QATAR UNIVERSITY. College of Engineering MEMBRANE DISTILLATION DESALINATION: WATER QUALITY AND ENERGY EFFICIENCY ANALYSIS.

Transcription

1 QATAR UNIVERSITY College of Engineering MEMBRANE DISTILLATION DESALINATION: WATER QUALITY AND ENERGY EFFICIENCY ANALYSIS A Thesis in Environmental Engineering by Ahmad M. Kayvani Fard Copyright 2013 Ahmad Kayvani Fard Submitted in partial fulfillment of the requirements for the degree of Master of Science August 2013

2 Copyrights 2013 Ahmad Kayvani Fard All rights reserved. No part of this material may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, except in the form of brief excerpts or quotations for review purpose, the written permission from the author. ii

3 Committee Page The thesis of Ahmad Kayvani Fard was reviewed and approved by the following: We, the committee members listed below accept and approve the Thesis/Dissertation of the student named above. To the best of this committee s knowledge, the Thesis/Dissertation conforms the requirements of Qatar University, and we endorse this Thesis/Dissertation for examination. Name Signature Date Name Signature Date Name Signature Date iii

4 Executive Summary More than 70% of the global population finds its fresh water from aquifers, rivers, and lakes. Global population estimated to approach 9 billion by 2050 and the standard of living of fast developing countries, such as Qatar, increases and the demand for fresh water is increasing dramatically. Qatar is located in an arid region where there is no source of surface fresh water give the very low precipitation per year. Qatar s primary source of fresh water is through seawater desalination. Amongst the major processes that are commercially available on the market, the most common large scale techniques are Multi-Stage Flash distillation (MSF), Multi Effect distillation (MED), and Reverse Osmosis (RO). Although commonly used, these three processes are highly expensive due to high energy input requirements and high operating costs associated with maintenance and stress induced on the systems in harsh alkaline media. Beside that cost, environmental food print of these desalination techniques are significant; from damaging marine eco-system, to huge land use, to release of tons of GHG and huge carbon footprint. Other less energy consuming techniques based on membrane separation are being sought to reduce both the carbon footprint and operating costs. Emerged in 1960s, membrane distillation is an alternative technology for water desalination attracting more attention since 1980s. MD process involves the evaporation of a hot feed, typically below boiling point of brine at standard conditions, by creating a water vapor pressure difference across the thickness of a porous, hydrophobic membrane. There are four configurations of membrane distillation including air gap membrane distillation (AGMD), direct contact membrane distillation (DCMD), sweep gas membrane distillation (SGMD), and vacuum membrane distillation (VMD). Main advantages of MD over commercially available technologies and specially RO are reduction of membrane and module stress due to absence of trans-membrane pressure, less impact of contaminant fouling on distillate due to transfer of only water vapor, utilization of low grade or waste heat from oil and gas industries to heat up the feed up to required temperature difference across the membrane, superior water quality, and relatively lower capital and operating cost. To achieve the objective of this study, state of the art flat-sheet cross-flow DCMD bench scale unit was designed, commissioned, and tested. The objective of this study is to analyze the iv

5 characteristics and morphology of a the membrane suitable for DCMD through SEM imaging and contact angle measurement, study the water quality of distillate produced by DCMD bench scale unit, and conduct an energy efficiency analysis of DCMD with varied process parameters. Comparison with available literature data is undertaken where appropriate and laboratory data is used to compare a DCMD distillate quality with that of other desalination techniques and standards. Membrane analysis showed that the PTFE membrane used has contact angle of 127º with highly porous surface supported with less porous and bigger pore size PP membrane. High contact angle of PTFE membrane shows the high selectivity of membrane to reject water and allow vapor to flow through membrane pores (hydrophobic). Study on the effect of feed solution (salinity) and temperature on water quality of distillate produced from ICP analysis showed that with any salinity and different feed temperature (up to 70 ºC) the electric conductivity of distillate is less than 5 μs/cm with 99.99% salt rejection and proved to be feasible and effective process capable of consistently producing high quality distillate from very high feed salinity solution (i.e mg/l TDS) even with substantial quality difference compared to other desalination methods such as RO and MSF. In terms of energy efficiency, feed flow rate is an important factor and doubling feed flow rate can increase the energy efficiency by about 2 folds. Similarly, increasing feed temperature, increases energy efficiency significantly due to an increase in driving force and hence flux. Effect of concentration and permeate temperature also has been studied and showed increasing these factor cause lead to a decline in energy efficiency. MD has not yet been commercialized and to be practically implemented and feasible, MD should use free and cheap waste energy such as industrial waste heat from flue gas or other sources. Other factors in improving feasibility of MD might be optimization of process conditions, preparation of novel membranes, module configurations as well as spacers support. v

6 Declaration Student Declaration To the best of my knowledge, the thesis contains no material previously published or written by another person or institution, except where due reference is made in the text of the thesis. The thesis contains no material which has been accepted for the award of any other degree in any university or other institution. Name Signature Date Supervisors Declaration To the best of my/our knowledge, the thesis conforms the requirements of Qatar University, and we/i endorse this thesis for examination. Name Signature Date Name Signature Date vi

7 Table of Content Committee Page... iii Executive Summary... iv Declaration... vi Table of Content... vii List of Table... x List of Figures... xi Acknowledgments... xiv Chapter 1: Introduction Objective and Scope of Work... 9 Chapter 2: Background and Literature Survey Thermal Desalination Multiple Effect Distillation (MED) Multi-Stage Flash Desalination (MSF) Membrane Desalination Reverse Osmosis (RO) Membrane Distillation (MD) Desalination Technologies: Advantages and Disadvantages Water Quality Process Effluent Brine Thermal efficiency and Energy Consumption Environmental Footprint of Desalination Process Green House Gas (GHG) Emissions Membrane Distillation Process: A Promising Solution State of the Art in Membrane Distillation Chapter 3: Direct Contact Membrane Distillation Theory Membrane Module and Configuration Heat and Mass Transfer in DCMD Heat Transfer vii

8 Mass transfer Concentration Polarization Temperature Polarization Liquid Entry Pressure Flow Turbulence and Flow Distribution Pressure Drop Membrane Physical Properties Membrane porosity Membrane Pore Size Membrane Thickness Pore Size Distribution Pore Tortuosity Thermal Conductivity Feed Temperature Water Quality Status of Quality of Distillate Produced Using MD Energy Efficiency Chapter 4: Approach and Methodology DCMD Bench Scale Unit Feed Solution Membrane Sheet DCMD Module Configuration (MD Block) Material of Construction Dimensions Auxiliary Equipment Pump Heaters and coolers Temperature and Pressure Measurement Flow Meter Conductivity Meter Digital Display viii

9 Weighing Balance Data Acquisition System Tubing Experimental Procedure Water Quality and Energy Efficiency Experiments Contact Angle measurement SEM Imaging Chapter 5: Results and Discussion Contact Angle Measurement SEM Characterization Reproducibility of DCMD Bench Scale Unit Water Quality Analysis Energy Efficiency Analysis Effect of Feed Flow Rate Effect of Feed Temperature Effect of Permeate Temperature Effect of Feed Concentration and Feature Chapter 6: Conclusion and Summary Chapter 7: Recommendations and Future Works References Appendix Appendix (A): Flux Profile Appendix (B): Operational Parameters of Different Experiments Nomenclature ix

10 List of Table Table 1: GCC population and water resource... 4 Table 2: GCC desalination capacity and its status regarding world capacity... 6 Table 3: existing and future planed desalination plants in GCC countries... 7 Table 4: Typical range of concentration versus desalination recovery Table 5: type of effluents from different desalination process Table 6: Thermal and electrical consumption of different desalination process Table 7: minimum energy requirement versus feed salinity Table 8: relevant airborne emissions produced by desalination technologies Table 9: contact angle and surface energy of some commonly used membranes Table 10: Thermal conductivity of different material Table 11: Salinity of different type of water Table 12: Salt content of sea water and brackish water Table 13: Average salt concentrations of different world water sources Table 14: Quality of drinking water in State of Qatar Table 15: ph and EC of feed solutions used in the experiments Table 16: Properties of PTFE membrane Table 17: Parameters used for experiments of DCMD system Table 18: number of experiments conducted with different feed solution and temperatures Table 19: Flow rate ranges for DCMD tests Table 20: Contact angle measurements for PTFE membrane Table 21: Contact angle measurements for PP membrane Table 22: Average flux of different feed solution at different temperature using flow rate of 1.5 L/min Table 23: Feed characteristic using ICP analysis Table 24: ICP detection limit Table 25: Distillate quality of DCMD test at ºC Table 26: Distillate quality of DCMD test at ºC Table 27: Water quality comparison between MD distillate and distillate from MSF thermal desalination plant (QEWC) x

11 List of Figures Figure 1: Different type of water around the globe and corresponding total percentage Figure 2: Human water consumption... 2 Figure 3: different types of desalination technologies around the world Figure 4: Trends of desalination capacity in the world and GCC region Figure 5: location and capacity of installed desalination plants in GCC states Figure 6: Schematic of typical MED process Figure 7: schematics diagram of MSF process with recirculation. A Figure 8: Reverse osmosis basic design Figure 9: Principle of the membrane distillation process Figure 10: Different MD configuration (a) DCMD, (b) AGMD, (c) SGMD, (d) VMD Figure 11: current and forecasted water demand in state of Qatar Figure 12: Resistance to heat transfer in MD Figure 13: Spacer properties and parameter Figure 14: Heat transfer across the membrane Figure 15: Parameters of Young s equation Figure 16: Process flow diagram of DCMD bench scale unit Figure 17: System setup in Qatar University s lab Figure 18: schematic of bottom and top plate Figure 19: DCMD bench scale cell and its parts Figure 20: explosion diagram of MD cell with spacer and membrane sheet Figure 21: Dimensions of the bottom plate (top view) Figure 22: Dimensions of the top plate (top view) Figure 23: Dimensions of top and bottom plate looking from side view Figure 24: cross section of feed and permeate inlet and outlet looking from side view Figure 25: Contact angle measuring device Figure 26: equilibrium state of a distilled water droplet on a flat-sheet membrane active layer (PTFE) Figure 27: equilibrium state of a distilled water droplet on a flat-sheet membrane support layer (PP) Figure 28: SEM picture of PTFE membrane and the thermal bonding of active layer to support layer with 8000 time magnification (left) and 2000 time magnification (right) Figure 29: SEM image of support layer of membrane (PP) with 200 (left) and 500 (right) time magnification Figure 30: SEM image of active layer of membrane (PTFE) with (left) and (right) time magnification Figure 31: SEM image of PTFE side of membrane and pore size and pore size distribution of the active layer of membrane xi

12 Figure 32: SEM image of PP side of membrane and pore size and pore size distribution of the support layer of membrane Figure 33: Pore size distribution of PTFE membrane (Active layer) Figure 34: SEM images of membrane thickness Figure 35: Flux profile of feed at different inlet temperature with constant permeate temperature (Tp=30 C, Q=1.5 L/min, thermal brine) Figure 36: Flux profile of feed at different inlet temperature with constant permeate temperature (Tp=20 C, Q=1.5 L/min, thermal brine) Figure 37: Flux profile of constant feed temperature and different permeate temperature (Tf=70 C, Q= 1.5 L/min, thermal brine) Figure 38: Flux profile of constant feed temperature and different permeate temperature (Tf=60 C, Q= 1.5 L/min, thermal brine) Figure 39: Effect of feed flow rate on DCMD energy efficiency (Tf=70C,Tp=30C, thermal brine) Figure 40: Effect of feed flow rate on permeate flux (Tf=70C,Tp=30C, thermal brine) Figure 41: Effect of feed flow rate on permeate flux (T f =70C,T p =30C, thermal brine) Figure 42: The relationship between temperature and water vapor pressure by Antoine equation Figure 43: Effect of feed temperature on permeates flux (Tp=30C, Q=1.5 L/min, Thermal brine) Figure 44: Linear effect of flux with feed temperature (Tp=30C, Q=1.5 L/min, Thermal brine) Figure 45: Effect of feed temperature on energy efficiency (Tp=30C, Q=1.5 L/min, Thermal brine) Figure 46: Effect of permeate temperature on flux (Tf=70C, Q=1.5 L/min, Thermal brine) Figure 47: Permeate flux of different cold side temperature over time (Tf=70C, Q=1.5 L/min, Thermal brine) Figure 48: Effect of permeate inlet temperature on energy efficiency (Tf=70C, Q=1.5 L/min, Thermal brine) Figure 49: Effect of feed concentration on permeates flux (Tf=70C, Tp=30 C, Q=1.5 L/min). 110 Figure 50: Effect of feed concentration on energy efficiency (Tf= 30 C, Tp=30 C, Q=1.5 L/min) Figure 51: Effect of feed concentration and feed inlet temperature on permeate flux (Tp=30 C, Q=1.5 L/min) Figure 52: Effect of feed concentration and feed inlet temperature on energy efficiency (Tp=30 C, Q=1.5 L/min) Figure 53: Effect of feed concentration and permeate inlet temperature on permeate flux (Tf=70 C, Q=1.5 L/min) Figure 54: Effect of feed concentration and permeate inlet temperature on energy efficiency (Tf=70 C, Q=1.5 L/min) xii

13 Figure 55: Flux profile (Tf= 70 C, Tp= 20 C, Q= 1.5 L/min, thermal brine) Figure 56: Flux profile (Tf= 70 C, Tp= 30 C, Q= 1.5 L/min, thermal brine) Figure 57: Flux profile (Tf= 70 C, Tp= 40 C, Q= 1.5 L/min, thermal brine) Figure 58: Flux profile (Tf= 65 C, Tp= 20 C, Q= 1.5 L/min, thermal brine) Figure 59: Flux profile (Tf= 65 C, Tp= 30 C, Q= 1.5 L/min, thermal brine) Figure 60: Flux profile (Tf= 65 C, Tp= 40 C, Q= 1.5 L/min, thermal brine) Figure 61: Flux profile (Tf= 60 C, Tp= 20 C, Q= 1.5 L/min, thermal brine) Figure 62: Flux profile (Tf= 60 C, Tp= 30 C, Q= 1.5 L/min, thermal brine) Figure 63: Flux profile (Tf= 60 C, Tp= 40 C, Q= 1.5 L/min, thermal brine) Figure 64: Flux profile (Tf= 70 C, Tp= 20 C, Q= 1.5 L/min, seawater) Figure 65: Flux profile (Tf= 70 C, Tp= 30 C, Q= 1.5 L/min, seawater) Figure 66: Flux profile (Tf= 70 C, Tp= 40 C, Q= 1.5 L/min, seawater) Figure 67: Flux profile (Tf= 65 C, Tp= 20 C, Q= 1.5 L/min, seawater) Figure 68: Flux profile (Tf= 65 C, Tp= 30 C, Q= 1.5 L/min, seawater) Figure 69: Flux profile (Tf= 65 C, Tp= 40 C, Q= 1.5 L/min, seawater) Figure 70: Flux profile (Tf= 60 C, Tp= 20 C, Q= 1.5 L/min, seawater) Figure 71: Flux profile (Tf= 60 C, Tp= 30 C, Q= 1.5 L/min, seawater) Figure 72: Flux profile (Tf= 60 C, Tp= 40 C, Q= 1.5 L/min, seawater) Figure 73: Flux profile (Tf= 70 C, Tp= 20 C, Q= 1.5 L/min, NaCl solution) Figure 74: Flux profile (Tf= 70 C, Tp= 30 C, Q= 1.5 L/min, NaCl solution) Figure 75: Flux profile (Tf= 70 C, Tp= 40 C, Q= 1.5 L/min, NaCl solution) Figure 76: Flux profile (Tf= 65 C, Tp= 20 C, Q= 1.5 L/min, NaCl solution) Figure 77: Flux profile (Tf= 65 C, Tp= 30 C, Q= 1.5 L/min, NaCl solution) Figure 78: Flux profile (Tf= 65 C, Tp= 40 C, Q= 1.5 L/min, NaCl solution) Figure 79: Flux profile (Tf= 60 C, Tp= 20 C, Q= 1.5 L/min, NaCl solution) Figure 80: Flux profile (Tf= 60 C, Tp= 30 C, Q= 1.5 L/min, NaCl solution) Figure 81: Flux profile (Tf= 60 C, Tp= 40 C, Q= 1.5 L/min, NaCl solution) xiii

