factors, scaling formation was limited in the RF-
growth in the bulk solution due to an increase in ion collision frequency VMD system and no distillate N.F decline was observed. Furthermore,
in the magnetic fields and rapid ions reactivity in the water, leading to the N.F increased over time as a result of mainly direct heating of the
the lack of ions in the solution to precipitate [29,30,45]. In the current feed on the surface of the membrane, the negligible TP along with
work, the RF-VMD system exhibits an increase in N.F rather than the limited heat loss, and breaking of water hydrogen-bonds due to the high
decline as seen in C-VMD as a result of CaSO4 scaling. mobility as discussed in our previous work [35]. The temperature of
The results of this study provide a qualitative description of the RF- Fe-CNTs layer exposed to the high frequency electromagnetic field
IH mechanism that leads to scaling mitigation and reduction. In order to increased over time until reaching a steady-state, therefore, the solution
address scaling phenomenon under RF-IH conditions, two plausible

4
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

Fig. 4. Morphological images of used membranes’ surface in (a) C-VMD; (b) RF-VMD systems, and EDX images of selected parts of membranes in (c) C-VMD; (d) RF-
VMD systems for the distillation of 35 g/L NaCl with 1.8 g/L CaSO4 solution. Feed velocity of 2.33 cm/min, vacuum pressure of 20 kPa, and feed inlet temperatures of
65 oC and 20 oC, respectively.

Fig. 5. Light microscopy images of the concentrated flow of (a) C-VMD; (b) RF-VMD, and ImageJ counting images for the concentrated flow of (c) C-VMD; (d)
RF-VMD.

5
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

in contact with the heated layer is heated while passing through it,
leading to higher distillate flux over time until it stabilizes (Fig. 3) [35].
A schematic illustration of the proposed RF-IH impact on CaSO4
crystals deposition and mitigation in comparison to the scaling in the C-
MD system is presented in Fig. 6.
To assess the influence of ions and concentrations on scaling, we
evaluated three solutions and measure the changes in N.F as a function
of time; (S1) a solution containing only 35 g/L of NaCl; (S2) a solution
containing a mixture of 35 g/L NaCl and 1.8 g/L CaSO4, and (S3) a so­
lution containing only 1.8 g/L CaSO4. All the experiments were done at
feed flow velocity of 2.33 cm/min, vacuum pressure of 20 kPa, and feed
inlet temperature of 20 oC. When assessed with the RF-VMD system,
there was no decline in N.F for all solutions (Fig. 7); therefore, scaling
formation can be considered negligible. The limited scale formation is
assumed to be a result of the high-frequency magnetic field as discussed
in the previous section. Furthermore, it should be noticed that for all the
feed solutions, distillate flux increased to values higher than the initial
Fig. 7. Influence of RF-VMD on N.F as a result of solution composition. Feed
flux value (i.e., J > J0, or N.F > 1, where J ¼ up to 7 kg/m2.h and J0 ¼ 3 velocity of 2.33 cm/min, vacuum pressure of 20 kPa, and feed inlet temperature
� 0.2 kg/m2.h) as a result of the heating process through the membrane of 20 oC.
module (i.e., thermally conducting membranes with limited TP) and
impact of the high frequency magnetic field on the mobility of ions,
increase N.F (S1), the presence of calcium and sulfate led to a negative
leading to enhanced water evaporation through breaking hydrogen
impact (S2). This is assumed to be a result of small crystals formed in the
bonds [54]. A maximum value of N.F ¼ 2.5 was measured for the so­
bulk solution as some of the sodium and chloride ions are presumed to
lution containing only NaCl (S1) but, in contrast to the expected increase
break/damage the CaSO4 crystals instead of weakening the water
in N.F with the increase in ionic strength, the presence of both salts (i.e.,
intermolecular forces (hydrogen bonds). The solution containing only
NaCl and CaSO4 in S2) led to a reduced N.F value (N.F ¼ 2). Further­
calcium and sulfate (S3) led to the lowest N.F value as a result of crystal
more, the solution containing only CaSO4 (S3) resulted in the lowest N.F
formation and a relatively low ionic strength in comparison to the pre­
value (N.F ¼ 1.4). While sodium and chloride ions have shown to
vious NaCl pure solution (S1). This demonstrates that both mobility and

Fig. 6. Schematic illustration of (a) CaSO4 scaling in the C-VMD system (b) declined CaSO4 scaling in the RF-VMD system.

