Desalination Technology for Management Water Issue in California. 15 pages

Desalination 360 (2015) 97–107

Contents lists available at ScienceDirect

Desalination

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

Synoptic analysis of direct contact membrane distillation performance in Qatar: A case study

Ahmad Kayvani Fard a,⁎, Yehia M. Manawi a, Tarik Rhadfi a, Khaled A. Mahmoud a, Majeda Khraisheh b, Farid Benyahia b

a Qatar Environment and Energy Research Institute, Qatar Foundation, PO Box 5825, Doha, Qatar b Department of Chemical Engineering, Qatar University, P.O. Box 2713, Doha, Qatar

H I G H L I G H T S

• The feasibility of DCMD to desalinate saline solution is studied. • Efficiency of DCMD is evaluated using GOR analysis. • Energy efficiency of DCMD is investigated against different operating parameters. • DCMD is capable of producing highly pure water at temperature as low as 70 °C.

⁎ Corresponding author. E-mail address: [email protected] (A. Kayvani Fard).

http://dx.doi.org/10.1016/j.desal.2015.01.016 0011-9164/© 2015 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history: Received 27 November 2014 Received in revised form 12 January 2015 Accepted 13 January 2015 Available online 20 January 2015

Keywords: Desalination Membrane distillation Direct contact membrane distillation Energy efficiency Gain Output Ratio Rejected brine Water quality

The aim of this study is to evaluate a bench scale direct contact membrane distillation (DCMD) performance using flat sheet polytetrafluoroethylene (PTFE) membrane at different inlet flow rates, temperatures, and salinity com- position. The effect of the various operating conditions on the quality of the distillate water produced was inves- tigated by Inductively Coupled Plasma/Optical Emission Spectrometry (ICP/OES) and ion chromatography (IC). Five different feed solutions were studied such as the Arabian Gulf seawater, rejected brine from Qatari thermal desalination plants and artificial brine. A permeate flux of 35.6 LMH can be produced at temperature difference of 50 °C between hot and cold sides. The thermal energy efficiency of the system was analyzed and reached up to 22%. The flow mode and turbulence promoter in feed channels were found to have a very important effect on the flux and energy efficiency. It was also noted that using spacers in the flow channels increase the distillate flux by more than 51% as compared with a spacer-free system. The investigation has shown that the highest Gain Output Ratio (GOR) value can be obtained when using less saline feed solution at the highest feed temper- ature, lowest permeate temperature and lowest flow rate. Finally, the salt rejection rate throughout the conduct- ed tests was very high (N99.9%) and almost independent of any studied operational parameters. Eventually, DCMD has proven to be a feasible and effective technology capable of consistently producing high quality distil- late from a very high salinity feed even with substantial quality difference compared to other desalination methods such as RO and MSF.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

Desalination is the main source of drinking water in the countries of the Gulf Cooperation Council (GCC). Among all the desalination plants worldwide, 50% are in the GCC region alone. Total desalination capacity of GCC in 2012 was 11 million cubic meter per day of fresh water forming around 45% of global desalinated water production [1]. The GCC countries use different desalination technologies such as multi stage flash (MSF), multi effect distillation (MED) and reverse osmosis (RO) to satisfy the demand and mitigate drought condition.

Due to the increasing demand of water especially from private and industrial sectors, production capacities for desalinated water still has to be increased to address the increasing demand for freshwater. Fur- ther water management measures including plans for better reuse of rejected water from desalination plants are required. RO is currently the most reliable desalination technology; however there are still major challenges to be resolved, such as membrane fouling, relatively low recovery, and relatively low removal of low molecular weight and uncharged contaminants [2]. Membrane distillation (MD) is an emerg- ing desalination technology which differs from other membrane tech- nologies in that the driving force for desalination is the difference in vapor pressure of water across the hydrophobic membrane, rather than total pressure. MD is a separation process using a porous

Table 1 GOR values for different desalination technologies.

Technology Heating steam temperature, °C TBT, °C GOR Reference

MSF 130 120 10–12 [9] 100 90 7–9 80 70 4–5

MED-TVC 150 70 12–15 120 70 10–12

LT-MED 90 80 10–13 80 70 8–10 70 60 6–8

AGMD – 0.9 [10] 0.97 [11] 0.8 [12]

VMD – 0.57 [13] 0.85 [14]

DCMD – 0.17 [13] 4.1 [15] 1.4 [16]

98 A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

hydrophobic membrane and its mechanism is divided into three stages: evaporation of water in feed side, transport of water vapor molecules to permeate side through hydrophobic membrane pore and condensation of vapor molecules in cold permeates side.

There are four different configurations of MD systems depending on the vapor pressure difference (between the hot and cold side) and the mode of vapor condensation, namely: direct contact membrane distilla- tion (DCMD), air gap membrane distillation (AGMD), sweeping gas membrane distillation (SGMD), and vacuum membrane distillation (VMD) [2–4]. In the DCMD configuration (Fig. 1), the temperature of the feed solution needs to be higher than that of the permeate solution to create a driving force for vapor transport across the hydrophobic micro-porous membrane. The feed temperature range required to evap- orate the water is not very large (40–80 °C) and this point makes the MD process more energy-efficient and open to waste heat utilization (using alternative heating and energy resources such as industrial waste heat).

A temperature difference as low as 10–20 °C between the warm and cold streams is sufficient to produce distillate water [5]. Industrial waste heat is a term used to describe the heat generated by fuel combustion or chemical reaction for the purpose of heating a process stream or reactor and now become useless and must be discharged to the environment. Low grade waste heat stream with temperature range below 200 °C and above 70 °C is required in order to elevate the feed temperature to a range of 50–80 °C [6]. Therefore, one of the major advantages of MD over other conventional processes such as RO or thermal processes is the lower operating temperature, which directly affects the energy consumption and by default, lowers the environmental impact [7].

According to Einav et al. [8], the amount of electricity necessary to produce 1 m3 of water from RO varies between 3.5–4.5 kWh/m3. Since there is no direct use of electricity in thermal desalination process, the efficiency of the heat used in thermal desalination is evaluated using the Gain Output Ratio (GOR). GOR is defined as the ratio between the amounts of water produced per unit mass of dry saturated steam sup- plied to the system.

GOR ¼ md ms

ð1Þ

where md is the mass flow rate of distillate and ms is the mass flow rate of saturated steam. GOR values of MD systems are significantly lower than their counterparts in the conventional desalination technol- ogies. The reason is that MD technology has a low flux due to the heat and mass resistance formed by the membrane and the surrounding boundary layers. The low mass flux could also be due to the operation at low temperatures when compared to conventional desalination which reduces the driving force of the process itself. Table 1 lists the GOR values of different thermal desalination technologies and temper- atures of the heating steam with different configurations of MD systems as reported by different literatures.