14 Acknowledgments This research would not have been possible without the guidance and the help of several individuals who in one way or another contributed and extended their valuable assistance in the preparation and completion of this project. My most sincere gratitude goes to my academic supervisors and mentors, Dr. Majeda Khraisheh and Prof. Farid Benyahia from Qatar University, for their professional advice and moral encouragement during my MSc. research. Their comments, advice and ideas certainly greatly helped me focus and organize my experiments and thoughts over the past 2 years. They have been my inspiration as I hurdle all the obstacles in the completion this research. Great deals appreciated go to the contribution of faculty and stuff of Chemical Engineering Department- Qatar University, for their patient in helping us doing ICP analysis and providing necessary facilities. I would like to thank staff of Global Water Sustainability Center-ConocoPhillips (GWSC) for their unlimited support and help doing chemical and physical lab analysis. Last but not least, I would like to thank Qatar University Central Laboratory Unit (CLU) for their support analyzing membrane using SEM imaging and also Texas A&M University-Qatar, for providing their facility for measuring contact angle of membrane. I will be always grateful and indebted to my family for their unconditional love and support; without you I wouldn t have come all that way. xiv

15 Chapter 1: Introduction Water is a key component in determining the quality of our lives. Today, people are concerned about the quality of the water they drink. Although water covers more than 70% of the Earth, only 1% of the Earth's water is available as a source of drinking. Yet, our society continues to contaminate this precious resource. Today, drinking water treatment at the point-of-use is no longer a luxury, it is a necessity! Consumers are taking matters into their own hands and are now determining the quality of the water they and their families will drink by installing a drinking water system that will give them clean, refreshing, and healthier water. Where does our water come from and how is it treated? this is an essential question in human life in which the quality of water is a core answer to this question. There are mainly three source of water for drinking purposes: surface water, underground water, and treated sea water. Surface Water is found in lakes, rivers, and reservoirs. Groundwater lies under the surface of the land, where it travels through and fills openings in the rocks. Sea water is treated by desalination to produce fresh water. Figure (1) shows available source of water globally and their percentage to total world s availability. As it can be seen from this figure, sea water is predominant in Gulf region whereas natural source of water such as rivers and brackish water is main source for human consumption in other parts of the world. 1

16 Figure 1: Different type of water around the globe and corresponding total percentage. Adapted from Corrado, Sommariva. "Basic Design of Thermal Desalination Process." Workshop. Global Water Sustainability Center. Qatar Science and Technology Park, Doha. 6 Dec The world's six billion people are appropriating 54 percent of all the accessible freshwater contained in rivers, lakes and underground aquifers. The breakdown of water use by human is shown in Figure (2). Industry, 22 Domestic use, 1.4 Irrigation, 70 Figure 2: Human water consumption [World Health Organization] 2

17 It is predicted from WHO that water use will be increased by 50% by 2025 in developing countries and 18% in developed countries. Since sea water desalination is one of the sources to produce fresh water for human consumption, different technologies are used such as MSF, MED, RO, and membrane desalination. Figure (3) shows different types of desalination technologies all around the globe. As it is shown in Figure (3), RO considered one of the dominant technologies in the world and thermal process are the ones coming after that. Literature review section will go through details of each of desalination technologies. Figure 3: different types of desalination technologies around the world. Adapted from Corrado, Sommariva. "Basic Design of Thermal Desalination Process." Workshop. Global Water Sustainability Center. Qatar Science and Technology Park, Doha. 6 Dec The Middle East is the worst region in terms of water resources in the world. Freshwater resources available in the region are less than 1% of the total available global freshwater. However, the region is home of 6% of the world s total population in which last reports indicates the high population growth rate. There are two challenges faced in terms of water resource in this 3

18 region. Firstly, natural water resources are close to zero in this region, and secondly, water consumption rates in the GCC region which is the biggest part of the Middle East region are one of the highest in the world. Natural Water resources in the GCC countries are generally limited due to the low average annual rainfall and high evaporation. Groundwater is encountered in shallow and deep aquifers with various potentialities. Main water resource is coming from desalination plants to substitute for water deficiency. In a recent study by Maplecroft [2], the GCC countries Bahrain, Qatar, Kuwait, and Saudi Arabia were rated as the world s most water-stressed countries, with the least available water per capita. Table (1) shows some statistic of 6 GCC countries in term of water demand and consumption. Table 1: GCC population and water resource [3] Country Area (km 2 ) Population (million) Surface Runoff (MCM) Rainwater (BCM) Annual evaporation (mm) Consumption per capita (L/d) KSA Kuwait Bahrain Qatar UAE Oman Total Avg. 400 One of the strategic options to fulfill current and future demand of water in GCC countries is to build desalination plants. From all the desalination plants in world, 50% of them are taking place in the GCC region. Due to increase in demand of water as a result of population growth and changing life style, capacity of desalination plants are in rapid growth and expected to increase by 35% in GCC countries and 25% globally [4]. As shown in Figure (4) GCC in general and Middle East is two major players in desalination market with huge demand and growth 4

19 throughout the years and expected to increase in future. Desalination process by itself is an environmental issue with huge environmental footprint such as brine discharges, very high energy consumption and atmospheric emission such as GHG. Figure 4: Trends of desalination capacity in the world and GCC region Adapted from Al. Ansari, Mohammed Saleh. "Concentrating Solar Power to Be Used in Seawater Desalination within the Gulf Cooperation Council." Energy and Environment Research 3 (2013): Total desalination capacity of Gulf States by 2012 is 11 million cubic meter per day of fresh water which is around 45% of global production of water by desalination [19]. The main producers of desalinated water in GCC countries are tabulated in Table (2). 5

20 Table 2: GCC desalination capacity and its status regarding world capacity [4] Country Percentage by world desalination capacity UAE 35% KSA 34% Kuwait 14% Qatar 8% Bahrain 5% Oman 4% GCC countries use different desalination technologies (MSF, MED and RO) to satisfying the demand and drought condition. Most of desalination plants run with natural gas as a fuel which makes it more efficient, less cost and more environmental safe than using the petroleum. Although natural gas has better efficiency compared to petroleum fuel, but both release huge amounts of GHG gases such as CO 2 and NO x which has some impact on the environment that might accelerate the increase in temperature and decrease the precipitation. Table (3) shows number of desalination plants in each GCC country and their capacity accordingly. 6

21 Table 3: existing and future planed desalination plants in GCC countries [5] Technology UAE Bahrain KSA Oman Qatar Kuwait Total MSF RO MED VC ED RO+MSF Total Figure (5) shows number and location of above desalination plants with its corresponding technology used and capacity. Figure 5: location and capacity of installed desalination plants in GCC states. Adapted from Lattemann, Sabine, and Thomas Hopner. "Environmental Impact and impact assessment of seawater desalination." Desalination 220 (2008):

22 In bigger scale, looking globally capacity for desalination of seawater is increasing dramatically. According to the latest report from the 24 th IDA Worldwide Desalting Plant Inventory the installed capacity for desalination of seawater approached 77.4 million m 3 /day by the end of 2012 which is distributed among desalination plants worldwide. 77% of these plants are located in Middle East and about two thirds of this water is produced by thermal processes whereas membrane desalination is the predominating process outside the region. Six percent of all plants are located in the Asia-Pacific region, 7% in the Americas, and the rest 10% in Europe [1]. Increasing and growth of desalination in GCC region and other part of the world have shifted attention to the role of desalination in alleviating water shortages. It has been proved that desalination technology has developed to a level where it can be seen as reliable source of water at a price comparable to water from conventional sources such as rivers or aquifer. As it is seen desalination will remain in GCC countries as the most feasible alternative to augment or meet future water supply requirements but many concerns rise over potential negative impacts on the environment such as the concentrate and chemical discharges to the marine environment, the emissions of air pollutants and the huge energy demand of the processes. Effect of traditional desalination technologies are long and among those are huge land use, impingement and entrainment of sea organism due to large feed intake, emission to atmosphere such as CO 2, NO x, and SO x due to considerable amount of energy needed to run the desalination process. A key concern is concentrated rejected brine and chemicals which may have adverse effect on water and sediment quality and damage marine life [6]. To reduce the effect of desalination process some solution has been raised and one of them is using Membrane distillation desalination as an alternative to augment fresh water supply with low energy cost, low expenditure, and minimum environmental footprint. Membrane distillation (MD), which is hybrid between thermal desalination and porous hydrophobic membrane as non-wetting contact media, is currently gaining increasing attention in membrane processes with significant advantages than most of traditional thermal desalination process and reverse osmosis (RO). MD is a separation process using a porous hydrophobic membrane consisting of three main steps: evaporation water in feed side, (ii) followed by transport of water vapor molecules to 8

23 permeate side through membrane pore and (iii) condensation of vapor molecules in cold permeates side. Depending on the way how vapor pressure difference as driving force and vapor condensation are employed, four different configurations of MD are available, namely: direct contact membrane distillation (DCMD), air gap membrane distillation (AGMD), sweeping gas membrane distillation (SGMD), and vacuum membrane distillation (VMD) Objective and Scope of Work This work presents a single stage DCMD process and its application using low grade waste heat from desalination or petrochemical plants as source of energy input and using rejected brine from desalination plants as a feed to reduce both environmental impact and carbon footprint of conventional desalination technologies and at the same time augment the production of fresh water. Quality of the produced water is studied throughout this research and compared with conventional desalination water. Also energy efficiency of the MD process is studied and compared with that of conventional desalination process. The main objectives of this work are: Measuring contact angle of polymeric membrane to ascertain its hydrophobicity Characterize the membrane by SEM imaging to determine its morphology and other derivative properties required for modeling and estimation purposes Evaluation of the effect of feed brine salinity and concentration on the quality of fresh water produced in membrane distillation desalination. Evaluation of the effect of different feed and permeate temperatures on quality of distillate produced Determination of the thermal efficiency of membrane distillation using Direct contact (DCMD) module by examining effect of temperature, flow rate, and concentration 9

24 Chapter 2: Background and Literature Survey In order to produce clean water for drinking and daily use from salt water an appropriate separation process must be developed. Since the separation of salt and water is not a spontaneous process, energy input has to be employed to reach the separation goal. In history, the first desalination technologies were based on the evaporation of pure water through the addition of heat provided by the sun or by combustion processes using some sort of fuel (i.e. burning wood). Although this principle is still applied to many of desalination processes but it is used in more complex manner and higher energy-integrity processes. Large scale commercial desalination technologies are categorized into two broad groups (as large commercial scales): Thermal desalination Membrane desalination Thermal family is those driven by usage of thermal heat through utilizing fossil fuel to evaporate sea water producing fresh water. On the other hand membrane family is those using membrane filter to filter sea water and produce fresh water Thermal Desalination The oldest technology to desalinate sea water or brackish water is thermal process in which the separation happens with phase change of the produced freshwater in form of water vapor. These processes need a significant amount of energy, due to high heat of vaporization of liquid water through producing steam by steam generator, waste heat boilers, and the extraction or backpressure steam from the turbine in power plant stations (co-generation plants) [7]. Commercially, two main types of evaporative desalination processes are used globally, Multiple Effect Distillation (MED) and Multi Stage Flash (MSF) desalination. MED and MSF systems are usually constructed as co-generation power-water plants. This is a convenient and economical since both power and water desalination processes require low pressure steam which can be obtained from power plant at low cost [8] Multiple Effect Distillation (MED) First MED plants were introduced in the industry in the 1960s with a first desalination plant of capacity of less than 500 m3/day. Most MED processes operate as low as temperatures of 70 º C 10

25 due to horizontal film configuration of evaporators where the feed seawater is sprayed on the outside surface of the hot tubes. Several advantages are encountered with this low operation temperature such as decreasing the rate of scale formation on the outside surface of the evaporator tubes and efficient combination with thermal or mechanical vapor compression [8]. Another advantage of MED systems is that their operation at low temperature allows for use of low grade energy (waste heat) from any process stream [8]. To construct MED system fairly inexpensive materials can be used for construction such as aluminum alloys for tubes and carbon steel for shells which can be considered as another advantage [8]. Generally, MED systems used in industries include up to 12 evaporation effects in average, where the energy needed to evaporate the first cell is extracted from cogeneration boilers. In single effect (SED) seawater is boiled producing steam which, by condensation, produces fresh water. By connecting many effects the process would be more efficient and economical. To achieve this, each effect must be operated at different pressures and temperature across the effects drops subsequently to about ºC for the last effect. In a MED plant, the first effect receives sea water which has been preheated in the tubes to boiling point and then the seawater is either sprayed or distributed onto the surface of evaporator tubes in a thin film to promote quick boiling and evaporation. The tubes are heated by steam flowing from a boiler, or other source, which is condensed on the opposite sides of the tubes. The condensate from the boiler steam is recycled to the boiler for reuse. Only a portion of the seawater in the first effect is evaporated and rest of it is transferred to second effect where it is again applied to a tube bundle and evaporated in similar manner to the first effect. These tubes are in turn being heated by the vapors created in the first effect. This vapor is condensed to fresh water product, while giving up heat to evaporate a portion of the remaining seawater feed in the next effect. This continues for several effects, with 8 or 16 effects in a typical large plant [7-8]. MED systems are operated in two modes: either standalone mode, where the water is heated by steam from the boiler, or thermal vapor compression mode, where part of the vapor formed in the last effect is compressed to the desired temperature and used to drive evaporation in the first effect. This heat integration leads to very high energy efficiency for the process, resulting in a performance ratio up to kg of distillate per kg of motive steam fed into the first effect. Figure (6) shows a typical MED process. 11

26 Steam 0.4 bar Condensing MED condenser Distillate pump MED TVC Thermo compressor condenser Steam 2.4 bar Distillate pump Figure 6: Schematic of typical MED process. Adapte from Cipollina, Andrea, Giorgio Micale, and Lucio Rizzuti. "Conventional Thermal Process." Seawater desalination conventional and renewable energy processes. Heidelberg: Springer, Multi-Stage Flash Desalination (MSF) The first MSF desalination process was introduced in the early 1950s. The main component in the MSF process is the flashing chamber or cell. In MSF system water is heated to 100 C or above and held under pressure until it is released into a vacuum chamber, in which water flashes into steam. To make the process more efficient, multiple stages are connected at successively lower pressures and evaporation and condensation is done in many stages, thus increasing efficiency. Seawater as feed is passed through heat exchanger tubing on the exterior of which water vapor, at progressively higher temperatures, is condensing. Finally, feed is passed through brine heater where energy for the process and heats the seawater to the maximum process temperature is supplied in form of steam from an external source (boilers). The seawater then transferred to the evaporator vessel where pressure is decreased, causing it to boil or flash. This process is repeated in many chambers, being the pressure reduced gradually so that flashing occurs at subsequently lower temperatures. The condensed vapor is fresh water [7-8]. 12

27 There are two main designs for the MSF process. The first design is the once-through system and the second is the brine circulation system. Figure (7) summarizes the MSF process [8]. Figure 7: schematics diagram of MSF process with recirculation. Adapted from Cipollina, Andrea, Giorgio Micale, and Lucio Rizzuti. "Conventional Thermal Process." Seawater desalination conventional and renewable energy processes. Heidelberg: Springer, Membrane Desalination Membranes and membrane processes were first introduced as an analytical tool in labs and then developed very quickly into industrial and commercial products and methods nowadays, membranes are utilized on a large scale to produce fresh water from sea and brackish water, to treat industrial effluents, to concentrate and filter mixtures in the food and drug industries, and to separate gases in petrochemical industries [9]. Reverse osmosis (RO) and membrane distillation desalination (MD) are two main type of process use membrane technology for water desalination. Reverse Osmosis (RO) technology account for around 50% of the world s total desalination capacity [8,10]. Membrane is defined as a barrier that restricts the transport of some components in a specific manner. Membrane operates in two modes: Physico-chemical characteristics mode: selective due to some, such as pore size, charged surface, etc. 13

28 Non-selective barrier mode: to separate and contact two adjacent phases [8] Based on thermodynamic equilibrium principles, mass and energy transfers occur at the same time. The basic of separation in membrane is differences in the transport rate of various species through the membrane which is mainly defined by driving force acting on each individual side of membrane. Based on nature of driving force and focusing only on membrane operations used in desalination, membrane technology is divided into three different categories, namely: Pressure-driven membrane processes, such as Reverse Osmosis (RO), Nanofiltration (NF), Ultrafiltration (UF) and Microfiltration (MF). Electrical potential-driven membrane processes, such as Electro-dialysis (ED). Temperature-driven membrane processes, such as Membrane Distillation (MD) and Membrane Crystallization (MCr). Main advantage of membrane based technologies are, their high efficiency and operational simplicity, high selectivity and permeability for the transport of specific components, compatibility between different membrane operations in integrated systems, low energy requirements, good stability under operating conditions, environment-compatibility, easy control and scale-up and great flexibility [8]. Among these, MD and RO are the main technologies used worldwide to produce fresh water [8] Reverse Osmosis (RO) Reverse Osmosis (RO) is a process that uses semipermeable spiral wound membranes and ion specific membrane to separate and remove dissolved solids, organic, pyrogens, submicron colloidal matter, color, nitrate and bacteria from water. Feed water is sent to the unit under pressure through the semi permeable membrane, where water goes through the pores of the membrane (small percentage (1%) of sea salt passes through the membrane, or leaks around the seals) and impurities in the water are rejected in the concentrated form and flushed to the drain. The membranes are called semi-permeable or selective since they reject the salt ions while letting the water molecules to pass through membrane pores. The operating pressure for RO process is dependent on feed salinity [7-8]. The RO process is simple in design and consists of feed, permeate and reject stream. For feed water it is necessary to provide pretreatment in order to remove inorganic solids and suspended 14

