A novel multi-effect membrane distillation (MEMD) process has been implemented to treat water containing four different inorganic solutes. The 4-stage MEMD module was developed based on the air-gap configuration. The influence of operating parameters like concentration, feed temperature, flow rate and operating time on permeate fluxes of zinc sulfate, sodium fluoride, magnesium chloride and sodium carbonate solutions was observed. Concentration had negligible effect on the MEMD's permeate flux, while its performance increased with increasing feed temperature and flow rate. Its separation efficiency was stable at more than99.91% throughout the experiment. In addition, its specific energy consumption after the recovery of the latent heat of vaporization and sensible heat of brine was measured at different component concentrations and found to be independent of the type of component.

INTRODUCTION

Membrane distillation (MD) is an emerging, non-isothermal membrane separation process for use when water is present as a major component in the feed to be separated. It uses thermal energy to provide a vapor phase of more volatile molecules present in both the feed and condensing (permeated vapor) streams. The driving force is the vapor pressure difference created by the temperature differential across the hydrophobic membrane surface (Lawson & Lloyd 1997; Curcio & Drioli 2005). The main features of MD are that it can be operated at low temperature and unpressurized, in comparison to conventional processes like reverse osmosis (RO). MD can also operate on low grade energy – e.g., solar or waste heat (Li et al. 2008; Manella et al. 2010; Pangarkar et al. 2016). Hence MD is expected to be cost-effective.

The traditional MD process has some limitations like high thermal energy consumption and low permeates flux, and no module designs were found in the literature. Due to this, it is not used in industry (Cipollina et al. 2012; Geng et al. 2014; Khalifa et al. 2015). Some researchers have modified MD modules using the multi-effect concept in MD configurations – e.g., vacuum multi-effect membrane distillation (V-MEMD), a process implemented by Memsys GmbH (Germany) for desalination (Lu et al. 2012; Zhao et al. 2013). Multi-effect membrane distillation (MEMD) has some advantages over traditional MD, like its high water production rates due to the multiple stages in each module, recovery of internal energy, high gain output ratio, low specific energy consumption due to the recovery of heat, and low cooling water consumption due to the use of fresh feed solutions for cooling.

Among the four major MD configurations, air-gap membrane distillation (AGMD) has advantages like low conductive heat losses and low potential for membrane wetting, and the possibility of internal heat recovery (Liu et al. 1998; Cipollina et al. 2012; Pangarkar & Deshmukh 2015). Because of this, the multi-effect concept was added to the AGMD configuration to develop a new MEMD module. In this study this 4-stage MEMD module based on AGMD has been applied successfully in a laboratory study to treat aqueous solutions containing inorganic components.

METHODS

MEMD module

The internal channels of the 4-stage MEMD module and water flow within it are shown in Figure 1. The module was made with an acrylic material using aluminum foil for the cooling plates. The laboratory module contains three feed channels, two cooling channels and four permeate or air gap channels. Each feed and coolant channel is 80 (l) × 100 (h) × 5 (d) mm, and the air gap thickness is about 2 mm. A common cooling channel is used in the module's two stages successively. Fresh feed was circulated through these channels to cool the permeate vapor. The vapor was condensed on the aluminum foil surface. A picture of the internal arrangement of this MEMD module is shown in Figure 2(a) with the assembled module shown in Figure 2(b).
Figure 1

Internal channels and water flow in the 4-stage MEMD module.

Figure 1

Internal channels and water flow in the 4-stage MEMD module.

Figure 2

(a) Internal arrangement of the MEMD module and (b) The assembled module.

Figure 2

(a) Internal arrangement of the MEMD module and (b) The assembled module.

Four, flat sheet, polytetrafluoroethylene hydrophobic membranes were used – supplied by Madhu Chemicals Pvt. Ltd. Mumbai (India). Membrane pore size, thickness and porosity were nominally 0.45 and175 μm, and 70%, respectively. The effective permeation membrane areas for single sheets and four membrane sheet sets are respectively about 0.008 and 0.032 m2.

Experimental setup and procedure

Figure 3 is a schematic of the MEMD process. The 20 liter first feed (and cooling) tank – 1 in Figure 3 – contained fresh feed water, which provides cooling when circulated through the module's cooling channels. The internal latent heat of vaporization is transferred to the feed/cooling water as the water vapor condenses. The sensible heat is recovered in the heat exchanger from the hot brine solution, after which external heat is supplied to the second feed tank (2 in Figure 3). The feed is circulated from the feed tank to the first feed channel by the 0.373 kW (0.5 hp) circulation pump. A Rota-meter is used to measure the feed flow rate. The feed and cooling channel inlet and outlet temperatures were measured using pt100 thermocouple sensors.
Figure 3

Schematic diagram of MEMD experimental setup.

Figure 3

Schematic diagram of MEMD experimental setup.

