University of Southern California Dissertations and Theses

Thermally driven water treatment with membrane distillation: membrane performance, waste heat integration, and cooling analysis

(USC Thesis Other)

Thermally driven water treatment with membrane distillation: membrane performance, waste heat integration, and cooling analysis

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Thermally driven water treatment with membrane
distillation: Membrane performance, waste heat
integration, and cooling analysis

A dissertation submitted in partial satisfaction of the requirements
for the degree of Doctor of Philosophy in
Environmental Engineering

Submitted by
Ryan D. Gustafson
May, 2019

Degree approved by the faculty of the
University of Southern California Graduate School

a
Amy E. Childress, Ph.D. (Ph.D. advisor)
a
Kelly T. Sanders, Ph.D.
a
Adam L. Smith, Ph.D.
a
Daniel McCurry, Ph.D.
b
Mohammed Beshir, Ph.D.

a
Viterbi School of Engineering; Sonny Astani Department of Civil and Environmental Engineering
b
Viterbi School of Engineering; Ming Hsieh Department of Electrical and Computer Engineering
i

Table of Contents
List of Tables ................................................................................................................................................ iii
List of Figures ............................................................................................................................................... iii
Acknowledgements ...................................................................................................................................... vi
Support ....................................................................................................................................................... vii
Chapter 1 Introduction .......................................................................................................................... 1
1.1 Background ................................................................................................................................... 1
1.1.1 Water, energy, and membrane distillation (MD) .................................................................. 1
1.1.2 MD Overview ........................................................................................................................ 1
1.1.3 Membrane morphology and operating conditions............................................................... 5
1.1.4 MD and Waste Heat .............................................................................................................. 6
1.1.5 MD and Cooling ..................................................................................................................... 7
1.2 Objectives and scope of work ....................................................................................................... 7
1.3 Dissertation organization .............................................................................................................. 9
Chapter 2 Morphological changes and creep recovery behavior of expanded
polytetrafluoroethylene (ePTFE) membrane used for membrane distillation ........................................... 10
2.0 Abstract ....................................................................................................................................... 10
2.1 Introduction ................................................................................................................................ 11
2.2 Materials and methods ............................................................................................................... 16
2.2.1 Membrane .......................................................................................................................... 16
2.2.2 MD experiments.................................................................................................................. 16
2.2.3 Membrane characterization ............................................................................................... 18
2.2.4 Creep recovery tests ........................................................................................................... 19
2.3 Results and discussion ................................................................................................................ 20
2.3.1 Water flux during MD experiments .................................................................................... 20
2.3.2 Morphological changes during MD experiments ................................................................ 21
2.3.3 Time- and temperature-dependent deformation during creep and recovery ................... 29
2.4 Conclusions ................................................................................................................................. 31
Chapter 3 Membrane distillation driven by intermittent and variable-temperature waste heat:
System arrangements for water production and heat storage .................................................................. 33
3.0 Abstract ....................................................................................................................................... 33
3.1 Introduction ................................................................................................................................ 34
3.2 Materials and methods ............................................................................................................... 39
ii

3.2.1 Modeling ............................................................................................................................. 39
3.2.2 Experimental ....................................................................................................................... 45
3.3 Results and discussion ................................................................................................................ 48
3.3.1 Modeling analyses of system arrangements ...................................................................... 48
3.3.2 Experimental analyses of system arrangements ................................................................ 57
3.4 Conclusions ................................................................................................................................. 62
Chapter 4 A critical review of cooling in membrane distillation ......................................................... 63
4.0 Abstract ....................................................................................................................................... 63
4.1 Introduction ................................................................................................................................ 63
4.2 MD system features that affect energy consumption ................................................................ 65
4.2.1 Heat recovery ...................................................................................................................... 65
4.2.2 Feed stream configuration .................................................................................................. 68
4.3 Energy metrics in MD .................................................................................................................. 70
4.4 Cooling approaches used in existing MD systems ...................................................................... 73
4.4.1 Replenishment cooling ........................................................................................................ 73
4.4.2 Auxiliary Cooling.................................................................................................................. 84
4.5 Factors affecting choice of cooling approach ........................................................................... 110
4.5.1 Cooling water availability .................................................................................................. 112
4.5.2 Regulatory environment ................................................................................................... 113
4.5.3 Water recovery and costs ................................................................................................. 113
4.6 Conclusions and future research needs .................................................................................... 115
Chapter 5 Conclusions ....................................................................................................................... 118
5.1 Research Synopsis ..................................................................................................................... 118
5.1.1 Summary of morphological changes and creep recovery in ePTFE MD membranes ....... 118
5.1.2 Summary of integration of MD with variable-temperature and intermittent waste heat
119
5.1.3 Summary of a critical review of cooling in MD ................................................................. 120
Chapter 6 References ........................................................................................................................ 121

iii

List of Tables
Table 2.1: Experimental matrix of long-term MD experiments ................................................... 18
Table 2.2: Surface porosity of virgin and used membranes ......................................................... 25
Table 2.3: Bulk porosity of virgin and used membranes .............................................................. 29
Table 3.1: Water flux variability .................................................................................................... 56
Table 3.2. Average temperatures and water flux ......................................................................... 60
Table 4.1: Summary of pilot-scale MD systems using only replenishment cooling ..................... 75
Table 4.2: Summary of pilot-scale MD systems using auxiliary cooling ....................................... 85
Table 4.3: Advantages and disadvantages of replenishment and auxiliary cooling ................... 111
Table 4.4: Advantages and disadvantages of different wet cooling approaches. ...................... 112

List of Figures
Figure 1.1: Classical membrane distillation (MD) configurations ................................................... 3
Figure 2.1: Creep recovery curves ................................................................................................ 15
Figure 2.2: Schematic of bench-scale DCMD system. ................................................................... 17
Figure 2.3: Water flux over time ................................................................................................... 21
Figure 2.4: Energy dispersive x-ray spectroscopy (EDS) spectra .................................................. 22
Figure 2.5: Scanning electron microscopy (SEM) images ............................................................. 23
Figure 2.6: Atomic force microscopy (AFM) images ..................................................................... 27
Figure 2.7: Cross-sectional scanning electron microscope (SEM) images .................................... 28
Figure 2.8: Creep recovery curves for virgin membrane .............................................................. 30
Figure 3.1: Temperature profiles from example low-grade waste heat ...................................... 36
iv

Figure 3.2: Direct and indirect system arrangements. ................................................................. 38
Figure 3.3. Schematic of bench-scale DCMD system .................................................................... 46
Figure 3.4. Model results for steady-state water flux and feed inlet temperature ..................... 50
Figure 3.5. Model results for steady-state feed tank temperature and rate of heat storage ..... 51
Figure 3.6. Model results for water flux and total product water w/ different intermittencies . 54
Figure 3.7: Model results for specific thermal energy consumption ........................................... 57
Fig. 3.8. Experimental results for water flux and total product water at 90% intermittency ...... 58
Fig. 3.9: Feed temperatures and water flux versus time .............................................................. 59
Figure 4.1: External heat recovery ................................................................................................ 66
Figure 4.2: Internal heat recovery ................................................................................................ 67
Figure 4.3: Single-pass and recirculating feed stream configurations ......................................... 69
Figure 4.4: Replenishment cooling ............................................................................................... 74
Figure 4.5: Diagram reproduced from Banat et al. showing replenishment cooling ................... 76
Figure 4.6: Diagram reproduced from Hagedorn et al. showing replenishment cooling............. 77
Figure 4.7: Diagram reproduced from Andrés-Mañas et al. showing replenishment cooling ..... 78
Figure 4.8: Diagram reproduced from Raluy et al. showing replenishment cooling .................... 81
Figure 4.9: Diagram reproduced from Hagedorn et al. showing replenishment cooling............. 83
Figure 4.10: Diagram of once-through cooling process. .............................................................. 86
Figure 4.11: Diagram reproduced from Zhao et al. showing once-through cooling .................... 88
Figure 4.12: Diagram reproduced from Dow et al. showing once-through cooling ..................... 91
Figure 4.13: Diagram of open-circuit wet cooling ........................................................................ 93
Figure 4.14: Diagram reproduced from Schwantes et al. showing open-circuit wet cooling ...... 95
v

Figure 4.15: Diagram of conventional closed-circuit wet cooling ................................................ 97
Figure 4.16: Diagram reproduced from Kabeel et al. showing closed-circuit wet cooling .......... 98
Figure 4.17: Diagram reproduced from Schwantes et al. showing closed-circuit wet cooling .. 100
Figure 4.18: Diagram reproduced from Cipollina et al. showing closed-circuit wet cooling ..... 101
Figure 4.19: Diagram reproduced from Wang et al. showing closed-circuit wet cooling .......... 103
Figure 4.20: Diagram reproduced from Wang et al. showing closed-circuit wet cooling .......... 103
Figure 4.21: Dry cooling process diagram using mechanical draft. ............................................ 105
Figure 4.22: Diagram reproduced from Guillén-Burrieza et al. showing dry cooling ................. 106
Figure 4.23: Diagram reproduced from Guillén-Burrieza et al. showing dry cooling ................. 108
Figure 4.24: Diagram reproduced from Dow et al. showing dry cooling ................................... 110

vi

Acknowledgements
I would like to thank my wife, Alexandra Gustafson, for her love and support throughout the
undertaking and performance of the research in this dissertation. She has helped me grow
every day and I wouldn’t be the person I am today without her. I would like to thank my parents
and older brothers, whom have always provided encouragement and support for me from the
very beginning. I would also like to thank my best friend, Oscar Aguilar, whom has been like a
brother to me through the years.
I want to thank my advisor, Dr. Amy Childress, for helping me grow as a researcher over the
these past few years. I received my initial training in membrane-based water research at her old
lab in Reno as an undergraduate student and have enjoyed working with her to build her new
lab at USC. I would also like to thank my undergraduate research advisor, Dr. Andrea Achilli, for
introducing me to the world of membrane-based water treatment and for preparing me for
graduate-level research. I also want to thank Dr. Sage Hiibel for helping me grow as a
researcher during my short time at the University of Nevada, Reno.
I would like to extend gratitude to my dissertation committee – Drs. Amy Childress, Adam
Smith, Kelly Sanders, Daniel McCurry, and Mohammed Beshir – for their valuable feedback
during my qualifying exam and dissertation defense. I also want to thank the other Ph.D.
students and research assistants in the Childress Research Group that I have worked closely
with during my time at USC: thank you to Dr. Chistopher Morrow, Allyson McGaughey, Lauren
Crawford, Samantha McVety, and Weijian Ding.
vii

Support
This work was supported by the Strategic Environmental Research and Development Program
(SERDP Project number ER-2237), the US Environmental Protection Agency Science to Achieve
Results (US EPA STAR Grant #83486701) program, a Viterbi Graduate School Ph.D. Fellowship,
and a National Water Research Institute and Southern California Salinity Coalition Fellowship.
Support was also provided through assistance by staff at USC facilities including the California
Core Center for Excellence in Nano Imaging (CNI), Center for Excellence in NanoBioPhysics, and
M.C. Gill Composites Center.

viii

“In every walk with Nature one receives far more than he seeks.”
― John Muir

1

Chapter 1 Introduction
1.1 Background
1.1.1 Water, energy, and membrane distillation (MD)
Growing populations have increased stress on water resources and climate change could
exacerbate water stress in some regions in the future [1, 2]. One way to address this problem is
to integrate water from non-traditional sources by embracing seawater desalination and
wastewater reuse technologies to help diversify a region’s water portfolio [2]. However, the
technologies that are used to treat these non-traditional water resources typically consume
large amounts of electrical energy, which could exacerbate problems related to climate change
[3, 4]. Membrane distillation (MD) is one technology that is poised to address these concerns
because it provides nearly 100% rejection of non-volatile contaminants (e.g., dissolved salts) [5]
and operates using minimal electrical energy by using low-grade thermal energy [6-8].
1.1.2 MD Overview
In MD, heated feed water (e.g., seawater or brackish water) is passed along one side of a
microporous hydrophobic membrane at ambient pressure, water is transported across the
membrane in the vapor phase, and the vapor condenses on a cool solid or liquid surface on the
other side of the membrane. This vapor-phase water transport across the membrane is in
contrast to the liquid-phase water transport that occurs in conventional pressure-driven
membrane processes (e.g., reverse osmosis, nanofiltration, ultrafiltration, and microfiltration).
MD is referred to as a thermally driven process because the vapor pressure driving force in MD
is mostly determined by the solution temperature. Higher salinity can reduce the vapor
2

pressure driving force but the effect is minimal compared to the driving force reduction at high
salinity in other membrane-based water treatment processes like reverse osmosis [9]. The
phase-change separation of MD results in very high rejection of non-volatile contaminants [5]
and operation at low pressure allows MD systems to be constructed largely of low-cost plastic
system components [10]. The characteristically high rejection of non-volatile feed contaminants
in MD and the minimal reduction in driving force at high salinity has resulted in significant
interest in MD for seawater and brackish water desalination [11-17] and treatment of
hypersaline solutions [18].
MD systems are characterized partly by their MD configuration, and the classical configurations
are direct contact MD (DCMD), air gap MD (AGMD), sweeping gas MD (SGMD), and vacuum MD
(VMD) (Fig. 1.1) [10, 19]. In each MD configuration, the feed solution (e.g., brackish water or
seawater) is circulated on one side of the membrane, but each configuration differs in the
design of the module on the condensation side of the membrane. In DCMD (Fig. 1.1a), a cool
pure water solution is circulated on the condensation side of the membrane, vapor condenses
upon contact with the pure water, and product water is collected as overflow from the
condensation-side tank or reservoir. In AGMD (Fig. 1.1b), an air gap exists on the condensation
side of the membrane and the water vapor that travels across the air gap is condensed on a
cool surface (i.e., condensation foil) within the module and collected as product water. In
SGMD (Fig. 1.1c), an inert sweeping gas is passed along the condensation side of the
membrane, the water vapor gathered by the sweeping gas is condensed on a cool surface
external to the membrane module, and the condensate is collected as product water. In VMD
(Fig. 1.1d), vacuum pressure is applied on the condensation side of the membrane, the water
3

vapor is condensed on a cool surface external to the membrane module, and the non-
condensable gases that are remaining are discharged to the surroundings.

Figure 1.1: Classical membrane distillation (MD) configurations: (a) direct contact MD (DCMD),
(b) air gap MD (AGMD), (c) sweeping gas MD (SGMD), and (d) vacuum MD (VMD).
DCMD is the most studied MD configuration [20], due to its simplicity of construction and high
water flux compared to other configurations [10, 19-21]. The drawback of DCMD is its high
4

conductive heat loss [10, 19-21], due to the direct contact of the condensation side of the
membrane with the cool distillate solution. VMD is the second most studied MD configuration
[20] and can provide higher water flux than DCMD at the same feed temperature because of
lower conductive heat losses and the ability to reduce the condensation-side vapor pressure to
almost zero by applying a strong vacuum [19]. The drawbacks of VMD include the required use
of an external condenser and higher risk of pore wetting [19-21]. AGMD is the third most
studied MD configuration [20] and is known for its high thermal energy efficiency [19-21]. The
drawback of AGMD is lower water flux than DCMD or VMD due to the higher mass transfer
resistance of the stagnant air in the gap [10, 19, 21]. SGMD is the least studied configuration
[20]. The drawbacks of SGMD include the required use of an external condenser [10, 19-21] and
the large condenser surface area required for condensation due to the large sweeping gas
volume that contains the water vapor [10]. Although DCMD is by far the most studied
configuration, most DCMD studies occur at bench- or lab-scale, while most pilot systems use
the AGMD configuration due to its low energy consumption [19, 20].
Several additional MD configurations have been developed more recently, including permeate
gap MD (PGMD, also known as water gap or liquid gap MD), material gap MD (MGMD),
conductive gap MD (CGMD), and vacuum AGMD (V-AGMD). The PGMD design is almost
identical to that of AGMD, except that the air gap is filled with water, resulting in lower mass
transfer resistance compared to an air-filled gap [22-24]. Francis et al. [25] tested different
materials in the gap under the name of MGMD – including polyurethane sponge, polypropylene
mesh, and sand – and found that a water-filled gap (i.e., PGMD) provided the highest water flux
of the materials tested. Conductive gap MD (CGMD) can also be thought of as a type of MGMD
5

in which the gap is filled with water and a material with a higher conductivity than water [22].
Swaminathan et al. [22] have evaluated CGMD with different types of metal mesh materials in
the gap, finding increasing water flux with increasing material conductivity. In V-AGMD, vacuum
pressure is applied to the air gap to remove non-condensable gases, reducing the resistance to
mass transfer in the gap to allow for higher water flux [26, 27]. In contrast to VMD, the vacuum
applied in V-AGMD is not low enough to control the condensation-side vapor pressure, so the
driving force remains thermal in nature [26].
1.1.3 Membrane morphology and operating conditions
While a number of different hydrophobic polymeric membrane materials can be used for MD
[28], expanded polytetrafluoroethylene (ePTFE) membranes are the most commonly used
membrane material in commercially available MD systems [21]. The surface mictrostructure of
ePTFE membranes is characterized by a network of polymer strands known as fibrils
interconnected by polymer nodes [29, 30]. Changes in surface microstructure of membranes
after use in MD systems has been previously observed alongside changes in system
performance (i.e., water flux) [31-34], but these changes were always observed in the presence
of foulants and/or scalants – leaving the contribution of changes in membrane surface
microstructure to changes in performance unquantified.
ePTFE membranes have also been observed to decrease in thickness when exposed to
pressures and temperatures within the operating range of MD [35-38], but the permanence of
these changes was not evaluated. Some MD researchers [34, 39-42] have suggested that creep
– increasing deformation under constant applied stress – may be the cause of changes in
membrane morphology (i.e., surface microstructure and thickness) [43]. If creep occurs for a
6

prolonged period of time, then the material will recover a portion of the deformation once the
stress is released, but permanent deformation could remain. No studies currently exist that
analyze the potential link between changes in membrane morphology and creep. The severity
of changes in membrane morphology has been shown to be affected by temperature [44], but
the permanence of these changes has not been evaluated.
1.1.4 MD and Waste Heat
MD is able to be operated with low electrical energy consumption because it can be driven by
thermal energy from low-grade waste heat sources in the power and industry sectors [6, 8, 39,
45-53]. Power sector sources of low-grade waste heat include flue gas and cooling water from
thermoelectric power plants [54, 55], while industrial sources include kiln gasses and cooling
streams in the cement and ceramic industries [56-60]. While many modeling and experimental
studies of waste-heat-driven MD can be found in the literature [6, 8, 39, 45-53], none of these
studies consider the variability or intermittency that is inherent to many waste heat sources.
Consideration of variability is important because it can affect the controllability and
predictability of product water availability of the MD system. Consideration of intermittency is
important because it can result in periodic loss of the thermal driving force required for MD
system operation. Integration of heat storage into MD systems would be required for
continuous operation when using an intermittent waste heat source, but role of heat storage in
MD systems driven by intermittent or variable-temperature waste heat sources has not been
assessed in the literature. Different system arrangements, based partly on the location of the
heat exchanger used to bring waste heat into the MD system, could serve as useful tools for
overcoming issues associated with integration of variable-temperature or intermittent waste
7

heat sources. However, comparisons between different MD system arrangements have not
been performed in the unique context of variable-temperature or intermittent waste heat.
1.1.5 MD and Cooling
While numerous studies of MD specifically focus on the energy required for heating [6, 8, 39,
45-53], none explicitly focus on the energy needs for cooling. Many bench and pilot-scale
studies of MD use electrical chillers or heat pumps (e.g., [61-65]) but these approaches would
not be implemented at full-scale due to their high electrical energy consumption. A number of
pilot-scale studies have been operated using realistic cooling approaches (e.g., wet or dry
cooling) [6, 8, 45, 47, 48, 51, 66-72], but few of these studies provide detailed information
regarding the performance of the cooling system. While the contribution of heat recovery
techniques used in some MD systems has been briefly noted in the literature to contribute to
cooling [48], the cooling functionality provided by heat recovery has not been thoroughly
assessed. While the effect of feed stream configuration on water flux and heating energy needs
has been previously addressed [73, 74], the effects of feed stream configuration on cooling
currently remains unaddressed.
1.2 Objectives and scope of work
The main objectives of this work were to better understand how MD’s unique operating
conditions can affect system performance through changes in membrane morphology over
time and to further develop the knowledgebase for effective integration of thermal energy
sources into MD systems. To investigate the relationship between changes in membrane
morphology and MD system performance over time, continuous long-term DCMD experiments
8

and short-term thermomechanical analyses were performed using a commercially available
ePTFE membrane. Three 30-day DCMD experiments were performed with different feed
temperature and salinity combinations, and changes in surface microstructure, porosity, and
thickness were compared between these experiments. Thermomechanical analyses consisted
of creep recovery tests performed at different temperatures to assess the ability of the
membrane to deform over time under low and constant applied stress and the contribution of
temperature to the extent of deformation that occurs. To investigate effective integration of
heating energy into MD systems, experimental and modeling analyses of waste-heat-driven
DCMD were performed. DCMD experiments and modeling were performed using simulated
intermittent waste heat and the ability of heat storage and different system arrangements to
improve system performance was assessed under these conditions. Experiments using
simulated variable-temperature waste heat were performed to assess how heat storage and
system arrangements can be used to improve performance when the waste heat source is
always on but variable in temperature. The effective integration of cooling into MD systems
was investigated by performing a critical review of cooling in the MD literature. This review
included development of a new linguistic framework for cooling approaches used in MD, critical
assessment of energy consumption metrics and experience with different cooling approaches in
existing pilot-scale MD systems, a review of the factors that affect the choice of cooling
approach for an MD system, and identification of future research opportunities regarding
cooling in MD. Pursuit of the objectives of this dissertation has helped to address important
issues associated with MD system operation outside of the laboratory, helping push the
technology closer to application in full-scale water treatment systems.
9

1.3 Dissertation organization
This dissertation is a compilation of papers written over the course of the dissertation research.
Chapter 2 is an entire paper that has been submitted for publication in the Journal of
Membrane Science; this research builds upon research previously published in the Journal of
Membrane Science [34] that I was a co-author on. Chapter 3 is an entire paper that has been
published in the journal Desalination [7]. Chapter 4 consists of an entire paper that has been
approved for submittal to the journal Applied Energy and is in the final stages of preparation.

