- Email: [email protected]
Field evaluation of membrane distillation technologies for desalination of highly saline brines
Desalination 351 (2014) 101–108
Contents lists available at ScienceDirect
Desalination journal homepage: www.elsevier.com/locate/desal
Field evaluation of membrane distillation technologies for desalination of highly saline brines Joel Minier-Matar a, Altaf Hussain a, Arnold Janson a, Farid Benyahia b, Samer Adham a,⁎ a b
ConocoPhillips Global Water Sustainability Centre (GWSC), Qatar Science and Technology Park (QSTP), P. O. Box 24750, Doha, Qatar Department of Chemical Engineering, Qatar University, P.O. Box 2713, Doha, Qatar
H I G H L I G H T S • • • • •
Two MD pilot units were tested side-by-side to treat seawater and thermal brines. MD was able to consistently produce high quality distillate (b 10 μs/cm). Operating conditions were optimized; unit A operated at 6.2 LMH and 52% recovery. Pretreatment is critical to avoid wetting and ensure proper system performance. Antifoam wets the MD membranes and activated carbon was an effective pretreatment.
a r t i c l e
i n f o
Article history: Received 14 May 2014 Received in revised form 20 July 2014 Accepted 23 July 2014 Available online xxxx Keywords: Membrane distillation Seawater Thermal desalination brine Pilot plant Thermal brine
a b s t r a c t Membrane distillation (MD) is a hybrid thermal-membrane desalination process that uses low-grade waste heat and hydrophobic membrane to produce high quality distillate. The MD process can treat highly saline brines that other conventional desalination processes cannot treat. These unique features of the MD process make it an ideal candidate to desalinate concentrated brines from thermal desalination plants to augment fresh water production from existing facilities. A consortium consisting of ConocoPhillips Global Water Sustainability Center, Qatar University, and Qatar Electricity & Water Company was formed to evaluate the application of MD for the desalination of concentrated brines from thermal plants. Five different MD technologies were evaluated and the two most suitable technologies were selected for field-testing. The pilot units A & B are based on multi-effect vacuum and air gap MD technologies, respectively. These units were tested side-by-side at a full-scale thermal desalination plant in Qatar. Pilot unit A showed a stable flux of 6.2 LMH under optimized conditions with excellent salt rejection (N99.9%). Pilot unit B achieved a distillate flux of 2.5 LMH and salt rejection greater than 98.9%. Overall, MD was shown to be a feasible technology to produce potable quality water from the brines discharged from thermal desalination plants. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Membrane distillation (MD) is an emerging hybrid thermalmembrane desalination process that uses a vapor pressure difference, created by a temperature gradient across a hydrophobic membrane, as the driving force to produce high quality distilled water (Fig. 1) [1]. A temperature difference as low as 10 °C–20 °C between the warm and cold streams is sufficient to produce distilled water under the right conditions. Key advantages of MD include: • Production of distilled water • Treatment of high salinity brines ⁎ Corresponding author. E-mail address: [email protected] (S. Adham).
http://dx.doi.org/10.1016/j.desal.2014.07.027 0011-9164/© 2014 Elsevier B.V. All rights reserved.
• Lower greenhouse gas emissions • Utilization of low grade waste heat or renewable energy resulting in lower operating costs • Minimization of capital equipment costs by using inexpensive plastics due to the ambient operating pressures and low operating temperatures. Other desalination processes that have been operating at a largescale in the Middle-East include reverse osmosis (RO), multi-stage flash (MSF) and multi-effect distillation (MED). MD has a potential niche application to desalinate concentrate brines from these processes. One reason is that the brine from thermal desalination plants contains pretreatment chemicals that assist in reducing the scaling on the MD membrane. In addition to scale minimization, using thermal brines as the feed water reduces the energy required as the feed is already significantly preheated.
102
J. Minier-Matar et al. / Desalination 351 (2014) 101–108
investigated and different pretreatment options were evaluated. The critical design parameters needed to improve process efficiency at a full-scale level were identified and optimized. 2. Materials and methods 2.1. Description of pilot units A Request for Proposal (RFP) for a 1 m3/d pilot unit was issued to five leading MD technology providers. After a technical and commercial assessment of the bids, two pilot systems were selected for field evaluation. The two systems represent different MD technologies: • Pilot unit A: vacuum multi-effect MD unit, with 4 effects (Fig. 2) • Pilot unit B: air gap MD unit, single effect (Fig. 3).
Fig. 1. Schematic of the MD process.
Furthermore, the location of thermal processes alongside with power plants results in the availability of potential low-grade waste heat (e.g. boiler blowdown, dump condenser) that can be used as the heat source. Moreover, by retrofitting existent thermal desalination plants with an MD plant, water production could be increased with minor capital investment and scarce potable water resources could be redirected for priority end uses. Most of the MD pilot studies available in the literature focus on seawater desalination, mainly by harnessing the solar energy either by photovoltaic panels or solar collectors with heat storage mediums [2–5]. Some researchers have also studied the optimization of MD systems for specific applications [6]. An example of an optimization study is the evaluation of multi-effect MD systems where the results showed a significant increase in the process energy efficiency [7]. Researchers have also studied issues related to scaling and fouling. The deposition of calcium carbonate on the surface of MD membranes has been investigated during the demineralization of lake water using a MD bench scale system. Results showed that the morphology of the formed deposit has a significant influence on the permeate flux decline and a higher flux was obtained with less compact deposit layers [8]. Other researchers have studied the mitigation of calcite and gypsum by the application of antiscalants. Results showed that a certain type of antiscalants extends the induction period for the nucleation of gypsum and calcite without affecting membrane performance [9]. In another study, the scaling kinetics of calcium carbonate were investigated. Induction time measurements were carried out using dynamic light scattering to identify the shifting between homogeneous and heterogeneous nucleation mechanisms as a function of supersaturation. Membrane cleaning was also investigated using a two-step process consisting of citric acid and sodium hydroxide [10]. Like any emerging process, MD has potential challenges associated with it: limited experience on scale-up, process design and pretreatment issues on seawater/brine sources. In order to advance the technical knowledge in those areas, a consortium consisting of the ConocoPhillips Global Water Sustainability Center (GWSC), Qatar University (QU), and Qatar Electricity & Water Company (QEWC) was formed to evaluate the feasibility of testing the MD process at a pilot scale for desalination of brines from local thermal plants. The primary objective of this research paper is to evaluate the performance of MD technologies on seawater and brines from thermal desalination plants. Two leading MD technologies were pilot tested to treat brines from thermal plants in Qatar. The present MD pilot study is unique as it addresses the desalination of concentrated thermal brine by evaluating two leading MD technologies side-by-side under continuous (24/7) operation at field conditions. The effect of the chemicals, from the MSF process, on the membranes was also
Each MD system was provided with pumps, immersion heaters, water chillers, tanks and controls in a 20 ISO shipping container. Both pilot units were equipped with software that allowed remote access and control. The specifications of the pilot units are summarized in Table 1. 2.2. Modes of operation Pilot unit A operates in a continuous flow, one pass mode. After the feed water enters the system, it is heated and then allowed to flow across the different effects in series. In the last effect, it is rejected as a concentrated stream into an external tank. Distillate is produced at each of the system's four effects and is collected in a separate external tank (Fig. 2). The latent energy released during condensation in each effect is transferred to the incoming feed of the subsequent effect. Cooling water is only applied to the final effect to remove the heat from the last effect. Pilot unit B operates in feed and bleed mode, i.e., the feed water enters the module and the concentrated feed is recirculated back into the feed tank while the distillate is collected separately in an external tank. Therefore, the feed water gets concentrated in the feed tank and once the concentrated water reaches a predetermined conductivity value, part of the tank is drained and fresh feed is added to maintain a constant conductivity/salt concentration (Fig. 3). 2.3. Chemical analysis The chemical compositions of the thermal brine, seawater and distillate from each of the units were analyzed using the analytical methods as listed in Table 2.
Fig. 2. Pilot unit A — schematic.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108
103
Table 2 Analytical methods. Parameter
Method
Instrument
Cations (Sodium, Magnesium, Calcium, Potassium) Anions (Chloride, Sulfate, Bromide)
Separation by IonPac® CS12A — 4 × 250 mm column with MSA eluent; ions measured by conductivity detector Separation by IonPac® AS19 — 4 × 250 mm column with EGC III KOH eluent; ions measured by conductivity detector Combustion method
Ion chromatography (ICS 3000, Dionex)
Total Organic Carbon (TOC) Strontium, Boron Conductivity pH
Fig. 3. Pilot unit B — schematic.