The performance of the 4-stage MEMD process was analyzed for various feed and operating conditions like concentration, temperature, flow rate, and process time. Details are given in Table 1. The investigation into permeate flux and rejection factors was based on the use of four inorganic components, i.e. zinc sulfate (ZnSO4), sodium fluoride (NaF), magnesium chloride (MgCl2) and sodium carbonate (Na2CO3). The concentration ranges used in the study are shown in Table 2.

Table 1

Experimental operating parameters used in 4-stage MEMD investigation

Operating parameter Range 
Feed temperature 40 to 80 °C 
Feed flow rate 0.3 to 1 L/min 
Coolant (fresh feed) temperature 27 °C 
Coolant flow rate in each channel 0.25 L/min 
Operating parameter Range 
Feed temperature 40 to 80 °C 
Feed flow rate 0.3 to 1 L/min 
Coolant (fresh feed) temperature 27 °C 
Coolant flow rate in each channel 0.25 L/min 
Table 2

Concentration ranges of individual components

  Concentration ranges (mg/L)
 
Components Minimum Maximum 
Zinc sulfate (ZnSO4150 20,150 
Sodium fluoride (NaF) 400 20,400 
Magnesium chloride (MgCl22,000 22,000 
Sodium carbonate (Na2CO3300 20,300 
  Concentration ranges (mg/L)
 
Components Minimum Maximum 
Zinc sulfate (ZnSO4150 20,150 
Sodium fluoride (NaF) 400 20,400 
Magnesium chloride (MgCl22,000 22,000 
Sodium carbonate (Na2CO3300 20,300 

Module performance parameters like permeate flux, rejection factor and specific energy consumption for feed heating were evaluated using Equations (1)–(3), respectively.

Permeate flux, JD (L/m2h): 
formula
1
Module separation efficiency: 
formula
2
Specific energy consumption with heat recovery: 
formula
3
where V (L) is the volume of permeate collected in time t (h), A (m2) is the membrane area, Cf and Cp are the feed and permeate concentrations respectively, mf and mD (Kg/s) are the mass flow rates of the feed and permeate waters respectively, Tf and T0 are the temperature of the feed entering the first feed channel, and after recovery of both the latent heat of vaporization and sensible heat of brine respectively, and Cpf (at 80 °C 4.18 KJ/kg °C) is the specific heat capacity of water.

RESULTS AND DISCUSSION

Effect of concentration on 4-stage MEMD performance

The effect of concentration of different component feed solutions on permeate flux from the 4-stage MEMD process is shown in Figure 4. The flux with distilled water was about 43.5 L/m2 h. On the other hand, the flux decreases slightly with increasing test solution concentration, from around 43.3 (150 mg/L) to 42.3 L/m2 h (20,150 mg/L) for ZnSO4, 42.9 (400 /L) to 42.3 L/m2 h (20,400 mg/L) for NaF, 42.8 (2,000 mg/L) to 41.9 L/m2 h (22,000 mg/L) for MgCl2 and 43.0 (300 mg/L) to 42.4 L/m2 h (20,300 mg/L) for Na2CO3. The flux was measured after 3 hours in each case at a feed temperature of about 80 °C, coolant temperature about 27 °C, feed flow rate 0.5 ± 0.01 L/min and coolant flow rate about 0.25 ± 0.01 L/min.
Figure 4

Effect of different concentrations on permeate flux (feed temp. = 80 °C, feed flow rate = 0.5 L/min, coolant temp. = 27 °C).

Figure 4

Effect of different concentrations on permeate flux (feed temp. = 80 °C, feed flow rate = 0.5 L/min, coolant temp. = 27 °C).

The reduction in MEMD permeate flux arises from the increase in latent heat of vaporization with increasing feed concentration. At the same time, both its viscosity and side boundary layer thickness increase, which increases the chances of membrane wetting. Water ‘activity’ in the feed is a function of temperature and decreases with increasing feed solute concentration, leading to a decline in permeate flux (Alkhudhiri et al. 2012). If the feed solution concentration increases, the mass and heat transfer coefficients decrease on the feed side, due to concentration polarization and a reduction in the surface membrane temperature. Hence the vapor pressure of the feed decreases, which leads to a decrease in the permeate flux (Lawson & Lloyd 1997).

Throughout the process in Figure 4, the permeate flux in each case were measured over 3 hour periods. As the feed concentration increased the permeate flux decreased slightly. Hence the feed concentration had no significant influence on permeate flux in the MEMD process over shorter periods. In the long term, membrane fouling will occur at higher solute concentrations.