10

Chapter 2 Morphological changes and creep recovery behavior of
expanded polytetrafluoroethylene (ePTFE) membrane used for
membrane distillation
2.0 Abstract
Existing studies of membrane distillation (MD) have shown changes in membrane surface
microstructure along with decreased system performance during long-term MD system
operation, however, changes in feed-side morphology during long-term operation have not
been demonstrated in the absence of foulants. Also, the net effect of morphological changes
alone on performance during long-term operation has not been determined. Three 30-day
direct contact MD experiments were run at different feed temperatures and salinities with an
expanded polytetrafluoroethylene (ePTFE) membrane. Results showed that changes in
membrane morphology (surface morphology, thickness, and bulk porosity) in the absence of
foulants can result in decreasing water flux over time. Higher temperature (65 °C) experiments
showed greater membrane morphology changes compared to lower temperature (45 °C)
experiments. Creep recovery tests performed at different temperatures demonstrated
membrane deformation over time under low and constant stress with the potential for
permanent deformation after stress removal; these tests also demonstrated that higher
temperatures result in greater deformation. Creep recovery test results were shown to provide
useful metrics for comparing the influence of time and temperature on changes in morphology
between different membranes, without the need for running long-term MD experiments to
assess this behavior.
11

2.1 Introduction
In direct contact membrane distillation (DCMD), a warm feed solution is circulated on the feed
side of a hydrophobic, microporous membrane, while cooler pure water flows on the distillate
side of the membrane. Energy for heating the feed solution is typically provided by low-grade
waste heat [6, 7, 39, 49] or solar energy [16, 17, 66, 72, 74, 75] to minimize electrical energy
use. The temperature difference across the membrane is the primary determinant of the vapor
pressure driving force, while feed solution salinity slightly reduces the vapor pressure driving
force. Even with very high feed solution salinity, the vapor pressure driving force is only
minimally diminished [18, 76]. This, combined with MD’s high rejection of non-volatile
contaminants [5], has led to applications of membrane distillation (MD) in the treatment of
wastewater [77], brackish water [14], seawater [11, 75], reverse osmosis brines [12, 78], and
hypersaline solutions [18].
Expanded polytetrafluoroethylene (ePTFE) membranes are the most commonly used
membrane in commercially available MD systems [21]. These membranes are the most
hydrophobic of the polymeric membranes used in MD [19, 28] and they are highly resistant to
chemical degradation [30]. ePTFE membranes are also desirable for use in MD systems because
of their wide availability due to their use in a multitude of applications outside of water
treatment [30]. The surface microstructure of ePTFE membranes is characterized by a unique
network of polymer strands referred to as “fibrils”, which meet at polymer “nodes” throughout
the material [29, 30].
12

Researchers have documented changes in membrane surface microstructure along with
decreased system performance during long-term MD system operation with different
membrane materials [31-34]; however, the contribution of changes in surface microstructure to
changes in performance were obscured by the presence of foulants and/or scalants. For
example, Guillen-Burrieza et al. [31] assessed the impacts of membrane cleaning procedures on
ePTFE membranes and compared surface microstructure between virgin and used/cleaned
membrane samples. Results showed changes in membrane surface microstructure on the feed
side of the used membranes, including broken fibrils, fibrils fused together, and broken nodes.
However, the presence of foulants and the use of chemical cleaning agents (e.g., sulfuric acid
and EDTA/Na
5
P
3
O
10
) made the cause of the observed changes unclear. Another limitation of
past studies is approximation of long-term continuous operation with intermittent operation
[31-33]. Intermittent operation may obscure the actual impacts of long-term continuous MD
system operation on membrane surface microstructure by allowing the membrane to recover
from stresses experienced during system operation.
In a preceding study by the authors of this paper [34], changes in ePTFE membrane surface
properties and MD system performance were observed during longer term operations (20 and
100 days). Foulants appeared on the feed side of the membrane despite using only a NaCl feed
solution and were suspected to have entered the system through a partially open feed auxiliary
tank. The presence of foulants prevented observation of feed-side membrane surface
microstructure; however, significant changes in distillate-side membrane surface
microstructure were observed. Fibril widening on the distillate side contributed to decreased
contact angle (i.e., decreased hydrophobicity) in both experiments and decreased surface
13

roughness in the 100-day experiment. Although these experiments demonstrated how changes
in distillate-side surface microstructure can negatively impact hydrophobicity, the contribution
of changes in feed-side surface microstructure to decreased performance (i.e., decreased water
flux) in the absence of foulants still remains unquantified in the MD literature. Additionally,
studies that have demonstrated changes in membrane surface microstructure and membrane
performance have not addressed the potential role that MD operating conditions (i.e., higher
temperature with low and constant stress applied to the membrane) may play in the severity of
morphological changes.
Saffarini et al. [44] analyzed the effects of temperature on ePTFE membranes by exposing
ePTFE membranes to air temperatures of 25, 35, and 50 °C. After only five minutes of exposure,
fibril length decreased with increasing temperature. These results demonstrated that
microstructural changes in the membrane surface occurred at temperatures within the
operating range of MD, but they did not indicate whether microstructural changes were
permanent. Permanent changes were, however, observed after annealing at 80 °C when nano-
scale changes occurred in the membrane’s underlying crystal structure. The influence of
thermal history on micro- and nano-scale properties of other forms of PTFE has also been
identified by other researchers [79-82]. While these studies do indicate that there is a
relationship between temperature and changes in micro- and nano-scale properties of PTFE,
exposure times to elevated temperatures were short (< 20 min) and permanency of
microstructural changes was not evaluated.
For ePTFE membranes in DCMD systems, the flow of feed and distillate solutions on each side
of the membrane induces a low pressure on each side (e.g., 0.1 MPa in [68]), resulting in the
14

constant application of compressive stress that can lead to membrane compaction [35-38].
Zhang et al. [35] investigated the effects of compressive stress on ePTFE membranes and found
34% decrease in thickness after short-term exposure to pressure of 0.07 MPa in a sealed flask
connected to a manometer at room temperature. In a later study [37], Zhang et al. performed
the same compression analyses with a modified setup allowing for testing at different
temperatures and showed that the ePTFE membrane was more compressible at higher
temperature (60 °C versus 21 or 45 °C). Whether or not changes in membrane thickness due to
compaction were permanent was not investigated, and the potential for the slow development
of further compaction under long-term stress exposure – known as creep – was not considered.
Creep is a phenomenon in which deformation (i.e., strain) occurs over time when applied stress
is held constant. During creep (solid red line in Fig. 1), time-independent strain (i.e., elastic
strain) occurs initially and is followed by increasing time-dependent strain (i.e., viscous strain)
at a decreasing or constant rate over time. If the stress is removed after a short time period
(Fig. 1a), then the material can return to its original dimensions (i.e., full strain recovery occurs);
if the stress is removed after a longer time period (Fig. 2.1b), then permanent (i.e., plastic)
deformation can remain in the material (i.e., partial strain recovery occurs). The series of creep
followed by recovery shown in Fig. 2.1 is known as a creep recovery test [43, 83]. Higher
temperature is associated with greater creep strain [84], which has been observed in dense
PTFE gaskets [85] and catalyst-coated Nafion membranes (which have a PTFE backbone) [43],
but to the best of the authors’ knowledge, creep recovery data for pure ePTFE membranes is
not available in the scientific literature. Creep has been suggested as the cause of observed
15

deformations in MD membranes [34, 39-42], but no analyses of the potential link between
observed MD membrane deformations and creep has been presented in the MD literature.

Figure 2.1: Creep recovery curves adapted from Desai et al. [86] for (a) full strain recovery, and
(b) partial strain recovery.
The main objective of the current study was to quantify changes in ePTFE membrane surface
morphology (microstructure, roughness, and porosity) and more broadly, membrane
morphology (surface morphology, thickness, and bulk porosity) during long-term MD
experiments in the absence of foulants, and to connect these morphological changes to
observed changes in MD membrane performance. A secondary objective was to assess the
contribution of MD operating conditions – including temperature and experiment duration – to
changes in membrane morphology. First, 30-day MD experiments were performed using ePTFE
membranes and three combinations of feed temperature and salinity. Changes in membrane
surface morphology and thickness were evaluated using scanning electron and atomic force
microscopy; these changes were used to evaluate changes in bulk membrane porosity and to
assess the influence of temperature on morphological changes. Creep recovery tests were
performed to evaluate the membrane’s tendency to shift away from virgin membrane
16

morphological characteristics over time and to further evaluate the degree to which
temperature influences the membrane’s tendency to deform.
2.2 Materials and methods
2.2.1 Membrane
Experiments were performed using a flat-sheet ePTFE membrane (Parker Performance
Materials, Lee’s Summit, MO, USA) that is hydrophobic, single-layer, and symmetric.
2.2.2 MD experiments
2.2.2.1 Experimental system
Long-term performance testing was performed using a custom-built, bench-scale DCMD system
(Fig. 2.2). The system was modified from the preceding study (i.e., McGaughey et al. [34]) in
which the heated auxiliary feed tank was partially open to the environment. To prevent dust
and other foulants from entering the feed solution, the heater was removed from the auxiliary
tank and the tank was fully covered. The inline heater in the feed circulation loop was sufficient
for maintaining constant inlet temperature. The same acrylic flat-sheet DCMD membrane
module and non-woven spacers (Sterlitech, Kent, WA) as used in the preceding study were used
in the current study.
17

Figure 2.2: Schematic of bench-scale DCMD system.
2.2.2.2 Experimental procedure
Three 30-day experiments were performed at the conditions shown in Table 2.1. All
experiments were performed with either 45 or 65 °C feed inlet temperature and 38 °C distillate
inlet temperature. A flow rate of 1.5 L/min was used on the feed and distillate sides of the
membrane. Feed solutions were made by dissolving NaCl (VWR, Radnor, PA) in deionized water
to produce a salinity of either 5 or 200 g/L; deionized water was used as the distillate solution.
Experiments are differentiated from each other throughout the current study by referring to
them with regard to their feed solution temperature and salinity; for example, “65 °C-5 g/L” is a
18

shorthand for the experiment with a 65 °C feed solution temperature and 5 g/L NaCl feed
solution concentration.
Table 2.1: Experimental matrix for three 30-day membrane distillation experiments. Feed and
distillate flow rates were 1.5 L/min.
Feed Inlet
Temperature (°C)
Distillate Inlet
Temperature (°C)
Feed Salinity
(g/L)
65 38 5
65 38 200
45 38 5

2.2.3 Membrane characterization
The feed and distillate sides of the membrane were characterized before and after each MD
experiment. All used membranes were rinsed with deionized water and desiccated for a
minimum of 24 hours prior to characterization. Analyses of virgin membranes were performed
on a section cut adjacent to the sample used in the experiment. Membrane cross-sections were
prepared using the cryo-snap freeze-fracture method [87].
Scanning electron microscopy (SEM) was performed using a field emission scanning electron
microscope (JSM-7001, Jeol USA, Huntington Beach, CA, USA) for two-dimensional
characterization of membrane morphology. Energy-dispersive x-ray spectroscopy (EDS) was
performed using an energy dispersive spectroscope (Apollo X, EDAX, Mahwah, NJ, USA)
attached to the scanning electron microscope for semi-quantitative characterization of
elemental composition. Membrane samples were sputter-coated in platinum and/or palladium
to render the surface conductive for SEM analyses. SEM images were analyzed with ImageJ
software (version 1.51j8, National Institutes of Health, Bethesda, MD, USA) to calculate
19

membrane thickness and surface porosity. Bulk porosity of membrane samples was calculated
using Equation 2.1 [88]:
𝘀 = 1 −
𝑚 𝜌 𝑝 𝐴𝛿

(2.1)
where 𝘀 is bulk porosity, 𝑚 is membrane sample mass (g), 𝜌 𝑝 is polymer density (g/cm
3
), 𝐴 is
membrane sample surface area, and 𝛿 is membrane sample thickness (cm). Polymer density is
assumed to be constant at 2.2 g/cm
3
[89, 90].
Atomic force microscopy (AFM) was performed using an atomic force microscope (Innova,
Bruker, Billerica, MA, USA) with etched silicon probes (NCHV-A, Bruker, Billerica, MA, USA) for
characterization of membrane surface roughness. AFM images were analyzed using Gwyddion
software (version 2.47, Czech Metrology Institute, Brno, Czech Republic) and surface roughness
was calculated using the two-dimensional statistical quantities tool. Surface roughness was
calculated based on five forward scans and five backward scans of 2 x 2 µm areas and the
resulting roughness values were averaged.
2.2.4 Creep recovery tests
Virgin membranes were tested using a dynamic mechanical analyzer (DMA) (Q800, TA
Instruments, New Castle, DE, USA). Samples of 6-mm width and 20-mm length were loaded into
the tensile clamp inside of the DMA’s furnace. Creep recovery tests were performed at 25 and
65 °C. In each test, the applied stress was held constant at 0.05 MPa for 1 hr, allowing the strain
in the sample to increase, then the stress was reduced to zero and the sample was allowed to
recover strain for a period of 2 hrs.
20

2.3 Results and discussion
2.3.1 Water flux during MD experiments
Graphs of water flux over time for the three experiments are shown in Fig. 2.3. Distillate
conductivity remained below 6.5 µS/cm in all three experiments, indicating that high rejection
of feed solutes was maintained throughout each experiment and no significant wetting
occurred. Because vapor pressure driving force is primarily a function of temperature [15],
water flux is lower in the 45 °C experiment (Fig. 2.3a) than the 65 °C experiments (Figs. 2.3b and
c). Water flux is lower in the 65 °C-200 g/L experiment (Fig. 2.3c) than in the 65 °C-5 g/L
experiment (Fig. 2.3b) because of the decrease in vapor pressure driving force caused by the
higher salinity of the feed solution [18]. In all three graphs of Fig. 2.3, water flux decreases over
time. Over the 30 days, water flux decreased by 3.7, 7.1, and 5.9% in the 45 °C-5 g/L, 65 °C-5
g/L, and 65 °C-200 g/L experiments, respectively. The gradual decrease in water flux observed
in each experiment was unexpected because there were no foulants in the solution chemistry
and the system was completely closed. Scalants could not have caused the decrease in water
flux because feed salinity was well below the solubility limit of sodium chloride (360 g/L). EDS
and SEM analyses were performed to verify the absence of foulants and scalants on the
membrane surfaces.
21

Figure 2.3: Water flux over time in experiments with (a) 45 °C feed inlet temperature and 5 g/L
NaCl feed concentration (45 °C-5 g/L), (b) 65 °C feed inlet temperature and 5 g/L NaCl feed
concentration (65 °C-5 g/L), and (c) 65 °C feed inlet temperature and 200 g/L NaCl feed
concentration (65 °C-200 g/L). Distillate inlet temperature was 38 °C and flow rates were 1.5
L/min.
2.3.2 Morphological changes during MD experiments
EDS spectra and SEM images for virgin and used membranes in the three experiments are
shown in Figs. 2.4 and 2.5. Because EDS spectra and SEM images for both the feed and distillate
sides of the virgin membrane were almost identical, a single EDS spectrum and SEM image is
used to represent both. EDS spectra for all virgin and used membranes (Fig. 2.4) showed a
carbon-to-fluorine ratio of approximately 0.5, as expected from the C
2
F
4
chemical formula of
PTFE. In addition to carbon and fluorine, only conductive coating (platinum and/or palladium)
was present; indicating neither foulants nor scalants were present.
22

Figure 2.4: Energy dispersive x-ray spectroscopy (EDS) spectra of: (a-c) virgin membranes, (d-f)
feed side of used membranes, and (g-i) distillate side of used membranes for experiments with
45 °C feed inlet temperature and 5 g/L NaCl feed concentration (45 °C-5 g/L), 65 °C feed inlet
temperature and 5 g/L NaCl feed concentration (65 °C-5 g/L), and 65 °C feed inlet temperature
and 200 g/L NaCl feed concentration (65 °C-200 g/L), respectively.
23

Figure 2.5: Scanning electron microscopy (SEM) images of: (a-c) virgin membrane, (d-f) feed side
of used membranes, and (g-i) distillate side of used membranes for experiments with 45 °C feed
inlet temperature and 5 g/L NaCl feed concentration (45 °C-5 g/L), 65 °C feed inlet temperature
and 5 g/L NaCl feed concentration (65 °C-5 g/L), and 65 °C feed inlet temperature and 200 g/L
NaCl feed concentration (65 °C-200 g/L), respectively. SEM images were taken at ×4000
magnification and correspond to the EDS spectra in Fig. 2.4.The SEM images in Fig. 2.5 further
24

confirm that neither foulants nor scalants were present on the membrane surface; however,
the images do reveal changes to the membrane surface microstructure. SEM images of the
virgin membranes (Fig. 2.5a-c) demonstrate the classic network of distinct nodes
interconnected by thin fibrils that is characteristic of ePTFE membrane surface microstructure
[34, 44]. SEM images of the feed side of used membranes in the 45 °C-5 g/L experiment (Fig.
2.5d) show some of the classic ePTFE surface microstructure, but also show widened fibrils and
possibly flattened nodes, resulting in less distinction between fibrils and nodes. The feed side of
the membranes used in the 65 °C experiments (Figs. 2.6e and f) shows more significant
deviation from the classic ePTFE surface microstructure. The surface microstructure of these
membranes consists of areas with nodes and fibrils appearing fused together into “aggregated
areas” with an underlying fibrous structure. The more severe changes for the 65 °C experiments
compared to the 45 °C experiment indicates that temperature plays an important role in
severity of microstructural changes to the membrane surface. Although higher salinity solutions
can result in greater shear stress on the membrane surface by increasing the viscosity [91], the
similarity in microstructural changes for the two different salinities in the 65 °C experiments
suggests that this does not contribute to significant differences in microstructural changes to
the membrane.
On the distillate side of the used membranes, fibril widening was observed for all three
experiments (Fig. 2.5g-i) and sections with flattened nodes were observed in both 65 °C
experiments (Fig. 2.5h-i). However, changes on the lower-temperature distillate side were
generally less severe than on the higher-temperature feed side, further supporting the idea
that higher temperatures increase the severity of changes in surface microstructure.
25

Small amounts of fibril widening and fusing were observed on the feed side of used ePTFE
membranes by Guillen-Burrieza et al. [31], however, these observations were made after
fouling and scaling, exposure to cleaning agents, and intermittent (rather than continuous)
operation – making the cause of the microstructural changes unclear. A significant amount of
fibril widening was also observed on the distillate side of ePTFE membranes by McGaughey et
al. [34]. However, the current study is the first to show that significant changes in feed-side
membrane surface microstructure occur after continuous operation under typical MD
operating conditions in as little as 30 days.
Differences in surface microstructure between virgin and used membranes were quantified by
comparing the proportion of black (i.e., void) space in each SEM image using surface porosity
(Table 2.2). Feed-side surface porosity decreased by 55, 80, and 89% in the 45 °C-5 g/L, 65 °C-5
g/L, and 65 °C-200 g/L experiments (Fig. 2.5). The greater changes in surface porosity in the 65
°C experiments than in the 45 °C experiment provide further support the assertion that
temperature influences changes in membrane surface microstructure. As expected, distillate-
side surface porosity decreased less than feed-side surface porosity; decreases of 41, 44, and
48% were measured for the 45 °C-5 g/L, 65 °C-5 g/L, and 65 °C-200 g/L experiments.
Table 2.2: Surface porosity of virgin and used membranes for experiments with feed-side
conditions of 45 °C feed inlet temperature and 5 g/L NaCl feed concentration (45 °C-5 g/L), 65 °C
feed inlet temperature and 5 g/L NaCl feed concentration (65 °C-5 g/L), and 65 °C feed inlet
temperature and 200 g/L NaCl feed concentration (65 °C-200 g/L).
Virgin Used Feed Used Distillate
Experiment
Surface
Porosity
Surface
Porosity % Change
Surface
Porosity % Change
45 °C-5 g/L 0.44 0.20 -55% 0.26 -41%
65 °C-5 g/L 0.42 0.08 -80% 0.23 -44%
65 °C-200 g/L 0.42 0.04 -89% 0.22 -48%
26

To verify flattening of nodes, surface roughness was quantified using AFM scans (Fig. 2.6).
Consistent with the trends in changes in surface microstructure observed in the SEM images,
these scans indicate a significant decrease in feed-side surface roughness in the 65 °C
experiments (Figs. 2.7c and d) and minimal decrease in surface roughness in the 45 °C
experiment (Fig. 2.6b). Feed-side surface roughness decreased by 23.9, 44.8, and 83.2% on the
smoothest areas of the membrane in the 45 °C-5 g/L, 65 °C-5 g/L, and 65 °C-200 g/L
experiments. The decreased surface roughness supports the assertion that some nodes are
flattened. Changes in surface roughness on the distillate side were statistically insignificant
according to T-test results (data not shown). In the preceding study [34], a 57% decrease in
distillate-side surface roughness was observed; this is likely due to the more than three-times
longer experiment duration (100 days).
27

Figure 2.6: Atomic force microscopy (AFM) images for feed side of (a) virgin membrane, and
used membranes in experiments with (b) 45 °C feed inlet temperature and 5 g/L NaCl feed
salinity (45 °C-5 g/L), (c) 65 °C feed inlet temperature and 5 g/L NaCl feed salinity (65 °C-5 g/L),
and (d) 65 °C feed inlet temperature and 200 g/L NaCl feed salinity (65 °C-200 g/L).
The presence of flattened nodes and widened fibrils could be indicative of membrane
compaction – a phenomenon known to occur during ePTFE membrane testing (e.g., [35, 37]).
For this reason, cross-sectional SEM images were taken of virgin and used membranes for
measurement of membrane thickness (Fig. 2.7). Membrane thickness decreased by 13.2, 39.1,
and 24.3% in the 45 °C-5 g/L (Fig. 2.7b), 65 °C-5 g/L (Fig. 2.7c), and 65 °C-200 g/L (Fig. 2.7d)
experiments. The greater decrease in thickness (i.e., greater compaction) of membranes in the
65 °C experiments than in the 45 °C experiment supports the theory that severity of changes in
membrane microstructure is greater at higher temperatures. Zhang et al. [35, 37] showed
comparable decreases in thickness of ePTFE membranes at the same pressures (0.04 MPa on
feed side and 0.03 MPa on distillate side) as in the current study, but those changes were
28

demonstrated at the time of exposure to compressive stress. The results of the current study
are the first to demonstrate a permanent decrease in ePTFE membrane thickness after the
compressive stress is removed.

Figure 2.7: Cross-sectional scanning electron microscope (SEM) images of (a) virgin membrane,
and used membranes from experiments operated with (b) 45 °C and 200 g/L NaCl (45 °C-5 g/L),
(b) 65 °C and 5 g/L NaCl (65 °C-5 g/L), and (d) 65 °C and 200 g/L NaCl (65 °C-200 g/L) feed
solutions.
Membrane thickness measurements were used to calculate changes in the membrane’s bulk
porosity in each experiment according to Equation 2.1. Compared to the virgin membrane, bulk
porosity decreased by 10, 20, and 14% in the 45 °C-5 g/L, 65 °C-5 g/L, and 65 °C-200 g/L
experiments (Table 2.3). The much smaller decrease in bulk porosity (Table 2.3) than in surface
porosity (Table 2.2) may partly explain why very large decreases in surface porosity did not
result in very large decreases in water flux (Fig. 2.3). Additionally, thinner membranes are
known to result in greater water flux than thicker membranes [60, 61], so the severity of
decrease in water flux caused by decreased porosity may be tempered by improvement in
membrane mass transfer characteristics caused by decreased thickness. Because the decrease
29

in water flux observed in Fig. 2.3 developed slowly over time, it seems likely that the decreases
in membrane porosity and thickness also develop slowly over time.
Table 2.3: Bulk porosity of virgin and used membranes for experiments with feed-side conditions
of 45 °C feed inlet temperature and 5 g/L NaCl feed concentration (45 °C-5 g/L), 65 °C feed inlet
temperature and 5 g/L NaCl feed concentration (65 °C-5 g/L), and 65 °C feed inlet temperature
and 200 g/L NaCl feed concentration (65 °C-200 g/L).
Bulk Porosity
Experiment Virgin Used % Change
45°C-5g/L 0.80 0.72 -10%
65°C-5g/L 0.80 0.64 -20%
65°C-200g/L 0.80 0.69 -14%

2.3.3 Time- and temperature-dependent deformation during creep and recovery
To evaluate relationships between membrane deformation (i.e., strain), length of time that
stress is applied, and temperature, creep recovery tests were performed on a virgin membrane
with an applied tensile stress of 0.05 MPa at temperatures of 25 and 65 °C (Fig. 2.8). An applied
stress of 0.05 MPa was selected because it is within the operating range of pressures expected
in an MD system (e.g., 0.1 MPa in [68]). In the creep recovery test, applied stress is held
constant (0.05 MPa in Fig. 2.9) during the creep period of the test (0 to 60 min) and then the
applied stress is removed to begin the strain recovery period (60 to 180 min). During the creep
period (0 to 60 min) at both temperatures, elastic (i.e., time-independent) deformation occurs
initially, followed by viscous (i.e., time-dependent) deformation at a decreasing rate over time.
During the strain recovery period (60 to 180 min) at both temperatures, elastic deformation
(i.e., strain) is recovered initially, followed by viscous strain recovery at a decreasing rate over
time.
30

Figure 2.8: Creep recovery curves for virgin membrane with 0.05 MPa tensile stress at (a) 25 °C
and (b) 65 °C.
Strain reached 5.5% before the stress was removed at 25 °C (Fig. 2.8a) and 15.9% before the
stress was removed at 65 °C (Fig. 2.8b), indicating that higher temperature allows for greater
deformation at the same applied stress and that deformation at a constant applied stress can
increase with time. These results corroborate the more severe changes in membrane
morphology observed in the 65 °C MD experiments compared to the 45 °C experiment. The
increasing deformation with time under constant applied stress (0 to 60 min) supports the
assertion that the morphological changes observed in membranes used during MD experiments
may develop over time, which would be consistent with the slowly decreasing water flux over
time observed in each experiment (Fig. 2.3).
The strain remaining in the sample at the end of the strain recovery period is 0.42% at 25 °C
(Fig. 2.8a) and 5.3% at 65 °C (Fig. 2.8b). The strain recovery rate at the end of the strain
recovery period is 0.01%/min at both 25 and 65 °C. If this strain recovery rate were to stay
31

constant, the sample would return to its original dimensions in 0.7 hr at 25 °C and 8.0 hr at 65
°C. However, the strain recovery rate could also continue to decrease even closer to 0%/min,
which would indicate permanent deformation. The creep recovery test results in Fig. 2.8
demonstrate how continuous application of stress to the membrane during regular operation –
even at a low value – could result in membrane deformation over time and that the
deformation could be permanent. These results are consistent with creep recovery test results
using Nafion membranes (which have a PTFE backbone) [43], but to the best of the authors’
knowledge, the results in Fig. 2.8 are the first creep recovery test results presented in the
scientific literature for pure ePTFE membranes.
2.4 Conclusions
The results of the current study show for the first time that feed-side ePTFE membrane surface
morphology (microstructure, roughness, and porosity) can change significantly during
continuous exposure to typical MD operating conditions (i.e., higher temperature with low and
constant stress application) in as little as 30 days. The feed side of the membrane is the first
barrier to transport of contaminants across the membrane, so understanding how this barrier
changes under typical MD operating conditions is important. Results of the current study also
showed that changes in membrane morphology (surface morphology, thickness, and bulk
porosity) alone (i.e., in the absence of foulants) can result in decreasing water flux over time –
connecting changes in morphology to changes in MD system performance. These results imply
that modeling approaches that assume membrane morphology remains constant over time
may overpredict water flux during long-term operation using ePTFE membranes. SEM analyses
and creep recovery test results in the current study showed that higher temperatures may
32

allow for greater deformation of ePTFE membranes under constant applied stress. Creep
recovery tests were shown to provide a useful and simple framework for characterizing (i) a
membrane’s tendency to shift away from virgin membrane morphological characteristics over
time and (ii) the degree to which temperature influences the membrane’s tendency to deform.
Creep recovery test results (e.g., maximum strain reached during creep and strain remaining in
the sample at the end of recovery) provide useful metrics for comparing the influence of time
and temperature on changes in morphology between different membranes, without the need
for running long-term MD experiments to assess this behavior.