The GOR values change according to the configuration of the MD sys- tem employed. AGMD and VMD have low GOR values while DCMD have higher values. Summers [17] have reported a GOR value of 4.1 which seems high relative to other GOR values of DCMD. Their system was a

Hot Channel

Cold Channel

Hot Feed Out (Brine)

Cold Feed In

Hydrophobic Membrane

Permeate Flow Direc�on

Hot Feed In

Cold Feed Out

Fig. 1. Schematic diagram of counter flow DCMD process.

multistage consisting of 8 stages. The optimum GOR value was 0.9 for a single stage system employing membrane distillation as a desalination technology and the Arabian Gulf water as feed. Some other studies have listed GOR values at different feed inlet hot temperatures without tack- ling the other operating conditions such as the permeate temperature, flow rate and salinity content.

Since there is no direct use of steam as source of heating to the MD, alternatively, the energy efficiency or thermal efficiency in MD can be defined as ratio of latent heat for evaporation of feed solution over total input heat to the system in addition to the conventional GOR calcu- lation method (Eq. (1)) [17]:

EE ¼ QL QT

ð2Þ

where, QT is the total heat input to the system and QL is the latent heat needed for evaporation. To date, there is little work done on energy efficiency of MD and especially DCMD in the open literature and the most of the works are focused on temperature polarization and heat transfer in MD without any explicit reference to energy efficiency. Data on GOR analysis of DCMD is limited in literature and most of the works are done on overall heat efficiency of the system or temperature polarization effect. The reported GOR in literature are either overestimated or analyzed at limited laboratory parameters. Yao et al. [18] studied the energy efficiency of AGMD with hollow fiber mem- brane. Geng et al. [19] used mathematical modeling to study AGMD also to find GOR of such system by finding trans-membrane tempera- ture and not overall system temperature. Lin et al. [20] studied module scale thermodynamic analysis of DCMD that includes the simultaneous heat and mass transfer processes. Thus, the main aim of this work is to investigate the potential and feasibility of DCMD process with respect to the energy efficiency and permeate water quality under various ex- perimental conditions and feed solutions using simple thermodynamic relations of mass and heat transfer based on overall system tempera- ture. Moreover, the feasibility of using brine from MSF plants as a low fouling feed for DCMD will also be investigated with a view for a pro- spective future large-scale application of this technology in Qatar. In this work, the concept of using low grade waste heat from different in- dustries in Qatar for freshwater production will be adopted. Since indus- trial waste heat can be found at various conditions (temperatures, flow rate, etc.) inside chemical plants, different sets of experiments will be conducted under similar conditions in the laboratory using electrical heaters to simulate such industrial conditions. The obtained results will help in determining potential production rates from various indus- trial activities in Qatar where waste heat can be utilized to produce freshwater through membrane distillation.

99A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

2. Theory and background

One way to enhance the flux is to increase the vapor pressure differ- ence across the membrane or to reduce temperature polarization. To achieve this, it is important to improve the convective heat transfer co- efficient in order to produce higher flux. The temperature polarization phenomenon occurs due to the presence of a thermal boundary layer which is located just adjacent to the membrane surface. This layer creates a heat resistance and makes the temperature at the liquid– membrane interface lower than that at the bulk of the feed and reduces the overall driving force [10]. Temperature polarization can be quanti- fied by the term temperature polarization coefficient (TPC) which is the ratio between the useful energy for mass transfer of vapors to the total energy invested in the process or in another words, it is a fraction of the trans-membrane temperature to the bulk temperature difference. This can be expressed mathematically as:

TPC ¼ Tm;f−Tm;p Tf−Tp

ð3Þ

where Tm,f and Tm,p are the feed and permeate membrane surface temperatures, respectively and Tf and Tp are the feed and permeate bulk temperatures, respectively. After conducting the bench scale tests under the various operating conditions, the amount of distillate pro- duced can be used to replace md in Eq. (1) which will be used to figure out the value of GOR at different operating conditions. Given the hot side mass flow rate (mh), outlet and inlet temperatures (Th,out and Th, in, respectively) and the specific heat capacity (Cp), the amount of heat energy supplied through the hot stream (Q) can be figured out from Eq. (4).

Q ¼ mhCp Th;out−Th;in � �

ð4Þ

The gained enthalpy from the heater (heating up the feed) or sim- ulated recoverable waste heat from the industrial streams must be converted into equivalent mass of steam (or ms) since there is no di- rect use of steam in MD processes. ms can be figured out by relating the gained enthalpy in Eq. (3) to that at standard conditions (100 °C and 1 atm). This steam at standard conditions is the mostly used heating steam in desalination plants to heat the incoming seawater and hence will be used to find the amounts of heating steam (ms) in MD process. The equation used to figure out the energy require- ment to heat steam at standard conditions can be mathematically expressed as Eq. (5):

Q ¼ msΔHref ð5Þ

where ΔHref is the latent heat of steam at standard conditions (100 °C and 1 atm) and has a value of 2326 kJ/kg [21]. By equating Eqs. (4) and (5), ms can be expressed as:

ms ¼ mhCp Th; out− Th; in

� �

ΔHref : ð6Þ

The value of GOR can be figured out more precisely by plugging in the values of md and ms at different operating conditions. Alternatively, for energy efficiency definition in Eq. (7), Total heat is equal to latent heat and heat lost by conduction and defined as:

EE ¼ QL QT

ð7Þ

QT ¼ QL þ QC ð8Þ

where, QC is heat lost by conduction and defined as:

QC ¼ kA δ

Tf−Tcoldð Þ ¼ h Tf −Tcoldð Þ ð9Þ

Where, k is thermal conductivity of membrane material, A is mem- brane area, δ is membrane thickness, Tf is feed inlet temperature, Tcold is cold side temperature, and h is heat transfer coefficient. Furthermore, the heat flux across the membrane from hot side to cold side consists of latent heat flown by water vapor and heat lost by conduction. The ener- gy consumption can be estimated using the energy balance of heating and cooling energy requirement as:

Qh ¼ mf Cpf Tfi−Tfoð Þ ð10Þ

Qc ¼ mdCpd Tdi−Tdoð Þ ð11Þ

where Qh and Qc are the heating and cooling energy, m is mass flow rate, Cp is specific heat, and T represents temperatures. Subscript f re- lates to feed side and d for the distillate side and i and o stand for inlet and outlet. Throughout studies conducted by Criscuoli et al. [13], the single stage DCMD showed energy efficiency between 27% up to a max- imum of 30% and they found that the low rate has no effect on thermal efficiency of MD. It was also found that temperature has a direct rela- tionship with energy efficiency. Simplifying the relation above and expanding Eq. (7), energy efficiency equation can be stated as:

EE ¼ QL QT

¼ Ai JΔHv ρpm

�

p Tdo−Tdið ÞCp ð12Þ

where Ai is effective membrane area, ρ is density and ṁp is the mass flow rate of the permeate side determined by experiment, J is permeate flux, ΔHv is latent heat of evaporation of water at average temperature and Cp is heat capacity of water at average temperature.

The permeate flux in membrane distillation is simplified in a shorter form equation as:

J ¼ CmembraneΔP ð13Þ

where Cmembrane is a membrane constant and is different for different material and membrane morphology, and it is related to a number of factor according to Eq. (14).