Effect of feed flow rate on MEMD performance

The effect of feed flow rate on permeate flux for different feed solution concentrations is presented in Figure 5. In this part of the work, the effect of different flow rates – e.g., 0.3, 0.4, 0.5, 0.6, 0.8 and 1 L/min – on permeate flux was studied. The flux was measured for 1 hour in each case. The module was drained and the membrane surface cleaned using de-ionized water. When the feed flow rate was increased from 0.3 to 1 L/min, the permeate flux increases of 26.5 to 29% for were observed for all feed solutions. Initially, when the feed flow rate was increased from 0.3 to 0.5 L/min, for all four solutes, the flux rate rose more than it did for feed flow rates from 0.5 to 1 L/min. The rise in permeate flux with feed flow rate is thought to be due to increased turbulence at the membrane surface, reduced heat transfer resistance in the boundary layer, and an increase in the mass transfer coefficient. An increase in feed flow rate reduces the difference between the bulk and membrane surface concentrations, thus reducing the concentration polarization effect. It also reduces the temperature difference between the feed and the membrane surface, causing a temperature polarization effect due to an increase in the Reynolds number (Calabro et al. 1994; Xu et al. 2009; Hwang et al. 2011; Al-Anezi et al. 2012).
Figure 5

Effect of feed flow rate on permeate flux for (a) ZnSO4, (b) NaF, (c) MgCl2 and (d) Na2CO3 (feed temp. = 80 °C, coolant temp. = 27 °C).

Figure 5

Effect of feed flow rate on permeate flux for (a) ZnSO4, (b) NaF, (c) MgCl2 and (d) Na2CO3 (feed temp. = 80 °C, coolant temp. = 27 °C).

Effect of feed temperature on MEMD performance

The MEMD's permeate flux and performance depends largely on feed temperature. The permeate fluxes for different aqueous solutions are presented in Figure 6. Increasing the feed temperature leads to an exponential rise in permeate flux because of the effect of temperature on vapor pressure – i.e., at low temperatures, vapor pressures are low, but they are higher at high temperatures. This increases the mass transfer driving force leading to the flux increase. The temperature polarization effect also decreases with increasing feed temperature (Srisurichan et al. 2006; Matheswaran et al. 2007; Khalifa et al. 2015). The results show that, when the feed temperature is increased from 40 to 80 °C, the difference in performance of the MEMD observed for the different solution types were small enough to be ignored.
Figure 6

Effect of MEMD feed temperature on permeate flux for (a) ZnSO4, (b) NaF, (c) MgCl2 and (d) Na2CO3 (feed flow rate = 0.5 L/min, coolant temp. = 27 °C).

Figure 6

Effect of MEMD feed temperature on permeate flux for (a) ZnSO4, (b) NaF, (c) MgCl2 and (d) Na2CO3 (feed flow rate = 0.5 L/min, coolant temp. = 27 °C).

MEMD process separation efficiency

During all experiments, the produced (permeate) water quality of the MEMD was measured. The individual component separation efficiency is presented in Figure 7. The separation efficiency of MEMD exceeded 99.91% for all solutions tested during the 3-hour tests. Also, it will be remain same over several months of MEMD test.
Figure 7

MEMD separation efficiency for different solute concentrations, by solute.

Figure 7

MEMD separation efficiency for different solute concentrations, by solute.

Long-term MEMD performance

A trial was carried out to check the feasibility of long-term use of the 4-stage MEMD process for water treatment. The unit was operated continuously 30 hours for each of the different aqueous solutions – see Figure 8. The feed and coolant temperatures were about 80 and 27 °C, with flow rates of about 0.5 and 0.25 L/min respectively. The permeate quality remained the same throughout the trials and, in each case, the flux reduction became more severe when the feed concentration increased. During the 30-hour operation, the permeate flux reduction for the low concentration solutions was negligible (<1%). There was a reduction in permeate flux for the high concentration solutions, however. The flux reduction observed were 7.94% for ZnSO4, 9.5% for NaF, 8.71% for MgCl2, and 7.24% for Na2CO3. It is thought most likely that the flux reduction was caused by membrane fouling at the high concentrations. This is due to the component's cake formation on the membrane surface. Hence the mass transfer resistance was increased caused the reduction in the permeate flux.
Figure 8

Effect of time on MEMD permeate flux for (a) ZnSO4, (b) NaF, (c) MgCl2 and (d) Na2CO3 (feed temp. = 80 °C, feed flow rate = 0.5 L/min, coolant temp. = 27 °C).

Figure 8

Effect of time on MEMD permeate flux for (a) ZnSO4, (b) NaF, (c) MgCl2 and (d) Na2CO3 (feed temp. = 80 °C, feed flow rate = 0.5 L/min, coolant temp. = 27 °C).