2.4. Operating conditions and testing program The operating conditions used during the experiments are summarized in Table 3. Those parameters were recommended by the manufacturer and used throughout the experiments (unless specified otherwise). 2.5. Preliminary testing Preliminary testing was performed at a temporary site at QU with the objective of ensuring that the pilot units met the design specifications before relocating them to the thermal desalination plant. These tests included baseline performance and the effect of temperature and salinity on flux. 2.5.1. Baseline performance test Baseline tests were conducted with tap water as feed using the operating conditions recommended by the manufacturer (Table 3); results are summarized in Table 4. Results showed that pilot unit A achieved distillate flux of 5 LMH with a temperature difference (ΔT) across the membrane of 25.6 °C (6.4 °C per effect; first effect temperature of 59.5 °C). Pilot unit B produced a flux 20% higher than unit A (6.1 vs. 5.0 LMH). Although the flux in pilot unit B was higher than that of A, it was achieved with a higher ΔT across the membranes (35 vs. 25.6 °C) and a higher temperature in the first effect (69 vs. 59.5 °C) 2.5.2. Effect of temperature Temperature is a critical operating parameter since it dictates the driving force across the membrane and therefore the distillate flux. Different operating temperatures were tested to determine the flux vs.
Table 1 MD pilot units' specification. Parameter
Pilot Unit A
Pilot Unit B
No. of modules & arrangement No. of effects per module Total membrane area (m2) Membrane characteristics
1 4 6.4 Pore size: 0.2 μm, Active layer: PTFE, Backing: PP Temperature, pressure, flow & conductivity Immersion heaters External water chiller 70–100 (vacuum)
2 in parallel 1 4.6
Operating parameters logged in PLC Feed water heating element Cooling water element Operating pressure (mbar)
Temperature, pressure, flow, conductivity & pH
1000 (ambient)
Inductively Coupled Plasma, Axial mode Immersing conductivity probe in the sample Immersing pH probe in the sample
Shimadzu TOC Analyzer TOC-V Thermo ICAP 6500 Orion 3 Star conductivity meter Orion 3 Star pH meter
vapor pressure (ΔVp) relationships for the two systems. Results confirmed that there is a linear correlation between the ΔVp across the membrane and the flux (Fig. 4). It was also observed that a 10 °C increase in feed temperature (from 60 °C to 70 °C) increased the flux by 44% while a 10 °C decrease in cooling water temperature (from 30 °C to 20 °C) only resulted in a 17% increase in flux. This is attributed to the exponential dependency of the vapor pressure as function of temperature [1]. Even at low ΔT, both systems produced considerable amounts of distillate; at a ΔT of only 12 °C, pilot unit B flux was 1.1 LMH, equivalent to a production rate of 120 L/d.
2.5.3. Effect of salt concentration The effect of salt concentration was evaluated using two different NaCl solutions (35, 70 g/L). As shown in Fig. 5, the flux obtained with 70 g/L NaCl is 20% lower than the flux obtained during the 35 g/L experiment, which is consistent with the results from bench-scale testing published earlier by the authors [11]. Pilot unit B showed similar behavior. A salt rejection of N 99.9% was achieved by both pilot units, thereby meeting the specifications provided by the manufacturer.
3. Results & discussion 3.1. Tests at desalination facility After the completion of the preliminary tests, the two pilot units were relocated to a local thermal desalination facility and tested, sideby-side. Both pilot units were operated with brine (~70 g/l TDS) from the thermal desalination plant continuously for 24 h a day, 7 days a week. The thermal brine was pretreated using a two-step filtration system (1 μm + Granular Activated Carbon) to remove particulates and organic contaminants present in the feed stream that could potentially affect the integrity of the membranes. The feed TDS remained constant in the presence/absence of the GAC filter. This conservative pretreatment gave both systems the best chance of success. Table 3 Standard operating conditions for each pilot unit. Parameter
Units
Pilot Unit A
Pilot Unit B
Temperature feed inlet (TF) Temperature hot side (THS) Temperature cold side (TCS) Feed flow rate (QF) Hot side flow rate (QHS) Cold side flow rate (QCS) Pressure (P)
°C °C °C L/min L/min L/min mbar
25–35 70 20 1.4 16.5 13.5 70
25–35 70 20 30 NA 40 1000
104
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 7
Table 4 Baseline performance for each pilot unit. Pilot unit A
Pilot unit B
Feed Flux Distillate conductivity range Distillate produced First effect inlet temperature ΔT Recovery based on volume
– LMH μS/cm L/day °C °C %
Tap water 5.0 ± 0.4 0.2–1.0 757 ± 77 59.5 ± 1.0 25.6 ± 1.6 39 ± 3.6
6.1 ± 0.3 1–10 670 ± 40 69 ± 1.0 35 ± 0.7 NA
6 5 4 3 2
As shown in Fig. 6, pilot unit A was able to consistently generate high quality product water (TDS b 10 mg/L) over a 12-day period, while maintaining a stable flux of 4.5 LMH. Due to the high salt concentrations, a 10% reduction in flux was observed when compared to the tap water baseline tests (5.0 LMH). The ΔT and vacuum pressure in Fig. 6 are 19.6 °C and 70 mBar, respectively. During the first 3 days of operation, the distillate conductivity was averaged at 2.5 μS/cm, after which a steady increase of conductivity was observed until it stabilized at ≈ 10 μS/cm. Recovery rate was 34% and the concentrated MD brine had a TDS of 112 g/L. Table 5 shows the water quality analyses of major anions and cations. Salt rejections greater than 99.9% were achieved. Results indicate that the pretreatment was effective in removing potential contaminants that could compromise the integrity of the membranes (see pretreatment Section 3.2). Using thermal brine as the feed water, pilot unit B generated a poorer product water quality (2530 μS/cm). As shown in Fig. 7, a significant decrease in flux (from 4 to 2.5 LMH) was observed throughout the testing period. Therefore, the system was stopped after 5 days of operation. Table 5 shows the water quality data; approximately 99% salt rejection was achieved (feed tank concentration was used as reference). The decrease in the flux could be attributed to the feed and bleed mode of operation; as the thermal brine in the feed tank concentrates, the water flux decreases. The poor salt rejection could possibly be due to a small mechanical leakage in the membrane modules, but that was not confirmed by the authors. The inspection of pilot unit B membrane modules was very challenging due to the heavy weight and design of the modules, which required the presence of the technology provider staff to make the autopsy. It was also learned during the study that the provider is developing a newer membrane module design with lighter weight, which makes them more attractive for full-scale installations. Thus, the project team decided to end the field testing program with pilot unit B, hoping for future testing opportunities with the new & lighter design membrane modules. After the tests were completed, both systems were flushed with tap water. It was observed that both systems regained their original flux, confirming that no severe membrane fouling occurred during the test
0 0
3
Pilot Unit A
2
Pilot Unit B
2
3
4
5
6
7
8
Fig. 5. Pilot unit A: effect of salt concentration on water flux. TF: 25 °C, THS: 70 °C, TCS: 30 °C, QF: 1.4 L/min, P: 70 mbar, ΔT35,000: 22.1 °C, ΔT70,000: 23.6 °C.
with thermal brine. For pilot unit B, the conductivity of the tap water used to flush the membranes was measured and no increase was observed after flushing; indicating that the poor rejection may not be due to scaling of the membranes. During the tap water cleaning, the distillate conductivity was also high; the rejection of the system B was approximately 70%. 3.2. Pretreatment evaluation To assess the impact of pretreatment on the performance of the system, tests were conducted by removing the granular activated carbon (GAC) filter from the upstream of the MD process. During the first 4 days of operation, pilot unit A was able to generate a distilled quality of conductivity b 1 μS/cm. The water quality started to deteriorate on the 5th day with a decrease in distillate flux; the distillate conductivity increased exponentially to ≈700 μS/cm as shown in Fig. 8. The increase in distillate conductivity and the slight decrease in flux were attributed to membrane wetting caused by residual chemicals present in the brine that were not removed by the cartridge filters. This test confirmed the importance of the activated carbon filter in removing process chemicals that could potentially compromise the hydrophobicity of the membranes. A laboratory investigation was carried out to study the different types of process chemicals (antiscalant/antifoam) that may affect the hydrophobicity of the membranes in pilot unit A. The membranes were tested using a bench-scale MD unit. The state-of-the-art unit used for this study is described in reference [11]. The first test was conducted by spiking a 70 g/L NaCl solution with antiscalant. As shown in
Distillate Flux, LMH
Distillate Flux (LMH)
4
1
Time, hours
7
5
70 g/L NaCl
1
8
6
35 g/L NaCl
7
1000000
6
100000
5
10000
4
1000
Flux Brine conductivity Feed conductivity
3
100
Distillate conductivity
2
10
1
1
Conductivity, µS/cm
Units
Distillate Flux, LMH
Parameter
1 0
0 0
0 0
50
100
150
200
250
1
2
3
4
5
6
7
8
9
10
11
12
Time, days
Vapor pressure difference (mbar) Fig. 4. Pilot units A & B: vapor pressure difference vs. water flux.