Cmembrane∝ daε tδ

ð14Þ

where, a, is an exponent factor within the range of 1 to 2. In this equation d is mean pore diameter of the membrane, ε is membrane po- rosity in percentage, t is proportion of conductive heat, and δ is mem- brane thickness [11].

ΔHV is calculated using Eq. (15) as function of temperature in Kelvin in the range of 273–373 K [22]:

ΔHV Tð Þ ¼ 1850:7 þ 2:8273T−1:6 � 10 −3T2: ð15Þ

The flux (J) in Eq. (12) is found experimentally by taking the differ- ence in the weight of distillate tank over certain time periods under given experimental conditions. Weighting balance is used to record the weight of water tank and explained in Materials and methods sec- tion. Distillate flux (measured as kg/m2h and reported as LMH) is calcu- lated as:

Distillate flux ¼ Difference in weight of distillate tank kgð Þ Membrane area m2

� � � Time hð Þ

: ð16Þ

100 A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

3. Materials and methods

3.1. DCMD bench scale setup

A bench scale plate-and-frame single stage DCMD module was de- signed, fabricated and tested in a system comprising solution feed flow metering and coolant recirculation flow metering, and associated ancillary equipment with data acquisition system (Fig. 2).

The hydrophobic flat sheet membrane used in this work was a PTFE membrane supported by a polypropylene (PP) sheet to provide it with strength and a degree of rigidity during the experiments. The PP sheet has also another advantage; decreasing the thermal conductivity of the membrane, which reduce the heat loss from the hot side through conduction. The flat sheet PTFE membranes laminated onto a PP non- woven layer (Sterlitech Corporation, US) had an active area of 0.014 m2 (of total area of 0.0203 m2), pore size of 0.22 μm, and thickness of 175 μm (both active layer and support layer). Details of the DCMD module and flow channel dimensions can be found elsewhere [23,24].

The feed velocities in both sides of the membrane were maintained constant and equal by using peristaltic pumps (Thermo Scientific model: FH100X, US). Highly pure deionized water (conductivity b 2 μS/cm) was used as the cold stream input to the MD unit. Temperature and pressure of the inlet and outlet streams of the membrane module

Fig. 2. Schematic diagram of DCMD bench s

were measured using thermo resistance RTDs (Model: RTD-NPT-72-E, Omega Engineering, UK) and pressure transducers (Model: PX309- 030GI, Omega Engineering, UK), respectively. The temperature of the feed solution was varied between 60 °C to 70 °C and controlled using heating circulators (Model: F32-MA, Julabo, Germany). Similarly, a chill- er was used to vary the distillate temperature in the range of 20–40 °C using refrigerated circulators (Model: F32-MA, Julabo, Germany). The digital data display system (Model: DP25B-E-230-A, Omega Engineer- ing, UK) was used to monitor pressure, temperature, flow rate, and con- ductivity of the system. The flux of distillate and mass of distillate produced were determined by the weight of the distillate using weighing balance (Model: VWR# 97035-640, Mettler Toledo) through active membrane area and time as described in Eq. (15). The concentra- tion of the feed was maintained constant over time by adding highly pure water in the feed reservoir to substitute water lost as distillate and distillate conductivity were monitored using conductivity indicator (model: 3433E8A, 10 cell constant, Hatch, USA). The data was acquired using a National Instruments data acquisition hardware (Chasis Model cDAQ9188; Module Model: NI-9219, National Instruments, US). The weight from the balance was acquired using a serial server (Model: NI ENET 232, National Instruments, USA). Data storage and processing was developed using LAB View data acquisition software. Flow rate of the feed and the distillate were varied from 0.5 l/min up to 3 l/min by

cale unit with module representation.

Table 3 Feed characteristic using ICP and IC analysis.

Parameters Unit Formula Brine Sea water NaCl solution

pH – – 8.38 7.60 6.98 TDS ppm TDS 58,000 41,600 100,000

Quality control meq anions 991 618 1683 meq cations 983 685 1797 meq ratio (anions/cations) 1.01 0.90 0.94

Anions, mg/L Fluoride F− b2.50 b2.5 b5.0 Chloride Cl− 32,127 19,661 59,605 Bromide Br− 46.0 32.0 b5.0 Phosphate PO4

3− b2.5 b2.5 b5.0 Nitrate NO3

− b2.5 b2.5 b5.0 Nitrite NO2

− b2.5 b2.5 b5.0 Sulfate SO4

2− 4025.0 3013.0 87.0 Thiosulfate S2O3

2− b2.5 b2.5 b5.0

Cations, mg/L Lithium Li+ b2.5 b2.50 b5.0 Sodium Na+ 18,434.0 12,858.0 41,228.0 Potassium K+ 491.0 343.0 b5.0 Magnesium Mg2+ 1738.0 1147.0 9.0 Calcium Ca2+ 52.0 459.0 51.0 Ammonium NH4

+ b2.5 b2.5 b5.0

Heavy metal, mg/L Aluminum Al b0.23 b0.23 b0.23 Boron B 4.85 3.80 b0.01 Cadmium Cd b0.35 b0.35 b0.35 Cobalt Co b0.01 b0.01 b0.01 Chromium Cr b0.01 b0.01 b0.01 Copper Cu b0.01 b0.01 b0.01

101A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

means of peristaltic pumps. Each experiment was repeated and the pa- rameters were recorded at least three times for reliability and quality assurance.

Three different feed solutions were used in this study, namely: brine, seawater, and synthetic NaCl solution. Seawater was collected from open intake of the Arabian Gulf; thermal reject brine was collected from Qatari MSF thermal desalination plant in Ras Abu Fantas, and syn- thetic NaCl Solution was prepared in the lab by dissolving ANALAR grade NaCl in deionized water. Conductivity and pH of the feed solu- tions are given in Table 2.

Turbulent promoters or spacers are thought to enhance the perfor- mance of MD systems by increasing the distillate flux produced. Throughout the experimental tests, polypropylene (PP) spacers (which has a 90° perpendicular filaments) were introduced in the flow channels of the DCMD system in order to study the effect of spacers on the performance of MD systems by conducting the test with/without spacer at the same operating conditions (feed and permeate tempera- ture of 70 °C and 30 °C, respectively, flow rate of 1.5 L/min for both hot and cold side, and brine as a feed solution).

3.2. Analytical methods

In order to investigate the performance of DCMD systems, analytical tools were utilized to investigate the quality of water as well as to char- acterize the membrane surface and morphology.

To observe the membrane cross-sections, membrane pores, bond- ing, and membrane backing support, the scanning electron microscope (SEM) imaging was carried out. FEI Quanta 200 Environmental Scanning Electron Microscope (ESEM) with a resolution of 5 nm and a magnification ×200K was used. To get the images and study the mor- phology of the membranes, a sample of membrane was cut and frozen in liquid nitrogen before being fractured. Cross section and surface of the membrane were sputtered with gold and then transferred to the mi- croscope for imaging.