29 solid and using high pressure pump given feed through semi permeable membrane. Depending upon the permeate where it is used necessary post treatment is recommended [11]. In the RO desalination process, a pressure greater than the osmotic pressure applied to the saline water will cause fresh water to flow through the membrane while holding back the solutes (salts). The higher the applied pressure above the osmotic pressure, the higher the rate of fresh water transports across the membranes [11]. Figure (8) shows a simple flow diagram of typical RO system. Figure 8: Reverse osmosis basic design. Adapted from Raluy, G, L Serra, and J Uche. "Life Cycle Assessment Of MSF, MED And RO Desalination Technologies." Energy (2006): Membrane Distillation (MD) Membrane distillation (MD) is a thermally driven technology developed over 60 years and first by Weyl and Findley [12-14] where hydrophobic membrane sheets are used to separate hot and cold stream of water and allows water to evaporate due to temperature difference. Nature of hydrophobic membrane allows water vapor to pass and rejects the liquid water to go through the membrane pores. Difference in temperature on two sides of membrane is driving force where cause water to vaporize and condense on the cold surface. The result of this physical-chemical operation is distillate with almost 99.99% salt rejection [15]. Unlike conventional desalination technology such as MSF and RO, MD does not suffer from salt entrainment which is nonvolatile [16]. 15

30 Figure 9: Principle of the membrane distillation process. Adapted from Charcosset, Catherine. "A Review Of Membrane Processes And Renewable Energies For Desalination." Desalination (2009): As shown in Figure (9) sea water flow on the hot side of membrane at around C and cold water flow on cold side at temperature of around 30 C. Since the membrane is selective to water vapor and rejects the ions, the result of this condensation is a distillate of very high purity which, unlike in conventional distillation, does not suffer from the entrainment of species which are nonvolatile (theoretically 100% of ion rejection) [15]. The process starts with passing the seawater or rejected brine from other sources on one side of the membrane at an elevated temperature, for example C. At the other side of the membrane, a lower temperature water at around C, creates a water vapor partial pressure difference between the two sides of the membrane and allows the evaporation through the membrane. The water vapor goes through the pores of membrane and condenses on the lowtemperature side and distillate is formed. Various types of MD have been known for several years: direct contact, air gap, sweeping gas and vacuum are the four forms currently available and will be discussed below. Different membrane configuration are used for cold and hot side but most common ones are Direct contact (DCMD), Air gap (AGMD), sweeping gas (SGMD), and vacuum (VMD) [17-19]. 16

31 Direct Contact MD In direct contact MD (Figure (10)), the feed solution is in direct contact with the membrane on one side and cold distillate (permeate) is in contact with another side of the membrane and first introduced by Lawson and Lloyd, Martinez-Diez and Florido-Diez, Phattarawik and Jiraratanon [20-22]. The temperature of the feed solution is higher than that of the permeate solution to create a driving force for vapor transport across the membrane. If the purpose of the process is to desalinate seawater, the permeate solution is fresh water. Because the membrane is the only barrier between both solutions the obtained water vapor fluxes in direct contact MD are relatively high. Unfortunately this is also true for the energy flux by heat conduction, so that heat losses in direct contact MD are also relatively high. DCMD is well operated for applications where aqueous solutions are needed to be concentrated [5, 20, 22-26]. Figure 10: Different MD configuration (a) DCMD, (b) AGMD, (c) SGMD, (d) VMD Air gap MD Air gap is first introduced by Jonsson et al. [27] and Banat [28] and only the feed solution is in contact with the membrane surface and the permeate is in contact of a cold surface where it condenses. In between there is an air gap between the membrane and the cold surface to reduce energy loss by heat conduction through the membrane. Resistance to mass transfer in AGMD is 17

32 higher than other configurations. Air gap MD is suitable for all direct contact MD applications. However, it is also suitable to separate other volatile substances, e.g. alcohols from an aqueous solution. This is not possible in direct contact MD, because those substances are likely to wet the membrane at permeate side due to lower surface tension and/ or smaller contact angle with the membrane. Since in the air gap MD, permeates is not in direct contact with the membrane, there is no danger of membrane wetting at the permeate side in this case [23-24] Sweep gas MD In sweep gas MD, which is also called membrane air stripping, the vapor at the permeate side of the membrane is removed by a sweep gas and then condensed in external condenser. Like air gap MD, it can also be used for removing volatile substances other than water. An advantage of using a sweep gas is that the resistance to mass transfer is less than that of AGMD. However, drawback is the dilution of the vapor by the sweep gas, which leads to higher demands on the condenser capacity. Or, if a relatively small flow of sweep gas is used, heat transferred across the membrane causes a temperature increase of the sweep gas. This forms a problem, because it leads to higher vapor pressures at the permeate side and thus a lower driving force [23-24, 29-33] Vacuum MD VMD is first developed by Bandini et al. and Sarti et.al. [34-35]. Instead of using sweep gas the vapor can also be removed by evacuation and subsequently external condensation. Vacuum system is applied to increase or establish the vapor pressure difference between the membrane sides Vacuum MD can be used for the separation of various aqueous mixtures with volatile compounds, and recently it has also been proposed as a means for seawater desalination [23-24]. 18

33 2.3- Desalination Technologies: Advantages and Disadvantages Despite the promising vision of desalination technology to provide the world with fresh and healthy water with minimal environmental, social, and technical issues and impacts; significant challenges such as water quality, feed water intake, environmental footprint, energy consumption, concentrate management, and social concern are faced Water Quality Desalination is known for its low recovery and has reported that water recovery ratio 30% to 35% is a typical value for most of desalination plants [1]. Operation of desalination plants is restricted to certain operation conditions and parameters and it tends to require extensive pretreatment, especially if the feed is taken from an open seawater intake. The major contributor to high expenses of desalination is due to characteristic of feed and properties such as: (1) degree of hardness, (2) turbidity, and (3) TDS. Those seawater properties give rise to three major problems in seawater desalination, which employ severe restriction and have distinct effects on the performance and efficiency of seawater desalination plants. Hardness One of the major operational issues in the desalination techniques whether thermal, RO, or membrane is high degree of hardness. As discussed earlier, the basic operation of desalination is based on separation of fresh water from saline water. By this the salts and hardness ions are left behind in the brine with the effect that both the solid concentration of brine and hardness concentrations are increased. Because hardness ions are soluble in seawater the increase in their concentration and under certain operation conditions could lead to their precipitation on the desalination equipment, e.g., tubes, membranes, etc., causing them to scale and therefore causing fouling which leads to reduction in thermal efficiency. Depending on the desalination technology and nature of operation, different types of salt scale could form: Alkaline soft scale made of CaCO 3 and Mg(OH) 2 Non-alkaline hard scale consisting of CaSO 4, or CaSO 4.½H 2 O or CaSO 4.2H 2 O [44] The formation of the non-alkaline scale increases by increasing temperature because CaSO 4 solubility decreases as the solution temperature is increased which makes the thermal desalination (MSF and MED) more affected with fouling issue [36]. 19

34 To prevent scale formation and fouling whether in tubing for thermal desalination or membrane surface for RO and MD certain anti-scalant additives are added to the feed, e.g., polyphosphate, polyphosponates or polycarboxylic acid and H 2 SO 4 or HCl [36]. Turbidity Another issue in seawater desalination is presence of impurities in seawater feed which increase turbidity. Feed turbidity due to the presence of macro particles and macro organisms such as mussels, barnacles, algae has to be reduced for both SWRO and thermal desalination plants. Removal of turbidity and fine particulates from feed to SWRO and MD process is a must but for thermal process is not critically required. Feed salinity The high TDS of the seawater feed constitutes a major problem to the most of desalination technologies especially RO. Below effect of salinity to each process is given in details. Thermal Process (MSF/MED) Increasing feed salinity affect the performance of thermal processes by increasing the boiling temperature or as it s called boiling point elevation (BPE). Boiling point elevation lowers the vapor temperature and as a result decreases the driving force for heat transfer. As a result increasing boiling point need more energy input to the system and as a result lower the performance of the system. Reverse osmosis (RO) in RO salinity contributes to increase in degree of hardness as sea water salinity increases in which it affect the feed osmotic pressure by increasing the ionic molar concentration, and, therefore, the seawater TDS are increased. For SWRO the applied pressure is directly used to overcome the osmotic pressure (π) and the remaining pressure is the net pressure driving water through the membrane. Hence, the product water quantity (Q p ) is directly related to net pressure and the less is the osmotic pressure Δπ the greater is the net pressure and, therefore, the greater is the amount of pressure available to drive the permeate water through the membrane and the greater is the quantity of product, which in turn, lowers the process energy requirement [36]. In reverse osmosis (RO) operations, when feed salinity increases it substantially reduce the driving force for mass transfer across the membrane and it significantly increase the salt leakage through the membrane pores to permeate side which will affect the produced water quality. presence of high salt concentration in the feed to RO also indorse concentration polarization, 20

35 scaling, compaction of a cake layer, and increased osmotic pressure that leads to reduced performance [37]. Membrane desalination (MD) Dissolved components reduce the vapor pressure of water and as a result, the vapor pressure of the water decreases as the salt concentration in the feed stream of MD processes increases, which lead to decrease in driving force for evaporation. Because of that, if salt concentration in the permeate stream is increased also, flux will be increased due to reduced vapor pressure in the permeate, a phenomenon used in the osmotic distillation process [37]. When feed salinity increases another boundary layer parallel to the temperature and velocity boundary layer develops next to the membrane interface, this concentration boundary layer, with the temperature boundary layer further reduces the driving force for evaporation [37]. This effect is one of the most significant advantages of the DCMD process for desalination where feed salinity has minimal effect on the performance of the system [37-38]. Results in other studies [37, 39-41] have shown total flux declines of 13 28% for MD systems operated with feed concentrations of g/l NaCl. Although both RO and MD operate using membrane technology but the effect of feed concentration on the performance is different. Concentration polarization presents in both RO and MD process and is defined as a layer formed near the surface of the membrane, where the water immediately next to the membrane surface becomes depleted in the permeating solute on the feed side of the membrane, and its concentration become less than that of the bulk fluid. On the other side, the concentration of the non-permeating component increases at the surface of membrane resulting in a concentration gradient nearby to the membrane surface. In RO processes, presence of concentration polarization tends to elevate the osmotic pressure at the membrane surface resulting in decrease of the driving force for mass transport but in MD process, concentration polarization lower the feed membrane interface partial vapor pressure which only insignificantly reduce the driving force for evaporation. As a result the pressure in RO reduce the product water flux due to compaction of the concentration boundary layer and slower shear flow; and in MD minimal effect is observed due to the absence of pressure and high turbulence [37]. Quality of water affects all stages of desalination and specially recovery of desalination process. The main issue to the effect of water quality to recovery of desalination process is the 21

36 concentration of salts in feed water (sea water). The best way to express this fact is through concentration factor which is defined as: (1) Where R is water recovery percentage of desalination process. Table (4) summarizes CF and recovery of different desalination process. Table 4: Typical range of concentration versus desalination recovery [1] Parameter MSF MED RO Recovery % CF To produce a metric cube of water different plants need different water quality intake to their process. According to Corrado [1], in term of water intake, RO need less amount of water compared to other thermal process Process Effluent Brine At the end of the desalination process, the saline water is separated into two streams: one with a low concentration of dissolved salts and inorganic materials (the potable water stream) and the other, containing the remaining dissolved salts (the concentrate stream or also called brine discharge). The amount of concentrate flow discharged from desalination plants varies generally from 15 to 85 percent of the feed flow, depending on the feed water salinity and technology used in the plant based ot its maximum recovery. Concentrate or brine as it defined in literature [42] is the byproduct from desalination. Brine is usually liquid materials that may contain up to 20% of the treated water. As defined by [42] concentrate stream with TDS level of higher than 36,000 mg/l is referred to as brine. Major parameters in brine are TDS, temperature, and specific weight (density). The brine depending on type of technology used may also have some amounts of certain chemicals used for pretreatment and post-treatment of either feed water or product water [42]. 22

37 Quantity of brine produced by desalination process directly proportional the desalination process recovery rate (product water/feed water) [42]. Generally, membrane based technologies have a higher recovery rate than distillation plants, resulting in a higher salt amount in the concentrate [42]. As shown earlier in Table (4) brine produced from RO can have more salt concentration than the receiving water compared to that of thermal process. The content of brine from desalination plants based on technology used is summarized in Table (5) below. Table 5: type of effluents from different desalination process [42] Technology type Discharge flow Content RO/ Thermal/MD RO Thermal Seawater concentrate Chlorination Antiscalants Filter backwash De-chlorination Coagulants Flocculants Corrosion inhibitors Antifoaming Salinity and Heat Chemical Chemical Suspended solids Chemical Chemical Chemical Chemical Chemical The types and the amounts of the chemicals used depend on the chosen technology and the required quality of the product water. Chemicals that are likely to be found in the brines include antiscaling materials, surfactants, and acids used for the lowering of ph. The salts returned to the sea are identical to those present in the feed water, but they are now present at a higher concentration. For RO process, the rejected brine has salt content of about 30-70%, or times more than original feed water concentration [43]. On the other hand for MSF plant the rejected 23

38 concentration reach up to l.5 times more than original concentration of feed water [43]. The chemicals used in the pretreatment of seawater used for desalination are mainly: NaOCl or free chlorine, used for chlorination, preventing biological growth (antifouling). FeCI used for the flocculation and removal of suspended matter from the water. SO or HCl, used for ph adjustment. SI-IMP (NaPO,), and similar materials, prevent scale formation on the pipes and on the membranes. NaHSO, used in order to neutralize any remains of chlorine in the feed water [42][43]. Above materials are chemicals which are approved by health organizations such as American EPA to be used in systems for drinking water. Among technologies available, due to high pretreatments required, RO system is not advised for areas having temperature higher than 40 ºC, high sea salinity level, high silt density, high bacteria activity, and high pollution [41] Thermal efficiency and Energy Consumption The minimum amount of energy needed to separate a saline solution into fresh water and rejected brine depends mainly on the salinity of the saline feed, regardless of the type of process and system configuration of the desalination process. In other words, different desalination technologies share a minimum energy requirement for driving the separation process, regardless of the type of system [9]. Thermal and reverse osmosis process For operation of MSF and MED desalination plant two types of energy are needed. The first and main type of energy required to operate the process is in form of low temperature heat which is fed into the system through the heat input section and is supplied as mean of steam generated in boilers or from power generation plants. The second form of energy is in form of electricity, which is needed to run the system s pumps [44]. In GCC countries MSF desalination plants are usually built as dual-purpose power/water production systems. However, in dual purpose plants low-temperature heat is usually received in form of steam from the power generation plant. The low-temperature heat contributes mainly to the total energy input to the system regardless of source of heat whether it is provided by power plant, waste heat recovery boiler, or fuel-fired boiler. 24

39 The electricity required in thermal process is mainly used to operate utilities such as brine recycle, brine blow down, distillate and condensate pumps, in addition to feed water transfer pumps, main intake pumps and other auxiliary pumps for chemical dosing. In RO pumps are the main source of energy consumption on the system which is mainly run by electricity. The correct type of pump and its efficiency for the pressure and flows of the system will have the most effect on the energy costs to operate the system. RO membrane energy consumption is related to site-specific salinity and temperature and other design-specific characteristics such as hydraulic loading rates (flux) and the percentage of feed water recovered. Cabassud et al. [39] stated that the energy consumption of RO varies between 4 kwh/m 3 for process with energy recovery and 12 kwh/m 3 for system with no energy recovery [41]. Table (6) below summarize the type of energy and amount of energy required for different type of desalination technique. Table 6: Thermal and electrical consumption of different desalination process [44] Process Thermal energy (kwh/m 3 ) Electrical energy (kwh/m 3 ) Total energy (kwh/m 3 ) MSF MED RO Usually and as discussed earlier, the main factor effecting the energy requirement is type of feed. In general, as the feed salt concentration increases, the amount of energy required to produce fresh water increases as well. Table (7) summarizes some value of energy corresponding to feed salinity. 25