Fig. 6. Pilot unit A: performance on thermal brine. TF: 32 °C, THS: 70 °C, TCS: 20 °C, QF: 1.4 L/min, P: 70 mbar, ΔT: 19.6 °C.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108
105
Table 5 Pilot units A & B: water quality on thermal brine. Pilot unit A
Pilot unit B
Parameter
Units thermal Brine (n = 4)
Distillate (n = 4)
Reject brine (n = 4)
Rejection (%)
Thermal brine (n = 1)
Distillate (n = 1)
Reject brine (n = 1)
Rejection (%)
TDS Chloride Sodium Sulfate Magnesium Potassium Calcium Bromide Strontium Boron TOC
mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
6 2 2 1.2 0.3 b0.2 b1 b0.1 b0.2 b0.2 b1
112,279 64,838 33,258 7264 3929 1193 1423 110 17 12 4
99.99 99.99 99.99 99.97 N99.99 N99.97 N99.87 N99.8 N98.2 N98 –
68,529 39,064 20,375 4996 2435 770 703 55 11 10 3
1472 840 496 69 49 19 17 1 b0.5 b0.5 b1
136,191 78,281 38,926 10,885 4565 1394 1328 156 21 16 5
98.92 98.93 98.73 99.37 98.93 98.62 98.71 99.04 N97.6 N96.8 –
71,031 40,988 21,310 4344 2519 779 775 64 11 8 3
Fig. 9, the antiscalant concentration was increased periodically from 2 mg/L to 800 mg/L and the distillate conductivity was monitored. There was no significant change in the distillate conductivity. This observation rules out the possibility of the antiscalant causing membrane wetting. A similar test was conducted by spiking the salt solution with the antifoam agent. The antifoam agent concentration was increased from 0.2 mg/L to 50 mg/L. The distillate conductivity sharply increased at an antifoam agent concentration of 50 mg/L as shown in Fig. 9. This confirms that the antifoam agent is the chemical responsible for the membrane wetting. In a typical MSF process, the antifoam agent is dosed at a concentration between 0.1 and 0.3 mg/L. When the pilot unit received 10,000 L of thermal brine after five days of operation, the membrane was exposed to at least 1000 mg of antifoam agent; this was sufficient to wet the membrane causing the deterioration of the distillate water quality discussed earlier. To further confirm that the antifoam agent was the chemical responsible for membrane wetting, membrane coupons were soaked in six individual beakers for five days. The beakers contained the following solutions: nanopure water, concentrated brine, 70 g/L NaCl, 70 g/L NaCl with 100 mg/L antiscalant, 70 g/L NaCl with 100 mg/L antifoam and 70 g/L NaCl with both antiscalant (100 mg/L) and antifoam (10 mg/L). The membranes were dried and a couple of nanopure water droplets were added on top of the membrane as shown in Fig. 10. Wetting was only observed in the coupons that were in contact with the antifoam agent; supporting the conclusion that this chemical was the main cause of membrane wetting. The removal of process chemicals by activated carbon was also investigated. 100 mL of a 1000 mg/L antifoam and antiscalant solutions
3.3. Performance on seawater The test with seawater was performed using pilot unit A with 1 μm filtration as pretreatment. The chlorinated seawater was collected at the intake, prior to the addition of antiscalant and antifoam agents. Results showed that the system was able to operate steadily for 8 days with distillate flux of 4.8 LMH (Fig. 11) and recovery of 30%. Since antiscalant was not added, the recovery was limited to 35% to avoid salt precipitation in the membranes and it was controlled by slightly increasing the feed flow rate. The TDS of the seawater was increased from 44 g/L to 66 g/L. As shown in Table 6, excellent salt rejection of greater than 99.9% was achieved for the major ions. These results demonstrated the capability of MD to desalinate seawater, generating excellent quality distillate. 3.4. Pilot unit A process optimization A set of experiments was conducted to determine the main parameters that influence the water production in pilot unit A. The parameters were optimized for maximum recovery, water production and energy efficiency. All tests were performed using thermal brine as a feed and a two-step pretreatment consisting of 1 μm + GAC filters.
Feed & Bleed mode 7
1000000
6
100000
6
100000
5
10000
5
10000
4
1000
4
1000
3
100
3
100
2
10
2
1
Brine conductivity
1
Feed conductivity
1
Distillate conductivity
0
0 1
2
3
4
5
1
Feed conductivity
Distillate conductivity
0
10
Flux
Flux
Conductivity, µ S/cm
1000000
Distillate Flux, LMH
7
Conductivity, µS/cm
Distillate Flux, LMH
Concentraon stage
were treated individually with 500 mg of GAC. After adding the GAC to the individual beakers, the solutions were stirred for 30 min and filtered using 0.22 μm Nylon filter. Total Organic Carbon (TOC) results showed that GAC removes 21% and 23% of the antiscalant and antifoam agent, respectively. Higher removal could be achieved by GAC packed column. The portion of the antifoam agent that was removed in this experiment may have been responsible for membrane wetting.
6
Time, days Fig. 7. Pilot unit B: performance on thermal brine. TF: 30 °C, THS: 70 °C, TCS: 20 °C, ΔT: 35 °C.
0
0 0
1
2
3
4
5
6
Time, days Fig. 8. Pilot unit A: Performance on thermal brine without activated carbon filter. TF: 32 °C, THS: 70 °C, TCS: 20 °C, QF:1.4 L/min, P: 70 mbar, ΔT: 21.4 °C.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 800
Antiscalant (AS)
600
Distillate Flux, LMH
Conductivity, µS/cm
700
Antifoaming (AF)
500 400 300 AS - 2 ppm AF - 0.2 ppm
200
AS - 10 ppm AF - 1 ppm
AS - 100 ppm AF - 10 ppm
AS - 800 ppm AF - 50 ppm
100
7
1000000
6
100000
5
10000 Flux
4
1000
Brine conductivity
Feed conductivity Distillate conductivity
3
100
2
10
1
1
0
Conductivity, µS/cm
106
0 0
1
2
3
4
5
6
7
8
9
Time, days
0 0
50
100
150
200
250
300
350
400
450
500
Time, min
Fig. 11. Pilot unit A: Performance on seawater. TF: 32 °C, THS: 70 °C, TCS: 20 °C, QF: 1.7 L/min, P: 70 mbar, ΔT: 21.9 °C.
Fig. 9. Impact of scale inhibitor and antifoam agent on pilot unit A membrane.
3.4.1. Effect of thermal brine temperature The objective of this test was to evaluate the effect of thermal brine temperature (prior to entering the system) on the membrane flux. The brine was preheated using an external tank fitted with immersed electrical heaters. This preheating permitted the system to be operated at ΔTs higher than the ΔTs used in earlier test (standard conditions). As expected, operation with preheated thermal brine resulted in an increase in distillate production. Experimental results are summarized in Fig. 12a and Table 7.
3.4.2. Effect of operating vacuum The effect of vacuum pressure on the system was also investigated. Three different pressures were tested: 70, 100 and 130 mbar. Lowering the vacuum resulted in higher flux as shown in Fig. 12b.
3.4.3. Feed water flow rate The objective of this test was to evaluate the effect of the feed flow rate on the % recovery. The target was to achieve a recovery of 50%. Prior to conducting the test, preliminary lab experiments were performed to confirm that operating the system at 50% recovery would not cause the salts to precipitate. Results indicated that the system could be operated at a maximum recovery of 70%, after which salts would start to precipitate. Such high recoveries can be achieved, without precipitation,
due to the antiscalant that is added to the thermal brine as part of the MSF process. Three feed flow rates (1.1 L/min, 1.2 L/min & 1.5 L/min) were investigated; flux remained constant at 5 LMH for the three flow rates tested. While operating the pilot unit at 1.1 L/min, the system achieved a recovery of 48% as shown in Fig. 12c. 3.5. Test at optimum conditions A final test was conducted using the optimized conditions summarized in Table 8. The feed temperature prior to entering the system was maintained at 50 °C since this is the temperature at which the thermal brine is typically rejected during the summer. The internal electric heater of the system was used to increase the temperature from 50 °C to 75 °C. Results showed that at optimum conditions, distillate production could be increased by 40% in comparison with standard conditions (690 L/day vs. 950 L/day). 52% recovery and distillate water with an average conductivity of ≈ 15 μS/cm were achieved at a stable flux of 6.2 LMH (Fig. 13). 4. Summary & conclusions In Qatar, the production of fresh water is primarily achieved using thermal desalination plants that are co-located with power plants; installations that provide both hot concentrated brines and low grade
Nanopure Water
Thermal Brine
70,000 mg/L NaCl
70,000 mg/L NaCl+ 100 mg/L Anscalant
70,000 mg/L NaCl + 100 mg/L Anfoam
70,000 mg/L NaCl+ 100 mg/L A + nscalant 10 mg/L Anfoam
Fig. 10. Image of membrane coupons from pilot unit A after soaking in different solutions.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 Table 6 Pilot unit A; water quality on seawater. Parameter
Units
TDS Chloride Sodium Sulfate Magnesium Potassium Calcium Bromide Strontium Boron TOC
mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
Table 7 Summary of results from optimization experiment.