Inductively Coupled Plasma/Optical Emission Spectrometry (ICP/ OES) is one of the most powerful analytical tools for the determination of heavy metals in aqueous samples. The technique is based on the spontaneous emission of photons from the excited atoms by the induc- tively coupled plasma. Liquid samples are introduced directly into the instrument and the sample solution is converted to an aerosol by the nebulizer before being transferred into the plasma state. At its core the inductively coupled plasma (ICP) sustains a temperature of approx- imately 10,000 K, so the aerosol is quickly vaporized, de-solved, decomposed, ionized and excited. Both the atomic and ionic excited state species may then relax to the ground state via the emission of pho- tons with a wavelength that can be used to identify the elements from which they initiated. The total number of photons is directly propor- tional to the concentration of the originating element in the sample.

Ion Chromatography (IC) is considered as the ideal method for ana- lyzing anions. It is a physicochemical method for separating mixtures of substances. The separation effect is based on repeated distribution be- tween two phases; one phase is stationary while the second, mobile phase moves in a particular direction. If a mixture of anions is subjected to chromatographic separation then distribution equilibrium is formed between the mobile and stationary phases for each individual anion. The anions can only be separated successfully when the distribution co- efficient, D, of the anion differs sufficiently from one another. According- ly anions with a larger distribution coefficient D will be held back

Table 2 feed solutions used in experiments.

Solution Source Concentration, ppm pH

Brine Desalination plant 58,000 8.38 Sea water Sea open intake (Arabian Gulf) 41,600 7.60 NaCl solution Prepared in lab Various concentration 7.00

(retained) more strongly than those with a smaller D. Eventually, the eluted anions could be detected by conductivity detector. Table 3 shows the characteristics of the three feed solutions used in this study.

4. Results and discussion

4.1. Membrane properties

The hydrophobic PTFE flat sheet membrane used in this work was supported by a PP sheet to provide it with strength and a degree of rigid- ity during the experiments. The PTFE membrane and PP support were thermally bonded and no adhesive or other materials were used to en- sure full compatibility with fresh water production. Fig. 3a and b show the pore size of the active layer (PTFE) and the support layer (PP), re- spectively. It is found that the pore sizes are smaller in the active layer, while more porous area is available in the support layer which re- sults in an increase in vapor transport by reducing the mass resistance. The inside walls of the pores increase the resistance to diffusion by de- creasing the momentum of the vapor molecules and having higher pores in support layer cause reduction in the diffusion resistance lead- ing to higher permeate flux. Also the porosity of the active layer is higher compared to the support layer which will increase the effective mass transfer area within the membrane [11].

The real thicknesses of the PTFE membrane and PP support layer to be around 30 μm and 140 μm, respectively. It is pivotal for the active layer to be as thin as possible to facilitate higher flux as membrane thickness of the active layer is inversely proportional to the permeate

Manganese Mn b0.01 b0.01 b0.01 Molybdenum Mo b0.05 b0.05 b0.05 Nickel Ni b0.01 b0.01 b0.01 Lead Pb b0.14 b0.14 b0.14 Antimony Sb b0.02 b0.02 b0.02 Selenium Se b0.08 b0.08 b0.08 Strontium Sr b0.37 b0.37 b0.37 Titanium Ti b0.04 b0.04 b0.04 Zinc Zn b0.28 b0.28 b0.28

Fig. 3. SEM picture of PTFE membrane with PP support: (A) pore size of PTFE active layer (B) pore size of support layer PP membrane.

102 A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

flux. On the other hand, a thicker membrane reduces the heat losses by conduction through the membrane. Accordingly a tradeoff analysis be- tween the flux and energy efficiency is required to get the most opti- mum thickness value. The optimum value of the membrane thickness was estimated in the range of 30–60 μm, which is in agreement with the values found in literature [25].

Distribution of pore sizes in a sample of membrane showed very nar- row pore size distribution with most of the pore being less than 0.22 μm. A 225 μm2 random membrane sample was analyzed and the maximum pore diameter obtained from the SEM imaging was 0.96 μm with only 13% of total pores. 83% of pores have size of less than 0.22 μm and almost 5% of pores exhibit pore with exactly pores diameter of 0.22 μm. This sharp pore size distribution reduces the potential water leakage and pore wetting through the membrane [26]. Porosity of the active side of the membrane and support layer is estimated to be around 85% and 90%, respectively. Contact angle measurements showed that the active side has contact angle of 126.95° and support layer has contact angle of 104.5°. Contact angle greater than 90° shows hydrophobicity of the membrane which is the key requirement for MD process to allow vapor to pass and repel liquid water [27]. Adnan et al. [28] tested differ- ent PTFE membrane from different source and contact angle of 126° up to 165° were reported in their studies. Higher contact angle in combina- tion with other factors such as smaller pore size, lower surface energy and higher surface tension lead to higher liquid entry pressure be great- er than the pressure difference at the membrane's liquid/vapor interface to prevent pore wetting. Pore wetting lead to penetration of liquid water and affect the quality of fresh water produced. Also, low contact angle leads to reduction in ability of membrane to reject non-volatile feed [25].

0

5

10

15

20

25

30

35

40

0 50 100 150 200 250 300 350

Fl u

x (L

M H

)

Time (min)

70-30 °C

65-30 °C

60-30 °C

70-20 °C

70-40 °C

Fig. 4. Flux profile of different feed (hot) inlet temperature and permeate (cold) temper- ature combinations (flow rate at both sides 1.5 L/min, feed solution: brine).

4.2. Effect of operating parameters

Experiments were performed at different feed and permeate side temperatures, feed flow rates, and feed solutions on bench scale unit. Fig. 4 depicts the distillate flux profiles at various feed inlet temperature and cold inlet temperature combinations at constant feed flow rate for the brine feed solution.

Tests with DCMD setup indicates that higher feed temperature en- hances the evaporation efficiency which is a ratio of the heat of evapo- ration to the total heat exchanged, although this may cause higher temperature polarization. In general, both temperature and feed flow rate have positive effects on the water vapor flux of DCMD. At a given temperature, the Reynolds number increases when increasing velocity of feed, which causes increase in the boundary layer heat transfer coef- ficient and the temperature polarization coefficient. Ultimately, it leads to an increase in the water vapor flux. Relative to the flow rate, feed temperature has a smaller effect on the Reynolds number at a given flow rate. This small dependency is due to limited changes in the density and viscosity of water in the given small temperature range. Although the increase of temperature slightly increases the Reynolds number, it drastically increases the pressure difference across the membrane sur- faces, which is the driving force and hence the flux. The experimental tests at different temperatures showed that the effect of feed tempera- ture on the distillate flux was higher than the effect of the permeate temperature. This is due to the exponential relationship between the vapor pressure of water and its temperature. Ideally, the vapor pressure of water changes more steeply at high temperature when compared with changes at low temperatures. Adham et al. [2] and Minier-matar et al. [5] used similar feed solution with same operative parameters and the similar conclusions were found. The result in this study com- pared to studies by Adham et al. [2] and Minier-matar et al. [5] shows better performance in terms of flux due to good insulation which avoids heat loss and reveal better recovery of the system. Optimization and tradeoff between the flow rate and feed temperature is an operative way to get high water vapor flux in DCMD.