40 Table 7: minimum energy requirement versus feed salinity [44] Salinity (ppm) Min energy (kwh/m 3 ) In Gulf region, since the sea water salinity is around ppm, the min energy required to produce a metric cube of water is around 0.71 kwh. The energy required to desalinate a metric cube of water varies from one plant to another and from technology to technology, and the reverse osmosis technology is the most energy efficient. As stated by Einav et al [43], in general the amount of electricity necessary to produce 1 m 3 of water varies between kwh/m 3. This means plant producing 100 million m 3 /y water would require an electrical output of MW of energy [43]. The efficiency of heat used in desalination is evaluated using Gain Output Ratio GOR and is defined as the ratio between the amounts of water produced per unit mass of dry saturated steam supplied to the system. (2) Typical GOR values for large-scale commercial MSF plants range between 8 and 10 kg/kg. The GOR value of 8 is a very common Figure for MSF plants in the GCC countries operating at TBT of approximately 91 C. As the TBT is increased to some 110 C for the same plant, GOR value reaches 8.6. For MED plants the typical GOR value is about 8.5 for TBT value of 91 C and it reaches 12 when the TBT increases to 110 C [44]. 26

41 2.4- Environmental Footprint of Desalination Process Building and operating a desalination plant have different environmental footprint. There are five major aspects to the impact of desalination plants on the environment which are discussed below: Land use: mainly plants are located next to the seashore for pumping stations rather than for recreation and tourism which has adverse effect on land usage and marine environment [43]. Impact on the aquifer: In case of constructing the desalination plant inland in order to minimize the impact on the beach, due to need of piping and pumping of sea water to the site and rejection of brine from the plant to the sea; Leakage from the pipes can be an environmental hazard to the aquifers [43]. Effect on the marine environment: return of rejected brine to the sea has an adverse effect on the marine. Although the brine contains materials, which is originally presents in the sea environment but its high specific weight and the potential presence of additional chemicals introduced in the pretreatment stage may harm the marine population in the area of the discharge of the brine [43]. Piping from sea and to the sea also has adverse effect to the sea environment. Impact of noise: using of pumps, turbines, and generators create noise pollution to the area near to the site [43]. Intensive use of energy: as described before, desalination technologies are energy intensive processes which directly affect the environment by producing pollutants such as CO x, NO x and SO x [43] Green House Gas (GHG) Emissions Among above factors the intensive use of energy is a concern in GCC countries due to higher energy consumption to produce one metric cube of water since the salinity is higher and this directly affect the environment through producing more pollution encountered with burning fossil fuels. As reported by literature [7,45-47] depending on type of fuel, type of process, boiler efficiency, salinity, and other operational parameters each desalination and power plant associated with air emission for producing a metric ton on fresh water. Table (8) summarizes some inventories from different studies for air emission per meter cube of water produced. 27

42 Table 8: relevant airborne emissions produced by desalination technologies (all units in kg/m 3 ) Reference Airborne MSF MED RO Raluy et al.[7] Raluy et al. [46] Raluy et al. [46] Bushnak [47] CO NO x SO x CO NO x SO x CO Dickie [45] CO As shown in Table (8), different ranges of emission depends mainly on type of technology and whether energy recovery is used or not and if the system is stand alone or cogeneration plant. Mainly cogeneration plants are releasing higher emission compared to only desalination plants due to huge energy used to produce electricity. In state of Qatar, potable water demand for drinking purposes is increasing every year by 10% which can lead to environmental issues. As shown in Figure (11) current and forecasted demand of water in Qatar is shown. If only an MSF plant is operated to produce fresh water and by doing simple calculation, one can estimate amount of CO 2 released per day in By numbers shown in Table (8) if we assume the plant produces 15 kg of CO 2 to produce one metric cube of water producing 361 MIGD of water in 2020 produces 24 million kg CO 2 per day. By looking at Qatar current status of CO 2 emission per capita which is highest in the world, the projection of water demand and its carbon footprint is significant and solutions have to be considered. 28

43 MIGD Year Figure 11: current and forecasted water demand in state of Qatar [48] The intensified use of energy by the desalination plant results in indirect environmental impacts, since the energy requirements of the plant increase the production of electricity and steam (heat), the burning of fuels and in turn the boost the process of global warming. However, the important fact is to recognize between reducing GHG emissions and reducing fossil fuel energy use. One great step to reach this goal is to use renewable energy such as wind or solar energy or through utilization of waste heat which comes from power plant or petrochemical industries in MD process. Utilizing waste heat has an advantage of leaving its footprint behind while it can be useful to be used in MD process where low temperature is needed and by doing so one can augment water production with same environmental footprint. The niche application of MD process in GCC countries would be by utilizing waste heat where no much capital and operating cost would be required for pumping, building infrastructure, and water intake while producing more water with same amount of water intake used before and same environmental and carbon footprints [15]. Due to high energy consumption, the desalination industry is exacerbating air pollution through NO x and SO 2 emissions. However, NO x emissions are decreasing due to technological upgrades and SO 2 emissions fluctuate depending if oil is used instead of natural gas. In addition, the water production sector is the second largest emitter of CO 2 and contributor to climate change after the 29

44 oil sector in GCC countries. Fossil fuel consumption in desalination plants is expected to continue to increase as new desalination capacity becomes operational with the increasing water demand Membrane Distillation Process: A Promising Solution As pointed earlier, one of the major advantages of MD over other conventional processes such as RO or thermal process is its lower operating temperature which directly affects the energy consumption [49]. Since MD is hybrid process of membrane technology which is driven by thermal process, its low temperature operation and on the other hand low water production made it an energy inefficient process compared to other desalination techniques as reported by number of studies [26,50]. For a process to become a good application in industry, high flux with moderate energy consumption is highly required. It worth mentioning that although the flux is low, but using renewable source of energy or waste heat energy from thermal process or other petrochemical process can make the process more efficient and also environment friendly process. Ultimately, it is the careful integration of MD in petrochemical plants and power plants/desalination that makes it efficient and convenient. Throughout the context, by conducting experiments the energy efficiency of the process is evaluated in more details State of the Art in Membrane Distillation In this section a literature survey of membrane distillation is presented and the publications with the corresponding research teams involved are reviewed. Membrane distillation is a relatively an emerging technique in water desalination and is still in the development and advancement stage. The MD concept is not new and has been known for almost fifty years but most of studies were done on pilot scale or bench scale units and there are no commercial scale plants available. The historical development and recent research on MD is reviewed in this section. MD technology evolved in different stages including the developments of suitable polymeric materials and their transformation into uniform membranes, configuration of the membrane, and development of heat and mass transfer models. Findley et al. in 1967 [12] was the pioneer in testing MD for the first time to separate volatile compounds from water. Membranes he used were not suitable for the tests and he concluded that MD technology can be feasible evaporation technique if considering lower cost compared to 30

45 other methods, extending life time of the process, and having higher temperature [12]. Findley et al. [13] enhanced the hydrophobicity of the membrane using Teflon material but it was not practical to use it as commercial application. By 1980s, MD emerged as one of the candidate for good separation techniques. Cheng et al. [51-52] in 1981 examined separation of fresh water from saline water and recommended use of composite membranes in which hydrophobic micro porous membrane is supported by a hydrophilic layer to prevent wetting and fouling caused by crystal growth of salt through membrane pores. Later, Gore [53] and Hanbury and Hodgkies [54] came up with different configuration as spiralwound module in which the feed enters a roll at the center. Using this module they could produce high purity water. The main advantage of this module was the recovery of latent heat of condensation as a source of feed preheating. In terms of heat and mass transfer modeling, Jonsson el al [28] were the first group introducing theoretical heat and mass transport model in the air gap system. Their model did not include the effect of temperature polarization and was only depended on the heat and mass transfer equations. Drioli et al [55-57] experimentally studied the fundamental features and parameters of membrane distillation such as flux, temperature, and concentration relationships. They concluded that membrane distillation can be a feasible technique to concentrate saline solution and producing fresh water. Drioli et al had done a research on dye solutions to investigate the possibility of using membrane distillation in waste water treatment process [58]. Also they carried out experiments to study the appropriateness of MD in the concentration of orange juice although viscosity of the fluid can be a limiting factor of the process [59]. They developed a model based on Knudsen diffusion to fit the experimental data by neglecting the concentration polarization effect. Gostoli, Sarti and coworkers have applied MD to desalination of saline water to produce pure water [60-63]. Different configurations were examined using PTFE flat sheet membranes. Their main objective was to consider the relation of temperature of both hot and cold side on flux of fresh water (permeate) produced. Different mathematical model were developed in these works [64]. 31

46 Gostoli, Sarti and coworkers also studied experimentally and theoretically the possible use of sweeping gas membrane distillation for desalination [65]. Schofield et al [66-68] investigated the direct contact configuration for producing distilled water from aqueous solutions. They considered temperature polarization effect on their semi-empirical models. Their main finding was reduction in flux due to vapor pressure reduction for saline systems. In 1987, Kimura et al. [69] have studied air gap membrane distillation for different aqueous solutions using PTFE membrane sheets. Analyses were based on molecular diffusion of water through stagnant air at atmospheric pressure. The main conclusion was that high feed viscosities reduce permeate flux. Using plate and frame module, Kubota et al. [70] performed experiments on seawater desalination by using a single stage process tested on different membrane sheets. Their main objective was to test heat efficiency of the process and how parameters affect it. They concluded that heat loss in the tested modules was large. Schneider et al. [71] studied the membrane morphology by looking at the pore size and porosity of membranes by testing the sodium chloride feed concentration and its effect on flux. The effect of pore size on liquid entry pressure was also investigated. Again in 1988, they realized that MD is still not competitive with large desalination techniques but can be used with cheaper waste heat available. Hogan et al. [72] in Australia tested the viability of MD operated with a solar powered plant for the supply of domestic drinking water in arid zones. A hollow fiber module integrated with a solar collector and heat exchanger was used and they found heat recovery of 60 to 80% with low feed flow rates and large heat transfer areas requirement. Sakai et al. [73] studied the effect of temperature and concentration polarization using PTFE flat sheet membranes by testing variety of feed solutions including pure water, NaCl solution and bovine plasma and blood. They noticed an increase in water vapor permeability. Kurokawa et al. [74] also studied the effect of concentration polarization using acidic solutions. They performed the test using PTFE membranes and an air gap module. They concluded that flux decreases when feed concentrations increased. The salt rejection factor was about 99.9%. 32

47 Wu et al. [75] and his group expected that MD technology will become an option in future to treat industrial waste water. Fujii et al. [76] have done experiments on studying the effect of pore size on the separation factor of dilute alcohols from aqueous mixtures. Direct contact system with membrane pore size smaller than those of microfiltration membranes were used in experimental setup. They concluded that flux will change by changing membrane morphology. Zarate et al. [77] used a stirring device in MD configuration with different PTFE membranes to study the effect of stirring on flux. NaCl solution was used and model was employed to show the result as function of temperature polarization. In 1993, Agashichev and Sivakov [78] developed a model in which it takes into account the temperature-concentration polarization effect based on mass and heat balance. Their model was capable of predicting the temperature and concentration at the membrane surface in plate and frame configuration. They were pioneer in developing such model. J. Phattaranawik et al. [79] investigated the effect of spacer in DCMD. They found that spacerfilled channels is shown to achieve fluxes 31 41% higher than without spacers. They concluded that the temperature polarization coefficients are substantially increased and approach to unity when the spacers are used in the channels. M. Gryta et al. [80] studied Wastewater treatment using DCMD for a feed containing NaCl and protein as well as the effluents produced during the regeneration of ion exchangers. They found the electrical conductivity of the distillate to be around 2 4 μs/cm. a major deposits which caused fouling were observed in this study due to feed characteristic. They suggested feed pretreatment before introducing the feed to the system. T.Y. Cath et al. [81] investigated MD performance using vacuum enhanced direct contact membrane distillation (DCMD) with a turbulent flow regime and with a feed water temperature of only 40 C. They found that the enhanced configurations offer less temperature polarization effects due to better mixing and better mass transport of water due to higher permeability through the membrane and due to a total pressure gradient across the membrane. The performance of the new configuration was investigated with NaCl and synthetic sea salt feed solutions and they claimed a salt rejection greater than 99.9% in almost all cases. 33

48 Criscuoli et al. [82] studied three different lab-made flat modules longitudinal, transversal and cross-flow membrane for carrying out DCMD and VMD experiments and their performance were compared in terms of trans-membrane fluxes, energy consumption/permeate flow rate ratios and evaporation efficiency. The transversal-flow module and longitudinal-flow module behaved similarly in terms of flux on the other hand, VMD performed better than the DCMD in terms of trans-membrane fluxes, energy consumption/permeate flow rate ratios, evaporation efficiency. A Criscuoli et al. [83] studied VMD to treat water containing different types of dyes and effect of different parameters such as feed temperature, feed flow rate, feed concentration, on the permeate flux and on rejection has been investigated. Their conclusion indicated that the permeate flux increases with feed temperature and flow rate, due to the higher vapor pressure and to the higher heat transfer coefficient, respectively, and that it has close relation with the chemical properties of dyes. M. Gryta [84] studied the bicarbonate deposition on the membrane surface operated with DC configuration. He found that the high feed temperature enhanced the decomposition of HCO3 ions. D Hou et al. [85] used DCMD with polyvinylidene fluoride membrane to remove fluoride from brackish groundwater. The maximum permeate flux they reached was 35.6 kg/(m 2 hr) was at feed temperature of 80 C and the cold distillate water at 20 C. The feed concentration had no significant impact on the permeate flux and the rejection in fluoride. The membrane module efficiency declined gradually, by increasing the concentration due to fouling formation of CaF 2 deposits on the membrane surface. P Pal et al. [86] used DCMD driven by solar energy using PTFE membrane with cross flow module to separate arsenic from contaminated groundwater. They claimed almost 100% arsenic removal without wetting membrane pore. The maximum flux achieved was kg/m 2 h. D. Winter et al [87] studies the full scale spiral wound MD-modules. They found that the salt rejection rate throughout all experiments is very high and they reached a high purity distillate production with the spiral wound modules. In double stage process the distillate conductivity was dropped down to 0.19 µs/cm. 34

49 In 2011, H.J. Hwang et al. [88] investigated module dimensions on performance of DCMD using PTFE (poly tetra fluoro ethylene) membrane. Different characteristic such as liquid entry pressure (LEP), contact angle (CA), pore diameter, effective porosity and pore size distribution, were employed in the study to develop a two dimensional (2D) model containing mass, energy, and momentum balance to predict permeate flux production. They found the fluxes increases as temperature and velocity increases, and they seem to reach maximum values asymptotically at high velocity. The values of mass transfer coefficients observed in this study were in the range of L/m 2 hpa. They concluded that the flux and vapor pressure differences decreased with an increase in the NaCl concentration due to polarization layers formed on the membrane. Bahmanyar et al. [89] studied the effect of operating conditions such as feed flow rate, temperature, and concentration on DCMD. They specifically studied effect of these parameters on temperature and concentration polarization. They developed a simultaneous mass and heat transfer model and solved numerically by MATLAB to predict the flux. They found using the model the optimal value for the membrane thickness is in the range of 30 to 60 µm. A. Alkhudhiri et al. [90] studied the influence of membrane pore size in MD performance using feed of different salts such as sodium chloride (NaCl), magnesium chloride (MgCl 2 ), sodium carbonate (Na 2 CO 3 ), and sodium sulphate (Na 2 SO 4 ). Also, the effect of feed concentration, feed temperature, coolant temperature and feed flow rate on permeate flux is studied. Conclusion has been reached that flux declines as the concentration of salt and coolant temperatures increase, and increases as the feed temperature and flow rate increase. Furthermore, the energy consumption measurements clarified that membrane pore size and feed concentration has no direct effect of energy consumption. A. Criscuoli et al [91] for the first time applied VMD to remove arsenic from water. The test showed that no arsenic was detected in the permeate water under different flux ranged between 3 and 12.5 kg/hm 2 at 20 C and 40 C, respectively. The efficiency of the system in terms of flux and arsenic rejection was not affected by arsenic concentration and type. E.K. Summers et al. [92] at MIT and his group at MASDAR investigated the energy efficiency of the three types of MD (DC, VMD, and AG). They have developed a model and was validated using experimental data and the conclusion have been reached that the GOR of single stage 35