6
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

bulk crystallization, which are affected by the magnetic field, have an result of the presence of Caþ2 and SO-2 4 ions and their crystals in the bulk
important role in the low or negligible scaling of RF-VMD systems. solution and their influence on hydrogen bonds breaking by Naþ and Cl-
ions as previously discussed.
SEM images of the membranes following the distillation of solutions
3.4. Impact of CaSO4 concentration
containing high CaSO4 concentration (3.6 g/L CaSO4) are shown in
Fig. 9 and include EDX analysis to confirm crystal composition. While
The influence of CaSO4 concentration at supersaturation on the
imaging of the membranes following C-VMD and RF-VMD shows the
scaling was assessed for both MD systems. CaSO4 concentration in the
presence of CaSO4 crystals, there is a significant difference in crystal size
feed solutions was varied between 1.8 and 3.6 g/L (saturation index (SI)
and concentration between the assessed systems. SEM analysis of the C-
of 0.78–0.8, eq. (1)).
VMD membrane (Fig. 9a) showed multiple large CaSO4 crystals, with
Based on the literature [15,55,56], the RO brine was initially used at
sizes larger than 200 μm. When comparing these results with the pre­
a concentration of 1.8 g/L CaSO4. Higher concentration (i.e. supersat­
vious data regarding the lower CaSO4 concentration (1.8 g/L CaSO4,
urated values such as 3.6 g/L) was assessed to describe extreme condi­
Fig. 4a), we see an increase in crystal size and overall number of crystals.
tions where the RO brine is concentrated as part of a brine management
This confirms the data obtained regarding the influence of CaSO4 con­
approach or when assessing wastewater such as produced water. While
centration and N.F decrease, as seen in Fig. 8a. While for the RF-VMD
the feed solution reached supersaturated conditions, it was continuously
system, no crystals were found on the membrane surface for low
mixed at high velocity (high rpm) to prevent precipitation of CaSO4 in
CaSO4 concentrations (1.8 g/L CaSO4, Fig. 4b), higher concentration of
the feed stank as done by Ge et al. [15] and Hou et al. [18]. The feed was
CaSO4 in the treated solutions resulted in sporadic small crystals with
not filtered to remove preformed crystals prior the experiments as this
sizes ranging between 20 and 180 μm, indicating that scaling formation
will impact the concentration of ions entering the MD system. According
was relatively minor (Fig. 9b). As solutions in both systems had similar
to Yan et al. [56], CMD systems are influenced by different fouling
ionic strength and compositions, and both solutions reached similar
mechanisms for filtered and unfiltered solutions as a result of the formed
average temperatures (65 oC), scaling mitigation is assumed to be a
crystals in the feed. It was shown that for hypersaline supersaturated
function of the magnetic field according to the proposed fouling miti­
solutions (e.g. 3.6 g/L CaSO4), crystallization of CaSO4 was based on
gation mechanisms. Briefly, the high mobility of Naþ and Cl- ions in the
homogenous nucleation in the bulk solution and heterogeneous nucle­
RF magnetic field limits crystal size and prevents them from precipi­
ation on the surface of membrane simultaneously, both impacting the
tating and growing on the membrane surface. Instead, multiple smaller
scaling behavior. The small bulk crystals speed up the homogeneous
crystals are removed at the concentrated stream as a result of shear.
nucleation which can easily migrate and deposit on the membrane
surface. Meanwhile, at the RF-VMD, the heating occurs while the solu­
tion is in the MD cell, therefore, it is expected to reach supersaturation 3.5. Impact of feed flow velocity on distillate flux
conditions in the flow cell and when in contact with the membrane.
Therefore, we did not filter the feed as we wanted to compare both Our previous work showed that in RF-VMD systems, feed flow ve­
systems (C-VMD and RF-VMD) at similar concentrations and assess locity impacts the temperature of the solution at the membrane-water
‘realistic’ conditions, where the system is exposed to both crystal for­ interface, as a result of contact time between liquid and heated mem­
mation mechanisms. brane (i.e. hydraulic retention time, HRT) [35]. Therefore, higher feed
In this study, as expected for supersaturation conditions (SI> 0), the velocity resulted in lower solution temperature and lower distillate flux.
C-VMD system (Fig. 8a) showed a decrease in N.F with the increase of In addition, flow velocity was hypothesized to impact scaling as the
Caþ2 and SO-2 4 concentrations due to crystals formation and growth. solubility of CaSO4 in water decreases with increasing temperature [16].
Following a 420 min filtration experiment, distillate flux was reduced by Moreover, at higher velocities, feed containing ions is exposed to the
more than 50% for all CaSO4 containing solutions with the maximum magnetic field for a shorter time, leading to limited influence of the
reduction for the highest concentration of CaSO4 (i.e., 3.6 g/L of CaSO4). magnetic field on the mobility and crystallization in the bulk. Therefore,
The N.F of the RF-VMD system, however, did not decrease over time there is a tradeoff between high flow rate and scaling. For a high flow
(Fig. 8b) and displayed a similar trend as previously shown (Fig. 7). velocity with low HRT, the solution is not heated as required due to the
Interestingly, while the high concentration of CaSO4 did not decrease N. limited contact time between the liquid and the membrane surface.
F with time, it decreased the maximum N.F value. For the solution Therefore, low distillate flux is expected. On the other hand, as tem­
containing 1.8 g/L of CaSO4, the maximum N.F obtained by the system peratures are relatively low, CaSO4 solubility is higher and there is
was 25% lower than the solution containing only NaCl, and for a higher limited crystallization on the membrane surface. In addition, at short
concentration of 3.6 g/L CaSO4, the N.F was reduced to about 44% of the HRTs, Naþ and Cl- ions are less impacted by the RF-IH and will do
N.F with only NaCl. The change in N.F values is hypothesized to be a limited damage to the formed crystals in addition to limited impact on

Fig. 8. Influence of CaSO4 feed concentration on the N.F of (a) C-VMD; (b) RF-VMD systems. Feed velocity of 2.33 cm/min, vacuum pressure of 20 kPa, and feed inlet
temperatures of 65 oC and 20 oC for C-VMD and RF-VMD, respectively.