The effect of feed flow rate on the permeate flux and energy efficien- cy (EE) of the bench scale system is illustrated in Fig. 5. The energy effi- ciency is computed using Eq. (12) and experimental data. The feed velocity positively affects the EE of the DCMD in the studied range, pri- marily through the increased flux. The effect of feed flow rate can be re- lated to the mass flux of the DCMD as indicated by a flux increase when increasing the feed flow rate (Fig. 5). In these experiments, the temper- ature in the liquid velocity variation is kept constant at 70 °C for feed (hot) side and 30 °C for cold (permeate) side when the rejected brine is used as feed solution. The increase is attributed to minimizing the boundary layer, created by temperature polarization, and maximizing the heat transfer coefficient. Moreover, at high feed rates, a more turbu- lent flow regime results causing the temperature at the membrane in- terface to approach the bulk feed temperature. The trans-membrane

Permeate Flux

Energy Efficiency

0

5

10

15

20

25

30

35

40

45

50

0

5

10

15

20

25

30

35

40

0 0.5 1 1.5 2 2.5 3 3.5

En er

g y

Effi ci

en cy

(% )

Pe rm

ea te

F lu

x (L

M H

)

Flow Rate (L/min)

Fig. 5. Effect of feed flow rate on the direct contact membrane distillation energy efficiency and permeate flux (Tf: 70 °C, Tp: 30 °C, brine).

103A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

temperature difference becomes higher thus lowering the temperature polarization effect. It should be noted that the increase in flux with the increase in flow rate appears to reach some asymptotic value at higher feed flow rate. According to the data in this study energy efficiency could vary from 21.7% for the best case and down to 11.6% for the sys- tem where the temperature difference between the hot and cold sides is at its minimum.

This study showed that the effect of the temperature on the perme- ate flux is higher than the effect of flow rate which is shown from the highest flux values obtained at the highest temperature and flow rate studied. Therefore, temperature is the most influencing factor affecting the permeate flux more than the flow rate does. The experimental re- sults found in this part were not different than the results published by others. They found that as the circulation velocity increases, the per- meate flux increases as well till the flux reaches an asymptotic value in which the flux was not increasing as the circulation velocity was in- creased [29–31].

In general, as the feed temperature increases the performance of DCMD by increasing the driving force, which is directly proportional to the vapor pressure as defined by Antoine equation. Fig. 6 confirms that the flux production increases by increasing the feed solution tem- perature. The energy efficiency of the DCMD system also increases with an increase in the feed temperature. The energy efficiency of sys- tem increased from 13.6% to 17% when feed temperature is increased

Permeate Flux

Energy Efficiency

4

6

8

10

12

14

16

18

10

15

20

25

30

35

55 60 65 70 75

En er

g y

Effi ci

en cy

(% )

Pe rm

ea te

F lu

x (L

M H

)

Feed Inlet Temperature ( °C)

Fig. 6. Effect of feed inlet temperature on the permeate flux and energy efficiency (Tp: 30 °C, Hot/Cold inlet flow rate: 1.5 L/min).

by 10 °C (60 °C to 70 °C). It is anticipated that increasing the feed tem- perature enhances the evaporation efficiency (the ratio of heat of evap- oration to total heat exchanged). In studies by Yao et al. [18] and Geng et al. [19] on AGMD, the efficiency of the system showed an increase of GOR for AGMD with the increase of inlet temperature. It was shown also that with increasing the temperature by 10 °C, increases GOR by 3 folds.

Higher vapor pressure can be obtained, in principal, by decreasing permeate side temperature. In contrast to feed temperature, increasing permeate side temperature, decreases the flux. Experimental results showed that for brine at constant feed side temperature of 70 °C, in- creasing permeate temperature decreases the flux. The effect is lower compared to increasing the feed temperature (Fig. 7). The effect of per- meate side temperature on energy efficiency of DCMD system is depicted in Fig. 7. As the flux increases with decreasing the permeate side temperature, the energy efficiency decreases by increasing the per- meate inlet temperature hence, decreasing heat transfer coefficient and increasing polarization effect. Energy efficiency of the system decreased from 21% when permeate temperature is kept at 20 °C and decreased to 14% when the temperature of the cold side rose to 40 °C. Similar effect was reported in literature [3,18,19].

On the other hand, effect of temperature on GOR analysis is depicted in Fig. 8 by showing the effect of hot and cold side's temperatures on GOR values. The plot shows that as the feed temperature increases, GOR values increase as there will be more distillate produced due to in- creased distillate flux at high temperatures. On the other hand, increas- ing permeate temperatures results in lowering the temperature difference between the hot and cold sides (and hence lowering the driv- ing force) and that results in a lower permeate flux. Not only that, in- creasing the permeate temperature will reduce the temperature difference between the hot inlet and outlet streams (ΔT in Eq. (5)) and hence lower GOR values.

Fig. 9 shows the relationship between GOR values and the feed flow rates at a feed and permeate temperature of 80 °C and 20 °C, respective- ly. It is shown in this figure that as the flow rate increases, GOR de- creases due to the increase in the heating requirement (high ms in GOR definition). Moreover, increasing the flow rate will result in less residence time inside the MD modules and that will result in lowering the distillate produced (md in Eq. (1)). Despite the fact that increasing the feed/permeate flow rate will have the effect of having a more effi- cient heat and mass transfer process (hence, increasing md), it seems that this effect has less impact on the GOR values and hence the overall effect would be lowering GOR values with the flow rate. The reason be- hind this behavior is that at low flow rates, low flux value was observed due to the presence of the “relatively thick” thermal boundary layer. As the circulation rate increases, the thermal boundary layer is expected to get thinner and hence higher fluxes until it reaches a value that is

Permeate Flux

Energy Efficiency

0

5

10

15

20

25

10

15

20

25

30

35

40

45

50

15 20 25 30 35 40 45

En er

g y

Effi ci

en cy

(% )

Pe rm

ea te

F lu

x (L

M H

)

Permeate Inlet Temperature ( °C)

Fig. 7. Effect of permeate inlet temperature on the permeate flux and energy efficiency of direct contact membrane distillation (Tf: 70 °C, Hot/Cold inlet flow rate: 1.5 L/min, feed: brine).

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

20 25 30 35 40

G O

R

Cold Inlet Temperature, oC

80 °C 75 °C 70 °C

65 °C 60 °C 55 °C

Fig. 8. GOR values at different Hot and Cold temperatures for brine solution. The flow rate was maintained at 1.5 L/min for both hot and cold sides.

15

20

25

30

35

40

0 50 100 150 200 250 300 350

Fl u

x (L

M H

)

Time (min)

Brine Sea water NaCl solu�on-100000 ppm NaCl Solu�on-58000 ppm NaCl Solu�on- 10000 ppm

Fig. 10. Effect of feed concentration on permeates flux (Tf: 70 °C, Tp: 30 °C and Hot/Cold flow rate: 1.5 L/min).

104 A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

independent of the circulation velocity. This directly affects the GOR value as flux has direct effect on GOR per mathematical equation de- scribed in Eq. (1).