33

Chapter 3 Membrane distillation driven by intermittent and variable-
temperature waste heat: System arrangements for water production
and heat storage
3.0 Abstract
The intermittency and variability inherent to many waste heat sources has largely been
overlooked in existing studies of membrane distillation (MD). In the current study, MD system
operation with intermittent and variable-temperature waste heat was assessed with two
system arrangements: “direct” and “indirect.” In the direct arrangement, the heat exchanger
and membrane module are in a single loop; in the indirect arrangement, they are in two
separate loops. Modeling results indicate the direct arrangement produced 17.7% more water
at 12.5% intermittency and the indirect arrangement produced 21.5% more water at 87.5%
intermittency, due to the indirect arrangement’s ability to store more heat when the waste
heat source is on. Waste heat variability was strongly reflected in water flux profiles, but the
indirect arrangement showed significantly less water flux variability – more than two times less
in modeling analyses with intermittent waste heat and 30.4% less in variable-temperature
experiments. Lower water flux variability in the indirect arrangement translates to better
system controllability, even when the direct arrangement produces more water. The
advantages of each arrangement identified in the current study give system designers key
information to improve water production, heat storage, and/or system control in different
waste heat scenarios.
34

3.1 Introduction
Membrane distillation (MD) is a thermally driven water treatment process that can be used in
desalination [11-17] and water reuse [77, 92, 93] applications. In direct contact MD (DCMD), a
warm feed solution (e.g., brackish water, seawater, wastewater, or other impaired water) is
passed along one side of a hydrophobic, microporous membrane and a cooler distillate solution
is passed along the other side. The temperature difference across the membrane induces a
transmembrane vapor pressure difference, which causes water vapor to pass through the
membrane pores and condense upon contact with the cool distillate stream. Benefits of MD
include the ability to treat hypersaline solutions with minimal decrease in driving force [18],
high rejection of non-volatile contaminants [5], the ability to use low-pressure system
components with low capital cost and low safety concerns [10], and compatibility with low-
grade waste heat as an energy source [8, 15, 50, 94, 95].
While definitions of “low-grade waste heat” differ between studies, the United States
Department of Energy defines it as waste heat having a temperature between 25 and 150 °C,
and estimates that 1222 trillion BTUs of low-grade waste heat are emitted each year in the
United States alone [96]. This represents a significant and untapped source of energy. In
particular, the fraction of low-grade waste heat emitted at temperatures less than 100 °C,
which is less useful for other heat recovery or waste-heat-driven processes, can be exploited by
MD. Low-grade waste heat is available from many different energy processes in the power and
industrial sectors. Power sector sources of low-grade waste heat include flue gas and cooling
water from thermoelectric power plants [54, 55], while industrial sources include kiln gasses
and cooling streams in the cement and ceramic industries [56-60].
35

The inherently cyclic conditions and variable loads in industrial processes and power generation
result in intermittent and variable supplies of waste heat. Thus, intermittency and variability are
key considerations for a sizable fraction of the low-grade waste heat available for MD.
Temperature profiles from example intermittent and variable-temperature waste heat sources
were characterized for the current study (Fig. 3.1). Fig. 3.1a shows the temperature profile of
flue gas from a natural-gas-fired boiler over a 24-hour period. The most striking characteristic of
the temperature profile is the temperature cycling, or intermittency, caused by the boiler
turning on and off depending on the heating demand. Fig. 3.1b shows the temperature profile
of an industrial autoclave discharge. The autoclave discharge temperature profile shows
significant temperature variability throughout the 24-hour period. Notwithstanding the data in
Fig. 3.1, it is common practice to quantify waste heat availability using an average temperature
and/or a single energy value that unintentionally obscures the variability and/or intermittency
by lumping energy availability into a single heat content value (for a finite time period) [59]. To
provide a more realistic assessment of waste-heat-driven MD system operation, variability and
intermittency must be decoupled from average temperature and bulk heat content values and
evaluated for their individual impacts on MD system operation and control.
36

Figure 3.1: Temperature profiles from example low-grade waste heat sources: (a) Intermittent
temperature profile of flue gas from an industrial gas-fired boiler and (b) variable temperature
profile from an industrial autoclave discharge.
Existing waste-heat-driven MD studies have clearly demonstrated the ability to drive MD
systems using waste heat and the potential economic competitiveness of such a strategy [6, 8,
39, 45-53], but have not considered temperature variability or intermittency. Heat source
variability was only mentioned in three studies of waste-heat-driven MD [8, 39, 51]; no
implications were discussed. Temperature variability in these studies may have been low and
therefore not particularly concerning, but the data in Fig. 3.1b provide clear evidence that
waste heat variability can be significant. Variability in waste heat source temperature should be
considered because it is likely to result in significant variability in water flux, which would result
in low predictability of product water availability and could necessitate the use of more
complex control systems. Waste heat intermittency (as shown in Fig. 3.1a) can have an even
more significant effect on waste-heat-driven MD system performance by causing periodic
37

losses of the thermal driving force required for continuous operation. To continuously operate
MD systems and maintain product water availability even when the waste heat source is off,
heat storage may be required. The limited literature discussing heat storage in MD systems is
found in solar-thermal MD literature [13, 66, 74].
Heat storage has been integrated into solar-thermal MD systems through two main system
arrangements: “direct” and “indirect.” The direct and indirect terminology, coined by Banat et
al. [75], is synonymous with the “single-loop” and “two-loop” terminology used in other solar-
thermal MD studies. In the direct arrangement, feed solution is circulated in a single loop from
the feed tank to the solar collector and through the membrane module [13, 68, 74, 75, 97].
Advantages of the direct arrangement include immediate delivery of the most recently heated
feed solution to the membrane module, as well as low capital cost and compact system size.
Disadvantages include the need for corrosion-resistant solar collectors that are much more
expensive and less widely available than conventional collectors [98, 99] and less system
controllability associated with using a single pump to control heat extraction and membrane
channel hydrodynamics. In the indirect arrangement, feed solution is simultaneously circulated
through the membrane module in a “membrane loop” and through the solar collector in a
separate “heat loop” [66, 70, 100-106]. The main advantage of the indirect arrangement is that
it separates heat extraction and membrane module flow control, allowing higher flow rates to
be used in the heat loop so that more heat can be extracted without affecting membrane
channel hydrodynamics. The main disadvantage of the indirect arrangement is the higher
capital cost associated with the extra pump required to have separate loops. If either the direct
38

or indirect arrangement is used with an intermittent heat source (e.g., solar or waste heat),
heat storage can be incorporated using the feed tank also as a heat storage tank.
All existing waste-heat-driven MD modeling and experimental studies have used the direct
arrangement (Fig. 3.2a); none of these studies has considered heat storage needs in their
analyses. The indirect arrangement (Fig. 3.2b) offers the potential for more heat extraction and
storage, due to the ability to operate at a higher flow rate in the heat loop than in the
membrane loop so that heat can be extracted and stored at a faster rate. A detailed
comparison of system arrangements in the context of waste-heat-driven MD systems has not
been made before, particularly in the unique context of intermittent or variable-temperature
waste heat. In comparison to intermittency in solar energy, which is generally predictable due
to the familiar and consistent bell-shaped curve of daily solar irradiance [75, 107], intermittency
in waste heat sources is more complex – varying from industry-to-industry, source-to-source,
and even day-to-day – making the design of waste-heat-driven MD systems more complex. This
extra complexity has not been considered in the existing MD literature.

Figure 3.2: (a) Direct and (b) indirect system arrangements.
39

The main objective of the current study was to evaluate direct and indirect arrangements of
MD driven by intermittent and variable-temperature waste heat sources. Modeling analyses
were used to assess water flux and heat storage behavior under different degrees of waste heat
intermittency. Experimental analyses were used to confirm modeling results for waste heat
intermittency and to evaluate the more thermodynamically complex scenario of variable-
temperature waste heat. A secondary objective was to evaluate how system arrangement can
affect system controllability and predictability of product water availability in MD systems
driven by intermittent or variable-temperature waste heat sources. To the best of the authors’
knowledge, the results of the current study represent the first in-depth analysis of the roles of
waste heat source characteristics (i.e., intermittency and variability) in MD system
performance.
3.2 Materials and methods
3.2.1 Modeling
A previously validated steady-state model of DCMD with a flat-sheet membrane [108] was used
to evaluate DCMD performance. The model uses a stepwise solution method in which the
membrane area is discretized into smaller subsections (steps). Model equations are applied to
the first step with the output from first-step calculations serving as input for the second step;
the process is repeated along the length of the membrane. The stepwise calculation process is
iterated until the error between known and calculated inlet conditions is below a specified
tolerance. The model was previously validated with a commercial polytetrafluoroethylene
(PTFE) membrane [108]; the same membrane is used in the current study and is described in
Section 2.2.1 below. Membrane length of 40.6 cm, membrane width of 27.9 cm, and channel
40

height of 0.7 mm were used in all simulations with counter-current flow through the membrane
module.
A steady-state model of a flat-plate heat exchanger was used to model the feed-side heat
exchanger. To account for the changing driving force along the length of the heat exchanger,
the heat exchanger area was discretized into smaller sections and model equations were
applied to successive sections of the heat exchanger, in the same manner as was used for the
DCMD model. The heat exchanger model was based on the log mean temperature difference
method [109] for counter-current flow, described by:
𝑞 𝑗 = 𝑚 ̇ ℎ , 𝑗 𝑐 𝑝 , ℎ , 𝑗 ( 𝑇 ℎ , 𝑖 , 𝑗 − 𝑇 ℎ , 𝑜 , 𝑗 ) (3.1)
𝑞 𝑗 = 𝑚 ̇ 𝑐 , 𝑗 𝑐 𝑝 , 𝑐 , 𝑗 (𝑇 𝑐 , 𝑜 , 𝑗 − 𝑇 𝑐 , 𝑖 , 𝑗 ) (3.2)
𝑞 𝑗 = 𝑈 𝑗 𝐴 𝑗 ∆ 𝑇 𝐿𝑀 , 𝑗 (3.3)
𝑈 𝑗 =
1
ℎ
ℎ , 𝑗 − 1
+ 𝑑 𝑝 /𝑘 𝑝 + ℎ
𝑐 , 𝑗 − 1
(3.4)
∆ 𝑇 𝐿𝑀 , 𝑗 =
( 𝑇 ℎ , 𝑖 , 𝑗 − 𝑇 𝑐 , 𝑜 , 𝑗 ) − ( 𝑇 ℎ , 𝑜 , 𝑗 − 𝑇 𝑐 , 𝑖 , 𝑗 )
ln [
𝑇 ℎ , 𝑖 , 𝑗 − 𝑇 𝑐 , 𝑜 , 𝑗 𝑇 ℎ , 𝑜 , 𝑗 − 𝑇 𝑐 , 𝑖 , 𝑗 ] ⁡

(3.5)
ℎ
ℎ / 𝑐 , 𝑗 = 𝑁𝑢
ℎ / 𝑐 , 𝑗 𝑘 ℎ / 𝑐 , 𝑗 𝐷 ℎ
(3.6)
𝑁𝑢
ℎ / 𝑐 , 𝑗 = 1 .67 𝑅𝑒
ℎ / 𝑐 , 𝑗 0 . 44
𝑃𝑟 ℎ / 𝑐 , 𝑗 0 . 5
⁡⁡(45 ≤ 𝑅𝑒 ≤ 300 ) (3.7)
𝑁𝑢
ℎ / 𝑐 , 𝑗 = 0 .405 𝑅𝑒
ℎ / 𝑐 , 𝑗 0 . 7
𝑃𝑟 ℎ / 𝑐 , 𝑗 0 . 5
⁡⁡(300 < 𝑅𝑒 ≤ 2000 ) (3.8)
41

𝑁𝑢
ℎ / 𝑐 , 𝑗 = 0 .84 𝑅𝑒
ℎ / 𝑐 , 𝑗 0 . 6
𝑃𝑟 ℎ / 𝑐 , 𝑗 0 . 5
⁡⁡(2000 < 𝑅𝑒 ≤ 20 ,000 ) (3.9)
𝑅𝑒
ℎ / 𝑐 , 𝑗 =
𝜌 ℎ / 𝑐 , 𝑗 𝑣 ℎ / 𝑐 , 𝑗 𝐷 ℎ
𝜇 ℎ / 𝑐 , 𝑗 (3.10)
𝑃𝑟 ℎ / 𝑐 , 𝑗 =
𝜇 ℎ / 𝑐 , 𝑗 𝐶 𝑝 , ℎ / 𝑐 , 𝑗 𝑘 ℎ / 𝑐 , 𝑗 (3.11)
where 𝑞 is heat transfer rate (W), 𝑚 ̇ is mass flow rate (kg/s), 𝑐 𝑝 is specific heat capacity (J kg
-1
K
-
1
), 𝑇 is temperature (°C), 𝑈 is overall heat transfer coefficient (W m
-2
K
-1
), 𝐴 is heat exchanger
plate surface area (m
2
), ∆ 𝑇 𝐿𝑀
is log mean temperature difference (°C), ℎ is convective heat
transfer coefficient (W m
-2
K
-1
), 𝑁𝑢 is dimensionless Nusselt number, 𝑅𝑒 is dimensionless
Reynolds number, 𝑃𝑟 is dimensionless Prandlt number, 𝑑 is heat exchanger plate thickness (m),
𝑘 is thermal conductivity (W m
-1
K
-1
), 𝐷 ℎ
is hydraulic diameter of the heat exchanger channel
(m), 𝜌 is density (kg m
-3
), 𝑣 is velocity (m/s), and 𝜇 is dynamic viscosity (kg s
-1
m
-1
). Subscripts
ℎ,⁡ 𝑐 , 𝑖 , 𝑜 , 𝑝 , and 𝑗 denote hot stream, cold stream, inlet, outlet, heat exchanger plate, and
section of the discretized heat exchanger, respectively. A waste heat source temperature of
87.6 °C and flow rate of 6 L/min was used in all simulations. A heat loop flow rate was used in
the indirect arrangement that was seven times higher than the membrane loop flow rate, for all
simulations. The heat exchanger design used in the model had a 30.5-cm long and 15.2-cm wide
plate with a channel height of 1 mm; the plate was modeled as titanium with a thickness of
2.24 mm and a thermal conductivity of 21.5 W m
-1
K
-1
[33]. This heat exchanger area was
selected to result in a membrane area-to-heat exchanger area ratio of approximately 2.5, as is
considered optimal in the literature (e.g., [99, 105]). Heat exchanger equations were iteratively
solved using a similar process to the DCMD model. Because the current study focused only on
42

the feed-side of the DCMD system, only the feed-side heat exchanger was modeled while the
distillate side was assumed to have a constant membrane module inlet temperature of 30 °C
for all simulations.
For each arrangement, a steady-state energy balance was performed on the feed tank to
determine feed tank temperature and corresponding feed solution temperatures that enter the
heat exchanger and membrane module. The energy balance on the feed tank for the direct
arrangement was modeled using:
𝐸 ̇ 𝑡 , 𝑖𝑛
= 𝐸 ̇ 𝑡 , 𝑜𝑢𝑡 (3.12)
𝑒 𝑚 , 𝑜 𝑚 ̇ 𝑚 , 𝑜 + 𝑒 𝑅 𝑚 ̇ 𝑤 = 𝑒 𝑡 𝑚 ̇ 𝐻𝑋 , 𝑖 (3.13)
where 𝐸 ̇ 𝑡 , 𝑖𝑛
is the rate of energy entering the feed tank (W), 𝐸 ̇ 𝑡 , 𝑜𝑢𝑡 is the rate of energy leaving
the feed tank (W), 𝑒 is specific enthalpy (J/kg), and 𝑚 ̇ is mass flow rate (kg/s). The subscripts 𝑡 ,
𝑚 , 𝑅 , 𝑤 and 𝐻𝑋 represent conditions of the feed tank, membrane module, feed tank
replenishment water, water produced by the membrane, and heat exchanger, respectively. The
feed tank was assumed to be replenished with distilled water at a rate equal to the water
production rate from the DCMD process. The energy balance on the feed tank for the indirect
arrangement was:
𝑒 𝑚 , 𝑜 𝑚 ̇ 𝑚 , 𝑜 + 𝑒 𝐻𝑋 , 𝑜 𝑚 ̇ 𝐻𝑋 , 𝑜 + 𝑒 𝑅 𝑚 ̇ 𝑤 = 𝑒 𝑡 𝑚 ̇ 𝑚 , 𝑖 + 𝑒 𝑡 𝑚 ̇ 𝐻𝑋 , 𝑖 (3.14)
Feed tank temperature was determined by iteratively solving Eq. (3.13) or (3.14). The DCMD,
heat exchanger, and feed-tank energy balance models were then coupled into an MD system
model.
43

The rate of heat storage in the MD system was modeled as the difference between the heat
entering the system through the heat exchanger and the heat leaving the system through the
membrane using:
𝑞 ̇ 𝐻𝑆 , 𝑡 = 𝑞 ̇ 𝐻𝑋 , 𝑡 − 𝑞 ̇ 𝑚 , 𝑡 (3.15)
𝑞 ̇ 𝐻𝑋 , 𝑡 = 𝑒 𝐻𝑋 , 𝑜 𝑚 ̇ 𝐻𝑋 , 𝑜 − 𝑒 𝐻𝑋 , 𝑖 𝑚 ̇ 𝐻𝑋 , 𝑖 (3.16)
𝑞 ̇ 𝑚 , 𝑡 = 𝑒 𝑚 , 𝑖 𝑚 ̇ 𝑚 , 𝑖 − 𝑒 𝑚 , 𝑜 𝑚 ̇ 𝑚 , 𝑜 (3.17)
where 𝑞 ̇ 𝐻𝑆 , 𝑡 is the rate of heat storage (W), 𝑞 ̇ 𝐻𝑋 , 𝑡 is the rate of heat input from the heat
exchanger (W), and 𝑞 ̇ 𝑚 , 𝑡 is the rate of heat loss from the feed solution across the membrane
(W) – all at a given feed tank temperature. The quantity of heat required to change the
temperature of solution in the feed tank was divided by the heat storage rate to determine the
time required to achieve that change in feed tank temperature:
𝑑𝑡 𝑡 =
𝑑𝑞 𝑡 ,𝑑𝑇
𝑞 ̇ 𝐻𝑆 , 𝑡 =
𝑉 𝑡 𝜌 𝑡 𝐶 𝑝 , 𝑡 𝑑𝑇 𝑡 𝑞 ̇ 𝐻𝑆 , 𝑡 (3.18)
where 𝑉 𝑡 is the feed tank volume, and 𝑑𝑡 𝑡 is the differential time (s) required to achieve a
differential temperature change 𝑑𝑇 𝑡 (K) with a differential quantity of heat input 𝑑𝑞 𝑡 ,𝑑𝑇
(J) – all
at a given feed tank temperature. The rate of heat storage for a specified feed tank
temperature was used with a specified feed tank temperature differential in Eq. (3.18) to
develop curves of water flux over time. The water flux versus time curves were integrated to
produce curves of product water volume over time.
44

Energy consumption was assessed using specific thermal energy consumption (STEC). STEC is
traditionally defined as the rate of heat transfer into the system across the heat exchanger,
divided by the water production rate at a point in time [6, 62, 110]. In systems that store heat
for later use, the traditional STEC definition would not correspond to the heat consumed to
produce water at a point in time, unless the system were at steady state, where heat
transferred into the system equals heat consumed by the system. In fact, STEC using the
traditional definition would be zero when operating from stored heat. In the current study, it is
proposed to define a new STEC based on the rate of heat transfer across the membrane
module, to characterize the heat consumed by the system to produce water at any time,
whether or not the system is operating at steady state or from stored heat. To differentiate
STEC defined for the membrane module from the traditional STEC, 𝑆𝑇𝐸𝐶 𝑚 (kWh/m
3
) is used to
denote STEC of the membrane module:
𝑆𝑇𝐸𝐶 𝑚 =
𝑞 ̇ 𝑚 𝐽 𝑤 𝐴 𝑚 (3.19)
where 𝐴 𝑚 is membrane area (m
2
). If heat recovery techniques are used (e.g., as in [110, 111]),
then 𝑆𝑇𝐸𝐶 𝑚 will be higher than the traditional STEC value, and the comparison of these values
would indicate the energy savings associated with heat recovery. The feed tank was assumed to
be well-mixed in all modeling analyses and heat losses to the surroundings were neglected.
Energy required to pump feed solution in the membrane and heat loops was also neglected
because calculations (data now shown) indicated that total pumping energy was always less
than 1.4% of total specific energy consumption for the conditions tested. This finding agrees
with the very low pumping energy consumption found by other MD researchers [62, 112].
45

3.2.2 Experimental
3.2.2.1 MD membrane and solution chemistries
A commercial flat-sheet PTFE membrane (Parker Performance Materials, Lee’s Summit, MO)
was used in all experiments; this was the same membrane used in the modeling analyses. The
membrane is hydrophobic, symmetric, does not have a support layer, has a thickness of 67 µm,
and has porosity of 80% [88, 113]. The membrane has an average pore size of 0.3 ± 0.17 µm,
nearly 90% of the pores have a diameter less than 0.5 µm, and 61% of the pores have diameters
of 0.2 ± 0.06 µm [34]. The feed solution was created by dissolving sodium chloride (NaCl) (VWR,
Radnor, PA) in deionized water to achieve a final solution concentration of 10 g/L NaCl;
deionized water was used as the distillate solution.
3.2.2.2 Experimental system
A schematic of the custom DCMD system used to perform the experiments is shown in Fig. 3.3.
All wetted parts in the feed stream were either non-metallic or titanium to resist corrosion
[114]. The membrane was housed in a custom-built acrylic module that provided an active
membrane area of 0.014 m
2
. A non-woven mesh spacer (Sterlitech, Kent, WA) was placed in the
membrane module feed channel and a woven tricot spacer (Hornwood Inc., Lilesville, NC) was
placed in the distillate channel for support and to promote turbulence [108]. A gear pump
(March Manufacturing Inc., Glenview, IL) circulated the feed solution in the heat loop in the
indirect arrangement, and a peristaltic pump (Cole Parmer, Vernon Hills, IL) circulated the feed
solution in the membrane loop in both arrangements. The distillate solution was circulated
using a gear pump (Micropump, Vancouver, WA). Rotameters (Omega, Norwalk, CT) measured
flow rates of the feed and distillate streams entering the membrane module. Resistance
46

temperature detectors (Omega) monitored the temperatures in the feed tank and at the inlet
and outlet of the membrane module.

Figure 3.3. Schematic of bench-scale DCMD system, shown in the direct system arrangement.
A recirculating chiller (Cole Parmer) cooled the distillate stream via a heat exchanger (Alfa Heat
Exchangers, Tampa, FL) and a recirculating dynamic-temperature bath (Cole Parmer) heated the
feed stream via a heat exchanger (Alfa Heat Exchangers). The dynamic-temperature bath
contains heating and cooling elements that work together to maintain high-resolution
temperature control. The feed-side heat exchanger was moved to test the direct and indirect
arrangements. The distillate heat exchanger was in the direct arrangement for all experiments.
In-line conductivity probes and transmitters (Cole Parmer) monitored the conductivity of the
47

feed and distillate streams over time, allowing membrane integrity to be verified throughout
the duration of each experiment. An auxiliary tank with a solenoid valve (Electric Solenoid
Valves, Islandia, NY) was used to periodically dose the feed tank with deionized water to
maintain constant concentration in the feed tank throughout experiments.
Water vapor that passed through the membrane and condensed into the distillate solution
overflowed the distillate tank and was collected in a product water reservoir. The mass of
product water was monitored over time using an analytical scale (Ohaus, Parsippany, NJ).
Water flux was calculated by dividing the change in product water volume by membrane area
and time interval. All data were recorded using a LabVIEW (National Instruments, Austin, TX)
monitoring and control program.
3.2.2.3 Experiments
All experiments were operated in counter-current flow through the membrane module and
heat exchangers. The chiller and dynamic-temperature bath both achieved and maintained
setpoint temperatures through use of their onboard proportional-integral-derivative
temperature controllers. The total feed solution volume was 43 L.
3.2.2.3.1 Intermittent heat source experiments
The effect of intermittent waste heat on performance of MD in direct and indirect
arrangements was tested using a temperature of 87.6 °C that was turned on and off. The
dynamic-temperature bath was set to a constant temperature of 87.6 °C using LabVIEW, and
the chiller was set to 30 °C. The system was allowed to reach steady-state operation and was
then operated for 0.5 hr, after which the simulated waste heat source (dynamic-temperature
48

bath) was turned off. In the indirect arrangement, the heat loop pump was also turned off and
the heat loop was isolated from the system via shut-off valves. The system continued to
operate for a period of 4.5 hr using the heat stored within the solution in the feed tank and
with the distillate chiller still on. In the direct arrangement, the flow rate was 0.5 L/min; in the
indirect arrangement, the membrane loop flow rate was 0.5 L/min and the heat loop flow rate
was 9 L/min.
3.2.2.3.2 Variable-temperature heat source experiments
The effect of variable-temperature waste heat on MD performance in direct and indirect
arrangements was tested using the variable temperature profile shown in Fig. 3.1b. This
temperature profile was input into LabVIEW and used to control the dynamic-temperature
bath; the chiller was set to 20 °C. The system was allowed to reach steady-state operation with
the first temperature in the waste heat source temperature profile (75.9 °C) and then the heat
source temperature was updated each minute, based on the pre-programmed variable-
temperature data. In both the direct and indirect arrangements, all flow rates were 1 L/min.
Experiments were performed for 24 hr.
3.3 Results and discussion
3.3.1 Modeling analyses of system arrangements
3.3.1.1 Water flux and inlet temperature with constant-temperature waste heat source
As a baseline comparison, steady-state water flux and feed inlet temperature as a function of
membrane loop flow rate were modeled for the direct and indirect arrangements (Fig. 3.4)
using a constant-temperature (non-variable) waste heat source that is always on (non-
49

intermittent). Increasing membrane loop flow rate increases steady-state water flux in both
arrangements (Fig. 3.4a) because of greater turbulence and reduced polarization at higher
velocities [10, 115]. Steady-state water flux in the direct arrangement is an average of 17.3%
higher than in the indirect arrangement because in the direct arrangement, the high-
temperature feed solution leaving the heat exchanger is immediately sent to the membrane. In
the indirect arrangement, the most recently heated water is mixed with lower-temperature
water in the feed tank and then sent to the membrane. The result is an average 6.7% higher
feed inlet temperature in the direct arrangement for the range of flow rates analyzed (Fig.
3.4b). The relatively small (6.7%) difference in average feed temperature results in the larger
(17.3%) difference in average water flux because of the exponential relationship between feed
temperature and vapor pressure driving force. Based solely on the objective of maximizing
water flux in this scenario, where the waste heat source has a constant temperature and is
always on (as is commonly assumed in the literature), the direct arrangement would be
preferred over the indirect arrangement. However, for real-world waste heat sources that are
intermittent, an additional objective must be considered: maximizing heat storage to continue
producing water when the waste heat source is off.