7
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

Fig. 9. EDX and SEM images of membranes surface at 3.6g/L CaSO4 feed concentration in: (a) C-VMD; (b) RF-VMD systems.

water evaporation. At low flow velocities, high solution temperatures Therefore, the overall distillate flux is relatively low. On the other hand,
are expected at the membrane-solution interface, and therefore, higher increasing the flow velocity to values higher than 6.75 cm/min led to a
distillate flux is expected. However, at higher temperatures, CaSO4 decrease in solution temperature due to shorter membrane-liquid con­
solubility decreases, and more scaling is expected. In addition, Naþ and tact time, at the same time, the lower temperature limited scaling due to
Cl- ions are expected to be impacted by the RF-IH magnetic field, thus, a higher CaSO4 solubility. As a result, the overall distillate flux is rela­
ion mobility can damage the formed crystals and decrease the scaling tively low. Overall, for RF-VMD systems, high HRT may positively in­
phenomenon. fluence the solution’s temperature and distillate flux but at the same
To validate our hypothesis, the influence of feed flow velocity on time have a negative impact CaSO4 solubility. HRT may also influence
CaSO4 scaling was assessed at a flow velocity range of 2.33–11.25 cm/ the mobility of ions, which was shown to impact crystallization and
min, corresponding to flow rates of 1–5 ml/min and a residence time of distillate flux. For the assessed RF-VMD system, a maximum distillate
206 to 43 s, respectively. Experiments were done using similar solutions flux was obtained at a feed flow velocity of 6.75 cm/min regardless of
as before (1.8 g/L and 3.6 g/L CaSO4 with 35 g/L NaCl). Results revealed the CaSO4 concentrations.
that increasing the flow velocity to the extent of 6.75 cm/min, improved
the N.F by 20% and 25% for 1.8 and 3.6 g/L feed concentrations,
3.6. Impact of concentration factor on distillate flux
respectively (Fig. 10). As expected, at feed velocities lower than 6.75
cm/min (about 3 ml/min), the solution temperature increases to higher
As CaSO4 solubility increases with the increase in ionic strength [57],
values due to the longer membrane-liquid contact time but at the same
we evaluated the influence of high salts concentrations on scaling in
time, the elevated temperature leads to a decreased CaSO4 solubility and
both systems to assess the influence of the RF-IH. The salt rejection was
increased scaling, even under the influence of RF-IH magnetic fields.
determined to be high (about 99%) in both systems. Therefore, the

Fig. 10. Influence of feed flow velocity on the N.F of RF-VMD system at 35 g/L NaCl with (a) 1.8 g/L; (b) 3.6 g/L CaSO4 feed solution.

8
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

concentration factor (CF) was calculated according to eq. (3) [20]. and our calculated values (Fig. 11a) show that increasing the ionic
strength to the extent of about 1.5 M (corresponding to a CF of about 2)
Cf
CF ¼ (3) increases CaSO4 solubility, while higher ionic strengths lead to lower
Ci
solubility or have limited effect. Fig. 11b addresses these values by
where Cf and Ci are the concentration of both salts in the feed solu­ displaying the CaSO4 concentrations and corresponding solubility
tion with recycled concentrate and the initial feed concentration (g/L), values at different CFs.
respectively. The values of CCfi were calculated based on the described For C-MD, the decrease in N.F can be correlated with an increase in
ratio in the previous studies [2,15,18], which are based on the ratio of CF (Fig. 11c) and attributed to the CaSO4 scaling and decrease in partial
cumulative permeate volume to the initial volume of feed solution (R) at vapor pressure [59]. Lower N.F values of salt solutions containing NaCl
any point. Therefore, we used the defined equation by Martinetti et al. in the absence of CaSO4 are only associated with the lower partial vapor
[2] and Hou et al. [18] which is CF ¼ 1 1 R to calculate CF values. pressure (i.e., driving force) because of the high NaCl concertation [59],
The solubility values (s) were calculated at a range of concentration whereas in the presence of CaSO4, the decrease of N.F can be a result of
factors (CFs) and ionic strengths with saturation indexes of 0.7–1 (eq. both lower vapor pressure and scaling at the membrane surface. At CF
(1), experimental section), according to eq. (4) [58]. below 2, both feed solutions (with and without CaSO4) had a similar N.F
decrease trend (Fig. 11c) while in the presence of CaSO4 the N.F is 10%
lower. The difference is attributed to the decrease in partial vapor
0 0
√I B C 2
log ​ s ¼ log ​ so þ 4S þ I I (4)
1 þ Asp √I 2 2 pressure due to the additional CaSO4 ions. At these CF values, the ionic
strength is lower than 1.5 mol/L (as shown in Fig. 11a) and the con­
where I is the ionic strength, which is equal to the molality of (NaCl þ 4 centration of CaSO4 in the feed solution is lower than the saturated
� molality of CaSO4), s is the solubility (mol/L) of CaSO4 at I, so is the solubility (Fig. 11b), leading to limited or no crystallization. At higher
solubility at I ¼ 0 (at 65 oC, ​ so ¼ 34.41 � 10 6 [58]), S is the limiting CF values (above CF of 3), the solution is above the solubility limit of
Debey-Hiickel slope (at 65 oC, S ¼ 0.55 [57]). Asp, B , and C are
0 0
CaSO4, leading to 20% more flux reduction in comparison to the solution
empirical parameters according to temperatures [58], where at 65 oC, containing only NaCl. This is hypothesized to be a result of scaling on the
membrane surface. The influence of CaSO4 scaling and supersaturation
Asp ¼ 1.564, ​ B ¼ 0, and C ¼ 0:009.
0 0