4.3. Effect of feed concentration

Feed salinity plays a significant role in energy efficiency and GOR of the system. Feed concentration directly affects the viscosity and mobil- ity of the fluid which in turn affects vapor pressure of the solution and increase the effect of polarization concentration thus decreasing perme- ate flux. Fig. 10 depicts the effect of feed brine concentration on flux of the system. With an increase in the feed concentration, a concentration boundary layer might develop at the membrane surface. Concentration boundary layer, together with the temperature boundary layer, reduces the driving force for water evaporation and decreases the flux by affect- ing the heat and mass transfer coefficients. Nonetheless, increasing the flow rate was found to enhance the turbulent flow conditions in the feed stream and decreases both boundary layers resulting in improved cell performance. Lawson and Lloyd [32] observed that concentration polarization had almost no effect on water flux over extended ranges of temperatures (40–80 °C) and feed concentrations when strong mixing was maintained in the feed stream. The most significant advan- tage of the DCMD process for desalination is the relatively minimal ef- fect of feed salt concentration on the performance of the system when compared to RO. Furthermore, the water vapor pressure decreases by increasing salt concentration while the viscosity of the solution in- creases at higher feed brine concentration which will decrease the heat transfer coefficient due to decreasing Reynolds number. Also due

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

G O

R

Flow Rate, L/min

Fig. 9. GOR values at different Feed flow rate for thermal reject brine solutions at a feed and permeate temperatures of 80 °C and 20 °C, respectively.

to increase in concentration polarization at feed side boundary layer, mass transfer coefficient decreases and this lead to decreasing flux and energy efficiency. The performance of RO is more affected by feed concentration compared to MD. In RO, as the feed concentration in- creases, the permeate flux will be reduced due to the reduction in the mass transfer across the membrane and that results in increasing the salt passage, concentration polarization, scale formation, and osmotic pressure (higher operating cost).

The effect of added chemicals to the feed water during the pretreat- ment process of MSF plant must also be highlighted. Several chemicals such as chlorine as bio-foulant, phosphonates, polyphosphate, polymaleic, poly-carboxylic acids, or a blend of several of these as anti-scalant, poly-othelyne ethylene oxide or a similar surfactant as anti-foaming agent, and other chemicals for anti-fouling purposes usu- ally exist in the feed water with high dosage [33]. As shown in Fig. 11, the permeate flux from the brine feed solution showed a higher value when compared to seawater. This is most likely due to the presence of these additives in the feed which could efficiently reduce the fouling de- spite the fact that it has higher salinity. Also the presence of these chemicals and non-ionic agents in the brine can reduce the viscosity of the solution and cause increase in flux and energy efficiency of the system. The physical properties of the feed have an effect on the bound- ary layer film adjacent to the membrane surface in the hot side and can lead to a slight increase in the flux as it is observed in Fig. 10. Decreasing viscosity also causes the heat transfer coefficient to increases due to in- crease in Reynolds number. Another fact which is also widely reported in literature [2,5,34,35] is the effect of membrane type and morphology on performance of the system. Adham et al. [2] reported that using the same feed with same operating condition with different types of

0

5

10

15

20

25

30

35

40

Brine Seawater NaCl Solu�on (100000 ppm)

NaCl Solu�on (58000 ppm)

NaCl Solu�on (10000 ppm)

Fl u

x, L

M H

70 ˚C 65 ˚C 60 ˚C

Fig. 11. Effect of feed concentration and feed (hot) inlet temperature on permeate flux of DCMD process (Tp: 30 °C, Hot/Cold flow rate: 1.5 L/min).

Table 5 Effect of feed concentration on energy efficiency and GOR of the system (Tf: 70 °C, Tp: 30 °C and Hot/Cold flow rate: 1.5 L/min).

Feed GOR

Brine 1.09 Sea water 0.84 NaCl solution (100,000 ppm) 0.66 NaCl solution (58,000 ppm) 2.1 NaCl solution (10,000 ppm) 2.4

105A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

membrane cause different flux and efficiency due to membrane physi- cal properties such as contact angle and material thermal conductivity.

Artificial brine prepared from NaCl with the same salinity showed a slightly higher performance compared to brine as it only contains sodi- um and chloride ions. Similar trend was observed by Adham et al. [2]. The trend can also be true for energy efficiency as decreasing the flux decreases the energy efficiency (Table 4), in accordance with the energy efficiency Eq. (12). As concentration of the feed increases, energy effi- ciency of the system decreases except in the case of brine (dosed with chemicals) which shows higher efficiency compared to seawater.

Table 5 shows the effect of salinity level and different feed solutions on GOR values. In this test, the temperature of the feed and permeate so- lutions were kept at 80 °C and 20 °C, respectively while the flow rate of both solutions was also kept at 1.5 L/min. The thermal rejected brine has the highest GOR value compared to seawater as it has already been dosed with anti-foulant (hence, no major fouling is encountered). On the other hand, NaCl solution with similar concentration as brine showed higher GOR since it contains no fouling or micro-organisms. Both the seawater and synthetic NaCl solutions with higher salinity come after the thermal rejected brine as they are not dosed with anti- foulant. The reason why seawater has a higher GOR value (when com- pared to NaCl solution with 100,000 ppm salinity) is it has lower salinity (higher md in Eq. (1)).

Seawater contains large number of macro-particles and micro- organism such as algae, mussels, etc. and the presence of such contam- inant in water can cause reduction of flux as seen before. To overcome this, pretreatment through Micro Filtration (MF) system with pore size of 1 μm and pressure of 1–2 bar might be recommended to remove the microorganisms and algae. This can help in increasing MD flux. From the calculations conducted in this study, it has been shown that the highest GOR value for bench scale can be achieved at the highest feed and lowest permeate temperatures (at a temperature combination of 80–20 °C with a GOR value of 2.4). When the GOR tests were also inves- tigated at the optimum flow rate, it has been found that the highest GOR (1.47) is encountered at the lowest flow rate (or 0.5 L/min). Conducting the same tests at different feed solutions showed that the optimum feed solution is the one has less salinity and preferably dosed with anti- scalant (from thermal rejected brine). Thermal reject brine as feed solu- tion with a flow rate of 0.5 L/min and a feed and permeate temperatures of 80 and 20 °C , respectively will result in having good performance possible (GOR value: 1.47) which will be manifested in the good distil- late produced (md) per unit of steam supplied (ms).

4.4. Effect of feed flow pattern and turbulence

Flow pattern and configuration of the system also play a major role in enhancing the flux. It is found that introducing turbulent promoters into the flow channels is responsible for destabilizing the laminar flow and promoting eddies; hence, the flow regime is no longer laminar. The ultimate result of that will not only be mass transfer enhancement but also heat and momentum transfer enhancement. The same finding was also reached in this research and depicted in the permeate profiles of both systems (with and without a spacer).

The flux from the system with and without spacer was compared and depicted in Fig. 12a. The turbulent promoter (spacer) was found

Table 4 Effect of feed concentration and different process temperature combinations on energy ef- ficiency of the system (Hot/Cold flow rate: 1.5 L/min).