Seawater (N = 1)
Distillate (N = 1)
Reject brine (N = 1)
Rejection (%)
44,701 25,531 13,421 3263 1336 511 454 36 7 6 1
9 4 4 0.3 0.4 0.3 b1 b0.1 b0.2 b0.2 b1
66,304 38,180 19,232 4927 2263 831 650 55 10 9 2
99.98 99.98 99.97 99.99 99.97 99.93 N99.7 N99.7 N97.2 N96.6 –
(a) 7 Disitllate Flux, LMH
6
Flux 5.7 Flux 5.0
5 4
Parameter
Units
Feed brine temperature Flux First stage temperature ΔT Operating vacuum Feed water flow rate Recovery Operating vacuum Feed brine temperature Flux First stage temperature ΔT Feed water flow rate Recovery Feed water flow rate Recovery Feed brine temperature Flux First stage temperature ΔT Operating vacuum
°C LMH °C °C mbar L/min % mbar °C LMH °C °C L/min % L/min % °C LMH °C °C mbar
Results 55 5.7 68.1 20.4 100 1.4 42 70 32 4.9 61.4 19.6 1.4 35 1.1 48 32 5.0 62 21.2 100
45 5.0 66.5 18.4 100 1.4 36 100 32 4.0 64.2 16.8 1.4 29 1.2 45 32 5.0 62 20.8 100
25 3.8 63.9 16.4 100 1.4 27 130 32 3.7 67.2 15.2 1.4 27 1.5 35 32 4.9 61.4 19.8 100
Flux 3.8
3 2 1 0 55
45
25
Thermal BrineInlet Temperature,°C
(b) 7 6
Disllate Flux, LMH
107
Flux: 4.9
5
Flux: 4.0
4
Flux: 3.7
3 2 1 0 70
100
130
Vacuum Pressure,mBar
(c)100%
waste heat as by-products. Membrane distillation (MD) is a hybrid thermal-membrane process that can use low grade waste heat and a hydrophobic membrane to produce a high quality distillate. The MD process can treat highly saline brines that the conventional membrane desalination processes are unable to treat efficiently. A consortium consisting of the ConocoPhillips Global Water Sustainability Center (GWSC), Qatar University (QU), and Qatar Electricity & Water Company (QEWC) was formed to evaluate the application of MD in desalinating brines from full-scale thermal desalination plants for sustainable augmentation of fresh water supplies in Qatar. The main objective was to investigate the performance of the MD technology to desalinate seawater and concentrated brine. Two leading pilot scale MD technologies, vacuum multi-effect (A) and air gap MD (B), were operated in parallel to treat seawater and thermal brine discharged from a local thermal desalination facility under field conditions. Below is a summary of the study's main conclusions: - Pilot unit A achieved distillate flux of 5 LMH on tap water with an average temperature difference (ΔT) across the membrane of 25.6 °C and a feed temperature in the first effect of 59.5 °C. - Pilot unit B produced a flux of 6.1 LMH on tap water with a ΔT across the membranes of 35 °C and a feed temperature in the first effect of 69 °C. - Results from both pilot units using tap water showed that the water flux is directly proportional to the applied vapor pressure difference across the MD membranes.
90% 80%
Table 8 Operating parameters and performance at standard and optimum conditions.
Recovery, %
70% 60%
40%
Parameter
Units
Standard conditions
Optimum conditions
Feed temperature (TF) Temperature hot side (THS) Temperature cold side (TCS) Feed flow rate (QF) Hot side flow rate (QHS) Cold side flow rate (QCS) Pressure (P) First effect inlet temperature ΔT Flux Recovery Production
°C °C °C L/min L/min L/min mbar °C °C LMH % L/day
32 70 20 1.4 16.5 13.5 70 61 19.6 4.5 34 690
50 75 20 1.25 16.5 13.5 55 65.2 26 6.2 52 950
Recovery 48%
50%
Recovery 45% Recovery 35%
30% 20% 10% 0% 1.5
1.2
1.1
Feed flow rate,L/min Fig. 12. Optimization results: (a) effect of feed temperature, (b) effect of vacuum pressure, (c) effect of feed flow rate. Table 7 shows a summary of the main operating parameters.
108
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 7
Acknowledgment
6
Distillate Flux, LMH
Standard (non-optimized)
5
Optimized
4 3 2
Opmized condions Feed Temperature : 50 C Feed Flow rate : 1.25 L/min Vacuum set point : 55 mbar Recovery : 52 %
1
Standard (non-opmized) condions Feed Temperature : 32 C Feed Flow rate : 1.4 L/min Vacuum set point : 76 mbar Recovery : 34 %
0 0
0.5
1
1.5
2
2.5
3
Time, days Fig. 13. Pilot unit A performance: optimized vs. standard conditions. Operating conditions are summarized in Table 8.
- The MD flux obtained while operating with 70 g/L synthetic NaCl solution was 20% lower than the flux obtained during operation with 35 g/L NaCl solution. - Pilot unit A was able to desalinate brines from thermal desalination plants and consistently produce a high quality distillate (conductivity b 10 μS/cm — N99.9% rejection). - Pilot unit A also showed stable flux (4.8 LMH) and produced high quality distillate when operated directly on seawater. - Under optimized conditions, pilot unit A operating on thermal desalination brines achieved a stable flux of 6.2 LMH and a recovery of 52%. - Pilot unit B showed stable performance while operating on thermal desalination brines with significant salt rejection of 98.9%. However, due to challenges experienced with system and the heavy weight of the membrane modules, it was not possible to complete the field testing program with that unit. - Pretreatment is critical due to the possible presence of process chemicals in the brine of the MSF process that could potentially cause MD membrane wetting. - It was found that the antifoam agent resulted in the wetting of MD membranes. Granular activated carbon proved to be an effective pretreatment. - Overall, MD was shown to be a feasible technology to produce potable quality water from the brines discharged from thermal desalination plants.
The research team would like to thank Mr. Fahad Al Muhannadi, General Manager of QEWC, and Mr. Abdulsattar Al Rasheed, CEO of Ras Abu Fontas Desalination Plant, for their support in the overall testing program. The research team thanks the efforts of Dr. Majeda Khraisheh from Qatar University, and research assistants Ahmad Fard and Yehia Manawi for facilitating the operations of the pilot unit at Qatar University. The team acknowledges the MD technology providers for technical support during the operation of the pilot units. The research team would also like to acknowledge their colleagues at the GWSC including, Mr. Raul Dores, Dr. Samir Gharfeh, Dr. Isik Turkmen, Dr. Nabin Upadhyay, Ms. Eman Al Shamari and Ms. Mary Katebah for conducting bench scale experiments and water analyses on the seawater and brine samples.
References [1] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive review, Desalination 287 (2012) 2–18. [2] R. Schwantes, A. Cipollina, F. Gross, J. Koschikowski, D. Pfeifle, M. Rolletschek, V. Subiela, Membrane distillation: solar and waste heat driven demonstration plants for desalination, Desalination 323 (2013) 93–106. [3] F. Banat, N. Jwaied, M. Rommel, J. Koschikowski, M. Wieghaus, Desalination by a “compact SMADES” autonomous solar powered membrane distillation unit, Desalination 217 (2007) 29–37. [4] A. Kullab, A. Martin, Membrane distillation and applications for water purification in thermal cogeneration plants, Sep. Purif. Technol. 76 (2011) 231–237. [5] E.G. Burrieza, G. Zaragoza, S.M. Cuevas, J. Blanco, Experimental evaluation of two pilot-scale membrane distillation module used for solar desalination, J. Membr. Sci. 409–410 (2012) 264–275. [6] A.E. Jansen, J.W. Assink, J.H. Hanemaaijer, J. van Medevoort, E. Van Sonsbeek, Development and pilot testing of full-scale membrane distillation modules for deployment of waste heat, Desalination 323 (2015) 55–65. [7] K. Zhao, W. Heinzl, M. Wenzel, S. Buttner, F. Bollen, G. Lange, S. Heinzl, N. Sarda, Experimental study of the memsys vacuum-multi-effect-membrane-distillation (V-MEMD) module, Desalination 323 (2013) 150–160. [8] Marek Gryta, Alkaline scaling in the membrane distillation process, Desalination 228 (2008) 128–134. [9] F. He, K. Sirkar, J. Gilron, Effect of antiscalants to mitigate membrane scaling by direct contact membrane distillation, J. Membr. Sci. 345 (2009) 53–58. [10] E. Curcio, X. Ji, G.D. Profio, A.O. Sulaiman, E. Fontananova, E. Drioli, Membrane distillation operated at high seawater concentration factors: Role of the membrane on CaCO3 scaling in presence of humic acid, J. Membr. Sci. 346 (2010) 263–269. [11] Samer Adham, Altaf Hussain, Joel Minier Matar, Raul Dores, Arnold Janson, Application of membrane distillation for desalting brines from thermal desalination plants, Desalination 314 (2013) 101–108.