can be seen in CF values between two and four, resulting in 55% flux
To evaluate the impact of the RF-IH at higher salt concentrations
reduction in comparison to the solution containing only NaCl which
such as found in concentrated RO brine or produced water and to assess
suffered from 40% flux reduction. Our results are consistent with data
how increased salinity may influence N.F and scale formation, the
from the literature [15]. Ju et al. showed that an increase in ionic
concentrate stream was recycled back into the feed tank following MD,
strength to more than 1.5mol/L (corresponding to the CF more than 3.5)
thus increasing the ions concentrations in the feed containing 35 g/L
led to CaSO4 scaling [15]. As expected, SEM imaging of the membrane
NaCl with/without 1.8 g/L CaSO4. Results of previous studies [15,57]
following C-VMD with CaSO4 at CF ¼ 4 exhibited very large CaSO4

Fig. 11. (a) Impact of ionic strength on the CaSO4 solubility; (b) Impact of CF on the solubility and concentration of CaSO4; (c) Impact of CF on N.F of C-VMD and
SEM image of membrane surface at CF ¼ 4 in the C-VMD; (d) Impact of CF on N.F of RF-VMD and SEM image of membrane surface at CF ¼ 4 in the RF-VMD.

9
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

crystals (with sizes ranging between 300 and 950 μm) on the membrane similar to the solution containing only NaCl, while a solution containing
surface (Fig. 11c). The large crystals and scaling layer on the membrane Naþ, Cl- and CaSO4 have lower N.F values. Accordingly, ions that will
surface can also increase the concentration boundary layer resistance on not contribute to precipitants formation, including additional Kþ, Cl-,
the membrane surface in addition to blocking the membrane pores to and Mgþ may enhance the distillate flux by a similar mechanism as
prevent vapor transport, leading to a notable decrease in the N.F [56]. described for Naþ and Cl- ions. Among all existent salts in the RO brine
In contrast, for the RF-VMD system, high-frequency magnetic fields solution, the tendency of CaSO4 to crystallize and form scaling is the
increased N.F values as previously shown [35]; this is especially seen highest as its solubility in water is very low. At 25 oC, the saturation
when the solution contains high salinity without CaSO4. Therefore, solubility of KCl, MgCl2, MgSO4, and NaCl in the water are 350.5, 540.6,
distillate flux was shown to increase with the increase in CF (Fig. 11d) 360, and 358.9 g/L, respectively, which are much higher than the sol­
and the N.F reached values of about 400% when the solution was ubility concentration of CaSO4 (2.4 g/L). As previously explained, we
concentrated four times (i.e., CF of four). This phenomenon was hy­ have not addressed calcium carbonate fouling as it was shown to form
pothesized to be a result of the high-frequency vibration of Naþ and Cl- scaling prior to the MD process, mainly by precipitation on heat ex­
ions, leading to a decrease in water molecules’ bonds [54] as previously changers and pipes [14]. In addition, the solubility of all salts in RO
discussed. In contrast to C-VMD, where CaSO4 ions in the feed led to a brine (except CaSO4 and CaCO3) increases with the increase in tem­
decrease in distillate flux due to scaling, the RF-IH process was shown to perature. Therefore, exposure and direct heating of RO brine by RF-IH is
mitigate membrane scaling, leading to constant and stabilized N.F over assumed to mainly form CaSO4 crystals. Furthermore, additional ions
time (Fig. 8b). A decrease of about 40% in N.F was measured for solu­ such as Kþ, Cl-, Naþ, and Mg2þ may have increased mobility under the
tions with CaSO4 in comparison to the solution containing only NaCl at RF-IH field and therefore may lead to higher distillate flux value and
high CF due to the negative impact of Caþ2 and SO-2 4 ions. Interestingly, physical breakdown of existing CaSO4 crystals. Images of the membrane
even at CF larger than two, the reduced value of N.F for feed solutions following RF-VMD using RO brines solutions (Fig. 12b), revealed limited
with CaSO4 is approximately constant, inferring that the ionic strength sporadic small CaSO4 crystals.
has a negligible impact on scaling procedure under the influence of an
RF-IH magnetic field. SEM images of the membranes following RF-VMD 4. Conclusions
experiments with CaSO4 solutions (at CF ¼ 4) show that although some
CaSO4 crystals can be detected at the membrane surface, they are rela­ CaSO4 scaling formation and mitigation were assessed in a novel RF-
tively small (<100 μm) and limited in number (Fig. 11d). VMD system using thermally conducting membranes modified by car­
As the obtained results were encouraging, we evaluated the system’s bon nanotubes functionalized with magnetic iron nanoparticles. In
ability to handle simulated RO brine, containing multiple ions at contrast to C-VMD, where CaSO4 crystallization and fouling led to a
elevated concentrations. decrease in distillate flux, RF-VMD showed limited fouling and no flux
decline. Membrane analysis following distillation showed only sporadic
small CaSO4 crystals on the membrane surface, even when treating so­
3.7. Treatment and scaling behavior of simulated RO brine lutions with high ionic strength. Furthermore, high concentrations of
small crystals were detected at the concentrate stream, indicating
Assessing scaling mitigation of RO brines is gaining interest as RO is limited deposition on the membrane surface. In addition, the presence of
currently the most common desalination approach and there is a additional ions typically found in concentrated RO brine leads to
growing understanding of the need for RO brine management. We used a enhanced distillate flux and limited scaling. The influence of different
synthetic solution containing 35 g/L NaCl, 1.80 g/L CaSO4, 2.91 g/L ions was assessed, showing a decrease in distillate flux as a function of
MgSO4, 4.52 g/L MgCl2, and 1.03 g/L KCl, with a total salinity of 46.5 g/ CaSO4 concentration and an increase in distillate flux as a result of
L to simulate RO brine according to Ge et al. [15]. Similar values were increased Naþ and Cl- ions. Results are attributed to the RF-IH magnetic
used for describing RO brine in the literature [55,56]. The RF-IH mag­ fields used for directly heating the membrane surface, leading to high
netic field is assumed to influence RO brine solutions containing mul­ mobility and collision of ions as a result of Lorentz force. These are
tiple ions in addition to Naþ, Cl-, Caþ2, and SO-2
4 . Fig. 12a presents N.F hypothesized to enhance distillation flux and diminish scaling by
values of synthetic RO in comparison to saline solutions with/without weakening intermolecular water bonds and damaging the formed CaSO4
CaSO4. Results indicate that the N.F values of the RO brine are relatively

