University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2021 Allyson L. McGaughey
The Roles of Surface and Pore Properties in Wetting
Resistance for Membrane Distillation Membranes
by
Allyson L. McGaughey
A Dissertation Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Doctor of Philosophy
Engineering (Environmental Engineering)
May 2021
ii

Acknowledgements
First, I would like to express my deep appreciation for my advisor, Professor Amy Childress, for her
guidance, insight, and support. Her mentorship, commitment to my success, and high standards for our
work have made me the researcher I am today, and I would be hard-pressed to imagine a better advisor. I
am both proud and grateful to have been a member of her research group.
I would also like to thank my dissertation committee: Professors Felipe de Barros, Malancha Gupta, Adam
L. Smith, and Qiming Wang, for their collaboration and feedback on my work, which has been significantly
improved by their insight. I would also like to extend thanks to my fellow graduate students and the
postdoctoral scholars that I worked closely with: Dr. Christopher Morrow, Dr. Ryan Gustafson, Sophia Plata,
Xin Wei, Weijian Ding, Sultan Alnajdi, and Shounak Joshi from the Childress lab, Dr. Prathamesh
Karandikar from the Gupta lab, Dr. Syeed Md Iskander from the Smith lab, as well as all the other members
of the environmental engineering lab at USC for their collaboration and companionship over the past five
years.
Finally, I would like to thank my family, my parents and my siblings for their guidance, support, and
encouragement throughout my education (and life), and my fiancé, Paul, for moving with me to pursue
graduate school, and for his love, patience, and support through these years. I look forward to sharing this
and future milestones with you.

Support
This work was supported by the National Science Foundation under grant number 1820389, the US
Strategic Environmental Research and Development Program under project number ER-2237, and the US
Environmental Protection Agency Science to Achieve Results under grant number 83486701, as well as by
an American Membrane Technology Association and Affordable Desalination Coalition fellowship and a
USC Viterbi Graduate School Ph.D. merit award. Characterization instrumentation and support was
provided by the University of Southern California Core Center of Excellence in Nano Imaging (CNI) and the
University of Southern California Center of Excellence in NanoBioPhysics.

iii

Table of Contents

Acknowledgements ....................................................................................................................................... ii

List of Figures ................................................................................................................................................ v
List of Tables ................................................................................................................................................ vii

Abstract ....................................................................................................................................................... viii

1. Introduction ............................................................................................................................................ 1
1.1 Background and Motivation ............................................................................................................ 1
1.2 Objectives and Scope of Work ....................................................................................................... 3
1.3 Dissertation Organization ............................................................................................................... 4

2. Effect of Long-term Operation on Membrane Surface Characteristics and Performance in Membrane
Distillation ...................................................................................................................................................... 5
Abstract...................................................................................................................................................... 5
2.1 Introduction and Background .............................................................................................................. 5
2.2 Materials and Methods ........................................................................................................................ 8
2.1.1 MD Membranes ............................................................................................................................ 8
2.1.2 Performance Characterization ...................................................................................................... 8
2.1.3 Membrane Characterization ......................................................................................................... 9
2.3 Results and Discussion ..................................................................................................................... 10
2.3.1 Effects of Long-term Operation ................................................................................................... 10
2.3.2. Effect of Short-term Operation with Scaling and Confirmation of Long-term results ................. 17
2.4 Conclusions and Implications ............................................................................................................ 20
Acknowledgements ................................................................................................................................. 21

3. Characterization of Fibril Roughness and Contact Angle to Predict Pore Wetting Resistance in
Membrane Distillation .................................................................................................................................. 22
Abstract.................................................................................................................................................... 22
3.1 Introduction and Background ............................................................................................................ 22
3.1.1 Characterization of Wetting Resistance ..................................................................................... 23
3.1.2 Hypothesis and Objectives ......................................................................................................... 27
3.2 Materials and Methods ...................................................................................................................... 27
3.2.1 Membranes ................................................................................................................................. 27
3.2.2 Material Characterization ............................................................................................................ 28
3.2.3 Hydrophobicity Characterization ................................................................................................. 29
3.2.4 MD Performance Characterization ............................................................................................. 29
3.2.5 LEP Characterization .................................................................................................................. 31
iv

3.3 Results and Discussion ..................................................................................................................... 32
3.3.1 Membrane Characterization ....................................................................................................... 32
3.3.2 LEP Characterization .................................................................................................................. 36
3.4 Conclusions and Implications ............................................................................................................ 38
Acknowledgements ................................................................................................................................. 40

4. Hydrophobicity versus Pore Size: Polymer Coatings to Improve Membrane Wetting Resistance for
Membrane Distillation .................................................................................................................................. 41
Abstract.................................................................................................................................................... 41
4.1 Introduction and Background ............................................................................................................ 41
4.1.1 Current Understanding of Membrane Wetting ............................................................................ 42
4.1.2 Membrane Modification .............................................................................................................. 45
4.1.3 Objectives ................................................................................................................................... 47
4.2 Materials and Methods ...................................................................................................................... 47
4.2.1 Membranes and Membrane Modification ................................................................................... 47
4.2.2 Characterization .......................................................................................................................... 48
4.3 Results and Discussion ..................................................................................................................... 51
4.3.1. Surface Characteristics of Uncoated and Coated Membranes ................................................. 51
4.3.2. Liquid Entry Pressure ................................................................................................................ 54
4.3.3. Scaling-induced Wetting Performance ...................................................................................... 55
4.3.4. Surfactant-induced Wetting Performance .................................................................................. 58
4.4 Conclusions and Implications ............................................................................................................ 60
Associated Content ................................................................................................................................. 61
Supporting Information......................................................................................................................... 61
Acknowledgements ................................................................................................................................. 61

5. Conclusions ......................................................................................................................................... 62

References .................................................................................................................................................. 65

Appendix A: Supplementary Information .................................................................................................... 74
A.1 Supplementary Information to Chapter 3 .......................................................................................... 74
A.1.1 Characterizing Wetting in MD ..................................................................................................... 74
A.2 Supplementary Information to Chapter 4 .......................................................................................... 77
A.2.1 Materials and Methods ............................................................................................................... 77
A.2.2 Membrane Characterization ....................................................................................................... 77
A.2.3 Supporting References ............................................................................................................... 81

v

List of Figures

Figure 1.1 Schematic illustrating membrane distillation process ................................................................. 2

Figure 2.1 Schematic of bench-scale DCMD system .................................................................................. 8
Figure 2.2 Water flux and distillate conductivity versus time for the long-term experiment ....................... 11
Figure 2.3 Representative FESEM images of (a) virgin membrane, (b) feed side of used membrane, and
(c) distillate side of used membrane and EDS spectra of (d) virgin membrane, (e) feed side of used
membrane, and (f) distillate side of used membrane for the long-term experiment ................................... 12
Figure 2.4 Representative tapping mode AFM images of (a) virgin membrane, feed side of used membrane,
and (c) distillate side of used membrane for the long-term experiment ..................................................... 16
Figure 2.5 Water flux and distillate conductivity versus time for the short-term experiment ...................... 18
Figure 2.6 Representative FESEM images of (a) virgin membrane, (b) feed side of used membrane, and
(c) distillate side of used membrane and EDS spectra of (d) virgin membrane, (e) feed side of used
membrane, and (f) distillate side of used membrane for the short-term experiment .................................. 19

Figure 3.1 Illustration of a) Young’s state and intrinsic contact angle 𝜃 0
on a smooth, hydrophobic material,
b) Cassie-Baxter state and measured contact angle 𝜃 𝑚 on an idealized membrane surface, and c) Wenzel
state and 𝜃 𝑚 on an idealized membrane surface........................................................................................ 25
Figure 3.2 Fibril contact angle 𝜃 𝑓 formed at the three-phase interface at a pore entrance (illustrated as a
cross-section of membrane fibrils) in the Wenzel state, with non-negligible fibril roughness ..................... 27
Figure 3.3 SEM images at x5,000 magnification (10-μm scale bar) showing membrane surface morphology
of the feed side of the unused a) QM050, b) QM022, and c) QL822 membranes and at x50,000
magnification (1-μm scale bar) showing fibril surface morphology of the d) QM050, e) QM022, and f) QL822
membranes ................................................................................................................................................. 28
Figure 3.4 Baseline MD performance of MD membranes in terms of a) water flux (𝐽 𝑊 ) and b) distillate
concentration (𝐶 𝑑 ) and salt flux (𝐽 𝑆 ) versus time ......................................................................................... 31
Figure 3.5 3-D AFM images and average surface roughness (Sa) of the feed side at 5 x 5 μm area for the
a) QM050, b) QM022, c) QL822 membranes, at the maximum pore scale for the d) QM050, e) QM022, and
f) QL822 membranes, and at the nominal pore scale for the g) QM050, h) QM022, and i) QL822
membranes ................................................................................................................................................. 33
Figure 3.6 Experimental LEP results for the QM022 membrane in terms of a) 𝛥𝑃 and 𝐶 𝑑 versus time, and
b) 𝛥 𝐶 𝑑 /𝛥𝑡 versus time ................................................................................................................................. 37
Figure 3.7 Illustration of the surface fibril contact angle 𝜃 𝑓 ,𝑆 at the feed side of a membrane and the internal
fibril contact angle 𝜃 𝑓 ,𝑖 formed at a pore throat within a membrane for a) an unwetted membrane and b) a
partially wetted membrane after fouling/scaling occurs .............................................................................. 39

Figure 4.1 Representative SEM images of a) top and b) bottom of the uncoated CA membrane; c) top and
d) bottom of the coated CA membrane; e) top and f) bottom of the uncoated PTFE membrane; and g) top
and h) bottom of the coated PTFE membrane ........................................................................................... 52
Figure 4.2 LEP distributions for the top side of the a) uncoated PTFE, b) coated PTFE, and c) coated CA
membranes ................................................................................................................................................. 55
vi

Figure 4.3 Performance of the uncoated PTFE, coated PTFE, and coated CA membranes with 5 M NaCl
feed solution in terms of a) water flux, b) distillate conductivity, and c) solute rejection ............................ 56
Figure 4.4 Sensitivity analysis of 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
for heterogeneous NaCl nucleation from 5 M solution at 25°C to
contact angle (𝜃 ) and surface porosity (𝜖 ) .................................................................................................. 58
Figure 4.5 Performance of uncoated PTFE, coated PTFE, and coated CA membranes with 200 ppm
surfactant in 1 M NaCl feed solution in terms of a) water flux, b) distillate conductivity, and c) solute
rejection ....................................................................................................................................................... 59
Figure 4.6 Sensitivity analysis of LEP to contact angle (𝜃 ) and pore diameter (𝑑 𝑝 ). Theoretical calculations
are based on pure water at 25 °C ............................................................................................................... 60

vii

List of Tables

Table 2.1 Contact angles measured on the virgin and used membrane before and after the long-term
experiment................................................................................................................................................... 13
Table 2.2 Roughness values for membrane used in the long-term experiment ......................................... 16
Table 2.3 Contact angles measured on membrane used in the short-term experiment ............................ 20

Table 3.1 Material characterization results for the three membranes. ....................................................... 29
Table 3.2 Measured contact angles (𝜃 𝑚 ) for all membranes. ..................................................................... 29
Table 3.3 Measured 𝑅 ′ values at the membrane (5-μm) scale and at the maximum and nominal pore scales
for the feed side of the unused membranes ............................................................................................... 34
Table 3.4 Wenzel contact angles at the membrane (5-μm) scale (𝜃 𝑊 ,𝑚 ) and fibril contact angles (𝜃 𝑓 ) at the
maximum and nominal pore scale for the feed side of the unused membranes ........................................ 35
Table 3.5 LEP results from the Kim-Harriott model based on the intrinsic contact angle (𝜃 0
), 𝜃 𝑓 , and
measured contact angle (𝜃 𝑚 ) for all membranes ........................................................................................ 36

Table 4.1 Root-mean-square 2-D surface roughness and contact angle values of uncoated and coated
membranes ................................................................................................................................................. 51
Table 4.2 Surface pore parameters of the uncoated and coated membranes. .......................................... 52
Table 4.3 Percent composition results from XPS analysis of uncoated and coated membranes .............. 54
Table 4.4 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
of the uncoated PTFE, coated PTFE, and coated CA membranes ............................. 56

viii

ABSTRACT

High-salinity waste streams produced during desalination of saline, brackish, and waste waters or from
industrial processes present unique treatment challenges, and their management is associated with high
capital and operating costs and high energy consumption. Membrane distillation (MD) is a thermally driven
membrane process that is promising for the treatment of complex, high-salinity streams. While reverse
osmosis is salinity-limited, MD is viable for hypersaline streams because MD relies on a driving force that
is only slightly decreased by salinity. While conventional thermal desalination processes are associated
with low energy efficiency and high cost, MD can more effectively utilize low-grade thermal energy, achieve
higher energy efficiency at smaller capacities, and require lower capital costs.
In MD, a warmer, saline stream flows on one side and a cooler, pure distillate stream flows on the other
side of a microporous, hydrophobic membrane. The vapor pressure difference across the membrane drives
water in the feed stream to evaporate, diffuse through the membrane pores, and condense into the cooler
distillate stream. The resistance of membrane pores to wetting results in a vapor gap that separates the
feed and distillate streams and prevents liquid penetration. Thus, membrane wetting resistance is critical
to prevent passage of nonvolatile solutes. High wetting resistance is particularly important for high-salinity
streams due to the potential for salt precipitation (scaling) on the membrane surface, which can cause pore
wetting. Surfactants and alcohols, which may be present in industrial waste streams, can also cause wetting
due to reduced surface tension at pore entrances. MD has not been widely implemented for the
management of challenging waste streams – in large part due to insufficient membrane wetting resistance
upon exposure to high salinity streams and reduced wetting resistance over long-term operation.
This dissertation provides fundamental understanding of pore wetting in MD and introduces new
characterization parameters to describe pore wetting resistance in situ. Wetting resistance depends on two
membrane properties: hydrophobicity and pore size. While the relationship between hydrophobicity, pore
size, and wetting resistance can appear simple, there are complexities that have not been previously
addressed. Membrane hydrophobicity may not be constant during long-term operation – due to fouling or
physical and chemical changes to the membrane itself. In the literature, hydrophobicity has frequently been
represented by the measured, feed-side contact angle of the membrane or the intrinsic contact angle of the
smooth membrane material; the roles of wetting state and fibril roughness at pore entrances have not been
considered with respect to wetting resistance. Liquid entry pressure (i.e., the transmembrane pressure
resulting in liquid penetration) has been represented by the largest membrane pore without considering the
distribution of pore sizes characteristic of real membranes. Finally, in the literature, efforts to improve
membrane wetting resistance have almost entirely focused on modifying the feed-side surface contact
ix

angle. The potential for improving wetting resistance by modifying distillate-side hydrophobicity, internal
hydrophobicity, pore size, and pore size distribution has not been considered.
In the first study, characterization of membrane hydrophobicity before and after long-term operation
revealed that hydrophobicity of polytetrafluoroethylene membranes does not remain constant during
operation, even in the absence of significant fouling. There is significant interest in recovering membrane
hydrophobicity after fouling has occurred. Typically, membrane hydrophobicity is assumed to be
recoverable if foulants can be removed. This was the first study to show that irreversible loss of membrane
hydrophobicity occurred after long-term use. Permanent reductions in hydrophobicity due to physical or
chemical changes to the membrane itself can result in permanent reductions in membrane performance
and lifetime. The effects of long-term operation may be especially significant for membranes with surface
coatings or thin selective layers, which may be more vulnerable than the internal layers and distillate-side
surface. At the same time, it was found that rejection remained high. When fouling is present, distillate-side
and internal hydrophobicity may be equally or more important than feed-side hydrophobicity, as the internal
layers and distillate-side surface are less affected by foulants that can effectively hydrophilize the feed-side
surface.
In the second study, the role of hydrophobicity in wetting resistance was further investigated. In using the
measured contact angle to indicate wetting resistance, it is implicitly assumed that macroscopic droplets
used for contact angle measurement are (1) representative of pore wetting that occurs on a microscopic
scale, and (2) in the same wetting state as feed solution at the membrane surface during MD operation (in
situ). In using the intrinsic contact angle to indicate wetting resistance, it is implicitly assumed that fibril
roughness is negligible. In this study, it is shown that the assumptions for using either measured or intrinsic
contact angle may not be valid for hydrophobic polytetrafluoroethylene membranes at typical operating
pressures – especially when membranes are exposed to low surface-tension feed solutions, fouling, or
scaling. The fibril contact angle of polytetrafluoroethylene membranes was introduced and calculated based
on the Wenzel model and measured fibril roughness. Results show that while measured contact angles
overestimate wetting resistance, the fibril contact angle may provide better prediction of in situ pore wetting
resistance for MD membranes. The fibril contact angle can be used to predict pore wetting resistance for
any hydrophobic membrane, based on intrinsic contact angle and fibril roughness. Furthermore, the fibril
contact angle may provide a quantitative estimation of the internal hydrophobicity of polytetrafluoroethylene
membranes. These results provide fundamental understanding of pore wetting resistance and have
important implications for the design of wetting-resistant MD membranes, especially for challenging
applications.
In the third study, the relative importance of hydrophobicity and pore size were investigated for surfactant-
and scaling-induced wetting in MD. In the literature, efforts to improve wetting resistance have focused
nearly entirely on modifying the feed-side measured contact angle; the potential for reduced pore size to
increase resistance to scaling- and surfactant-induced wetting has not been investigated. Results show that
x

while reduced pore size/porosity and increased surface hydrophobicity resulted in similar resistance to
surfactant-induced wetting, reduced pore size/porosity provided greater wetting resistance to scaling-
induced wetting – likely due to the prevention or reduction of internal scaling. However, because reduced
porosity negatively impacts permeability, a permeability/wetting resistance trade-off results, which is
especially relevant for high-salinity streams. Also, liquid-entry-pressure distribution, arising from pore-size
distribution, was introduced. Results show that the minimum liquid entry pressure, which corresponds to
the wetting resistance of the single, largest pore, does not necessarily characterize wetting resistance in
situ. In scaling-induced wetting, salt crystals may not form at the largest pore; in surfactant-induced wetting,
module flow patterns may result in concentration polarization and reduced surface tension that drive wetting
at pores other than the largest pore. In both cases, multiple pores likely become wetted before solute
passage is significant enough to detect wetting. Therefore, the liquid entry pressure distribution must be
considered to fully describe the wetting resistance of membranes in situ.
This work combined fundamental scientific principles with engineering approaches to achieve systematic
understanding of wetting resistance in MD. The results provide understanding of wetting mechanisms,
membrane wetting resistance, and membrane durability in MD – both during long-term operation and for
challenging, high-salinity streams. This work is increasingly relevant as freshwater scarcity and impairment
drive suppliers to exploit alternative sources, including challenging, high-salinity streams. Although MD is
theoretically well-suited to treating high-salinity streams, a lack of ability to predict and prevent membrane
wetting has limited adoption. These results not only further fundamental understanding of wetting resistance
but can also guide MD membrane design. Increasing the roughness of surface and internal fibrils to
increase fibril hydrophobicity may enable wetting resistance even if feed side hydrophobicity is reduced –
either slowly, during long-term operation, or due to fouling/scaling. Reducing surface pore size while
maintaining high porosity and low thickness may enable MD membranes to achieve both high permeability
and high wetting resistance – especially for high-salinity streams. Finally, controlling pore size distributions
to skew liquid entry pressure distributions towards higher values may enable more robust wetting resistance
for challenging applications. Robust wetting resistance is critical to make MD viable as a sustainable
method for the management and reclamation of challenging waste streams.

1

CHAPTER 1.

1. Introduction
1.1 Background and Motivation
Desalination of challenging, high-salinity brines has attracted recent interest, both for reclamation of
valuable freshwater resources and to reduce negative environmental impacts and costs associated with
brine disposal [1-3]. Brine management can be challenging and costly, accounting for 5-33% of total
desalination costs [4]. Global desalination currently results in the production of approximately 141.5 million
m
3
/day of brine [5]; while brine quality varies depending on the feed water, recovery rate, and chemical
additives used during treatment, ~70 g/kg salinity is typical of seawater desalination [3, 6]. While coastal
desalination brines can often be discharged to an ocean or other receiving body, inland brines are especially
challenging, often requiring costly alternatives such as deep well injection, evaporation ponds, or other zero
liquid discharge processes. In the US, shale gas production results in the inland generation of large volumes
of produced water with elevated salinities of up to 400 g/kg [6]. Brine disposal is also associated with
negative environmental impacts. Due to high salinity and density, direct disposal of brines can result in
depletion of dissolved oxygen and transport of chemical contaminants to the ocean floor; deep well injection
has been associated with seismicity, increased regulation, and limitations on available sites; evaporation
ponds require large amounts of available land and may contaminate groundwater [1-7]. Costs,
environmental concerns, and regulation of brine disposal has driven investment in the treatment and reuse
of high-salinity brines.
Currently, the most energy efficient process for desalination is reverse osmosis (RO) [1, 8]; RO accounts
for nearly 70% of global desalination capacity [1, 5]. In RO, an applied hydraulic pressure exceeding the
osmotic pressure forces water from a saline stream through a semipermeable membrane. For high salinity
solutions, applied hydraulic pressures to achieve sufficient water flux exceed practical limits of system
components [6, 8, 9]. Thus, RO is salinity limited – currently, RO cannot be used to treat streams with
salinities greater than ~80 g/kg [1] and may not be economical for produced water with salinity greater than
45 g/kg [7]. Thermal processes that rely on a phase change to desalinate water can overcome this limitation
because, compared to osmotic pressure, water vapor pressure is only slightly dependent on salinity [10].
Thermal (or other phase-change based) desalination processes are therefore necessary to manage high-
salinity brines [1, 6, 11] – however, the energy efficiency of thermal processes is inherently lower than that
of RO because the energy associated with evaporation (i.e., the latent heat) generally exceeds the specific
Gibbs free energy of separation by orders of magnitude [1, 8]. Therefore, interest in thermal technologies
that can treat high salinity brines while mitigating energy consumption and costs, is growing.
2

MD is a thermally driven separation process that offers several unique benefits relative to conventional
thermal processes, such as multi-effect and multi-stage flash distillation, used for high-salinity treatment
applications. MD can utilize low-grade thermal energy, such as “waste” heat from the power and industrial
sectors, low-grade solar thermal energy, and geothermal power [1, 6, 12]. MD has a small footprint, is
modular, and has relatively low capital costs [1, 6]. MD can also be more energy efficient than competitive
thermal desalination technologies, especially for smaller system capacities [1]. MD is also associated with
less severe fouling than RO, because MD utilizes a porous membrane under no applied hydraulic pressure
[1, 6]. MD has therefore attracted interest for the management of desalination brines, produced waters, and
other high-salinity streams [1, 4, 6, 13, 14]. Still, MD has not been widely implemented due to remaining
challenges, including pore wetting at high salinities and uncertainties about the robustness and
sustainability of MD for long-term performance.
In MD, a saline feed stream and pure distillate stream flow on either side of a hydrophobic, microporous
membrane (Figure 1.1). The temperature difference between the two streams results in a transmembrane
vapor pressure difference, which drives water from the feed stream to evaporate, diffuse through membrane
pores, and condense into the distillate stream. When unwetted, the membrane pores support a vapor gap
between the feed and distillate streams that prevents passage of nonvolatile solutes and particulate matter
present in the feed stream. When wetted, the feed and distillate streams can freely mix and solute rejection
is lost [1, 13, 14]. Thus, membrane wetting resistance is critical to the MD process. Wetting resistance is
typically characterized by membrane hydrophobicity (via the measured or intrinsic contact angle of a
membrane) and by liquid entry pressure (i.e., the minimum transmembrane hydraulic pressure at which
pore wetting occurs).

Figure 1.1 Schematic illustrating membrane distillation process.

3

1.2 Objectives and Scope of Work
The overall objective of this dissertation is to provide fundamental understanding of pore wetting in MD,
under different wetting conditions and during long-term operation, and to develop new characterization
methods and parameters to describe pore wetting resistance in situ. Wetting resistance depends on two
membrane properties: hydrophobicity and pore size; and while the relationship between hydrophobicity,
pore size, and wetting resistance can appear simple, there are several complexities that have not been
previously addressed. Membrane hydrophobicity may not be constant during long-term operation – due to
fouling or physical and chemical changes to the membrane itself. In the literature, hydrophobicity has
frequently been represented by the measured, feed-side surface contact angle of an unused membrane or
by the intrinsic contact angle of the smooth, ideal membrane material; the roles of wetting state and fibril
roughness at pore entrances have not yet been considered with respect to wetting resistance. Furthermore,
wetting resistance, characterized by liquid entry pressure, has been represented by the maximum pore size
without considering the distribution of pore sizes characteristic of real membranes. Finally, in the literature,
efforts to improve membrane wetting resistance have almost entirely focused on modifying the feed-side
surface contact angle. The potential for improving wetting resistance by modifying distillate-side
hydrophobicity, internal hydrophobicity, pore size, and pore size distribution has not been considered.
In the first study, the long-term performance of an MD membrane was tested and membrane hydrophobicity
was characterized before and after use. The objective of the study was to evaluate changes in membrane
performance, surface morphology, and surface hydrophobicity after long-term operation with saline feed
solutions, in order to isolate the effect of salinity on performance and membrane properties. Both the feed
and distillate side of the membrane – which has been previously overlooked – were characterized before
and after long-term operation. Results provide further understanding of the impact of long-term operation
on membrane surface characteristics, the potential relevance of distillate-side and internal membrane
characteristics to rejection, and the implications of observed changes to MD membrane design.
In the second study, the role of hydrophobicity in wetting resistance was further explored. Either measured
or intrinsic contact angles have been used to indicate wetting resistance in the literature. However, for
measured contact angles may reflect macroscopic wetting in a different wetting state than occurs during in
situ MD, and intrinsic contact angles do not consider fibril roughness at pore entrances. The objective of
this study was to develop a method to characterize the wetting resistance at the pore entrance of MD
membranes using the fibril contact angle, calculated based on the Wenzel model and measured fibril
roughness. The relevance of the fibril contact angle to wetting resistance in situ is investigated by comparing
model and experimental results for liquid entry pressure based on the intrinsic, fibril, and measured contact
angles.
In the third study, the objective was to, for the first time, investigate the relative importance of hydrophobicity
and pore size under surfactant- and scaling-induced wetting conditions via initiated chemical vapor
deposition of hydrophobic coatings on commercial membrane substrates. Coating conformality was
4

compared for hydrophilic and hydrophobic substrates, and the uncoated and coated membranes were
tested to assess their performance in MD. For the first time, the effectiveness of reduced pore size/porosity
was compared to increased hydrophobicity to prevent scaling- and surfactant-induced wetting in MD. A
novel liquid entry pressure distribution parameter, based on the pore size distribution, was introduced in
order to more comprehensively characterize wetting resistance than by minimum liquid entry pressure
values alone. Lastly, the permeability/wetting resistance trade-off for MD membranes is considered with
respect to membrane modifications that improve resistance to scaling- and surfactant-induced wetting.
1.3 Dissertation Organization
This dissertation consists of a compilation of papers written during the course of the dissertation research.
Chapter 2 is an entire paper manuscript that was published in the Journal of Membrane Science. Chapter
3 is a draft manuscript for a paper in the final stages of preparation. Chapter 4 is an entire paper manuscript
that was published in the journal ACS Applied Polymer Materials.

5

CHAPTER 2.

2. Effect of Long-term Operation on Membrane Surface Characteristics
and Performance in Membrane Distillation

Reprinted (adapted) with permission from A. McGaughey, R. Gustafson, and A. Childress, "Effect of long-
term operation on membrane surface characteristics and performance in membrane distillation," Journal of
Membrane Science, vol. 543, pp. 143-150, 2017. Copyright 2017 Elsevier.