50

Figure 3.4. Model results for steady-state (a) water flux and (b) feed inlet temperature in direct
and indirect system arrangements when using a constant-temperature (non-variable) waste
heat source that is always on (non-intermittent). Waste heat source temperature was 87.6 °C
when on and distillate inlet temperature was 30 °C.
3.3.1.2 Feed tank temperature and heat storage with constant-temperature waste heat source
When evaluating heat storage for a given MD system (with a fixed feed tank size), feed tank
temperature is the main determinant of the quantity of heat stored. Thus, in the current study,
feed tank temperature is used as the measure of quantity of heat stored. Fig. 3.5a shows
steady-state feed tank temperature (i.e., quantity of heat stored) as a function of membrane
loop flow rate, over the same range of flow rates shown in Fig. 3.4. The indirect arrangement
results in an average 13.3% higher steady-state feed tank temperature than the direct
arrangement, indicating that the indirect arrangement stores a greater quantity of heat at
steady state than the direct arrangement. Comparison of Figs. 3.5a and 3.4a demonstrates an
important tradeoff between the two arrangements at steady state: when the waste heat source
is on, the direct arrangement results in greater water flux while the indirect arrangement
results in greater feed tank temperature (i.e., greater quantity of heat stored).
51

Figure 3.5. Model results for (a) steady-state feed tank temperature (i.e., quantity of heat
stored) as a function of membrane loop flow rate, and (b) rate of heat storage as a function of
feed tank temperature when using a membrane loop flow rate of 2.8 L/min in direct and indirect
system arrangements. An 87.6 °C, constant-temperature (non-variable) waste heat source that
is always on (non-intermittent) was used with a distillate temperature of 30 °C. Points a and b
identify steady-state feed tank temperature in each arrangement, and points c and d identify
points of comparison for rate of heat storage at 35 °C.
The rate of heat storage was also modeled for both arrangements as a function of feed tank
temperature when using a membrane loop flow rate of 2.8 L/min (Fig. 3.5b). Until the system
reaches steady state, some fraction of heat input from the waste heat source is used to
immediately produce water while the remaining fraction is stored in the feed solution – leading
to increasing feed tank temperature (i.e., heat storage). When the feed tank temperature
reaches its steady-state value (points a and b in Fig. 3.5b), the rate of heat input from the waste
heat source is equal to the rate of heat loss across the membrane and no additional heat is
stored (i.e., all heat brought into the system is used solely for water production). The rate of
heat storage in the indirect arrangement is 138% higher than in the direct arrangement at a
feed tank temperature of 35 °C (comparison of points c and d in Fig. 3.5b), and the indirect
52

arrangement has a higher rate of heat storage at all feed tank temperatures. These results
indicate that the indirect arrangement collects and stores waste heat at a faster rate than the
direct arrangement. This faster rate of heat collection and storage allows for the cultivation of
as much heat as possible in the period of time that the waste heat source is on. Greater storage
of heat means more heat is available to produce water at times that best align with water
demand – similar to demand response in the energy sector.
3.3.1.3 Total volume of water produced, water flux variability, and energy consumption over time
with intermittent waste heat source
Fig. 3.6 shows modeling results for water flux (Figs. 3.6a and b) and product water volume (Fig.
3.6c) over time when using the same waste heat source (i.e., same temperature and flow rate)
with two different intermittencies (i.e., percentages of time off). A given degree of
intermittency can be achieved with various on/off frequencies; the analyses in the current
study consider intermittency with one on/off cycle in each 24-hour period. As before, the waste
heat source is modeled as having a constant temperature of 87.6 °C when it is on. The results in
Fig. 3.6 are shown after an initial 96-hour equilibration period that started with a feed tank
temperature of 35 °C. In Figs. 3.6a and b, a large difference in the MD system’s water flux
behavior is observed between the direct and indirect arrangements when the waste heat
source turns off and on. When the waste heat source turns off, water flux in the direct
arrangement drops suddenly because the lower-temperature feed tank – rather than the heat
exchanger – becomes the heat source. On the other hand, water flux in the indirect
arrangement decreases gradually from its last value when the waste heat source turns off
because the feed tank always serves as the heat source in the indirect arrangement, whether
53

the tank is being heated through the heat exchanger (waste heat source on) or not (waste heat
source off). When the waste heat source turns back on, similar behavior is observed as when
the waste heat source turned off, but in the opposite direction (i.e., the water flux jumps
suddenly in the direct arrangement and increases gradually in the indirect arrangement). All
existing modeling and experimental waste-heat-driven MD studies use the direct arrangement,
yet the significant drop/jump in water flux when the waste heat source turns off/on has not
been previously identified. The drop in water flux in the direct arrangement at the off/on
transition has implications for the time that the system is available to meet a specified water
flux target. For example, if using the MD system under the conditions shown in Fig. 3.6b (87.5%
intermittency) with a minimum flux requirement of 4 L m
-2
h
-1
to meet consumer demands, the
direct arrangement would only be able to meet that requirement for the first 58% of the time
that the waste heat source is off. However, the indirect arrangement would meet the 4 L m
-2
h
-1

requirement at all times.

54

Figure 3.6. Model results for (a) water flux with 12.5% intermittency in direct and indirect
system arrangements, (b) water flux with 87.5% intermittency in direct and indirect system
arrangements, and (c) product water volume in the direct and indirect arrangements with 12.5%
and 87.5% intermittency as a function of time. Points a, b, c, and d identify points of comparison
for total product water volume between arrangements and intermittencies. Waste heat source
temperature was a constant (non-variable) value of 87.6 °C when on, distillate inlet
temperature was 30 °C, and membrane loop flow rates were 2.8 L/min.
Comparison of the data in Figs. 3.6a and b confirms that higher intermittency results in lower
average water flux from both arrangements. The impacts of lower water fluxes can be seen by
comparing product water volume over time for each arrangement (Fig. 3.6c). For the same
waste heat source temperature and flow rate, the direct arrangement at 12.5% intermittency
produces more than four times the water volume at 87.5% intermittency (comparison of points
a and b in Fig. 3.6c). Similar results are obtained in the indirect arrangement (comparison of
points c and d in Fig. 3.6d), indicating that intermittency is an extremely important – yet,
consistently overlooked – factor that must be considered in the design of waste-heat-driven
MD systems. Interestingly, intermittency also affects which arrangement produces the greatest
product water volume in a given time period. For 12.5% intermittency, the direct arrangement
produces 17.7% more water than the indirect arrangement (comparison of points a and c in Fig.
3.6c), and for 87.5% intermittency, the indirect arrangement produces 21.5% more water than
55

the direct arrangement (comparison of points b and d in Fig. 3.6c). Hence, the direct
arrangement may be more favorable for waste heat sources with less intermittency, while the
indirect arrangement would be preferred for highly intermittent waste heat sources.
The quantitative differences in total volume of water produced over time by each arrangement
in Fig. 3.6 are summarized in Table 3.1 along with differences in water flux variability (i.e.,
standard deviation) over time. Variability in water flux is more than two times greater in the
direct arrangement than in the indirect arrangement for both intermittencies. Water flux
variability is an important consideration when comparing alternative system designs because
higher water flux variability results in less predictable product water availability and may
necessitate more complex control systems. Considering the lower water flux variability from
the indirect arrangement – whether using a low- or high-intermittency waste heat source – the
indirect arrangement may be favored in applications where more predictable product water
availability and simpler system controls are priorities, even if more water is produced in the
direct arrangement.

56

Table 3.1: Water flux variability and total product water volume in a 96-hour model simulation
with different degrees of intermittency when using a membrane loop flow rate of 2.8 L/min,
constant (non-variable) waste heat source temperature of 87.6 °C when on, and distillate inlet
temperature of 30 °C.

Arrangement
% Difference
Between
Arrangements Direct Indirect
12.5% Intermittency
Water Flux Variability (L m
-2
h
-1
) 7.88 2.08 279%
Total Product Water Volume (L) 344 293 17.7%
87.5% Intermittency
Water Flux Variability (L m
-2
h
-1
) 8.69 3.31 162%
Total Product Water Volume (L) 75.1 91.3 21.5%
% Difference Between
Intermittencies
Water Flux Variability (L m
-2
h
-1
) 10.2% 59.2%
Total Product Water Volume (L) 358% 221%

Energy consumption also changes as the conditions in the feed stream change over time due to
addition of heat to the system when the waste heat source is on and consumption of heat
when the waste heat source is either off or on. The data in Fig. 3.7 show 𝑆 𝑇𝐸𝐶 𝑚 for a 24-hour
subset of the simulation data shown in Fig. 3.6. 𝑆𝑇𝐸𝐶 𝑚 is lower for the direct arrangement
when the waste heat source is on and lower for the indirect arrangement when the waste heat
source is off. Lower 𝑆𝑇𝐸𝐶 𝑚 is associated with higher water flux caused by more efficient use of
thermal energy at higher temperatures due to the exponential relationship between vapor
pressure driving force and temperature. At lower intermittency (Fig. 3.7a), there is very little
difference in average 𝑆𝑇𝐸𝐶 𝑚 between arrangements. At higher intermittency (Fig. 3.7b), there
is a modestly lower (7.1%) average 𝑆𝑇𝐸𝐶 𝑚 in the indirect arrangement. While the choice of
arrangement does affect energy consumption, the effect is small because greater water flux,
which reduces 𝑆𝑇𝐸𝐶 𝑚 , is offset by a corresponding increase in heat transferred across the
membrane.
57

Figure 3.7: Model results for specific thermal energy consumption of the membrane module
(STEC
m
) with (a) 12.5% intermittency, and (b) 87.5% intermittency as a function of time in direct
and indirect system arrangements. Waste heat source temperature was a constant (non-
variable) value of 87.6 °C when on, distillate inlet temperature was 30 °C, and membrane loop
flow rates were 2.8 L/min.
3.3.2 Experimental analyses of system arrangements
3.3.2.1 Water flux and total volume of water produced with intermittent waste heat source
Experiments were performed with a simulated intermittent waste heat source to confirm the
trends observed in modeling results. Water flux and product water volume over time are shown
in Fig. 3.8 for a simulated waste heat source with 90% intermittency and a constant
temperature of 87.6 °C when the heat source is on. The system was initially allowed to reach
steady state with the waste heat source on and the system operated for 0.5 hr; the waste heat
source was then turned off and the system operated from stored heat for 4.5 hr. The
experimental results confirm model results, which showed that the direct arrangement
operates at higher water flux when the waste heat source is on and the indirect arrangement
operates at higher water flux when the waste heat source is off. The sudden drop in water flux
shown in model results for the direct arrangement when the waste heat source turned off was
58

also observed experimentally. The higher water flux from the indirect arrangement during the
period when the waste heat source is off resulted in 41.5% more product water volume at the
end of the 5-hr experiment – confirming model results, which showed that the indirect
arrangement produces more water than the direct arrangement when using a waste heat
source with high intermittency.

Fig. 3.8. Experimental results for (a) water flux and (b) product water volume over time in direct
and indirect system arrangements when using a waste heat source with 90% intermittency.
Membrane loop flow rates were 0.5 L/min, the waste heat source temperature was a constant
(non-variable) value of 87.6 °C when on, and cooling source temperature was 30 °C.
3.3.2.2 Water flux variability with variable-temperature waste heat source
Waste heat source temperature was held constant in the modeling and experimental analyses
shown thus far, but temperature is likely to vary when a waste heat source is on (e.g., 3.Fig. 1b).
Because variable-temperature waste heat is much more thermodynamically complex than
constant-temperature waste heat, modeling analyses could be prohibitive, so MD system
performance with variable-temperature waste heat was evaluated experimentally using the
real-world waste-heat data from Fig. 3.1b; results are shown in Fig. 3.9. Variability in the waste
59

heat source temperature profile (square-shaped data points) is almost directly mirrored in the
module inlet temperature (X-shaped data points), feed tank temperature (circle-shaped data
points), and water flux (triangle-shaped data points) profiles for both arrangements. The strong
relationship between variability in waste heat source temperature and variability in water flux
demonstrated by the data in Fig. 3.9 indicates that availability of product water will be less
predictable and system control will be more complex compared to a system using a constant-
temperature heat source.

Fig. 3.9: Feed temperatures and water flux versus time for (a) direct and (b) indirect system
arrangement experiments, using a variable-temperature waste heat source, membrane loop
flow rates of 1 L/min, and distillate cooling source temperature of 20 °C. Thickness of vertical
gray lines at hour 18.6 identify extent of delay between peak waste heat source temperature
and peak water flux.
Variability in temperature and water flux data is quantified using standard deviation and is
shown in Table 3.2. In the indirect arrangement (Fig. 3.9b), water flux response to variability in
waste heat source temperature is dampened by 30.4% (Table 3.2) relative to the direct
arrangement (Fig. 3.9a). Dampening occurs in the indirect arrangement because the feed tank
60

acts as a buffer tank that allows mixing of the most recently heated solution in a large tank
volume to reduce high-frequency temperature variations [116, 117]. Lower feed-temperature
variability results in a more stable water flux profile in the indirect arrangement, allowing for
greater system control while also resulting in higher predictability of product water availability.
Table 3.2. Average temperatures and water flux, with standard deviation as a measure of
variability, for direct and indirect system arrangement experiments using a variable-
temperature waste heat source, 1 L/min membrane loop flow rates, and distillate cooling source
temperature of 20 °C.
Arrangement
Temperature (°C)
Water Flux
(L m
-2
h
-1
)
Waste Heat
Source Feed Inlet Feed Tank
Direct 71.6 ± 8.1 63.8 ± 6.3 58.4 ± 4.5 22.9 ± 6.0
Indirect 71.6 ± 8.1 58.4 ± 5.1 60.2 ± 5.3 18.7 ± 4.6

The indirect arrangement also provides a delayed response to changes in waste heat source
temperature. For example, the sudden change in waste heat source temperature (square-
shaped data points) at hour 18.6 in Fig. 3.9 causes an immediate water flux response in the
direct arrangement (Fig. 3.9a, thin vertical grey bar), but a delayed water flux response of 0.3 hr
in the indirect arrangement (Fig. 3.9b, thicker vertical grey bar). The delayed response occurs
due to non-ideal mixing in the feed tank, and can be used to improve system controllability by
providing a forecasting window that allows time for system software to adjust operating
variables (e.g., flow rates or number of membrane modules in use) to meet water production
goals. This delayed response would provide an additional degree of control to mitigate
expected decreases in water flux based on observed decreases in waste heat source
temperature. Variability dampening and the forecasting window can both be increased by
increasing feed tank volume in the indirect arrangement. Increasing mixing in the feed tank will
61

also increase variability dampening but will decrease the forecasting window. Ideally, feed tank
volume and degree of mixing would be adjusted to balance forecasting and variability
dampening capabilities when using the indirect arrangement. While the indirect arrangement
provides forecasting and variability dampening capabilities that are not available in the direct
arrangement, these capabilities come at the cost of 22.5% lower average water flux. The
tradeoff between better system control and higher water flux should be considered when
selecting an arrangement.
Results of the current study have shown that selection of a preferred arrangement for
particular waste heat source characteristics (i.e., intermittency and/or variability) will be a
function of water flux/production requirements and operator preferences regarding system
controllability and predictability of product water availability. However, cost considerations will
also affect the choice of arrangement. In terms of capital cost, the indirect arrangement
requires an additional pump and more piping. Energy consumption associated with operation
of the additional pump will be negligible compared to thermal energy consumption [62, 112],
especially given that additional pumping is only required when the waste heat source is on. The
slightly higher cost of the indirect arrangement can be justified in applications where better
heat storage, greater water production with highly intermittent heat sources, or better system
controllability is desired. Lastly, larger tanks required for heat storage or variability dampening
would also result in a marginally higher system cost and larger system footprint. The additional
cost and footprint of larger tanks would have to be weighed against the value of better system
control or additional energy extraction in a particular application.
62

3.4 Conclusions
Modeling and experimental analyses were used to evaluate MD system response to waste heat
intermittency and variability, revealing tradeoffs between water production and system
operation/control for direct and indirect system arrangements. The direct arrangement was
shown to achieve greater water flux when the waste heat source was on, and the indirect
arrangement was shown to achieve greater water flux when the waste heat source was off –
regardless of the degree of waste heat intermittency or variability. The indirect arrangement
operates at higher water flux when the waste heat source is off because it stores a greater
quantity of heat at a faster rate while the waste heat source is on. When operating with an
intermittent waste heat source, the indirect arrangement produces more water when the
degree of intermittency is high (e.g., 87.5%) and the direct arrangement produces more water
when the degree of intermittency is low (e.g., 12.5%). When using a variable-temperature
waste heat source, the direct arrangement achieved greater water flux, but the direct
arrangement provided a forecasting feature that improves system controllability and
predictability of product water availability. The indirect arrangement’s better system
controllability is further enforced by its greater dampening of water flux variability that results
from intermittency or variability in waste heat source temperature over time. Results from the
current study indicate that ideal MD system design will depend strongly on waste heat source
characteristics (i.e., intermittency and variability) and application-specific constraints regarding
system control and predictability of product water availability.

63

Chapter 4 A critical review of cooling in membrane distillation
4.0 Abstract
Research on energy needs in membrane distillation (MD) systems has focused almost
exclusively on the thermal energy required for heating. However, thermal energy for cooling
must also be supplied in order to provide the thermal gradient across the membrane that
drives water treatment in MD. The lack of information on cooling needs in MD systems is
addressed in this manuscript by presenting a review of the existing pilot-scale studies that used
realistic cooling approaches with low electrical energy consumption. A new linguistic
framework is proposed in which the cooling provided by heat recovery is referred to as
“replenishment cooling” and cooling approaches using at least one external fluid (e.g., once-
through, wet, and dry cooling) are referred to as “auxiliary cooling.” Factors that affect
selection of a cooling approach for a particular application were reviewed, including cooling
water availability, regulatory environment, water recovery, and cost. Experience of MD
researchers with these cooling approaches was critically assessed, identifying future research
opportunities on the need for (i) more detailed experimental data on operation with only
replenishment cooling in high overall water recovery systems, (ii) more detailed information on
cooling system sizing, (iii) justification for cooling approach selection, and (iv) an accurate
accounting of the relative amounts of cooling provided by replenishment and auxiliary cooling
in systems using both approaches.
4.1 Introduction
The energy required for MD system operation is primarily for heating and cooling. Electrical
heaters and chillers are often used in bench-scale studies of MD (e.g., [34, 88, 108, 113, 118])
64

because they are highly controllable. However, due to the high latent heat of vaporization of
water (2256 J/kg), an economically competitive MD system cannot use electrical energy to
provide the thermal energy needed for heating and cooling in a real-world system. The need for
heating has been addressed by the use of solar thermal energy [66, 68, 70, 71, 75, 103, 106,
118] or low-grade waste heat from industrial or power generation processes (e.g., boiler flue
gas or diesel engine coolant) [6-8, 45, 49, 51].
The energy needed for cooling and the methods for providing the required cooling in MD are
given very little attention in the existing MD literature. Many bench and pilot-scale studies of
MD are done using chillers or heat pumps that have high electrical energy consumption (e.g.,
[61-65]). Some pilot-scale MD systems have been shown to operate with realistic and low
electrical energy cooling approaches [6, 8, 45, 47, 48, 51, 66-72], but the role of cooling is given
very little attention in these studies. In most of these studies, justification of why a particular
cooling approach was selected is not provided and the performance of the cooling system is not
described. While multiple energy metrics are available to characterize energy consumption or
efficiency for heating [6, 7, 62], such metrics not regularly used for cooling.
The main objective of this review is to provide a critical evaluation of the role of cooling in pilot-
scale MD systems. First, an overview and critical analysis of factors that affect energy needs in
MD and metrics used for energy consumption/efficiency in MD is provided. Next, low electrical
energy cooling approaches used in previous pilot studies are placed within a new linguistic
framework and the operational experience with those cooling methods is critically evaluated.
Lastly, factors that affect selection of an appropriate cooling approach are discussed and future
research needs regarding cooling in MD are proposed.
65

4.2 MD system features that affect energy consumption
4.2.1 Heat recovery
Thermal efficiency of the MD configurations has been improved through the use of heat
recovery techniques, and the use of heat recovery techniques has also been noted to reduce
cooling needs [48]. As feed water passes the membrane in any MD configuration, the latent
heat of vaporization is absorbed from the feed water and transferred to the condensation side;
simultaneously, heat is transferred from the feed side to the condensation side due to
conduction through the membrane [108]. This behavior causes a decrease in the feed
temperature as it passes along the length of the membrane and an increase in the coolant or
distillate stream temperature as it passes through the coolant channel or along the other side
of the membrane.
The heat lost across the membrane can be recovered through external or internal heat
recovery. External heat recovery can be used with DCMD, VMD, and SGMD. In DCMD (Fig. 4.1),
a heat exchanger that is external to the membrane module is placed between the cool feed and
the warm distillate leaving the module, allowing heat lost to the distillate stream to be used to
preheat the feed before it is heated by the primary heat source [110, 119, 120]. In VMD and
SGMD, external heat recovery is implemented by using the incoming cool feed as the cooling
fluid in the condenser [121].
66

Figure 4.1: External heat recovery in direct contact membrane distillation (DCMD).
In AGMD (Fig. 4.2a) and PGMD, internal heat recovery is implemented by passing cool feed
water into the coolant channel where it absorbs heat through the condensation foil, then
passing the preheated feed through the heat source, and finally passing the fully heated feed by
the membrane. The AGMD or PGMD module integrated with internal heat recovery is the form
used in the majority of recent AGMD and PGMD studies [8, 22-24, 39, 48, 75], so all future
references to AGMD or PGMD should be assumed to use internal heat recovery, unless stated
otherwise. Internal heat recovery has also been integrated into a multistage V-AGMD module
design developed by Memsys and known as vacuum multi-effect MD (V-MEMD) (Fig. 4.2b).
Heat is recovered in V-MEMD by passing cool feed water through the coolant channel at one
end of the membrane module (the condenser), then passing the preheated feed through the
67

heat source, and finally passing the feed through a number of modules in series with each
successive module serving as coolant for the module before it [67, 69].