Contents lists available at ScienceDirect
Desalination journal homepage: www.elsevier.com/locate/desal
Field evaluation of membrane distillation technologies for desalination of highly saline brines Joel Minier-Matar a, Altaf Hussain a, Arnold Janson a, Farid Benyahia b, Samer Adham a,⁎ a b
ConocoPhillips Global Water Sustainability Centre (GWSC), Qatar Science and Technology Park (QSTP), P. O. Box 24750, Doha, Qatar Department of Chemical Engineering, Qatar University, P.O. Box 2713, Doha, Qatar
H I G H L I G H T S • • • • •
Two MD pilot units were tested side-by-side to treat seawater and thermal brines. MD was able to consistently produce high quality distillate (b 10 μs/cm). Operating conditions were optimized; unit A operated at 6.2 LMH and 52% recovery. Pretreatment is critical to avoid wetting and ensure proper system performance. Antifoam wets the MD membranes and activated carbon was an effective pretreatment.
a r t i c l e
i n f o
Article history: Received 14 May 2014 Received in revised form 20 July 2014 Accepted 23 July 2014 Available online xxxx Keywords: Membrane distillation Seawater Thermal desalination brine Pilot plant Thermal brine
a b s t r a c t Membrane distillation (MD) is a hybrid thermal-membrane desalination process that uses low-grade waste heat and hydrophobic membrane to produce high quality distillate. The MD process can treat highly saline brines that other conventional desalination processes cannot treat. These unique features of the MD process make it an ideal candidate to desalinate concentrated brines from thermal desalination plants to augment fresh water production from existing facilities. A consortium consisting of ConocoPhillips Global Water Sustainability Center, Qatar University, and Qatar Electricity & Water Company was formed to evaluate the application of MD for the desalination of concentrated brines from thermal plants. Five different MD technologies were evaluated and the two most suitable technologies were selected for field-testing. The pilot units A & B are based on multi-effect vacuum and air gap MD technologies, respectively. These units were tested side-by-side at a full-scale thermal desalination plant in Qatar. Pilot unit A showed a stable flux of 6.2 LMH under optimized conditions with excellent salt rejection (N99.9%). Pilot unit B achieved a distillate flux of 2.5 LMH and salt rejection greater than 98.9%. Overall, MD was shown to be a feasible technology to produce potable quality water from the brines discharged from thermal desalination plants. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Membrane distillation (MD) is an emerging hybrid thermalmembrane desalination process that uses a vapor pressure difference, created by a temperature gradient across a hydrophobic membrane, as the driving force to produce high quality distilled water (Fig. 1) [1]. A temperature difference as low as 10 °C–20 °C between the warm and cold streams is sufficient to produce distilled water under the right conditions. Key advantages of MD include: • Production of distilled water • Treatment of high salinity brines ⁎ Corresponding author. E-mail address: [email protected] (S. Adham).
http://dx.doi.org/10.1016/j.desal.2014.07.027 0011-9164/© 2014 Elsevier B.V. All rights reserved.
• Lower greenhouse gas emissions • Utilization of low grade waste heat or renewable energy resulting in lower operating costs • Minimization of capital equipment costs by using inexpensive plastics due to the ambient operating pressures and low operating temperatures. Other desalination processes that have been operating at a largescale in the Middle-East include reverse osmosis (RO), multi-stage flash (MSF) and multi-effect distillation (MED). MD has a potential niche application to desalinate concentrate brines from these processes. One reason is that the brine from thermal desalination plants contains pretreatment chemicals that assist in reducing the scaling on the MD membrane. In addition to scale minimization, using thermal brines as the feed water reduces the energy required as the feed is already significantly preheated.
102
J. Minier-Matar et al. / Desalination 351 (2014) 101–108
investigated and different pretreatment options were evaluated. The critical design parameters needed to improve process efficiency at a full-scale level were identified and optimized. 2. Materials and methods 2.1. Description of pilot units A Request for Proposal (RFP) for a 1 m3/d pilot unit was issued to five leading MD technology providers. After a technical and commercial assessment of the bids, two pilot systems were selected for field evaluation. The two systems represent different MD technologies: • Pilot unit A: vacuum multi-effect MD unit, with 4 effects (Fig. 2) • Pilot unit B: air gap MD unit, single effect (Fig. 3).
Fig. 1. Schematic of the MD process.
Furthermore, the location of thermal processes alongside with power plants results in the availability of potential low-grade waste heat (e.g. boiler blowdown, dump condenser) that can be used as the heat source. Moreover, by retrofitting existent thermal desalination plants with an MD plant, water production could be increased with minor capital investment and scarce potable water resources could be redirected for priority end uses. Most of the MD pilot studies available in the literature focus on seawater desalination, mainly by harnessing the solar energy either by photovoltaic panels or solar collectors with heat storage mediums [2–5]. Some researchers have also studied the optimization of MD systems for specific applications [6]. An example of an optimization study is the evaluation of multi-effect MD systems where the results showed a significant increase in the process energy efficiency [7]. Researchers have also studied issues related to scaling and fouling. The deposition of calcium carbonate on the surface of MD membranes has been investigated during the demineralization of lake water using a MD bench scale system. Results showed that the morphology of the formed deposit has a significant influence on the permeate flux decline and a higher flux was obtained with less compact deposit layers [8]. Other researchers have studied the mitigation of calcite and gypsum by the application of antiscalants. Results showed that a certain type of antiscalants extends the induction period for the nucleation of gypsum and calcite without affecting membrane performance [9]. In another study, the scaling kinetics of calcium carbonate were investigated. Induction time measurements were carried out using dynamic light scattering to identify the shifting between homogeneous and heterogeneous nucleation mechanisms as a function of supersaturation. Membrane cleaning was also investigated using a two-step process consisting of citric acid and sodium hydroxide [10]. Like any emerging process, MD has potential challenges associated with it: limited experience on scale-up, process design and pretreatment issues on seawater/brine sources. In order to advance the technical knowledge in those areas, a consortium consisting of the ConocoPhillips Global Water Sustainability Center (GWSC), Qatar University (QU), and Qatar Electricity & Water Company (QEWC) was formed to evaluate the feasibility of testing the MD process at a pilot scale for desalination of brines from local thermal plants. The primary objective of this research paper is to evaluate the performance of MD technologies on seawater and brines from thermal desalination plants. Two leading MD technologies were pilot tested to treat brines from thermal plants in Qatar. The present MD pilot study is unique as it addresses the desalination of concentrated thermal brine by evaluating two leading MD technologies side-by-side under continuous (24/7) operation at field conditions. The effect of the chemicals, from the MSF process, on the membranes was also
Each MD system was provided with pumps, immersion heaters, water chillers, tanks and controls in a 20 ISO shipping container. Both pilot units were equipped with software that allowed remote access and control. The specifications of the pilot units are summarized in Table 1. 2.2. Modes of operation Pilot unit A operates in a continuous flow, one pass mode. After the feed water enters the system, it is heated and then allowed to flow across the different effects in series. In the last effect, it is rejected as a concentrated stream into an external tank. Distillate is produced at each of the system's four effects and is collected in a separate external tank (Fig. 2). The latent energy released during condensation in each effect is transferred to the incoming feed of the subsequent effect. Cooling water is only applied to the final effect to remove the heat from the last effect. Pilot unit B operates in feed and bleed mode, i.e., the feed water enters the module and the concentrated feed is recirculated back into the feed tank while the distillate is collected separately in an external tank. Therefore, the feed water gets concentrated in the feed tank and once the concentrated water reaches a predetermined conductivity value, part of the tank is drained and fresh feed is added to maintain a constant conductivity/salt concentration (Fig. 3). 2.3. Chemical analysis The chemical compositions of the thermal brine, seawater and distillate from each of the units were analyzed using the analytical methods as listed in Table 2.
Fig. 2. Pilot unit A — schematic.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108
103
Table 2 Analytical methods. Parameter
Method
Instrument
Cations (Sodium, Magnesium, Calcium, Potassium) Anions (Chloride, Sulfate, Bromide)
Separation by IonPac® CS12A — 4 × 250 mm column with MSA eluent; ions measured by conductivity detector Separation by IonPac® AS19 — 4 × 250 mm column with EGC III KOH eluent; ions measured by conductivity detector Combustion method
Ion chromatography (ICS 3000, Dionex)
Total Organic Carbon (TOC) Strontium, Boron Conductivity pH
Fig. 3. Pilot unit B — schematic.