Abstract
In this study, significant changes to surface morphology and decreased surface hydrophobicity were
observed on both the feed and distillate sides of membrane distillation membranes after 100 days of
operation. Contact angles decreased by 56% and 26% on the feed and distillate sides, respectively. Surface
roughness also decreased by 92% and 57% on the feed and distillate sides, respectively. Moderate
morphological changes were also observed after 20 days of operation. While decreased hydrophobicity
and surface roughness on the feed side were associated with fouling/scaling deposits and not changes to
the actual membrane surface, decreased hydrophobicity and surface roughness on the distillate side
indicated changes to the actual membrane surface. Often, membrane hydrophobicity is assumed to be
recoverable if foulants can be removed; however, if membrane hydrophobicity decreases due to physical
changes in the membrane surface, hydrophobicity may not be fully recoverable and membrane lifetime may
be reduced. Despite significant reductions in feed-side hydrophobicity, distillate conductivity remained low,
indicating that other membrane characteristics, such as distillate-side and internal or pore wall
hydrophobicity, may play an important role in maintaining rejection during long-term operation.

2.1 Introduction and Background
Membrane distillation (MD) is a thermally driven water treatment process whereby a warmer feed stream
flows on one side of a hydrophobic, microporous membrane and a cooler distillate stream flows on the
other side. The temperature difference across the membrane creates a vapor pressure difference that
causes liquid water to evaporate from the feed stream, pass through the membrane pores, and condense
into the distillate stream. MD has the potential to produce high quality water while operating at low
temperature differences that are achievable using waste heat and/or renewable energy sources [10, 15,
16]. In direct contact MD (DCMD), the simplest configuration of MD, both liquid streams directly contact
the membrane surfaces. Non-volatile species are prevented from passing through the membrane under
normal operation if membrane hydrophobicity is sufficient to prevent liquid from wetting the membrane
pores [17, 18]. Thus, membrane hydrophobicity is a key parameter for assuring rejection of non-volatile
species.
6

MD membranes are typically polymeric, often composed of polytetrafluoroethylene (PTFE), polyvinylidene
fluoride (PVDF), and/or polypropylene (PP) [18]. Hydrophobicity is usually highest for PTFE membranes
[17, 18]. PTFE membranes are commonly characterized by a surface structure consisting of strands of
polymer material called fibrils, which connect at junctions or “nodes” [19]. In general, the surface
morphology of a PTFE membrane can be characterized by surface roughness and by the size, shape, and
form of the fibrils, nodes, and pores that make up the surface [19, 20]. Membrane fibrils also contribute to
surface roughness [21]. Surface roughness, in turn, affects surface hydrophobicity [22] and membrane
performance.
Because the vapor pressure driving force in MD does not decrease significantly for feed streams with higher
total dissolved solids (TDS) concentrations, MD membranes are being considered for a range of moderate-
to high-salinity applications [10, 23-26]. Some of these applications have complex solution chemistries
containing organic species; for this reason, they are often considered high-fouling applications. In these
cases, the presence of fouling and scaling can obfuscate observation of the actual membrane surface;
however, characterization of the membrane surface may not be critical as the feed solution would more
likely interact with the foulant/scalant layer than with the membrane itself [27]. On the other hand, in low-
fouling applications, characterizing changes in membrane surface morphology with time is critical for
understanding long-term performance. Low-fouling applications may be defined as applications with feed
streams that consist primarily of inorganic contaminants that are not present at high enough concentrations
to exceed their solubility and form scale; these applications include effluent polishing (removal of targeted
contaminants that pass through other treatment processes) and re-concentration of draw solutions used in
forward osmosis systems [23, 28, 29]. To achieve sustainable, long-term performance in these applications,
membrane hydrophobicity must be maintained; however, there are few long-term studies in the literature
to support or dispute the frequent, implicit assumption that hydrophobicity remains constant in the absence
of fouling.
Most bench-scale MD studies have durations on the order of minutes to hours and focus on short-term flux
or initial membrane performance. Few bench-scale studies with low-fouling feed solutions have been
carried out over long time periods [30-38]. Of these, only a small subset [34, 36, 39] mention possible
changes in membrane surface characteristics that occur during operation and suggest a relationship
between changes in membrane surface morphology and performance. No studies have systematically
evaluated relatively unfouled MD membrane surfaces after exposure to low-fouling solutions.
In pilot-scale studies, operation over longer time periods is more common; however, due to the complexities
of extracting membranes from modules to perform autopsies, consideration of the relationship between
long-term performance and membrane surface characteristics is even less common than in bench-scale
studies, with a few exceptions. In 2011, Guillen-Burrieza et al. [40] studied the performance and efficiency
of a pilot-scale solar-powered MD system treating marine salt solutions of 1 and 35 g/L TDS with
commercial PTFE membranes. For a period of four months, the system was operated during daylight hours
7

and shut down overnight. Distillate conductivity increased due to membrane wetting and salt passage. An
integrity test showed that the membrane lacked holes, suggesting to the authors that wetting occurred due
to a decrease in membrane hydrophobicity, although surface characterization was not able to be performed
to verify this [40]. In 2014, Guillen-Burrieza et al. [41] examined membrane performance and the efficacy
of cleaning procedures for a PTFE-PP composite membrane used in a pilot-scale solar MD system to treat
a 35 g/L marine salt solution. The system was operated during the day and shut down overnight for a total
of 17 days. Pore wetting occurred, shown by increasing distillate conductivity. Various membrane surface
characteristics, including contact angle and surface morphology, were compared for the virgin and used
membranes. Scaling on the feed side of the membrane was associated with fibril damage [41]. Cleaning
procedures were applied at the laboratory scale to samples of the used membrane. Contact angle recovery
varied with cleaning solution but did not reach virgin membrane values; it was suggested that membrane
structural damage due to operation and/or cleaning procedures had occurred [41]. To the best of the
authors’ knowledge, these are the only studies that have examined how membrane surface characteristics
are affected by long-term operation and in turn, affect long-term salt rejection.
No long-term studies have yet examined the role of distillate-side surface characteristics in MD
performance. Studies on composite hydrophobic-hydrophilic MD membranes have demonstrated good
short-term performance and durability (e.g., [42-46]), and improved performance has been observed for an
MD membrane with hydrophilic feed-side and hydrophobic distillate-side surfaces [45]. This may indicate
that distillate-side characteristics or internal material characteristics are relevant to the pore wetting
process; however, this possibility has not yet been fully explored in the literature. Internal or pore wall
hydrophobicity of aquaporin nanopores and carbon nanotubes has previously been studied theoretically
via molecular dynamics simulations (e.g., [47, 48]). The concept has also been mentioned for
inhomogeneous fuel cell materials [49, 50]. However, internal hydrophobicity of porous materials is not yet
experimentally accessible [50]. Guillen-Burrieza et al. [41] found that membranes with a thin PTFE layer on
a less hydrophobic PP support were significantly wetted during long-term use despite a feed-side contact
angle greater than 90°. The authors suggested that this may have been due to active layer compression
facilitating liquid “bridging” through membrane pores [41]. Thus, simply having a hydrophobic layer facing
the feed solution may not be enough to ensure that membrane pores remain dry, especially for long-term
operation.
The overall objective of this study is to evaluate changes in membrane surface hydrophobicity, surface
morphology, and performance after long-term operation. Only saline feed solutions are used in order to
clearly identify the effect of moderate- to high-salinity solutions on long-term MD membrane performance.
Furthermore, the distillate side of the membrane, which has been largely overlooked in previous studies, is
characterized after long-term operation. The results of this study are used to better understand the impact
of long-term operation on membrane surface characteristics, the potential relevance of distillate-side
8

membrane characteristics to long-term salt rejection, and the implications of these phenomena for MD
membrane design.

2.2 Materials and Methods
2.1.1 MD Membranes
Commercial flat-sheet expanded polytetrafluoroethylene (PTFE) microfiltration membranes (Clarcor,
Franklin, TN) were used for all long-term testing. The membranes are hydrophobic, single-layer, and
symmetric. The average pore size of the membrane surface was found to be 0.3 ± 0.17 μm, based on
analysis of ten field emission scanning electron microscope (FESEM) images using ImageJ software
(version 1.49, National Institutes of Health, Bethesda, MD). Nearly 90% of the pores were less than 0.5 μm
in diameter and 61% of the pores had diameters of 0.2 ± 0.06 μm.
2.1.2 Performance Characterization
Membrane performance was tested in a bench-scale DCMD system designed for continuous long-term
operation. A schematic of the system is shown in Figure 2.1. To eliminate corrosion and rust in the system
during long-term experiments, as has been reported in other long-term MD studies (e.g., [41], [51]), all
system components were non-metallic with the exception of a titanium heat exchanger, which is highly
resistant to corrosion [52], on the distillate side.

Figure 2.1 Schematic of bench-scale DCMD system.
For all performance tests, the feed and distillate solutions were circulated at 1.5 L/min using gear pumps
(Pulsafeeder, Rochester, NY.) The feed and distillate temperature set points were held at 65 and 35 °C
using proportional integral derivative controllers for the inline heater and chiller. Feed concentrations of 35
and 200 g/L NaCl were used for the long- and short-term experiments, respectively. ACS-grade NaCl
9

(VWR, Radnor, PA) and deionized (DI) water were used for all solutions. Membranes were housed in a
custom-built, clear, acrylic module, providing continuous visual observation of the membrane, with an active
membrane area of 0.014 m
2
. 1.7-mm mesh spacers (Sterlitech, Kent, WA) were used on either side of the
membrane for support and to increase turbulence in both the feed and distillate channels [53].
Flow rate and conductivity were continuously monitored using digital flow meters (Omega, Norwalk, CT)
and in-line conductivity probes and transmitters (Cole Parmer, Vernon Hills, IL) on both the feed and
distillate streams, as shown in Figure 2.1. Conductivity measurements were used to characterize salt
rejection. Resistance temperature detectors (Omega, Norwalk, CT) were used to measure the temperature
at the inlets and outlets of the membrane module on both the feed and distillate sides. Flux through the
membrane was monitored by continuously weighing distillate overflow on an analytical balance (Ohaus,
Parsippany, NJ). Data from the probes and scale were recorded at 10-min intervals using a data acquisition
and control program developed in LabView (National Instruments, Austin, TX).
2.1.3 Membrane Characterization
Contact angle, morphology, chemical composition, and roughness of the membrane surface were
characterized prior to and after both experiments. Contact angles of the membrane surface were measured
using the sessile drop method and a goniometer (Model 500, Ramé-Hart, Succasunna, NJ). Prior to use,
contact angles were measured on a minimum of three virgin membrane samples; measured values did not
vary significantly for the different samples or with the side of the symmetric membrane. Following use,
contact angles were measured on three samples of approximately 1 ⨯ 2 cm cut from the membrane. For
all contact angle measurements, 5-μL droplets of DI water were carefully deposited onto the membrane
surface using an automated pipette. Droplets were deposited on different areas of each sample and
reported values represent an average of at least ten measurements.
Surface morphology of both sides of the virgin and used membrane was evaluated using an FESEM (JSM-
7001, Jeol USA, Huntington Beach, CA). Prior to imaging, samples were sputter-coated in gold and
palladium to render the surface conductive. Energy-dispersive X-ray spectroscopy (EDS) was also used for
semi-quantitative characterization of the chemical composition of the membrane surface. Membrane
samples were cut adjacent to the membrane sample cut for performance testing. FESEM-EDS results
showed that there were no significant differences in surface morphology or elemental composition between
the feed and distillate sides of unused membranes. Following performance testing, FESEM and EDS
measurements on both the feed and distillate sides were taken on samples of approximately 1 ⨯ 1 cm cut
from the used membrane. ImageJ software was used to analyze FESEM images for fibril/node spacing.
Membrane surface roughness was characterized using an atomic force microscope (AFM) (Innova, Bruker,
Billerica, MA). Initial measurements were taken on both sides of the unused membrane, with samples cut
adjacent to the membrane sample for performance testing. Measurements on the unused membrane
showed that there were no significant differences in roughness between different sides of the unused
10

membrane. Following performance testing, roughness measurements were taken on samples of
approximately 1 ⨯ 1 cm cut from the used membrane. Samples were affixed to mounting disks using
double-sided carbon tape and 10 ⨯ 10 μm

areas of each sample were scanned in tapping mode using an
antimony-doped silicon probe (FMV-A, Bruker, Billerica, MA). Images were analyzed using Gwyddion
software (version 2.47, Czech Metrology Institute, Brno, Czech Republic) and roughness parameters were
obtained by averaging horizontal line profiles for each image.

2.3 Results and Discussion
2.3.1 Effects of Long-term Operation
Membrane Performance
Water flux and distillate conductivity versus elapsed time for the long-term (100-day) experiment with 35
g/L NaCl feed solution are shown in Figure 2.2. Flux decreased fairly linearly with time over the 100 days
of continuous operation, with a correlation coefficient of 0.9; overall, flux decreased by 16%. The cause for
the decrease in flux was not apparent at first; eventually, a brown-colored fouling layer became visible on
the membrane surface. It should be noted that low-fouling feed solutions are not necessarily “clean” and at
least some degree of fouling is likely to occur over long time periods. Although no foulants were added to
the feed solution, trace quantities of organic matter may have been present in the DI water used to make
the NaCl feed solution; also, dust may have entered the feed solution via the auxiliary feed tank, which was
partially open to the environment. Flux decline was likely due to gradual accumulation of these trace
foulants on the surface. The hydrophobicity of the membrane and elevated operating temperatures used in
MD could result in greater attraction of organics to the membrane surface than in membrane processes
using hydrophilic membranes or operating at ambient temperatures [54-56].

11

Figure 2.2 Water flux and distillate conductivity versus time for the
long-term experiment with 35 g/L NaCl feed solution and average
feed and distillate temperatures of 63 and 40 °C measured at the
module. Flow rates on the feed and distillate sides were
maintained at 1.5 L/min throughout the experiment. Distillate
conductivity is shown as a moving average.
F
A brief decrease in flux occurred at the end of day 40 due to a malfunction in the chiller, which shut off for
8.5 hours overnight (data omitted). When the chiller restarted, there was a slightly elevated driving force,
resulting in the subsequent brief increase in flux shown in Figure 2.2. The chiller also briefly malfunctioned
on day 68 (data omitted); flux did not significantly increase afterward. Although the distillate temperature
increased by 14 °C during both events, a temperature difference of at least 10 °C was maintained. Both
events were brief and the system rapidly returned to steady state after each. In previous studies, temporary
system shut-downs have been associated with membrane wetting due to vapor condensation within the
membrane pores (e.g., [21, 31, 41, 57]). This was not observed in the current study, in which the system
was not shut down and membrane wetting did not occur.
Feed conductivity averaged 89 ± 3.1 mS/cm and distillate conductivity remained below 8 μS/cm for the
duration of the experiment. Salt rejection based on conductivity was always greater than 99.8%. At day 58,
distillate conductivity, which was relatively constant for the first 50 days of operation, began to slightly
increase. At the same time, there was a slight (< 1 L/m
2
h) increase in water flux. This indicates the possibility
of a small volume of liquid feed solution passing through the membrane, simultaneously increasing the
measured water flux and distillate conductivity. However, as shown by the low distillate conductivity and
12

high rejection maintained throughout the experiment, it is likely that most of the membrane pores remained
dry.
Membrane Fouling and Surface Morphology
Representative FESEM images and EDS spectra for the membrane are shown in Figure 2.3. A
representative image of one side of the virgin membrane surface was selected (Figure 2.3a) as the feed
and distillate sides did not differ significantly. FESEM images of the feed side of the used membrane (Figure
2.3b) confirmed that the fouling layer fully covered portions of the membrane. FESEM images of the
distillate side of the used membrane (Figure 2.3c) showed no external foulants and a similar morphology
as the virgin membrane but with increased fibril and node widths. ImageJ software was used to analyze the
void space in the FESEM images of the virgin and distillate side of the used membrane. Void space
decreased by an average of 70% after long-term operation. This was supported by visual observation of
the images; the virgin membrane has a greater number of fibrils as well as more fibrils with thin diameters
(Figure 2.3a). In contrast, the used membrane has wider fibrils and larger, flattened nodes (Figure 2.3c).
Fibril aggregation was also reported by Saffarini et al. [19] when the observed heat-treated PTFE
membranes had shorter fibrils relative to virgin membranes, perhaps due to aggregation of nodes. Fibril
and node changes were attributed to temperature effects [19]. Morphology changes were also observed by

Figure 2.3 Representative FESEM images of (a) virgin membrane, (b) feed side of used membrane, and
(c) distillate side of used membrane and EDS spectra of (d) virgin membrane, (e) feed side of used
membrane, and (f) distillate side of used membrane for the long-term experimentwith 35 g/L NaCl feed
solution and average feed and distillate temperatures of 63 and 40 °C measured at the module. FESEM
images were taken at ⨯ 4,000 magnification; EDS spectra were taken at ⨯ 250 magnification to characterize
a larger area of the heterogeneous fouling deposits. EDS spectra were taken in triplicate for each sample
and the resulting average atomic percent composition values are shown. Values for the virgin membrane
represent an average of both sides of the symmetric membrane.
13

Barbe et al. [58] for PP membranes soaked in water for 24 to 72 hours; these changes were attributed to
water entry into pores resulting in pore expansion. Interestingly, no changes were observed when using a
CaCl2 solution [58].
In the EDS spectrum for the virgin membrane (Figure 2.3d), the fluorine peak represents 73% while the
carbon peak represents 25% on an atomic basis, reflecting the structure of the PTFE molecule [59]. In the
EDS spectrum for the feed side of the used membrane (Figure 2.3e), the fluorine peak represented only
0.4% on an atomic basis; in some scans, it was below detection limits. This indicates that portions of the
membrane surface were fully covered by the fouling layer. The greater amount of carbon relative to fluorine
and the appearance of oxygen indicated that organic matter was present in the fouling layer. Hydrogen
cannot be identified by EDS; however, the carbon to oxygen ratio (55% C to 20% O by weight) is somewhat
similar to that of natural organic matter (40-60% C to 30-50% O by weight) [60, 61]. Although nitrogen and
sulfur did not appear, these elements can be 0 to less than 1% by weight of natural organic matter [61]. A
small peak for silica also appeared in the feed-side EDS spectra. Silica may have been present in very
small concentrations in the DI water used in the feed solution; it is also possible that silica came from
silicone components (e.g., sealant) used in the system although the reported temperature resistance of all
silicone components was well above operating temperatures. Carbon associated with the silica deposits as
silicone may also account for the higher C:O ratio relative to natural organic matter. EDS results for the
distillate side of the used membrane (Figure 2.3f) showed it was free of external foulants. Higher
magnification EDS scans of the node area between fibrils confirmed that these areas were composed only
of membrane material.
Membrane Hydrophobicity
Contact angles measured on the virgin and used membrane for the long-term experiment are shown in
Table 2.1.
Table 2.1 Contact angles measured on the virgin and used membrane before and after the long-term
experimentwith 35 g/L NaCl feed solution and average feed and distillate temperatures of 63 and 40 °C
measured at the module.
Membrane Side Water Contact Angle
Virgin
Feed 140 ± 4.0°
Distillate 140 ± 3.6°
Used
Feed 61 ± 9.1°
Distillate 104 ± 14°

14

The contact angle of the feed side of the membrane was reduced by 56%, making the surface much more
hydrophilic after use; this was not unexpected due to the fouling layer on the surface. Numerous previous
studies have noted that fouling on the feed side of the membrane promotes wetting because the more
hydrophilic foulant layer provides a path for liquid to enter the membrane pores [34, 37, 41, 55, 62-64].
Given the presence of the fouling layer, no conclusions regarding the change in contact angle of the actual
membrane surface could be made. The contact angle of the fouled membrane as it was during operation
represented the “effective hydrophobicity” of the membrane surface, as experienced by the feed solution.
As foulants began to adhere to the membrane surface, the virgin membrane surface no longer interfaced
with the feed solution; instead, the feed solution was in contact with a more heterogeneous and hydrophilic
surface composed of PTFE and foulants. Interestingly, despite the effective hydrophilization of the feed-
side surface of the membrane, salt rejection remained high and constant. This suggests that another
mechanism, or combination of mechanisms, may have played a significant role in maintaining rejection.
Surprisingly, the contact angle of the distillate side of the membrane was reduced 26% during use. The
lack of fouling on the distillate side, as confirmed by FESEM and EDS analyses, means the reduced contact
angle is entirely due to physical changes in the membrane itself. Although to the best of the authors’
knowledge reductions in distillate-side contact angles have not been reported in the literature, PP
membranes soaked in brine solution have been found to be hydrophilized due to the formation of hydrophilic
groups [63]. Also, heat treatment of PTFE membranes has been found to result in a lower measured liquid
entry pressure (LEP), or pressure at which liquid can penetrate membrane pores [19].
Contact angles can be related to LEP according to
𝐿𝐸𝑃 =
−2𝛽 𝛾 𝑙 cos( 𝜃 )
𝑟 𝑝 ,𝑚𝑎𝑥
( 2.1)
where 𝛽 is a parameter describing pore geometry, 𝛾 𝑙 is surface tension of the wetting liquid, 𝜃 is contact
angle between the wetting liquid and surface, and 𝑟 𝑝 ,𝑚𝑎𝑥
is maximum pore radius [28]. LEP measurements
of QM022 membranes resulted in an LEP of 450 kPa for pure water at 25 °C [28]. The decrease in contact
angle on the feed side from 140 to 61° corresponds to a decrease in LEP from 450 to 0 kPa, indicating that
the feed side of the membrane would readily be wetted at ambient conditions. On the distillate side, the
decrease in contact angle from 140 to 104° corresponds to a decrease in LEP from 450 to 142 kPa.
Although this change is significant, a distillate-side LEP of 142 kPa would still likely be high enough to
prevent significant wetting given that testing in a separate experiment (using stainless-steel pressure
probes) found operating pressures for this system to be consistently less than 70 kPa.
Despite possible surface wetting caused by the fouling layer, liquid did not pass through the membrane.
With a wetted surface layer, the feed solution vapor interface could move inward and may be located at an
inner layer of the membrane or within membrane pores [35]. This suggests that internal hydrophobicity may
be important in maintaining rejection, especially when a fouling layer is present. The membranes used in
15

this study are symmetric and composed of a single layer of PTFE; however, the internal hydrophobicity
may differ from that of the membrane surface. It has been previously noted that the hydrophobicity of the
interior of PVDF membranes may not be equal to the surface hydrophobicity [65]. If internal hydrophobicity
differs from surface hydrophobicity, it may contribute uniquely to rejection when surface hydrophilization
occurs during long-term operation. Also, although improved performance and rejection have been
demonstrated many times for MD membranes with superhydrophobic surface coatings or thin surface
layers (e.g., [66], [67], [68]), their long-term performance is less certain. Gryta [35] also suggested that a
substantial hydrophobic layer may be necessary to maintain rejection based on surface wetting studies.
When the feed side surface is hydrophilized, internal hydrophobicity and distillate-side surface
hydrophobicity together may be necessary to maintain a vapor-filled layer and prevent wetting via bridging
of liquid from the feed and distillate streams. However, a detailed study of the role of internal hydrophobicity
has not yet been performed in the MD literature.
Membrane Surface Roughness
RMS roughness values obtained from Gwyddion image analysis software are shown in Table 2.2 and AFM
scans of the virgin and used membrane are shown in Figure 2.4. A representative image of one side of the
virgin membrane surface was selected as the two sides did not differ significantly. Membrane fouling on the
feed side of the used membrane led to a reduction in roughness of approximately 92% relative to the virgin
membrane (Table 2.2). This can be observed both in the AFM images (comparison of Figure 2.4a and b)
and FESEM images (comparison of Figure 2.3a and b). The fouling deposits effectively smoothed the
membrane surface on the feed side. A similar effect was observed by Guillen-Burrieza et al. [21]; soaking
and intermittently drying PTFE membranes in seawater for 5 to 20 days resulted in fouling and scaling
deposits that filled in spaces between fibrils and effectively smoothed the membrane surface. In the current
study, it is interesting to note that the morphological changes on the distillate side of the used membrane
were also characterized by a significant reduction in roughness – approximately 57% relative to the virgin
membrane (Table 2.2). As increased roughness correlates with increased hydrophobicity [22], decreased
surface roughness is likely to have contributed to the decreased contact angles seen in Table 2.1.

16

Table 2.2 Roughness values for membrane used in the long-term experimentwith 35 g/L NaCl feed solution
and average feed and distillate temperatures of 63 and 40 °C measured at the module.
Membrane Side RMS Roughness (nm)
Virgin
Feed 31.0 ± 5.40
Distillate 27.1 ± 7.58
Used
Feed 3.0 ± 0.88
Distillate 14.5 ± 3.11

Figure 2.4 Representative tapping mode AFM images of (a) virgin membrane, feed side of used membrane,
and (c) distillate side of used membrane for the long-term experimentwith 35 g/L NaCl feed solutionand
average feed and distillate temperatures of 63 and 40 °C measured at the module.

For the distillate side of the used membrane, changes in membrane surface hydrophobicity and morphology
(Figure 2.4c) may have been due to temperature effects or pore wetting, as proposed by previous studies
[19, 58]. Saffarini et al. [19] observed minor shifts in the surface structure of PTFE membranes at
temperatures < 50 °C in air as well as decreased LEP with increasing temperature. The morphological
changes may have also been due to shear forces exerted on the membrane surfaces by the transverse
flow of liquid over the surface. Computational fluid dynamics models have found shear stress values of
approximately 1 to 7 Pa, depending on spacer configuration, at similar cross-flow velocities [69].
Morphological changes may also have been due to hydraulic pressure effects on both sides of the
membrane that caused compression over time even though only low pressures (less than 70 kPa) are
required for recirculation. These forces are likely orders of magnitude lower than the Young’s modulus of
the membrane, as similar polymer membranes have shown yield stress values of 7.5 MPa [70]. Although
shear forces and hydraulic pressure were likely low in this study, long-term operation may result in
permanent deformation over time via creep, especially at elevated temperatures [71]. Polymeric
membranes are known to have relatively low creep resistance [72]. Within the MD literature, creep and
membrane compaction have been only briefly mentioned as potential risks of degassing feed solution, as
causes of minor flux decline, and as benefits of mechanically strong membrane materials [73-75]. However,
17

creep, especially due to the combined effect of shear forces, hydraulic pressure, and temperature, has not
been previously reported in the literature to the best of the authors’ knowledge.
2.3.2. Effect of Short-term Operation with Scaling and Confirmation of Long-term
results
To corroborate the results of the long-term study and to further investigate membrane resistance to wetting
under scaling conditions, a short-term (20-day) experiment with 200 g/L NaCl feed solution was performed,
with membrane scaling induced after 15 days. Operating conditions were otherwise identical to the previous
long-term experiment. Water flux and distillate conductivity versus elapsed time for the short-term
experiment are shown in Figure 2.5.
Flux decreased fairly linearly with time over the first 15 days of continuous operation, with a correlation
coefficient of 0.7; overall, flux decreased by less than 5%. This was similar to the long-term experiment with
35 g/L NaCl, in which flux declined by 3% over the first 15 days. After day 15, the solenoid valve loop
maintaining constant feed concentration was shut off and the feed solution was allowed to concentrate in
batch mode. Operating conditions, including temperatures, were otherwise held constant. Due to the
elevated concentration, a visible layer of crystallized salt (scale) developed on the membrane surface and
caused rapid and severe flux decline, reducing and eventually preventing water flux through the membrane.
After replenishing the feed tank with DI water, flux was partially recovered to approximately 34% of the
initial flux and remained constant for the final five days of operation. This demonstrated that some portion
of the fouling was reversible, although the irreversible portion was associated with the majority of the flux
decline.