Feed 70–30 °C

65–30 °C

60–30 °C

70–20 °C

70–40 °C

Brine 17.003 14.289 13.616 20.259 14.019 Sea water 16.665 11.296 10.708 18.557 13.018 NaCl solution (100,000 ppm) 12.870 10.131 9.443 18.046 9.513 NaCl solution (58,000 ppm) 17.856 14.810 13.841 20.986 14.861 NaCl solution (10,000 ppm) 18.240 15.141 14.112 21.681 15.134

to increase the mass transfer of the MD systems when compared with the spacer-free systems. Spacer destabilizes the flow and creates eddy currents in the laminar regime so that momentum, heat, and mass transfer are enhanced. This can be easily observed if the permeate fluxes obtained from the two system were compared. The distillate flux was enhanced by around 51% when a spacer is used which shows the signif- icance of spacers as an enhancement tool to the MD flux. Energy effi- ciency of the system was also reduced accordingly by reduction in flux as shown in Fig. 12b.

By altering the flow hydrodynamic to the turbulent flow regime, the net-like spacers help in increasing the film heat transfer coefficient (h) and hence enhance the mass flux of the MD system. The hot side heat transfer coefficient of the spacer-free channel was found to be 1190; however, when a spacer was introduced, the heat transfer coeffi- cient at the hot side increased to 2163. This accounts for about 82% in- crease in the heat transfer coefficient only with a spacer. The Reynolds number was estimated to be 1105 in the absence of spacers and 6379 when a spacer is used which agrees well with the results found by Phattaranawik et al. [22] However, The flow regime of our system lies in the turbulent region couturier to Phattaranawik et al. where they found the flow regime of the circulating solutions falls in the lower tran- sition region between the laminar and turbulent regimes. This can be simply explained by the use of smaller and thinner (shallower) flow channels in pour case which would further increase the inertia over the viscous forces as opposed to the system studied by Phattaranawik et al. [22]. Fluids flowing through flow channels with Reynolds numbers below 2000 are considered laminar while those above that are consid- ered turbulent. This means that the application of spacers have changed the flow regime inside the flow channel from being laminar into turbulent.

The effect of flow mode on the permeate flux was also studied by having a co-current and counter current system in Fig. 12a. In the co- current mode, the same counter-current system used earlier was used but the inlet and outlet streams were reversed so there was a co- current flow mode within the system. Fig. 12a show the permeate flux from both system. The effect of having a more efficient heat transfer sys- tem on the distillate flux was also explored when two different flow patterns (counter and co current) were run and the permeate flux was measured. A higher distillate flux value was associated with the more efficient heat transfer flow mode (counter current). Energy effi- ciency of such system is shown in Fig. 12b.

We have reported previously [22,36] on the temperature profile of the co-current mode that the hot and cold streams enter the same ter- minal (same side of the MD module). Initially the temperature gradient between the hot and cold streams are large but as they flow through the MD system, the temperature gradient steeply decreases which results in lower permeate fluxes throughout the last stages. On the other hand, in the counter current mode, the hot and cold streams enter at opposite sides of the MD module. This results in an intermediate temperature gradient which will remain constant throughout the flow which will re- sult in a higher permeate flux and hence GOR and energy efficiency compared to the co-current system. Hereafter, counter current flow pat- tern is more favored over the co-current one due to the enhanced heat and mass transfer efficiency. When spacers are used, the temperature on the feed side of the membrane surface will be increased and the tem- perature on the permeate side will be reduced which will result in a

(a) (b)

15

20

25

30

0 100 200 300

Pe rm

ea te

F lu

x, L

M H

Time, min

Baseline Test Co-Current mode Without Spacer

0

5

10

15

20

Baseline Test Without spacer CO-Current flow mode

En er

g y

Effi ci

en cy

, %

Fig. 12. (A) Permeate flux and (B) energy efficiency for the baseline test (with spacer and counter-current mode), without a spacer, and co-current flow mode (Tf: 70 °C, Tp: 30 °C, Hot/Cold flow rate: 1.5 L/min, feed: brine).

106 A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

larger driving force due to higher temperature difference and hence higher fluxes and reduction of the temperature polarization coefficient.

4.5. Water quality analysis

Water quality is a major factor in investigating the performance of desalination technologies. Quality of the distillate produced from DCMD bench scale unit is analyzed using Inductive Coupled Plasma (ICP) and Ion Chromatography (IC) and the salt rejection was measured based on difference between initial feed concentration and final distil- late concentration using Eq. (16).

Y ¼ Cf−Cp

Cf � 100 ð16Þ

where, Cf and Cp are the salt concentration of feed and salt concen- tration of permeate, respectively.

The distillate produced by DCMD in this study is highly pure and al- most there was no salt passed to the distillate through the membrane. Using Eq. (16), the salt rejection for all salts are 99.9%. The details of dis- tillate quality analysis are found in Table S3 and S4 in the Supplementa- ry material section. The feed temperature has no significant effect on the quality of distillate produced. This is due to the hydrophobicity of mem- brane which is not a function of temperature and regardless of feed temperature the vapor produced is pure and salts are rejected at the membrane surface. Further discussion on water quality is available in the Supplementary material section.

5. Conclusions

We have successfully evaluated a single stage Direct Contact Mem- brane Distillation (DCMD) bench scale with various feed and permeate side conditions using PTFE flat-sheet membranes. Energy efficiency, GOR performance, and water quality of distillate for the system were analyzed at different operating conditions and it was found that the highest GOR value can be obtained at low salinity feed and a flow rate of 0.5 L/min. In other words, the highest GOR value can be achieved at the lowest permeate and highest feed temperatures. Moreover, the lower the flow rates of both feed and permeate solutions, the higher the GOR value. Feed solutions that are dosed with anti-scalant, anti- fouling, and anti-foaming have been found to achieve higher GOR values. Although feed salinity plays an important role in flux control of the system, anti-fouling dosing can enhance the flux of system even at higher salinity as it has been observed with the case of seawater and brine. Salt rejection of 99.9% was achieved with all feed solutions at dif- ferent salinities and electrical conductivity of distillate in all case was less than 5 μS/cm. MD showed a promising application to be used as a source to produce fresh water for industries using waste heat with

minimum capital cost, minimum operating cost, and yet reduced envi- ronmental impact and damage to ecosystem. DCMD has proven to be a feasible and effective technology capable of consistently producing high quality produced water from a very high salinity feed even with substantial quality difference compared to other desalination methods such as RO and MSF.

Acknowledgments

The authors would like to express their gratitude to Qatar University and ConocoPhillips GWSC for their financial support of this work under grants QUST-CENG Fall 12/13-22 and QUEX-CWT-11/12-5 respectively. The authors would like also to thank Mr. Joel Minier Matar of ConocoPhillips GWSC and Dr. Ashraf S. Hassan of QEERI for his technical support.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.desal.2015.01.016.

References

[1] M. Al Ansari, Concentrating solar power to be used in seawater desalination within the Gulf Cooperation Council, Energy Environ. Res. 3 (1) (2013) 10–23.