50 VMD systems is limited to less than 1, even with ideal heat recovery/recirculation of the brine reject. On the other hand AGMD and DCMD have the potential for high GOR if properly optimized. Y. Huo et al. [93] used direct contact membrane distillation to study the removal of methyl orange from an aqueous solution. They found that the water flux was steady and no membrane fouling was observed throughout the study. J. Zhang et al. [94] and his research group were pioneers in developing a new model for vacuum membrane distillation. This model measures the gas permeability as a function of membrane length for modeling the flux. The model prediction and experimental results were almost identical with 10% error. S. Adham et al. [15] and his research group at GWSC studied the feasibility of using MD technology to desalinate brines from thermal desalination plants. Flat direct contact MD bench scale unit was used with different MD membranes was compared under various different operating conditions using synthetic saline solutions, brine from a thermal desalination plant and seawater from the Persian Gulf. The study showed the water produced has high quality distillate (conductivity below 10 μs/cm) from high salinity brines (70,000 mg/l TDS) from a thermal desalination plant. They suggested pilot plant testing with a consortium consisting of QU, QEWC and KAHRAMA to study the feasibility of MD in bigger scale. G. Chen et al. [95] studied DCMD with gas bubbling and examined its effect on the MD performance especially at elevated salt concentrations in the feed steam. The study showed 26% flux increment when feed was concentrated from 18% to saturation also gas bubbling can delay the occurrence of flux decline due to crystal deposition. G. Chen et al. were the first to introduce bubbling into configuration of MD. 36

51 Chapter 3: Direct Contact Membrane Distillation Theory The performance of direct contact membrane distillation depends on many factors but three factors have major impact on DCMD performance and water quality produced. These factors are: Membrane physical properties The operation conditions: guiding principles The module design and configuration 3.1- Membrane Module and Configuration As discussed earlier MD has four main configurations in terms of permeate side configurations: direct contact (DCMD), air gap (AGMD), sweep gas (SGMD), and vacuum (VMD). Because of its higher flux and simpler structure and assembly, bench scale DCMD is widely used and studied in literature compared to AG and SG [96]. One of the main drawbacks of DCMD is its low energy efficiency. In general membrane material has low thermal conductivity but the driving force for mass transfer also contributes to conductive heat lost due to the small membrane thickness. Due to this high loss, DCMD relatively has low thermal efficiency (defined as the fraction of heat energy used for evaporation) [97-98]. In AGMD the controlling factor for mass and heat transfer is the air gap [99] because of high thermal and mass transfer resistances. In AGMD compared to DCMD, more energy is used to evaporate water. Also, compared to DCMD, the AGMD has lower flux with the same temperature difference between the feed and permeate streams as that of DCMD, due to the high mass transfer resistance across the air gap [100]. In SGMD, because of the reduced vapor pressure on the permeate side of the membrane; mass transfer rate is higher than AGMD. Heat loss through the membrane is lower in SGMD compared to AGMD. Another drawback of SGMD is its higher investment and running cost because of using air blower and air compressor [101]. In VMD, driving force is higher than other three configurations at the same feed temperature, due to reduction in the vapor pressure at the cold side. It appears from literature [15, 79, 96, 100, ] that DCMD is a promising MD technology with more than half of the published references for MD based on DCMD for applications where the major feed component is water such as desalination. 37

52 Although each one of above configuration has its own pros and cons, but direct contact membrane desalination is considered the optimum one in terms of publicity and operation. It has been considered as the configuration in this study due to its simplicity, ease of operation, possibility of using waste heat for its operation, and its application in wide range of industries. Major applications of DCMD are desalination process, waste water treatment, and concentration of aqueous solutions. One of major advantages of DCMD is its ability to operate at low temperature in range of 40 ºC to 80 ºC [105]. More advantages are complete rejection of nonvolatile components, low operating pressure (nearly atmospheric pressure), and low vapor space [102]. According to Dow et al. flux of permeate produced is higher in DCMD compared to other MD configurations. It appears from the literature that average flux measured by DCMD is in the range of 3.6 to 40 L/m 2 hr, with the range being strongly dependent on input temperature difference used [103]. In terms of operational characteristic, literature reported that it has four major advantages over other configuration and they are: The feed need less pre-treatment It can be operated at constant pressure Less affected by membrane fouling It allows heat recovery [107] However, beside advantages mentioned above, DCMD has different disadvantages such as: Low permeate flux compared to other desalination and separation process Mass transfer resistance due to air trapped within the membrane Huge heat lost by conduction [108] In terms of heat lost, Al-Khudhiri et al. [108] reported that heat lost may happen in three ways. The first form is due to trapped air within the membrane, second form is due to heat lose by conduction, and lastly by temperature polarization. Temperature polarization is a phenomenon happens by creating thermal boundary layer on both sides of membrane due to difference in temperature at cold and hot side of the membrane system [108]. In most cases effect of heat transfer to the total heat transfer resistance is measured by temperature polarization. 38

53 3.2- Heat and Mass Transfer in DCMD In MD processes, heat and mass transfers are coupled together in the same direction from the hot side to the cold side [109]. As described in Figure (8) the feed temperature, T f, drops across the feed side boundary layer to T m,f at the membrane surface. Some water evaporates and is transported through the membrane. Simultaneously, heat is conducted through the membrane to the cold (permeate) side. The cold flow temperature T p increases across the permeate boundary layer to T m,p at the membrane surface on the cold side as water vapor condenses into the fresh water stream and gains heat from the feed side. The driving force is, therefore, the vapor pressure difference between T m,f and T m,p, which is less than the vapor pressure difference between T f and T p. This phenomenon is called temperature polarization which was described earlier in this context Heat Transfer Heat transfer in MD which is transfer of heat from feed to permeate side consist of two steps [106]: First, the heat transfers from the hot side to the cold side across the membrane as sensible heat and latent heat, so as to form the temperature difference between boundary layer and bulk flow; second, the heat transfers from the bulk flow of the feed to the boundary layer via heat convection, due to the temperature difference arising from the first step. In the first step, as shown in Figure 8, the sensible heat is conducted through the membrane to the cold side, and the latent heat is carried by the water vapor, which is evaporated at the interface between the hot stream and membrane pores and is condensed at the interface between the pores and cold stream for DCMD [26]. Heat transfer in MD, as a result, can be categorized into three categories. (1) convectional heat transfer in the feed boundary layer, Q f,conv. and the heat transferred due to mass transfer across the feed thermal boundary layer, Q f,m.t. ; (2) combination of both conductive heat transfer through the membrane, Q m,cond. and heat transferred because of water vapor migration through the membrane pores, Q m,m.t.; (3) convectional heat transfer in the thermal permeate boundary layer, Q p,conv., and the heat transferred due to mass transfer across the permeate thermal boundary layer, Q p,m.t. Figure (12) shows resistance to heat transfer in MD [1010]. 39

54 Figure 12: Resistance to heat transfer in MD These heat transfer relations can be expressed mathematically as: Heat Through the feed solution thermal boundary layer: (3) Heat Through the membrane: (4) Heat Through the permeate thermal boundary layer (5) In the above equations, h f is the feed boundary layer heat transfer coefficient, h p is the permeate boundary layer heat transfer coefficient. J is the permeate flux; H v is the enthalpy of the vapor, which is evaluated at the average temperature of feed and membrane/feed interface. J or flux is written in linear as function of water vapor pressure and membrane distillation coefficient, C; (6) Membrane distillation coefficient, C, is function of membrane morphology; like the porosity, pore size, membrane tortuosity, membrane thickness and also vapor properties such as molecular weight, mean free path, and mean membrane temperature [ ]. In above equations H V and H L are enthalpy of gas and liquid water, respectively. H V is calculated using equation (7) as function of temperature in Kelvin in range of K [111,1023]: (7) And H L is calculated using equation (8) as function of temperature in Kelvin in range of K [ ]: (8) At steady state and by assuming isoenthalpic flow of vapor for q v, the heat transfer equations are expressed as: 40

55 (9) By equity of equation 3, 4, and 5, membrane surface temperature on both cold and hot side can be calculated as [111]: { } (10) { } (11) From above equation, the temperature polarization seems to decrease with increasing convective heat transfer coefficient and increase with increasing in mass flux and temperature. One way to enhance convective heat transfer coefficient is through introducing mesh like spacers to enhance turbulence and consequently increase the flux. By introducing spacer, the correlation changes and the new correlation will be: { } (12) { } (13) Where script s indicates presence of spacer in the feed channel. Film heat transfer coefficient for the spacer filled channel, and, can be modified using correlations below: And: ( ) (14) ( ) ( ( )) (15) In above equation Re, stands for Reynolds number and for spacer filled channel can be expressed as: (16) Where, u s is the velocity when the spacer fills the channel, and and are density and the viscosity of water, respectively. 41

56 Velocity can be calculated by: (17) Where Q is volumetric flow rate, is spacer voidage and A is the cross sectional area of empty channel [113]. The hydraulic diameter,, can be obtained using; ( ) (18) Where is thickness of the spacer and is the specific surface of the spacer and calculated by: (19) Where is filament diameter of the spacer [79]. Figure (13) shows how the above parameters can be calculated using the spacers. Figure 13: Spacer properties and parameter. Adapted from Phattaranawik, J. "Heat Transport And Membrane Distillation Coefficients In Direct Contact Membrane Distillation." Journal of Membrane Science (2003):

57 In AGMD, an air gap is interposed between the membrane and the cooling plate, and the percentage of sensible heat loss is less than that in DCMD [ ], but the stagnant air gap also increases the resistance to the mass transfer. Instead of the stagnant air gap, a striping gas is used in SGMD, which boosts the mass transfer and provides good resistance to sensible heat transfer, but there is more energy consumption from the blower and/or condenser if the permeate is the product [ ]. In VMD, the sensible heat loss can even be neglected, if a very low vacuum is employed in the permeate chamber, but it would not be as competitive as DCMD and AGMD if the thermal energy cannot be recovered from the external condenser Mass transfer Transfer of mass in the DCMD process has three main levels: 1. Vaporization of the hot feed from the liquid/gas interface, 2. cross of the vapor due to vapor pressure difference from the hot interface to the cold interface through the membrane pores 3. Condensation of vapor into the cold side stream [119]. Due to above mechanism, there are two major factors controlling the mass transfer: vapor pressure difference, and the permeability of the membrane. Camacho et al has described the mass transfer in membrane as limiting step in the MD taking into account good fluid dynamics conditions on both sides of the membrane [119]. The effect of the physical properties on membrane permeability includes: 1. Membrane porosity: The effective area where mass transfer occurs is less than the total membrane area due to the fact that the membrane is not 100% porous. 2. Tortuosity: Membrane s pore path is not straight through the membrane and it will create extra path length for vapor transport greater than the membrane thickness. 3. Diffusion resistance: The inside walls of the pores increase the resistance to diffusion by decreasing the momentum of the vapor molecules [119]. The mechanism of mass transfer in the DCMD is governed by three basic mass transport phenomena are combination of these three known as 1. Knudsen-diffusion (K): Knudsen diffusion takes place when pore sizes are small resulting in collision between the molecules and the inside walls of membrane [120]. 43

58 2. Poiseuille-flow (P): the gas molecules act as continuous fluid driven by pressure gradient [108]. 3. Molecular-diffusion (M): molecular diffusion occurs when molecular moves and they correspond to each other as a result of concentration difference [108]. 4. Transition mechanism (combination between above three mechanisms) [106,121]. The Knudsen number (Kn) is used to indicate the dominant mass transfer mechanism in the pores and defined mathematically as: (20) Where, d is the mean pore size of the membrane; and l (also denoted as λ) is the mean free path of the molecules defined by Kuhn and Forstering [122] and Albert and Silbey [123] as: (21) Where k b is the Boltzman constant and is equal to J/K, σ w and σ a are the collision diameters for water vapor and air respectively and reported in literature as m for water vapor and m [119]. T is the mean temperature in the pores, and M w and M a are the molecular weights of water and air respectively. Mean free path of molecule also defined by as [124]: Where, k b is the Boltzman constant, T and P are the mean temperature and pressure in the pores. According to Camacho et al. [119] and Al-Obidini et al. [125] for typical operation of membrane distillation at 60 C and using membrane pore size in range of 0.2 to 1.0 μm, the mean free path of the water vapor in the membrane pores is 0.11 µm. therefore, Knudsen number (Kn) will be in the range of 0.5 to 0.1 [119]. Knudsen number is an indication of which mass transfer mechanism is dominant. When Kn is greater than one (Kn > 1) Knudsen diffusion is dominant over other mass transfer mechanism meaning the mean free path of water vapor is larger than that of membrane pores and resulting in more molecule-pore wall collision compared to molecule-molecule collision. In this case, mass transfer which is reported by Khayet et al. [126] is defined as: (22) ( ) (23) 44

59 Where, ε is membrane porosity, τ is pore tortuosity, r is pore radius, δ is membrane thickness and Mw is molecular weight of water and other factors are constants. If Kn < 0.01, and mean free path (λ) is less than membrane pore distance (100λ<dp), ordinary molecular diffusion is dominant mass transfer mechanism which represents the vapor diffusion through stagnant air film and its transport equation is defined as: (24) Where P air is pressure within the membrane pore, D is diffusion coefficient, and P is the total pressure inside the pores which is equal to partial pressure of air and water vapor [108]. Schofield et al. [127] described the vapor diffusion flux as: (25) Where P and Pair are average air and gas pressure respectively. In case Kn is 0.01<kn<1 and λ<dp<100λ, both Knudsen and ordinary molecular diffusion take place at the same time due to collision of water vapor molecules and also diffusion of vapor through the air film [108,126]. Alkhudhiri et al. [108] defined this type of mass transfer as: [( ( ) ) ( ) ] (26) PD is diffusivity of water vapor through stagnant air and found by definition [128]: (27) According to Lawson and Lloyd [23] when pore size are less than 0.5 µm most dominant mass transfer mechanism are Knudsen and ordinary molecular diffusion. Guijt et al. [129] pointed out that molecular diffusion is only happening when the pore sizes are big and oppositely, Zhongwei et al. [120] stated that Poiseuille transfer is dominant in system with membrane having large pores. On the other hand, Schofield et al. [130] and Fane et al. [121] reported that Knudsen and Poiseuille mass transfer are happening when DCMD system is degased and aerated. Two other popular mass transfer models are widely used in membrane distillation mass transfer studies, which are Schofield s model [127,130] and the dusty-gas model for DCMD [119]. Dusty-Gas model [ ] is the mostly used model for DCMD configuration and it assumes the porous membrane to be an array of dust particles held fixed in space, and the dust particles in terms of the classical kinetic theory of gases are supposed to be giant molecules in the 45

60 interactions between gas and surface [119]. Based on this model, Knudsen-viscous transition region is dominant mass transfer mechanism and is defined as [119]: [( ) ] (28) In which (29) And (30) To simplify all of mentioned models and correlations, flux in membrane distillation is simplified in a shorter form equation as: (31) Where is membrane constant and is different for different materil and pore sizes and it is related to different factor by: (32) Where a is an exponent factor and is range of 1 to 2. In this equation d is mean pore diameter of the membrane, ε is membrane porosity in percentage, t is proportion of conductive heat, and δ is membrane thickness [119]. It can be realized that by increasing pore sizes and porosity it the flux for MD can be increased and oppositely, by reducing the tortuosity and thickness of the membrane, flux can decrease. However, reducing the thickness of the membrane also increases the sensible heat loss from the hot side to the cold side, which leads to a reduction of water flux, because of reduction of vapor pressure difference [119]. This can be reduced by increasing the membrane porosity which reduce the heat transfer coefficient (λ/b) of the membrane and consequently reduces the conductive heat transfer across the membrane [119]. When designing membrane distillation module main points have to be considered such as: Concentration polarization The temperature polarization Uniform flow distribution The pressure drop 46

61 The liquid entry pressure Flow turbulence Feed temperature [133] Each one of the above factors directly or indirectly affects the performance of the membrane thermal (heat transfer) and flux performance (mass transfer). Next section reports each one of the above in details Concentration Polarization Since membrane process is used as a separation media and some the membrane has the ability to transport one species and reject some other ones there will be a buildup of concentration at the membrane surface compared to the bulk concentration. The retained solutes accumulate at the membrane surface where their concentrations gradually increase at the interface and decrease at the bulk. Concentration build-up generates a diffusive flow back to the bulk of the feed. At steady state, the flow of the convective solute to the membrane surface is balanced by the diffusive flux flow of the solute from the membrane neighborhood to the bulk. In summary, the solute concentration is higher at the interface of the membrane than that in the bulk. This phenomenon, referred to concentration polarization. Concentration polarization tends to reduce the transport rates across the membrane because, for a given pressure difference and since the transport rates are inversely proportional to the concentration difference across the membrane this will affect the flux and tend to reduce the permeate flux. The effect of concentration polarization has been reported by Mulder, Banat and Lawson et al. and [20,27,134]. Banat [27] found a slight decrease of 6% in the permeate flux when the concentration of the salt increased from 1 to 10 wt% (a 10-fold increase in concentration reduced the flux by 6%) Temperature Polarization As mentioned earlier, MD process is hybrid between thermal and membrane process which undergo both mass and heat transfer. The main driving force in MD process is difference in temperature between hot side (feed) and cold side (permeate) of membrane which cause vapor pressure and drive the process. The direction of heat transfer in MD process is from feed side which is at higher temperature to the permeate side which is colder than feed side and three main transfers is occurring at the same time; Figure (14): 47