Figure 4.2: Internal heat recovery in (a) air gap membrane distillation (AGMD), and (b) vacuum
multi-effect membrane distillation (V-MEMD).
68

4.2.2 Feed stream configuration
In addition to the MD configuration selected for a system, the feed stream configuration and
the degree of overall water recovery (i.e., percent of feed water converted to product water)
must be selected. The feed stream can be arranged in a single-pass configuration or a
recirculating configuration, as shown for AGMD in Fig. 4.3. In a single-pass configuration (Fig.
4.3a), a given volume of feed water only passes through a membrane module once and brine is
continuously discharged, making it an inherently steady-state process when all inputs are
constant [73]. Single-pass water recovery is a maximum of 7-9% for any MD configuration, so a
staged module design (Fig. 4.3a) can be used to increase overall water recovery by passing the
brine from one module into the coolant channel of the next module in series [73, 111].
Interstage cooling can be used to maintain a high driving force in each successive module, as
shown in Fig. 4.3a, or the system can be operated without interstage cooling to provide a less
capital- and energy-intensive design with the tradeoff of a lower water production rate [111].
69

Figure 4.3: (a) Single-pass, and (b) recirculating feed stream configurations shown for air gap
membrane distillation (AGMD). Single-pass is shown with multiple stages in series to illustrate
the ability to operate at high overall water recovery.
In a recirculating feed stream configuration (Fig. 4.3b), a given volume of feed water passes
through the membrane module more than once by returning brine exiting the module to the
feed channel inlet after being reheated [122]. When using a recirculating feed stream
configuration, the system can be operated in batch, semi-batch, or partial brine discharge
70

mode. In batch mode, an initial feed volume is concentrated over time until the desired water
recovery is reached, the feed tank is emptied and replenished with a new volume of feed water,
and the process is repeated. In semi-batch mode, operation is the same as batch but the feed
tank is periodically or continually replenished with new feed water while the water is being
treated. If continuous replenishment is used in semi-batch mode, the replenishment rate is
equal to the water production rate so that the feed tank volume remains constant. If periodic
replenishment is used in semi-batch mode, the feed tank volume will increase or decrease over
time depending on the frequency and volume of replenishment compared to the water
production rate and operation time. Batch and semi-batch modes are inherently unsteady
processes, due to increasing feed concentration over time. In partial brine discharge mode,
operation is the same as semi-batch, but a portion of the brine is continuously discharged,
allowing for steady-state operation when using continuous replenishment and constant inputs.
4.3 Energy metrics in MD
Thermal energy consumption is the most dominant form of energy consumption in MD systems
[6, 62]. The thermal energy consumed from an external source is quantified in MD systems with
specific thermal energy consumption ( 𝑆𝑇𝐸𝐶 ) [6, 7, 62], but STEC has typically only been used to
quantify heating energy consumption in MD systems – neglecting cooling energy consumption.
Overall 𝑆𝑇𝐸𝐶 ( 𝑂𝑆𝑇𝐸𝐶 ) can be separated into 𝑆𝑇𝐸𝐶 for heating ( 𝑆𝑇𝐸𝐶 ℎ
) and 𝑆𝑇𝐸𝐶 for cooling
( 𝑆𝑇𝐸𝐶 𝑐 ) according to the following:
𝑂𝑆𝑇𝐸𝐶 = 𝑆𝑇𝐸𝐶 ℎ
+ 𝑆𝑇𝐸𝐶 𝑐 (1)
71

𝑆𝑇𝐸𝐶 ℎ
=
𝑄 ̇ ℎ
𝐽 𝑤 𝐴 𝑚
(2)
𝑆𝑇𝐸𝐶 𝑐 =
𝑄 ̇ 𝑐 𝐽 𝑤 𝐴 𝑚
(3)
𝑄 ̇ ℎ
= 𝑚 ̇ ℎ 𝑒 , 𝑜𝑢𝑡 ℎ
ℎ 𝑒 , 𝑜𝑢𝑡 − 𝑚 ̇ ℎ 𝑒 , 𝑖𝑛
ℎ
ℎ 𝑒 , 𝑖𝑛
(4)
𝑄 ̇ 𝑐 = 𝑚 ̇ 𝑐𝑒 , 𝑖𝑛
ℎ
𝑐𝑒 , 𝑖𝑛
− 𝑚 ̇ 𝑐𝑒 , 𝑜𝑢𝑡 ℎ
𝑐𝑒 , 𝑜𝑢𝑡 (5)
where 𝑄 ̇ is rate of thermal energy consumption (W), 𝐽 𝑤 is water flux (L·m
-2
h
-1
), 𝐴 𝑚 is membrane
area (m
2
), 𝑚 ̇ is mass flow rate (kg/s), ℎ is enthalpy (J·kg
-1
), and the subscripts ℎ, 𝑐 , ℎ 𝑒 , 𝑐𝑒 , 𝑖𝑛 ,
and 𝑜𝑢𝑡 stand for heating, cooling, heating equipment, cooling equipment, entering, and
exiting, respectively. The heating equipment is typically a heat exchanger in which the entering
and exiting mass flow rates are equal, but the cooling equipment may involve evaporation of
the stream being cooled (e.g., open-circuit wet cooling tower [8]) resulting in differences in
entering and exiting mass flow rates. While 𝑆𝑇𝐸𝐶 characterizes the thermal energy
consumption of an MD system, gained output ratio ( 𝐺𝑂𝑅 ) has been used to characterize the
thermal energy efficiency [8, 22]. 𝐺𝑂𝑅 is defined as follows:
𝐺𝑂𝑅 =
𝐽 𝑤 𝐴 𝑚 ∆ 𝐻 𝑣 𝑄 ̇ ℎ

(6)
where ∆ 𝐻 𝑣 is the latent heat of vaporization (J/kg).
Eqs. 1 – 6 apply to steady-state operating conditions or to an instant in time when a system is
operating under unsteady conditions. Because systems operating with batch and semi-batch
feed stream configurations with non-continuous replenishment always operate under unsteady
conditions, the 𝑆𝑇𝐸𝐶 and 𝐺𝑂𝑅 in these systems will change over the cycle time [73], so
average 𝑆𝑇𝐸𝐶 and 𝐺𝑂𝑅 are used to characterize the cycle :
72

𝑆𝑇𝐸𝐶 ̅ ̅ ̅ ̅ ̅ ̅ ̅
ℎ
=
∫ 𝑄 ̇ ℎ
( 𝑡 )𝑑𝑡 𝜏 0
∫ 𝐽 𝑤 (𝑡 )𝐴 𝑚 𝑑𝑡 𝜏 0

(7)
𝐺𝑂𝑅 ̅ ̅ ̅ ̅ ̅ ̅
=
∫ 𝐽 𝑤 (𝑡 )𝐴 𝑚 ∆ 𝐻 𝑣 (𝑡 )𝑑𝑡 𝜏 0
∫ 𝑄 ̇ ℎ
(𝑡 )𝑑𝑡 𝜏 0

(8)
where 𝑡 is time (s) and 𝜏 is cycle time (s).
Although it is possible for a price to be charged for waste heat [50], the cost of thermal energy
for heating in MD systems mostly derives from the capital cost of the heating equipment. To
minimize the capital cost of heating equipment, the size of heating equipment (e.g., heat
exchanger or solar collector size) must be minimized by minimizing 𝑆𝑇𝐸𝐶 ℎ
. 𝑆𝑇𝐸𝐶 ℎ
minimization
can be achieved by optimizing membrane characteristics, membrane module design, and heat
recovery design [10]. Electrical energy consumption associated with heating arises from the
pressure drop across the heat exchanger that must be overcome by the feed circulation pump
according to [62]:
𝐸 ̇ 𝑝 =
𝑉 ̇ 𝑤 ∆ 𝑃 𝜂 𝑝
(9)
where 𝐸 ̇ 𝑝 is electrical energy consumed for pumping (W), 𝑉 ̇ 𝑤 is volumetric flow rate of water
(m
3
/s), ∆ 𝑃 is pressure drop (Pa), and 𝜂 𝑝 is the pump efficiency. Energy and capital costs will also
be incurred for heat source fluids that need to be circulated on the heat-source side of the heat
exchanger. For example, a solar thermal system that requires circulation of the heat transfer
fluid or a liquid waste heat source that must be pumped from a reservoir and through the heat
exchanger. If external heat recovery is used, capital costs will be associated with the heat
exchanger size and electrical energy costs will be associated with the pressure drop across the
73

heat exchanger. Because heat recovery facilitates recycling of heat within the MD system, no
𝑆𝑇𝐸𝐶 is associated with it. As with 𝑆𝑇𝐸𝐶 ℎ
, 𝑆𝑇𝐸𝐶 𝑐 must also be minimized in order to minimize
cooling system size (e.g., cooling tower size or heat exchanger size) for minimizing system cost.
4.4 Cooling approaches used in existing MD systems
A new linguistic framework is proposed to describe the cooling approaches used in MD.
Replenishment cooling is proposed to describe the cooling that occurs due to heat recovery and
auxiliary cooling is proposed to describe cooling that is achieved by use of at least one fluid
(e.g., water or air) that will not be used as the feed solution in the MD module (e.g., once-
through, wet, or dry cooling). The performance of replenishment cooling and different types of
auxiliary cooling approaches that have been previously used in pilot-scale MD systems is
described in Sections 4.4.1 and 4.4.2.
4.4.1 Replenishment cooling
4.4.1.1 Theory and energy consumption
Replenishment cooling occurs when heat recovery is used with a cool feed water source,
resulting in heat being recycled within the MD system rather than needing to be rejected to the
surroundings by use of an auxiliary cooling approach. Replenishment cooling can be
implemented with a single-pass or recirculating feed stream configuration and can be used with
or without auxiliary cooling, as shown for AGMD in Fig. 4.4. Because the feed stream serves as
the coolant, an additional coolant pump is not required for replenishment cooling and the feed
circulation pump provides the necessary circulation. The pumping energy associated with the
pressure drop across the replenishment cooling heat exchanger or channel can be thought of as
74

energy consumption required for heat recovery, so that no electrical energy is consumed for
replenishment cooling.

Figure 4.4: Replenishment cooling implemented with air gap membrane distillation (AGMD) in
(a) single-pass, and (b) recirculating feed stream configurations.
Although replenishment cooling can be used in conjunction with auxiliary cooling approaches,
the systems discussed in Sections 4.4.1.2-4.4.1.4 are those that did not also use auxiliary
cooling. Because MD system operation with a single-pass feed stream configuration is simpler
than with a recirculating feed stream configuration, systems using a single-pass configuration
are discussed first, followed by those using a recirculating configuration in semi-batch and
partial brine discharge modes. The pilot-scale MD systems that have used only replenishment
cooling are summarized in Table 4.1 and described in detail in Sections 4.4.1.2-4.4.1.4.

75

Table 4.1: Summary of pilot-scale membrane distillation systems that use replenishment cooling
without auxiliary cooling.
Citation
MD
Configuration
Feed Stream
Configuration Operation Mode Location
Banat et al. [66] PGMD Single-Pass n/a Jordan
Hagedorn et al. [48]
("Helgoland” System)
PGMD Single-Pass n/a Germany
Andrés-Mañas et al. [67] V-MEMD Single-Pass n/a Spain
Raluy et al. [68] PGMD Recirculating Semi-Batch Spain
Hagedorn et al. [48]
(“Pre-Industrial” System)
DCMD Recirculating Partial Brine Discharge Ocean

4.4.1.2 Single-pass feed stream configuration
Pilot MD systems that have been operated using replenishment cooling in a single-pass feed
stream configuration include systems by Banat et al. [66] and Hagedorn et al. [48]. Banat et al.
[66] operated an AGMD seawater desalination pilot plant with solar thermal and photovoltaic
energy (Fig. 4.5), using four membrane modules in parallel, each with 10 m
2
membrane area.
While no direct discussion of cooling was provided, data from the first few months of operation
was presented, showing successful operation using replenishment cooling with average daily
water flux between 6 and 26 L·m
-2
h
-1
. Hagedorn et al. [48] operated a PGMD seawater
desalination pilot plant (“Helgoland” system) using an electrical heater for heat energy and two
membrane modules in parallel (Fig. 4.6), each with 4.5 m
2
membrane area. System operation
was characterized with hot feed water and cool seawater temperature profiles for most of a
300-day operation period and a relatively stable water production rate of approximately 0.5
m
3
/day from each module for the last approximately 100 days. Seawater temperature was
shown to vary between approximately 5 and 19 °C, likely due to seasonal variation, but this
variation did not affect the water production rate in the data shown because an electric heater
76

was used to maintain a constant hot feed temperature. The Helgoland pilot study by Hagedorn
et al. [48] and the pilot study by Banat et al. [66] both indicate successful operation of a pilot-
scale MD system using only replenishment cooling in a single-pass feed stream configuration,
but these systems provide very little detail regarding the control and effects of cooling on
system performance.

Figure 4.5: Diagram reproduced from Banat et al. [66] showing replenishment cooling used in a
pilot-scale permeate gap membrane distillation (PGMD) system with a single-pass feed stream
configuration.
77

Figure 4.6: Diagram reproduced from Hagedorn et al. [48] showing replenishment cooling used
in a pilot-scale permeate gap membrane distillation (PGMD) system with a single-pass feed
stream configuration. Numbers in diagram identify system components: (1) Feedwater tank, (2)
sulfuric acid dosing unit, (3) feed pump, (4) flow meter, (5) gravel filter, (6) cartridge filter, (7)
automatic backflush filter, (8) MD modules, (9) freshwater tank, (10) freshwater pump, and (11)
heat exchanger.
Andrés-Mañas et al. [67] operated a pilot-scale MD system for seawater desalination using
replenishment cooling in a single-pass feed stream configuration, but also provided detailed
information regarding the role of cooling in the operation of their system. The system used a V-
MEMD module with 6.4 m
2
of membrane area and was driven by solar thermal heat (Fig. 4.7).
One of the unique aspects of this system was the use of a much higher replenishment flow rate
than was necessary solely to maintain flow in the membrane module. This high replenishment
flow rate was passed through the condenser to accomplish heat recovery and to maximize
cooling for vapor condensation. Because the flow through the condenser was much larger than
necessary for flow through the membrane module, 50-80% of the flow leaving the condenser
was discharged back to the sea, while the rest of the pre-heated solution was sent to the solar
78

heat exchanger and then the membrane module. This “high-rate” replenishment cooling
method helps maximize cooling to maximize water production but it also results in most of the
recovered heat being disposed of, although this drawback was not discussed.

Figure 4.7: Diagram reproduced from Andrés-Mañas et al. [67] showing high-rate replenishment
cooling used in a pilot-scale vacuum multi-effect membrane distillation (V-MEMD) system with a
single-pass feed stream configuration.
Despite the unique approach to maximizing cooling used by Andrés-Mañas et al. [67], the
authors concluded that the main limitation of the system was its limited cooling capacity. The
authors explained that the cooling capacity becomes limited when using a combination of high
feed hot temperature and high feed flow rate. Under these conditions, too much water vapor is
produced in the condenser, increasing the pressure in the condenser, which limits the vacuum
79

suction. Because flow of feed, and therefore replenishment cooling water, is determined by the
degree of vacuum suction, the lower suction at higher feed temperatures and flow rates results
in lower cooling flow and therefore decreased water flux. The authors proposed methods for
improving the cooling performance consisting of using a circulation pump separate from the
vacuum pump for the feed/cooling stream and increasing the condenser heat exchanger area.
Although seawater cooling source temperature is a constraint that cannot be controlled
(average of 20 °C during the tests shown in the study), the authors also demonstrated 44%
greater water flux when using a lower seawater temperature (16 °C) compared to using a
higher seawater temperature (24 °C). Despite complications with cooling at the extremes of
system operation capacity, the system was operated successfully between 2.4 and 8.5 L·m
-2
h
-1

over a range of feed hot temperatures between 60 and 80 °C and flow rates between 90 and
180 L/h.
4.4.1.3 Recirculating feed-stream configuration in semi-batch mode
Raluy et al. [68] operated a PGMD seawater desalination pilot plant using periodic
replenishment cooling with a recirculating feed stream configuration in semi-batch mode (Fig.
4.8). The system used solar thermal energy to supply heat to a single PGMD module with either
8.5 m
2
membrane area (2005 and 2008 data) or 10 m
2
membrane area (2009-2010 data). The
authors described a process of operation in which the 500 L feed tank is filled with an initial
batch volume of cool seawater each day, this volume is treated throughout the day, and then
the tank is emptied after the day’s operation is completed. When operating the system in 2005
and 2008, the feed tank was typically replenished once per day, but the authors indicated that
the increasing tank temperature caused by feed recirculation depleted the driving force so
80

much that it resulted in inefficient operation. Considering this, the authors changed the system
to allow for 3-4 replenishments with cool seawater per day after changing the membrane
module in 2009 so that the driving force could be maintained at a higher value. The change in
replenishment schedule was accompanied by significant improvements in the system’s
performance – resulting in the lowest average distillate conductivity and the highest water
production and energy efficiency out of all years of operation. While the authors did attribute
the improved performance in the 2009-2010 data to the new replenishment schedule, it should
be noted that the average solar irradiation in 2009-2010 was also 4.2 and 13.2% higher than in
2005 and 2008, respectively, and these differences in solar irradiation may also have
contributed to the improved performance in 2009-2010. Additionally, the authors did briefly
note that small amounts of water were removed from the tank at times to control the salinity
of the feed solution, but they did not indicate the frequency, rate, or volume of this feed tank
brine discharge, and they did not indicate how this brine discharge may have affected the
system’s performance. The system was successfully operated for approximately 8-9 hrs on
sunny days and fewer hours on cloudy days, with average water production of 47, 44, and 76
L/day in 2005, 2008, and 2009-2010, respectively.
81

Figure 4.8: Diagram reproduced from Raluy et al. [68] showing replenishment cooling used in a
pilot-scale permeate gap membrane distillation (PGMD) system with a recirculating feed stream
configuration operated in semi-batch mode.
An important lesson learned from the study by Raluy et al. [68] is that the cooling needed to
operate an MD system in batch/semi-batch treatment mode can simply be contained within a
large initial volume of cool feed water. While the study indicated that a greater frequency or
rate of replenishment with cool feed water will result in greater water production, it is also
clear that a larger and lower-temperature initial volume will result in greater water production.
The results of the study indicated that the cooling provided by the initial volume of cool feed
water will eventually be depleted, so it is important to note that as membrane area increases
for the same tank volume, the cooling inherent to the initial volume will be depleted more
quickly. Replenishment with cool feed water can be used to decrease the rate of cooling
depletion, but the authors note that higher frequency of replenishment also results in the
tradeoff of lower water recovery.
82

4.4.1.4 Recirculating feed stream configuration in partial brine discharge mode
Hagedorn et al. [48] operated a DCMD seawater desalination pilot plant (named the “pre-
industrial” system) that used continuous replenishment cooling with a recirculating feed stream
configuration in partial brine discharge mode (Fig. 4.9). The system used 27.5 m
2
membrane
area and was designed to be driven by waste heat from diesel generators on board a cargo
ship. The system was designed with a valve that can be used to adjust the brine recycle rate,
allowing water recovery to be controlled and allowing the system to be operated with partial
brine discharge (i.e., partial brine recycle) or in a single-pass configuration (i.e., without brine
recycle). A valve was also installed in the waste heat line to adjust the waste heat flow rate so
that a constant membrane module feed inlet temperature could be achieved. The brine recycle
valve can be adjusted to provide a degree of control over the “feed mix” temperature by mixing
warm brine with cool replenishment seawater (see Fig. 4.9). Because the feed mix temperature
is also the cooling source temperature for the distillate stream, controlling the feed mix
temperature also results in a degree of control over the distillate inlet temperature. However,
this degree of control is limited to temperatures between the brine and seawater
temperatures.
83

Figure 4.9: Diagram reproduced from Hagedorn et al. [48] showing replenishment cooling used
in a pilot-scale direct contact membrane distillation (DCMD) system with a recirculating feed
stream configuration operated in partial brine discharge mode. PT: pre-treatment consisting of
backwash filter, cartridge filter, and antiscalant dosing unit, RX: heat exchanger for external
heat recovery, HX: heat exchanger for the waste heat source.
While the authors state that the combination of brine recycle and replenishment cooling shown
in Fig. 4.8 “has been tested and feasibility was proven,” graphs or tables of data under different
operating conditions were not presented. Instead, a single scenario was described in which a
brine recycle ratio of 0.19 was used. In this scenario, brine recycle caused a slightly elevated
feed mix temperature above the replenishment seawater temperature of 20.6-25.0
o
C, but
water production only dropped from 2.1 m
3
/day without brine recycle to 2.0 m
3
/day with brine
recycle. Further, the brine recycle scenario was said to have reduced STEC
h
by 8% compared to
operation with a 19 °C feed mix temperature and no brine recycle. Although no further
experimental results were presented for operation with brine recycle, detailed modeling results
were presented for the brine recycle scenario.
84

Modeling results for water production rate, STEC
h
, and water recovery were presented as a
function of seawater temperature with brine recycle adjusted to achieve a feed mix
temperature of 30
o
C. This modeling data showed that the system could maintain constant
water production rate and STEC
h
with a seawater temperature up to 30
o
C, at the expense of
decreasing water recovery because less brine could be recycled as the seawater temperature
was increased. For seawater temperatures above 30
o
C, no brine is recycled (configuration
becomes single-pass) and the feed mix temperature becomes equal to the increasing seawater
temperature, causing the water production rate to drop and STEC
h
to increase. The decreasing
water production rate and increasing STEC
h
with increasing seawater temperature when
operating the system without brine recycle (i.e., single-pass configuration) was also
demonstrated with experimental data over a seawater temperature range of 14 to 33
o
C.
4.4.2 Auxiliary Cooling
Auxiliary cooling approaches that have been used in existing MD systems include once-through,
dry, open-circuit wet, and closed-circuit wet cooling. Some of the systems that use auxiliary
cooling also use supplementary cooling that can be accomplished by one of the other auxiliary
cooling approaches or through replenishment cooling. A summary of the pilot-scale MD
systems that use auxiliary cooling approaches is shown in Table 4.2, and these systems are
discussed in-detail in Sections 4.4.2.1-4.4.2.4.
85

Table 4.2: Summary of pilot-scale membrane distillation systems that use auxiliary cooling.
Cooling Approach Citation
MD
Configuration
Feed Stream
Configuration
Operation
Mode Supplementary Cooling Location
Once-Through Zhao et al. [69] V-MEMD Single-pass n/a None Unspecified
Once-Through Dow et al. [6] DCMD Recirculating Partial Brine
Discharge
None Australia
Once-Through Xu et al. [45] VMD Recirculating Semi-Batch None Ocean
Dry Dow et al. [51] DCMD Recirculating Unspecified Once-Through Australia
Dry Guillén-Burrieza et al.
[70]
AGMD (Without
Heat Recovery)
Recirculating Batch None Spain
Dry Guillén-Burrieza et al.
[71]
AGMD Recirculating Batch None Spain
Open-Circuit Wet Schwantes et al. [8]
("Namibia System")
PGMD Recirculating Partial Brine
Discharge
Replenishment Namibia
Closed-Circuit Wet Kabeel et al. [72] DCMD Recirculating Batch None Egypt
Closed-Circuit Wet Schwantes et al. [8]
("Pantelleria System")
PGMD Recirculating Partial Brine
Discharge
Once-Through
Replenishment
Italy
Closed-Circuit Wet Schwantes et al. [8]
("Gran Canary System")
PGMD Recirculating* Partial Brine
Discharge
Once-Through
Replenishment
Spain
Closed-Circuit Wet Wang et al. [47] DCMD Recirculating Partial Brine
Discharge
None China
*(data only presented for single-pass)
86

4.4.2.1 Once-through cooling
4.4.2.1.1 Theory and energy consumption
Once-through cooling transfers sensible heat through conduction and convection from a warm
process water stream to a cooler water stream from outside of the system by use of a heat
exchanger (Fig. 4.10) [123, 124]. Electrical energy is consumed due to the pumping energy
required to overcome the pressure drop across the heat exchanger, as described by Eqn. 9.
Advantages of once-through cooling include the use of a generally lower temperature cooling
source and greater simplicity than wet or dry cooling, which allows once-through cooling
systems to provide the better performance than wet and dry cooling [123]. Additionally, once-
through cooling has lower water consumption than wet cooling [125]. Disadvantages include
much greater impingement and entrainment of aquatic life on intake structures, more thermal
pollution, and much greater water withdrawals than wet and dry cooling [124, 125].