[2] S. Adham, A. Hussain, J.M. Matar, R. Dores, A. Janson, Application of membrane dis- tillation for desalting brines from thermal desalination plants, Desalination 314 (2013) 101–108.

[3] S. Alobaidani, E. Curcio, F. Macedonio, G. Diprofio, H. Alhinai, E. Drioli, Potential of membrane distillation in seawater desalination: thermal efficiency, sensitivity study and cost estimation, J. Membr. Sci. 323 (1) (2008) 85–98.

[4] M. Tomaszewska, Membrane distillation — examples of applications in technology and environmental protection, Pol. J. Environ. Stud. 9 (1) (2000) 27–36.

[5] J. Minier-Matar, A. Hussain, A. Janson, F. Benyahia, S. Adham, Field evaluation of membrane technologies for desalination of highly saline brines, Desalination 351 (2014) 101–108.

[6] C. Gostoli, G.C. Sarti, S. Matulli, Low temperature distillation through hydrophobic membranes, Sep. Sci. Technol. 22 (1987) 855–872.

[7] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive review, Desalination 287 (2011) 2–18.

[8] R. Einav, K. Harussi, D. Perry, The footprint of the desalination processes on the environment, Desalination 152 (1-3) (2003) 141–154.

[9] R. Deng, L. Xie, H. Lin, J. Liu, W. Han, Integration of thermal energy and seawater desalination, Energy 35 (2010) 4368–4374.

[10] F. Banat, Desalination by a “compact SMADES” autonomous solar-powered membrane distillation unit, Desalination 217 (2007) 29–37.

[11] H.E. Fath, S.M. Elsherbiny, A.A. Hassan, M. Rommel, M. Wieghaus, J. Koschikowski, M. Vatansever, PV and thermally driven small-scale, stand-alone solar desalination systems with very low maintenance needs, Desalination 225 (2008) 58–69.

[12] E. Guillen-Burrieza, J. Blanco, G. Zaragoza, D.C. Alarcon, P. Palenzuela, M. Ibarra, W. Gernjak, Experimental analysis of an air gap membrane distillation solar desalina- tion pilot system, J. Membr. Sci. 379 (2011) 386–396.

[13] A. Criscuoli, M. Carnevale, E. Drioli, Evaluation of energy requirements in membrane distillation, Chem. Eng. Process. 47 (7) (2008) 1098–1105.

107A. Kayvani Fard et al. / Desalination 360 (2015) 97–107

[14] X. Wang, L. Zhang, H. Yang, H. Chen, Feasibility research of potable water production via solar-heated hollow fiber membrane distillation system, Desalination 247 (2009) 403–411.

[15] H. Lee, F. He, L. Song, J. Gilron, K.K. Sirkar, Desalination with a cascade of crossflow hollow fiber membrane distillation devices integrated with a heat exchanger, AICHE J. 57 (2011) 1780–1795.

[16] G. Zuo, R. Wang, R. Field, A.G. Fane, Energy efficiency evaluation and economic anal- yses of direct contact membrane distillation system using aspen plus, Desalination 283 (2011) 237–244.

[17] E.K. Summers, Energy efficiency comparison of single-stage membrane distillation (MD) desalination cycles in different configurations, Desalination 290 (2011) 54–66.

[18] Y. Qin Yao, Y. Yuan, L. Liu, F. He, Y. Wu, A continuous-effect membrane distillation process based on hollow fiber AGMD module with internal latent-heat recovery, AICHE J. 59 (4) (2013).

[19] H. Geng, H. Wu, P. Li, Q. He, Study on a new air gap membrane distillation module for desalination, Desalination 334 (2014) 29–38.

[20] S. Lin, N. Yip, M. Elimelech, Direct contact membrane distillation with heat recovery: thermodynamic insights from module scale modeling, J. Membr. Sci. 453 (2014) 498–515.

[21] C. Summariva, Desalination in the Gulf, International Desalination Association (IDA), Perugia, 2011.

[22] J. Phattaranawik, R. Jiraratananon, A.G. Fane, C. Halim, Mass flux enhancement using spacer filled channels in direct contact membrane distillation, J. Membr. Sci. 187 (2001) 193–201.

[23] A. Kayvani Fard, Approach and methodology, Membrane Distillation Desalination: An Emerging Technology, LAPLAMBERT Academic Publishing, Germany, 2014. 54–77.

[24] Y.M. Manawi, K. Majeda, A. Kayvani Fard, F. Benyahia, S. Adham, Effect of operation- al parameters on distillate flux in direct contact membrane distillation (DCMD): comparison between experimental and model predicted performance, Desalination 336 (2014) 110–120.

[25] L.M. Camacho, L. Dumée, J.Z. Zhang, Jun-de Li, M. Duke, J. Gomez, S. Gray, Advances in membrane distillation for water desalination and purification applications, Water 5 (2013) 94–196.

[26] H.J. Hwang, H. Ke, G. Stephen, Z. Jianhua, S.M. Il, Direct contact membrane distilla- tion (DCMD): experimental study on the commercial PTFE membrane and model- ing, J. Membr. Sci. 371 (1–2) (2011) 90–98.

[27] J. Zhang, N. Dow, M. Duke, E. Ostarcevic, J.-D. Li, S. Gray, Identification of material and physical features of membrane distillation membranes for high performance desalination, J. Membr. Sci. 349 (2010) 295–303.

[28] A. Sharmiza, M. Hoang, H. Wang, Commercial PTFE membranes for membrane dis- tillation application: effect of microstructure and support material, Desalination 284 (2012) 297–308.

[29] J.M. Rodriquez-Maroto, L. Martinez-Diez, Bulk and measured temperatures in direct contact membrane distillation, J. Membr. Sci. 250 (2005) 141–149.

[30] S. Bouguccha, R. Chouilkh, M. Dhahbi, Numerical study of the coupled heat and mass transfer in membrane distillation, Desalination 152 (2002) 245–252.

[31] M. Khayet, M.P. Godino, J.I. Mengual, Possibility of nuclear desalination through var- ious membrane distillation configurations: a comparative study, Int. J. Nucl. Desalinat. 1 (1) (2003) 30–46.

[32] K. Lawson, D. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 1–25. [33] M. Darwish, A. Hassabou, B. Shomar, Using seawater reverse osmosis (SWRO)

desalting system for less environmental impacts in Qatar, Desalination 309 (2013) 113–124.

[34] J. Li, Y. Guana, F. Chenga, Y. Liuc, Treatment of high salinity brines by direct contact membrane distillation: effect of membrane characteristics and salinity, Chemosphere (2015). http://dx.doi.org/10.1016/j.chemosphere.2014.12.006.

[35] J. Mericq, S. Laborie, C. Cabassud, Vacuum membrane distillation of seawater reverse osmosis brines, Water Res. 44 (2010) 5260–5273.

[36] Y. Manawi, M. Khraisheh, A. Fard, F. Benyahia, S. Adham, A predictive model for the assessment of the temperature polarization effect in direct contact membrane distil- lation desalination of high salinity feed, Desalination 341 (2014) 38–49.