2.4. Operating conditions and testing program The operating conditions used during the experiments are summarized in Table 3. Those parameters were recommended by the manufacturer and used throughout the experiments (unless specified otherwise). 2.5. Preliminary testing Preliminary testing was performed at a temporary site at QU with the objective of ensuring that the pilot units met the design specifications before relocating them to the thermal desalination plant. These tests included baseline performance and the effect of temperature and salinity on flux. 2.5.1. Baseline performance test Baseline tests were conducted with tap water as feed using the operating conditions recommended by the manufacturer (Table 3); results are summarized in Table 4. Results showed that pilot unit A achieved distillate flux of 5 LMH with a temperature difference (ΔT) across the membrane of 25.6 °C (6.4 °C per effect; first effect temperature of 59.5 °C). Pilot unit B produced a flux 20% higher than unit A (6.1 vs. 5.0 LMH). Although the flux in pilot unit B was higher than that of A, it was achieved with a higher ΔT across the membranes (35 vs. 25.6 °C) and a higher temperature in the first effect (69 vs. 59.5 °C) 2.5.2. Effect of temperature Temperature is a critical operating parameter since it dictates the driving force across the membrane and therefore the distillate flux. Different operating temperatures were tested to determine the flux vs.
Table 1 MD pilot units' specification. Parameter
Pilot Unit A
Pilot Unit B
No. of modules & arrangement No. of effects per module Total membrane area (m2) Membrane characteristics
1 4 6.4 Pore size: 0.2 μm, Active layer: PTFE, Backing: PP Temperature, pressure, flow & conductivity Immersion heaters External water chiller 70–100 (vacuum)
2 in parallel 1 4.6
Operating parameters logged in PLC Feed water heating element Cooling water element Operating pressure (mbar)
Temperature, pressure, flow, conductivity & pH
1000 (ambient)
Inductively Coupled Plasma, Axial mode Immersing conductivity probe in the sample Immersing pH probe in the sample
Shimadzu TOC Analyzer TOC-V Thermo ICAP 6500 Orion 3 Star conductivity meter Orion 3 Star pH meter
vapor pressure (ΔVp) relationships for the two systems. Results confirmed that there is a linear correlation between the ΔVp across the membrane and the flux (Fig. 4). It was also observed that a 10 °C increase in feed temperature (from 60 °C to 70 °C) increased the flux by 44% while a 10 °C decrease in cooling water temperature (from 30 °C to 20 °C) only resulted in a 17% increase in flux. This is attributed to the exponential dependency of the vapor pressure as function of temperature [1]. Even at low ΔT, both systems produced considerable amounts of distillate; at a ΔT of only 12 °C, pilot unit B flux was 1.1 LMH, equivalent to a production rate of 120 L/d.
2.5.3. Effect of salt concentration The effect of salt concentration was evaluated using two different NaCl solutions (35, 70 g/L). As shown in Fig. 5, the flux obtained with 70 g/L NaCl is 20% lower than the flux obtained during the 35 g/L experiment, which is consistent with the results from bench-scale testing published earlier by the authors [11]. Pilot unit B showed similar behavior. A salt rejection of N 99.9% was achieved by both pilot units, thereby meeting the specifications provided by the manufacturer.
3. Results & discussion 3.1. Tests at desalination facility After the completion of the preliminary tests, the two pilot units were relocated to a local thermal desalination facility and tested, sideby-side. Both pilot units were operated with brine (~70 g/l TDS) from the thermal desalination plant continuously for 24 h a day, 7 days a week. The thermal brine was pretreated using a two-step filtration system (1 μm + Granular Activated Carbon) to remove particulates and organic contaminants present in the feed stream that could potentially affect the integrity of the membranes. The feed TDS remained constant in the presence/absence of the GAC filter. This conservative pretreatment gave both systems the best chance of success. Table 3 Standard operating conditions for each pilot unit. Parameter
Units
Pilot Unit A
Pilot Unit B
Temperature feed inlet (TF) Temperature hot side (THS) Temperature cold side (TCS) Feed flow rate (QF) Hot side flow rate (QHS) Cold side flow rate (QCS) Pressure (P)
°C °C °C L/min L/min L/min mbar
25–35 70 20 1.4 16.5 13.5 70
25–35 70 20 30 NA 40 1000
104
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 7
Table 4 Baseline performance for each pilot unit. Pilot unit A
Pilot unit B
Feed Flux Distillate conductivity range Distillate produced First effect inlet temperature ΔT Recovery based on volume
– LMH μS/cm L/day °C °C %
Tap water 5.0 ± 0.4 0.2–1.0 757 ± 77 59.5 ± 1.0 25.6 ± 1.6 39 ± 3.6
6.1 ± 0.3 1–10 670 ± 40 69 ± 1.0 35 ± 0.7 NA
6 5 4 3 2
As shown in Fig. 6, pilot unit A was able to consistently generate high quality product water (TDS b 10 mg/L) over a 12-day period, while maintaining a stable flux of 4.5 LMH. Due to the high salt concentrations, a 10% reduction in flux was observed when compared to the tap water baseline tests (5.0 LMH). The ΔT and vacuum pressure in Fig. 6 are 19.6 °C and 70 mBar, respectively. During the first 3 days of operation, the distillate conductivity was averaged at 2.5 μS/cm, after which a steady increase of conductivity was observed until it stabilized at ≈ 10 μS/cm. Recovery rate was 34% and the concentrated MD brine had a TDS of 112 g/L. Table 5 shows the water quality analyses of major anions and cations. Salt rejections greater than 99.9% were achieved. Results indicate that the pretreatment was effective in removing potential contaminants that could compromise the integrity of the membranes (see pretreatment Section 3.2). Using thermal brine as the feed water, pilot unit B generated a poorer product water quality (2530 μS/cm). As shown in Fig. 7, a significant decrease in flux (from 4 to 2.5 LMH) was observed throughout the testing period. Therefore, the system was stopped after 5 days of operation. Table 5 shows the water quality data; approximately 99% salt rejection was achieved (feed tank concentration was used as reference). The decrease in the flux could be attributed to the feed and bleed mode of operation; as the thermal brine in the feed tank concentrates, the water flux decreases. The poor salt rejection could possibly be due to a small mechanical leakage in the membrane modules, but that was not confirmed by the authors. The inspection of pilot unit B membrane modules was very challenging due to the heavy weight and design of the modules, which required the presence of the technology provider staff to make the autopsy. It was also learned during the study that the provider is developing a newer membrane module design with lighter weight, which makes them more attractive for full-scale installations. Thus, the project team decided to end the field testing program with pilot unit B, hoping for future testing opportunities with the new & lighter design membrane modules. After the tests were completed, both systems were flushed with tap water. It was observed that both systems regained their original flux, confirming that no severe membrane fouling occurred during the test
0 0
3
Pilot Unit A
2
Pilot Unit B
2
3
4
5
6
7
8
Fig. 5. Pilot unit A: effect of salt concentration on water flux. TF: 25 °C, THS: 70 °C, TCS: 30 °C, QF: 1.4 L/min, P: 70 mbar, ΔT35,000: 22.1 °C, ΔT70,000: 23.6 °C.
with thermal brine. For pilot unit B, the conductivity of the tap water used to flush the membranes was measured and no increase was observed after flushing; indicating that the poor rejection may not be due to scaling of the membranes. During the tap water cleaning, the distillate conductivity was also high; the rejection of the system B was approximately 70%. 3.2. Pretreatment evaluation To assess the impact of pretreatment on the performance of the system, tests were conducted by removing the granular activated carbon (GAC) filter from the upstream of the MD process. During the first 4 days of operation, pilot unit A was able to generate a distilled quality of conductivity b 1 μS/cm. The water quality started to deteriorate on the 5th day with a decrease in distillate flux; the distillate conductivity increased exponentially to ≈700 μS/cm as shown in Fig. 8. The increase in distillate conductivity and the slight decrease in flux were attributed to membrane wetting caused by residual chemicals present in the brine that were not removed by the cartridge filters. This test confirmed the importance of the activated carbon filter in removing process chemicals that could potentially compromise the hydrophobicity of the membranes. A laboratory investigation was carried out to study the different types of process chemicals (antiscalant/antifoam) that may affect the hydrophobicity of the membranes in pilot unit A. The membranes were tested using a bench-scale MD unit. The state-of-the-art unit used for this study is described in reference [11]. The first test was conducted by spiking a 70 g/L NaCl solution with antiscalant. As shown in
Distillate Flux, LMH
Distillate Flux (LMH)
4
1
Time, hours
7
5
70 g/L NaCl
1
8
6
35 g/L NaCl
7
1000000
6
100000
5
10000
4
1000
Flux Brine conductivity Feed conductivity
3
100
Distillate conductivity
2
10
1
1
Conductivity, µS/cm
Units
Distillate Flux, LMH
Parameter
1 0
0 0
0 0
50
100
150
200
250
1
2
3
4
5
6
7
8
9
10
11
12
Time, days
Vapor pressure difference (mbar) Fig. 4. Pilot units A & B: vapor pressure difference vs. water flux.