62 From the feed bulk to the feed membrane interface (heat transfer through the feed boundary layer) from 0 to 1; From the feed membrane interface to the membrane distillate interface (heat transfer across the membrane) from 1 to 2; From the membrane distillate interface to the distillate bulk (heat transfer through the distillate boundary layer) from 2 to 3 [100]. Figure 14: Heat transfer across the membrane Due to boundary layer presenting at both side of membrane surface (1 and 2), temperature at point 1 is lower than temperature at 0 and temperature at point 2 is higher than point 3. This is due to phenomena called temperature polarization and quantified as temperature polarization coefficient. TPC is ratio between useful energy for mass transfer of vapors to the total energy invested in the process or in another words it is a fraction of the trans-membrane temperature to the bulk temperature difference and mathematically written as [24,26,37,104,108]: (33) Where T mf is the interfacial feed temperature, T mp the interfacial permeate temperature, T f the bulk feed temperature, and T p is the bulks permeate temperature. Both evaporation and 48

63 condensation rate depends on interfacial temperature (T mf and T mp ) and since vapor pressure is function of temperature mainly, it is always at great favor to keep the difference between the two temperatures as high as possible and keeping the τ close to unity. Although it is difficult to keep the τ equal to 1 but TPC of 0.2 to 0.9 is reported in literatures [37]. Temperature polarization has a more effect on the flux compared to concentration polarization. Schofield et al. [66] found that the temperature polarization coefficient can reach 0.6 at a feed temperature of 60 o C i.e. the permeate flux is overestimated by 40% when the temperature polarization is ignored. Therefore, modeling the MD process must consider the temperature polarization. TPC can be enhanced by improving membrane module design and operation parameters such as increasing low mixing and flow turbulence. By increasing flow mixing, thermal boundary layer is reduced and thereby difference between interfacial temperatures is increased leading to higher vapor pressure and consequently more flux Liquid Entry Pressure Due to use of hydrophobic membrane in MD process, feed liquid must not go through the membrane pores and cause membrane wetting. This fact is related to the applied pressure on the membrane in which it has not to exceed a limit known as liquid entry pressure (LEP). LEP is a function of the maximum pore size and the membrane hydrophobicity. It is directly related to feed concentration and the presence of organic solutes, which usually reduce the LEP [14]. LEP can be estimated using the Laplace (cantor) equation reported by numbers of literature which relate maximum pore size to other operational parameters such as: [23-24, 37,135]: (34) where P f is hydraulic pressure on the feed side, P d is the and permeate side hydraulic pressure, B is a geometric pore coefficient, γ l is liquid surface tension, θ is contact angle of water with membrane surface and r max is the maximum pore size. The contact angle is phenomena resulted due to free energy of the surface and the liquid, solid, and vapor. Contact angle in membrane science is introduced to describe the relative hydrophobicity of a membrane surface. In very strong hydrophilic membrane which is membrane with ability to allow liquid to enter the pores, the liquid is attracted to the solid surface and the droplet will completely spread out throughout the membrane solid surface and 49

64 the contact angle will be close to 0. Contact angle of between 0 to 90 is given to less strong hydrophilic membranes [136]. For hydrophobic membrane which is membrane with tendency to resist liquid entering the pores the contact angle will be larger than 90. The contact angle of hydrophobic membrane varies between 90 to 150 or even nearly 180. On the hydrophobic membrane surfaces, water droplets simply rest on the surface, without actually wetting the membrane surface [136]. Contact angle of membrane is estimated using Young s equation as: (35) Where γ l is liquid interfacial tension, γ s is solid interfacial tension and γ lv is vapor liquid interfacial tension. Figure (15) describes each one of above in details. Figure 15: Parameters of Young s equation It worth mentioning that in Young s equation, the contact surface is assumed to be completely flat. Table (9) summarizes contact angle and surface energy of some popular membrane used in MD process. Table 9: contact angle and surface energy of some commonly used membranes [108] Membrane material Contact angle Surface energy (x10 3 N/m) Polytetrafluoroethylene (PTFE) 108 to Polyvinylidenefluride (PVDF) Polypropylene (PP)

65 According to the Laplace (cantor) equation and considering all parameters, a membrane having higher contact angle (high hydrophobicity), smaller pore size, lower surface energy and high surface tension for the feed solution will have higher LEP value. According to [18] a membrane with pore size of about 0.2 µm, the LEP is around 2-4 atm and for pore size of around 0.45 µm, the LEP decreases is up to 1 atm [37]. In operating MD, special care has to be taken in order to avoid membrane wetting which allows water to penetrate the membrane pores and terminate evaporation process and cause problems and affect the water quality. Set of test are conducted to measure the contact angle of membrane sheet used in this study and will be shown in later chapters Flow Turbulence and Flow Distribution One way to enhance the flux is to increase the vapor pressure difference across the membrane or to reduce temperature polarization [ ]. To achieve this, it is important to improve the convective heat transfer coefficient to produce more flux according to heat balance equation described in heat transfer section. The convective heat transfer coefficient can be written as Equation (36) [138]: ( ) (36) Where λf is thermal conductivity of the feed, and ( ) is temperature gradient in thermal boundary layer of the feed. This can be achieved by improving the design of flow passage, membrane arrangement, and by applying turbulence. Promoters like mesh spacers are designed to generate turbulence at the surface of membrane in which, it affect the convective heat transfer coefficient by reducing the thickness of the thermal boundary layer. However, the hydrodynamic pressure has a square relationship to the flow rate [119], and the increased pressure will reduce the effect of increasing turbulence if the membrane is compressible [102,140]. To create better mixing, high flow velocities are suggested in which it both improve mixing and lower the temperature polarization effect [23-24,37]. Increasing velocity on the other hand will create pressure drop in which it is recommended to work under optimum conditions which is tradeoff between pressure drop, velocity, and good turbulence. Increasing pressure cause membrane wetting as discussed earlier. Both flow mixing and turbulence enhance the MD 51

66 flux as reported by literature [37]. The presence of turbulence promoters, e.g., net-like spacers or zigzag spacers shown schematically in Figure (5), above [ ] can effectively reduce the thickness of the thermal boundary layer and improve α f [119]. It is also important that high heat transfer rates are achieved with a low pressure drop in the channels where the feed solution and cooling liquid are flowing [ ,141, ]. From reported data [137], it is found that the temperature polarization coefficient of spacer-filled channels falls in the range of , in comparison with a temperature polarization coefficient for flowing channels without spacers. The effect of Reynolds number on heat transfer for the spacer-filled flat channels is presented and discussed in [137, 141, 145]. It is also noticed that the influence of turbulence on flux becomes less at higher turbulence levels. Therefore, it is necessary to control turbulence within an adequate range to reduce the energy cost associated with pumping Pressure Drop Pressure drop is a natural phenomenon happens as a result of flow of fluid in a channel due to a resistance to flow imposed by the walls and the fluid itself [37]. To keep the fluid moving and flowing in the channel, a minimum pressure must be maintained at the entrance to the flow channel. The pressure drop is defined as [24,37]: (37) Where f is the friction factor, d the hydraulic diameter of the flow channel, ρ the fluid density, and u is the fluid velocity. Precautions have to be carried while designing and operating MD to avoid membrane wetting in such a way to keep the pressure lower than LEP Membrane Physical Properties Another important factor to be considered in MD operation and design is membrane physical properties. The basic requirements for membrane properties for MD purpose are [ ]: 52

67 An acceptable membrane thickness, in which trade-off between membrane permeability and thermal resistance has to be considered to increase flux and decrease thermal resistance. Reasonable pore size and uniform distribution of pore size, limited by the LEP required of the membrane. High hydrophobicity and low surface energy low thermal conductivity high porosity Membrane Porosity Membrane porosity or membrane void volume is one of the major parameters affecting MD process. Membranes having higher porosity reveal greater surface area for evaporation and consequently higher flux. Most of MD membrane porosity in average lies between 30% and 85% [24]. According to Martinez-Diez et al. [24] higher membrane porosity, results in higher permeate fluxes regardless of which MD configuration is used. Another advantage of higher membrane porosity is its lower conductive heat loss because the conductive heat transfer coefficient of the gases entrapped within the membrane pores is an order of magnitude smaller than that of the hydrophobic polymer used for membrane preparation [23-24]. Heat transfer coefficient is related to membrane porosity according to the equation [23]: (38) Where ϵ is membrane porosity, h mg is heat transfer coefficient of the gas (vapor), and h ms heat transfer coefficient of the membrane solid. Value of h mg is always less than value of h ms in order of magnitude therefore value of h m can be reduced by increasing membrane porosity ϵ Membrane Pore Size The typical pore size used in MD application is ranging from 100 nm to 1µm in which increasing pore size increases flux [24]. The membranes used for MD application mainly have pore size in range of 0.2 and 1 µm [23, ]. Pore size has direct relation with type of mass transfer in the MD process. As reported by literature, increasing flux due to increase in pore size is related to type of mass transfer of being Knudsen diffusion for membranes having small pore size and Knudsen-viscous for those having larger pore size [24]. To enhance the flux by playing with 53

68 membrane pore size it is better to have more likely Knudsen diffusion in case of small pore size and Knudsen-viscous transition where the membrane pore size are large [37,153]. To avoid membrane wettability optimum value is needed to be determined for each MD application Membrane Thickness In MD process the permeate flux is inversely proportional to the membrane thickness. Again thickness of membrane has to be evaluated to get the optimum value in which to have high permeated flux and less heat loss by conduction. To have high flux, the membrane should be as thin as possible. On the other hand, to have better heat efficiency the membrane should be as thick as possible due to the fact that in MD heat loss by conduction takes place through the membrane [23,103,154]. According to Hsu et al. [103] the optimum value for membrane thickness is estimated to be in range of µm Pore Size Distribution Membranes in general have distribution of different pore sizes rather than a uniform pore size. As a result, more than one heat and mass transfer mechanisms can take place at the same time. In general pore size distribution has no significant impact on MD performance [103] Pore Tortuosity Membrane tortuosity is defined as the average length of the pores compared to the membrane thickness [24]. In MD studies and literature a value of 2 is frequently assumed for tortuosity factor [23, ]. In general, the MD flux is inversely proportional to the membrane thickness times its tortuosity. It s very difficult to measure the tortuosity of the membrane and usually a constant value is assumed in calculating the MD flux Thermal Conductivity The heat through membrane transferred via conduction through membrane material and heat through vapor transport. The heat by conduction is the mean of transfer of heat by membrane material and thermal conductivity is an important factor to find out this effect. In general, high thermal conductivities tends to increases sensible heat transfer across the membrane and as a result cause reduction in vapor flux due to lowered interface temperature difference. Generally, thermal conductivity of membrane is function of thermal conductivity of air (k G ) and polymeric material (k P ). These two factors are also of function of porosity of material (ε) and thickness (δ) 54

69 of membrane. According to Phattaranawik et al. [22] Isostrain or parallel model is the best model to calculate the membrane thermal conductivity by assuming parallel heat flow through air and membrane material, and it is appropriate for estimating the thermal conductivity as the tortuosity approaches 1. PTFE membranes have been estimated to have tortuosity s of 1.1 [119], and hence Isostrain approach to estimate thermal conductivity is suitable for PTFE membranes. (39) And k G and k P are calculated by correlation introduced by S. S. Ibrahim et al. and Lawson and lloyd [23, 111] as function of temperature in kelvin in range of K; (40) (41) (42) (43) (44) Thermal conductivities of water vapor, air, and different membrane polymer are listed in Table (10). Table 10: Thermal conductivity of different material [111] Thermal conductivity Range (wm -1 K -1 ) Water vapor Air PVDF PTFE PP at 298 K and 0.03 at 348 K at 298 K and at 348 K at 296 K and 0.21 at 348 K at 296 K and 0.29 at 348 K at 296 K and 0.2 at 348 K The most common materials cited by most of the literatures for application of desalination of water used for MD membranes are polytetrafluoroethylene (PTFE), polypropylene (PP) and polyvinylidenefluoride (PVDF) [159]. The porosity of the membranes used is in the range of 55

70 0.06 to 0.95, the pore size is in the range of 0.2 to 1.0 μm, and the thickness is in the range of 0.04 to 0.25 mm [160]. Among all of the mentioned materials, PTFE has the best hydrophobicity (largest contact angle with water), good chemical and thermal stability and oxidation resistance, but it has the highest conductivity which will cause greater heat loss through PTFE membranes; PVDF has good hydrophobicity, thermal resistance and mechanical strength and can be easily prepared into membranes with versatile pore structures by different methods; PP also exhibits good thermal and chemical resistance [161]. Membranes are manufactured using different methods and among them three are well used in industries and they are: Sintering, stretching, and phase inversion [ ]. In sintering method polymeric powder is pushed between a film or plate and sintered below the melting point of the polymer. The resulted membrane sheet has porosity in the range of 10 40% and pore sizes are in the range of 0.2 to 20 μm. PTFE membranes are produced using this method [ ]. Another method is stretching technology in which membrane sheets are produced by extrusion from a polymeric powder at operating temperatures close to the melting point in conjuction coupled with a rapid draw-down. Using this method, membranes produced have pore sizes in the range of μm and porosity of about 90%. PP and PTFE membranes are produced using stretching technology [160, ]. To produce PVDF membranes, Phase inversion method is used. In this process, the polymer is dissolved in a solvent [169] and spread as a μm thick film on proper supports, such as nonwoven polyester, PP backing material or PP scrim backing [167], and an appropriate precipitant (typically water) is added to split the homogeneous solution film into two phases (a solid polymer rich phase and a liquid rich phase). The produced membrane has a pore size in the range of 0.2 to 20 μm, and porosity of approximately 80% [167] Feed Temperature As MD is driven by vapor pressures which vary exponentially with the stream temperature, the flux is affected greatly by the feed temperature. Furthermore, since the heat loss through thermal conduction is linear to the temperature difference across the membrane as according to Equation (3), the proportion of energy used for evaporation will increase as the feed temperature increases 56

71 [119]. However, an increase of temperature polarization due to the high flux and greater heat and mass transfer was also observed with rising temperature [119], but this can be reduced by using turbulence promoters such as spacers [119] Water Quality The quantity and quality of different water type represents significant opportunity for implementation of different desalination technologies to augment water resource worldwide and more focused in GCC region [170]. Usually quality of water is reported as total dissolved solid or salinity. Water salinity is defined and categorized by salt concentration and ranges from fresh, to brackish, to saline water. Most non-seawater resources have salinity up to 10 ppt (parts per thousand). Seawater salinity ranges from 35 to 45 ppt in total dissolved salts (TDS). Table (11) below summarizes the parts per thousand salinity definitions for water. Worth mentioning, seawater salinity has to be reduced approximately one hundred times lower to be considered fresh drinking water. This ratio predicts the large amount of work, or energy, demanded to produce fresh water. Table 11: Salinity of different type of water (units in ppt) [109] Fresh water Brackish water Saline water Brine < >50 Brackish water is defined as type of water that exceeds the secondary drinking water standards of 500 mg/l of TDS or the WHO standards of 1000 mg/l of TDS [109]. In GCC region the Dammam aquifer is the main source of brackish water and has salinity in range of 2500 to mg/l of TDS. The main contributor of the salinity in this aquifer is due to presence of salts such as chloride (Cl), sulfate (SO 4 ), sodium (Na), and Calcium (Ca). Characteristic of brackish water in general is tabulated in Table (12) below. 57

72 Table 12: Salt content of sea water and brackish water [ ] Element Sea Water Concentration (mg/l) Brackish water Cl SO Ca Na Mg K Seawater varies considerably in quality (e.g. salinity, temperature) depending on the geographical location of the sea. Typical TDS concentrations of seawater can range from less than 35,000 mg/l to greater than 45,000 mg/l reaching TDS levels of up to 50,000 mg/l [119]. A summary of several feed water sources and associated TDS concentrations is shown in Table (13). The values indicate average salinity. 58

73 Table 13: Average salt concentrations of different world water sources [109] Water body Salinity, ppm Pacific ocean Mediterranean sea Atlantic ocean Red sea Persian Gulf At the end of any desalination process, the saline water (feed) after being processed is separated into two streams: one with a low concentration of dissolved salts and inorganic materials (the potable water stream) and the other, containing the remaining dissolved salts (the concentrate stream or also called brine discharge). The amount of concentrate flow discharged from desalination plants varies generally from 15 to 85 percent of the feed flow, depending on the feed water salinity and technology used in the plant. The term brine is usually used in the desalination literature for seawater concentrate with higher salinity content > 36,000 mg/l [41], whereas the more general term concentrate can be used for any concentrated stream generated from either brackish or seawater. Brines usually contains different chemicals used to treat the water based on the technology used and mainly can include antiscaling materials, surfactants, and acids used for the lowering of PH. The salts returned to the sea are identical to those present in the feed water, but they are now present at a higher concentration. The types and the amounts of the chemicals used depend on the chosen technology and the required quality of the product water. The World Health Organization (WHO) states that a permissible salinity limit for potable drinking water is 0.5 ppt and 1.0 ppt under limited consumption. The US Environmental Protection Agency (EPA) states that drinking water with TDS greater than 500 mg/l (0.5 ppt) can be distasteful. In Qatar, drinking water quality is established based on WHO and EU standards. Table (14) summarizes some standards for major ions in drinking water in State of Qatar. 59

74 Table 14: Quality of drinking water in State of Qatar [173] Parameter Unit Guide level Max. Permissible limit ph Total Dissolved Solids (TDS) mg/l Alkalinity mg/l HCO Chloride mg/l Cl Total Hardness mg/l CaCO Chlorine Residual mg/l Cl Fluoride mg/l F Sulphate mg/l SO Calcium mg/l Ca Copper mg/l Cu 1 2 Sodium mg/l Na Iron mg/l Fe Manganese mg/l Mg Magnesium mg/l Mg Aluminum mg/l Al Nitrate mg/l NO

75 Status of Quality of Distillate Produced Using MD As per objective of this work, the primary goal is to find and identify the quality of distillate produced using membrane distillation (MD) under different parameters such as different feed solution, different temperature, and different flow rates and effect of each one on water quality produced. Many researcher developed experiments to analyze the quality of distillate produced in MD. S. T. Hsu et al [104] studied the effect of NaCl solution and real sea water in permeate conductivity using Millipore PTFE membrane with different pore size (0.2 and 0.5 µm). They found that measured conductivities of produced water from either DCMD or AGMD are in range of 7-12 µs/cm [104]. Although the flux was affected by type of feed water (sea water showed lower flux compared to NaCl solution) but quality of water remained the same throughout all the experiments. Another group of researcher led by Cath et al. [37] studied different feed such as NaCl and synthetic sea water as feed and they found that the salt rejection by DCMD is around 99.9% for their experiments. They have used three different membranes and all the membranes showed the identical salt rejection of 99.9% [37]. C. Feng et al. [174] used different NaCl salinity solution to find the salt rejection percentage in AGMD using PVDF membrane. They found the salt rejection of 98.7% up to 99.9% with feed solution of 1%wt NaCl, 3.5%wt NaCl, and 6%wt Nacl solutions. Their analysis of distillate showed the concentration of salt in the distillate is around ppm which is less than the limit for drinking water [174]. Pal et al. [86] and criscuoli et al. [91] investigated removal of arsenic from groundwater using solar driven MD (SDMD) by using PTFE membrane in DCMD configuration. Their studies showed 100% arsenic removal for ground water containing 0.6 mg/l of arsenic. Hou et al. [85] conducted research on purification of Nacl, MgSO 4, and CaCl 2 aqueous solution by DCMD at ph of 4 and 9 using solution with conductivities of 14 and 10 ms/cm. the results showed rejection efficiency of % and also indicated that the membrane had no selective rejection for different non-volatile solute and the ph had no major impact on the salt rejection of membrane. Conductivity of distillate was found to be around 3 µs/cm [85]. Khayet et al. [175] studied also the effect of different membrane on salt rejection and found that rejection of 99.95% can be achieved with all the type of membrane. 61

76 Kullab et al. [176] used pilot plants operated with waste heat from power cogeneration plant in Sweden with feed having conductivity of around 467 µs/cm. the configuration of the pilot plant was AGMD and the distillate quality they found was in range of 1-3 µs/cm. Karakulski, M. Gryta [177], K. He et al. [178], Alkhudhiri et al [90], and Adham et al. [15] investigated the effect of different feed solution on quality of water and all reported conductivities in range of 1-3 µs/cm. Removal of Boron in water is studied by D. Hou et al. [179] and the arrived to a conclusion that using DCMD process had high boron removal efficiency (>99.8%) and the permeate boron was below the maximum permissible level even at feed concentration as high as 750 mg/l. In this piece of work, full analysis of distillate will be presented and the water samples are analyzed using Inductive Coupled Plasma (ICP). Generally, salt rejection percentage of the system is defined as [175]: (45) Where, and are salt concentration of feed and salt concentration of permeate, respectively Energy Efficiency The energy efficiency in MD is defined as ratio of latent heat for evaporation over total input heat to the system [180]: (46) Where, is total heat input to the system and is latent heat needed for evaporation. Total heat is equal to latent heat and heat lost by conduction and defined as: (47) Where, is heat lost by conduction. To this date, there is no much work done on energy efficiency of MD in literature and main focus was on temperature polarization and heat transfer in MD and only few researches were done on energy efficiency [82]. Criscuoli et al. [82] was one of the pioneers on doing some studies on energy efficiency of MD by calculating the energy consumption using energy balance of heating and cooling energy requirement as: 62

77 (48) (49) Where and are the heating and cooling energy, m is mass flow rate, is specific heat, and T are temperatures. Subscript f relates to feed side and d for the distillate side and I and o stand for inlet and outlet temperatures [82]. Throughout their studies, the DCMD showed energy efficiency between 27% up to maximum of 30% and they found that flow rate has no effect on thermal efficiency of MD but on the other hand temperature has directly relation with energy efficiency [82]. Their test with VMD showed efficiency of around 75% [100]. Similar to the approach used by [180], Bui et al. [50] used the definition to find energy efficiency in DCMD and they defined the efficiency as: (50) Where is effective membrane area, ρ is density and Q is volumetric flow rate of permeate side. They concluded that feed velocity is an important factor affecting EE as it is steadily enhanced the EE of DCMD [82]. Feed concentration showed a negative effect on EE due to change in vapor pressure of the feed [82]. Another approach to define efficiency of a desalination system including MD is GOR (gained output ratio) of the system and defined as ratio of latent heat of evaporation of a unit mass of product water to the amount of energy used by a desalination system to produce that unit mass of product [92]: (51) Different researcher reported a value for GOR in which they ranges from 0.17 for Criscuoli et al. [133], 4.1 for Lee et al. [181], 1.4 for Zuo et al. [182], and values less than 1 for E. K. Summers et al. [92]. E. K. Summers et al. [150] indicated that VMD has lower GOR compared to DCMD and AGMD due to local condensation which allows heat recovery and as result higher heat efficiency for the system. Alkhudhiri et al. [108] indicated that % of the overall heat to the system is considered to be latent heat and the rest of 20-40% is lost by conduction. 63

78 Chapter 4: Approach and Methodology In this section, details of experimental setup and methodology used are provided. First of all the details of overall DCMD bench scale unit is given and then of membrane, MD test configuration, and auxiliary equipment is explained in more details DCMD Bench Scale Unit Figure (16) illustrate a schematic diagram of a flat sheet DCMD process for the PTFE membrane overall process flow diagram. The bench scale unit consists of balance (1), hot and cold water bath (2), peristaltic pumps (3), reservoir tanks (4), DCMD cell (5), and temperature and pressure sensors (6,7). Figure 16: Process flow diagram of DCMD bench scale unit 64

79 As it can be seen from Figure (16), four pressure transducers ports were installed at the inlets and outlets of hot and cold streams to monitor pressure drop at the inlet and outlet of each stream. Similarly, four temperature thermocouple probes are installed at the inlet and outlet of cold and hot stream to monitor inlet and outlet temperature out and in of the MD cell. Permeate conductivity is measured by installing conductivity probe meter at the outlet of cold stream from MD cell. Hot water from feed reservoir pumped via peristaltic pump to hot bath and after being heated by hot bath, it goes to MD cell and recycled back to the feed. Similarly, cold water from cold water reservoir also pumped to cold bath using peristaltic pump and after being cooled by water bath send to MD cell and recycled to permeate tank. Flux of permeate is measured using balance by measuring initial weight of water and instantaneous weight of water at time each 30 seconds using data acquisition (DAQ) system. Weight of feed, all the pressure and temperature, and conductivity of permeate also being recorded into computer using DAQ system and saved as CSV Excel file for further processing. The system is fully automated as all the data is recorded automatically to reduce the error of noting down the values. All the values of pressure, weight, temperature, and conductivity are displayed for the user using Omega digital display to keep track of data by user at any given time. Figure (17) shows the assembled system in the lab with displays and data acquisition system. 65

80 Figure 17: System setup in Qatar University s lab The details of each device and equipment are given in sections below Feed Solution Three different feed solutions to the DCMD system were used in this study. Solutions are: thermal brine, sea water, and synthetic NaCl solution. Details of each feed solution are given below in Table (15). 66

81 Table 15: ph and EC of feed solutions used in the experiments Conductivity Solution Source ph (ms/cm) Thermal Rejected Ras Abu-Fentas (RAF) Thermal Brine Desalination Plant (QEWC) Sea Water Sea Open Intake (Persian Gulf) NaCl Solution Demineralize water+ NaCl (100g in 1 liter) 4.3- Membrane Sheet PTFE membrane filters contain a membrane made of pure PTFE laminated onto a polypropylene non-woven layer is used in experiments. Membranes are purchased from Sterlitech Corporation in US. The main properties of the membrane are tabulated in Table (16). Table 16: Properties of PTFE membrane Description Specification Pore size 0.22 μm Thickness 175 µm thick Pore Size Range 0.2 to 1.0 micron Diameter 13 mm to 142 mm circles, rectangles or rolls Sterilization Auto-clavable up to 130 C The flat sheet membrane is used due to its higher flux compared to hollow fiber sheets [119]. The reported flux from flat sheet membranes in literature is typically L m 2 h 1 [119] at operating temperature of 60 C and 20 C. The polymeric membrane used in the experiment consists of, a thin active layer made of PTFE and a porous support layer made of PP. This structure is able to provide sufficient mechanical strength for the membrane to enable the active layer to be manufactured as thin as possible, which reduces the mass transfer resistance. PTFE membranes are more appropriate for application of membrane distillation, since they have thinner active layer and support layer compared to PVDF and PP membrane. 67

82 The porosity of the membranes used is in the range of 0.7 to 0.75, the pore size is 0.22 μm, and the thickness is 175 µm thick. Thermal conductivity of the PTFE membrane is reported in literature as Wm -1 K -1 [183] and calculated using equation (39) (40) (41) and (43). PTFE is used due to its higher hydrophobicity (largest contact angle with water), good chemical and thermal stability and oxidation resistance, but it has the highest conductivity which will cause greater heat transfer through PTFE membranes [119]. The membrane coupons used in the experimental setup has length of 19.1 cm and width of 14 cm DCMD Module Configuration (MD Block) MD module consists of two Teflon (PTFE) blocks with feed channel and O-ring. The bottom plate is used to carry the hot feed in counter-current flow and top plate is used to carry the permeate and cold flow in similar flow pattern. Figure (19) shows the bottom and top plates. Figure 18: schematic of bottom and top plate A single piece of rectangular membrane coupon is mounted in the Bottom plate on top of the inner O-ring to avoid leakage of the salty water into fresh water flowing at the top plate. Four 68

83 Guide Pins are used in the bottom plate to assure proper alignment of the membrane sheet and proper orientation of the top and bottom plates. The bottom and top plates are assembled together using 4 pieces of Allen screw bolt and nuts. Washers are used on two sides of the plates when fitting the screws to distribute the load due to flexibility of the Teflon blocks. Double Vitton material O-rings are used at two different layers of the bottom plate to provide a leak-proof seal. Two Feed Inlets and two outlets are assembled on bottom and top plates using National Pipe Thread (NPT). NPT is a tapered thread and this type of thread pull tight and therefore make a fluid-tight seal. Feed flows through the feed inlet and pass through manifold made in middle of the plate and pass through the membrane cavity on goes out through other manifold on the other side of the plate and flows out of the plate. The fluid flows tangentially across the Membrane surface due to opposite location of the inlet and outlet which force the fluid to pass all the way from inlet to the other side of the block and continues sweeping over the membrane and collects in the manifold and then flows through the feed outlet and into a tank. The permeates flow through the top plate, in the same manner of the feed flow. The permeate flows to the center of the top plate, goes through the manifold, and then flows out through the Permeate Outlet connection into a user-supplied tank. Spacer is kept in feed manifold in the membrane cavity to enhance turbulence of the feed flow. Figure (20) shows the exact parts on real DCMD cell. 69

84 Figure 19: DCMD bench scale cell and its parts The order of placing the membrane and spacer into the MD blocks are shown in Figure (21). Figure 20: explosion diagram of MD cell with spacer and membrane sheet 70

85 4.5- Material of Construction The MD cell is designed to limit the actual fluid flow to cell body. The cell body is made of Teflon (PTFE) material. Teflon is used due to its high advantages over other polymeric materials such as: Resistant to many chemicals Weather and UV resistance Non stick Outstanding performance at extreme temperatures (withstands temperatures of 260 ºC and cryogenic temperatures of -24 ºC and still has the same chemical properties.) Low coefficient of friction Non-wetting (hydrophobic) Excellent optical properties Very good insulating material (compared to stainless steel) Since MD technology still is not an energy efficient technology, using insulating material to prevent heat lose is essential. The coefficient of the thermal conductivity of PTFE (Teflon) does not vary with the temperature and it is relatively high compared to polycarbonate, PVDF, and PP, so that PTFE can be considered to be a good insulating material. O-rings used for making the cell leak-proof are made of Viton which is a synthetic rubber with diameter of the 2mm and installed in the Bottom plate. Viton has a lower probability of catching fire and withstand very high temperatures (up to 200 ºC). Fittings for inlet and outlet are purchased from Parker USA made of PFA (Perfluoroalkoxy) Teflon nuts and ETFE (Ethylene tetra-fluoro-ethylene) body. These two materials are corrosion resistant and can withstand high temperature (up to 200 ºC) and high pressure (100 psi) Dimensions The overall size of each plate are 233x182.8x30 mm. when assembled together the size of the cell increases to 233x182.8x60 mm. the full dimension of the cells are shown in Figure (22) and Figure (23) for bottom and top plate, respectively. All units are in mm. 71

86 Figure 21: Dimensions of the bottom plate (top view) Figure 22: Dimensions of the top plate (top view) 72

87 The depth of the membrane cavity is 2 mm and depth for O-ring groves also made 2 mm to fit the 2 mm diameter O-rings tight in order to avoid leakage. Figure (24) shows the dimensions in more details looking from side view. Figure 23: Dimensions of top and bottom plate looking from side view The length of the manifold and feed inlet for both feed and permeate side is 132 mm connected to 3/8 FNPT fittings as shown in Figure (25). 73

88 Figure 24: cross section of feed and permeate inlet and outlet looking from side view 4.7- Auxiliary Equipment To operate MD unit different auxiliary equipment are used including pumps, heater, cooler, sensors, balances, and software. Below, function and type of each equipment is given in more details Pump Peristaltic Pumps are used to pump and deliver feed and distillate water to the membrane. The reason behind using Peristaltic Pumps is its accuracy and ease of controlling flow rate, gentle pumping action, and no contact between fluid and pipe interior which both affect the water quality and also clean-up of the pump itself. Another good advantage of peristaltic pumps is zero internal backflow giving accurate dosing without slip The Peristaltic Pumps used in the setup is of Thermo Scientific model: FH100X which has flow capacity of 14 to 4000 ml/min run with rpm of 4 to 400 and has accuracy of ±0.25%. For operation of the pump, Silicone Peristaltic Pump Tubing model BioPharm Silicone GP Tubing, 9.5mm ID, Thermo Fisher, US is used which can tolerate temperature range of -60 C to +232 C. 74

89 Heaters and coolers The temperature of the feed and distillate water is controlled by two refrigerated/ heating circulators Model: F32-MA, Julabo, Germany. The refrigerated/heating bath works in temperature range of -35 to 200 ºC with temperature stability of ±0.02 ºC. for the heating bath due to use of corrosive material (sea water and brine) special coil made of Hastelloy C-276 (Nickel-Molybdenum-Chromium alloy with the addition of Tungsten) is used. Hastelloy is excellent corrosion resistance used in a wide range of corrosive media Temperature and Pressure Measurement Pressure and temperature of the inlet and outlet streams are measured using pressure transducers and thermo resistance RTDs respectively. Pressure are measured using Pressure transducers Model: PX GI, Omega Engineering, UK and can handle temperature of -40 to 85 C and pressure range of 1 to 5 psi with accuracy of ±0.25%. Temperatures are measured using thermo resistance RTDs (3 wires) Model: RTD-NPT-72-E, Omega Engineering, UK which capable of handling pressure up to 2500 psi and maximum temperature of 230 C. both of the sensors are used as NPT fitting to provide leak-free environment while installed online Flow Meter Flow of the hot side will be measured using magmeter model: FMG82, Omega Engineering, UK with NPT fitting to avoid leakage while installed online. Magmeter is a volumetric flow meter which does not have any moving parts. Flow meter is made of HDPE casing and PVDF material for the electrodes which withstand temperature of 0 to 93 C and pressure up to 150 psi. the flow rate range is from 0.03 GPM to 3 GPM max with accuracy of ±1%, ±0.002 GPM of reading across rated range. Distillate production flow rate is measured using liquid flow sensor model: FPR-1506, Omega Engineering, UK made of PTFE and works in flow range of 5 to 5.0 L/min under temperature range of 0 to 70 C and max pressure of 10 psi. The sensor uses Pelton-type turbine wheel to determine the flow rate of the liquid. 75

90 Conductivity Meter The water quality of distillate is measured by conductivity meter with RTD temperature element model: 3433E8A, 10 cell constant, Hatch Company, US. Temperature element is built into conductivity probe tip for fast response to changes in temperature with ±0.1ºC accuracy. The range of operation is from pure water with conductivity of µs/cm or 18.2 MΩ up to saline water of 200,000 µs/cm with accuracy of ± 2 % of reading above 200 μs/cm. it fits online with NPT fitting to avoid leakage and work for flow rates of 0-3 m/s maximum, temperature of -20 to 200 C, and pressure of 6.8 bar at 150 C. The probe is connected to Universal Controller model: LXV , Hatch Company, US Digital Display The data are accessed by digital display meter Model: DP25B-E-230-A, Omega Engineering, UK. This digital meter is able to convert the received voltage of thermocouple, RTDs, pressure transducers and converts it to real Pressure and temperature units. It basically converts analog data to digital data readable by PC Weighing Balance To determine flux of distillate produced and to maintain constant concentration of feed water weighing balances Model: VWR# , Mettler Toledo is used. Weight of distillate produced over active membrane area and time s measured to determine the flux of distillate water and the feed concentration is maintained constant by adding pure water (deionized water) in the feed tank to replace water lost as distillate by weighting the feed tank and observing the difference. These balances can catty weights in range of 820/4200 g with readability of 0.01/0.1 g Data Acquisition System The data was acquired and recorded using a National Instruments data acquisition hardware (Chasis Model cdaq9188; Module Model: NI-9219, National Instruments, US which is 4- channel universal C Series module used for measuring several signals from sensors such as strain gages, RTDs, thermocouples, load cells, and other powered sensors. It can withstand voltage range of -60 to 60 V and current range of A to A. 76

91 Universal C-series is connected to Compact DAQ USB chassis model: NI cdaq-9178 National Instruments, US for measurement system due to its function for analog input, analog output, digital I/O, and counter/timer measurement system. A serial server Model: NI ENET 232, National Instruments, USA which is Ethernet serial interfaces provide additional RS232, RS422, and RS485 ports to computer through standard Ethernet networks is used to acquire the weight from the balance. The data was stored and processed using LAB View data acquisition software Tubing Tygon Silicone Tubing with internal diameter of 3/8" and outside diameter of 1/2" is used to connect auxiliary units together via PTFE fittings from Parker, US. Tube is made of silicone which can withstand temperature range of 80 to 200 C and maximum pressure of 4 psi Experimental Procedure Different experiments are conduced to meet the objective of the study. Experiments are designed in a way to measure the water quality and energy efficiency analysis of the system. Experiments details are given below Water Quality and Energy Efficiency Experiments First group of experiments were conducted with identical pressures and velocities (flow rate) of the feed and permeate streams. Pressure and velocity is controlled by adjusting pump rpm. Different feed with different conductivity are used to study the quality of the permeate. Sea water, thermal rejected brine, and synthetic NaCl solution at ppm NaCl concentration are used as feeds used in the experiments. Table (17) shows parameters of the experiments. 77

92 Table 17: Parameters used for experiments of DCMD system Parameter Range Pressure (atm) 1 Feed/ Permeate Flow rate (L/min) 1.5 Hot side temperature (ºC) Cold side temperature (ºC) Sea water 65 Feed conductivity Thermal brine 76.8 (ms/cm) NaCl solution Taking closer look to the experiments above number of experiments are summarized in Table (18), below. Table 18: number of experiments conducted with different feed solution and temperatures Feed Feed at 70 ºC Feed at 65 ºC Feed at 60 ºC Thermal Brine Sea water NaCl solution Another set of experiments with thermal brine and constant temperature of 70 ºC for hot side and 30 ºC for cold side is conducted to study the effect of flow rate on energy efficiency. In these experiments, the pressure, temperature and feed solutions are kept constant. Table (19) shows the parameters for these experiments. 78

93 Table 19: Flow rate ranges for DCMD tests Experiment Parameter Feed/Permeate Flow rate (L/min) 2.5 Temperature: ºC 2.0 Pressure: 1 atm 1.5 Feed type: thermal brine Contact Angle measurement Widely used methods of contact angle measurement are the sessile drop method (including captive bubble method), the Wilhelmy plate method, and the capillary rise method. Among mentioned methods, the sessile drop method is the widely used ones in the most of the literature. A contact angle analyzer (KRÜSS, DSA30, Germany) was used. Figure (25) shows a schematic diagram of the method used. Figure 25: Contact angle measuring device. Adapted from Park, Byung Hyun, Myong-Hwa Lee, Sang Bum Kim, and Young Min Jo. "Evaluation Of The Surface Properties Of PTFE Foam Coating Filter Media Using XPS And Contact Angle Measurements." Applied Surface Science (2011):

94 In sessile drop a liquid droplet is deposited onto the membrane surface by using an I -shaped needle. DI water was used as a liquid for both the methods. The contact angle is calculated from a digital video image of the drop on the membrane using image-processing software, which allowed the estimation of the contact angle from the circle fitting of the drop using the sessile drop method. Theoretically, contact angles measured for sessile drop method are described by the Young equation, assuming that surface is smooth and homogeneous [184]. Membrane sample preparation is an important step for contact angle measurement. To start, the membrane sample was kept dry at 40 C in the vacuum oven. Later the measurement starts as follow: a dangling droplet of 6 μl of DI water at the end of needle is cautiously deposited to membrane surface to avoid the effect of falling force by gravity. After this moment, all processes have to be recorded for 500 frames (approximately 16 s) at least. Contact angle values, calculated by DSA100 software (KRÜSS, DSA30, Germany), is determined as the averaged values during the periods of frames (about s) due to error, mainly because of evaporation of the liquid. To reduce the error and to guarantee a statistical reliability, multiple measurements were taken and averaged SEM Imaging Scanning electron microscopy (SEM) is a microscopic method capable of producing very high magnification images of a membrane surface. Due to the manner in which the image is created, SEM images have a characteristic three-dimensional appearance and are useful for judging the surface structure of the sample. To observe the membrane cross sections, membrane pores, bonding, and membrane backing support, the SEM imaging is undertaken in Central Laboratory Unit (CLU)-Qatar University. FEI Quanta 200 Environmental Scanning Electron Microscope (ESEM) with a resolution of 5 nm and a magnification X200K used for this purpose. To get the images and study the morphology of the membranes, sample of membrane was frozen in liquid nitrogen and then fractured. Cross section and surface of the membrane were sputtered with gold and then transferred to the microscope for imaging. The imaging process is done by CLU-QU labs. 80

95 Chapter 5: Results and Discussion This chapter covers the results and discussion of experiments designed to investigate different aspects of direct contact membrane distillation (DCMD) process applied to water desalination. Results are divided into different sub-sections including: 1. Contact angle measurement 2. SEM images of membrane to study the membrane pores 3. Reproducibility of DCMD bench scale unit 4. Water Quality Tests 5. Energy Efficiency Analysis The key parameters measured in the last two categories were the flux, quality of water passing through the membrane pores, and energy efficiency of the DCMD system Contact Angle Measurement Contact angle of polymer is commonly used to estimate the hydrophobicity and wetting properties of polymer surface. As mentioned earlier, larger contact angle represents hydrophobic surface while smaller angle represents hydrophilic surface. A Contact Angle Measuring Instrument DSA30 from KRUSS GmbH was used to measure the contact angle of a PTFE membrane using the sessile drop method as follows: deposition of a liquid droplet onto the membrane surface using an I-shaped needle and DI water was used as a liquid and the angle of the drop with the membrane is measured using the Young equation, assuming that surface is smooth and homogeneous [185]. Five readings were measured and an average was obtained from the results. Figure (27) and (28) shows the equilibrium state of a distilled water droplet on a flat-sheet membrane for PTFE and PP membrane, respectively. 81

96 Figure 26: equilibrium state of a distilled water droplet on a flat-sheet membrane active layer (PTFE) Figure 27: equilibrium state of a distilled water droplet on a flat-sheet membrane support layer (PP) 82

97 Contact angle of both sides are measured and reported in Table (20) for PTFE layer (active side) and Table (21) shows the contact angle for PP layer (support side). Table 20: Contact angle measurements for PTFE membrane Run ϴ (Left) [deg] ϴ (Right) [deg] ϴ (Average) [deg] ± ± ± ± ± 5.05 Table 21: Contact angle measurements for PP membrane Run ϴ (Left) [deg] ϴ (Right) [deg] ϴ (Average) [deg] ± ± ± ± ± 2.93 As shown in Table (20) and (21), the high contact angles obtained can be attributed to the high hydrophobicity of the membrane. The hydrophobic nature of membrane permit only vapor to pass and rejects water, ensuring high selectivity in the process of MD. It can also be seen that both sides of the membrane are hydrophobic since the contact angle of both sides are higher than 90 degree but hydrophobicity of PTFE side is higher than that of PP side. Contact angles measured are in consistent with the values reported in literature [160]. Higher contact angle in combination with other factors such as smaller pore size, lower surface energy and higher surface tension lead to higher liquid entry pressure be greater than the pressure difference at the membrane s liquid/vapor interface to prevent pore wetting. Pore wetting lead to penetration of liquid water and affect the quality of fresh water produced. Also, low contact angle leads to reduction in ability of membrane to reject non-volatile feed [119]. According to literature [119], the flux of MD was found to be dependent with the surface contact angle. Membranes with lower surface energy (lower contact angle) have less tendency for pore wetting. Moreover, as the membrane hydrophobicity increases, thermal conductivity of the membrane decreases [119]. This is desirable in DCMD operation, since it reduces the heat losses 83

98 by conduction across the membrane and avoids the establishment of strong heat polarization layers on the membrane interface and in the membrane pores [119]. Similar tests were conducted in literature and similarly showed the same result. Hwang et al. [88] conducted the same measurement on PTFE commercial membranes and value of 122º was reported. Another study by Zhang et al. [160] using PTFE and PVDF membranes indicated that PTFE membranes can have contact angle up to 140º and average of 126º is commonly reported in most of literature. Zhu et al. [186] reported that the contact angle can be affected by the membrane surface composition, pore size and roughness. According to their studies, contact angle of PTFE hollow fibers membrane is in range of 125º which is almost to the results found in these report. Wenzel [187] considered analytical relationship between roughness and contact angle and concluded that for hydrophobic surfaces, the contact angle increases with increasing roughness according to the Wenzel equation: (50) Where is the measured contact angle and is what the contact angle would be if the surface were flat. and denote the effective and projected (flat) surface area, respectively. According to his study, contact angle of 122 º were reported for PTFE membrane. Adnan et al [188] tested different PTFE membrane from different source and contact angle of 126 º up to 165 º were reported in their studies. 84

99 5.2- SEM Characterization To observe the membrane cross sections, membrane pores, bonding, and membrane backing support, the SEM imaging is undertaken in Central laboratory Unit-Qatar University. FEI Quanta 200 Environmental Scanning Electron Microscope (ESEM) with a resolution of 5 nm and a magnification X200K used for this purpose. To get the images and study the morphology of the membranes, sample of membrane was frozen in liquid nitrogen and then fractured. Cross section and surface of the membrane were sputtered with gold and then transferred to the microscope for imaging. As shown in Figure (28) membrane active side (PTFE) is bonded to the support layer (PP) using thermal bonding and there is no glue or adhesive used to attach the two layers. This can be good as temperature wont effect the membrane bonding and quality of water is not affected. Figure 28: SEM picture of PTFE membrane and the thermal bonding of active layer to support layer with 8000 time magnification (left) and 2000 time magnification (right) Another SEM images were one to observe the surface of active layer which is made of PTFE and the support layer which is made of PP. As shown in Figure (29) and Figure (30) it is clear that the active layer has more pores compared to the support layer and the pore size are bigger in PP side compared to active layer. 85

100 Figure 29: SEM image of support layer of membrane (PP) with 200 (left) and 500 (right) time magnification Figure 30: SEM image of active layer of membrane (PTFE) with (left) and (right) time magnification To check the pore size difference of active layer and support layer another set of images as shown in Figure (31) and Figure (32) is taken. As can be seen in Figure (32), the pore sizes are 86

101 smaller in active layer but it has more porous area compared to the support layer in Figure (32). This can be helpful to have bigger pore size in support layer for ease of vapor transport. Figure 31: SEM image of PTFE side of membrane and pore size and pore size distribution of the active layer of membrane Figure 32: SEM image of PP side of membrane and pore size and pore size distribution of the support layer of membrane 87

102 Percentage (%) < > 0.22 Pore Size Distribution (μm) Figure 33: Pore size distribution of PTFE membrane (Active layer) This result of pore size distribution in Figure (33) shows a very narrow pore size distribution, with a mean pore size of 0.22 μm. Sample of 225 µm random membrane sample were analyzed for data in Figure (33) and the maximum pore diameter obtained from the SEM imaging was 0.96 μm with only 13 % of total pores. Figure (33) indicates that the sharp pore size distribution reduces the potential water leakage and pore wetting through the membrane and is in constant with results in study done by Hwang et al. [88]. According to literature [119] the flux is reported to be essentially increasing with the average pore size of the membrane. Pores of commercial PTFE and PVDF membranes most often falls within a 200 to 400 nm range as also proven in the SEM images in Figure (32). It might be seen in some literature that the larger pore size membranes exhibit lower flux can be related to other membrane parameters such as their larger thickness. A study by Adnan et al. [188] indicates the supported membranes give lower flux in comparison to the non-supported membrane due blockage resulted by some of the pores of membrane, and hence reduction of the porosity. Their study showed a 50% reduction in porosity of in a scrim backing membrane using SEM images. In addition to porosity, they found that the thickness could be another factor that contributes to reduction in flux. In addition, support materials absorb some heat supplied by the feed, thus increasing temperature polarization and hence reducing the flux. 88

103 Not only pore size but porosity also seems to be critical and membranes with porosity higher than 80% show the best performance [119]. As can be seen from Figure (32) PTFE side of the membrane has higher porosity compared to PP side and this will help more transport of vapor to the other side. Some of the very highest porosity materials tested in MD literature [advance] showed much lower flux compared to the less porous materials due to effect of other structural parameters such as pore size, or thermal conductivity. The presence of membrane support material affects the surface temperatures of the membrane (Equations (12) and (13)). Of all the mechanisms, only the Knudsen diffusion is affected by the average membrane temperature. Due to effect on membrane surface temperature, the membrane heat conduction is affected by the surface temperatures and hence the energy efficiency and flux. Due to the high porosity of the membrane, the effective thermal conductivity of the membrane is not considerably affected by the difference in the thermal conductivities of the membrane and support materials since the most of the pores are filled with air and water vapor [185]. Thickness of membrane also was studied through SEM images and shown in Figure (34). Figure 34: SEM images of membrane thickness Analyzing the thickness of membrane is the most difficult parameters to evaluate experimentally. The thickness measurements are typically performed either with micrometers or through scanning electron microscopy image analysis. These techniques calculate the membrane thickness over small surface areas and as a result they are not always representative of the whole 89