Fig. 12. (a) Impact of RF-VMD on N.F of RO brine compared to saline solutions with and without CaSO4; (b) EDX and SEM images of membranes surface of RO brine
in RF-VMD system.

10
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

crystals. The RF-VMD system was also assessed as a function of feed flow performance, efficiency, and salt rejection, Separ. Purif. Technol. 170 (2016)
78–91, https://doi.org/10.1016/j.seppur.2016.06.028.
velocity, showing optimized results at low-velocity values (6.75 cm/
[20] T. Zou, X. Dong, G. Kang, M. Zhou, M. Li, Y. Cao, Fouling behavior and scaling
min). Following these encouraging results, we suggest that RF-VMD may mitigation strategy of CaSO4 in submerged vacuum membrane distillation,
possibly be utilized to substitute C-VMD processes when treating solu­ Desalination 425 (2018) 86–93, https://doi.org/10.1016/j.desal.2017.10.015.
tion with high salinity and scaling tendency. [21] H.C. Duong, M. Duke, S. Gray, P. Cooper, L.D. Nghiem, Membrane scaling and
prevention techniques during seawater desalination by air gap membrane
distillation, Desalination 397 (2016) 92–100, https://doi.org/10.1016/j.
Declaration of competing interest desal.2016.06.025.
[22] P. Zhang, P. Kn€ otig, S. Gray, M. Duke, Scale reduction and cleaning techniques
during direct contact membrane distillation of seawater reverse osmosis brine,
The authors declare that they have no known competing financial Desalination 374 (2015) 20–30, https://doi.org/10.1016/j.desal.2015.07.005.
interests or personal relationships that could have appeared to influence [23] M. Gryta, Long-term performance of membrane distillation process, J. Membr. Sci.
the work reported in this paper. 265 (2005) 153–159, https://doi.org/10.1016/j.memsci.2005.04.049.
[24] A. Razmjou, E. Arifin, G. Dong, J. Mansouri, V. Chen, Superhydrophobic
modification of TiO2 nanocomposite PVDF membranes for applications in
CRediT authorship contribution statement membrane distillation, J. Membr. Sci. 415–416 (2012) 850–863, https://doi.org/
10.1016/j.memsci.2012.06.004.
[25] H.K. Shon, S.H. Kim, S. Vigneswaran, R. Ben Aim, S. Lee, J. Cho, Physicochemical
Arezou Anvari: Writing - original draft, Conceptualization, Meth­ pretreatment of seawater: fouling reduction and membrane characterization,
odology, Investigation, Visualization. Kartikeya M. Kekre: Methodol­ Desalination 238 (2009) 10–21, https://doi.org/10.1016/j.desal.2008.01.030.
ogy, Investigation. Avner Ronen: Supervision, Writing - review & [26] J. Wang, D. Qu, M. Tie, H. Ren, X. Peng, Z. Luan, Effect of coagulation
pretreatment on membrane distillation process for desalination of recirculating
editing, Conceptualization, Methodology. cooling water, Separ. Purif. Technol. 64 (2008) 108–115, https://doi.org/
10.1016/j.seppur.2008.07.022.
References [27] L. Tang, A. Iddya, X. Zhu, A.V. Dudchenko, W. Duan, C. Turchi, J. Vanneste, T.
Y. Cath, D. Jassby, Enhanced flux and electrochemical cleaning of silicate scaling
on carbon nanotube-coated membrane distillation membranes treating geothermal
[1] J. Larsson, Membrane Distillation Study of Membrane Distillation Unit at the Duke
brines, ACS Appl. Mater. Interfaces 9 (2017) 38594–38605, https://doi.org/
Smart Home, Duke in Collaboration with KTH, 2010. http://smarthome.duke.ed
10.1021/acsami.7b12615.
u/sites/smarthome.duke.edu/files/Larsson_MDreport.pdf.
[28] M.S. Humoud, S. Roy, Scaling reduction in carbon nanotube-immobilized
[2] C.R. Martinetti, A.E. Childress, T.Y. Cath, High recovery of concentrated RO brines
membrane during membrane distillation, Water 11 (2019) 2588, https://doi.org/
using forward osmosis and membrane distillation, J. Membr. Sci. 331 (2009)