Fig. 6. Pilot unit A: performance on thermal brine. TF: 32 °C, THS: 70 °C, TCS: 20 °C, QF: 1.4 L/min, P: 70 mbar, ΔT: 19.6 °C.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108
105
Table 5 Pilot units A & B: water quality on thermal brine. Pilot unit A
Pilot unit B
Parameter
Units thermal Brine (n = 4)
Distillate (n = 4)
Reject brine (n = 4)
Rejection (%)
Thermal brine (n = 1)
Distillate (n = 1)
Reject brine (n = 1)
Rejection (%)
TDS Chloride Sodium Sulfate Magnesium Potassium Calcium Bromide Strontium Boron TOC
mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
6 2 2 1.2 0.3 b0.2 b1 b0.1 b0.2 b0.2 b1
112,279 64,838 33,258 7264 3929 1193 1423 110 17 12 4
99.99 99.99 99.99 99.97 N99.99 N99.97 N99.87 N99.8 N98.2 N98 –
68,529 39,064 20,375 4996 2435 770 703 55 11 10 3
1472 840 496 69 49 19 17 1 b0.5 b0.5 b1
136,191 78,281 38,926 10,885 4565 1394 1328 156 21 16 5
98.92 98.93 98.73 99.37 98.93 98.62 98.71 99.04 N97.6 N96.8 –
71,031 40,988 21,310 4344 2519 779 775 64 11 8 3
Fig. 9, the antiscalant concentration was increased periodically from 2 mg/L to 800 mg/L and the distillate conductivity was monitored. There was no significant change in the distillate conductivity. This observation rules out the possibility of the antiscalant causing membrane wetting. A similar test was conducted by spiking the salt solution with the antifoam agent. The antifoam agent concentration was increased from 0.2 mg/L to 50 mg/L. The distillate conductivity sharply increased at an antifoam agent concentration of 50 mg/L as shown in Fig. 9. This confirms that the antifoam agent is the chemical responsible for the membrane wetting. In a typical MSF process, the antifoam agent is dosed at a concentration between 0.1 and 0.3 mg/L. When the pilot unit received 10,000 L of thermal brine after five days of operation, the membrane was exposed to at least 1000 mg of antifoam agent; this was sufficient to wet the membrane causing the deterioration of the distillate water quality discussed earlier. To further confirm that the antifoam agent was the chemical responsible for membrane wetting, membrane coupons were soaked in six individual beakers for five days. The beakers contained the following solutions: nanopure water, concentrated brine, 70 g/L NaCl, 70 g/L NaCl with 100 mg/L antiscalant, 70 g/L NaCl with 100 mg/L antifoam and 70 g/L NaCl with both antiscalant (100 mg/L) and antifoam (10 mg/L). The membranes were dried and a couple of nanopure water droplets were added on top of the membrane as shown in Fig. 10. Wetting was only observed in the coupons that were in contact with the antifoam agent; supporting the conclusion that this chemical was the main cause of membrane wetting. The removal of process chemicals by activated carbon was also investigated. 100 mL of a 1000 mg/L antifoam and antiscalant solutions
3.3. Performance on seawater The test with seawater was performed using pilot unit A with 1 μm filtration as pretreatment. The chlorinated seawater was collected at the intake, prior to the addition of antiscalant and antifoam agents. Results showed that the system was able to operate steadily for 8 days with distillate flux of 4.8 LMH (Fig. 11) and recovery of 30%. Since antiscalant was not added, the recovery was limited to 35% to avoid salt precipitation in the membranes and it was controlled by slightly increasing the feed flow rate. The TDS of the seawater was increased from 44 g/L to 66 g/L. As shown in Table 6, excellent salt rejection of greater than 99.9% was achieved for the major ions. These results demonstrated the capability of MD to desalinate seawater, generating excellent quality distillate. 3.4. Pilot unit A process optimization A set of experiments was conducted to determine the main parameters that influence the water production in pilot unit A. The parameters were optimized for maximum recovery, water production and energy efficiency. All tests were performed using thermal brine as a feed and a two-step pretreatment consisting of 1 μm + GAC filters.
Feed & Bleed mode 7
1000000
6
100000
6
100000
5
10000
5
10000
4
1000
4
1000
3
100
3
100
2
10
2
1
Brine conductivity
1
Feed conductivity
1
Distillate conductivity
0
0 1
2
3
4
5
1
Feed conductivity
Distillate conductivity
0
10
Flux
Flux
Conductivity, µ S/cm
1000000
Distillate Flux, LMH
7
Conductivity, µS/cm
Distillate Flux, LMH
Concentraon stage
were treated individually with 500 mg of GAC. After adding the GAC to the individual beakers, the solutions were stirred for 30 min and filtered using 0.22 μm Nylon filter. Total Organic Carbon (TOC) results showed that GAC removes 21% and 23% of the antiscalant and antifoam agent, respectively. Higher removal could be achieved by GAC packed column. The portion of the antifoam agent that was removed in this experiment may have been responsible for membrane wetting.
6
Time, days Fig. 7. Pilot unit B: performance on thermal brine. TF: 30 °C, THS: 70 °C, TCS: 20 °C, ΔT: 35 °C.
0
0 0
1
2
3
4
5
6
Time, days Fig. 8. Pilot unit A: Performance on thermal brine without activated carbon filter. TF: 32 °C, THS: 70 °C, TCS: 20 °C, QF:1.4 L/min, P: 70 mbar, ΔT: 21.4 °C.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 800
Antiscalant (AS)
600
Distillate Flux, LMH
Conductivity, µS/cm
700
Antifoaming (AF)
500 400 300 AS - 2 ppm AF - 0.2 ppm
200
AS - 10 ppm AF - 1 ppm
AS - 100 ppm AF - 10 ppm
AS - 800 ppm AF - 50 ppm
100
7
1000000
6
100000
5
10000 Flux
4
1000
Brine conductivity
Feed conductivity Distillate conductivity
3
100
2
10
1
1
0
Conductivity, µS/cm
106
0 0
1
2
3
4
5
6
7
8
9
Time, days
0 0
50
100
150
200
250
300
350
400
450
500
Time, min
Fig. 11. Pilot unit A: Performance on seawater. TF: 32 °C, THS: 70 °C, TCS: 20 °C, QF: 1.7 L/min, P: 70 mbar, ΔT: 21.9 °C.
Fig. 9. Impact of scale inhibitor and antifoam agent on pilot unit A membrane.
3.4.1. Effect of thermal brine temperature The objective of this test was to evaluate the effect of thermal brine temperature (prior to entering the system) on the membrane flux. The brine was preheated using an external tank fitted with immersed electrical heaters. This preheating permitted the system to be operated at ΔTs higher than the ΔTs used in earlier test (standard conditions). As expected, operation with preheated thermal brine resulted in an increase in distillate production. Experimental results are summarized in Fig. 12a and Table 7.
3.4.2. Effect of operating vacuum The effect of vacuum pressure on the system was also investigated. Three different pressures were tested: 70, 100 and 130 mbar. Lowering the vacuum resulted in higher flux as shown in Fig. 12b.
3.4.3. Feed water flow rate The objective of this test was to evaluate the effect of the feed flow rate on the % recovery. The target was to achieve a recovery of 50%. Prior to conducting the test, preliminary lab experiments were performed to confirm that operating the system at 50% recovery would not cause the salts to precipitate. Results indicated that the system could be operated at a maximum recovery of 70%, after which salts would start to precipitate. Such high recoveries can be achieved, without precipitation,
due to the antiscalant that is added to the thermal brine as part of the MSF process. Three feed flow rates (1.1 L/min, 1.2 L/min & 1.5 L/min) were investigated; flux remained constant at 5 LMH for the three flow rates tested. While operating the pilot unit at 1.1 L/min, the system achieved a recovery of 48% as shown in Fig. 12c. 3.5. Test at optimum conditions A final test was conducted using the optimized conditions summarized in Table 8. The feed temperature prior to entering the system was maintained at 50 °C since this is the temperature at which the thermal brine is typically rejected during the summer. The internal electric heater of the system was used to increase the temperature from 50 °C to 75 °C. Results showed that at optimum conditions, distillate production could be increased by 40% in comparison with standard conditions (690 L/day vs. 950 L/day). 52% recovery and distillate water with an average conductivity of ≈ 15 μS/cm were achieved at a stable flux of 6.2 LMH (Fig. 13). 4. Summary & conclusions In Qatar, the production of fresh water is primarily achieved using thermal desalination plants that are co-located with power plants; installations that provide both hot concentrated brines and low grade
Nanopure Water
Thermal Brine
70,000 mg/L NaCl
70,000 mg/L NaCl+ 100 mg/L Anscalant
70,000 mg/L NaCl + 100 mg/L Anfoam
70,000 mg/L NaCl+ 100 mg/L A + nscalant 10 mg/L Anfoam
Fig. 10. Image of membrane coupons from pilot unit A after soaking in different solutions.
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 Table 6 Pilot unit A; water quality on seawater. Parameter
Units
TDS Chloride Sodium Sulfate Magnesium Potassium Calcium Bromide Strontium Boron TOC
mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
Table 7 Summary of results from optimization experiment.