18

Figure 2.5 Water flux and distillate conductivity versus time for the
short-term experiment with 200 g/L NaCl feed concentration and
average feed and distillate temperatures of 63 and 40 °C
measured at the module. Distillate conductivity is shown as a
moving average.
The virgin and used membrane surfaces were characterized by FESEM and EDS, as shown in Figure 2.6.
A representative image of one side of the virgin membrane surface was selected as the two sides did not
differ significantly.
The feed side of the used membrane was nearly completely covered by fouling deposits (Figure 2.6b),
which were composed of carbon, sodium, and chlorine (Figure 2.6e). The EDS spectrum also showed
oxygen and fluorine, although there was significantly more carbon relative to fluorine, indicating that organic
fouling occurred (as in the long-term experiment). A very small peak for silica was present, likely originating
from the same source as the long-term experiment. The appearance of nickel on the feed side was
surprising, especially given the lack of nickel in the long-term experiment; it is possible that this originated
from impurities in the ACS-reagent-grade NaCl used for feed solutions, which contained ~2 mg/L heavy
metals. The development of the scaling layer at the end of 15 days resulted in a slight increase in distillate
conductivity (Figure 2.5), indicating that some pore wetting occurred. The scaling layer likely promoted pore
wetting by decreasing the effective hydrophobicity of the surface [21, 34, 37, 41, 55, 57, 63, 64]. The
increased salinity of the feed solution may also have promoted membrane wetting. NaCl solutions result in
a lower contact angle relative to pure water, but NaCl solutions also have a slightly higher surface tension
than pure water; the overall effect on LEP has been found to be fairly small [19] but may be significant at
19

higher salinities. In this study, despite severe scaling of the membrane and elevated feed solution
concentration, the distillate conductivity remained quite low as in the long-term experiment.

Figure 2.6 Representative FESEM images of (a) virgin membrane, (b) feed side of used membrane, and
(c) distillate side of used membrane and EDS spectra of (d) virgin membrane, (e) feed side of used
membrane, and (f) distillate side of used membrane for the short-term experiment with 200 g/L NaCl feed
concentration and average feed and distillate temperatures of 63 and 40 °C measured at the module.
Distillate conductivity is shown as a moving average. All images and EDS spectra were taken at x 4,000
magnification and the resulting atomic percent composition values are shown. Value for the virgin
membrane represent an average of both sides.
The distillate side of the used membrane also had a small amount of sodium and chlorine present on the
surface, indicating that salt penetrated to the distillate side of the membrane due to wetting of some
membrane pores; crystallization likely occurred during desiccation. The lack of oxygen and the similar
carbon to fluorine ratio as the virgin membrane indicated that essentially no organic foulant material was
present on the distillate side of the membrane. As in the long-term experiment, there was a distinct change
in morphology of the distillate side of the used membrane (comparing Figure 2.6a and c). The number of
thin membrane fibrils decreased and there was an increase in the apparent node or aggregated fibril area,
although ImageJ analysis showed no significant decrease in void space between the virgin and used
membrane.
Contact angles measured on the virgin and used membrane in the short-term experiment are presented in
Table 2.3.

20

Table 2.3 Contact angles measured on membrane used in the short-term experimentwith 200 g/L NaCl
feed concentration and average feed and distillate temperatures of 63 and 40 °C measured at the module.
Membrane Side Water Contact Angle
Virgin
Feed 140 ± 3.96 °
Distillate 140 ± 3.63 °
Used
Feed side 49.1 ± 3.18 °
Distillate side 114 ± 6.35 °

As before, the contact angle decreased on both sides of the membrane. On the feed side of the used
membrane the contact angle was reduced by 65%, making the surface much more hydrophilic; this was
expected due to the visible layer of salt crystals on the surface. The contact angle on the distillate side of
the used membrane decreased by 19% relative to the virgin membrane. The significant decrease after only
20 days of use was unexpected; however, this change was less than that observed in the long-term
experiment, suggesting that the distillate-side contact angle decreases with increasing duration of use.
Morphology changes were also more significant after 100 days of operation than after 20 days. The results
of the short-term experiment corroborated the findings of the long-term experiment and supported the
likelihood that duration of use affects membrane characteristics even when membranes are exposed only
to salt solutions and trace foulants. The observations made from comparing 100-day to 20-day operation
further supported the likelihood that feed-side surface hydrophobicity is not an isolated parameter and other
factors, specifically distillate-side surface hydrophobicity and internal hydrophobicity, may play key roles in
maintaining salt rejection.

2.4 Conclusions and Implications
In this study, significant changes to MD membrane surface morphology, decreased membrane
hydrophobicity, and decreased surface roughness were observed after long-term use. Distillate conductivity
remained low even after significant decrease in the feed-side contact angle; this suggests that maintaining
distillate-side hydrophobicity and/or internal hydrophobicity may be more important for long-term
performance than has been previously suggested. Furthermore, fouling was observed even when “clean”
feed solutions were used. When fouling is present, distillate-side and internal hydrophobicity may be equally
or more important than feed-side hydrophobicity as the distillate surface, internal layers, and pore walls are
less affected by foulants that can effectively hydrophilize the surface. Effects of long-term operation may
be especially significant for membranes with surface coatings or thin active layers, which may be more
vulnerable than internal material and distillate-side membrane surfaces.
Furthermore, there is significant interest in recovering membrane hydrophobicity after fouling has occurred.
Typically, membrane hydrophobicity is assumed to be recoverable if foulants can be removed; however, if
21

membrane hydrophobicity is decreased due to physical or chemical changes in the membrane itself,
hydrophobicity may not be recoverable. Permanent reductions in hydrophobicity can result in permanent
reductions in membrane performance and lifetime.

Acknowledgements
The authors would like to acknowledge funding support for this study from the Strategic Environmental
Research and Development Program (SERDP Project Number ER-2237) and from the US Environmental
Protection Agency Science to Achieve Results (US EPA STAR Grant #83486701). Additionally, the authors
would like to acknowledge support to this project from two fellowships awarded to R.D. Gustafson: a Viterbi
Graduate School Ph.D. Fellowship and a National Water Research Institute and Southern California Salinity
Coalition fellowship. The FESEM-EDS and AFM images and spectra were acquired at the University of
Southern California Center for Electron Microscopy and Microanalysis and Center for Excellence in
NanoBioPhysics, respectively.

22

CHAPTER 3.

3. Characterization of Fibril Roughness and Contact Angle to Predict Pore
Wetting Resistance in Membrane Distillation

Abstract
In characterizing and testing membrane distillation (MD) membranes, measured or intrinsic contact angles
are generally used to indicate or calculate wetting resistance. For conventional, hydrophobic membranes,
contact angles are typically measured in a Cassie-Baxter state; however, it has been shown that
conventional MD membranes treating feed solutions at typical operating pressures are usually in the
Wenzel state – especially for low-surface tension feed solutions and/or when the membrane is fouled or
scaled. In these cases, the characteristics of the pore entrance are expected to be more important to wetting
resistance than the characteristics of the membrane surface. The objective of this study was to develop a
method to characterize the wetting resistance at the pore entrance of microporous polytetrafluoroethylene
membranes. As these membranes consist of a network of polymer fibrils, zooming in to the scale of interest
(i.e., to individual pore entrances) would mean characterizing the contact angle on an individual,
microscopic fibril, which has significant experimental challenges. Instead, fibril contact angle (𝜃 𝑓 ) was
calculated based on the Wenzel model and another newly introduced parameter: fibril roughness. Fibril
roughness was measured via atomic force microscopy scans of membranes at relevant length scales.
Results show that fibril roughness can be significantly lower than membrane roughness. Due to differences
in roughness and wetting state, 𝜃 𝑓 values are significantly lower than measured contact angles, and
because fibril roughness is non-negligible, 𝜃 𝑓 values are significantly greater than intrinsic contact angles.
Comparison of model and experimental results for liquid entry pressure show that 𝜃 𝑓 may provide better
prediction of in situ pore wetting resistance for MD membranes. Fibril roughness may be as or more
important than membrane roughness to impart pore wetting resistance, especially for challenging
applications. Furthermore, 𝜃 𝑓 can serve as an indicator of pore wetting resistance in situ for any hydrophobic
membrane based on intrinsic contact angles and fibril roughness.

3.1 Introduction and Background
High-salinity waste streams produced during desalination of saline, brackish, and waste waters or from
industrial processes present unique treatment challenges; currently, their management is associated with
high capital and operating costs and high energy consumption [1-7]. Membrane distillation (MD) is a
thermally driven membrane process that is promising for the treatment of challenging, high-salinity streams.
While reverse osmosis is salinity-limited, MD relies on a driving force that is only slightly decreased by
salinity [1, 13]. While conventional thermal desalination processes are associated with low energy efficiency
23

and high cost, MD can effectively utilize low-grade thermal energy, achieve higher energy efficiency at
smaller capacities, and require lower capital costs [1]. In MD, a warmer, saline stream flows on one side
and a cooler, pure distillate stream flows on the other side of a microporous, hydrophobic membrane. The
partial vapor pressure difference across the membrane drives water in the feed stream to evaporate and
diffuse through the unwetted membrane pores, condensing into the cooler distillate stream [13, 76]. When
the membrane pores are unwetted, a vapor gap is maintained between the feed and distillate streams and
passage of unwanted solutes into the distillate stream is prevented. When the membrane pores are wetted,
the feed and distillate streams freely mix through the membrane. Thus, pore wetting must be prevented to
support phase separation and achieve solute rejection [13]. Full and partial wetting of membrane pores also
impact water flux. Partial wetting (i.e., when liquid partially penetrates a pore, resulting in a thinner vapor
gap between the feed and distillate streams [13, 77, 78]) may increase temperature polarization, reducing
the effective driving force for water flux, but also has been suggested to reduce the diffusion path length for
water vapor (i.e., reduce the effective thickness of the membrane), which increases water flux [14, 35, 36,
78, 79]. Partial wetting can also increase vapor flux because the liquid-vapor interfacial area within the
interconnected pores is greater than that at isolated pore entrances of the membrane surface [35, 78, 79].
This reflects the trade-off between wetting resistance and vapor flux that was described by Wang et al. [79]
and Li et al. [78]. A wetting resistance/permeability trade-off was also observed in our previous paper (i.e.,
McGaughey et al. [77]) where reduced pore size resulted in higher wetting resistance and lower
permeability. Reduced pore size may also inhibit internal scaling, increasing wetting resistance specifically
for high-salinity streams [77].
Membrane wetting resistance is key to performance – and wetting resistance depends on membrane
properties; specifically, pore size and hydrophobicity. Pore sizes between 0.2 and 0.45 μm are often used
for MD membranes [28]. To achieve hydrophobicity, low-surface-energy polymers are typically used;;
common polymers used include poly(vinylidene fluoride), polypropylene, polyethylene,
polydimethylsiloxane, and polytetrafluoroethylene (PTFE) [80-82]. PTFE membranes are typically the most
hydrophobic and PTFE is the most commonly used material for MD membranes [18, 83, 84]. PTFE
membrane morphology is typically characterized by polymer strands, called fibrils, that intersect at junctions
or “nodes”; pores are formed between fibrils [19, 85].
3.1.1 Characterization of Wetting Resistance
Liquid Entry Pressure
Wetting resistance of MD membranes is often characterized by measured or modeled liquid entry pressure
(LEP) values. LEP is defined as the minimum transmembrane pressure at which liquid enters a pore. For
hydrophobic membranes, LEP is typically reported as the minimum LEP, or the lowest transmembrane
pressure at which wetting is observed; this corresponds to wetting of the largest pore. LEP is most
frequently modeled by the Young-Laplace equation [86]:
24

𝐿𝐸𝑃 =
−4𝐵 𝛾 𝑙𝑣
cos 𝜃 𝑑 𝑝 (3.1)
where 𝐵 is a pore geometry factor, 𝛾 𝑙𝑣
is liquid-vapor interfacial tension, 𝜃 is contact angle, and 𝑑 𝑝 is pore
size. The largest pore size (𝑑 𝑝 ,𝑚𝑎𝑥
) is used to calculate the minimum LEP. The Young-Laplace model
assumes cylindrical pores with constant radius of curvature. Microporous PTFE membranes are generally
not characterized by cylindrical pores with constant radii. Kim and Harriott [85] developed an equation to
describe wetting through a microporous PTFE membrane. Kim and Harriot modeled the PTFE membrane
as a grid of intersecting fibrils of diameter 𝑑 𝑓 that form pore openings of diameter 𝑑 𝑝 [85] (as illustrated in
Figure 3.1b and c). LEP in the Kim and Harriott equation is given by [85]:
𝐿𝐸𝑃 =
−4𝛾 𝑙𝑣
𝑑 𝑝 cos( 𝜃 − 𝛼 )
1 +
𝑑 𝑓 𝑑 𝑝 ( 1 − cos 𝛼 )
(3.2)
where 𝛼 is the angle of the liquid meniscus. Surface pore diameters are assumed to control wetting
resistance because the membrane bulk is characterized by a highly interconnected pore space. This model
was found to agree with experimental results for a hollow-fiber PTFE membrane [85] as well as flat sheet
PTFE, polyvinylidene fluoride, and nylon membranes [87].
In both the Young-Laplace and Kim-Harriott models, and in general indicate wetting resistance in MD, 𝜃
has been defined sometimes as the intrinsic contact angle (e.g., [13, 79, 87-89]) and sometimes as the
measured (macroscopic) contact angle (e.g., [1, 18, 35, 83, 85, 87, 90-92]). Yet, discrepancies between
expected wetting resistance and in situ wetting resistance have been observed (e.g., [87, 93-95]).
Hydrophobicity
The intrinsic contact angle (𝜃 0
) is the contact angle formed by droplets on the surface of a smooth,
homogeneous surface under ideal conditions (Young’s state) (Figure 3.1a). 𝜃 0
is given by Young’s equation
[96]:
𝛾 𝑙𝑣
cos 𝜃 0
= 𝛾 𝑠𝑣
− 𝛾 𝑠𝑙
(3.3)
where 𝛾 is interfacial energy and subscripts 𝑠 , 𝑙 , and 𝑣 refer to the solid, liquid, and vapor phases,
respectively. Implicit in using the intrinsic contact angle to characterize wetting resistance and LEP, is the
assumption fibril walls are perfectly smooth at pore entrances. For membranes with smooth pore entrances
and smooth fibrils, the intrinsic contact angle may provide good approximation of the actual contact angle
at pore entrances. However, for membranes with rough fibrils, the intrinsic contact angle is likely to provide
a less accurate prediction. In one study, Taylor et al. [95] observed that for commercial membranes,
surfactant-induced wetting depended more on membrane material than measured contact angle or pore
size [95]; this may indicate that either the intrinsic contact angle or another hydrophobicity parameter is
more important to wetting resistance than the measured contact angle, though this was not discussed.
25

On rough surfaces, water droplets can be in a Cassie-Baxter state, sitting atop solid surface features and
trapped air bubbles (Figure 3.1b) or in a Wenzel state, with liquid filling surface grooves (Figure 3.1c). In
the Cassie-Baxter state, the measured contact angle (𝜃 𝑚 ) is given by [96]:
cos 𝜃 𝑚 = 𝜙 𝑠 ( cos 𝜃 0
+ 1)− 1 ( 3.4)
where 𝜙 𝑠 is the fraction of liquid in contact with the solid (versus the fraction in contact with trapped air)
[96]. The Cassie-Baxter model assumes that 𝜙 𝑠 is constant and does not depend on where the droplet is
located on the surface [97]. In the Wenzel state, the measured contact angle is given by [96]:
cos 𝜃 𝑚 = 𝑅 ′cos 𝜃 0
(3.5)
where 𝑅 ′ is the roughness coefficient, or the ratio between the actual and projected surface area [96]. As
for the Cassie-Baxter model, the Wenzel model assumes that 𝑅 ′ is constant and independent of droplet
location and that the droplet size is large relative to the roughness scale [97]. For real surfaces that are
hydrophobic, the measured contact angle increases with increasing surface roughness. In the Cassie-
Baxter state, increasing roughness results in more trapped air beneath the droplet and lower 𝜙 𝑠 ; in the
Wenzel state, increasing roughness increases the interfacial solid surface area and 𝑅 ′, geometrically
increasing the measured contact angle [96]. The Cassie-Baxter state is more stable for surfaces that have
higher contact angle and roughness values, while the Wenzel state is more stable for surfaces that have
moderate-to-low contact angles, when air bubbles are not trapped at the surface [96, 98].

Figure 3.1 Illustration of a) Young’s state and intrinsic contact angle 𝜃 0
on a smooth, hydrophobic material,
b) Cassie-Baxter state and measured contact angle 𝜃 𝑚 on an idealized membrane surface, and c) Wenzel
state and 𝜃 𝑚 on an idealized membrane surface. Idealized membrane surfaces are illustrated as a cross-
section of intersecting fibrils that form pore openings Only a few rows of intersecting fibrils are illustrated
for simplicity. It should also be noted that in real membranes, fibril diameter, spacing, and orientation
typically vary more widely than shown here.
26

Implicit in using measured contact angles to indicate wetting resistance or calculate LEP are two key
assumptions: (1) that macroscopic droplets used for contact angle measurement are representative of pore
wetting resistance that occurs on a microscopic scale, and (2) that the wetting state of water droplets used
for contact angle measurement is the same as the wetting state of feed solution at the membrane surface
during MD (i.e., the in situ wetting state)
In sessile droplet contact angle measurement, typical droplet volumes (i.e., 5 μL) result in droplet diameters
on the order of 1 – 2 mm; these droplets span hundreds to thousands of fibrils and pore openings. Measured
contact angles depend on the membrane surface roughness formed by many fibrils and pore openings.
However, contact angles at pore entrances depend on microscopic fibril roughness. Roughness parameters
are inherently dependent on length scale and can vary significantly with length scale for dense and porous
membranes [99-102]. Roughness parameters that affect contact angle (𝜙 𝑠 , in the Cassie-Baxter state, and
𝑅 ′, in the Wenzel state) may also depend on length scale, resulting in significant differences between the
measured contact angle and the effective contact angle in situ. In one study that used the measured contact
angle to characterize wetting resistance in MD, it was explicitly assumed that surface roughness is
representative of pore wall roughness [91]; however, the assumption was not tested, nor was the wetting
state mentioned.
For membranes with relatively high contact angles (such as PTFE membranes), droplets used for contact
angle measurement are likely in a Cassie-Baxter state. While it has been suggested that omniphobic and
superhydrophobic membranes operate in a Cassie-Baxter state [14, 78, 79], it has been shown that
hydrophobic polyvinylidene fluoride and quartz fiber membranes operate in a Wenzel state under typical
MD conditions [78, 79]. Furthermore, because applying relatively low pressures (~250 Pa) can drive an
irreversible transition into the Wenzel state on rough surfaces [96], typical hydraulic pressures used in MD
systems (on the order of 10-100 kPa [84]) likely force the feed solution into a Wenzel state for many MD
membranes – both on the macroscopic scale (partially wetting pores) and the microscopic scale (wetting
fibril surfaces). Therefore, measured contact angles may not be suitable to characterize wetting resistance
in situ. Yet, to our knowledge, differences between the wetting state for measured contact angles and the
in-situ wetting state – either at the macroscopic or microscopic scales – have not been previously discussed
for MD.
Furthermore, to our knowledge no experimental investigations of the measured contact angle at pore
entrances or on pore walls have been reported for hydrophobic membranes. Experimental investigation of
wetting and contact angles at the microscale is challenging. Environmental scanning electron microscopy
(SEM) has been used to investigate microscale wetting for droplet radii as low as 10 μm [103]; however,
controlling droplet orientation for accurate measurement is challenging [104]. While AFM can be used to
image smaller water droplets on surfaces, deformation of soft surfaces, evaporation, contamination, and
preferential deposition or condensation of water droplets affect results [103].
27

3.1.2 Hypothesis and Objectives
We hypothesize that, for hydrophobic MD membranes, wetting resistance in situ depends on a microscopic
fibril contact angle (𝜃 𝑓 ) formed at the air, feed, and fibril surface interface at pore entrances (Figure 3.2).
Non-negligible fibril roughness is expected to result in 𝜃 𝑓 values that are unique from intrinsic contact
angles. Differences in roughness with length scale (i.e., differences between the fibril roughness and
membrane roughness) and differences between the in situ wetting state and the wetting state of measured
contact angles are expected to result in 𝜃 𝑓 values that are also unique from measured contact angles. 𝜃 𝑓
may provide better understanding and more accurate prediction of wetting resistance in situ, which is critical
for maintaining separation integrity – especially for challenging feed streams.
The objective of this study is to develop a method to calculate a representative value of the contact angle
at pore entrances (i.e., 𝜃 𝑓 ) for microporous PTFE membranes, based on the Wenzel model, the intrinsic
contact angle of PTFE, and measured fibril roughness. We also aim to investigate the role of 𝜃 𝑓 in wetting
resistance by comparing LEP model results based on the intrinsic, fibril, and measured contact angles to
experimentally measured LEP values for the PTFE membranes.

Figure 3.2 Fibril contact angle 𝜃 𝑓 formed
at the three-phase interface at a pore
entrance (illustrated as a cross-section of
membrane fibrils) in the Wenzel state,
with non-negligible fibril roughness.

3.2 Materials and Methods
3.2.1 Membranes
Three flat-sheet, microporous membranes from Parker Performance Materials (Lee’s Summit, MO, USA)
were used. Two of the membranes were symmetric, single-layer PTFE membranes: QM050, with a nominal
pore size of 0.05 μm and QM022, with a nominal pore size of 0.2 μm. One of the membranes was a
supported membrane with a PTFE active layer on a polypropylene support layer: QL822, with a nominal
pore size of 0.45 μm.
28

3.2.2 Material Characterization
A micrometer with an accuracy of ± 1 μm (MDC-1 PX, Mitutoyo, Kawasaki, Japan) was used to characterize
membrane thickness. For the single-layer membranes, bulk porosity was characterized via the gravimetric
method [23]. For supported membranes, bulk porosity was not measured because the support layer could
not be fully removed without damaging the selective layer, preventing accurate use of the gravimetric
method.
2-D surface morphology was characterized using a field-emission scanning electron microscope (Nova
NanoSEM 450, FEI, Hillsboro, OR, USA). ImageJ software (version 1.52, National Institutes of Health,
Bethesda, MD, USA) was used to analyze scanning electron microscopy (SEM) images to characterize
surface pore size distribution and surface porosity; reported results represent an average of results for three
arbitrary areas of each membranes’ surface. Figure 3.3a-c show SEM images of the intersecting polymer
fibrils that form the membrane surfaces; Figure 3.3d-f show the fibril surface morphology.

Figure 3.3 SEM images at x5,000 magnification (10-μm scale bar) showing membrane surface morphology
of the feed side of the unused a) QM050, b) QM022, and c) QL822 membranes and at x50,000
magnification (1-μm scale bar) showing fibril surface morphology of the d) QM050, e) QM022, and f) QL822
membranes.
3-D surface morphology was characterized using an atomic force microscope (Dimension Icon, Bruker,
Billerica, MA, USA) at three relevant length scales: the 5-μm scale (spanning multiple fibrils and pores,
termed the membrane scale), the maximum pore scale for each membrane, and the nominal pore scale for
each membrane. The maximum and nominal pore scales were selected to represent fibril roughness at the
entrance of the largest pore (i.e., the pore of lowest wetting resistance) and at entrances of typical pores
for each membrane. Fibril diameters, 𝑑 𝑓 /𝑑 𝑝 ratios, roughness parameters, and 𝑅 ′ values were calculated
using Gwyddion software (version 2.55, Czech Metrology Institute, Brno, Czech Republic). Reported values
represent an average of results from at least three scans of arbitrary areas on the membrane surface. For
all parameters, statistical significance was evaluated by performing two-sided t-tests assuming unequal
variances. Table 3.1 summarizes material characterization results for the three membranes.

29

Table 3.1 Material characterization results for the three membranes.
Membrane QM050 QM022 QL822
Bulk porosity (%) 84 ± 0.5 87 ± 0.2 70-85**
Thickness (μm) 61 ± 7 68 ± 7 180 ± 6
Maximum 𝑑 𝑝 (μm) 1.04 ± 0.1 1.24 ± 0.1 2.43 ± 0.04
Average 𝑑 𝑓 (μm) 0.35 ± 0.3 1.3 ± 1.1 0.32 ± 0.2
* ± values represent standard deviations
** indicates values are manufacturer-reported
3.2.3 Hydrophobicity Characterization
A value of 105° was used for the intrinsic contact angle of PTFE; this was based on reported intrinsic values,
which vary between 100 and 110° [98, 105-107]. Measured contact angles were characterized according
to the sessile drop method (with 5-μL drop volume) using a goniometer (Model 260, ramé-hart, Succasunna,
NJ, USA). Reported values represent an average of at least ten measurements. Table 3.2 shows measured
contact angles for each membrane; all membranes have similar measured contact angles.
Table 3.2 Measured contact angles (𝜃 𝑚 ) for all membranes.
Membrane QM050 QM022 QL822
𝜃 𝑚 ( °) 144 ± 5 142 ± 7 144 ± 4
* ± values represent standard deviations
Microscopic contact angle values were calculated using the Wenzel model (equation 3.5) at the membrane
(5-μm) scale (𝜃 𝑊 ,𝑚 ) and at the maximum and nominal pore scale (𝜃 𝑓 ) for each membrane, based on the
intrinsic contact angle of PTFE and measured 𝑅 ′ values. For 𝜃 𝑓 , it is assumed that the fibril surfaces are
wetted and that air bubbles are not trapped in surface grooves (i.e., that feed solution is in a Wenzel state).
It should also be noted that the Wenzel model assumes that line tension (i.e., energy per unit length
associated with the contact line at the three-phase interface) is negligible. For macroscopic droplets, line
tension is typically negligible and, specifically for PTFE surfaces, Włoch et al. [108] showed that line tension
is also negligible to the nanoscale, and the Wenzel and Cassie-Baxter models describe nanoscale wetting
behavior well. For all parameters, statistical significance was evaluated by performing two-sided t-tests
assuming unequal variances.
3.2.4 MD Performance Characterization
Membrane performance was characterized using a bench-scale MD system described in a previous study
(i.e., McGaughey et al. [77]). Membrane coupons were installed in a custom-built membrane module with
an active area of 20 cm
2
with mesh spacers (Sterlitech, Kent, WA, USA) placed in both the feed and distillate
channels. For all membranes, the side of the membrane facing the feed stream is referred to as the feed
30

side and the side facing the distillate is referred to as the distillate side. For supported membranes, the
selective layer was installed facing the feed stream and the support layer was installed facing the distillate
stream. Data were recorded using a customized data acquisition and control program (LabView, National
Instruments, Austin, TX, USA).
MD performance was evaluated at a constant feed-side temperature of 53 °C and a constant distillate-side
temperature of 18 °C. 1 M NaCl was used as the feed solution and deionized water was used as the distillate
solution. Transmembrane pressure was held constant at 0 kPa. Water flux was calculated according to
𝐽 𝑊 =
𝑚 𝑑 𝑎 𝑚 Δ𝑡 ⁄ ( 3.6)
where 𝐽 𝑊 is water flux (kg/m
2
s) and 𝑚 𝑑 (kg) is change in mass of the distillate overflow per membrane area
𝑎 𝑚 (m
2
) during time Δ𝑡 (s). Salt flux was calculated according to
𝐽 𝑆 =
𝐶 𝑑 ,𝑡 2
𝑉 𝑑 ,𝑡 2
− 𝐶 𝑑 ,𝑡 1
𝑉 𝑑 ,𝑡 1
𝑎 𝑚 ( 𝑡 2
− 𝑡 1
)
( 3.7)
where 𝐽 𝑆 is salt flux (mM/h) 𝐶 𝑑 is distillate solute concentration (mM), 𝑉 𝑑 is distillate volume (L), and
subscripts 𝑡 𝑖 refer to time I (h). Salt concentrations were calculated from measured conductivities using
standard curves prepared with NaCl solutions of known concentrations. Seven concentrations with
conductivities between the minimum and maximum of the distillate conductivity probe (0 and 200 μS/cm)
and seven additional concentrations with conductivities between the minimum and maximum of the feed
conductivity probe (0 and 200 μS/cm) were selected. For both probes, three samples of each concentration
were prepared and used to generate three standard curves, which were averaged. All standard curve R
2

values were at or above a minimum of 0.997. Baseline MD performance data for the three membranes are
shown in Figure 3.4. Averaging over all membranes between 3 – 24 h operation, water flux was 23 ± 2
L/m
2
h (Figure 3.4a). For all three membranes, salt flux was near-zero and distillate concentration remained
low (Figure 3.4b and c), showing that wetting did not occur.

31

Figure 3.4 Baseline MD performance of MD membranes in terms of a) water flux (𝐽 𝑊 ) and b) distillate
concentration (𝐶 𝑑 ) and salt flux (𝐽 𝑆 ) versus time. Experiments were performed for 24 h with a feed solution
of 1 M NaCl at 52 ± 1 °C and a distillate solution of deionized water at 17 ± 1 °C. The average
transmembrane pressure difference was 0 ± 3 kPa.
3.2.5 LEP Characterization
The Kim and Harriott model [85] (equation 3.2) was used to calculate LEP values, using the surface tension
of pure water at 25 °C (0.072 N/m) for 𝛾 𝑙𝑣
. The ratio of fibril diameter to pore diameter (𝑑 𝑓 /𝑑 𝑝 ) was
determined from AFM results and Gwyddion software according to the method validated by Guillen-Burrieza
et al. [87] Minimum LEP values were calculated using maximum pore diameters.
Calculated LEP values were verified by experimental LEP values, which were measured in the bench-scale
MD system at constant operating temperatures of 25 °C and constant flow rates of 0.76 L/min on both the
feed and distillate sides. 0.1 M NaCl was used as the feed solution in order to ensure measurable solute
passage and maintain a surface tension (i.e., 0.072 N/m) equivalent to that of pure water. When comparing
measured LEP values to modeled LEP values, it was also assumed that contact angles for 0.1 M NaCl
were not significantly different than for pure water. Deionized water was used as the distillate solution. Feed
stream pressure was increased stepwise using a needle valve (McMaster-Carr, Elmhurst, IL, USA). Δ𝑃 was
calculated as the feed stream pressure (i.e., the average of measured pressures at the inlet and outlet of
the module on the feed-side) minus the distillate stream pressure (i.e., the average of measured pressures
at the inlet and outlet of the module on the distillate side). LEP was defined as the highest Δ𝑃 reached prior
to or at the time that membrane wetting first occurred.

32

3.3 Results and Discussion
3.3.1 Membrane Characterization
Figure 3.5 shows 3-D images and average 2-D roughness (Sa) measured on the (unused) feed side of each
membrane at the membrane (5-μm) scale and at the maximum and nominal pore scales for the three
membranes. Comparing Figure 3.5a-c to Figure 3.5d-f, Sa at the maximum pore scale is not significantly
lower than Sa at the membrane scale for any membrane. However, Sa at the nominal pore scale is 181,
117, and 102%, lower than Sa at the membrane scale for the QM050, QM022, and QL822 respectively;
differences are statistically significant for all membranes at α = 0.05. These results indicate that the surface
roughness of an area spanning multiple fibrils and pores, as relevant to the measured contact angle, is
significantly greater than the surface roughness of a single fibril, as is relevant to the meniscus of feed
solution formed at a pore entrance during wetting in MD.

33

Figure 3.5 3-D AFM images and average surface roughness (Sa) of the feed side at 5 x 5 μm area for the
a) QM050, b) QM022, c) QL822 membranes, at the maximum pore scale for the d) QM050, e) QM022, and
f) QL822 membranes, and at the nominal pore scale for the g) QM050, h) QM022, and i) QL822
membranes.
Table 3.3 shows values of the Wenzel model roughness parameter 𝑅 ′ for the three membranes at the
membrane (5-μm) scale and at the maximum and nominal pore scale for the three membranes. If droplets
used for measured contact angles were assumed to be in the Wenzel state, all three membranes would
have relatively high 𝑅 ′ value of approximately 3. Since measured 𝑅 ′ values at the membrane scale are all
significantly less than 3 at α = 0.05, these results support the assertion that the sessile droplets used for
macroscopic contact angle measurement are in the Cassie-Baxter state for all three membranes.
As for Sa, 𝑅 ′ at the maximum pore scale are not significantly different than 𝑅 ′ at the membrane scale for
any membrane. Despite statistically significant differences between Sa at the membrane scale and nominal
pore scale for all membranes, 𝑅 ′ values at the membrane scale are not significantly different than 𝑅 ′ at the
nominal pore scale for the QM050 and QL822 membranes. However, for the QM022 membrane, 𝑅 ′ at the
nominal pore scale is 24% smaller than 𝑅 ′ at the membrane scale; the difference is statistically significant
34

at α = 0.05. In other words, for the QM022 membrane, the 𝑅 ′ value relevant to a meniscus of liquid
penetrating a typical pore is significantly smaller than that relevant to a droplet spanning multiple fibrils and
pore openings, as is used for sessile drop contact angle measurements. Also, for all membranes, the 𝑅 ′
value relevant to pore wetting for both a typical pore and the largest pore is greater than that of a perfectly
smooth surface, as corresponds to the intrinsic contact angle (i.e., 𝑅 ′
≡ 1); differences are statistically
significant at α = 0.05.
Table 3.3 Measured 𝑅 ′ values at the membrane (5-μm) scale and at the maximum and nominal pore scales
for the feed side of the unused membranes.
𝑹 ′
Scale QM050 QM022 QL822
membrane 1.7 ± 0.04 2.2 ± 0.3 1.9 ± 0.2
maximum pore 1.5 ± 0.1 2.0 ± 0.3 1.8 ± 0.1
nominal pore 1.3 ± 0.1 1.7 ± 0.2 2.1 ± 0.3
* ± values represent standard deviations
Table 3.4 shows calculated contact angles determined at the membrane (5-μm) scale (𝜃 𝑊 ,𝑚 ) and at the
maximum and nominal pore scales (𝜃 𝑓 ) for the three membranes, based on 𝑅 ′
results and the Wenzel
model (equation 3.5). At all scales, effective droplet sizes are large relative to the roughness scale,
satisfying the assumption made by the Wenzel model. For the QM050, QM022, and QL822 membranes,
calculated 𝜃 𝑊 ,𝑚 values are 21, 13, and 18% smaller than the measured contact angles, respectively;
differences are statistically significant at α = 0.05. These differences also support the assertion that sessile
droplets used for contact angle measurement are in the Cassie-Baxter state. As discussed in Section 3.1.1,
the in situ wetting state is likely the Wenzel state: it has been shown that hydrophobic polyvinylidene fluoride
and quartz fiber membranes operate in a Wenzel state under typical MD conditions [78, 79] and typical MD
operating pressures (10-100 kPa [84]) likely result in a Wenzel state for feed solution at pore entrances for
many MD membranes.

35

Table 3.4 Wenzel contact angles at the membrane (5-μm) scale (𝜃 𝑊 ,𝑚 ) and fibril contact angles (𝜃 𝑓 ) at the
maximum and nominal pore scale for the feed side of the unused membranes.
QM050 QM022 QL822
Scale 𝜽 𝑾 ,𝒎 ( °) 𝜽 𝑾 ,𝒎 ( °) 𝜽 𝑾 ,𝒎 ( °)
membrane 117 ± 0.01 125 ± 0.05 120 ± 0.03
𝜽 𝒇 ( °) 𝜽 𝒇 ( °) 𝜽 𝒇 ( °)
maximum pore 113 ± 0.03 121 ± 0.06 117 ± 0.02
nominal pore 109 ± 0.02 117 ± 0.04 124 ± 0.07
* ± values represent standard deviations
For all membranes, calculated 𝜃 𝑓 values lie between the intrinsic and measured contact angles. At both the
maximum and nominal pore scales, calculated 𝜃 𝑓 values are significantly larger than intrinsic contact angles
and significantly smaller than measured contact angles for all membranes (Table 3.4); differences are
statistically significant at α = 0.05. The differences between calculated 𝜃 𝑓 values and the intrinsic and
measured contact angles are likely due to (1) differences between the assumed in situ wetting state
(Wenzel state) and the wetting state between droplets used for sessile drop contact angle measurement
(Cassie-Baxter state), and (2) differences between the fibril roughness at the maximum and nominal pore
scales, the membrane surface roughness, and the surface roughness of an ideal, smooth solid (𝑅 ′
≡ 1).
Results show that calculated 𝜃 𝑓 values vary between different membranes of the same material (Table 3.4),
due to differences in fibril roughness and 𝑅 ′ (Table 3.3). At the nominal pore scale, calculated 𝜃 𝑓 values for
the QM050 and QM022 membranes are slightly smaller than 𝜃 𝑓 values at the maximum pore scale.
Conversely, for the QL822 membrane, 𝜃 𝑓 at the nominal pore scale is slightly larger than both 𝜃 𝑓 at the
maximum pore scale and 𝜃 𝑊 ,𝑚 ; this is due to the slightly larger 𝑅 ′
value at the nominal pore scale for
membrane QL822 (Table 3.3). Thus, depending on differences in measured fibril roughness with scale,
calculated 𝜃 𝑓 values do not necessarily decrease with scale. It should also be noted that the Wenzel model
assumes that surface roughness is homogeneous; significant surface roughness heterogeneity – especially
patterned surface roughness – may result in deviations in 𝜃 𝑓 with location on the membrane surface.

36

3.3.2 LEP Characterization
LEP Model Results
Table 3.5 shows LEP results from the Kim-Harriott model using the measured, fibril, and intrinsic contact
angles. To characterize minimum LEP, the maximum pore size was used for 𝑑 𝑝 , and 𝜃 𝑓 at the maximum
pore scale were used for all membranes. As expected, LEP values decrease with increasing membrane
pore size. Also, LEP values calculated using 𝜃 𝑓 are intermediate, lying between LEP values calculated
using the measured and intrinsic contact angles for all membranes. For the QM050 and QL822 membranes,
due to the relatively low calculated values of 𝜃 𝑓 at the maximum pore scale, LEP values calculated using
𝜃 𝑓 are nearer to the LEP values calculated using the intrinsic contact angle than to values calculated using
the measured contact angles. For the QM022 membrane, the LEP value calculated using 𝜃 𝑓 is nearly
equidistant from the LEP values calculated using the intrinsic and measured contact angles. This is likely
due to the higher fibril roughness, 𝑅 𝑓 ′
, and 𝜃 𝑓 values at the maximum pore scale for the QM022 membrane
compared to the QM050 and QL822 membranes (Table 3.3). Thus, LEP values based on 𝜃 𝑓 can vary for
membranes of the same material, not only due to differences in pore size but also due to differences in fibril
roughness.
Table 3.5 LEP results from the Kim-Harriott model based on the intrinsic contact angle (𝜃 0
), 𝜃 𝑓 , and
measured contact angle (𝜃 𝑚 ) for all membranes.
Model LEP (kPa)
Contact angle QM050 QM022 QL822
𝜽 𝒎 236 194 103
𝜽 𝒇 160 150 77
𝜽 𝟎 137 112 65

Experimental LEP Results
To determine whether calculated 𝜃 𝑓 , intrinsic contact angles, or measured contact angles are better
indicators of pore wetting resistance in MD, model and experimental LEP results were compared for the
QM022 membrane. Figure 3.6a shows experimental LEP results for the QM022 membrane in terms of Δ𝑃
and 𝐶 𝑑 versus time. The minimum change detected by the distillate conductivity probe is large relative to
the total change in distillate conductivity that occurred during the experiment. As a result, 𝐶 𝑑 appears to
increase stepwise (Figure 3.6a).

37

Figure 3.6 Experimental LEP results for the QM022 membrane in terms
of a) 𝛥𝑃 and 𝐶 𝑑 versus time, and b) 𝛥 𝐶 𝑑 /𝛥𝑡 versus time. Wetting
boundaries are indicated by dashed lines. Model LEP values based on 𝜃 0
,
𝜃 𝑓 at the maximum pore scale, and 𝜃 𝑚 are indicated on the y-axis in a)
with shaded bars representing the range in LEP resulting from the range
in maximum pore sizes measured from SEM images.
Figure 3.6b shows Δ𝐶 𝑑 /Δ𝑡 versus time. 𝐶 𝑑 increases initially from t = 0 to 1.4 h (Figure 3.6a) with decreasing
Δ𝐶 𝑑 /Δ𝑡 (Figure 3.6b), indicating decreasing solute flux (as discussed in Appendix A.1, Supplementary
Information). This shows that 𝐶 𝑑 is increasing due to solute mixing rather than pore wetting (Appendix A.1).
At t = 1.4 h, an increase in 𝐶 𝑑 occurs when Δ𝐶 𝑑 /Δ𝑡 is equal to the Δ𝐶 𝑑 /Δ𝑡 at the previous point (t = 1.1 h)
(Figure 3.6a and b). This may indicate the beginning of a linear trend in 𝐶 𝑑 , characteristic of constant wetting
(Appendix A.1); therefore, a potential wetting boundary is placed at t = 1.4 h. However, it is notable that
Δ𝐶 𝑑 /Δ𝑡 remained constant despite an increase in Δ𝑃 . When wetting occurs in LEP experiments, liquid and
solute flux are pressure driven. As Δ𝑃 increases beyond the LEP, increasing liquid and solute flux would
be expected, resulting in exponentially increasing 𝐶 𝑑 and Δ𝐶 𝑑 /Δ𝑡 . Also, additional smaller pores with slightly
higher LEP may become wetted, further increasing solute flux and Δ𝐶 𝑑 /Δ𝑡 . This is characteristic of
progressive wetting (Appendix A.1). Solute mixing can also result in an apparent linear increase in 𝐶 𝑑
initially, with constant Δ𝐶 𝑑 /Δ𝑡 (Appendix A.1, Figure A.1.1a and b). Therefore, the increase in 𝐶 𝑑 at t = 1.4
h in Figure 3.6a is likely due to continued solute mixing.
At t = 1.7 h, where an increase in both 𝐶 𝑑 and Δ𝐶 𝑑 /Δ𝑡 occur (Figure 3.6a and b), followed by increasing
Δ𝐶 𝑑 /Δ𝑡 with increasing Δ𝑃 (Figure 3.6b). As expected, 𝐶 𝑑 increases exponentially as Δ𝑃 increases after t =
1.7 h, indicating progressive wetting (Appendix A.1). Therefore, a second wetting boundary is placed at t =
1.7 h. Because it is not conclusive whether wetting began at t = 1.4 or 1.7 h, the experimental LEP value is
presented as a range based on the maximum Δ𝑃 values reached prior to t = 1.4 and 1.7 h.
The experimental LEP range is equal to or lower than that expected for LEP based on 𝜃 𝑓 , but greater than
that expected for LEP based on the intrinsic contact angle. This result suggests that calculated 𝜃 𝑓 values
give a more accurate representation of hydrophobicity at pore entrances in situ than the measured contact
angle and, therefore, that LEP based on 𝜃 𝑓 more accurately characterizes pore wetting resistance in situ.
38

The intrinsic contact angle may also give an accurate prediction of hydrophobicity and pore wetting
resistance than the measured contact angle, especially for membranes with smooth fibrils or pore walls.

3.4 Conclusions and Implications
In MD, measured or intrinsic contact angles are generally used to characterize membrane wetting
resistance in situ. However, discrepancies between actual and predicted pore wetting resistance may arise
due to differences in the wetting state and roughness relevant to measured contact angles, intrinsic contact
angles, and the fibril contact angles formed at pore entrances in situ during MD. In this study, the surface
properties and performance of three microporous PTFE membranes (QM050, QM022, and QL822, with
nominal pore sizes of 0.05, 0.2, and 0.45 μm, respectively) were characterized. Measured 𝑅 ′ and Wenzel
contact angle values at the membrane scale show that water droplets used for sessile drop contact angle
measurement are likely in a Cassie-Baxter state for all three membranes. However, at typical hydraulic
pressures in MD, feed solution is likely in a Wenzel state at the membrane surface. Therefore, measured
contact angles likely do not characterize wetting resistance in situ for membranes other than
superhydrophobic or omniphobic membranes that are known to operate in a stable Cassie-Baxter state
during MD.
Results show that measured fibril roughness and 𝑅 ′
at length scales relevant to the meniscus of feed
solution at pore entrances can be significantly lower than membrane roughness. For the first time, 𝜃 𝑓 is
calculated for hydrophobic PTFE membranes, based on the Wenzel model, measured fibril roughness, and
the intrinsic contact angle of PTFE. 𝜃 𝑓 values are unique: because fibril roughness is non-negligible, 𝜃 𝑓
values are significantly greater than intrinsic contact angles, and because the fibril roughness and assumed
in situ wetting state differ from the membrane roughness and wetting state of macroscopic droplets, 𝜃 𝑓
values are significantly lower than measured contact angles. Comparison of model and experimental LEP
results for the QM022 membrane suggest that 𝜃 𝑓 may better describe pore wetting resistance in situ than
the measured contact angle. Because 𝜃 𝑓 depends on fibril roughness, fibril roughness may be equally – or
more – important than membrane surface roughness to pore wetting resistance during MD. Furthermore,
𝜃 𝑓 can be used to easily characterize pore wetting resistance for any hydrophobic membrane and feed
solution, using known or measured intrinsic contact angles and fibril roughness. These results have
important implications for the design of wetting-resistant membranes, especially for low-surface-tension
feed solutions and applications where fouling and/or scaling are expected (i.e., when the feed solution is
expected to be in the Wenzel state at the membrane surface).
These results also have implications for internal hydrophobicity characterization. A few authors have
mentioned the concept of internal hydrophobicity or an internal contact angle to define wetting resistance
for MD membranes [87, 109, 110]. Researchers have also suggested that the internal roughness or
hydrophobicity of MD membranes may differ from surface values [87, 109, 111, 112]. Still, to the best of
our knowledge, no prior studies have quantified an internal contact angle for MD membranes. If fibril
39

roughness is assumed to be similar for both surface and internal fibrils, θ
f
values characterized in the current
study provide an estimation of the internal hydrophobicity.
Internal hydrophobicity may play an important role in wetting resistance after partial wetting occurs (Figure
3.7). Various studies have shown that MD membranes can operate in a partially wetted state; for example,
composite hydrophobic-hydrophilic membranes can operate with the hydrophilic surface facing the distillate
[44, 113-131] or feed stream [45, 132], even for multiple days of operation [132]. In our previous study (i.e.,
McGaughey et al. [109]) during long-term (100-day) operation, fouling-induced hydrophilization of the feed
side surface of a PTFE membrane that resulted in effectively zero LEP did not affect rejection; wetting was
not observed [109]. These results suggested that partial wetting occurred to some internal layer of the
membrane where pore throats resisted further penetration (e.g., Figure 3.7b). When fouling/scaling reduce
surface hydrophobicity and cause partial wetting, the internal and distillate-side fibril hydrophobicity, which
are less impacted by fouling and scaling, may play important roles in maintaining wetting resistance. The
results of this study further understanding of pore wetting during MD and provide new hydrophobicity
parameters (i.e., fibril roughness and 𝜃 𝑓 ) that may better describe pore wetting resistance in MD as well as
other processes that rely on the resistance of a porous material to liquid penetration.

Figure 3.7 Illustration of the surface fibril contact angle 𝜃 𝑓 ,𝑆 at
the feed side of a membrane and the internal fibril contact
angle 𝜃 𝑓 ,𝑖 formed at a pore throat within a membrane for a) an
unwetted membrane and b) a partially wetted membrane after
fouling/scaling occurs. Membranes are illustrated as cross-
sections of fibril mats; for simplicity, only a few rows of fibril
mats are shown, as indicated by the axis break symbol. Also,
as noted previously, in real membranes the fibril diameter,
spacing, orientation, and morphology may vary more widely
than shown here.

40

Acknowledgements
This material is based on work supported by the National Science Foundation (grant number 1820389).
The authors would also like to acknowledge support from a fellowship awarded to A. L. McGaughey by the
American Membrane Technology Association and the Affordable Desalination Coalition. AFM and SEM
data were collected at the University of Southern California Center of Excellence in NanoBioPhysics and
the University of Southern California Core Center of Excellence in Nano Imaging (CNI), respectively.

41

CHAPTER 4.

4. Hydrophobicity versus Pore Size: Polymer Coatings to Improve
Membrane Wetting Resistance for Membrane Distillation

Reprinted (adapted) with permission from A. L. McGaughey, P. Karandikar, M. Gupta, and A. E. Childress,
"Hydrophobicity versus Pore Size: Polymer Coatings to Improve Membrane Wetting Resistance for
Membrane Distillation," ACS Applied Polymer Materials, vol. 2, pp. 1256-1267, 2020. Copyright 2020
American Chemical Society.

Abstract
Initiated chemical vapor deposition (iCVD) was used to coat two porous substrates (hydrophilic cellulose
acetate (CA) and hydrophobic polytetrafluoroethylene (PTFE)) with a cross-linked fluoropolymer to improve
membrane wetting resistance. The coated CA membrane was superhydrophobic and symmetric. The
coated PTFE membrane was hydrophobic and asymmetric, with smaller pore size and lower porosity on
the top surface than on the bottom surface. Membrane performance was tested in membrane distillation
experiments with (1) a high-salinity feed solution and (2) a surfactant-containing feed solution. In both
cases, the coated membrane had higher wetting resistance than the uncoated membranes. Notably, wetting
resistances were better predicted by LEP distributions than by minimum LEP values. When LEP
distributions were skewed toward high LEP values (i.e., when small pores with high LEP were greater in
number), significant (measurable) salt passage did not occur. For the high-salinity feed solution, the coated
PTFE membrane had greater wetting resistance than the coated CA membrane; thus, reduced surface
pore size/porosity (which may reduce intrapore scaling) was more effective than increased surface
hydrophobicity (which may reduce surface nucleation) in preventing scaling-induced wetting. Reduced pore
size/porosity was equally as effective as increased hydrophobicity in resisting surfactant-induced wetting.
However, reduced porosity can negatively impact water flux; this represents a permeability/wetting
resistance trade-off in membrane distillation – especially for high-salinity applications. Membrane and/or
membrane coating properties must be optimized to overcome this permeability/wetting resistance trade-off
and make MD viable for the treatment of challenging streams. Then, increasing hydrophobicity may not be
necessary to impart high wetting resistance to porous membranes. These results are important for future
membrane design, especially as manufacturers seek to replace perfluorinated materials with
environmentally friendly alternatives.

4.1 Introduction and Background
Membrane distillation (MD) is a promising option for applications requiring treatment of high-salinity
wastewaters, high water recovery, or brine concentration [1, 24]. Unlike reverse osmosis, MD is only slightly
42

affected by feed water salinity because MD relies on a vapor pressure driving force [10, 28]; also, MD can
achieve high rejection of low-molecular-weight solutes that may pass through reverse osmosis membranes
[23, 25, 133]. Unlike conventional thermal distillation, MD can easily integrate with low-grade waste heat
and solar thermal power [1, 15, 16, 134, 135]; MD can also be more energy-efficient at small system
capacities [1].
Generally, microporous, hydrophobic, polymer membranes are used in MD; these are often composed of
polytetrafluoroethylene (PTFE), polyvinylidene fluoride, or polypropylene [84]. The vapor pressure
difference across the membrane drives water from a warmer, saline feed stream to evaporate at the
membrane surface and diffuse through the membrane pores [1]. In MD, permeability and water flux are
known to increase with increasing membrane porosity and to decrease with increasing membrane thickness
[1]. Increasing membrane pore size also increases permeability [1], although researchers have observed
that water flux is less sensitive to pore size than to porosity and thickness, especially at pore sizes less
than 0.3 μm [28].
When the membrane pores are unwetted, a vapor gap separates the feed and distillate streams and
prevents the passage of nonvolatile solutes [1, 25]. However, pore wetting does occur and is frequently
associated with salt precipitation (scaling) on the membrane surface [64] and/or the presence of surfactants
in the feed stream [14, 88]. Pore wetting and subsequent solute passage can be predicted by the wetting
resistance of a membrane. Increased wetting resistance is a key goal of many studies on new MD
membrane materials and/or coatings (e.g., Su et al. [136], Karanikola et al. [137], Wang and Lin [138], Boo
et al. [139], Servi and colleagues [140, 141], Warsinger et al. [142]).
4.1.1 Current Understanding of Membrane Wetting
Wetting can occur partially through a membrane or fully, through its depth. In partial wetting, the feed
solution enters the pore and reduces the vapor gap thickness. This can lead to full wetting via liquid bridging
across the thin vapor gap [35]. When the pore is fully wetted (i.e., there is no vapor gap), the feed and
distillate solutions can mix and solutes can pass through the membrane [14].
Expectations of wetting resistance are often based on membrane hydrophobicity, which is typically
characterized by surface contact angle (𝜃 ) measurement. Generally, a membrane is expected not to wet if
it is hydrophobic (i.e., has low surface energy with 𝜃 > 90°) and is expected to wet if it is hydrophilic (i.e.,
has high surface energy with 𝜃 < 90°). The maximum achievable contact angle for static water on a smooth
surface (i.e., the maximum contact angle due to material surface energy alone) is 120° [143, 144]. However,
contact angle measurements also depend on surface roughness [145]; on hydrophobic surfaces, increased
roughness increases the measured contact angle [144, 146]. Low-energy surfaces that are rough can be
“superhydrophobic”, (i.e., have 𝜃 ≥ 150°). As polymeric membrane surfaces are generally rough,
superhydrophobic MD membranes are most often produced using low-surface-energy materials or coatings
that reduce surface energy (e.g., Su et al. [136], Karanikola et al. [137], Servi and colleagues [140, 141],
43

Warsinger et al. [142] Sadeghi et al. [145], Munirasu et al. [147], Wang et al. [148], Li et al. [149], Liao et
al. [150], Guo et al. [151], Dong and colleagues [152, 153], Hamzah and Leo [154], Zhu et al. [155], Cong
and Guo [156], Meng and colleagues [157, 158], Lu et al. [67], Xiao and colleagues [159, 160]); some of
these coatings also increase surface roughness. A few researchers (e.g., Munirasu et al. [147], Wang et al.
[148], Li et al. [149]) have also produced superhydrophobic MD membranes specifically via patterned
surface microstructures on polymer membranes.
Membrane wetting resistance is also often quantified by liquid entry pressure (LEP), and it is assumed that
if the transmembrane pressure exceeds the LEP, the feed solution will flood the membrane pore(s). In the
MD literature, LEP is most frequently modeled by the Young-Laplace equation for a single pore:
𝐿𝐸𝑃 =
−2𝐵 𝛾 𝑙𝑣
cos( 𝜃 )
𝑟 ( 4.1)
where 𝐵 is a pore geometry factor (equal to one for cylindrical pores), 𝛾 𝑙𝑣
is liquid-vapor surface tension of
the wetting phase, and 𝑟 is pore radius. 𝜃 has been defined as the surface contact angle [18], the surface
advancing contact angle [110], the intrinsic contact angle (i.e., between the feed solution and a smooth
sample of membrane material) [161], and the intrinsic advancing contact angle [141].
For PTFE membranes, which typically consist of strands of polymer material (fibrils) connected at nodes
[84], Kim and Harriott [85] developed a model to describe wetting through a grid of fibrils of radius 𝑅 that
form pores of radius 𝑟 :
𝐿𝐸𝑃 =
−2𝛾 𝑙𝑣
𝑟 cos( 𝜃 − 𝛼 )
1 +
𝑅 𝑟 ( 1 − cos 𝛼 )
( 4.2)
where 𝛼 is the angle formed by the meniscus of the wetting liquid penetrating the pore. According to the
Kim and Harriott model [85], wetting resistance depends on the surface pore size. It is assumed that large
surface pores connect to various internal pores without bottlenecks. For membranes characterized by
interconnected internal pore structures and relatively high (e.g., 50%) porosities, such as the PTFE and CA
membranes used in the current study, internal bottlenecks are unlikely [85]. Guillen-Burrieza et al. [161]
found good agreement between experimental and theoretical results for the Kim-Harriott model using a
range of membranes, with 𝑅 /𝑟 determined from atomic force microscopy (AFM) results and 𝜃 as the intrinsic
advancing contact angle.
In previous studies, membrane LEP has typically been defined as a single value - the minimum LEP - which
represents wetting of the largest membrane pore. However, typical membranes are comprised of a range
of pore sizes and are characterized by a pore size distribution. During MD operation, wetting can only be
detected if sufficient solute passage occurs. It is likely that in real systems, wetting of multiple pores - rather
than the single largest pore - occurs before wetting is detected and becomes disruptive. Similarly, the
experimental LEP depends on the sensitivity of the measured response parameter. Previous researchers
have used various response parameters to detect wetting and LEP exceedance, such as continuous flow
[28], an increase in distillate conductivity [141], and visual observation of a droplet [161] or droplets [85]. If
44

distillate flowrate is used, for example, enough pores must be wetted for the flow to be detectable. Again,
wetting of multiple pores – rather than the single largest pore – may occur before wetting is detected.
In real systems, the theoretical minimum LEP (representing wetting of the single largest pore) may not be
of practical importance. Expressing LEP as a distribution, analogous to the pore size distribution, may
provide greater understanding and ability to consistently characterize wetting resistance. LEP distributions
have not previously been reported in the literature, to the best of our knowledge.
Scaling-Induced Wetting
Scaling hydrophilizes the membrane surface and reduces the LEP. Scaling occurs due to salt crystal
nucleation and growth. Nucleation will theoretically occur in any supersaturated system given sufficient time
[162]. Nucleation occurs when the Gibbs free energy barrier to nucleation (Δ𝐺 ∗
) is exceeded. Nucleation is
either primary or secondary (i.e., nucleation on existing crystals). Primary nucleation occurs either
homogeneously, in pure solution, or heterogeneously on a foreign surface (e.g., a membrane surface) [162].
The resistance of the membrane to scaling is generally quantified by the Gibbs free energy barrier to
heterogeneous nucleation on the membrane surface (Δ𝐺 𝑀 ∗
) [137, 162, 163].
Although surface scaling is often associated with wetting in MD [14], some researchers suggest that surface
scaling alone does not lead to wetting but that internal scaling is required for pore wetting to occur [35, 164].
Intrapore scaling causes the pores to have hydrophilic surfaces that may facilitate liquid penetration [35,
164]. Intrapore scaling can begin with surface scaling, via secondary nucleation on existing crystals, or it
can occur directly via intrapore nucleation, which is when crystals nucleate in pores or on pore walls.
Nucleation occurs in the aqueous phase [162]; therefore, intrapore nucleation can only occur when the
pores are at least partially wetted.
In previous studies, researchers have mainly focused on increasing the surface hydrophobicity of MD
membranes to reduce scaling and scaling-induced wetting [64] (e.g., Su et al. [136], Karanikola et al. [137],
Meng and colleagues [157, 158], Xiao and colleagues [159, 160]). It has been suggested that increasing
the surface hydrophobicity can prevent or delay scaling-induced wetting by increasing Δ𝐺 𝑀 ∗
[64]. This has
also been observed experimentally; in 2018, Karanikola et al. [137] developed a superhydrophobic
“slippery” membrane with high Δ𝐺 𝑀 ∗
and in 2019, Xiao et al. [159] developed a superhydrophobic,
micropillared slippery membrane. Under scaling conditions, the superhydrophobic, “slippery” membranes
outperformed commercial membranes, but eventually wetted. Interestingly, by comparing between
superhydrophobic and superhydrophobic, “slippery” membranes, Xiao et al. [160] observed that that Δ𝐺 𝑀 ∗

may be less important than feed solution hydrodynamics, which were affected by the surface patterns on
the slippery membrane, for controlling scaling (and, therefore, scaling-induced wetting).
Also, in MD, it has been reported that membrane scaling is dominated by crystal deposition (e.g., deposition
of crystals formed by secondary nucleation on surfaces in solution, such as colloidal particles) rather than
heterogeneous nucleation on the membrane itself [142]. In this case, scaling could occur regardless of Δ𝐺 𝑀 ∗
.
45

and the efficacy of superhydrophobic membranes would be limited. This is important for real systems with
more complex feed streams. In these cases, an alternative membrane modification strategy may be
necessary. Gryta [35] found that scaling-induced wetting could be mitigated by a low-porosity surface layer,
which sterically restricted intrapore crystal growth. However, the low-surface-porosity membrane was not
compared to a superhydrophobic membrane. To the best of our knowledge, our current study represents
the first comparison of the efficacy of superhydrophobic and small-pore-size/low-porosity membranes to
prevent scaling-induced wetting in MD. We hypothesize that reduced surface pore size and/or porosity
(which may reduce intrapore scaling) can be as - or more - effective than increased surface hydrophobicity
(which may reduce surface nucleation) in preventing scaling-induced wetting.
Surfactant-Induced Wetting
When considering applications requiring treatment of challenging streams (e.g., desalination of produced
waters from oil and gas operations), the presence of surfactants in the feed solution is also a concern [139].
Surfactant molecules can adsorb to a hydrophobic membrane surface with the hydrophilic head outward,
which effectively hydrophilizes the surface [89, 165]. Surfactants can also reduce the surface tension of the
feed solution, which reduces the LEP. If the LEP is sufficiently low at a pore entrance, partial wetting will
occur. After partial wetting occurs, surfactant molecules can readily adsorb to and hydrophilize the wetted
portion of the pore walls. As additional surfactant molecules are transported to the liquid-vapor interface,
the local surfactant concentration increases and the surface tension of the feed solution decreases further.
This can eventually lead to full wetting of the membrane [89].
Similar to scaling-induced wetting prevention, most previous researchers have focused on modifying
membrane surface hydrophobicity to prevent surfactant-induced wetting. Superhydrophobic membranes
and omniphobic membranes (i.e., hydrophobic membranes with re-entrant or concave pore openings that
provide further wetting resistance) have generally shown improved resistance to surfactant-induced wetting
during short-term operation (e.g., [138, 139, 166]). Superoleophobic (i.e., highly oil-repellant) membranes
and hydrophilic/oleophobic coatings on omniphobic substrates have also prevented surfactant-induced
wetting for up to 10 h (e.g., [167, 168]). Amphiphobic membranes (i.e., membranes that are both
hydrophobic and oleophobic) have shown improved resistance to wetting by oil- and surfactant-containing
feed solutions [169]. However, the ability of small-pore-size and/or low-porosity membranes to mitigate
surfactant-induced wetting has not been studied, to the best of our knowledge. We hypothesize that
reducing surface pore size and/or porosity may be as effective as increasing surface hydrophobicity to
prevent surfactant-induced wetting.
4.1.2 Membrane Modification
Notably, previous studies have focused on improving wetting resistance by increasing surface
hydrophobicity, likely in an effort to maintain high water flux. Improving wetting resistance by reducing the
membrane pore size may reduce water flux – especially if reducing the pore size also reduces the porosity.
This represents a permeability/wetting resistance trade-off in MD. The permeability/wetting resistance
46

trade-off in MD is similar to the permeability/selectivity trade-off in other membrane processes (e.g., reverse
osmosis) [170]. However, in MD, solute selectivity is lost when pore wetting occurs: this represents a step
change in selectivity that is unlike the continuous decrease in selectivity with increasing permeability that
exists for processes like reverse osmosis. Another example of the permeability/wetting resistance trade-off
in MD is operation at low flux when treating high-salinity feed solutions, which is done to reduce
concentration polarization and scaling that reduce membrane wetting resistance [14, 24].
Superhydrophobic membranes are most commonly fabricated by coating substrates with low-surface-
energy materials (e.g., Su et al. [136], Karanikola et al. [137], Servi and colleagues [140, 141], Warsinger
et al. [142], Sadeghi et al. [145], Liao et al. [150], Guo et al. [151], Dong and colleagues [152, 153], Hamzah
and Leo [154], Zhu et al. [155], Cong and Guo [156], Meng and colleagues [157, 158], Lu et al. [67]).
However, many coatings previously considered to be “gold standards” for imparting superhydrophobicity
consist of long-chain, perfluorinated compounds (e.g., [137, 141, 142, 145, 151-156]) that have been listed
as precursors to perfluorooctanoic acid (PFOA) [171]. PFOA is an environmentally persistent, toxic
compound of increasing regulatory concern [141]. Although some superhydrophobic membranes have
been fabricated using fluorinated compounds not listed as PFOA precursors, [67, 136, 140, 150, 157, 158]
these membranes have not been challenged by scaling- and surfactant-induced wetting conditions. To
replace perfluorinated materials with suitable alternatives for membrane processes, it is important to
mechanistically understand how the imparted properties delay or prevent wetting as well as identify
properties that may be more desirable.
Initiated chemical vapor deposition (iCVD) is a vapor-phase process that can be used to produce a wide
range of polymer coatings on textured substrates. iCVD does not require organic solvents, which makes it
more environmentally friendly and simpler to operate [143, 172] than other methods to produce
superhydrophobic surfaces, which are mostly solvent-based [173]. Recently, iCVD was shown to be viable
for roll-to-roll processing, making it an attractive option for large-scale functionalization of membranes [174].
In iCVD, vapor-phase monomer and initiator radicals diffuse to and adsorb on a cooled substrate from the
top down; polymerization occurs on the surfaces and pore walls of the cooled substrate [140, 143, 172].
For porous substrates, conformality (i.e., the evenness of coating thickness from the top to the bottom of
the substrate) is a key parameter. Nonconformal films, for which coating thickness decreases with depth,
are formed when initiator radicals and/or monomer molecules are depleted at pore entrances [140, 172]. In
iCVD, coating conformality generally increases with decreasing fractional monomer saturation in the reactor
[140, 172], and non-conformality generally increases with increasing aspect (length-to-diameter) ratio of
the substrate pores [140]. Coating conformality can affect permeability. For coated membranes with the
same surface pore size, permeability increases as conformality decreases. However, increasing pore
density has a greater effect on increasing permeability than reducing conformality [140]. The effect of
conformality on membrane wetting resistance for scaling- and surfactant-induced wetting conditions has
not been investigated, to the best of our knowledge.
47

4.1.3 Objectives
In this study, we use iCVD to coat a hydrophilic CA and hydrophobic PTFE substrate with
poly(1H,1H,2H,2H-perfluorodecyl acrylate-co-ethylene glycol diacrylate) (P(PFDA-co-EGDA)), a
crosslinked low-surface-energy material. Poly(1H,1H,2H,2H-perfluorodecyl acrylate) (PPFDA) coatings
have been previously shown to be highly effective in MD applications [141, 142, 151] but these coatings,
like P(PFDA-co-EGDA), are now considered unsuitable for commercial use due to PFOA concerns [141].
We compare coating conformality on the hydrophilic and hydrophobic substrates and test the uncoated and
coated membranes to assess their ability to prevent scaling- and surfactant-induced wetting in MD. For the
first time, the effectiveness of reduced pore size/porosity is compared with increased hydrophobicity to
prevent scaling- and surfactant-induced wetting. The LEP distribution concept is proposed and evaluated
to characterize membrane wetting resistance. Surface and performance characterization are performed to
identify key membrane modification strategies to increase wetting resistance, challenging past emphasis
on increasing surface hydrophobicity alone. Lastly, the permeability/wetting resistance trade-off in MD is
considered with respect to membrane modifications that improve resistance to wetting but may reduce
permeability.

4.2 Materials and Methods
4.2.1 Membranes and Membrane Modification
Two commercial flat-sheet membranes were used as substrates: an expanded polytetrafluoroethylene
(PTFE) membrane (Parker Performance Materials, Lees Summit, MO) and a cellulose acetate (CA)
membrane (Sterlitech, Kent, WA). The PTFE membrane is hydrophobic and symmetric, with a pore size of
0.2 μm, a bulk porosity of 84 ± 0.5%, and a thickness of 67 ± 5 μm. The CA membrane is hydrophilic and
symmetric, with a pore size of 0.2 μm, a bulk porosity of 55 ± 1%, and a thickness of 83 ± 2 μm. Bulk
porosity was measured according to the gravimetric method.[23] Reported values represent an average of
measurements for three samples of each membrane. For uncoated membranes, polymer densities of 2.2
and 1.3 g/cm
3
were used for PTFE and CA. For coated membranes, average polymer densities were
assumed to be similar to those of the corresponding substrates (due to the relatively small amount of coating
added). Membrane thicknesses were measured using a micrometer (MDC-1 PX, Mitutoyo, Kawasaki,
Japan) with an accuracy of ± 1 μm. Reported values represent an average of at least eight measurements
at different locations on each sample. The uncoated PTFE membrane, which has performed well during
long-term operation in previous studies (i.e., Gustafson et al. [84], McGaughey et al. [109]), was tested in
MD experiments for comparison with the coated membranes. The uncoated CA membrane was not used
for comparison because the pores would readily wet under MD conditions.
Both substrates were coated with the crosslinked fluoropolymer P(PFDA-co-EGDA). Crosslinking increases
the coating durability [175]. The coatings were fabricated using 1H,1H,2H,2H-perfluorodecyl acrylate
(PFDA) monomer (SynQuest, Alachua, FL), ethylene glycol diacrylate (EGDA) crosslinker (Polysciences,
48

Inc., Warrington, PA), and di-tert-butyl peroxide (TBPO) (98%) (Sigma-Aldrich, St. Louis, MO) initiator; all
were used as received without further purification.
The P(PFDA-co-EGDA) coating was deposited onto the substrates using a custom designed iCVD reactor
(GVD Corporation, Cambridge, MA) that is 250 mm in diameter and 48 mm in height. For each deposition,
a 6 x 6 cm substrate coupon was taped on the reactor stage, which was cooled to 30 °C via a recirculating
chiller. The PFDA and EGDA monomer jars were heated to 45 and 30 °C, respectively; the monomer lines
were heated to 15 °C above the jar temperatures to provide flow rates of 0.1 sccm for each. The TBPO
initiator jar was held at room temperature and a mass flow controller (MKS 1479A, Andover, MA) was used
to provide an initiator flow rate of 0.3 sccm. The filament array (80% Ni, 20% Cr) (Omega Engineering,
Stamford, CT) was resistively heated to 250 °C. A pressure transducer (Baratron capacitance manometer
622A01TDE, MKS) was used to measure the reactor pressure and an automated butterfly valve was used
to maintain a reactor pressure of 90 milliTorr, which resulted in a deposition rate of 30 nm/min. Deposition
rate was monitored on a reference silicon wafer in real-time using an in-situ laser interferometer (Helium-
Neon, 633 nm) (Industrial Fiber Optics, Tempe, AZ). Coatings were deposited to a final thickness of 400
nm.
4.2.2 Characterization
Surface Characterization
Contact angles (θ) were measured using a goniometer (Model 260, ramé-hart, Succasunna, NJ) and the
sessile drop method with 5-μL droplets. Reported values represent an average of at least five
measurements on the sample surface. Surface morphology was characterized using a field-emission
scanning electron microscope (JSM-7001, Jeol USA, Huntington Beach CA). Samples were sputter-coated
for 60 s prior to imaging. ImageJ software (version 1.49, National Institutes of Health, Bethesda MD) was
used to determine surface porosity (ϵ) and surface pore size (dp). Reported values represent averaged data
from SEM images of three arbitrary areas of both sides of each membrane. Surface pore size distributions
are reported as histograms of surface pore sizes normalized by probability density using MATLAB software
(version R2018b, MathWorks, Natick, MA); three histograms (corresponding to the three arbitrary areas
imaged) were obtained for each surface. For all distributions, the skewness, or asymmetry of the
distribution, was also calculated using MATLAB software. Positive skewness of a distribution indicates that
the distribution is weighted towards smaller values and negative skewness indicates that the distribution is
weighted towards larger values.
Surface roughness was analyzed using an atomic force microscope (Innova, Bruker, Billerica MA).
Gwyddion software (version 2.47, Czech Metrology Institute, Brno, Czech Republic) was used to calculate
surface roughness. Reported values represent averaged data from scans of three arbitrary areas of each
surface. Membrane surface elemental compositions were analyzed using x-ray photoelectron spectroscopy
(XPS) (Axis Ultra DLD, Kratos Analytical, Manchester, UK) with a monochromatic Al K-alpha x-ray source.
49

Survey spectra were collected over a binding energy range of 0 to 900 eV with a step size of 1 eV. The
XPS data were analyzed using CasaXPS software (Casa Software, Teignmouth, UK).
Δ𝐺 𝑀 ∗
was calculated according to [137, 163]:
Δ𝐺 𝑀 ∗
= Δ𝐺 ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 ∗
(
1
4
)( 2 + cos 𝜃 ) ( 1 − cos 𝜃 )
2
(1 − 𝜖 ( 1 + cos 𝜃 )
2
( 1 − cos 𝜃 )
2
)
3
( 4.3)
LEP was calculated according to the Kim and Harriott [85] model in equation (4.2), where 𝜃 is the intrinsic
advancing contact angle. For the PTFE membrane, intrinsic advancing contact angle values reported by
Morra et al. [105] were used; for the coated membranes, intrinsic advancing contact angle values were
measured by coating P(PFDA-co-EGDA) on a silicon wafer and measuring the advancing contact angle.
𝑅 /𝑟 was determined from AFM results and Gwyddion software according to the method validated by
Guillen-Burrieza et al. [161] Surface tension was taken as that of deionized water at 25 °C.
Theoretical minimum LEP values were calculated using the average maximum surface pore radius of each
membrane. LEP distributions were determined using all surface pore sizes (i.e., the surface pore size
distribution) from each area. LEP distributions were presented as histograms of LEP values normalized by
probability density. LEP distributions were determined for three arbitrary areas of each membrane surface.
Experimental LEP values were measured by gradually increasing the applied pressure on a saline solution
in contact with the feed side of a membrane sample placed in a pressure cell (Advantec MFS, Dublin, CA),
with deionized water on the distillate side. The experimental LEP was determined as the minimum pressure
at which an increase in distillate conductivity was detected.
Performance Characterization
Membrane performance was tested using a custom-built, bench-scale MD system (Supporting Information,
Figure A.2.1) and a custom-made acrylic module with an active area of 20 cm
2
. Mesh spacers (65 mm,
Sterlitech) were placed in the feed and distillate channels. For all experiments, the top side of the
membranes faced the feed solution.
Feed and distillate solutions were recirculated counter-currently at constant flow rates of 0.76 L/min using
a peristaltic pump (Masterflex, Cole Parmer, Vernon Hills, IL) on the feed side and a gear pump (Micropump,
Vancouver, WA) on the distillate side. Flow rates were monitored using manual flow meters (Omega
Engineering). Feed stream temperatures were controlled using a recirculating heater and distillate stream
temperatures were controlled using a recirculating chiller (Cole Parmer). Feed and distillate stream
temperatures and pressures were measured at the module inlet and outlet using in-line resistance
temperature detectors (Omega Engineering) and pressure probes (feed-side, Omega Engineering;
distillate-side, Cole-Parmer). Feed and distillate stream conductivities were continuously monitored using
in-line conductivity probes (Cole-Parmer). Non-metallic or titanium components were used on the feed side
to prevent corrosion. Data were recorded at 30-s intervals using a custom data acquisition and control
50

program developed in LabView (National Instruments, Austin, TX). Distillate overflow was continuously
monitored with a digital balance (Ohaus, Parsippany, NJ) and flux ( 𝐽 𝑤 in kg/m
2
s) was calculated as
𝐽 𝑤 =
𝑚 𝑑 𝑎 𝑚 Δ𝑡 ( 4.4)
where 𝑚 𝑑 is the change in distillate mass (kg) during time interval Δ𝑡 (s) and 𝑎 𝑚 is the membrane area (m
2
).
Rejection (𝑅 , unitless) was calculated according to
𝑅 = 1 −
𝐶 𝐷 ,𝑡 2
𝑉 𝐷 ,𝑡 2
− 𝐶 𝐷 ,𝑡 1
𝑉 𝐷 ,𝑡 1
𝐶 𝐹 ,𝑡 2
𝑉 𝐷 ,𝑡 2
− 𝐶 𝐹 ,𝑡 1
𝑉 𝐷 ,𝑡 1
( 4.5)
where 𝐶 𝐷 ,𝑡 is distillate concentration (mol/L NaCl) at time 𝑡 , 𝑉 𝐷 ,𝑡 is distillate volume at time 𝑡 (m
3
), and 𝐶 𝐹 ,𝑡
is feed concentration (mol/L NaCl) at time 𝑡 [134].
Scaling Experiments
For experiments evaluating scaling of the membrane, 5 M NaCl was used as the feed solution. ACS grade
NaCl (VWR) was used to prepare all solutions. The bench-scale system was operated in batch mode (i.e.,
without replenishment or dilution of feed solution) and experiments were terminated when distillate
conductivity exceeded 200 μS/cm or after 12 h. Operating temperatures and pressures were taken after
temperatures stabilized (i.e., after 1 h operation) until the end of each experiment. Average operating feed
and distillate stream temperatures were 52 ± 1 and 17 ± 0.8 °C. Average feed and distillate stream
pressures were 13 ± 1 and 12 ± 0.7 kPa; thus, the impact of hydraulic pressure difference on wetting and
salt flux were assumed negligible.
Surfactant Experiments
For experiments evaluating surfactant fouling of the membrane, 200 ppm anionic surfactant in 1 M NaCl
was used as the feed solution. Laboratory grade Triton X-100 (Sigma-Aldrich) and ACS grade NaCl (VWR)
were used to prepare all solutions. The bench-scale system was operated with constant feed solution
concentration for 5 h; if wetting did not occur by that time, the system was switched to batch mode to
increase surfactant concentration. Experiments were terminated when distillate conductivity exceeded 200
μS/cm or after 25 h. Operating temperatures were taken after temperatures stabilized (i.e., after 1 h of
operation) until the end of each experiment. Average feed and distillate stream temperatures were 52 ± 0.7
and 18 ± 0.9 °C. Average feed and distillate stream pressures were 12 ± 0.8 and 12 ± 0.9 kPa; thus, the
impact of hydraulic pressure difference on wetting and salt flux were assumed negligible.
In all cases, statistical significances were evaluated by performing two-sided t-tests at a significance level
(α) of 0.05. Unequal variances were assumed.

51

4.3 Results and Discussion
4.3.1. Surface Characteristics of Uncoated and Coated Membranes
Surface roughness and contact angle values for both sides of the uncoated and coated membranes are
shown in Table 4.1. For both uncoated and coated membranes, surface roughness values are greater for
the CA membranes than the PTFE membranes; the differences in surface roughness between the CA and
PTFE membranes are statistically significant (as confirmed by t-tests with α = 0.05). While the standard
deviations shown in Table 4.1 are relatively high, indicating spatial variability, the coefficients of variation
(i.e., the ratio of the standard deviation to the mean; not shown in Table 4.1) are comparable to those of
previous studies (e.g., Koyuncu et al. [176]).
Table 4.1 Root-mean-square 2-D surface roughness and contact angle values of uncoated and coated
membranes. The ± values in parentheses represent standard deviations.
Membrane
Surface roughness (nm) Contact angle (°)
Top Bottom Top Bottom
CA 468 (±156) 313 (±101) 0 0
PTFE 214 (±49) 99 (±3) 149 (±3) 145 (±2)

Coated CA 471 (±152) 438 (±99) 157 (±5) 158 (±8)
Coated PTFE 222 (±15) 115 (±33) 143 (±3) 145 (±7)

For all membranes, top and bottom contact angle measurements are similar (Table 4.1). The uncoated CA
membrane has a contact angle of zero on each side since water droplets are readily absorbed by the
hydrophilic membrane and a static sessile drop contact angle could not be measured. The generally greater
contact angles on the coated CA membranes compared to the coated PTFE membranes, despite both
membranes having the same coating, are likely due to the inherently rougher surface of the CA membrane,
as greater surface roughness results in higher contact angles on hydrophobic surfaces.[144-146]
Differences in contact angle between the CA and PTFE membranes are statistically significant (as
confirmed by t-tests with α = 0.05).
Representative SEM images for both sides of the uncoated and coated membranes are shown in Figure
4.1. Surface porosity and surface pore size values for all membranes are shown in Table 4.2.

52

Figure 4.1 Representative SEM images of a) top and b) bottom of the uncoated CA membrane; c) top and
d) bottom of the coated CA membrane; e) top and f) bottom of the uncoated PTFE membrane; and g) top
and h) bottom of the coated PTFE membrane. All images taken at x 4,000 magnification.
Table 4.2 Surface pore parameters of the uncoated and coated membranes. For the uncoated membranes,
parameters are shown as an average of both sides of the membranes since they are symmetric. For the
coated membranes, parameters are shown separately for each side due to the nature of the iCVD coating
technique. The ± values in parentheses represent standard deviations.
Membrane
Surface porosity
(%)
Average surface pore
size (nm)
Maximum surface pore
size (μm)
CA 27 (±8) 159 (±38) 3.2 (±0.9)
PTFE 21 (±2) 133 (±5) 1.3 (±0.1)

Top Bottom Top Bottom Top Bottom
Coated CA 21.0 (±4) 25.8 (±3) 132 (±4) 132 (±5) 3.4 (±1) 2.8 (±0.4)
Coated PTFE 3.5 (±2) 26 (±1) 64 (±5) 125 (±4) 1.4 (±0.3) 1.5 (±0.2)

Visual comparison of the top and bottom sides of the uncoated CA membrane (i.e., comparison of Figure
4.1a and b), shows similar morphology on both sides of the membrane. Comparison of the top and bottom
sides of the coated CA membranes (i.e., comparison of Figure 4.1c and d) also shows similar morphology.
Data in Table 4.2 show that the coating process did not significantly change the surface porosity, average
surface pore size, or maximum surface pore size of either side of the CA membrane, as confirmed by t-
tests with α = 0.05. The lack of significant visual or measured differences between the top and bottom sides
of the coated CA membrane indicates that the substrate was evenly coated from the top to the bottom; in
other words, the P(PFDA-co-EGDA) coating appeared to be conformal in the z-direction.
53

Similar to the uncoated CA membrane, visual comparison of Figure 4.1e and f shows similar morphology
on both sides of the uncoated PTFE membrane. However, for the coated PTFE membrane, it can be seen
on the top side (Figure 4.1g) that the fibrils and nodes have increased thicknesses, whereas on the bottom
side (Figure 4.1h) they appear relatively similar to the uncoated PTFE membrane. Thus, the P(PFDA-co-
EGDA) coating thickness appeared to be greater on the top side than on the bottom side, suggesting non-
conformality in the z-direction. Data in Table 4.2 support the visual observations: the top-side average
surface pore size and porosity decreased significantly after coating, but the bottom-side porosity and
average pore size did not (as confirmed by t-tests with α = 0.05). The maximum surface pore size did not
change significantly on either side. Pore size distribution results are shown in the Supporting Information
(Section A.2.2). For the coated membranes, bulk porosities were approximately 42 ± 3% and 66 ± 1%,
respectively. As expected, the bulk porosities of the membranes decreased after the membranes were
coated. For all membranes, surface porosity and average pore size values are lower than the corresponding
bulk values likely due to larger voids present in the bulk of the membrane (e.g., Supporting Information
Figure A.2.2).
Elemental composition results from XPS analysis of the uncoated and coated membranes are shown in
Table 4.3. For the uncoated CA membrane, only carbon and oxygen were detected, as expected. For the
coated CA membrane, fluorine was detected, indicating the presence of the P(PFDA-co-EGDA) coating;
XPS results were similar on both sides of the coated CA membrane, indicating that the coating is conformal.
For the uncoated PTFE membrane, carbon, oxygen, fluorine, and silicon were detected. On the top side of
the coated PTFE membrane, a higher fluorine percentage and lack of silicon indicate that the coating
thickness exceeded the XPS penetration depth of 5 nm; on the bottom side, the appearance of the silicon
peak indicates that the penetration depth exceeded the coating thickness (i.e., the coating is less than 5
nm thick). These results indicate that the coating on the PTFE membrane is thicker on the top than on the
bottom (i.e., the coating is nonconformal). Also, less fluorine was detected on the bottom side of the coated
PTFE membrane than on the top side, which may indicate a compositional gradient in the copolymer
coating in the z-direction (i.e., a higher PFDA:EGDA ratio on the top side than on the bottom side). The
XPS results are consistent with the SEM results (Figure 4.1), pore parameter results (Table 4.2), and pore
size distribution results (Supporting Information, Figure A.2.3).

54

Table 4.3 Percent composition results from XPS analysis of uncoated and coated membranes. For
uncoated membranes, results are shown for a single side representing the symmetric material. For the
coated membranes, results are shown separately for each side due to the nature of the iCVD coating
technique.
Membrane
Percent composition (%)
Carbon Oxygen Fluorine Silicon
CA 52 48 -* -*
PTFE 50 21 24 5

Top Bottom Top Bottom Top Bottom Top Bottom
Coated CA 41 43 6 6 53 50 -* -*
Coated PTFE 32 59 7.5 26 60 13 -* 2.6
* - indicates element was not detected
As discussed in Section 4.1.2, conformality in iCVD is affected by the fractional saturation of monomer in
the reactor and by the aspect ratio of the substrate pores [140, 172]. In the present study, an identical
coating material and identical operating parameters (including the fractional monomer saturation) were
used. Regarding aspect ratio: due to the greater thickness of the CA substrate (83 ± 2 μm) than of the PTFE
substrate (67 ± 5 μm), the aspect ratio of the CA substrate is slightly higher, which was expected to result
in less conformal coatings compared to the PTFE substrate. However, the opposite was observed: the CA
substrate coating is more conformal than the PTFE substrate coating. Therefore, it appears that substrate
surface energy may affect conformality. The monomer appears to have had greater affinity for the high-
surface-energy (~40 mN/m [177]) CA substrate than the low-surface-energy (~21 mN/m [178]) PTFE
substrate and, therefore, adsorbed more uniformly on the CA substrate.
4.3.2. Liquid Entry Pressure
Figure 4.2 shows the theoretical minimum LEP and LEP distribution for the top (feed) side of the
membranes. Despite the higher contact angle of the coated CA membrane compared to the coated PTFE
membrane (Table 4.1), the coated CA membrane has a lower minimum LEP due to its higher maximum
pore size (as calculated by the Young-Laplace and Kim-Harriott LEP models and shown in Table 4.2).
Based on minimum LEP results, the coated CA membrane is more likely to wet than the uncoated PTFE
and coated PTFE membranes.
To verify the theoretical LEP results, the experimental LEP was measured for the uncoated PTFE
membrane. Based on four samples, the experimental minimum LEP was 102 ± 7 kPa; this value agrees
well with theoretical results (Figure 4.2a). As experimental LEPs of commercial membranes used for MD
are typically between 100 and 440 kPa [28]; the values in the current study are at the lower end of this
range.

55

Figure 4.2 LEP distributions for the top side of the a) uncoated PTFE, b) coated PTFE, and c) coated CA
membranes. For each membrane, results are shown as three overlaid histograms normalized by probability
density, as discussed in Section 2.
For LEP measurements of the uncoated PTFE membrane, the experimental minimum LEP was determined
when a slight increase in distillate conductivity occurred. The rate of change of distillate conductivity
increased as pressure was increased further. At approximately 345 kPa, which agrees with previously
reported values [28], more significant salt passage and distillate flow (as may be detectable in an MD
system) occurred. This higher LEP value likely represents wetting of multiple pores. As discussed in Section
1.1, in LEP systems - and in real MD systems - wetting of multiple pores likely occurs before wetting is
detected. Therefore, the theoretical minimum LEP (representing wetting of the largest pore) may not be of
practical importance.
Because LEP is dependent on pore size and membranes are characterized by a pore size distribution,
expressing LEP as a distribution will provide greater ability to characterize wetting resistance. The
skewness of the LEP distributions, which results directly from pore size distribution skewness (Figure
A.2.3), can be used to indicate membrane wetting resistance. The uncoated PTFE membrane (Figure 4.2a)
has a positively skewed LEP distribution and both coated membranes have negatively skewed LEP
distributions (Figure 4.2b and c). Negatively skewed LEP distributions indicate that the distribution is
weighted toward higher LEPs or that there is a greater frequency of surface pores with high LEP, which
may indicate higher overall wetting resistance. For all membranes, differences in skewness were
statistically significant at α = 0.05. The greater likelihood of the uncoated PTFE membrane to wet than
either of the coated membranes is in contrast with the minimum LEP results (Figure 4.2), which suggest
that the coated CA membrane is most likely to wet.
4.3.3. Scaling-induced Wetting Performance
Performance results under scaling-induced wetting conditions are shown in Figure 4.3. In terms of water
flux, both coated membranes had lower fluxes than the uncoated PTFE membrane but operated stably
(Figure 4.3a). The lower water flux observed for the coated PTFE membrane relative to the uncoated PTFE
56

membrane is likely in large part due to its lower bulk and surface porosity. The lower water flux for the
coated CA membrane is likely due to its lower bulk porosity and greater thickness. For the uncoated PTFE
membrane, distillate conductivity sharply increased (Figure 4.3b) and conductivity rejection decreased
(Figure 4.3c) (i.e., wetting occurred) after approximately 1.3 h operation. Wetting was likely due to
concentration polarization resulting in exceedance of the saturation concentration of NaCl (5.5 M at 53 °C)
and subsequent scaling [64]. Scaling at the membrane surface could then lead to intrapore crystal growth,
which provides a hydrophilic surface that facilitates liquid entry into the pores.

Figure 4.3 Performance of the uncoated PTFE, coated PTFE, and coated CA membranes with 5 M NaCl
feed solution in terms of a) water flux, b) distillate conductivity, and c) solute rejection.
Both coated membranes experienced delayed wetting and salt passage relative to the uncoated PTFE
membrane. 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
results for the uncoated PTFE and coated membranes are shown in Table 4.4. The
slightly higher 𝛥𝐺
𝑀 ,𝑁𝑎𝐶𝑙 ∗
of the coated CA membrane compared to the uncoated PTFE membrane (Table
4.4) may have delayed surface nucleation, and hence, scaling and wetting; the difference is statistically
significant at α = 0.05. However, after ~4.3 h operation, wetting did occur for the coated CA membrane
(Figure 4.3b and c). Heterogeneous nucleation and/or deposition of crystals nucleated in bulk solution may
have resulted in surface scaling and led to wetting. Notably, 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
for the coated CA membrane is near
the theoretical maximum 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
(455 MJ/mol) obtained at a contact angle of 180° and a surface porosity
of zero; thus, additional increases in contact angle (or reductions in surface porosity) would result in minimal
increases in 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
.
Table 4.4 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
of the uncoated PTFE, coated PTFE, and coated CA membranes; theoretical
calculations are based on a concentration of 5 M NaCl and a temperature of 25 °C.
Membrane 𝚫 𝑮 𝑴 ,𝑵𝒂𝑪𝒍 ∗
(MJ/mol)
PTFE 447
Coated PTFE 445
Coated CA 454
57

The coated PTFE membrane did not wet over the duration of the experiment (Figure 4.3b and c).
Interestingly, this was the case even though the coated PTFE membrane has a similar 𝛥𝐺
𝑀 ,𝑁𝑎𝐶𝑙 ∗
to the
uncoated PTFE membrane (Table 4.4). Gryta [35] suggests that small pore size (relative to crystal size)
can restrict intrapore crystal growth. However, size exclusion likely will not prevent all intrapore scaling. For
example, for sodium chloride, critical nucleus sizes as small as ~10 molecules have been reported [162];
these could form in small pores. However, this mechanism of intrapore nucleation can only occur in already-
wetted pores. We suggest that the higher LEP of the coated PTFE membrane compared to the uncoated
PTFE membrane (Figure 4.2) may prevent partial wetting and, therefore, intrapore nucleation and crystal
growth – regardless of size exclusion effects. Notably, the LEP of the uncoated PTFE membrane is high
because of its small average surface pore size (Table 4.2); this is not due to size exclusion but because
smaller pores mean that the curvature of the meniscus of the penetrating solution is greater, which
increases the Laplace pressure of the meniscus [179] and, therefore, the LEP. By preventing wetting, the
higher LEP at the entrance of smaller pores may prevent intrapore nucleation and crystal growth regardless
of size exclusion effects. Surface roughness effects may also contribute; as noted by Meng et al. [158],
higher surface roughness may increase the number of available nucleation sites on a membrane. The
coated CA membrane is more rough than the coated PTFE membrane (Table 4.1).
To more broadly examine the effectiveness of increased hydrophobicity and Δ𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
as a strategy to
reduce scaling-induced wetting, a sensitivity analysis of 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
was performed (Figure 4.4). Ranges of
contact angle (90 to 180°) and surface porosity (0 to 1) were selected to bracket the membranes used in
this study. It can be seen that above a contact angle of 140°, 𝛥 𝐺 𝑀 ,𝑁𝑎 𝐶 𝑙 ∗
approaches a constant value.
Therefore, increasing hydrophobicity to increase 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
of a membrane is inherently limited as a scaling
mitigation strategy.
58

Figure 4.4 Sensitivity analysis of 𝛥 𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
for
heterogeneous NaCl nucleation from 5 M solution at 25°C
to contact angle (𝜃 ) and surface porosity (𝜖 ). Theoretical
calculations are based on a concentration of 5 M NaCl
and a temperature of 25 °C.

4.3.4. Surfactant-induced Wetting Performance
Performance results under surfactant-induced wetting conditions are shown in Figure 4.5. Similar to scaling-
induced wetting conditions, both coated membranes had lower water fluxes than the uncoated PTFE
membrane but operated stably (Figure 4.5a). Also similar to the scaling-induced wetting conditions, the
lower water flux observed for the coated PTFE membrane is likely due to its lower porosity relative to the
uncoated PTFE membrane, and the lower water flux observed for the coated CA membrane is likely due
to its lower bulk porosity and greater thickness compared to the uncoated PTFE membrane. For the
uncoated PTFE membrane, wetting occurred after approximately 1.3 h operation (Figure 4.5b and c), likely
due to reduced surface tension at the membrane surface resulting from concentration polarization [88].
59

Figure 4.5 Performance of uncoated PTFE, coated PTFE, and coated CA membranes with 200 ppm
surfactant in 1 M NaCl feed solution in terms of a) water flux, b) distillate conductivity, and c) solute rejection.

No wetting was observed for either coated membrane (Figure 4.5b and c). Unlike for scaling-induced
wetting conditions, the difference in surface pore size and porosity between the coated membranes (Table
4.2) did not correspond to a difference in surfactant-induced wetting resistance. Instead, the more
hydrophobic coated CA membrane and the smaller surface pore size/porosity coated PTFE membrane had
equivalent wetting resistance over the duration tested.
Also, the difference in minimum LEP values for the coated membranes (Figure 4.2) did not correspond to
a difference in wetting resistance. Instead, surfactant-induced wetting results corresponded to expectations
based on LEP distributions: while the uncoated PTFE membrane had a positively skewed LEP distribution
(Figure 4.2a), both coated membranes had negatively skewed LEP distributions (Figure 4.2b and c). In
other words, both coated membranes had fewer large surface pores with low LEPs, which may have
prevented measurable wetting. Additionally, large surface pores located adjacent to each other, which can
facilitate wetting [64], may be less likely in the coated membranes due to the low frequency of large surface
pores.
To more broadly examine the impact of pore size and contact angle on wetting resistance, a sensitivity
analysis of LEP to contact angle and pore diameter was performed (Figure 4.6). The contact angle range
(90 to 180°) was selected to represent the full range of hydrophobic membranes; the pore size range (0.05
to 0.3 μm) was selected to represent average pore sizes typically used for MD membranes. Figure 4.6
illustrates how LEP generally increases with increasing contact angle (and, therefore, with decreasing
surface energy) and how LEP generally increases with decreasing pore size, as defined by the Young-
Laplace model. Small pore diameters are necessary to access the highest LEP values because LEP is
inversely proportional to pore diameter. Despite this, as discussed in Section 1, the majority of MD
membrane modification strategies have focused on increasing membrane surface hydrophobicity to
60

increase wetting resistance. This is likely because reducing porosity and pore size can negatively affect
water flux [18, 28].

Figure 4.6 Sensitivity analysis of LEP to contact angle (𝜃 )
and pore diameter (𝑑 𝑝 ). Theoretical calculations are
based on pure water at 25 °C.

4.4 Conclusions and Implications
The CA membrane was coated conformally and became symmetric and superhydrophobic. The PTFE
membrane was coated nonconformally and became asymmetric, with smaller pore size and lower porosity
on the top than on the bottom surface. The difference in conformality may be linked to substrate surface
energy: for the CA membrane, higher substrate surface energy may have resulted in greater affinity for the
monomer, leading to more uniform monomer concentration within pores and therefore formation of a more
conformal polymer layer. In MD performance testing, both coated membranes outperformed the uncoated
PTFE membrane. However, key differences between the performances of the coated membranes were
observed.
Under scaling-induced wetting conditions, the coated PTFE membrane, due to its smaller surface pore size
and porosity, had higher wetting resistance than the coated CA membrane, despite the higher Δ𝐺 𝑀 ,𝑁𝑎𝐶𝑙 ∗
of
the coated CA membrane. While increasing surface hydrophobicity can delay surface nucleation, surface
scaling can also occur by deposition of salt crystals formed elsewhere. Prevention of intrapore scaling may
be necessary to prevent wetting in longer-term operation. Small surface pore sizes and/or porosity may
more effectively prevent intrapore scaling due to high LEP at pore entrances, which may prevent partial
wetting and, therefore, intrapore nucleation. Under surfactant-induced wetting conditions, the coated
membranes performed equally well. The reduced surface pore size and porosity of the coated PTFE
membrane were as effective as the increased surface hydrophobicity of the coated CA membrane. Also,
61

wetting resistances were better predicted by LEP distributions than by minimum LEP values. When LEP
distributions are (negatively) skewed towards higher LEP values, small pores with high LEP are greater in
number and significant, measurable salt passage may be prevented.
The results of this study challenge past emphasis on increasing surface hydrophobicity alone to improve
the wetting resistance of MD membranes. Reducing surface pore size and porosity can be more effective
than increasing surface hydrophobicity and may be a viable strategy to impart high wetting resistance
without relying on perfluorinated coating materials. However, importantly, reducing surface porosity and
pore size can negatively impact flux, resulting in a permeability/wetting resistance trade-off in MD –
especially for high-salinity applications. If key properties are optimized (e.g., pore density is maximized
and/or the thickness of the small-pore-size/porosity surface layer or membrane is minimized), it may be
possible to overcome the permeability/wetting resistance trade-off. When coating membranes to achieve
high wetting resistance without sacrificing water flux, coating conformality could be minimized [140]. It has
also been suggested that permeability can be maintained by increasing the pore density and/or by reducing
the thickness of the small-pore-size surface layer [140], but the effect on wetting resistance for challenging
solutions has not yet been considered. Thin, self-standing, dense hydrophobic membranes (e.g., Mejia
Mendez et al. [180]) and hydrophobic membranes with thin, low porosity surface layers (e.g., Shaulsky et
al. [181] and Nejati et al. [182]) could overcome the permeability/wetting resistance trade-off in MD;
however, these membranes have not yet been tested under wetting conditions.

Associated Content
Supporting Information
Schematic of bench-scale MD system used for performance characterization, cross-sectional SEM results,
surface pore size distributions of uncoated and coated membranes, and SEM and energy dispersive x-ray
spectroscopy analysis results for used membranes after scaling- and surfactant-induced wetting
performance tests.

Acknowledgements
This material is based on work supported by the National Science Foundation under grant number 1820389.
We would also like to acknowledge funding support from an American Membrane Technology Association
and Affordable Desalination Coalition fellowship awarded to A. L. McGaughey. SEM and XPS data were
collected at the University of Southern California Core Center of Excellence in Nano Imaging (CNI) and
AFM data was collected at the University of Southern California Center of Excellence in NanoBiophysics.

62

CHAPTER 5.

5. Conclusions
In this work, three studies investigating pore wetting in MD were performed to elucidate the roles of surface
and pore properties in membrane wetting during MD. New parameters were introduced to better describe
pore wetting resistance in situ. First, the effect of long-term operation on the performance and
hydrophobicity of an MD membrane was investigated; for the first time, it was shown that irreversible
changes in hydrophobicity can occur after long-term operation, even in the absence of significant fouling.
Also, equally important, it was shown that rejection remained high despite loss of feed-side hydrophobicity:
this was the first indication that both distillate surface and internal hydrophobicity can play important roles
in membrane wetting resistance. Second, to further investigate the role of hydrophobicity in wetting
resistance, the fibril hydrophobicity at pore entrances of hydrophobic, polytetrafluoroethylene membranes
was characterized based on the Wenzel model and measured fibril roughness. It was shown that calculated
fibril contact angles may better describe in situ pore wetting resistance than measured contact angles.
Finally, the roles of hydrophobicity and pore size in wetting resistance were compared and a new wetting
resistance parameter – the liquid entry pressure distribution – was introduced to characterize wetting
resistance based on pore size distribution rather than maximum pore size. Detailed conclusions from each
study are discussed below.
In the first study, significant changes to membrane surface morphology and hydrophobicity were observed
on both the feed and distillate sides of MD membranes subjected to 100 days of continuous operation. On
the feed side, the measured contact angle decreased by 56% and roughness decreased by 92%. On the
distillate side, the measured contact angle decreased by 26% and roughness decreased by 57%. Moderate
morphological changes and reductions in contact angle by 65 and 19% also occurred on the feed and
distillate sides of membranes subjected to 20 days of continuous operation and scaling. Fouling was
observed despite the use of “clean” feed solutions. While fouling/scaling deposits on the feed side of used
membranes were expected to result in reduced hydrophobicity, changes on the distillate side occurred in
the absence of fouling, indicating that the membrane surface itself was changed. For the first time, it was
shown that irreversible losses of membrane hydrophobicity can occur after long-term use. While membrane
hydrophobicity has previously been assumed to be recoverable if foulants can be fully removed,
hydrophobicity losses due to physical or chemical changes in the membrane material may not be reversible.
These results have implications for membrane durability and lifetime. When feed-side surface
hydrophobicity is impaired due to fouling or scaling, distillate-side and internal hydrophobicity – which are
less affected by fouling/scaling deposits – may become equally or more important to membrane wetting
resistance. The effects of long-term operation may be especially relevant to membranes that rely on a thin
63

surface coating or active layer to impart wetting resistance, as feed-side surface layers may be more
vulnerable than the internal layers and distillate-side surface of the membrane.
In the second study, the implicit assumptions inherent in using either the intrinsic or measured contact
angles to indicate and calculate wetting resistance of MD membranes were examined. For
polytetrafluoroethylene membranes, measured contact angles reflect macroscopic wetting in the Cassie-
Baxter wetting state – yet literature results show that membranes likely operate in a Wenzel state during
MD. It was shown that measured fibril roughness can be significantly less than membrane roughness (i.e.,
roughness of an area spanning multiple fibrils and pore openings) but is also non-negligible. For the first
time, a fibril-scale contact angle was characterized, based on the Wenzel model, intrinsic contact angles,
and measured fibril roughness. The relevance of intrinsic, fibril, and measured contact angles to in situ pore
wetting resistance was assessed by comparing model and experimental liquid entry pressure results.
Results show that while measured contact angles overestimate pore wetting resistance, the fibril contact
angle may better describe pore wetting resistance in situ. The fibril contact angle can be used to
characterize pore wetting resistance for any hydrophobic membrane or feed solution based on intrinsic
contact angles and fibril roughness. Furthermore, the fibril contact angle may provide a quantitative
estimation of the internal hydrophobicity of MD membranes, which has not previously been characterized.
These results provide fundamental understanding of hydrophobicity and wetting resistance parameters
relevant to pore wetting for MD, as well as other processes that rely on the resistance of a porous material
to liquid penetration.
In the third study, the relative roles of hydrophobicity and pore size – as well as pore size distribution – in
wetting resistance were assessed, both for surfactant- and scaling-induced wetting. Performance
characterization of modified membranes for a high-salinity feed solution (resulting in scaling) and for a
surfactant-containing feed solution (resulting in reduced surface tension) showed that membranes with
increased surface hydrophobicity and membranes with reduced pore size/porosity had equivalent
resistance to surfactant-induced wetting but different resistances to scaling-induced wetting. Reduced pore
size/porosity was found to be more effective than increased surface hydrophobicity to prevent scaling-
induced wetting. While increased surface hydrophobicity is associated with reduced nucleation potential
and delayed scaling on the membrane surface, deposition of formed crystals likely eventually induces
wetting. On the other hand, reduced surface pore size/porosity may more effectively mitigate internal
scaling by increasing the liquid entry pressure at pore openings and preventing nucleation and deposition
of crystals within membrane pores. This may prevent wetting for a longer duration. However, due to the
negative impact of reduced pore size/porosity on permeability and water flux, this represents a wetting
resistance/permeability trade-off in MD that is especially important for high-salinity streams. For MD to
become viable for the treatment of challenging brines, this trade-off must be overcome (for example, by
fabricating materials with thin, small-pore-size surface layers or materials with densely packed small pores
to achieve high surface porosity). Finally, a liquid entry pressure distribution parameter was introduced to
64

characterize wetting resistance based on the pore size distribution rather than based on the maximum pore
size alone. Membranes with liquid entry pressure distributions that are skewed towards higher values (i.e.,
membranes with pore size distributions that are skewed towards smaller pores with higher liquid entry
pressures) may prevent significant, measurable salt passage in situ during MD. Furthermore, while
minimum liquid entry pressure describes wetting resistance of the single, largest pore, wetting may initiate
at other pores during MD operation. In scaling-induced wetting, salt crystals may form or deposit at pores
other than the single largest pore. In surfactant-induced wetting, stagnant areas in membrane modules and
concentration polarization may result in reduced surface tension and wetting at membrane pores other than
the largest pore. Therefore, the liquid entry pressure distribution must be considered to fully describe the
wetting resistance of membranes in situ.
This work combined fundamental scientific principles with engineering approaches to achieve systematic
understanding of wetting mechanisms, membrane wetting resistance, and membrane durability in MD –
both during long-term operation and for challenging, high-salinity streams. This work is increasingly relevant
as freshwater scarcity and impairment drive suppliers to exploit alternative sources, including challenging,
high-salinity streams. The results of this dissertation not only further fundamental understanding of pore
wetting during MD but can also guide membrane material selection and design for improved wetting
resistance in MD. Increasing the roughness of surface and internal fibrils to increase fibril hydrophobicity
may enable wetting resistance if feed side hydrophobicity is reduced – either slowly, during long-term
operation, or due to fouling and/or scaling. Reducing surface pore size while maintaining high bulk porosity
and/or low membrane thickness may enable MD membranes to achieve both high permeability and high
wetting resistance – especially for high-salinity streams. Finally, controlling pore size distributions to
increase the skewness of liquid entry pressure distributions toward higher values may enable more robust
wetting resistance in situ. Robust wetting resistance is critical to make MD viable as a sustainable method
for management and reclamation of challenging waste streams.

65

References
[1] A. Deshmukh, C. Boo, V. Karanikola, S. Lin, A. P. Straub, T. Tong, et al., "Membrane distillation at
the water-energy nexus: limits, opportunities, and challenges," Energy & Environmental Science,
vol. 11, pp. 1177-1196, 2018.
[2] X. Wei, Z. M. Binger, A. Achilli, K. T. Sanders, and A. E. Childress, "A modeling framework to
evaluate blending of seawater and treated wastewater streams for synergistic desalination and
potable reuse," Water research, vol. 170, p. 115282, 2020.
[3] M. O. Mavukkandy, C. M. Chabib, I. Mustafa, A. Al Ghaferi, and F. AlMarzooqi, "Brine management
in desalination industry: From waste to resources generation," Desalination, vol. 472, p. 114187,
2019.
[4] J. Morillo, J. Usero, D. Rosado, H. El Bakouri, A. Riaza, and F.-J. Bernaola, "Comparative study of
brine management technologies for desalination plants," Desalination, vol. 336, pp. 32-49, 2014.
[5] E. Jones, M. Qadir, M. T. van Vliet, V. Smakhtin, and S.-m. Kang, "The state of desalination and
brine production: A global outlook," Science of the Total Environment, vol. 657, pp. 1343-1356,
2019.
[6] D. L. Shaffer, L. H. Arias Chavez, M. Ben-Sasson, S. Romero-Vargas Castrilló n, N. Y. Yip, and M.
Elimelech, "Desalination and reuse of high-salinity shale gas produced water: drivers, technologies,
and future directions," Environmental science & technology, vol. 47, pp. 9569-9583, 2013.
[7] H. Chang, T. Li, B. Liu, R. D. Vidic, M. Elimelech, and J. C. Crittenden, "Potential and implemented
membrane-based technologies for the treatment and reuse of flowback and produced water from
shale gas and oil plays: A review," Desalination, vol. 455, pp. 34-57, 2019.
[8] S. Lin, "Energy Efficiency of Desalination: Fundamental Insights from Intuitive Interpretation,"
Environmental Science & Technology, vol. 54, pp. 76-84, 2019.
[9] M. Elimelech and W. A. Phillip, "The future of seawater desalination: energy, technology, and the
environment," science, vol. 333, pp. 712-717, 2011.
[10] T. Y. Cath, V. D. Adams, and A. E. Childress, "Experimental study of desalination using direct
contact membrane distillation: a new approach to flux enhancement," Journal of Membrane
Science, vol. 228, pp. 5-16, 2004.
[11] L. F. Greenlee, D. F. Lawler, B. D. Freeman, B. Marrot, and P. Moulin, "Reverse osmosis
desalination: water sources, technology, and today's challenges," Water research, vol. 43, pp.
2317-2348, 2009.
[12] M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, B. J. Marinas, and A. M. Mayes,
"Science and technology for water purification in the coming decades," in Nanoscience and
technology: a collection of reviews from nature Journals, ed: World Scientific, 2010, pp. 337-346.
[13] T. Horseman, Y. Yin, K. S. Christie, Z. Wang, T. Tong, and S. Lin, "Wetting, Scaling, and Fouling
in Membrane Distillation: State-of-the-Art Insights on Fundamental Mechanisms and Mitigation
Strategies," ACS ES&T Engineering, 2020.
[14] M. Rezaei, D. M. Warsinger, J. H. Lienhard V, M. C. Duke, T. Matsuura, and W. M. Samhaber,
"Wetting phenomena in membrane distillation: Mechanisms, reversal, and prevention," Water
Research, vol. 139, pp. 329-352, 2018/08/01/ 2018.
[15] F. Suárez, S. W. Tyler, and 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, vol. 44, pp. 4601-4615, 2010/08/01/ 2010.
[16] F. Suárez, J. A. Ruskowitz, S. W. Tyler, and A. E. Childress, "Renewable water: direct contact
membrane distillation coupled with solar ponds," Applied energy, vol. 158, pp. 532-539, 2015.
[17] L. Camacho, L. Dumée, J. Zhang, J.-d. Li, M. Duke, J. Gomez, et al., "Advances in Membrane
Distillation for Water Desalination and Purification Applications," Water, vol. 5, p. 94, 2013.
[18] E. Drioli, A. Ali, and F. Macedonio, "Membrane distillation: Recent developments and perspectives,"
Desalination, vol. 356, pp. 56-84, 1/15/ 2015.
[19] R. B. Saffarini, B. Mansoor, R. Thomas, and H. A. Arafat, "Effect of temperature-dependent
microstructure evolution on pore wetting in PTFE membranes under membrane distillation
conditions," Journal of membrane science, vol. 429, pp. 282-294, 2013.
[20] K. J. Kim, A. G. Fane, C. J. D. Fell, T. Suzuki, and M. R. Dickson, "Quantitative microscopic study
of surface characteristics of ultrafiltration membranes," Journal of Membrane Science, vol. 54, pp.
89-102, 1990/11/15/ 1990.
66

[21] E. Guillen-Burrieza, R. Thomas, B. Mansoor, D. Johnson, N. Hilal, and H. Arafat, "Effect of dry-out
on the fouling of PVDF and PTFE membranes under conditions simulating intermittent seawater
membrane distillation (SWMD)," Journal of Membrane Science, vol. 438, pp. 126-139, 7/1/ 2013.
[22] D. Quéré, "Wetting and roughness," Annu. Rev. Mater. Res., vol. 38, pp. 71-99, 2008.
[23] G. Rao, S. R. Hiibel, A. Achilli, and A. E. Childress, "Factors contributing to flux improvement in
vacuum-enhanced direct contact membrane distillation," Desalination, vol. 367, pp. 197-205, 7/1/
2015.
[24] C. R. Martinetti, A. E. Childress, and T. Y. Cath, "High recovery of concentrated RO brines using
forward osmosis and membrane distillation," Journal of Membrane Science, vol. 331, pp. 31-39,
2009/04/01/ 2009.
[25] T. Y. Cath, D. Adams, and 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, vol. 257, pp. 111-119,
2005/07/15/ 2005.
[26] T. Y. Cath, S. Gormly, E. G. Beaudry, M. T. Flynn, V. D. Adams, and A. E. Childress, "Membrane
contactor processes for wastewater reclamation in space: Part I. Direct osmotic concentration as
pretreatment for reverse osmosis," Journal of Membrane Science, vol. 257, pp. 85-98, 7/15/ 2005.
[27] D. L. Shaffer, M. E. Tousley, and M. Elimelech, "Influence of polyamide membrane surface
chemistry on gypsum scaling behavior," Journal of Membrane Science, vol. 525, pp. 249-256,
2017.
[28] G. Rao, S. R. Hiibel, and A. E. Childress, "Simplified flux prediction in direct-contact membrane
distillation using a membrane structural parameter," Desalination, vol. 351, pp. 151-162, 10/15/
2014.
[29] T. Y. Cath, A. E. Childress, and C. R. Martinetti, "Combined membrane-distillation-forward-osmosis
systems and methods of use," ed: Google Patents, 2011.
[30] K. Schneider, W. Hölz, R. Wollbeck, and S. Ripperger, "Membranes and modules for
transmembrane distillation," Journal of Membrane Science, vol. 39, pp. 25-42, 1988/10/01 1988.
[31] F. A. Banat and J. Simandl, "Theoretical and experimental study in membrane distillation,"
Desalination, vol. 95, pp. 39-52, 1994.
[32] K. Karakulski, M. Gryta, and A. Morawski, "Membrane processes used for potable water quality
improvement," Desalination, vol. 145, pp. 315-319, 2002.
[33] C. Feng, K. C. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna, et al., "Production of
drinking water from saline water by air-gap membrane distillation using polyvinylidene fluoride
nanofiber membrane," Journal of Membrane Science, vol. 311, pp. 1-6, 3/20/ 2008.
[34] M. Gryta, "Long-term performance of membrane distillation process," Journal of Membrane
Science, vol. 265, pp. 153-159, 2005.
[35] M. Gryta, "Influence of polypropylene membrane surface porosity on the performance of membrane
distillation process," Journal of Membrane Science, vol. 287, pp. 67-78, 2007.
[36] M. Gryta and M. Barancewicz, "Influence of morphology of PVDF capillary membranes on the
performance of direct contact membrane distillation," Journal of Membrane Science, vol. 358, pp.
158-167, 2010.
[37] M. Gryta, "Wettability of polypropylene capillary membranes during the membrane distillation
process," Chemical Papers, vol. 66, pp. 92-98, 2012.
[38] M. M. A. Shirazi, A. Kargari, and M. J. A. Shirazi, "Direct contact membrane distillation for seawater
desalination," Desalination and Water Treatment, vol. 49, pp. 368-375, 2012/11/01 2012.
[39] M. El Fray and M. Gryta, "Environmental fracture of polypropylene membranes used in membrane
distillation process," Polimery, vol. 53, pp. 865-870, 2008.
[40] E. Guillén-Burrieza, J. Blanco, G. Zaragoza, D.-C. Alarcón, P. Palenzuela, M. Ibarra, et al.,
"Experimental analysis of an air gap membrane distillation solar desalination pilot system," Journal
of Membrane Science, vol. 379, pp. 386-396, 9/1/ 2011.
[41] E. Guillen-Burrieza, A. Ruiz-Aguirre, G. Zaragoza, and H. A. Arafat, "Membrane fouling and
cleaning in long term plant-scale membrane distillation operations," Journal of Membrane Science,
vol. 468, pp. 360-372, 10/15/ 2014.
[42] M. M. Teoh, T.-S. Chung, and Y. S. Yeo, "Dual-layer PVDF/PTFE composite hollow fibers with a
thin macrovoid-free selective layer for water production via membrane distillation," Chemical
Engineering Journal, vol. 171, pp. 684-691, 7/1/ 2011.
67

[43] P. Peng, A. G. Fane, and X. Li, "Desalination by membrane distillation adopting a hydrophilic
membrane," Desalination, vol. 173, pp. 45-54, 2005/03/01 2005.
[44] F. Edwie, M. M. Teoh, and T.-S. Chung, "Effects of additives on dual-layer hydrophobic–hydrophilic
PVDF hollow fiber membranes for membrane distillation and continuous performance," Chemical
Engineering Science, vol. 68, pp. 567-578, 2012/01/22/ 2012.
[45] M. R. Bilad, F. A. Al Marzooqi, and H. A. Arafat, "New Concept for Dual-Layer
Hydrophilic/Hydrophobic Composite Membrane for Membrane Distillation," Journal of Membrane
and Separation Technology, vol. 4, pp. 122-133, 2015.
[46] M. Khayet, J. I. Mengual, and T. Matsuura, "Porous hydrophobic/hydrophilic composite
membranes: Application in desalination using direct contact membrane distillation," Journal of
Membrane Science, vol. 252, pp. 101-113, 4/15/ 2005.
[47] C. Han, D. Tang, and D. Kim, "Molecular dynamics simulation on the effect of pore hydrophobicity
on water transport through aquaporin-mimic nanopores," Colloids and Surfaces A:
Physicochemical and Engineering Aspects, vol. 481, pp. 38-42, 2015/09/20/ 2015.
[48] S. Sriraman, I. G. Kevrekidis, and G. Hummer, "Coarse nonlinear dynamics and metastability of
filling-emptying transitions: water in carbon nanotubes," Physical review letters, vol. 95, p. 130603,
2005.
[49] Y.-I. Chou, Z.-Y. Siao, Y.-F. Chen, L.-Y. Sung, W.-M. Yang, and C.-C. Wang, "Water permeation
analysis on gas diffusion layers of proton exchange membrane fuel cells for Teflon-coating
annotation," Journal of Power Sources, vol. 195, pp. 536-540, 2010/01/15/ 2010.
[50] W. Song, H. Yu, L. Hao, Z. Miao, B. Yi, and Z. Shao, "A new hydrophobic thin film catalyst layer for
PEMFC," Solid State Ionics, vol. 181, pp. 453-458, 2010/03/29/ 2010.
[51] M. Gryta, "Effect of iron oxides scaling on the MD process performance," Desalination, vol. 216,
pp. 88-102, 10/5/ 2007.
[52] F. Mansfeld, G. Liu, H. Xiao, C. H. Tsai, and B. J. Little, "The corrosion behavior of copper alloys,
stainless steels and titanium in seawater," Corrosion Science, vol. 36, pp. 2063-2095, 1994/12/01/
1994.
[53] R. D. Gustafson, J. R. Murphy, and A. Achilli, "A stepwise model of direct contact membrane
distillation for application to large-scale systems: Experimental results and model predictions,"
Desalination, vol. 378, pp. 14-27, 2016/01/15/ 2016.
[54] G. Naidu, S. Jeong, and S. Vigneswaran, "Interaction of humic substances on fouling in membrane
distillation for seawater desalination," Chemical Engineering Journal, vol. 262, pp. 946-957,
2015/02/15/ 2015.
[55] L. D. Tijing, Y. C. Woo, J.-S. Choi, S. Lee, S.-H. Kim, and H. K. Shon, "Fouling and its control in
membrane distillation—A review," Journal of Membrane Science, vol. 475, pp. 215-244, 2/1/ 2015.
[56] M. Gryta, "Fouling in direct contact membrane distillation process," Journal of Membrane Science,
vol. 325, pp. 383-394, 11/15/ 2008.
[57] J. Gilron, Y. Ladizansky, and E. Korin, "Silica fouling in direct contact membrane distillation,"
Industrial & Engineering Chemistry Research, vol. 52, pp. 10521-10529, 2013.
[58] A. Barbe, P. Hogan, and R. Johnson, "Surface morphology changes during initial usage of
hydrophobic, microporous polypropylene membranes," Journal of Membrane Science, vol. 172, pp.
149-156, 2000.
[59] "National Institute of Standards and Technology Chemisty WebBook," NIST, Ed., 69 ed: U.S.
Secretary of Commerce, 2016.
[60] N. Hilal, A. F. Ismail, T. Matsuura, and D. Oatley-Radcliffe, Membrane Characterization: Elsevier,
2017.
[61] M. Sillanpää, "Chapter 1 - General Introduction," in Natural Organic Matter in Water, ed:
Butterworth-Heinemann, 2015, pp. 1-15.
[62] G. Naidu, S. Jeong, S.-J. Kim, I. S. Kim, and S. Vigneswaran, "Organic fouling behavior in direct
contact membrane distillation," Desalination, vol. 347, pp. 230-239, 8/15/ 2014.
[63] M. Gryta, J. Grzechulska-Damszel, A. Markowska, and K. Karakulski, "The influence of
polypropylene degradation on the membrane wettability during membrane distillation," Journal of
Membrane Science, vol. 326, pp. 493-502, 2009.
[64] D. M. Warsinger, J. Swaminathan, E. Guillen-Burrieza, H. A. Arafat, and J. H. Lienhard V, "Scaling
and fouling in membrane distillation for desalination applications: A review," Desalination, vol. 356,
pp. 294-313, 1/15/ 2015.
68

[65] E. Guillen-Burrieza, M. O. Mavukkandy, M. R. Bilad, and H. A. Arafat, "Understanding wetting
phenomena in membrane distillation and how operational parameters can affect it," Journal of
Membrane Science, vol. 515, pp. 163-174, 10/1/ 2016.
[66] Y. Liao, R. Wang, and A. G. Fane, "Engineering superhydrophobic surface on poly(vinylidene
fluoride) nanofiber membranes for direct contact membrane distillation," Journal of Membrane
Science, vol. 440, pp. 77-87, 8/1/ 2013.
[67] K.-J. Lu, J. Zuo, and T.-S. Chung, "Tri-bore PVDF hollow fibers with a super-hydrophobic coating
for membrane distillation," Journal of Membrane Science, vol. 514, pp. 165-175, 2016.
[68] J. Zhang, Z. Song, B. Li, Q. Wang, and S. Wang, "Fabrication and characterization of
superhydrophobic poly (vinylidene fluoride) membrane for direct contact membrane distillation,"
Desalination, vol. 324, pp. 1-9, 2013.
[69] M. Shakaib, S. M. F. Hasani, I. Ahmed, and R. M. Yunus, "A CFD study on the effect of spacer
orientation on temperature polarization in membrane distillation modules," Desalination, vol. 284,
pp. 332-340, 1/4/ 2012.
[70] X. Li, X. Yu, C. Cheng, L. Deng, M. Wang, and X. Wang, "Electrospun superhydrophobic
organic/inorganic composite nanofibrous membranes for membrane distillation," ACS applied
materials & interfaces, vol. 7, pp. 21919-21930, 2015.
[71] A. Sadeghi Alavijeh, R. M. H. Khorasany, A. Habisch, G. G. Wang, and E. Kjeang, "Creep
properties of catalyst coated membranes for polymer electrolyte fuel cells," Journal of Power
Sources, vol. 285, pp. 16-28, 7/1/ 2015.
[72] E. Chabanon, D. Mangin, and C. Charcosset, "Membranes and crystallization processes: State of
the art and prospects," Journal of Membrane Science, vol. 509, pp. 57-67, 2016.
[73] H. Maab, A. Al Saadi, L. Francis, S. Livazovic, N. Ghafour, G. L. Amy, et al., "Polyazole hollow fiber
membranes for direct contact membrane distillation," Industrial & Engineering Chemistry Research,
vol. 52, pp. 10425-10429, 2013.
[74] B. Li and K. K. Sirkar, "Novel membrane and device for direct contact membrane distillation-based
desalination process," Industrial & engineering chemistry research, vol. 43, pp. 5300-5309, 2004.
[75] A. E. Jansen, J. W. Assink, J. H. Hanemaaijer, J. van Medevoort, and E. van Sonsbeek,
"Development and pilot testing of full-scale membrane distillation modules for deployment of waste
heat," Desalination, vol. 323, pp. 55-65, 8/15/ 2013.
[76] A. K. Pabby and A. M. Sastre, "State-of-the-art review on hollow fibre contactor technology and
membrane-based extraction processes," Journal of Membrane Science, vol. 430, pp. 263-303,
2013/03/01/ 2013.
[77] A. L. McGaughey, P. Karandikar, M. Gupta, and A. E. Childress, "Hydrophobicity versus Pore Size:
Polymer Coatings to Improve Membrane Wetting Resistance for Membrane Distillation," ACS
Applied Polymer Materials, vol. 2, pp. 1256-1267, 2020.
[78] C. Li, X. Li, X. Du, Y. Zhang, W. Wang, T. Tong, et al., "Elucidating the trade-off between membrane
wetting resistance and water vapor flux in membrane distillation," Environmental Science &
Technology, vol. 54, pp. 10333-10341, 2020.
[79] W. Wang, X. Du, H. Vahabi, S. Zhao, Y. Yin, A. K. Kota, et al., "Trade-off in membrane distillation
with monolithic omniphobic membranes," Nature communications, vol. 10, pp. 1-9, 2019.
[80] B. Van der Bruggen, C. Vandecasteele, T. Van Gestel, W. Doyen, and R. Leysen, "A review of
pressure‐driven membrane processes in wastewater treatment and drinking water production,"
Environmental progress, vol. 22, pp. 46-56, 2003.
[81] S. Mosadegh-Sedghi, D. Rodrigue, J. Brisson, and M. C. Iliuta, "Wetting phenomenon in membrane
contactors–causes and prevention," Journal of Membrane Science, vol. 452, pp. 332-353, 2014.
[82] D. Hou, D. Jassby, R. Nerenberg, and Z. J. Ren, "Hydrophobic Gas Transfer Membranes for
Wastewater Treatment and Resource Recovery," Environmental science & technology, vol. 53, pp.
11618-11635, 2019.
[83] L. M. Camacho, L. Dumée, J. Zhang, J.-d. Li, M. Duke, J. Gomez, et al., "Advances in membrane
distillation for water desalination and purification applications," Water, vol. 5, pp. 94-196, 2013.
[84] R. D. Gustafson, A. L. McGaughey, W. Ding, S. C. McVety, and A. E. Childress, "Morphological
changes and creep recovery behavior of expanded polytetrafluoroethylene (ePTFE) membranes
used for membrane distillation," Journal of Membrane Science, vol. 584, pp. 236-245, 2019.
[85] B.-S. Kim and P. Harriott, "Critical entry pressure for liquids in hydrophobic membranes," Journal
of Colloid and Interface Science, vol. 115, pp. 1-8, 1987/01/01/ 1987.
69

[86] E. W. Washburn, "The dynamics of capillary flow," Physical review, vol. 17, p. 273, 1921.
[87] E. Guillen-Burrieza, A. Servi, B. S. Lalia, and H. A. Arafat, "Membrane structure and surface
morphology impact on the wetting of MD membranes," Journal of Membrane Science, vol. 483, pp.
94-103, 2015.
[88] Z. Wang, Y. Chen, X. Sun, R. Duddu, and S. Lin, "Mechanism of pore wetting in membrane
distillation with alcohol vs. surfactant," Journal of Membrane Science, vol. 559, pp. 183-195,
2018/08/01/ 2018.
[89] Z. Wang, Y. Chen, and S. Lin, "Kinetic model for surfactant-induced pore wetting in membrane
distillation," Journal of Membrane Science, vol. 564, pp. 275-288, 2018/10/15/ 2018.
[90] G. Rácz, S. Kerker, Z. Kovács, G. Vatai, M. Ebrahimi, and P. Czermak, "Theoretical and
experimental approaches of liquid entry pressure determination in membrane distillation
processes," Periodica Polytechnica Chemical Engineering, vol. 58, pp. 81-91, 2014.
[91] F. Zha, A. Fane, C. Fell, and R. Schofield, "Critical displacement pressure of a supported liquid
membrane," Journal of membrane science, vol. 75, pp. 69-80, 1992.
[92] A. T. Servi, E. Guillen-Burrieza, D. M. Warsinger, W. Livernois, K. Notarangelo, J. Kharraz, et al.,
"The effects of iCVD film thickness and conformality on the permeability and wetting of MD
membranes," Journal of Membrane Science, vol. 523, pp. 470-479, 2017/02/01/ 2017.
[93] M. d. C. García-Payo, M. A. Izquierdo-Gil, and C. Fernández-Pineda, "Wetting study of hydrophobic
membranes via liquid entry pressure measurements with aqueous alcohol solutions," Journal of
colloid and interface science, vol. 230, pp. 420-431, 2000.
[94] G. Rácz, S. Kerker, Z. Kovács, G. Vatai, M. Ebrahimi, and P. Czermak, "Theoretical and
experimental approaches of liquid entry pressure determination in membrane distillation
processes," Periodica Polytechnica. Chemical Engineering, vol. 58, p. 81, 2014.
[95] C. R. Taylor, P. Ahmadiannamini, and S. R. Hiibel, "Identifying pore wetting thresholds of
surfactants in direct contact membrane distillation," Separation and Purification Technology, vol.
217, pp. 17-23, 2019/06/15/ 2019.
[96] A. Lafuma and D. Quéré, "Superhydrophobic states," Nature materials, vol. 2, pp. 457-460, 2003.
[97] G. McHale, "Cassie and Wenzel: were they really so wrong?," Langmuir, vol. 23, pp. 8200-8205,
2007.
[98] J. C. Berg, An introduction to interfaces & colloids: the bridge to nanoscience: World Scientific,
2010.
[99] D. Johnson and N. Hilal, "Polymer membranes–Fractal characteristics and determination of
roughness scaling exponents," Journal of Membrane Science, vol. 570, pp. 9-22, 2019.
[100] S. Feng, G. Yu, X. Cai, M. Eulade, H. Lin, J. Chen, et al., "Effects of fractal roughness of membrane
surfaces on interfacial interactions associated with membrane fouling in a membrane bioreactor,"
Bioresource technology, vol. 244, pp. 560-568, 2017.
[101] P. C. Y. Wong, Y.-N. Kwon, and C. S. Criddle, "Use of atomic force microscopy and fractal geometry
to characterize the roughness of nano-, micro-, and ultrafiltration membranes," Journal of
Membrane Science, vol. 340, pp. 117-132, 2009.
[102] M. Zhang, J. Chen, Y. Ma, L. Shen, Y. He, and H. Lin, "Fractal reconstruction of rough membrane
surface related with membrane fouling in a membrane bioreactor," Bioresource technology, vol.
216, pp. 817-823, 2016.
[103] Y. Yuan and T. R. Lee, "Contact angle and wetting properties," in Surface science techniques, ed:
Springer, 2013, pp. 3-34.
[104] V. Gurau, M. J. Bluemle, E. S. De Castro, Y.-M. Tsou, J. A. Mann, and T. A. Zawodzinski,
"Characterization of transport properties in gas diffusion layers for proton exchange membrane fuel
cells: 1. Wettability (internal contact angle to water and surface energy of GDL fibers)," Journal of
Power Sources, vol. 160, pp. 1156-1162, 2006/10/06/ 2006.
[105] M. Morra, E. Occhiello, and F. Garbassi, "Surface characterization of plasma‐treated PTFE,"
Surface and Interface analysis, vol. 16, pp. 412-417, 1990.
[106] T. Urai, M. Kamai, and H. Fujii, "Estimation of intrinsic contact angle of various liquids on PTFE by
utilizing ultrasonic vibration," Journal of Materials Engineering and Performance, vol. 25, pp. 3384-
3389, 2016.
[107] K. Grundke, T. Bogumil, T. Gietzelt, H.-J. Jacobasch, D. Kwok, and A. Neumann, "Wetting
measurements on smooth, rough and porous solid surfaces," in Interfaces, Surfactants and
Colloids in Engineering, ed: Springer, 1996, pp. 58-68.
70

[108] J. Włoch, A. P. Terzyk, M. Wiś niewski, and P. Kowalczyk, "Nanoscale water contact angle on
Polytetrafluoroethylene surfaces characterized by molecular Dynamics–Atomic force microscopy
imaging," Langmuir, vol. 34, pp. 4526-4534, 2018.
[109] A. McGaughey, R. Gustafson, and A. Childress, "Effect of long-term operation on membrane
surface characteristics and performance in membrane distillation," Journal of Membrane Science,
vol. 543, pp. 143-150, 2017.
[110] F. F. Zha, A. G. Fane, C. J. D. Fell, and R. W. Schofield, "Critical displacement pressure of a
supported liquid membrane," Journal of Membrane Science, vol. 75, pp. 69-80, 1992/12/16/ 1992.
[111] M. Qtaishat and T. Matsuura, "Modelling of pore wetting in membrane distillation compared with
pervaporation," in Pervaporation, Vapour Permeation and Membrane Distillation, ed: Elsevier,
2015, pp. 385-413.
[112] E. Guillen-Burrieza, M. Mavukkandy, M. Bilad, and H. Arafat, "Understanding wetting phenomena
in membrane distillation and how operational parameters can affect it," Journal of Membrane
Science, vol. 515, pp. 163-174, 2016.
[113] Y. Wu, Y. Kong, X. Lin, W. Liu, and J. Xu, "Surface-modified hydrophilic membranes in membrane
distillation," Journal of Membrane Science, vol. 72, pp. 189-196, 1992/09/04/ 1992.
[114] M. Khayet, T. Matsuura, and J. I. Mengual, "Porous hydrophobic/hydrophilic composite
membranes: Estimation of the hydrophobic-layer thickness," Journal of Membrane Science, vol.
266, pp. 68-79, 2005/12/01/ 2005.
[115] M. Khayet, J. I. Mengual, and T. Matsuura, "Porous hydrophobic/hydrophilic composite
membranes," Journal of Membrane Science, vol. 252, pp. 101-113, 2005/04/15/ 2005.
[116] M. Khayet, T. Matsuura, J. I. Mengual, and M. Qtaishat, "Design of novel direct contact membrane
distillation membranes," Desalination, vol. 192, pp. 105-111, 2006/05/10/ 2006.
[117] S. Bonyadi and T. S. Chung, "Flux enhancement in membrane distillation by fabrication of dual
layer hydrophilic–hydrophobic hollow fiber membranes," Journal of Membrane Science, vol. 306,
pp. 134-146, 2007/12/01/ 2007.
[118] M. Qtaishat, M. Khayet, and T. Matsuura, "Novel porous composite hydrophobic/hydrophilic
polysulfone membranes for desalination by direct contact membrane distillation," Journal of
Membrane Science, vol. 341, pp. 139-148, 2009/09/30/ 2009.
[119] M. Qtaishat, M. Khayet, and T. Matsuura, "Guidelines for preparation of higher flux
hydrophobic/hydrophilic composite membranes for membrane distillation," Journal of Membrane
Science, vol. 329, pp. 193-200, 2009/03/05/ 2009.
[120] M. Qtaishat, D. Rana, M. Khayet, and T. Matsuura, "Preparation and characterization of novel
hydrophobic/hydrophilic polyetherimide composite membranes for desalination by direct contact
membrane distillation," Journal of Membrane Science, vol. 327, pp. 264-273, 2009/02/05/ 2009.
[121] M. Qtaishat, D. Rana, T. Matsuura, and M. Khayet, "Effect of surface modifying macromolecules
stoichiometric ratio on composite hydrophobic/hydrophilic membranes characteristics and
performance in direct contact membrane distillation," AIChE journal, vol. 55, pp. 3145-3151, 2009.
[122] M. Su, M. M. Teoh, K. Y. Wang, J. Su, and T.-S. Chung, "Effect of inner-layer thermal conductivity
on flux enhancement of dual-layer hollow fiber membranes in direct contact membrane distillation,"
Journal of Membrane Science, vol. 364, pp. 278-289, 2010/11/15/ 2010.
[123] M. Essalhi and M. Khayet, "Surface segregation of fluorinated modifying macromolecule for
hydrophobic/hydrophilic membrane preparation and application in air gap and direct contact
membrane distillation," Journal of Membrane Science, vol. 417, pp. 163-173, 2012/11/01/ 2012.
[124] M. Essalhi and M. Khayet, "Application of a porous composite hydrophobic/hydrophilic membrane
in desalination by air gap and liquid gap membrane distillation: A comparative study," Separation
and Purification Technology, vol. 133, pp. 176-186, 2014.
[125] J.-G. Lee, Y.-D. Kim, W.-S. Kim, L. Francis, G. Amy, and N. Ghaffour, "Performance modeling of
direct contact membrane distillation (DCMD) seawater desalination process using a commercial
composite membrane," Journal of Membrane Science, vol. 478, pp. 85-95, 2015/03/15/ 2015.
[126] J. Zhu, L. Jiang, and T. Matsuura, "New insights into fabrication of hydrophobic/hydrophilic
composite hollow fibers for direct contact membrane distillation," Chemical Engineering Science,
vol. 137, pp. 79-90, 2015.
[127] S. Ragunath, S. Roy, and S. Mitra, "Selective hydrophilization of the permeate surface to enhance
flux in membrane distillation," Separation and Purification Technology, vol. 170, pp. 427-433, 10/1/
2016.
71

[128] X. Feng, L. Y. Jiang, T. Matsuura, and P. Wu, "Fabrication of hydrophobic/hydrophilic composite
hollow fibers for DCMD: Influence of dope formulation and external coagulant," Desalination, vol.
401, pp. 53-63, 1/2/ 2017.
[129] A. Figoli, C. Ursino, F. Galiano, E. Di Nicolò, P. Campanelli, M. C. Carnevale, et al., "Innovative
hydrophobic coating of perfluoropolyether (PFPE) on commercial hydrophilic membranes for
DCMD application," Journal of Membrane Science, vol. 522, pp. 192-201, 1/15/ 2017.
[130] Y. C. Woo, L. D. Tijing, M. J. Park, M. Yao, J.-S. Choi, S. Lee, et al., "Electrospun dual-layer
nonwoven membrane for desalination by air gap membrane distillation," Desalination, vol. 403, pp.
187-198, 2/1/ 2017.
[131] J. Zuo, T.-S. Chung, G. S. O’Brien, and W. Kosar, "Hydrophobic/hydrophilic PVDF/Ultem® dual-
layer hollow fiber membranes with enhanced mechanical properties for vacuum membrane
distillation," Journal of Membrane Science, vol. 523, pp. 103-110, 2/1/ 2017.
[132] P. Peng, A. Fane, and X. Li, "Desalination by membrane distillation adopting a hydrophilic
membrane," Desalination, vol. 173, pp. 45-54, 2005.
[133] K. A. Salls, D. Won, E. P. Kolodziej, A. E. Childress, and S. R. Hiibel, "Evaluation of semi-volatile
contaminant transport in a novel, gas-tight direct contact membrane distillation system,"
Desalination, vol. 427, pp. 35-41, 2018.
[134] C. P. Morrow, N. M. Furtaw, J. R. Murphy, A. Achilli, E. A. Marchand, S. R. Hiibel, et al., "Integrating
an aerobic/anoxic osmotic membrane bioreactor with membrane distillation for potable reuse,"
Desalination, vol. 432, pp. 46-54, 2018/04/15/ 2018.
[135] R. D. Gustafson, S. R. Hiibel, and A. E. Childress, "Membrane distillation driven by intermittent and
variable-temperature waste heat: System arrangements for water production and heat storage,"
Desalination, vol. 448, pp. 49-59, 2018.
[136] C. Su, T. Horseman, H. Cao, K. Christie, Y. Li, and S. Lin, "Robust superhydrophobic membrane
for membrane distillation with excellent scaling resistance," Environmental science & technology,
vol. 53, pp. 11801-11809, 2019.
[137] V. Karanikola, C. Boo, J. Rolf, and M. Elimelech, "Engineered Slippery Surface to Mitigate Gypsum
Scaling in Membrane Distillation for Treatment of Hypersaline Industrial Wastewaters,"
Environmental science & technology, vol. 52, pp. 14362-14370, 2018.
[138] Z. Wang and S. Lin, "Membrane fouling and wetting in membrane distillation and their mitigation
by novel membranes with special wettability," Water Research, vol. 112, pp. 38-47, 2017/04/01/
2017.
[139] C. Boo, J. Lee, and M. Elimelech, "Omniphobic Polyvinylidene Fluoride (PVDF) Membrane for
Desalination of Shale Gas Produced Water by Membrane Distillation," Environmental Science &
Technology, vol. 50, pp. 12275-12282, 2016/11/15 2016.
[140] A. T. Servi, E. Guillen-Burrieza, D. M. Warsinger, W. Livernois, K. Notarangelo, J. Kharraz, et al.,
"The effects of iCVD film thickness and conformality on the permeability and wetting of MD
membranes," Journal of Membrane Science, vol. 523, pp. 470-479, Feb 1 2017.
[141] A. T. Servi, J. Kharraz, D. Klee, K. Notarangelo, B. Eyob, E. Guillen-Burrieza, et al., "A systematic
study of the impact of hydrophobicity on the wetting of MD membranes," Journal of Membrane
Science, vol. 520, pp. 850-859, 12/15/ 2016.
[142] D. M. Warsinger, A. Servi, S. Van Belleghem, J. Gonzalez, J. Swaminathan, J. Kharraz, et al.,
"Combining air recharging and membrane superhydrophobicity for fouling prevention in membrane
distillation," Journal of Membrane Science, vol. 505, pp. 241-252, 2016.
[143] M. Gupta and K. K. Gleason, "Initiated chemical vapor deposition of poly (1H, 1H, 2H, 2H-
perfluorodecyl acrylate) thin films," Langmuir, vol. 22, pp. 10047-10052, 2006.
[144] H. Y. Erbil, A. L. Demirel, Y. Avcı, and O. Mert, "Transformation of a simple plastic into a
superhydrophobic surface," Science, vol. 299, pp. 1377-1380, 2003.
[145] I. Sadeghi, N. Govinna, P. Cebe, and A. Asatekin, "Superoleophilic, Mechanically Strong
Electrospun Membranes for Fast and Efficient Gravity-Driven Oil/Water Separation," Acs Applied
Polymer Materials, vol. 1, pp. 765-776, Apr 2019.
[146] Y. Tian and L. Jiang, "Wetting: Intrinsically robust hydrophobicity," Nature materials, vol. 12, p. 291,
2013.
[147] S. Munirasu, F. Banat, A. A. Durrani, and M. A. Haija, "Intrinsically superhydrophobic PVDF
membrane by phase inversion for membrane distillation," Desalination, vol. 417, pp. 77-86, 2017.
72

[148] Z. Wang, Y. Tang, and B. Li, "Excellent wetting resistance and anti-fouling performance of PVDF
membrane modified with superhydrophobic papillae-like surfaces," Journal of Membrane Science,
vol. 540, pp. 401-410, 2017/10/15/ 2017.
[149] X. Li, C. Wang, Y. Yang, X. Wang, M. Zhu, and B. S. Hsiao, "Dual-biomimetic superhydrophobic
electrospun polystyrene nanofibrous membranes for membrane distillation," ACS applied materials
& interfaces, vol. 6, pp. 2423-2430, 2014.
[150] Y. Liao, C.-H. Loh, R. Wang, and A. G. Fane, "Electrospun superhydrophobic membranes with
unique structures for membrane distillation," ACS applied materials & interfaces, vol. 6, pp. 16035-
16048, 2014.
[151] F. Guo, A. Servi, A. Liu, K. K. Gleason, and G. C. Rutledge, "Desalination by membrane distillation
using electrospun polyamide fiber membranes with surface fluorination by chemical vapor
deposition," ACS applied materials & interfaces, vol. 7, pp. 8225-8232, 2015.
[152] Z.-Q. Dong, B.-J. Wang, X.-h. Ma, Y.-M. Wei, and Z.-L. Xu, "FAS grafted electrospun poly (vinyl
alcohol) nanofiber membranes with robust superhydrophobicity for membrane distillation," ACS
applied materials & interfaces, vol. 7, pp. 22652-22659, 2015.
[153] Z.-Q. Dong, X.-H. Ma, Z.-L. Xu, and Z.-Y. Gu, "Superhydrophobic modification of PVDF–SiO 2
electrospun nanofiber membranes for vacuum membrane distillation," RSC Advances, vol. 5, pp.
67962-67970, 2015.
[154] N. Hamzah and C. Leo, "Fouling prevention in the membrane distillation of phenolic-rich solution
using superhydrophobic PVDF membrane incorporated with TiO2 nanoparticles," Separation and
Purification Technology, vol. 167, pp. 79-87, 2016.
[155] Z. Zhu, Y. Liu, H. Hou, W. Shi, F. Qu, F. Cui, et al., "Dual-bioinspired design for constructing
membranes with superhydrophobicity for direct contact membrane distillation," Environmental
science & technology, vol. 52, pp. 3027-3036, 2018.
[156] S. Cong and F. Guo, "Janus Nanofibrous Membranes for Desalination by Air Gap Membrane
Distillation," Acs Applied Polymer Materials, vol. 1, pp. 3443-3451, Dec 2019.
[157] S. Meng, Y. Ye, J. Mansouri, and V. Chen, "Fouling and crystallisation behaviour of
superhydrophobic nano-composite PVDF membranes in direct contact membrane distillation,"
Journal of Membrane Science, vol. 463, pp. 102-112, 2014/08/01/ 2014.
[158] S. Meng, Y. Ye, J. Mansouri, and V. Chen, "Crystallization behavior of salts during membrane
distillation with hydrophobic and superhydrophobic capillary membranes," Journal of Membrane
Science, vol. 473, pp. 165-176, 1/1/ 2015.
[159] Z. Xiao, R. Zheng, Y. Liu, H. He, X. Yuan, Y. Ji, et al., "Slippery for scaling resistance in membrane
distillation: a novel porous micropillared superhydrophobic surface," Water Research, vol. 155, pp.
152-161, 2019/02/01/ 2019.
[160] Z. Xiao, Z. Li, H. Guo, Y. Liu, Y. Wang, H. Yin, et al., "Scaling mitigation in membrane distillation:
From superhydrophobic to slippery," Desalination, vol. 466, pp. 36-43, 2019/09/15/ 2019.
[161] E. Guillen-Burrieza, A. Servi, B. S. Lalia, and H. A. Arafat, "Membrane structure and surface
morphology impact on the wetting of MD membranes," Journal of Membrane Science, vol. 483, pp.
94-103, 2015/06/01/ 2015.
[162] J. W. Mullin, Crystallization: Elsevier, 2001.
[163] E. Curcio, E. Fontananova, G. Di Profio, and E. Drioli, "Influence of the structural properties of poly
(vinylidene fluoride) membranes on the heterogeneous nucleation rate of protein crystals," The
Journal of Physical Chemistry B, vol. 110, pp. 12438-12445, 2006.
[164] T. Horseman, C. Su, K. S. Christie, and S. Lin, "Highly Effective Scaling Mitigation in Membrane
Distillation Using a Superhydrophobic Membrane with Gas Purging," Environmental Science &
Technology Letters, vol. 6, pp. 423-429, 2019.
[165] A. E. Childress and M. Elimelech, "Effect of solution chemistry on the surface charge of polymeric
reverse osmosis and nanofiltration membranes," Journal of membrane science, vol. 119, pp. 253-
268, 1996.
[166] M. Rezaei, D. M. Warsinger, J. H. Lienhard V, and W. M. Samhaber, "Wetting prevention in
membrane distillation through superhydrophobicity and recharging an air layer on the membrane
surface," Journal of Membrane Science, vol. 530, pp. 42-52, 2017/05/15/ 2017.
[167] N. G. P. Chew, S. Zhao, C. Malde, and R. Wang, "Superoleophobic surface modification for robust
membrane distillation performance," Journal of Membrane Science, vol. 541, pp. 162-173,
2017/11/01/ 2017.
73

[168] Y.-X. Huang, Z. Wang, J. Jin, and S. Lin, "Novel Janus membrane for membrane distillation with
simultaneous fouling and wetting resistance," Environmental science & technology, vol. 51, pp.
13304-13310, 2017.
[169] X. Lu, Y. Peng, L. Ge, R. Lin, Z. Zhu, and S. Liu, "Amphiphobic PVDF composite membranes for
anti-fouling direct contact membrane distillation," Journal of Membrane Science, vol. 505, pp. 61-
69, 2016/05/01/ 2016.
[170] H. B. Park, J. Kamcev, L. M. Robeson, M. Elimelech, and B. D. Freeman, "Maximizing the right
stuff: The trade-off between membrane permeability and selectivity," Science, vol. 356, p.
eaab0530, 2017.
[171] OECD, "Preliminary Lists of PFOS, PFAS, PFOA and Related Compounds and Chemicals that
May Degrade to PFCA," OECD Papers, vol. 6/11, 2006.
[172] A. Asatekin and K. K. Gleason, "Polymeric nanopore membranes for hydrophobicity-based
separations by conformal initiated chemical vapor deposition," Nano letters, vol. 11, pp. 677-686,
2010.
[173] S. S. Latthe, A. B. Gurav, C. S. Maruti, and R. S. Vhatkar, "Recent progress in preparation of
superhydrophobic surfaces: a review," Journal of Surface Engineered Materials and Advanced
Technology, vol. 2, p. 76, 2012.
[174] C. Cheng and M. Gupta, "Roll-to-Roll Surface Modification of Cellulose Paper via Initiated Chemical
Vapor Deposition," Industrial & Engineering Chemistry Research, vol. 57, pp. 11675-11680, 2018.
[175] C. T. Riche, B. C. Marin, N. Malmstadt, and M. Gupta, "Vapor deposition of cross-linked
fluoropolymer barrier coatings onto pre-assembled microfluidic devices," Lab on a Chip, vol. 11,
pp. 3049-3052, 2011.
[176] I. Koyuncu, J. Brant, A. Lüttge, and M. R. Wiesner, "A comparison of vertical scanning
interferometry (VSI) and atomic force microscopy (AFM) for characterizing membrane surface
topography," Journal of Membrane Science, vol. 278, pp. 410-417, 2006/07/05/ 2006.
[177] R. Good, M. Chaudhury, and C. Van Oss, "Fundamentals of adhesion," LH Lee (Ed.), p. 153, 1991.
[178] B. Jańczuk and T. Białopiotrowicz, "The total surface free energy and the contact angle in the case
of low energetic solids," Journal of colloid and interface science, vol. 140, pp. 362-372, 1990.
[179] P.-G. De Gennes, F. Brochard-Wyart, and D. Quéré, Capillarity and wetting phenomena: drops,
bubbles, pearls, waves: Springer Science & Business Media, 2004.
[180] D. L. Mejia Mendez, C. Castel, C. Lemaitre, and E. Favre, "Membrane distillation (MD) processes
for water desalination applications. Can dense selfstanding membranes compete with microporous
hydrophobic materials?," Chemical Engineering Science, vol. 188, pp. 84-96, 2018/10/12/ 2018.
[181] E. Shaulsky, V. Karanikola, A. P. Straub, A. Deshmukh, I. Zucker, and M. Elimelech, "Asymmetric
membranes for membrane distillation and thermo-osmotic energy conversion," Desalination, vol.
452, pp. 141-148, 2019/02/15/ 2019.
[182] S. Nejati, C. Boo, C. O. Osuji, and M. Elimelech, "Engineering flat sheet microporous PVDF films
for membrane distillation," Journal of Membrane Science, vol. 492, pp. 355-363, 2015/10/15/ 2015.

74

Appendix A: Supplementary Information
A.1 Supplementary Information to Chapter 3
A.1.1 Characterizing Wetting in MD
In bench-scale MD systems, pore wetting is often detected by an increase in the distillate conductivity and,
therefore, the distillate solute concentration (𝐶 𝑑 ). 𝐶 𝑑 is suitable for pore wetting detection because 𝐶 𝑑 can
be measured using an inline conductivity probe with higher sensitivity than by using a flowmeter or balance
to measure liquid flux (via flow rate or mass change of the distillate stream). Both flowmeters and balances
can be affected by fluctuations in pumping rate and pressure changes in the system, especially during start-
up of the bench-scale system.
However, 𝐶 𝑑 can also fluctuate due to factors other than pore wetting. To determine when wetting occurs
based on solute passage, increases in 𝐶 𝑑 due to pore wetting must be distinguished from other fluctuations.
In MD, we have observed that increases in 𝐶 𝑑 are generally characterized by one of three cases (Figure
A.1.1): (1) increasing 𝐶 𝑑 tapering towards a constant concentration (e.g., Figure A.1.1a and b), (2) linearly
increasing 𝐶 𝑑 (e.g., Figure A.1.1c and d), or (3) exponentially increasing 𝐶 𝑑 (e.g., Figure A.1.1e and f). While
Case (1) represents solute mixing, only cases (2) and (3) represent pore wetting.
In case 1, 𝐶 𝑑 increases at first and then approaches a constant concentration (Figure A.1.1a). Both 𝐽 𝑆 and
the slope of the concentration (Δ𝐶 𝑑 /Δ𝑡 ) decrease with fluctuations; Δ𝐶 𝑑 /Δ𝑡 decreases from 0.05 at t = 0 h
to 0.03 at t = 4.7 h and to zero after t = 13 h (Figure A.1.1b). The decrease in 𝐽 𝑆 and Δ𝐶 𝑑 /Δ𝑡 over time
indicates that increases in 𝐶 𝑑 are likely due to pumping of distillate solution through the system loop causing
mixing between the deionized water added into the system at the beginning of the experiment with stagnant,
higher-salinity liquid present in the distillate loop prior to system start-up. The slow rate of change in 𝐶 𝑑
suggests gradual mixing is occurring, eventually resulting in constant 𝐶 𝑑 after 13 h operation. Continuous
leakage, as would occur during pore wetting, was not observed, as evidenced by the decrease in 𝐽 𝑆 over
time. Therefore, case 1 is not considered to be pore wetting and is termed “solute mixing”.
In case 2, 𝐶 𝑑 increases linearly from time t = 0 h, with constant Δ𝐶 𝑑 /Δ𝑡 versus time and constant 𝐽 𝑆 over
time (Figure A.1.1b and e). Constant 𝐽 𝑆 over time indicates a constant wetting rate; in this case at a much
greater rate than observed in the solute mixing example shown in Case 1 (Figure A.1.1a and d). This is
indicative of a membrane with insufficient wetting resistance, resulting in instantaneous wetting and
constant liquid flux through membrane pores, pinholes, or areas of insufficient wetting resistance, and is
termed “constant wetting”. If allowed to operate for a longer duration, as distillate overflow is continuously
removed, the distillate solution would eventually be fully replaced by the liquid and vapor flux through the
membrane. At this point, 𝐶 𝑑 would approach a constant value equal to the concentration of the combined
liquid and condensed vapor fluxing through the membrane. This is readily distinguishable from Case 1
because constant 𝐶 𝑑 will be reached only when the water produced is at least equal to the total initial volume
of the distillate stream (i.e., 2.14 L). Also, wetting rates and 𝐽 𝑆 would typically expected to be much higher
75

for liquid flux of high-concentration feed solution through a membrane with insufficient wetting resistance
when compared to solute mixing in the distillate stream of deionized water with low-concentration stagnant
solution present in the distillate loop; this can be seen by comparing the magnitude of 𝐶 𝑑 , 𝐽 𝑆 , and operating
times between Figure A.1.1a and b.
In case 3, 𝐶 𝑑 is initially near 0 mM and then begins increasing exponentially after t = 1.3 h (Figure A.1.1c);
increasing Δ𝐶 𝑑 /Δ𝑡 and 𝐽 𝑆 can be seen in Figure A.1.1f and c, respectively. This is indicative of wetting of a
greater and greater number of pores, resulting in exponentially increasing liquid flux and 𝐽 𝑆 , and is termed
“progressive wetting”. In this case, wetting is driven by scaling of the membrane surface, which may have
resulted in wetting at a single scaling site followed by progressive wetting of neighboring pores and/or the
subsequent development of additional scaling sites that caused further pore wetting. If allowed to operate
over a longer duration, progressive wetting will theoretically result in complete wetting of all membrane
pores. At long operating times, the distillate solution would eventually be replaced by liquid flux through the
wetted membrane, resulting in a constant value of 𝐶 𝑑 equal to the feed stream concentration. This is readily
distinguishable from Case 1 because constant 𝐶 𝑑 will be reached only when the water produced is at least
equal to the total initial volume of the distillate stream (i.e., 2.14 L); also, the magnitude of 𝐶 𝑑 reached would
typically greatly exceed the maximum value measurable by the distillate conductivity probe (i.e., 1.8 mM
NaCl). Thus, in characterizing wetting during the MD process, it is most informative to examine 𝐶 𝑑 , 𝐽 𝑆 , and
water produced data simultaneously.
Unlike MD wetting experiments, which are operated with a vapor pressure difference but no hydraulic
pressure difference across the membrane, during LEP experiments there is an applied hydraulic pressure
difference but no vapor pressure difference across the membrane. Therefore, only pressure-driven liquid
flux occurs when Δ𝑃 exceeds the membrane LEP. At Δ𝑃 near the minimum membrane LEP, when few
pores are wetted, liquid flux may be very small, below measurable values and affected by system
fluctuations (as discussed at the beginning of this section). Because small volume changes could not be
reliably measured, according to equation 3.7 𝐽 𝑆 cannot be reliably determined. We therefore propose that
Δ𝐶 𝑑 /Δ𝑡 can be used as a proxy for 𝐽 𝑆 . By comparing Δ𝐶 𝑑 /Δ𝑡 results with the corresponding 𝐽 𝑆 results from
Figure A.1.1, it can be seen that the shape of the Δ𝐶 𝑑 /Δ𝑡 curve matches that of the 𝐽 𝑆 curve: both decrease
stepwise in Figure A.1.1a and b, both remain constant in Figure A.1.1c and d, and both increase
exponentially in Figure A.1.1e and f.

76

Figure A.1.1 Example distillate conductivity (𝐶 𝑑 ) and salt flux (𝐽 𝑆 ) versus time data a), b) Case 1, where 𝐶 𝑑
initially increases and tapers to a constant concentration (solute mixing); c), d) Case 2, where 𝐶 𝑑 increases
linearly with time (constant wetting), and e), f) Case 3, where 𝐶 𝑑 is initially stable and then increases
exponentially with time (progressive wetting). All membranes were operated with feed and distillate side
temperatures of 53 and 18 °C, deionized water as distillate solution, flow rates of 0.76 L/min on either side
of the membrane, and ≤ 1 kPa transmembrane pressure.

77

A.2 Supplementary Information to Chapter 4
This material is published as supplementary information to a paper published in the journal ACS Applied
Polymer Materials [77].
A.2.1 Materials and Methods
Figure A.2.1 shows a schematic of the bench-scale membrane distillation (MD) apparatus used for all
experimental tests.

Figure A.2.1 Schematic of bench-scale direct contact MD system used for performance characterization.

A.2.2 Membrane Characterization
Cross-sectional Scanning Electron Microscopy
Cross-sections of membranes for imaging were prepared according to the “cryo-snap” freeze fracture
method [1]. Figure A.2.2 shows a representative cross-sectional scanning electron micrograph of the
uncoated PTFE membrane, illustrating the presence of relatively large void spaces in the bulk material.

Figure A.2.2 Scanning electron micrograph of cross-
section of uncoated PTFE membrane at x950
magnification.

78

Surface Pore Size Distributions
Surface pore size distributions for the top (feed side) of the membranes are shown in Figure A.2.3. For the
CA membrane (Figure A.2.3a and b), skewness did not change significantly after coating. For the PTFE
membrane (Figure A.2.3c and d), skewness was significantly different after coating, indicating that the pore
size distribution shifted towards smaller pores. This result agreed with the changes observed in SEM
images (Figure 4.1) and surface pore parameters (Table 4.2). All statistical significances were confirmed
by t-tests with α = 0.05.

Figure A.2.3 Surface pore size distributions for the top (feed
side) of the a) uncoated CA membrane, b) coated CA
membrane, c) uncoated PTFE membrane, and d) coated
PTFE membrane. For each membrane, results are shown as
three overlaid histograms normalized by probability density.
Surface Characterization of Used Membranes
The surface morphology of the used membranes was characterized using field emission scanning electron
microscopy (SEM) (JSM-7001, Jeol USA, Huntington Beach CA) with integrated energy dispersive x-ray
spectroscopy (EDS). Samples were sputter-coated for 60 s prior to imaging. To avoid the development of
artefacts caused by evaporation of leftover feed solution from the membrane surface during the drying
process, all membranes were rinsed prior to SEM/EDS characterization in deionized water. Rinsing may
also have resulted in removal of some crystals formed during MD tests.
Figure A.2.4 shows representative SEM results of used membranes. After use, the membrane morphology
of all membranes was qualitatively similar to that of the corresponding unused membrane. Also, as
expected, fouling/scaling is visible in some areas. Rinsing likely resulted in some removal of loosely
adhered and readily dissolved crystals after use; however, some remaining crystals were observed.
79

Figure A.2.4 SEM images of a) feed side, with inset of NaCl crystal on membrane surface, and b) distillate
side of uncoated PTFE membrane after use with 5 M NaCl feed solution, c) feed side and d) distillate side
of uncoated PTFE membrane after use with 1 M NaCl and 200 ppm surfactant feed solution, e) feed side
and f) distillate side of coated PTFE membrane after use with 5 M NaCl feed solution, g) feed side and h)
distillate side of coated PTFE membrane after use with 1 M NaCl and 200 ppm surfactant feed solution, i)
feed side and j) distillate side of coated CA membrane after use with 5 M NaCl feed solution, and k) feed
side and l) distillate side of coated CA membrane after use with 1 M NaCl and 200 ppm surfactant feed
solution. All images were taken at 4,000x magnification, excepting the inset to figure a), which was taken
at 30,000x magnification.
Figure A.2.5 shows representative EDS results of used membranes. EDS results also indicate that scaling
was detected in some areas. SEM and EDS results suggest that scaling and wetting occurred during MD
tests, which is in agreement with the performance results.
80

Figure A.2.5 Energy dispersive x-ray spectroscopy of a) feed side and b) distillate side of uncoated PTFE
membrane after use with 5 M NaCl feed solution, c) feed side and d) distillate side of uncoated PTFE
membrane after use with 1 M NaCl and 200 ppm surfactant feed solution, e) feed side and f) distillate side
of coated PTFE membrane after use with 5 M NaCl feed solution, g) feed side and h) distillate side of
coated PTFE membrane after use with 1 M NaCl and 200 ppm surfactant feed solution, i) feed side and j)
distillate side of coated CA membrane after use with 5 M NaCl feed solution, and k) feed side and l) distillate
side of coated CA membrane after use with 1 M NaCl and 200 ppm surfactant feed solution.

For the scaling-induced wetting scenario, a few SEM images and EDS scans (data not shown) indicated
that calcium-based crystals were present on the membrane surface. These were likely due to impurities in
the ACS-grade NaCl used to prepare the feed solutions. Calcium salts can be less soluble than NaCl
crystals and therefore may be more likely to remain after rinsing. A recent study by Su et al. [2] found that
superhydrophobic membranes may only delay, and not prevent, gypsum scaling. Although only a few
calcium crystals were observed, these could have aggravated pore wetting for the coated CA membranes.
These crystals did not apparently affect the coated PTFE membrane. Therefore, membranes with reduced
surface pore sizes may be more resilient to wetting caused by different scalants than superhydrophobic
membranes. Further study is needed to investigate this.

81

A.2.3 Supporting References
[1] R. R. Ferlita, D. Phipps, J. Safarik, and D. H. Yeh, "Cryo‐snap: A simple modified freeze‐fracture
method for SEM imaging of membrane cross‐sections," Environmental progress, vol. 27, pp. 204-
209, 2008.

[2] C. Su, T. Horseman, H. Cao, K. Christie, Y. Li, and S. Lin, "Robust superhydrophobic membrane
for membrane distillation with excellent scaling resistance," Environmental science & technology,
vol. 53, pp. 11801-11809, 2019.

Abstract (if available)

Abstract High-salinity waste streams produced during desalination of saline, brackish, and waste waters or from industrial processes present unique treatment challenges, and their management is associated with high capital and operating costs and high energy consumption. Membrane distillation (MD) is a thermally driven membrane process that is promising for the treatment of complex, high-salinity streams. While reverse osmosis is salinity-limited, MD is viable for hypersaline streams because MD relies on a driving force that is only slightly decreased by salinity. While conventional thermal desalination processes are associated with low energy efficiency and high cost, MD can more effectively utilize low-grade thermal energy, achieve higher energy efficiency at smaller capacities, and require lower capital costs. ❧ In MD, a warmer, saline stream flows on one side and a cooler, pure distillate stream flows on the other side of a microporous, hydrophobic membrane. The vapor pressure difference across the membrane drives water in the feed stream to evaporate, diffuse through the membrane pores, and condense into the cooler distillate stream. The resistance of membrane pores to wetting results in a vapor gap that separates the feed and distillate streams and prevents liquid penetration. Thus, membrane wetting resistance is critical to prevent passage of nonvolatile solutes. High wetting resistance is particularly important for high-salinity streams due to the potential for salt precipitation (scaling) on the membrane surface, which can cause pore wetting. Surfactants and alcohols, which may be present in industrial waste streams, can also cause wetting due to reduced surface tension at pore entrances. MD has not been widely implemented for the management of challenging waste streamsㅡin large part due to insufficient membrane wetting resistance upon exposure to high salinity streams and reduced wetting resistance over long-term operation. ❧ This dissertation provides fundamental understanding of pore wetting in MD and introduces new characterization parameters to describe pore wetting resistance in situ. Wetting resistance depends on two membrane properties: hydrophobicity and pore size. While the relationship between hydrophobicity, pore size, and wetting resistance can appear simple, there are complexities that have not been previously addressed. Membrane hydrophobicity may not be constant during long-term operationㅡdue to fouling or physical and chemical changes to the membrane itself. In the literature, hydrophobicity has frequently been represented by the measured, feed-side contact angle of the membrane or the intrinsic contact angle of the smooth membrane material

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

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

PDF

Fate of antibiotic resistance in anaerobic membrane bioreactors

PDF

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

PDF

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

PDF

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

PDF

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

PDF

Wastewater-based epidemiology for emerging biological contaminants

PDF

Vapor phase deposition of dense and porous polymer coatings and membranes for increased sustainability and practical applications

PDF

Development of carbon molecular-sieve membranes with tunable properties: modification of the pore size and surface affinity

PDF

A flow-through membrane reactor for destruction of a chemical warfare simulant

PDF

Optimizing biomembrane reactor systems for water reclamation and reuse applications

PDF

Material characterization of next generation shape memory alloys (Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si) for use in bridges in seismic regions

PDF

Machine-learning approaches for modeling of complex materials and media

PDF

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

Asset Metadata

Creator McGaughey, Allyson Lee (author)

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

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Engineering (Environmental Engineering)

Publication Date 02/22/2021

Defense Date 12/17/2020

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

Tag desalination,hydrophobicity,membrane characterization,membrane distillation,membrane fouling,OAI-PMH Harvest,water treatment,wetting resistance

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Childress, Amy (committee chair), de Barros, Felipe (committee member), Gupta, Malancha (committee member), Smith, Adam (committee member), Wang, Qiming (committee member)

Creator Email allysonmcgaughey@gmail.com,amcgaugh@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-420366

Unique identifier UC11667187

Identifier etd-McGaugheyA-9283.pdf (filename),usctheses-c89-420366 (legacy record id)

Legacy Identifier etd-McGaugheyA-9283.pdf

Dmrecord 420366

Document Type Dissertation

Rights McGaughey, Allyson Lee

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