10.3390/w11122588.
31–39, https://doi.org/10.1016/j.memsci.2009.01.003.
[29] A. Shahryari, M. Pakshir, Influence of a modulated electromagnetic field on fouling
[3] M. Khayet, C. Cojocaru, M.C. García-Payo, Experimental design and optimization
in a double-pipe heat exchanger, J. Mater. Process. Technol. 203 (2008) 389–395,
of asymmetric flat-sheet membranes prepared for direct contact membrane
https://doi.org/10.1016/j.jmatprotec.2007.10.048.
distillation, J. Membr. Sci. 351 (2010) 234–245, https://doi.org/10.1016/j.
[30] X. Xiaokai, Research on the electromagnetic anti-fouling technology for heat
memsci.2010.01.057.
transfer enhancement, Appl. Therm. Eng. 28 (2008) 889–894, https://doi.org/
[4] S. Adham, A. Hussain, J.M. Matar, R. Dores, A. Janson, Application of Membrane
10.1016/j.applthermaleng.2007.07.006.
Distillation for desalting brines from thermal desalination plants, Desalination 314
[31] M.S. Humoud, W. Intrchom, S. Roy, S. Mitra, Reduction of scaling in microwave
(2013) 101–108, https://doi.org/10.1016/j.desal.2013.01.003.
induced membrane distillation on a carbon nanotube immobilized membrane,
[5] L.D. Tijing, Y. Chul, J. Choi, S. Lee, S. Kim, H. Kyong, Fouling and its control in
Environ. Sci. Water Res. Technol. 5 (2019) 1012–1021, https://doi.org/10.1039/
membrane distillation — a review, J. Membr. Sci. 475 (2015) 215–244, https://
c9ew00153k.
doi.org/10.1016/j.memsci.2014.09.042.
[32] Z. Ji, J. Wang, D. Hou, Z. Yin, Z. Luan, Effect of microwave irradiation on vacuum
[6] D.M. Warsinger, J. Swaminathan, E. Guillen-Burrieza, H.A. Arafat, J.H. Lienhard V,
membrane distillation, J. Membr. Sci. 429 (2013) 473–479, https://doi.org/
Scaling and fouling in membrane distillation for desalination applications: a
10.1016/j.memsci.2012.11.041.
review, Desalination 356 (2015) 294–313, https://doi.org/10.1016/j.
[33] Z. Ji, J. Wang, Z. Yin, D. Hou, Z. Luan, Effect of microwave irradiation on typical
desal.2014.06.031.
inorganic salts crystallization in membrane distillation process, J. Membr. Sci. 455
[7] M. Gryta, Alkaline scaling in the membrane distillation process, Desalination 228
(2014) 24–30, https://doi.org/10.1016/j.memsci.2013.12.064.
(2008) 128–134, https://doi.org/10.1016/j.desal.2007.10.004.
[34] S. Roy, M.S. Humoud, W. Intrchom, S. Mitra, Microwave-Induced desalination via
[8] F. He, K.K. Sirkar, J. Gilron, Studies on scaling of membranes in desalination by
direct contact membrane distillation, ACS Sustain. Chem. Eng. 6 (2018) 626–632,
direct contact membrane distillation: CaCO3 and mixed CaCO3/CaSO4 systems,
https://doi.org/10.1021/acssuschemeng.7b02950.
Chem. Eng. Sci. 64 (2009) 1844–1859, https://doi.org/10.1016/j.
[35] A. Anvari, K.M. Kekre, A. Azimi Yancheshme, Y. Yao, A. Ronen, Membrane
ces.2008.12.036.
distillation of high salinity water by induction heated thermally conducting
[9] S. Srisurichan, R. Jiraratananon, A.G. Fane, Mass transfer mechanisms and
membranes, J. Membr. Sci. 589 (2019), 117253, https://doi.org/10.1016/j.
transport resistances in direct contact membrane distillation process, J. Membr.
memsci.2019.117253.
Sci. 277 (2006) 186–194, https://doi.org/10.1016/j.memsci.2005.10.028.
[36] J. Bizzozero, C. Gosselin, K.L. Scrivener, Expansion mechanisms in calcium
[10] E. Curcio, X. Ji, G. Di Profio, A.O. Sulaiman, E. Fontananova, E. Drioli, Membrane
aluminate and sulfoaluminate systems with calcium sulfate, Cement Concr. Res. 56
distillation operated at high seawater concentration factors: role of the membrane
(2014) 190–202, https://doi.org/10.1016/j.cemconres.2013.11.011.
on CaCO3 scaling in presence of humic acid, J. Membr. Sci. 346 (2010) 263–269,