Seawater (N = 1)
Distillate (N = 1)
Reject brine (N = 1)
Rejection (%)
44,701 25,531 13,421 3263 1336 511 454 36 7 6 1
9 4 4 0.3 0.4 0.3 b1 b0.1 b0.2 b0.2 b1
66,304 38,180 19,232 4927 2263 831 650 55 10 9 2
99.98 99.98 99.97 99.99 99.97 99.93 N99.7 N99.7 N97.2 N96.6 –
(a) 7 Disitllate Flux, LMH
6
Flux 5.7 Flux 5.0
5 4
Parameter
Units
Feed brine temperature Flux First stage temperature ΔT Operating vacuum Feed water flow rate Recovery Operating vacuum Feed brine temperature Flux First stage temperature ΔT Feed water flow rate Recovery Feed water flow rate Recovery Feed brine temperature Flux First stage temperature ΔT Operating vacuum
°C LMH °C °C mbar L/min % mbar °C LMH °C °C L/min % L/min % °C LMH °C °C mbar
Results 55 5.7 68.1 20.4 100 1.4 42 70 32 4.9 61.4 19.6 1.4 35 1.1 48 32 5.0 62 21.2 100
45 5.0 66.5 18.4 100 1.4 36 100 32 4.0 64.2 16.8 1.4 29 1.2 45 32 5.0 62 20.8 100
25 3.8 63.9 16.4 100 1.4 27 130 32 3.7 67.2 15.2 1.4 27 1.5 35 32 4.9 61.4 19.8 100
Flux 3.8
3 2 1 0 55
45
25
Thermal BrineInlet Temperature,°C
(b) 7 6
Disllate Flux, LMH
107
Flux: 4.9
5
Flux: 4.0
4
Flux: 3.7
3 2 1 0 70
100
130
Vacuum Pressure,mBar
(c)100%
waste heat as by-products. Membrane distillation (MD) is a hybrid thermal-membrane process that can use low grade waste heat and a hydrophobic membrane to produce a high quality distillate. The MD process can treat highly saline brines that the conventional membrane desalination processes are unable to treat efficiently. A consortium consisting of the ConocoPhillips Global Water Sustainability Center (GWSC), Qatar University (QU), and Qatar Electricity & Water Company (QEWC) was formed to evaluate the application of MD in desalinating brines from full-scale thermal desalination plants for sustainable augmentation of fresh water supplies in Qatar. The main objective was to investigate the performance of the MD technology to desalinate seawater and concentrated brine. Two leading pilot scale MD technologies, vacuum multi-effect (A) and air gap MD (B), were operated in parallel to treat seawater and thermal brine discharged from a local thermal desalination facility under field conditions. Below is a summary of the study's main conclusions: - Pilot unit A achieved distillate flux of 5 LMH on tap water with an average temperature difference (ΔT) across the membrane of 25.6 °C and a feed temperature in the first effect of 59.5 °C. - Pilot unit B produced a flux of 6.1 LMH on tap water with a ΔT across the membranes of 35 °C and a feed temperature in the first effect of 69 °C. - Results from both pilot units using tap water showed that the water flux is directly proportional to the applied vapor pressure difference across the MD membranes.
90% 80%
Table 8 Operating parameters and performance at standard and optimum conditions.
Recovery, %
70% 60%
40%
Parameter
Units
Standard conditions
Optimum conditions
Feed temperature (TF) Temperature hot side (THS) Temperature cold side (TCS) Feed flow rate (QF) Hot side flow rate (QHS) Cold side flow rate (QCS) Pressure (P) First effect inlet temperature ΔT Flux Recovery Production
°C °C °C L/min L/min L/min mbar °C °C LMH % L/day
32 70 20 1.4 16.5 13.5 70 61 19.6 4.5 34 690
50 75 20 1.25 16.5 13.5 55 65.2 26 6.2 52 950
Recovery 48%
50%
Recovery 45% Recovery 35%
30% 20% 10% 0% 1.5
1.2
1.1
Feed flow rate,L/min Fig. 12. Optimization results: (a) effect of feed temperature, (b) effect of vacuum pressure, (c) effect of feed flow rate. Table 7 shows a summary of the main operating parameters.
108
J. Minier-Matar et al. / Desalination 351 (2014) 101–108 7
Acknowledgment
6
Distillate Flux, LMH
Standard (non-optimized)
5
Optimized
4 3 2
Opmized condions Feed Temperature : 50 C Feed Flow rate : 1.25 L/min Vacuum set point : 55 mbar Recovery : 52 %
1
Standard (non-opmized) condions Feed Temperature : 32 C Feed Flow rate : 1.4 L/min Vacuum set point : 76 mbar Recovery : 34 %
0 0
0.5
1
1.5
2
2.5
3
Time, days Fig. 13. Pilot unit A performance: optimized vs. standard conditions. Operating conditions are summarized in Table 8.
- The MD flux obtained while operating with 70 g/L synthetic NaCl solution was 20% lower than the flux obtained during operation with 35 g/L NaCl solution. - Pilot unit A was able to desalinate brines from thermal desalination plants and consistently produce a high quality distillate (conductivity b 10 μS/cm — N99.9% rejection). - Pilot unit A also showed stable flux (4.8 LMH) and produced high quality distillate when operated directly on seawater. - Under optimized conditions, pilot unit A operating on thermal desalination brines achieved a stable flux of 6.2 LMH and a recovery of 52%. - Pilot unit B showed stable performance while operating on thermal desalination brines with significant salt rejection of 98.9%. However, due to challenges experienced with system and the heavy weight of the membrane modules, it was not possible to complete the field testing program with that unit. - Pretreatment is critical due to the possible presence of process chemicals in the brine of the MSF process that could potentially cause MD membrane wetting. - It was found that the antifoam agent resulted in the wetting of MD membranes. Granular activated carbon proved to be an effective pretreatment. - Overall, MD was shown to be a feasible technology to produce potable quality water from the brines discharged from thermal desalination plants.
The research team would like to thank Mr. Fahad Al Muhannadi, General Manager of QEWC, and Mr. Abdulsattar Al Rasheed, CEO of Ras Abu Fontas Desalination Plant, for their support in the overall testing program. The research team thanks the efforts of Dr. Majeda Khraisheh from Qatar University, and research assistants Ahmad Fard and Yehia Manawi for facilitating the operations of the pilot unit at Qatar University. The team acknowledges the MD technology providers for technical support during the operation of the pilot units. The research team would also like to acknowledge their colleagues at the GWSC including, Mr. Raul Dores, Dr. Samir Gharfeh, Dr. Isik Turkmen, Dr. Nabin Upadhyay, Ms. Eman Al Shamari and Ms. Mary Katebah for conducting bench scale experiments and water analyses on the seawater and brine samples.
References [1] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensive review, Desalination 287 (2012) 2–18. [2] R. Schwantes, A. Cipollina, F. Gross, J. Koschikowski, D. Pfeifle, M. Rolletschek, V. Subiela, Membrane distillation: solar and waste heat driven demonstration plants for desalination, Desalination 323 (2013) 93–106. [3] F. Banat, N. Jwaied, M. Rommel, J. Koschikowski, M. Wieghaus, Desalination by a “compact SMADES” autonomous solar powered membrane distillation unit, Desalination 217 (2007) 29–37. [4] A. Kullab, A. Martin, Membrane distillation and applications for water purification in thermal cogeneration plants, Sep. Purif. Technol. 76 (2011) 231–237. [5] E.G. Burrieza, G. Zaragoza, S.M. Cuevas, J. Blanco, Experimental evaluation of two pilot-scale membrane distillation module used for solar desalination, J. Membr. Sci. 409–410 (2012) 264–275. [6] A.E. Jansen, J.W. Assink, J.H. Hanemaaijer, J. van Medevoort, E. Van Sonsbeek, Development and pilot testing of full-scale membrane distillation modules for deployment of waste heat, Desalination 323 (2015) 55–65. [7] K. Zhao, W. Heinzl, M. Wenzel, S. Buttner, F. Bollen, G. Lange, S. Heinzl, N. Sarda, Experimental study of the memsys vacuum-multi-effect-membrane-distillation (V-MEMD) module, Desalination 323 (2013) 150–160. [8] Marek Gryta, Alkaline scaling in the membrane distillation process, Desalination 228 (2008) 128–134. [9] F. He, K. Sirkar, J. Gilron, Effect of antiscalants to mitigate membrane scaling by direct contact membrane distillation, J. Membr. Sci. 345 (2009) 53–58. [10] E. Curcio, X. Ji, G.D. Profio, A.O. Sulaiman, E. Fontananova, E. Drioli, Membrane distillation operated at high seawater concentration factors: Role of the membrane on CaCO3 scaling in presence of humic acid, J. Membr. Sci. 346 (2010) 263–269. [11] Samer Adham, Altaf Hussain, Joel Minier Matar, Raul Dores, Arnold Janson, Application of membrane distillation for desalting brines from thermal desalination plants, Desalination 314 (2013) 101–108.