Figure 4.10: Diagram of once-through cooling process.
87

4.4.2.1.2 Usage in existing MD systems
Zhao et al. [69] operated a V-MEMD pilot system that used once-through cooling with a single-
pass feed stream configuration (Fig. 4.11). The system was tested in the field using solar
thermal collectors to provide the heat energy and was also tested in the lab using a diesel
heater to provide the heat energy. The solar-thermal-driven installation was used for seawater
desalination and used membrane area of 5 m
2
. Freshwater was circulated through the
condenser in a recirculating loop and seawater was circulated through a heat exchanger
connected to that loop to provide once-through cooling. Graphs of temperature and pressure in
each stage were provided for a 280-min operating period and the average water flux for this
period was stated to be approximately 7 L·m
-2
h
-1
. Variability in the temperatures and pressures
was said to be caused by variability in heat input due to normal diurnal solar variation and due
to passage of clouds over the solar thermal collectors.
88

Figure 4.11: Diagram reproduced from Zhao et al. [69] showing once-through cooling used in a
pilot-scale vacuum multi-effect membrane distillation (V-MEMD) system with a single-pass feed
stream configuration.
The influence of cooling source characteristics on performance of the solar-thermal-driven
installation was not discussed, but a detailed analysis of cooling source characteristics was
performed for the diesel-heater-driven installation in the lab. Tests evaluating cooling in the lab
system were done by controlling the inner heating and cooling loop fluid characteristics shown
in Fig. 4.9: heating source temperature (T1_2 in Fig. 4.11) and flow rate (F1 in Fig. 4.11), and
cooling source temperature (T6_1 in Fig. 4.11) and flow rate (F6 in Fig. 4.11). A test was
performed with constant heat source temperature (60 °C) and flow rate (15 L/min), constant
cooling source flow rate (9 L/min), and constant feed inlet temperature (25 °C), while adjusting
cooling source temperature. Decreasing cooling source temperature from 40 to 25 °C (37.5%
decrease) resulted in increasing water flux from 1.5 to 2.8 L·m
-2
h
-1
(86.7% increase). GOR
increased from 2.1 to 2.8 (33.3% increase) over the same temperature range, likely due to
89

water flux increasing at a faster rate than heat consumption. While methods for increasing GOR
in MD systems tend to focus on increasing heat recovery or improving module design, the
results presented by Zhao et al. [69] indicate that control of cooling temperature – if possible –
can be another useful method for controlling GOR.
Zhao et al. [69] also performed a test with constant heat source temperature (60 °C) and flow
rate (15 L/min), constant cooling source temperature (25 °C), and constant feed inlet
temperature (25 °C), while adjusting cooling source flow rate. Water flux increased only
moderately from 2.7 to 3.4 L·m
-2
h
-1
(25.9% increase) with a large increase in cooling source flow
rate from 6 to 16 L/min (167% increase), although the water flux curve had already plateaued
at 3.4 L·m
-2
h
-1
upon reaching a cooling source flow rate of 12 L/min, indicating diminishing
returns at higher flow rates. Diminishing returns were also observed with regard to GOR, which
increased from 3.0 to 3.4 over the same temperature range and also plateaued around 12
L/min. These results indicate that cooling source temperature has a much larger influence over
water flux and GOR than cooling source flow rate. However, cooling source flow rate is much
easier to control than cooling source temperature, and care must be taken when adjusting
cooling source flow rate for improving flux or GOR because of the risk for diminishing returns.
Dow et al. [6] operated a DCMD pilot system for treatment of ion exchange column
wastewater, using once-through cooling with saline estuary water (Fig. 4.12). A recirculating
feed stream configuration was used in partial brine discharge mode to allow the system to
reach high water recovery (85-90%). Heat energy was provided by low-grade waste heat from a
natural-gas-fired power plant and the system used membrane area of 0.67 m
2
. The system was
operated over an 80-day period, with water flux ranging mostly between 2 and 5 L·m
-2
h
-1
.
90

Increasing ambient temperature over time resulted in increasing estuary temperature (i.e.,
cooling source temperature) over time, resulting in a slight increase in distillate inlet
temperature and a slight decrease in water flux over time during periods when feed inlet
temperature was relatively constant. The authors also noted sections with variability in feed
inlet temperature caused by variability in waste heat source temperature. These results
highlight one of the main hurdles that MD system designers face: developing a system design
that can overcome variability in heating and cooling source temperatures in a simple and cost-
effective manner to provide predictable system performance at all times. Another issue
concerning MD system design related to cooling that was raised in the study by Dow et al. [6]
was corrosion resistance of the cooling heat exchanger. After 36 days of operation, the copper-
brazed cooling heat exchanger failed due to corrosion, allowing saline estuary water to enter
the distillate side of the DCMD system. The heat exchanger failure experienced by Dow et al. [6]
highlights one of the tradeoffs of using relatively simple once-through cooling with saline water:
the heat exchanger must be constructed from more expensive corrosion-resistant materials like
titanium (e.g., [70, 101]).
91

Figure 4.12: Diagram reproduced from Dow et al. [6] showing once-through cooling used in a
pilot-scale direct contact membrane distillation (DCMD) system with a recirculating feed stream
configuration operated in partial brine discharge mode.
The system used by Xu et al. [45] also demonstrated operation with once-through water
cooling, but provided much less detail regarding operation of the cooling portion of the system;
this system is briefly described to provide a full assessment of how cooling has been
implemented in pilot-scale MD systems. Xu et al. [45] operated a VMD seawater desalination
pilot system on a ship, driven by waste heat from diesel engines, with a recirculating feed
stream configuration. The system was operated in semi-batch mode with membrane area of
12.3 m
2
. The only data provided was a graph of water flux as a function of feed inlet
temperature, showing water flux increasing from 3.0 L·m
-2
h
-1
with an inlet temperature of 45 °C
to 4.2 L·m
-2
h
-1
with an inlet temperature of 55 °C. No information was given regarding the
effect of cooling system operation on MD system performance.
92

4.4.2.2 Open-circuit wet cooling
4.4.2.2.1 Theory and energy consumption
In an open-circuit wet cooling tower (Fig. 4.13), the warm water stream to be cooled is sprayed
downward over a heat exchange surface called “fill” and dry air is passed through the fill. Air
flow in the opposite direction of the water flow is called “counterflow” (Fig. 4.13) and air flow
perpendicular to the water flow direction is called “crossflow”. Air flow aided by the use of a
fan is known as “mechanical draft”, which can be implemented using “forced draft” with the
fan at the inlet of the air or “induced draft” (Fig. 4.13) with the fan at the outlet of the air. Air
flow could also be induced through the use of large towers that rely on the chimney effect,
known as “natural draft” towers, but these towers must be hundreds of feet tall to create a
suitable draft, so mechanical draft towers are generally preferred based on capital cost and
aesthetic considerations [124]. When using mechanical draft, electrical energy is consumed by
the fan according to:
𝐸 ̇ 𝑓 =
𝑉 ̇ 𝑎 ∆ 𝑃 𝜂 𝑓
(10)
where 𝐸 ̇ 𝑓 is electrical energy consumed for fan operation (W), 𝑉 ̇ 𝑎 is volumetric flow rate of air
(m
3
/s), ∆ 𝑃 is pressure drop (Pa), and 𝜂 𝑓 is the fan efficiency. Because pumps are already used
to circulate water through the MD system, pumping energy associated with the cooling tower is
only derived from the additional pressure drop across the cooling tower, and can be calculated
from Equation 9.
93

Figure 4.13: Diagram of open-circuit wet cooling tower using mechanical draft.
In the open-circuit wet cooling tower, the water and air are brought into direct contact with
each other, and cooling occurs mostly by latent heat transfer due to evaporation and to a lesser
degree by sensible heat transfer [126, 127]. Latent heat is transferred from the warm water to
evaporate part of the water, and this absorption of latent heat causes the temperature of the
remaining water to decrease. The evaporated water, which contains the latent heat transferred
from liquid water, is entrained by the air passing through the tower, causing the air to leave the
tower with higher humidity. Sensible heat transfer also occurs due to conduction and
convection from warm water to cooler air. The air enters the cooling tower with a dry bulb
temperature and a relative humidity less than 100%, providing the medium for water to be
94

evaporated into. Wet cooling allows the process water to approach the wet bulb temperature
of the incoming air, which is lower than the dry bulb temperature of the incoming air [123].
As a given volume of water is recirculated through the cooling tower, the total dissolved solids
in the tower will increase over time due to the loss of water, requiring a portion of the water to
be discharged from the tower in a process known as “blowdown.” To account for water lost
from the system due to blowdown, evaporation, and drift (i.e., liquid water entrained in air
leaving the tower), makeup water is added to the tower periodically [124]. Advantages of wet
cooling towers in general include much less water withdrawals, impingement and entrainment
of aquatic life, and thermal pollution concerns than once-through cooling [123, 128, 129].
Disadvantages of wet cooling towers in general include higher water consumption (due to
evaporation) than once-through cooling and higher capital and variable costs than once-
through cooling [123, 128, 129].
4.4.2.2.2 Usage in existing MD systems
Schwantes et al. [8] operated a pilot-scale PGMD system with an open-circuit cooling tower in a
remote region of Amarika, Namibia. The system was used for brackish water desalination,
obtaining heat energy from solar thermal collectors and electrical energy from solar
photovoltaic collectors (Fig. 4.14). The system used 168 m
2
membrane area and was operated
with a recirculating feed stream configuration in partial brine discharge mode.
95

Figure 4.14: Diagram reproduced from Schwantes et al. [8] showing cooling with an open-circuit
cooling tower used in a pilot-scale permeate gap membrane distillation (PGMD) system.
Data for a day of the plant’s operation was presented, showing some variation in water
production rate due mostly to variation in solar energy input throughout the day. The system
was operated with a setpoint heat source temperature of 75 °C, which resulted in a feed hot
inlet temperature of approximately 73 °C for the majority of the system’s operating time;
although, feed hot temperature decreased during the evening as the supply of stored heat was
consumed. This allowed the system to produce water at approximately 150 L/h for the majority
of the operating time. The cooling system appears to have provided sufficient cooling capacity
because the cold feed inlet temperature remained very close to 25 °C throughout the day, with
only minor variation between approximately 22 and 28 °C. However, cooling was also
contributed by feed water replenishment. The degree to which replenishment contributed to
cooling was not made clear because the mode (continuous vs. periodic), frequency, and rate of
replenishment and brine discharge were not indicated. It is important to note that the
contribution of replenishment to cooling in an MD system must be separated from the
96

contribution of auxiliary cooling in order to properly assess the cooling performance of the
auxiliary cooling approach. Otherwise, a high brine discharge rate – and correspondingly high
replenishment rate – can be used to provide a large degree of cooling in an MD system with a
recirculating feed stream configuration, negating the main purpose of recirculation: high water
recovery.
The use of an open-circuit wet cooling tower will also contribute to faster concentration of the
feed solution, which is beneficial for brine volume minimization but also results in partial loss of
the pure-water fraction of the feed water brought into the MD system. While the direct use of
brine in the cooling tower minimizes the cooling system size, the cooling tower fill will scale
more quickly than a traditional cooling tower using water with lower total dissolved solids and
therefore need to be operated with lower concentration cycles (i.e., lower water recovery and
higher brine discharge) [130, 131]. However, this interplay between feed concentration, water
recovery, and brine discharge/blowdown was not discussed. Additionally, direct contact
between MD system water and cooling air may expose the MD system water to environmental
contaminants that will foul system components, but this was also not discussed.
4.4.2.3 Closed-circuit wet cooling
4.4.2.3.1 Theory and energy consumption
A closed-circuit wet cooling tower operates in the same way as an open-circuit wet cooling
tower, but the cooling air does not come into contact with the process water when using a
closed-circuit [132-134]. The process water is kept separate from the cooling air by use of a
heat exchanger (Fig. 4.15), which can be internal (Fig. 4.15a) or external (Fig. 4.15b) to the
cooling tower. Closed-circuit cooling towers have the same advantages and disadvantages as
97

open-circuit wet cooling compared to once-through and dry cooling, but there are also
advantages and disadvantages of closed-circuit wet cooling compared to open-circuit. The
advantage of closed-circuit over open-circuit is the separation of process water from direct
contact with potentially debris- or contaminant-carrying air, but the disadvantage is the higher
capital cost due to the large heat exchanger area required to achieve the same cooling effect as
open-circuit [133, 134]. In a conventional closed-circuit wet cooling tower (Fig. 4.15), a separate
water stream from the one being cooled is provided for evaporation. In some newer closed-
circuit wet cooling systems [8, 135], part of the process stream being cooled is used for
evaporation. In a conventional closed-circuit wet cooling tower, electrical energy is consumed
for operation of the cooling water pump and for overcoming the pressure drop associated with
the cooling tower heat exchanger in the process water loop; energy consumption for these
pumps is dictated by Equation 9. Electrical energy consumption associated with fan operation is
dictated by Equation 10.

Figure 4.15: Diagram of conventional closed-circuit wet cooling tower using (a) internal heat
exchanger, and (b) external heat exchanger with mechanical draft.
98

4.4.2.3.2 Usage in existing MD systems
Kabeel et al. [72] operated a pilot-scale DCMD system using a closed-circuit wet cooling tower
with a recirculating feed stream configuration (Fig. 4.16). The system was operated in batch
mode, used 1 m
2
membrane area, and obtained heat energy from solar thermal collectors. The
feed solution was a synthetic salt solution with 10 g/L salinity, but the specific solutes
contributing to salinity were not indicated. Air temperature and solar radiation data were
presented for 4 separate days in the month of June from 8 AM to 7 PM. Air temperature was
measured on 2 of the 4 days, and these temperature distributions were almost identical,
ranging between approximately 30 and 40 °C throughout the day; solar radiation data
presented for the other 2 days were almost identical as well.

Figure 4.16: Diagram reproduced from Kabeel et al. [72] showing a pilot-scale air gap
membrane distillation (AGMD) system operating with a recirculating feed stream configuration
and using a closed-circuit wet cooling tower.
99

Experimental data was presented for MD system operation with two feed flow rates when
using the cooling unit and when not using the cooling unit. These four sets of experimental
conditions tested are assumed to have been subject to the same air temperature and solar
radiation distributions (those previously discussed), but the authors did not clearly indicate this
in the study. The different feed flow rates were shown to have a negligible impact on the
distillate inlet temperature, but operation with the cooling system was shown to provide 10 to
17 °C lower distillate inlet temperature than when operating without the cooling system. The
MD system was still able to produce water without the cooling system because a large (250 L)
distillate tank volume – initially filled with cool water – provided an additional source of cooling.
However, the distillate inlet temperature distributions without the cooling system still show the
distillate inlet temperature decreasing over time after approximately 5 hours of operation.
Comparison of the air temperature distribution to distillate inlet temperature distribution
reveals that air temperature is always cooler than distillate inlet temperature after 5 hours of
operation, indicating that the surrounding air acted as an additional source of cooling, although
this was not discussed by the authors.
Schwantes et al. [8] operated a pilot-scale PGMD system for seawater desalination using a
combination of closed-circuit wet cooling and once-through cooling (Fig. 4.17). The system,
known as the “Pantelleria” system, used 120 m
2
membrane area and heat energy was provided
by waste heat from a diesel generator. A non-conventional closed-circuit wet cooling approach
known as a brine evaporative cooler condenser (BECC) was used; the design and operation of
the BECC system is described in-detail by Cipollina et al. [135]. The BECC (Fig. 4.18) operates by
splitting the warm brine leaving the MD module into a bypass brine stream and a recirculating
100

brine stream. The recirculating brine stream is circulated through a channel that is sealed-off
from the surroundings by thin plastic sheets. The bypass brine flows through channels on the
other sides of the plastic sheets and dry air is also blown through these channels, which causes
bypass brine to evaporate into the air from the plastic sheet surface, providing evaporative
cooling to the recirculating brine on the other side of the sheet. The bypass brine leaving the
BECC is disposed of while the cooled recirculating brine stream is returned to the feed tank for
further treatment. The use of brine from the MD system as the source of water for evaporation
eliminates the need for the secondary cooling water stream that was required in the
conventional closed-circuit cooling tower design used by Kabeel et al. [72].

Figure 4.17: Diagram reproduced from Schwantes et al. [8] showing a pilot-scale permeate gap
membrane distillation (PGMD) system using a non-conventional closed-circuit cooling tower
design known as a brine evaporative cooler condenser.
101

Figure 4.18: Diagram reproduced from Cipollina et al. [135] of non-conventional closed-circuit
wet cooling tower design known as a brine evaporative cooler condenser.
Purported benefits of the BECC design include corrosion resistance due to the use of plastic
materials, ease of scale and/or fouling removal due to the planar design and use of plastic
materials, and system modularity. Lab testing of the design was performed by Cipollina et al.
[135] with freshwater (rather than brine) at 45 °C and was shown to provide brine cooling of 6
to 9 °C with recirculating brine flow rates of 1 to 3 L/min, compared to only 1 to 2 °C of cooling
without evaporation. The BECC was used in the Pantelleria system as the primary source of
cooling, but once-through seawater cooling was also included in the system design as an
option. Water production and temperature data was presented for a day of operation and
relatively steady values of 4500 m
3
/h water production and 25 °C cold channel inlet
temperature were obtained. Although the authors indicated that the system could be operated
with only the BECC, only once-through seawater cooling, or both the BECC and once-through
seawater cooling, they did not indicate which of these cooling modes was used during
collection of the presented data or which mode was preferred. Further, the authors did not
102

indicate whether scaling or fouling in the BECC was an issue when using brine rather than the
freshwater typically used for evaporative cooling. In the same study by Schwantes et al. [8], the
system diagram was presented for a different pilot-scale PGMD system (the “Gran Canary”
system), which was said to use an experimental brine cooler. The simple diagram for the
experimental brine cooler in the Gran Canary system was identical to that shown for the
Pantelleria system, but the data presented for the Gran Canary system was collected when
using the system in a single-pass feed stream configuration (i.e., with replenishment cooling
and without the brine cooler).
Wang et al. [47] operated a lab-scale DCMD system integrated with a closed-circuit wet cooling
tower (Fig. 4.19). The system uses a recirculating feed stream configuration with continuous
brine discharge or “blowdown,” and the brine discharge is used as the makeup water for the
cooling tower; tap water was used as the feed solution. An electric heater was used to provide
the heat energy and the system used membrane area of 2 m
2
. The lab-scale system was
proposed as a test for the feasibility of integrating MD into recirculating cooling tower systems
between the condenser, where waste heat is removed from the primary industrial process (e.g.,
stream Rankine cycle for thermoelectric power production), and the cooling tower, where the
heat is rejected to the atmosphere (Fig. 4.20).
103

Figure 4.19: Diagram reproduced from Wang et al. [47] showing lab-scale direct contact
membrane distillation (DCMD) system with a recirculating feed stream configuration using a
closed-circuit wet cooling tower with DCMD system blowdown used as makeup water.

Figure 4.20: Diagram reproduced from Wang et al. [47] showing a direct contact membrane
distillation (DCMD) system integrated into the closed-circuit wet cooling tower system for an
industrial process.
Wang et al. [47] indicated that the system was able to achieve steady-state operation within a
24-hr period, and steady-state data for water production rate, flow rates, blowdown rates,
104

temperatures, and concentration factors were presented for a range of operating conditions.
Makeup water (i.e., replenishment water) temperature was 15 °C, relative humidity was 65%,
and wet bulb temperature was 18.9 °C. With feed inlet temperature ranging between 49.8 and
71.2 °C, the system was able to maintain the cooling water temperature between 22.6 and 23.7
°C. This cooling was due solely to the use of the cooling tower because replenishment cooling
(i.e., heat recovery) was not used in this system, indicating the viability of a closed-circuit wet
cooling tower for MD system cooling. Although the presented data showed how a DCMD
system can successfully operate with only a closed-circuit wet cooling tower over a wide range
of feed inlet temperatures, the study did not show how system performance would vary with
different wet bulb temperatures of the cooling air. Additionally, the system was only operated
with tap water feed solution, and scaling and/or fouling may be much more concerning when
treating solutions with much higher total dissolved solids (e.g., seawater or brackish water).
4.4.2.4 Dry cooling
4.4.2.4.1 Theory and energy consumption
In a dry cooling tower (Fig. 4.21), warm process water is circulated through the inside of
thermally conductive heat exchanger tubes while cool air is passed along the outside of the
tubes, resulting in the process water leaving the tower at a cooler temperature [124, 130]. Heat
exchangers are typically made of finned tubes to increase the heat transfer area [124, 130]. The
cooling process occurs through the transfer of sensible heat from the warm process water to
the cool air via conduction and convection by using mechanical draft fans. Dry cooling allows
the process water to approach the dry bulb temperature of the cool air [123]. Electrical energy
consumption of the fan is dictated by Equation 10. Because pumps are already used to circulate
105

water through the MD system, pumping energy associated with the dry cooling tower is only
derived from the additional pressure drop across the cooling tower heat exchanger, and can be
calculated from Equation 9. As with closed-circuit wet cooling, dry cooling has the advantage
over open-circuit wet cooling of the physical separation between process water and cooling air.
Advantages of dry cooling also include significantly lower water consumption and withdrawals
than wet and once-through cooling [125], but much higher capital and variable costs than wet
or once-through cooling [128, 136].

Figure 4.21: Dry cooling process diagram using mechanical draft.
4.4.2.4.2 Usage in existing MD systems
Guillén-Burrieza et al. [70] operated a pilot-scale AGMD system with dry cooling and heat
energy obtained through solar thermal collectors (Fig. 4.22). The AGMD modules did not
include internal heat recovery and separate circulation streams were used for the hot and cold
106

channels in the membrane module. The system used 2.8 m
2
membrane area and a recirculating
feed stream configuration operated in batch mode. Experimental water flux results were
presented as a function of temperature difference between hot and cold inlet temperatures
when using synthetic marine salt solutions of 1 and 35 g/L total dissolved solids. Water flux was
between 0.5 and 6.7 L·m
-2
h
-1
for temperature differences between 10 and 65 °C when using a 1-
g/L feed solution, and water flux values were slightly lower when using a 35 g/L feed solution.

Figure 4.22: Diagram reproduced from Guillén-Burrieza et al. [70] showing a pilot-scale air gap
membrane distillation (AGMD) system (without internal heat recovery) with a recirculating feed
stream configuration using dry cooling.
Guillén-Burrieza et al. [70] indicated difficulty in controlling hot stream inlet temperature due
to solar heat variability, and even greater difficulty in controlling cold stream inlet temperature
during summer months because of high air temperature up to 40 °C. Experiments were
operated only during sunlight hours and without replenishment, the hot and cold circuit fluids
were emptied at the end of each day, and the hot and cold circuits were refilled at the start of
each day. The refilling of the 2 m
3
cold tank with fresh water – likely to be cool – at the start of
each day likely contributed to cooling by providing a large heat sink volume that did not require
air cooling until later in the day, although this possibility was not discussed. Guillén-Burrieza et
107

al. [70] also presented modeling results that evaluated the impacts of hot and cold stream
temperatures and flow rates on water flux. Cold stream inlet temperature and particularly cold
stream flow rate were shown to have much less influence on water flux than hot stream
temperatures and flow rates; similar results have been found by other researchers [137, 138].
The authors indicate that cold stream temperature and flow rate have low impact on water flux
because the heat transfer coefficient of the air gap dominates the overall heat transfer
coefficient; other researchers have found similar results [9, 138, 139]. These results indicate
that cold stream characteristics will likely have much greater influence over water flux in
DCMD, where heat losses due to conduction are the highest of all MD configurations [10, 19-
21].
In a later study by Guillén-Burrieza et al. [71], a pilot-scale AGMD system (including internal
heat recovery) was also operated with air cooling and tested with a 35 g/L synthetic seawater
solution (Fig. 4.23). The system used 9 m
2
membrane area with a recirculating feed stream
configuration and obtained heat energy from solar thermal collectors. A graph of water flux and
inlet temperatures over time for operation during a sunny and clear day in July was presented,
showing a slight increase in the cold stream inlet temperature from approximately 30 to 35 °C
over the 5-hour period of water production. The increase in cold stream inlet temperature over
time corresponded with increasing hot stream inlet temperature over time from approximately
40 to 75 °C. Although the air cooling system was able to maintain the cold stream inlet
temperature near the target temperature of 30 °C, the reason for the increasing deviation from
the target temperature throughout the experiment was not discussed. The heat exchanger or
fan in the dry cooling system could have been undersized for the target operating conditions, or
108

air temperature may have been too high to allow for such a low cold stream inlet temperature
when operating with higher hot stream inlet temperatures. Additionally, the system was
operated in batch mode, emptied at the end of each day, and refilled at the start of each day,
so a large initial feed tank volume likely contributed to cooling in the earlier hours of operation
while contributing much less cooling in the later hours of operation as the feed tank heated up
throughout the day. This diminished contribution of cooling provided by the initial volume may
also explain the increasing deviation from the target cold stream inlet temperature throughout
the duration of the experiment, unless a very small feed tank volume was used (feed tank
volume was not given).

Figure 4.23: Diagram reproduced from Guillén-Burrieza et al. [71] showing a pilot-scale air gap
membrane distillation (AGMD) system with a recirculating feed stream configuration using dry
cooling.
109

Dow et al. [51] operated a pilot-scale DCMD system for the treatment of textile plant
wastewater effluent, using a combination of dry cooling and once-through cooling (Fig. 4.24). A
recirculating feed stream configuration was used, but the brine discharge mode was not
indicated. A fractionator was continuously operated in parallel with the MD system to remove
organics and foam, thereby controlling wetting potential of the feed throughout the trial. The
system used 6.38 m
2
membrane area and heat energy was obtained by absorbing heat from the
steam condensate return line of the textile plant. The system was operated periodically, for
multiple days at a time, over a 3-month period. Significant variability in temperature and flow
rate of the steam condensate resulted in feed inlet temperature varying between
approximately 30 and 60 °C and water flux varying between 2 and 8 L·m
-2
h
-1
. The cooling
system was able to maintain the distillate inlet temperature between approximately 15 and 28
°C. The authors indicated that only air cooling was used during the winter and a combination of
air cooling and once-through water cooling was used during the summer in order to maintain
target distillate inlet temperature throughout the trial. Because heat recovery was not used in
this system, replenishment did not contribute to cooling, and all cooling was due to either air
cooling or the combination of air cooling and once-through water cooling.
110

Figure 4.24: Diagram reproduced from Dow et al. [51] showing a pilot-scale direct contact
membrane distillation (DCMD) system with a recirculating feed stream configuration using dry
cooling along with once-through cooling.
4.5 Factors affecting choice of cooling approach
A summary of the advantages and disadvantages of replenishment cooling, once-through
cooling, wet cooling, and dry cooling is provided in Table 4.3. A summary of the advantages and
disadvantages of open-circuit wet cooling, conventional closed-circuit wet cooling, and the
BECC (non-conventional closed-circuit wet cooling) is provided in Table 4.4. The relative
advantages and disadvantages of the different cooling approaches provided in Tables 4.3 and
4.4 are discussed in Sections 4.5.1 – 4.5.3.