[37] C.F. de Lannoy, D. Jassby, D.D. Davis, M.R. Wiesner, A highly electrically
https://doi.org/10.1016/j.memsci.2009.09.044.
conductive polymer-multiwalled carbon nanotube nanocomposite membrane,
[11] L. Wang, B. Li, X. Gao, Q. Wang, J. Lu, Y. Wang, S. Wang, Study of membrane
J. Membr. Sci. (2012) 415–416, https://doi.org/10.1016/j.memsci.2012.05.061,
fouling in cross-flow vacuum membrane distillation, Separ. Purif. Technol. 122
718–724.
(2014) 133–143, https://doi.org/10.1016/j.seppur.2013.10.031.
[38] A.V. Dudchenko, J. Rolf, K. Russell, W. Duan, D. Jassby, Organic fouling inhibition
[12] M. Gryta, Calcium sulphate scaling in membrane distillation process, Chem. Pap.
on electrically conducting carbon nanotube-polyvinyl alcohol composite
63 (2009) 146–151, https://doi.org/10.2478/s11696-008-0095-y.
ultrafiltration membranes, J. Membr. Sci. 468 (2014) 1–10, https://doi.org/
[13] L.D. Nghiem, T. Cath, A scaling mitigation approach during direct contact
10.1016/j.memsci.2014.05.041.
membrane distillation, Separ. Purif. Technol. 80 (2011) 315–322, https://doi.org/
[39] A.V. Dudchenko, C. Chen, A. Cardenas, J. Rolf, D. Jassby, Frequency-dependent
10.1016/j.seppur.2011.05.013.
stability of CNT Joule heaters in ionizable media and desalination processes, Nat.
[14] F. He, J. Gilron, H. Lee, L. Song, K.K. Sirkar, Potential for scaling by sparingly
Nanotechnol. 12 (2017) 557–563, https://doi.org/10.1038/nnano.2017.102.
soluble salts in crossflow DCMD, J. Membr. Sci. 311 (2008) 68–80, https://doi.org/
[40] A. Anvari, A.A. Yancheshme, F. Rekaabdar, M. Hemmati, M. Tavakolmoghadam,
10.1016/j.memsci.2007.11.056.
A. Safekordi, PVDF/PAN blend membrane: preparation, characterization and
[15] J. Ge, Y. Peng, Z. Li, P. Chen, S. Wang, Membrane fouling and wetting in a DCMD
fouling analysis, J. Polym. Environ. 25 (2017) 1348–1358, https://doi.org/
process for RO brine concentration, Desalination 344 (2014) 97–107, https://doi.
10.1007/s10924-016-0889-x.
org/10.1016/j.desal.2014.03.017.
[41] H. Geng, J. Wang, C. Zhang, P. Li, H. Chang, High water recovery of RO brine using
[16] D.M. Warsinger, E.W. Tow, J. Swaminathan, J.H. Lienhard V, Theoretical
multi-stage air gap membrane distillation, Desalination 355 (2015) 178–185,
framework for predicting inorganic fouling in membrane distillation and
https://doi.org/10.1016/j.desal.2014.10.038.
experimental validation with calcium sulfate, J. Membr. Sci. 528 (2017) 381–390,
[42] W.L. Marshall, R. Slusher, E.V. Jones, Aqueous systems at high temperature. XIV.
https://doi.org/10.1016/j.memsci.2017.01.031.
Solubility and thermodynamic relationships for CaSO4 in NaCl-H2O solutions from
[17] G. Naidu, S. Jeong, Y. Choi, E. Jang, T.M. Hwang, S. Vigneswaran, Application of
40⋅ to 200⋅ C., 0 to 4 molal NaCl, J. Chem. Eng. Data 9 (1964) 187–191, https://
vacuum membrane distillation for small scale drinking water production,
doi.org/10.1021/je60021a011.
Desalination 354 (2014) 53–61, https://doi.org/10.1016/j.desal.2014.09.026.
[43] S. Roy, M. Bhadra, S. Mitra, Enhanced desalination via functionalized carbon
[18] D. Hou, Z. Wang, G. Li, H. Fan, J. Wang, H. Huang, Ultrasonic assisted direct
nanotube immobilized membrane in direct contact membrane distillation, Separ.
contact membrane distillation hybrid process for membrane scaling mitigation,
Purif. Technol. 136 (2014) 58–65, https://doi.org/10.1016/j.seppur.2014.08.009.
Desalination 375 (2015) 33–39, https://doi.org/10.1016/j.desal.2015.07.018.
[19] J.A. Bush, J. Vanneste, T.Y. Cath, Membrane distillation for concentration of
hypersaline brines from the Great Salt Lake: effects of scaling and fouling on

11
A. Anvari et al. Journal of Membrane Science 600 (2020) 117859

[44] K.U.G. Raju, G. Atkinson, The thermodynamics of “scale” mineral solubilities. 3. [52] J. Li, J. Liu, T. Yang, C. Xiao, Quantitative study of the effect of electromagnetic
Calcium sulfate in aqueous NaCl, J. Chem. Eng. Data 35 (1990) 361–367, https:// field on scale deposition on nanofiltration membranes via UTDR, Water Res. 41
doi.org/10.1021/je00061a038. (2007) 4595–4610, https://doi.org/10.1016/j.watres.2007.06.041.
[45] H.E.L. Madsen, Influence of magnetic field on the precipitation of some inorganic [53] M.A. Salman, M. Safar, G.A. Nuwaibit, The effect of magnetic treatment on
salts, J. Cryst. Growth 152 (1995) 94–100, https://doi.org/10.1016/0022-0248 retarding scaling deposition, TOJSAT 5 (2015) 62–77.
(95)00103-4. [54] K.T. Chang, C.I. Weng, An investigation into the structure of aqueous NaCl
[46] M. Rouina, H.R. Kariminia, S.A. Mousavi, E. Shahryari, Effect of electromagnetic electrolyte solutions under magnetic fields, Comput. Mater. Sci. 43 (2008)
field on membrane fouling in reverse osmosis process, Desalination 395 (2016) 1048–1055, https://doi.org/10.1016/j.commatsci.2008.02.020.
41–45, https://doi.org/10.1016/j.desal.2016.05.009. [55] J.P. Mericq, S. Laborie, C. Cabassud, Vacuum membrane distillation of seawater
[47] N.S. Zaidi, J. Sohaili, K. Muda, M. Sillanp€ aa
€, Magnetic field application and its reverse osmosis brines, Water Res. 44 (2010) 5260–5273, https://doi.org/
potential in water and wastewater treatment systems, Separ. Purif. Rev. 43 (2014) 10.1016/j.watres.2010.06.052.
206–240, https://doi.org/10.1080/15422119.2013.794148. [56] Z. Yan, H. Yang, F. Qu, H. Yu, H. Liang, G. Li, J. Ma, Reverse osmosis brine
[48] A. Luo, N. Lior, Critical review of membrane distillation performance criteria, treatment using direct contact membrane distillation: effects of feed temperature
Desalin. Water Treat. 57 (2016) 20093–20140, https://doi.org/10.1080/ and velocity, Desalination 423 (2017) 149–156, https://doi.org/10.1016/j.
19443994.2016.1152637. desal.2017.09.010.
[49] C. Piyadasa, H.F. Ridgway, T.R. Yeager, M.B. Stewart, C. Pelekani, S.R. Gray, J. [57] J.A.M. Meijer, G.M. Van Rosmalen, Solubilities and supersaturations of calcium
D. Orbell, The application of electromagnetic fields to the control of the scaling and sulfate and its hydrates in seawater, Desalination 51 (1984) 255–305, https://doi.
biofouling of reverse osmosis membranes - a review, Desalination 418 (2017) org/10.1016/0011-9164(84)87002-2.
19–34, https://doi.org/10.1016/j.desal.2017.05.017. [58] W.L. Marshall, R. Slusher, Thermodynamics of calcium sulfate dihydrate in
[50] G.Z. Ramon, E.M.V. Hoek, On the enhanced drag force induced by permeation aqueous sodium chloride solutions, 0-110� , J. Phys. Chem. 70 (1966) 4015–4027,
through a filtration membrane, J. Membr. Sci. (2012) 392–393, https://doi.org/ https://doi.org/10.1021/j100884a044.
10.1016/j.memsci.2011.10.056, 1–8. [59] M. Qtaishat, T. Matsuura, B. Kruczek, M. Khayet, Heat and mass transfer analysis in
[51] I.B. Silva, J.C. Queiroz Neto, D.F.S. Petri, The effect of magnetic field on ion direct contact membrane distillation, Desalination 219 (2008) 272–292, https://
hydration and sulfate scale formation, Colloids Surfaces A Physicochem. Eng. Asp. doi.org/10.1016/j.desal.2007.05.019.
465 (2015) 175–183, https://doi.org/10.1016/j.colsurfa.2014.10.054.

12