The membrane fouling was removed by washing the membrane. If it was washed with de-ionized water, the fouling agent was removed easily from the membrane surface by simple flushing because the deposit adhered only loosely. Conditions were unpressurized so no hard or firm layer formed. The washing procedure (with de-ionized water) was applied only to the MEMD module's feed channels. Figure 8 shows that the permeate flux was restored almost to its original level when the module was washed. The difference between the initial and post-washing fluxes was less than 0.5% for all four concentrated solutions. It is clear from this that washing is effective in removing fouling. The modules needed to be washed 12 or 14 times, depending on the membrane surface morphology, so some fouling may have remained, perhaps because the membrane's surface morphology changed after repeated washing.

Energy consumption

Figure 9 shows the energy consumption in relation to permeate flux rate for different feed solutions. The consumption was calculated after recovery of the latent heat of vaporization and sensible heat of the hot brine. After heat recovery into the feed water, the latter's temperature increases and the external heat requirement is lower. The specific energy consumption of the 4-stage MEMD for the low and high range concentrated solutions was about 0.53 and 0.548 kWh/kg respectively.
Figure 9

Energy consumption for different solutions and solute concentrations (feed temp. = 80 °C, feed flow rate = 0.5 L/min).

Figure 9

Energy consumption for different solutions and solute concentrations (feed temp. = 80 °C, feed flow rate = 0.5 L/min).

CONCLUSIONS

A study was performed using a 4-stage MEMD unit with four different inorganic solutions (ZnSO4, NaF, MgCl2 and Na2CO3). The MEMD's performance in terms of permeate flux was measured at different solute concentrations, feed temperatures, and feed flow rates. Permeate flux decreased slightly with increasing solute concentration, but increased with increasing feed temperature and flow rate. The MEMD unit performed well in individual cycles of 30 hours. It would be useful to know its true long term performance but that would require operations over several months at the least. Flux decline at high solute concentrations occurred quite early within the 30-hour trial period due to the deposition of precipitated components on the membrane surface. When the membrane was washed, the permeate flux was restored to initial levels (less than 0.5% difference between initial and post-washing). The separation efficiency – the rejection factor – was almost stable (more than 99.91%) for the four different solutions tested. Specific energy consumption is independent of the solute being separated. The specific energy consumption increased at high concentration solutions for each solute type.

ACKNOWLEDGEMENT

We are deeply indebted to the Chemical Engineering Department of College of Engineering and Technology, Akola which is affiliated to the Sant Gadge Baba Amravati University, Amravati (India) for making the laboratory facility available for this research.

REFERENCES

REFERENCES
Al-Anezi
A. A.
Sharif
A. O.
Sanduk
M. I.
Khan
A. R.
2012
Experimental investigation of heat and mass transfer in tubular membrane distillation module for desalination
.
ISRN Chemical Engineering
2012
,
1
8
.
Cipollina
A.
Disparti
M. G.
Tambarlni
A.
Micale
G.
2012
Development of a membrane distillation module for solar energy seawater desalination
.
Chemical Engineering Research and Design
90
(
12
),
2101
2121
.
Curcio
E.
Drioli
E.
2005
Membrane distillation and related operations - a review
.
Separation and Purification Reviews
34
,
35
86
.
Lawson
K. W.
Lloyd
D. R.
1997
Membrane distillation
.
J. Membr. Sci.
124
,
1
25
.
Li
N. N.
Fane
A. G.
Ho
W. S.
Matsuura
T.
2008
Advanced Membrane Technology and Applications
.
John Wiley and Sons, Inc.
,
Hoboken, New Jersey
.
Liu
G. L.
Zhu
C.
Cheung
C. S.
Leung
C. W.
1998
Theoretical and experimental studies on air gap membrane distillation
.
Heat Mass Transfer
34
,
329
335
.
Lu
S. J.
Gao
Q. J.
Lu
X. L.
2012
Device and process study on vacuum multiple-effect membrane distillation
.
Advanced Materials Research
573
,
120
125
.
Manella
G. A.
Carrubba
L.
Brucato
V.
2010
Characterization of hydrophobic polymeric membranes for membrane distillation process
.
Int. J. Material Forming
3
(
1
),
563
566
.
Matheswaran
M.
Kwon
T.
Kim
J.
Moon
I.
2007
Factors affecting flux and water separation performance in air gap membrane distillation
.
J. Ind. Eng. Chem.
13
(
6
),
965
970
.
Pangarkar
B. L.
Deshmukh
S. K.
Sapkal
V. S.
Sapkal
R. S.
2016
Review of membrane distillation process for water purification
.
Desalination and Water Treatment
57
(
7
),
2959
2981
.
Xu
Z.
Pan
Y.
Yu
Y.
2009
CFD Simulation on membrane distillation of NaCl solution
.
Front. Chem. Eng. China
3
(
3
),
293
297
.
Zhao
K.
Heinzel
W.
Wenzel
M.
Buttner
S.
Bollen
F.
Lange
G.
Heinzl
S.
Sarda
N.
2013
Experimental study of the memsys vacuum-multi-effect membrane distillation (V-MEMD) module
.
Desalination
323
,
150
160
.