111

Table 4.3: Advantages and disadvantages of low electrical energy cooling approaches.
Cooling
Approach Advantages Disadvantages
Replenishment
Cooling
Also provides heat recovery/reduced
heating energy consumption
Diminishing thermal driving force over
time with feed recirculation
Can provide all necessary cooling in
single-pass configuration
Interdependence of cooling controls
and feed controls
Not limited by cooling water availability
Not limited by cooling water regulations
Low capital and operating costs
Simple operation
Once-Through
Cooling
Best performance of all auxiliary cooling Requires availability of large external
cooling water source
Lowest capital and variable costs of all
auxiliary cooling
Regulatory concerns regarding
impingement and entrainment
Simple operation Regulatory concerns regarding
thermal pollution
Wet Cooling Better performance than dry cooling Poorer performance than once-
through cooling
Lower water withdrawals than once-
through cooling
Higher water consumption than once-
through cooling
Lower capital and variable costs than dry
cooling
Higher capital and variable costs than
once-through cooling
Can approach lower temperature (wet-
bulb) than dry cooling

Most complex operation of all
auxiliary cooling
Requires external cooling water
source for conventional closed-circuit
wet cooling
Potential regulatory concerns
regarding blowdown disposal
Possibility of direct contact between
process water and potentially
contaminated cooling air
Dry Cooling Lowest water withdrawals and
consumption of all auxiliary cooling

Poorest performance of all auxiliary
cooling
Highest capital and operating costs of
all auxiliary cooling
Cannot approach as low temperature
as wet cooling

112

Table 4.4: Advantages and disadvantages of different wet cooling options.
Cooling
Approach Advantages Disadvantages
Open-Circuit
Wet Cooling
Lower capital cost than conventional
closed-loop wet cooling

Process water in direct contact with
potentially contaminated cooling air
Scaling concerns may limit achievable
water recovery of membrane distillation
Conventional
Closed-Circuit
Wet Cooling
Higher capital costs than open-circuit
wet cooling
Regulatory concerns regarding blowdown
disposal

Process water does not contact
potentially contaminated cooling air
Brine
Evaporative
Cooler
Concentrator
Process water does not contact
potentially contaminated cooling air
Requires further demonstration of long-
term performance with high salinity
process water

4.5.1 Cooling water availability
If a large source of cooling water is not available near the MD system location, then once-
through water cooling will not be an option because of the large water withdrawals required
[123-125]. An external water source must also be available for conventional closed-circuit wet
cooling in order to support the high water consumption and low water withdrawals required
[123-125]. Even if a suitable source of cooling water for once-through or conventional closed
circuit wet cooling is available, water availability can still become an issue in times of drought
[124, 140], particularly if the cooling water source is also the feed water source. In locations
where cooling water is not available, replenishment cooling, dry cooling, or a BECC will be the
only options available.
113

4.5.2 Regulatory environment
Laws specific to a proposed system location will affect how an MD system must be designed,
including the type of cooling system used. For example, amendment 316 (b) to the Clean Water
Act in the United States dictates that facilities that have a design intake flow greater than 2
million gallons per day and use at least 25% of that flow exclusively for cooling must use wet or
dry cooling to minimize impingement and entrainment of aquatic organisms [130]. Regulations
like this would prevent the use of once-through cooling due to its large water withdrawals [125]
and limit the cooling approaches for an MD system to replenishment cooling, wet cooling, or
dry cooling due to their lower water withdrawals [125]. Once-through and conventional closed-
circuit wet cooling may also become limited due to regulatory environment under the
circumstances of a drought, when policies of curtailment can be put in place [140] or when the
water level of a cooling water source falls below the intake structure [124]. In locations where
water availability is limited and discharge of cooling tower blowdown is severely regulated,
blowdown disposal costs may require the use of dry cooling instead of wet cooling [123].
4.5.3 Water recovery and costs
Replenishment cooling is the lowest-cost cooling approach when using a single-pass feed
stream configuration with low water recovery (i.e., without staging in series) because the costs
associated with replenishment cooling are already included in the costs of heat recovery for
minimizing heating energy consumption. Although the cost of operating with only
replenishment cooling using a single-pass feed stream configuration without staging is low, the
inherently low water recovery associated with this configuration would require much greater
water withdrawals than a system operating at high water recovery with a staged single-pass or
114

recirculating feed stream configuration. The greater withdrawals of low recovery system
designs may result in them facing greater regulatory issues to minimize impingement and
entrainment concerns.
In systems operating with high water recovery, auxiliary cooling will be necessary to maintain
high thermal driving force over time with a recirculating configuration or in the final stages of a
series of modules in a single-pass configuration (see Section 4.4.1.3). If sufficient cooling water
is available and water withdrawals are not constrained due to regulations, then once-through
water cooling will be the lower-cost and better-performing option compared to dry or wet
cooling [128]. Once-through cooling has lower costs than dry and wet cooling because it is a
much simpler cooling approach that also allows for generally cooler temperatures to be
achieved [123, 124]. If regulations prevent the use of once-through cooling, then wet cooling is
considered the next lowest-cost and best-performing option [128]. Wet cooling has lower costs
than dry cooling because wet cooling can approach a lower temperature (i.e., wet-bulb
temperature) than dry cooling (i.e., dry-bulb temperature), which allows the wet cooling system
size to be smaller [136]. Although dry cooling is the highest cost cooling approach and cannot
achieve as low temperature as once-through or wet cooling, the cooling fluid used for dry
cooling (i.e., air) is available at every site location and may be the only option if water
availability and/or regulatory environment limit other cooling approaches.
Open-circuit wet cooling systems have lower cost than closed-circuit wet cooling systems, but
introduce concerns regarding direct contact of potentially contaminated cooling air with MD
streams [133, 134]. Additionally, the use of an open-circuit wet cooling tower may limit the MD
system to low water recovery due to scaling concerns that require frequent brine
115

discharge/blowdown when treating high salinity feed solutions (e.g., seawater). Cost data is not
available for the BECC [135] because it is only in the early stages of development and not
commercially available. Additionally, a more detailed assessment of the BECC’s long-term
performance with a high salinity feed (e.g., seawater) may be required before wider adoption.
Nonetheless, the BECC may provide a useful cooling alternative for MD systems with its
benefits of maintaining separation between cooling air and MD streams and not requiring an
external cooling water source.
4.6 Conclusions and future research needs
A critical analysis of the information available on cooling in MD systems has been performed for
the first time, allowing for identification of future research opportunities on cooling in MD that
will help lead to improved MD system design. A new linguistic framework was proposed that
allows the valuable contribution of heat recovery to cooling to be more formally acknowledged
through usage of the term “replenishment cooling.” This new linguistic framework also
acknowledges the secondary role that cooling approaches like once-through, wet, or dry
cooling should play in relation to replenishment cooling by identifying them as “auxiliary
cooling” approaches. The case is made for the quantification of specific thermal energy
consumption for cooling ( 𝑆𝑇𝐸𝐶 𝑐 ), as is already common practice for heating, to provide a
quantitative measure of cooling needs that is proportional to cooling system size. 𝑆𝑇𝐸𝐶 𝑐 will be
useful for sizing cooling systems and providing a system metric that can be targeted to minimize
system size/cost.
116

Replenishment cooling was identified as the lowest-cost option for cooling when using a single-
pass feed stream configuration with low water recovery, but opportunities for future studies of
replenishment cooling in MD have also been identified. While a small number of studies have
demonstrated successful operation of MD systems using only replenishment cooling with a
single-pass feed stream configuration and low water recovery, this low water recovery of MD
would require very large water withdrawals compared to operating at high overall water
recovery with a staged single-pass or recirculating feed stream configuration. The limited
number of studies showing operation only with replenishment cooling and a recirculating feed
stream configuration provided very little experimental data characterizing the system’s
performance – particularly regarding the relationship between water recovery, replenishment
rate, and brine discharge rate/frequency. Future studies of MD with replenishment cooling
should provide detailed time-varying performance data that includes water recovery,
replenishment rate, and brine discharge rate/frequency data in addition to the water flux and
temperature data typically provided.
Analysis of the performance of different auxiliary cooling approaches in MD systems has
allowed for identification of future research opportunities on their effective integration with
MD systems. The key problems identified with once-through and dry cooling approaches used
in MD systems was the lack of information regarding sizing of cooling equipment, the design
operating conditions of the cooling system, and evaluation of these cooling systems’
performance over a range of climatic and MD system operation conditions. These problems
also need to be addressed for wet cooling systems used in MD. However, the relationships
between scaling, water recovery, and brine discharge need to be researched in much greater
117

detail in order to accurately assess the potential for long-term use of wet cooling in MD
systems. When an auxiliary cooling approach is used in an MD system, justification for selection
of a particular cooling approach should be provided on the basis of cooling water availability,
regulatory environment, and relative performance/cost between systems.
MD system tanks were found to play an important role in cooling and this contribution should
be accounted for in future studies. A large tank volume may provide a large heat sink, but this
large tank volume may be unrealistic in a scaled-up scenario, so reasonable proportionality
between tank size and membrane area should be analyzed in future studies. Addressing the
research opportunities identified through this critical review of cooling in MD will provide a
more complete understanding of the performance capabilities of MD, providing the necessary
groundwork for wider commercial adoption of MD in the future.

118

Chapter 5 Conclusions
5.1 Research Synopsis
This dissertation represents investigations into the long-term performance and energy
integration issues of MD systems. These investigations include: (1) morphological changes and
creep recovery in ePTFE MD membranes, (2) integration of MD with variable-temperature and
intermittent waste heat, and (3) a critical review of cooling in MD. The research described in
this dissertation addresses important issues that must be taken into consideration in order for
MD to be more widely adopted in small-scale water treatment applications and eventually in
large-scale water treatment applications.
5.1.1 Summary of morphological changes and creep recovery in ePTFE MD membranes
Three 30-day DCMD experiments were operated using an ePTFE membrane with different
combinations of feed temperature and salinity in the absence of foulants. Membranes were
analyzed with SEM and AFM before and after experiments to assess changes in morphology
(i.e., surface microstructure, surface porosity, surface roughness, thickness, and bulk porosity).
Significant deviations from the classic fibril-and-node microstructure were found on the feed
side of the membrane in the two higher temperature (65 °C) experiments, while less-significant
changes were found in the lower temperature (45 °C) experiment. These changes were
quantified by calculating changes in surface porosity; decrease in feed-side surface porosity was
between 80 and 89% in the 65 °C experiments and was 55% in the 45 °C experiment. Thickness
was shown to decrease between 24 and 39% in the 65 °C experiments and by 13% in the 45 °C
experiment. Changes in thickness were used to calculate changes in bulk porosity; decrease in
bulk porosity was between 14 and 20% in the 65 °C experiments and was 10% in the 45 °C
119

experiment. Water flux was shown to decrease slightly over time in all three experiments. The
absence of foulants during these experiments indicated that the observed changes in
morphology likely caused the decrease in water flux.
Creep recovery tests were performed with a low applied stress within the operating range of
MD (0.05 MPa) at 25 and 65 °C. Results showed that the ePTFE membrane tested does deform
over time under constant and low applied stress – confirming the creep behavior proposed by
previous researchers. Creep was shown to be more severe at higher temperature, with 16%
peak strain at 65 °C and 6% peak strain at 25 °C. Deformation caused by creep was shown to be
partially recovered at both temperatures, but also indicated the potential for permanent
deformation to remain in the membrane. Both microscopy and creep recovery results indicated
that the ePTFE membrane tested experiences more significant deformation when exposed to
higher temperatures. While microscopy analyses provide high resolution data for the
assessment of changes in membrane morphology at different temperatures during long-term
MD experiments, creep recovery tests were shown to provide a simple and short-term method
for assessing this behavior.
5.1.2 Summary of integration of MD with variable-temperature and intermittent waste heat
Direct and indirect DCMD system arrangements were evaluated and compared based on their
ability to integrate variable-temperature and intermittent waste heat sources into MD systems.
Modeling results were performed using intermittent waste heat, showing that the direct
arrangement operates at higher water flux when the waste heat source is on but the indirect
arrangement operates at higher water flux when the waste heat source is off. Modeling results
also showed that the indirect arrangement stores a greater amount of heat at a faster rate than
120

the direct arrangement when the waste heat source is on. This behavior was apparent in
additional modeling results showing the indirect arrangement producing 21.5% more water
when using a waste heat source with 87.5% intermittency (high intermittency) and the direct
arrangement producing 17.7% more water when using a waste heat source with 12.5%
intermittency (low intermittency); similar results were found experimentally.
Experimental analyses were also performed with the more thermodynamically complex
scenario of variable-temperature waste heat using real-world waste heat data to drive the MD
process. These experiments showed that variability in waste heat source temperature is
strongly reflected in water flux profiles, but that water flux variability in the indirect
arrangement was dampened by 30.4% relative to the direct arrangement. The direct
arrangement provided 22.5% higher water flux than the indirect arrangement with the variable-
temperature heat source profile. However, the indirect arrangement was also shown to provide
a delayed water flux response to variability in waste heat source temperature, demonstrating
the indirect arrangements better controllability and predictability of product water availability
than the direct arrangement.
5.1.3 Summary of a critical review of cooling in MD
A critical review of the literature on pilot-scale MD systems was performed to isolate important
information on the performance and assessment of cooling systems used for MD. Important
factors that affect energy use in MD systems were reviewed, including heat recovery and feed
stream configuration. The energy metrics used to assess the thermal energy consumption for
heating were reviewed and proposed to be applied also to cooling in MD, providing a useful
metric ( 𝑆𝑇𝐸𝐶 𝑐 ) to target for cooling system size/cost minimization. A new linguistic framework
121

was proposed that acknowledges the important role that heat recovery plays in cooling and this
type of cooling was categorized as “replenishment cooling”; cooling approaches using at least
one external fluid for cooling that is not the feed (e.g., once-through, wet, or dry cooling) were
categorized as “auxiliary cooling.”
The experience of MD researchers in using replenishment and auxiliary cooling in pilot-scale
MD systems was summarized and critically assessed. Through this critical assessment, the
ability to operate only with replenishment cooling was identified and the advantages and
disadvantages of this approach were outlined. Additionally, the important and unacknowledged
role that system tanks have played in cooling in MD systems was described. Factors that affect
selection of a cooling approach for a particular application – based on cooling water availability,
regulatory environment, water recovery, and cost – were discussed and opportunities for
future research on cooling in MD were proposed. Future research needs identified regarding
cooling in MD include the need for (i) more detailed experimental data on operation with only
replenishment cooling in high overall water recovery systems, (ii) more detailed information on
cooling system sizing, (iii) justification for cooling approach selection, and (iv) an accurate
accounting of the relative amounts of cooling provided by replenishment and auxiliary cooling
in systems using both approaches.
Chapter 6 References
[1] J. Schewe, J. Heinke, D. Gerten, I. Haddeland, N.W. Arnell, D.B. Clark, R. Dankers, S. Eisner, B.M.
Fekete, F.J. Colón-González, S.N. Gosling, H. Kim, X. Liu, Y. Masaki, F.T. Portmann, Y. Satoh, T. Stacke, Q.
Tang, Y. Wada, D. Wisser, T. Albrecht, K. Frieler, F. Piontek, L. Warszawski, P. Kabat, Multimodel
assessment of water scarcity under climate change, Proceedings of the National Academy of Sciences,
(2013) 201222460.
122

[2] S.K. Tanaka, T. Zhu, J.R. Lund, R.E. Howitt, M.W. Jenkins, M.A. Pulido, M. Tauber, R.S. Ritzema, I.C.
Ferreira, Climate Warming and Water Management Adaptation for California, Climatic Change, 76
(2006) 361-387.
[3] S.G.S.A. Rothausen, D. Conway, Greenhouse-gas emissions from energy use in the water sector,
Nature Climate Change, 1 (2011) 210.
[4] B. Bates, Z. Kundzewicz, S. Wu, Climate change and water, Intergovernmental Panel on Climate
Change Secretariat, 2008.
[5] K.A. Salls, D. Won, E.P. Kolodziej, A.E. Childress, S.R. Hiibel, Evaluation of semi-volatile contaminant
transport in a novel, gas-tight direct contact membrane distillation system, Desalination, 427 (2018) 35-
41.
[6] N. Dow, S. Gray, J.-d. Li, J. Zhang, E. Ostarcevic, A. Liubinas, P. Atherton, G. Roeszler, A. Gibbs, M.
Duke, Pilot trial of membrane distillation driven by low grade waste heat: Membrane fouling and energy
assessment, Desalination, 391 (2016) 30-42.
[7] R.D. Gustafson, S.R. Hiibel, A.E. Childress, Membrane distillation driven by intermittent and variable-
temperature waste heat: System arrangements for water production and heat storage, Desalination,
448 (2018) 49-59.
[8] R. Schwantes, A. Cipollina, F. Gross, J. Koschikowski, D. Pfeifle, M. Rolletschek, V. Subiela, Membrane
distillation: Solar and waste heat driven demonstration plants for desalination, Desalination, 323 (2013)
93-106.
[9] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, A framework for better understanding membrane
distillation separation process, Journal of Membrane Science, 285 (2006) 4-29.
[10] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: A comprehensive review, Desalination,
287 (2012) 2-18.
[11] S.T. Hsu, K.T. Cheng, J.S. Chiou, Seawater desalination by direct contact membrane distillation,
Desalination, 143 (2002) 279-287.
[12] C.R. Martinetti, A.E. Childress, T.Y. Cath, High recovery of concentrated RO brines using forward
osmosis and membrane distillation, Journal of Membrane Science, 331 (2009) 31-39.
[13] J. Koschikowski, M. Wieghaus, M. Rommel, Solar thermal-driven desalination plants based on
membrane distillation, Desalination, 156 (2003) 295-304.
[14] T.D. Dao, S. Laborie, C. Cabassud, Direct As(III) removal from brackish groundwater by vacuum
membrane distillation: Effect of organic matter and salts on membrane fouling, Separation and
Purification Technology, 157 (2016) 35-44.
[15] T.Y. Cath, V.D. Adams, A.E. Childress, Experimental study of desalination using direct contact
membrane distillation: a new approach to flux enhancement, Journal of Membrane Science, 228 (2004)
5-16.
[16] F. Suárez, J.A. Ruskowitz, S.W. Tyler, A.E. Childress, Renewable water: Direct contact membrane
distillation coupled with solar ponds, Applied Energy, 158 (2015) 532-539.
[17] F. Suarez, S.W. Tyler, A.E. Childress, A theoretical study of a direct contact membrane distillation
system coupled to a salt-gradient solar pond for terminal lakes reclamation, Water research, 44 (2010)
4601-4615.
[18] K.L. Hickenbottom, T.Y. Cath, Sustainable operation of membrane distillation for enhancement of
mineral recovery from hypersaline solutions, Journal of Membrane Science, 454 (2014) 426-435.
[19] M.L. Camacho, L. Dumée, J. Zhang, J.-d. Li, M. Duke, J. Gomez, S. Gray, Advances in Membrane
Distillation for Water Desalination and Purification Applications, Water, 5 (2013).
[20] N. Thomas, M.O. Mavukkandy, S. Loutatidou, H.A. Arafat, Membrane distillation research &
implementation: Lessons from the past five decades, Separation and Purification Technology, 189 (2017)
108-127.
123

[21] P. Wang, T.-S. Chung, Recent advances in membrane distillation processes: Membrane
development, configuration design and application exploring, Journal of Membrane Science, 474 (2015)
39-56.
[22] J. Swaminathan, H.W. Chung, D.M. Warsinger, F.A. AlMarzooqi, H.A. Arafat, J.H. Lienhard V, Energy
efficiency of permeate gap and novel conductive gap membrane distillation, Journal of Membrane
Science, 502 (2016) 171-178.
[23] L. Cheng, Y. Zhao, P. Li, W. Li, F. Wang, Comparative study of air gap and permeate gap membrane
distillation using internal heat recovery hollow fiber membrane module, Desalination, 426 (2018) 42-49.
[24] D. Winter, J. Koschikowski, M. Wieghaus, Desalination using membrane distillation: Experimental
studies on full scale spiral wound modules, Journal of Membrane Science, 375 (2011) 104-112.
[25] L. Francis, N. Ghaffour, A.A. Alsaadi, G.L. Amy, Material gap membrane distillation: A new design for
water vapor flux enhancement, Journal of Membrane Science, 448 (2013) 240-247.
[26] R. Schwantes, K. Chavan, D. Winter, C. Felsmann, J. Pfafferott, Techno-economic comparison of
membrane distillation and MVC in a zero liquid discharge application, Desalination, 428 (2018) 50-68.
[27] M.A.E.-R. Abu-Zeid, L. Zhang, W.-Y. Jin, T. Feng, Y. Wu, H.-L. Chen, L.a. Hou, Improving the
performance of the air gap membrane distillation process by using a supplementary vacuum pump,
Desalination, 384 (2016) 31-42.
[28] E. Drioli, A. Ali, F. Macedonio, Membrane distillation: Recent developments and perspectives,
Desalination, 356 (2015) 56-84.
[29] E. Guillen-Burrieza, A. Servi, B.S. Lalia, H.A. Arafat, Membrane structure and surface morphology
impact on the wetting of MD membranes, Journal of Membrane Science, 483 (2015) 94-103.
[30] S. Ebnesajjad, Expanded PTFE Applications Handbook: Technology, Manufacturing and Applications,
William Andrew, 2016.
[31] E. Guillen-Burrieza, A. Ruiz-Aguirre, G. Zaragoza, H.A. Arafat, Membrane fouling and cleaning in long
term plant-scale membrane distillation operations, Journal of Membrane Science, 468 (2014) 360-372.
[32] M. Gryta, Long-term performance of membrane distillation process, Journal of Membrane Science,
265 (2005) 153-159.
[33] M. Gryta, M. Barancewicz, Influence of morphology of PVDF capillary membranes on the
performance of direct contact membrane distillation, Journal of Membrane Science, 358 (2010) 158-
167.
[34] A.L. McGaughey, R.D. Gustafson, A.E. Childress, Effect of long-term operation on membrane surface
characteristics and performance in membrane distillation, Journal of Membrane Science, 543 (2017)
143-150.
[35] J. Zhang, J.-D. Li, S. Gray, Effect of applied pressure on performance of PTFE membrane in DCMD,
Journal of Membrane Science, 369 (2011) 514-525.
[36] K.W. Lawson, M.S. Hall, D.R. Lloyd, Compaction of microporous membranes used in membrane
distillation. I. Effect on gas permeability, Journal of Membrane Science, 101 (1995) 99-108.
[37] J. Zhang, S. Gray, J.-D. Li, Modelling heat and mass transfers in DCMD using compressible
membranes, Journal of Membrane Science, 387-388 (2012) 7-16.
[38] I. Hitsov, L. Eykens, W.D. Schepper, K.D. Sitter, C. Dotremont, I. Nopens, Full-scale direct contact
membrane distillation (DCMD) model including membrane compaction effects, Journal of Membrane
Science, 524 (2017) 245-256.
[39] A.E. Jansen, J.W. Assink, J.H. Hanemaaijer, J. van Medevoort, E. van Sonsbeek, Development and
pilot testing of full-scale membrane distillation modules for deployment of waste heat, Desalination, 323
(2013) 55-65.
[40] B. Li, K.K. Sirkar, Novel Membrane and Device for Direct Contact Membrane Distillation-Based
Desalination Process, Industrial & Engineering Chemistry Research, 43 (2004) 5300-5309.
124

[41] L. Song, B. Li, K.K. Sirkar, J.L. Gilron, Direct Contact Membrane Distillation-Based Desalination:
Novel Membranes, Devices, Larger-Scale Studies, and a Model, Industrial & Engineering Chemistry
Research, 46 (2007) 2307-2323.
[42] H. Maab, A. Al Saadi, L. Francis, S. Livazovic, N. Ghafour, G.L. Amy, S.P. Nunes, Polyazole hollow fiber
membranes for direct contact membrane distillation, Industrial & Engineering Chemistry Research, 52
(2013) 10425-10429.
[43] A.S. Alavijeh, R.M. Khorasany, A. Habisch, G.G. Wang, E. Kjeang, Creep properties of catalyst coated
membranes for polymer electrolyte fuel cells, Journal of Power Sources, 285 (2015) 16-28.
[44] R.B. Saffarini, B. Mansoor, R. Thomas, H.A. Arafat, Effect of temperature-dependent microstructure
evolution on pore wetting in PTFE membranes under membrane distillation conditions, Journal of
Membrane Science, 429 (2013) 282-294.
[45] Y. Xu, B.-k. Zhu, Y.-y. Xu, Pilot test of vacuum membrane distillation for seawater desalination on a
ship, Desalination, 189 (2006) 165-169.
[46] S. Tavakkoli, O.R. Lokare, R.D. Vidic, V. Khanna, A techno-economic assessment of membrane
distillation for treatment of Marcellus shale produced water, Desalination, 416 (2017) 24-34.
[47] J. Wang, B. Fan, Z. Luan, D. Qu, X. Peng, D. Hou, Integration of direct contact membrane distillation
and recirculating cooling water system for pure water production, Journal of Cleaner Production, 16
(2008) 1847-1855.
[48] A. Hagedorn, G. Fieg, D. Winter, J. Koschikowski, T. Mann, Methodical design and operation of
membrane distillation plants for desalination, Chemical Engineering Research and Design, 125 (2017)
265-281.
[49] O.R. Lokare, S. Tavakkoli, G. Rodriguez, V. Khanna, R.D. Vidic, Integrating membrane distillation with
waste heat from natural gas compressor stations for produced water treatment in Pennsylvania,
Desalination, 413 (2017) 144-153.
[50] U.K. Kesieme, N. Milne, H. Aral, C.Y. Cheng, M. Duke, Economic analysis of desalination technologies
in the context of carbon pricing, and opportunities for membrane distillation, Desalination, 323 (2013)
66-74.
[51] N. Dow, J.V. García, L. Niadoo, N. Milne, J. Zhang, S. Gray, M. Duke, Demonstration of membrane
distillation on textile waste water: assessment of long term performance, membrane cleaning and waste
heat integration, Environmental Science: Water Research & Technology, 3 (2017) 433-449.
[52] A. Deshmukh, C. Boo, V. Karanikola, S. Lin, A.P. Straub, T. Tong, D.M. Warsinger, M. Elimelech,
Membrane distillation at the water-energy nexus: limits, opportunities, and challenges, Energy &
Environmental Science, (2018).
[53] R. Long, X. Lai, Z. Liu, W. Liu, Direct contact membrane distillation system for waste heat recovery:
Modelling and multi-objective optimization, Energy, 148 (2018) 1060-1068.
[54] S. Tavakkoli, O.R. Lokare, R.D. Vidic, V. Khanna, Systems-Level Analysis of Waste Heat Recovery
Opportunities from Natural Gas Compressor Stations in the United States, ACS Sustainable Chemistry &
Engineering, 4 (2016) 3618-3626.
[55] D.B. Gingerich, M.S. Mauter, Quantity, Quality, and Availability of Waste Heat from United States
Thermal Power Generation, Environmental Science & Technology, 49 (2015) 8297-8306.
[56] B. Peris, J. Navarro-Esbrí, F. Molés, A. Mota-Babiloni, Experimental study of an ORC (organic Rankine
cycle) for low grade waste heat recovery in a ceramic industry, Energy, 85 (2015) 534-542.
[57] F. Campana, M. Bianchi, L. Branchini, A. De Pascale, A. Peretto, M. Baresi, A. Fermi, N. Rossetti, R.
Vescovo, ORC waste heat recovery in European energy intensive industries: Energy and GHG savings,
Energy Conversion and Management, 76 (2013) 244-252.
[58] B.F. Tchanche, G. Lambrinos, A. Frangoudakis, G. Papadakis, Low-grade heat conversion into power
using organic Rankine cycles – A review of various applications, Renewable and Sustainable Energy
Reviews, 15 (2011) 3963-3979.
125

[59] Y. Ammar, S. Joyce, R. Norman, Y. Wang, A.P. Roskilly, Low grade thermal energy sources and uses
from the process industry in the UK, Applied Energy, 89 (2012) 3-20.
[60] M. Papapetrou, G. Kosmadakis, A. Cipollina, U. La Commare, G. Micale, Industrial waste heat:
Estimation of the technically available resource in the EU per industrial sector, temperature level and
country, Applied Thermal Engineering, 138 (2018) 207-216.
[61] A. Chafidz, S. Al-Zahrani, M.N. Al-Otaibi, C.F. Hoong, T.F. Lai, M. Prabu, Portable and integrated
solar-driven desalination system using membrane distillation for arid remote areas in Saudi Arabia,
Desalination, 345 (2014) 36-49.
[62] H.C. Duong, A.R. Chivas, B. Nelemans, M. Duke, S. Gray, T.Y. Cath, L.D. Nghiem, Treatment of RO
brine from CSG produced water by spiral-wound air gap membrane distillation — A pilot study,
Desalination, 366 (2015) 121-129.
[63] D. Woldemariam, A. Kullab, U. Fortkamp, J. Magner, H. Royen, A. Martin, Membrane distillation
pilot plant trials with pharmaceutical residues and energy demand analysis, Chemical Engineering
Journal, 306 (2016) 471-483.
[64] A. Chafidz, E.D. Kerme, I. Wazeer, Y. Khalid, A. Ajbar, S.M. Al-Zahrani, Design and fabrication of a
portable and hybrid solar-powered membrane distillation system, Journal of Cleaner Production, 133
(2016) 631-647.
[65] S. Bouguecha, M. Dhahbi, Fluidised bed crystalliser and air gap membrane distillation as a solution
to geothermal water desalination, Desalination, 152 (2003) 237-244.
[66] F. Banat, N. Jwaied, M. Rommel, J. Koschikowski, M. Wieghaus, Performance evaluation of the
“large SMADES” autonomous desalination solar-driven membrane distillation plant in Aqaba, Jordan,
Desalination, 217 (2007) 17-28.
[67] J.A. Andrés-Mañas, A. Ruiz-Aguirre, F.G. Acién, G. Zaragoza, Assessment of a pilot system for
seawater desalination based on vacuum multi-effect membrane distillation with enhanced heat
recovery, Desalination, 443 (2018) 110-121.
[68] R.G. Raluy, R. Schwantes, V.J. Subiela, B. Peñate, G. Melián, J.R. Betancort, Operational experience
of a solar membrane distillation demonstration plant in Pozo Izquierdo-Gran Canaria Island (Spain),
Desalination, 290 (2012) 1-13.
[69] K. Zhao, W. Heinzl, M. Wenzel, S. Büttner, F. Bollen, G. Lange, S. Heinzl, N. Sarda, Experimental
study of the memsys vacuum-multi-effect-membrane-distillation (V-MEMD) module, Desalination, 323
(2013) 150-160.
[70] E. Guillén-Burrieza, J. Blanco, G. Zaragoza, D.-C. Alarcón, P. Palenzuela, M. Ibarra, W. Gernjak,
Experimental analysis of an air gap membrane distillation solar desalination pilot system, Journal of
Membrane Science, 379 (2011) 386-396.
[71] E. Guillén-Burrieza, G. Zaragoza, S. Miralles-Cuevas, J. Blanco, Experimental evaluation of two pilot-
scale membrane distillation modules used for solar desalination, Journal of Membrane Science, 409-410
(2012) 264-275.
[72] A.E. Kabeel, M. Abdelgaied, E.M.S. El-Said, Study of a solar-driven membrane distillation system:
Evaporative cooling effect on performance enhancement, Renewable Energy, 106 (2017) 192-200.
[73] J. Swaminathan, J.H. Lienhard, Design and operation of membrane distillation with feed
recirculation for high recovery brine concentration, Desalination, 445 (2018) 51-62.
[74] H.C. Duong, L. Xia, Z. Ma, P. Cooper, W. Ela, L.D. Nghiem, Assessing the performance of solar
thermal driven membrane distillation for seawater desalination by computer simulation, Journal of
Membrane Science, 542 (2017) 133-142.
[75] F. Banat, N. Jwaied, M. Rommel, J. Koschikowski, M. Wieghaus, Desalination by a “compact
SMADES” autonomous solarpowered membrane distillation unit, Desalination, 217 (2007) 29-37.
[76] K.W. Lawson, D.R. Lloyd, Membrane distillation, Journal of Membrane Science, 124 (1997) 1-25.
126

[77] T.Y. Cath, D. Adams, A.E. Childress, Membrane contactor processes for wastewater reclamation in
space: II. Combined direct osmosis, osmotic distillation, and membrane distillation for treatment of
metabolic wastewater, Journal of Membrane Science, 257 (2005) 111-119.
[78] J. Ge, Y. Peng, Z. Li, P. Chen, S. Wang, Membrane fouling and wetting in a DCMD process for RO
brine concentration, Desalination, 344 (2014) 97-107.
[79] V. Villani, R. Pucciariello, G. Ajroldi, Calorimetric study of the room‐temperature transitions of
polytetrafluoroethylene: The influence of thermal history, Journal of Polymer Science Part B: Polymer
Physics, 29 (1991) 1255-1259.
[80] P. Rae, D. Dattelbaum, The properties of poly (tetrafluoroethylene)(PTFE) in compression, Polymer,
45 (2004) 7615-7625.
[81] L. Andena, M. Rink, F. Polastri, Simulation of PTFE sintering: Thermal stresses and deformation
behavior, Polymer Engineering & Science, 44 (2004) 1368-1378.
[82] S. Fai Lau, H. Suzuki, B. Wunderlich, The thermodynamic properties of polytetrafluoroethylene,
Journal of Polymer Science: Polymer Physics Edition, 22 (1984) 379-405.
[83] P.W. Majsztrik, A.B. Bocarsly, J.B. Benziger, Viscoelastic Response of Nafion. Effects of Temperature
and Hydration on Tensile Creep, Macromolecules, 41 (2008) 9849-9862.
[84] L.W. McKeen, Chapter 1 - Introduction to Plastics and Elastomers, in: L.W. McKeen (Ed.) The Effect
of Creep and Other Time Related Factors on Plastics and Elastomers (Second Edition), William Andrew
Publishing, Boston, 2009, pp. 1-31.
[85] X. Zheng, X. Wen, W. Wang, J. Gao, W. Lin, L. Ma, J. Yu, Creep-ratcheting behavior of PTFE gaskets
under various temperatures, Polymer Testing, 60 (2017) 229-235.
[86] C.S. Desai, S. Sane, J. Jenson, Constitutive Modeling Including Creep- and Rate-Dependent Behavior
and Testing of Glacial Tills for Prediction of Motion of Glaciers, International Journal of Geomechanics,
11 (2011) 465-476.
[87] R.R. Ferlita, D. Phipps, J. Safarik, D.H. Yeh, Cryo-snap: A simple modified freeze-fracture method for
SEM imaging of membrane cross-sections, Environmental Progress, 27 (2008) 204-209.
[88] G. Rao, S.R. Hiibel, A. Achilli, A.E. Childress, Factors contributing to flux improvement in vacuum-
enhanced direct contact membrane distillation, Desalination, 367 (2015) 197-205.
[89] R. Mishra, S.P. Tripathy, D. Sinha, K.K. Dwivedi, S. Ghosh, D.T. Khathing, M. Müller, D. Fink, W.H.
Chung, Optical and electrical properties of some electron and proton irradiated polymers, Nuclear
Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms,
168 (2000) 59-64.
[90] Q.-l. Huang, C.-f. Xiao, X.-y. Hu, X.-f. Li, Study on the effects and properties of hydrophobic
poly(tetrafluoroethylene) membrane, Desalination, 277 (2011) 187-192.
[91] F.M. White, Fluid Mechanics, 6 ed., McGraw-Hill, New York, NY, USA, 2008.
[92] N.C. Nguyen, H.T. Nguyen, S.-S. Chen, H.H. Ngo, W. Guo, W.H. Chan, S.S. Ray, C.-W. Li, H.-T. Hsu, A
novel osmosis membrane bioreactor-membrane distillation hybrid system for wastewater treatment
and reuse, Bioresource Technology, 209 (2016) 8-15.
[93] C.P. Morrow, N.M. Furtaw, J.R. Murphy, A. Achilli, E.A. Marchand, S.R. Hiibel, A.E. Childress,
Integrating an aerobic/anoxic osmotic membrane bioreactor with membrane distillation for potable
reuse, Desalination, 432 (2018) 46-54.
[94] G. Zakrzewska-Trznadel, M. Harasimowicz, A.G. Chmielewski, Concentration of radioactive
components in liquid low-level radioactive waste by membrane distillation, Journal of Membrane
Science, 163 (1999) 257-264.
[95] G.W. Meindersma, C.M. Guijt, A.B. de Haan, Desalination and water recycling by air gap membrane
distillation, Desalination, 187 (2006) 291-301.
[96] I. Johnson, W.T. Choate, A. Davidson, Waste heat recovery: Technology and opportunities in US
industry, in, BCS, Inc., Laurel, MD (United States), 2008.
127

[97] X. Wang, L. Zhang, H. Yang, H. Chen, Feasibility research of potable water production via solar-
heated hollow fiber membrane distillation system, Desalination, 247 (2009) 403-411.
[98] R. Porrazzo, A. Cipollina, M. Galluzzo, G. Micale, A neural network-based optimizing control system
for a seawater-desalination solar-powered membrane distillation unit, Computers & Chemical
Engineering, 54 (2013) 79-96.
[99] F. Banat, N. Jwaied, Economic evaluation of desalination by small-scale autonomous solar-powered
membrane distillation units, Desalination, 220 (2008) 566-573.
[100] J.D. Gil, L. Roca, A. Ruiz-Aguirre, G. Zaragoza, M. Berenguel, Optimal operation of a Solar
Membrane Distillation pilot plant via Nonlinear Model Predictive Control, Computers & Chemical
Engineering, 109 (2018) 151-165.
[101] Y. Wang, Z. Xu, N. Lior, H. Zeng, An experimental study of solar thermal vacuum membrane
distillation desalination, Desalination and Water Treatment, 53 (2015) 887-897.
[102] Y.-D. Kim, K. Thu, N. Ghaffour, K. Choon Ng, Performance investigation of a solar-assisted direct
contact membrane distillation system, Journal of Membrane Science, 427 (2013) 345-364.
[103] H. Chang, C.-L. Chang, C.-Y. Hung, T.-W. Cheng, C.-D. Ho, Optimization Study of Small-Scale Solar
Membrane Distillation Desalination Systems (s-SMDDS), International Journal of Environmental
Research and Public Health, 11 (2014).
[104] J.-G. Lee, W.-S. Kim, J.-S. Choi, N. Ghaffour, Y.-D. Kim, Dynamic solar-powered multi-stage direct
contact membrane distillation system: Concept design, modeling and simulation, Desalination, 435
(2018) 278-292.
[105] P.A. Hogan, Sudjito, A.G. Fane, G.L. Morrison, Desalination by solar heated membrane distillation,
Desalination, 81 (1991) 81-90.
[106] Z. Ding, L. Liu, M.S. El-Bourawi, R. Ma, Analysis of a solar-powered membrane distillation system,
Desalination, 172 (2005) 27-40.
[107] L.W. Hu, H.Z. Gong, D. Jun Yu, Q. Gao, N. Gao, M. Wang, Y. Yan, Y. Wang, J. Yu, Y. Liu, Diurnal
Variations in Solar Ultraviolet Radiation on Horizontal and Vertical Plane, Iranian Journal of Public
Health, 39 (2010) 70-81.
[108] R.D. Gustafson, J.R. Murphy, A. Achilli, A stepwise model of direct contact membrane distillation
for application to large-scale systems: Experimental results and model predictions, Desalination, 378
(2016) 14-27.
[109] F.P. Incropera, D.P. DeWitt, Introduction to Heat Transfer, 4th ed., John Wiley & Sons, 2002.
[110] S. Lin, N.Y. Yip, M. Elimelech, Direct contact membrane distillation with heat recovery:
Thermodynamic insights from module scale modeling, Journal of Membrane Science, 453 (2014) 498-
515.
[111] F. He, J. Gilron, K.K. Sirkar, High water recovery in direct contact membrane distillation using a
series of cascades, Desalination, 323 (2013) 48-54.
[112] C. Cabassud, D. Wirth, Membrane distillation for water desalination: How to chose an appropriate
membrane?, Desalination, 157 (2003) 307-314.
[113] G. Rao, S.R. Hiibel, A.E. Childress, Simplified flux prediction in direct-contact membrane distillation
using a membrane structural parameter, Desalination, 351 (2014) 151-162.
[114] F. Mansfeld, G. Liu, H. Xiao, C.H. Tsai, B.J. Little, The corrosion behavior of copper alloys, stainless
steels and titanium in seawater, Corrosion Science, 36 (1994) 2063-2095.
[115] H.J. Hwang, K. He, S. Gray, J. Zhang, I.S. Moon, Direct contact membrane distillation (DCMD):
Experimental study on the commercial PTFE membrane and modeling, Journal of Membrane Science,
371 (2011) 90-98.
[116] A. Faanes, S. Skogestad, Buffer Tank Design for Acceptable Control Performance, Industrial &
Engineering Chemistry Research, 42 (2003) 2198-2208.
128

[117] A. Faanes, S. Skogestad, A systematic approach to the design of buffer tanks, Computers &
Chemical Engineering, 24 (2000) 1395-1401.
[118] H.C. Duong, P. Cooper, B. Nelemans, T.Y. Cath, L.D. Nghiem, Evaluating energy consumption of air
gap membrane distillation for seawater desalination at pilot scale level, Separation and Purification
Technology, 166 (2016) 55-62.
[119] H. Lee, F. He, L. Song, J. Gilron, K.K. Sirkar, Desalination with a cascade of cross-flow hollow fiber
membrane distillation devices integrated with a heat exchanger, AIChE Journal, 57 (2011) 1780-1795.
[120] J. Gilron, L. Song, K.K. Sirkar, Design for Cascade of Crossflow Direct Contact Membrane
Distillation, Industrial & Engineering Chemistry Research, 46 (2007) 2324-2334.
[121] Y.-D. Kim, K. Thu, S.-H. Choi, Solar-assisted multi-stage vacuum membrane distillation system with
heat recovery unit, Desalination, 367 (2015) 161-171.
[122] M. Bindels, N. Brand, B. Nelemans, Modeling of semibatch air gap membrane distillation,
Desalination, 430 (2018) 98-106.
[123] R. Tawney, Z. Khan, J. Zachary, Economic and Performance Evaluation of Heat Sink Options in
Combined Cycle Applications, Journal of Engineering for Gas Turbines and Power, 127 (2005) 397-403.
[124] M.J. Rutberg, Modeling water use at thermoelectric power plants, in, Massachusetts Institute of
Technology, 2012.
[125] J. Macknick, R. Newmark, G. Heath, K.C. Hallett, Operational water consumption and withdrawal
factors for electricity generating technologies: a review of existing literature, Environmental Research
Letters, 7 (2012) 045802.
[126] A.K.M. Mohiuddin, K. Kant, Knowledge base for the systematic design of wet cooling towers. Part
I: Selection and tower characteristics, International Journal of Refrigeration, 19 (1996) 43-51.
[127] G. Barigozzi, A. Perdichizzi, S. Ravelli, Wet and dry cooling systems optimization applied to a
modern waste-to-energy cogeneration heat and power plant, Applied Energy, 88 (2011) 1366-1376.
[128] J. Price, M. Zeyringer, D. Konadu, Z. Sobral Mourão, A. Moore, E. Sharp, Low carbon electricity
systems for Great Britain in 2050: An energy-land-water perspective, Applied Energy, 228 (2018) 928-
941.
[129] J. Macknick, R. Newmark, G. Heath, K. Hallett, A review of operational water consumption and
withdrawal factors for electricity generating technologies, Contract, 303 (2011) 3000.
[130] A. Delgado Martín, Water Footprint of Electric Power Generation: Modeling its use and analyzing
options for a water-scarce future, in, Massachusetts Institute of Technology, 2012.
[131] M.H. Sharqawy, V.J.H. Lienhard, S.M. Zubair, On Thermal Performance of Seawater Cooling
Towers, Journal of Engineering for Gas Turbines and Power, 133 (2010) 043001-043001-043007.
[132] M.M.A. Sarker, E. Kim, C.G. Moon, J.I. Yoon, Performance characteristics of the hybrid closed
circuit cooling tower, Energy and Buildings, 40 (2008) 1529-1535.
[133] C.J. Chen, R.Z. Wang, Z.Z. Xia, J.K. Kiplagat, Study on a silica gel–water adsorption chiller integrated
with a closed wet cooling tower, International Journal of Thermal Sciences, 49 (2010) 611-620.
[134] Z.Z. Xia, C.J. Chen, R.Z. Wang, Numerical simulation of a closed wet cooling tower with novel
design, International Journal of Heat and Mass Transfer, 54 (2011) 2367-2374.
[135] A. Cipollina, G. Micale, L. Rizzuti, A brine evaporative cooler/concentrator for autonomous thermal
desalination units, Desalination and Water Treatment, 31 (2011) 269-278.
[136] J.G. Bustamante, A.S. Rattner, S. Garimella, Achieving near-water-cooled power plant performance
with air-cooled condensers, Applied Thermal Engineering, 105 (2016) 362-371.
[137] K. Ohta, I. Hayano, T. Okabe, T. Goto, S. Kimura, H. Ohya, Membrane distillation with fluoro-carbon
membranes, Desalination, 81 (1991) 107-115.
[138] F.A. Banat, J. Simandl, Theoretical and experimental study in membrane distillation, Desalination,
95 (1994) 39-52.
129

[139] F.A. Banat, J. Simandl, Desalination by Membrane Distillation: A Parametric Study, Separation
Science and Technology, 33 (1998) 201-226.
[140] R.A.M. Peer, K.T. Sanders, The water consequences of a transitioning US power sector, Applied
Energy, 210 (2018) 613-622.

Abstract (if available)

Abstract Water stress is expected to increase throughout the world and over time due to growing populations and may be exacerbated by climate change. This problem has motivated interest in water treatment technologies that allow for production of drinking water from non‐traditional water sources through desalination and water reuse. Membrane distillation (MD) has experienced increasing interest for use in desalination and water reuse systems because of its high rejection of non‐volatile contaminants (e.g., salts) and because its driving force is minimally affected by high salinity. MD is a thermally driven water treatment technology and this thermal driving force can be supplied by solar thermal energy or low‐grade waste heat—avoiding the high electrical energy consumption of technologies traditionally used for desalination and water reuse. Although there is significant potential for MD in desalination and water reuse applications, issues associated with long‐term membrane integrity and thermal energy integration must still be addressed. Research on long‐term membrane integrity was accomplished by performing multiple 30‐day continuous MD experiments alongside short‐term thermomechanical creep recovery testing using expanded polytetrafluoroethylene (ePTFE) membranes. Results showed that long‐term use of ePTFE membranes in MD results in reduced water flux over time and significant changes in membrane morphology

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The roles of surface and pore properties in wetting resistance for membrane distillation membranes

PDF

Integrating systems of desalination and potable reuse: reduced energy consumption for increased water supply

PDF

Water desalination: real-time membrane characterization for performance prediction and system analysis for energetic enhancement

PDF

Advancing energy recovery from food waste using anaerobic biotechnologies: performance and microbial ecology

PDF

Wastewater reclamation and potable reuse with novel processes: membrane performance and system integration

PDF

Developing frameworks to quantify the operational and environmental performance of energy systems within the context of climate change

PDF

Fate of antibiotic resistance in anaerobic membrane bioreactors

PDF

Bioelectrochemical treatment of anaerobic process effluents: mitigation of dissolved methane and sulfide

PDF

The formation, fate and transformation of nitromethane in potable reuse processes

PDF

Integrated technologies, blending schemes, and reuse practices to address contaminant and energy challenges in water reclamation

PDF

Catalytic oxidation and precursor identification of disinfection byproducts in recycled water

PDF

High temperature latent heat thermal energy storage to augment solar thermal propulsion for microsatellites

PDF

Identifying and mitigating the effects of urban heat islands in California

PDF

Optimizing biomembrane reactor systems for water reclamation and reuse applications

PDF

Developing high-resolution spatiotemporal methods to model and quantify water use for energy

PDF

Electrokinetic transport of Cr(VI) and integration with zero-valent iron nanoparticle and microbial fuel cell technologies for aquifer remediation

$Hydraulic fracturing and the environment: risk assessment for groundwater contamination from well casing failure$

PDF

Hydraulic fracturing and the environment: risk assessment for groundwater contamination from well casing failure

PDF

Molecular-scale studies of mechanical phenomena at the interface between two solid surfaces: from high performance friction to superlubricity and flash heating

PDF

Investigating the role of urban emission sources on redox-active PM compounds and the chemical analysis of the standardized diesel exhaust particles

PDF

Modeling and analysis of propulsion systems and components for electrified commercial aircraft

Asset Metadata

Creator Gustafson, Ryan David (author)

Core Title Thermally driven water treatment with membrane distillation: membrane performance, waste heat integration, and cooling analysis

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Engineering (Environmental Engineering)

Publication Date 08/22/2019

Defense Date 01/23/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag desalination,membrane distillation,OAI-PMH Harvest,PTFE,waste heat

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Childress, Amy (committee chair), Beshir, Mohammed (committee member), McCurry, Daniel (committee member), Sanders, Kelly T. (committee member), Smith, Adam L. (committee member)

Creator Email rdgustaf@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-128134

Unique identifier UC11675308

Identifier etd-GustafsonR-7112.pdf (filename),usctheses-c89-128134 (legacy record id)

Legacy Identifier etd-GustafsonR-7112.pdf

Dmrecord 128134

Document Type Dissertation

Format application/pdf (imt)

Rights Gustafson, Ryan David

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA