«23...?

. . it.
2:)

.1
2..
17..

, 2-.
«ﬂunwémmﬁm.

. (l
. 2.331.

2
{-5: .

 

*2
:36“

l‘
n.
.1 .
Sand..."
25 if.
‘1... .3... c.
u. tniell A
11-22. «2.2.2923... ‘3.
.31. 11... A19...
(31

.JVJ: A z. 2. fr. 2.. ..u...1.....£.:...w.. . L. 1.3.2. . . . .. W2. twaﬁvk. .1...x-r
u .32 x g. ”www— QMW.” 2.: 62¢;er a: gﬁé

.1; .1 7......? 4.22.... 2.12.3.2: I . 2...... .. .. .
2%.... .2 2.22.22.22.32. 2 2.2.. .2. 2......- ,,

h
t \-

:z;.- I. :2
. .. 7.04.. ...I r . 1.

 

Jooé

This is to certify that the
dissertation entitled

Novel Polymer Films for Separations in Nanofiltration

presented by
Matthew D. Miller

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemistry

flu/[Ado

Major Professor’s Signi/

5/08/05“

Date

 

 

MSU Is an Afﬁrmative ActlomEquol Opportunity mum

 

_——

 

LIBRARY
Michigan State
University

 

 

 

.7—

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2E cJCIHaEateDuejndd-pJS

Novel Polymer Films for Separations in Nanoﬁltration

BY

Matthew D. Miller

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2005

ABSTRACT

Novel Polymer Films for Separations in Nanoﬁltration

BY

Matthew D. Miller

Nanoﬁltraticn (NF) is a powerful separation technique, capable of
operation in both large- and small-scale applications. Despite extensive
developments in the NF ﬁeld, increased permeate ﬂuxes as well as greater
control over membrane properties are constant objectives. A common target for
NF performance enhancement is the membrane, which is the selective barrier
between the feed and permeate solutions. In this dissertation I detail how the
deposition of ultrathin polymer ﬁlms on porous supports yields selective, high ﬂux
membranes.

To form composite NF membranes with ultrathin polymer skins, I employ
alternating adsorption of polycations and polyanions on a porous support.
Separations can be optimized by varying the constituent polyelectrolytes, and in
general, the use of polycations and polyanions with lower charge densities allows
greater passage of larger analytes, presumably because ionic crosslinking
decreases with decreasing charge density. In situ ellipsometry conﬁrms that
lower charge densities result in highly water-swollen films. Careful selection of
polyelectrolytes results in membranes capable of separating salts, sugars, or,

remarkably, even proteins such as myoglobin and bovine serum albumin.

Additionally, membrane transport characteristics such as rejection and solution
flux can be optimized by simple changes in the ﬁlm deposition process.
Moreover, water ﬂuxes through these ﬁlms are 1.5-5 times greater than through

commercial NF membranes.

To my mom and dad.

ACKNOWLEDGMENTS

I would first like to thank Professor Bruening for his guidance and
assistance throughout the past four years. His ability to always have time for
questions and discussion has made my tenure in graduate school an educational
and enjoyable experience.

I also want to thank the members of the Bruening research group, past
and present, as well as Dr. Baker, J.B., Bao, and Ying for their fruitful and
insightful discussions. I especially thank Dan, Brian, Alex, and Rick for their
friendship and conversation on topics ranging from the newest video cards to the
sorry state of the Mets/Cubsfl'igers/Lions (depending on the year, of course—
unless we were talking about the Tigers or Lions, which are of late always pretty
sorry). I also greatly appreciate the initial nanoﬁltration research performed in the
Bruening Group by Jeremy, Brian, and Xiaoyun; as well as their guidance in
teaching me the equipment and analysis techniques.

I particularly want to thank my family (including Michelle’s family!) and
friends for all their love and support; they have kept me going strong for over 24
years. You have made me strive to be my best every day. Also, special thanks
to mom and dad for not taking away my chemistry set after any of the countless
mishaps in the garage. I promise to pay you back for all the structural damage to
the house... eventually.

Last but certainly not least, I want to thank my ﬁancée Michelle for

selﬂessly supporting me—her encouragement and presence lights my days.

TABLE OF CONTENTS

List of Tables ..................................................................................................... viii
List of Figures .................................................................................................... x
Key to Abbreviations .......................................................................................... xiv
Chapter 1: Introduction ..................................................................................... 1
1.1: Membranes and Nanoﬁltration ....................................................... 2
1.2: Multilayer Polyelectrolyte Films as Skin Layers in NF
Membranes .............................................................................. 14
1.3: Outline of this Dissertation ............................................................. 20
1.4: Works Cited ................................................................................... 24
Chapter 2: The Nanoﬁltration Properties of Multilayer Polyelectrolyte
Membranes through Variation of Film Composition .................. 32
2.1: Introduction .................................................................................... 33
2.2: Experimental .................................................................................. 36
2.3: Results and Discussion .................................................................. 42
2.4: Conclusions ................................................................................... 63
2.5: Acknowledgements ........................................................................ 63
2.6: Works Cited ................................................................................... 64
Chapter 3: Correlation of the Swelling and Permeability of
Polyelectrolyte Multilayer Films ................................................ 67
3.1: Introduction .................................................................................... 68
3.2: Experimental .................................................................................. 71
3.3: Results and Discussion .................................................................. 75
3.4: Conclusions ................................................................................... 96
3.5: Acknowledgements ........................................................................ 97
3.6: Works Cited ................................................................................... 98

vi

Chapter 4: Modified Polyelectrolyte Multilayer Films for Enhanced

Nanoﬁltration of Neutral and Charged Molecules ............................................ 101
4.1: Introduction ................................................................................. 102
4.2: Experimental ............................................................................... 106
4.3: Results and Discussion ............................................................... 111
4.4: Conclusions ................................................................................ 122
4.5: Acknowledgements ..................................................................... 123
4.6: Works Cited ................................................................................ 124
Chapter 5: Conclusions and Future Work ...................................................... 127
5.1: Conclusions ............................................................................................ 127
5.2: Future Work ............................................................................................ 128
5.3: Works Cited ............................................................................................ 132

vii

LIST OF TABLES

Table 2.1: Molecular weights, aqueous diffusion coefﬁcients (D), and
Stokes’ radii (rs) of the neutral molecules used in transport
studies. .......................................................................................... 44

Table 2.2: Flux and selectivity values for diffusion dialysis of glycerol,
glucose, sucrose, and rafﬁnose through porous alumina
coated with various PSS/PDADMAC ﬁlms deposited from
0.5 M NaCI. ................................................................................... 46

Table 2.3: Flux and selectivity values for diffusion dialysis of glycerol,
glucose, sucrose, and rafﬁnose through bare porous alumina
and alumina coated with various PSS/PDADMAC ﬁlms
deposited from 0.1 M NaCl. ........................................................... 50

Table 2.4: Rejections, water ﬂuxes, and selectivities from nanoﬁltration
of a series of neutral molecules using porous alumina coated
with a variety of PEM ﬁlms. ........................................................... 51

Table 2.5: Rejections, water ﬂuxes, and selectivities from nanoﬁltration
of a series of neutral molecules using porous alumina coated
with PSS/PDADMAC ﬁlms deposited from 0.1 M NaCl. ................ 54

Table 2.6: Rejections, water ﬂuxes, and selectivities from nanoﬁltration
of sucrose and NaCl by porous alumina coated with
PSS/PDADMAC ﬁlms. ................................................................... 61

Table 3.1: Ellipsometric thicknesses of PEMs under nitrogen (<5% RH),
water, or ethanol, and percent swelling in the two solvents. .......... 82

Table 3.2: Percent rejection in nanoﬁltration of glucose, sucrose, and
rafﬁnose dissolved in water. PSS/PDADMAC and
HA/chitosan data are from Chapter 2. The swelling values
are from analogous ﬁlms (Table 3.1). ............................................ 91

Table 3.3: Fluxes in diffusion dialysis of glucose, sucrose, and rafﬂnose
(dissolved in water) through porous alumina coated with
various polyelectrolyte ﬁlms. The bare alumina and
PSS/PDADMAC data are from Chapter 2. All ﬁlm
thicknesses were measured in a <5% RH nitrogen
atmosphere except for that of [HA/chitosan18HA, which is an
estimate based upon the thickness of 8-bilayer HA/chitosan
ﬁlms. .............................................................................................. 92

viii

Table 3.4: Fluxes in diffusion dialysis of 140 uM glucose, sucrose, and
rafﬁnose in ethanol through porous alumina coated with
various polyelectrolyte films. All ﬁlm thicknesses were
measured in 3 <5% RH nitrogen atmosphere except for that
of [HA/chitosanjaHA, which is an estimate based upon the
thickness of 8-bilayer HA/chitosan ﬁlms. ....................................... 95

Table 4.1: Rejections, solution ﬂuxes, and selectivities from
nanoﬁltration of a series of neutral molecules using porous
alumina coated with a variety of multilayer polyelectrolyte
ﬁlms.6'7. 0.5 M NaCl was used in deposition of the
PSS/PDADMAC layers. Films are given numbers to aid in
discussion .................................................................................... 108

Table 4.2: Rejections, solution ﬂuxes, and selectivities from
nanofiltration of sucrose and NaCl by porous alumina coated
with a variety of PEM ﬁlms arranged in order of increasing
NaCl/sucrose selectivity. 0.5 M NaCl was used in deposition
of the PSS/PDADMAC layers. Films are numbered to aid in
discussion .................................................................................... 1 16

Table 4.3: Rejections, solution ﬂuxes, and selectivities from
nanoﬁltration of BT and sodium sulfate using porous alumina
coated with a variety of PEM ﬁlms. All PSS/PDADMAC
layers were deposited from 0.5 M NaCI unless indicated
otherwise. Films are numbered to aid in discussion. .................. 119

Table 4.4: Rejections, solution ﬂuxes, and selectivities from
nanoﬁltration of BT and potassium phosphate at pH 8 using
porous alumina coated with PEM ﬁlms. All PSS/PDADMAC
layers were deposited from 0.5 M NaCl. Films are
numbered to aid in discussion. .................................................... 121

LIST OF FIGURES

Figure 1.1: Schematic depiction of a pressure-driven, size-selective
membrane separation ...................................................................... 2

Figure 1.2: Representation of the concentration proﬁle in NF. Charge or
size exclusion of the solute at the membrane-feed interface
results in rejection. .......................................................................... 5

Figure 1.3: Illustration depicting the size-exclusion model described in

Equation 1.6 .................................................................................... 7
Figure 1.4: Schematic description of the layer-by-layer assembly of

polyelectrolyte multilayer ﬁlms. Counter-ions and

intertwining of polymers are not shown for the sake of

clarity. ............................................................................................ 15
Figure 2.1: Schematic showing the NF unit used in this dissertation ............... 40

Figure 2.2: Schematic diagram of a “nanoﬁltration cell” from Figure 2.1
showing the assembly of the cell components. From top to
bottom: (1) and (2) are the inlet and outlet ports, (3) are the
threads, (4) is the rubber O-ring, (5) is the membrane, (6) is
the stainless steel frit that the membrane is placed on, (7) is
the cap that holds the frit, and (8) is the bottom part of the
cell that screws into the threads (3). This ﬁgure is adapted
from Stanton et al. The arrow indicates flow direction
through the cell. ............................................................................. 41

Figure 2.3: Structures of polyelectrolytes used in this study ............................ 43

Figure 2.4: SEM images of porous alumina coated with polyelectrolyte
ﬁlms. (a) cross section of 3 ~30 nm thick, 5-bilayer
PSS/PDADMAC ﬁlm deposited from 0.5 M NaCl. (b) top-
down view of a 4.5-bilayer PSS/chitosan ﬁlm. ............................... 47

Figure 2.5: NaCI flux and NaCI/Sucrose selectivity from DD through
porous alumina coated with PSS/PDADMAC films deposited
from 0.5 M NaCl. The arrows indicate which y-axis pertains
to each data set. Flux was normalized by dividing by the
source-phase concentration. ......................................................... 58

Figure 2.6: NaCl ﬂux and NaCl/Sucrose selectivity from diffusion dialysis
through porous alumina coated with PSS/PDADMAC ﬁlms
deposited from 0.1 M NaCl. The arrows indicate which y-
axis pertains to each data set. Flux was normalized by
dividing by the source-phase concentration. ................................. 59

Figure 3.1: Diagram depicting the model of light reﬂection used to ﬁt
ellipsometric data. The n and k terms describe the real and
imaginary parts of the complex refractive indices, while d
represents the layer thickness. The layers are not drawn to
scale. ............................................................................................. 76

Figure 3.2: Calculated W and A values for a SilSiOzlﬁlm system (water
as the ambient medium and a 2.0 nm silicon oxide layer) as
a function of the refractive index and thickness of the ﬁlm.
The simulation was performed at a wavelength of 450.5 nm
where the optical constants are: water - n=1.3395, silicon
oxide - n=1.4644, k=0, and silicon - n=4.7108, and k=0.0963.
The point with a thickness of 55 nm and a refractive index of
1.44 represents a 9.5-bilayer PSS/PDADAMC ﬁlm deposited
from 0.1 M NaCl, and the black lines through that point show
the i0.3° uncertainty in KP and A measurements. Enclosed
data points correspond to identical thicknesses at different
refractive indices ............................................................................ 77

Figure 3.3: Calculated LP and A values for the system depicted in Figure
3.1 (nitrogen with <5°/o relative humidity as the ambient
medium and a 2.0 nm silicon oxide layer) as a function of the
refractive index, ng, and thickness, d2, of the ﬁlm. The
simulation was performed at a wavelength of 450.5 nm
where the optical constants are n1=1.000, n3=1.4644, k3=0,
n4=4.7108, and k4 = 0.0963. The point with a thickness of 25
nm and a refractive index of 1.54 represents a 9.5-bilayer
PSS/PDADAMC ﬁlm deposited from 0.1 M NaCl, and the
black lines through that point show the i0.3° uncertainty in l~lJ
and A measurements. Enclosed data points correspond to
identical thicknesses at different refractive indices. ....................... 78

xi

Figure 3.4: Calculated llJ and A values for the system depicted in Figure
3.1 (ethanol as the ambient medium and a 2.0 nm silicon
oxide layer) as a function of the refractive index, n2, and
thickness, d2, of the ﬁlm. The simulation was performed at a
wavelength of 480.2 nm where the optical constants are
n1=1.3645, n3=1.4636, k3=0, n4=4.412, and k4 = 0.0629. The
point with a thickness of 50 nm and a refractive index of 1.47
represents a 9.5-bilayer PSS/PDADAMC ﬁlm deposited from
0.1 M NaCl, and the black lines through that point show the
:0.3° uncertainty in W and A measurements. Enclosed data
points correspond to identical thicknesses at different
refractive indices ............................................................................ 79

Figure 3.5: A 3 x 25 um AFM topographical image and the derived
average line scan for thickness analysis of a
[PSS/PDADMACho film deposited from 0.1 M NaCl. The
region between the two black arrows on the left determines
the average ﬁlm + wafer height, and the area between the
two dark grey arrows on the right determines the height of
the scratched Si wafer. The average PEM thickness for this
particular scan is 28.3 nm .............................................................. 81

Figure 3.6: The experimental (dashed line) and generated (ﬁt, solid line)
W and A ellipsometric data for a dry (under <5% RH
nitrogen) 10-bilayer PSS/PDADMAC ﬁlm deposited from 0.1
M NaCl. This ﬁt corresponds to a film that is 24 nm thick and
has a refractive index of 1.541 at 450.5 nm ................................... 83

Figure 3.7: The experimental (dashed line) and generated (ﬁt, solid line)
W and A ellipsometric data for a water-submerged 8-bilayer
HA/chitosan ﬁlm. This ﬁt corresponds to a ﬁlm that is 124
nm thick and has a refractive index of 1.389 at 450.5 nm. ............ 84

Figure 3.8: The experimental (dashed line) and generated (ﬁt, solid line)
KP and A ellipsometric data for an ethanol-submerged 9.5-
bilayer PSS/PDADMAC ﬁlm deposited from 0.1 M NaCl.
This ﬁt corresponds to a ﬁlm that is 52 nm thick and has a
refractive index of 1.453 at 480.2 nm. ........................................... 85

Figure 3.9: Structures of the polyelectrolytes used in Chapter 3. .................... 87

xii

Figure 4.1: Schematic illustration of the concept of hybrid polyelectrolyte
membranes containing PSS, PDADMAC, and PAH
deposited onto a porous alumina support. The color-coded
chemical structures of the individual polyelectrolytes are also
shown. (The porous support is actually several orders of
magnitude thicker than the polyelectrolyte ﬁlm.) .......................... 104

Figure 4.2: Cross-sectional SEM image of a [PSS/PDADMAC]3-
[PSS/PAH]PSS* ﬁlm deposited on porous alumina. The ﬁnal
PSS layer was deposited from 2.5 M MI'ICiz. ............................... 112

Figure 5.1: Simpliﬁed mechanism for the crosslinking reaction of HA and
chitosan by EDC/NHS coupling. .................................................. 129

xiii

 

LIST OF ABBREVIATIONS

Abbreviation . Deﬁnition

AFM ...................... Atomic Force Microscopy

DD .......................... Diffusion Dialysis

EDC ....................... 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide
FT-IR ...................... Fourier Transform Infrared

HA .......................... Hyaluronic Acid

Mw ......................................... Molecular Weight

MWCO ................... Molecular Weight Cutoff

NF .......................... Nanoﬁltration

NHS ....................... N-hydroxysuccinimide

PAA ........................ Poly(acrylic acid)

PAH ....................... Poly(allylamine hydrochloride)
PDADMAC ............. Poly(diallyldimethylammonium chloride)
PEI ......................... Poly(ethylene imine)

PEM ....................... Polyelectrolyte Multilayer

PSS ........................ Poly(styrene sulfonate)

RH .......................... Relative Humidity

RO ......................... Reverse Osmosis

SEM ....................... Scanning Electron Microscope

xiv

Chapter 1

INTRODUCTION

Chemical separations are essential processes in diverse applications
ranging from chromatographic analysis to petroleum reﬁning to the isolation of

1-3

pharmacologically active plant components. Techniques employed in such

separations include distillation,“5 recrystallization,7'8 centrifugation,9‘11

15-17 18-20

sublimation,”14 dialysis, and chromatography. This research focuses
speciﬁcally on nanoﬁltration (NF), which is a membrane-based process similar to
reverse osmosis. Membrane separations are often employed in large-scale

industrial processes such as desalination21 and gas separations,”24

though they
are also important in small-scale applications such as membrane introduction
mass spectrometry.25'26 These separations rely on the interaction of chemical
compounds with a selective phase, the membrane, to effect separations. The
selective phase is the foundation of these separations and should be amenable
to key improvements that employ novel ﬁlm chemistry.

This dissertation explores the use of multilayer polyelectrolyte multilayer
(PEM) ﬁlms as the discriminating layer in membranes. The minimal thickness of
the polyelectrolyte films allows ﬂuxes that are 1.5 to 5-fold greater than those
through commercial membranes, and Chapter 2 discusses how careful selection
of the constituent polycations and polyanions in these ﬁlms allows development

of membranes with a wide range of molecular weight cutoffs. Multilayer

polyelectrolyte ﬁlms are particularly promising for sugar and salt/sugar

separations. Chapter 3 discusses how the ellipsometrically measured swelling
behavior of these ﬁlms relates to their transport properties, and Chapter 4
investigates how slight changes to polyelectrolyte deposition systems can
signiﬁcantly enhance several practical separations.

To put these results in context, this introduction first brieﬂy discusses
separation mechanisms in the area of NF as well as some membrane synthesis
methods. The next section describes prior research on the formation and
structure of multilayer polyelectrolyte ﬁlms. Finally, an outline for the other

chapters of the dissertation is given.

1.1 Membranes and Nanof‘rltration
Membrane systems utilize a

discriminating layer (the membrane)

 

O o t
to allow selective transport between *0 mp " a
’ e 9 P e
two phases as shown in Figure 1.1. 9C C)
9o o AP c
Several forces can drive transport® O >
across this layer, including pressure JD 5 @
_ . @09 6C; o
differences, e.g., nanoﬁltration 0Q Q
27 - 21 \
(NF), reverse osmosrs (R0), and I
gas separations; 22 24 a
membrane
concentration gradient, e.g.,

Figure 1.1: Schematic depiction of
dialysis;28 or even an electrical apressure-driven, size-selective
membrane separation

29-31

potential difference, e.g., electrodialysis. Membrane separations can occur

between two liquid phases,32 from a liquid to a gas phase (pervaporation),33"35
and between two gas-phases.”38 This work focuses on liquid-liquid separations
that include diffusion dialysis and NF, but the primary emphasis is on NF
because of its higher ﬂuxes and greater practicality. In this technique, pressure
drives a solvent (with some accompanying solutes) across a membrane against
a concentration gradient. NF is similar to reverse osmosis, but NF membranes
are more permeable so lower pressures can be applied to achieve similar
fluxes.32'39'4° The use of lower pressure makes NF more economical than R0 for
separations that do not require high NaCl rejections.

The term “nanoﬁltration” was initially coined by the membrane company
FilmtecTM in the mid-19803, but the name was retroactively applied to
separations with water ﬂuxes from 0.2-2 m3/(m2 day bar) and NaCl rejections of
2043004,.“42 The molecular weight cutoff (cho, solute molecular weight
needed to achieve <10°/o passage of the solute through the membrane) of NF
membranes is generally between ZOO-10000 g/mol,43 so NF applications include
selective removal of molecules such as sugars,44 herbicides,45 pesticides,46 and
dyes.47 The largest application of NF is the softening of water (removal of Mg2+
and Ca2+ ions), and some plants have been built that can process 40 million
gallons of water per day.48 Despite these applications, improved membranes
with higher permeabilities, greater stabilities and lower propensities for fouling

would certainly be beneﬁcial for expanding the scope of this technique. Hence

the goal of this work is to develop a versatile method for forming ultrathin ﬁlms
that are capable of a wide range of high-ﬂux separations.

Separation Mechanisms in NF. NF membranes rely primarily on two
mechanisms for selective transport: charge exclusion and sieving. In both
mechanisms, transport is often modeled by assuming equilibrium at the feed and
permeate interfaces and allowing transport within the membrane to occur by
convection and diffusion (Figure 1.2). In charge exclusion (also called Donnan
exclusion), a high density of charge on the membrane surface results in
exclusion of species in solution with a charge of the same Sign.”53 Because
exclusion increases with the charge on the species, this mechanism can
separate singly and doubly charged species such as chloride and sulfate50'53 or

sodium and magnesium ions.49

To understand how a charged membrane rejects ions, consider a

membrane exposed to a solution of a single binary salt, AxBy, where 2A is the

charge on the cation and 23 is the charge on the anion. Because of the ﬁxed
charge on the membrane, the concentrations of mobile cations and anions within
the membrane are not the same. At equilibrium, this creates an electrical
potential (the Donnan potential), which is described by Equation 1.1,54 where
W00“ is the Donnan potential, R is the gas constant, T is temperature, F is

Faraday’s constant, aM is the activity of the ion in the membrane, and a8 is

RT as RT a8
w =———ln—Bandw z—ln—A
Don (ZBFJ [33]] Don {ZAF] [3%] (1'1)

 

Membrane

  

 

 

   

 

Solute _ 9:8
........ a

2..

(D

O

O

3

O

(D

a

Q.

5.

3

Feed Side Permeate
Side

Figure 1.2: Representation of the concentration proﬁle in NF. Charge or
size exclusion of the solute at the membrane-feed interface results in
rejection.

the activity of the ion in solution. This equation applies to both the cation and
anion of the salt. Equating the Donnan potential for each species and assuming
that solutions are sufﬁciently dilute that concentrations, c, equal activities yields
Equation 1.2.

RT cg RT of,
—'" _M =—“ 77 (1-2)
ZBF CB ZAF CA

The assumption of charge neutrality both in solution and in the membrane results
in Equations 1.3 and 1.4, respectively, where c8 is the species concentration in
bulk solution, cM is the species concentration in the membrane, and CW is the

concentration of ﬁxed charge due to the membrane material.
8 S
|zA|CA =|zB|Cj3 (1.3)
M M M
IZAICA =|ZB|CB +|Zx|Cx (1-4)
Finally, substituting Equations 1.3 and 1.4 into Equation 1.2 leads to the
distribution coefﬁcient for the anion shown in Equation 1.5. A similar expression
can be derived for the cation. In the case of a divalent and monovalent sodium
salt and a negatively charged membrane, this equation will result in a much
smaller distribution coefﬁcient for the divalent anion.

9g: |ZB|Cg A
c3 |zB|cg” +|zx|c,"§’

ZB

ZA (1.5)

 

 

For neutral molecules, the major separation mechanism in porous
membranes is sieving, in which transport into membrane pores depends on
solute dimensions. The most important variables for predicting the rejection
properties of a sieving membrane are the radii of the solutes and membrane
pores.55 Figure 1.3 describes a very
simple partition model with spherical
solutes and cylindrical pores used to
estimate steric—based membrane /
rejection, where r is the radius of the .
solute and R is the radius of the pore

(assuming a uniform pore-size). This X,

 

model assumes that the closest a

"~l.~- .
5‘. .

Figure 1.3: Illustration depicting the
size-exclusion model described in
Equation 1.6

solute can approach the pore wall is
the radius of that solute, r. Thus, the
center of the solute can only occupy a fraction of the cross-sectional area, (1-
r/R)2 (also called CD). Essentially, the ratio of solute to pore radius determines an
effective partition coefﬁcient, CD, that dictates how well molecules can enter the
membrane. lf transport to the permeate side is primarily due to convection, the
partition coefﬁcient can also predict rejection as described by Equation 1.6.41'44'55
Hindered convection or diffusion will result in a more complicated model of

transport.“4'56'57

2
Percent Rejection =[1—{1—é] ] * 100% (1.6)

Equation 1.7 (a form of Fick’s ﬁrst law) describes steady state diffusion
through a membrane, where j,- is the solute flux, D; is the diffusion coefﬁcient of
the solute through the membrane, c,;x=o and cm)», are the concentrations of the
solute in the membrane at the feed and permeate sides of the ﬁlm, respectively,

and Ax is the thickness of the membrane.

 

C- _ —C' _
jiz—Di l,x—0Axl,x_Ax (1.7)

However, the concentrations of solute at the membrane/solution interfaces are
controlled by (D, where cm and c,-,,, are the concentration of the solute in the feed

and permeate, as shown in Equation 1.8.

 

C- _ C-
(I): l,x—O= l,x=Ax (1.8)

Cif Cl.p

Also, the diffusion coefﬁcient in Equation 1.7 is a combination of the diffusion
coefﬁcient of the solute at inﬁnite dilution, D,;,-,,f, as well as the hindrance factor for
diffusion, Kid, and the ﬁlm porosity, g, as shown in Equation 1.9.

Di = Dl,iani,d8 (1 -9)
Note that Km. assuming a homogeneous velocity across the membrane pores, is
similar to the enhanced drag coefﬁcient, K’, which is a function of the ratio of
solute to pore radius, and expressions for calculating this value are available in
literature."“'57 Substitution of Equations 1.8 and 1.9 into Equation 1.7 results in
Equation 1.10, an expression for diffusive ﬂux through a membrane.

. (“Cir—C; )
fl = TDiinr Kid 5 ' [p

 

(1.10)

Solvent ﬂux in pressure driven processes like NF is described by Equation
1.11, where 27 is the solvent viscosity, 2' is the pore tortuosity, and AP is the

pressure drop across the membrane.

_ aRZAP

JV —
8771AX

 

(1.11)

Including both hindered diffusion and hindered convection, Km, the expression for

solute ﬂux in NF is shown in Equation 1.12.

. dc-
Ii = Ki,cClJv _ Dl,iani,d8 ‘d—j“ (1-12)

The hindered convection term can be approximated by the lag coefﬁcient, G,
which is also a function of the ratio of solute to pore radius. Again, expressions
for calculating this value are available in the literature.44 Finally, integration of
Equation 1.12 with the boundary conditions in Equation 1.13, results in an
expression for rejection that combines hindered diffusion and convection with the

ratio of solute to membrane pore radius, Equation 1.14.44'57

The partition
expression does not factor into the solute concentration on the permeate side of
the membrane in Equation 1.13 because that side is not stirred with the NF

system discussed in this dissertation (only the feed side is stirred with our cross-

flow equipment).

 

Cif (I) = Cl,x=o§ Ci,p = Cl,x=Ax (1-13)
- - Ki C(D
Rejectlon =1— K ' (1.14)
1—exp[——i'3ﬂ](1—K,,C)
Ki,d Di,infg

While the previous theoretical discussion describes the pore-ﬂow model, it
is also possible that a solution-diffusion mechanism inﬂuences solute and solvent
transport. In this model, solute and solvent dissolve in the ﬁlm, cross the
membrane via diffusion, then desorb into the permeate. In the case of solution-
diffusion, Equation 1.15 describes the solvent flux in NF (Ji), where D is the
diffusion coefﬁcient of solvent in the membrane, Ki is the sorption coefﬁcient of
the solvent, ci is the concentration of the solvent, A1: is the osmotic pressure, R is
the gas constant, T is temperature, and vi is the molar volume of the solvent.

_ DiKiCiVi(AP—A7T)
_ AXRT

 

J,- (1.15)

Equation 1.16 describes the solution-diffusion solute ﬂux through the membrane
where Dj is the diffusion coefﬁcient of the solute, K is the sorption coefﬁcient of
the solute, and Ac is the concentration change of the solute across the
membrane.

J j = 91559.1. (1.16)
A more in-depth derivation of these equations is beyond the scope of this
dissertation, though it is available elsewhere.41

It is often difﬁcult to determine the contribution of solution-diffusion or
transport through pores to the solute transport through the membranes described
in this dissertation. Baker states that the transition between the two mechanisms
occurs when the effective radii of membrane pores are between 0.25-O.5 nm.41

The pore size of the membranes described in this dissertation are likely above

this threshold. A previous modeling study with similar, more rejecting ﬁlms found

10

the average pore radius of the membranes to be between 0.4-0.5 nm.44 An
investigation of the local solute environment using a technique like ﬂuorescence

lifetime measurement58

may elucidate if transport occurs through the water-ﬁlled
pores or along the polymer backbone. Tedeschi et al. performed a similar study
with PEMs utilizing pyrene fluorescence as a polarity sensitive probe,59 but
unfortunately most of their data were collected when the ﬁlms were under varying
degrees of relative humidity and not immersed in water.

Synthesis of NF membranes. Sieving properties of membranes are a function of
pore size, and one of the largest factors that affect this variable is the method of
membrane synthesis. Most current NF membranes are made by interfacial
polymerization, phase inversion, or surface modiﬁcation of preexisting
membranes.60 lnterfacial polymerization is the process of loading a porous
support with a reactive species (such as a diamine) dissolved in solvent A, and
then immersing it in a complementary reactive species (eg, a di-acid chloride)
dissolved in solvent B. The two solvents are immiscible so the polymerization
occurs only at the solution A/solution B interface.60 Such membranes are
advantageous in that very little material is needed in the thin skin layer, so
expensive, high-performance polymers can be employed to make the skins.
Many recent membranes made by this process utilize trimesoylchloride as one of
the reactive speciess”53 This monomer is popular because it possesses three
active sites that can be reacted to varying degrees depending on stoichiometric

control. The unreacted groups are subsequently hydrolyzed, lending additional

control over the hydrophilicity to the membrane.‘50 The amine co-reactant is often

11

m— or p—phenylenediamine,62'63 however, bipiperidine or bisphenol derivatives are
often used to increase chemical resistances“56

Phase inversion is another popular technique, which yields asymmetrically
skinned membranes. In general, these structures are made by the precipitation
of a solvated polymer to form a membrane whose surface has very small pores
and sits on top of a porous, spongy bulk. Some of the methods for producing
these ﬁlms include immersion of dissolved polymer into a solvent in which the
polymer is not soluble, removal of solvent from a solution of a polymer in a
solvent/non-solvent system, temperature reduction, and placement of a cast ﬁlm
in an atmosphere that contains non—solvent saturated with a solvent. The
porosity of these ﬁlms is controlled by a combination of the polymer type, casting
solution, post-casting treatment, coagulation method, and post-precipitation
treatment.“60

Loeb and Sourirajan pioneered the phase inversion technique and
specialized in making membranes from cellulose acetate.67 Since cellulose
acetate membranes suffer from chemical instability,60 asymmetric membranes
have since been produced from several different polymers, including
polyamides,68 polyimides,69 sulfonated polysulfone,70 and brominated
poly(phenylene oxide).71 Many recent phase inversion membranes involve co-

polymers of poly(vinylidene ﬂuoride) (PVDF).72'77

For example, Jeon and co-
workers cast a mixture of poly(vinylidene ﬂuoride—co—hexaﬂuoropropylene) and
poly(ethylene oxide-co-ethylene carbonate) to form high ﬂux, highly stable

membranes for use in a polymer electrolyte system.76 Another novel system by

12

Zhai et al. involves membranes made from a co-polymer of PVDF and 2-(2-
bromoisobutyryloxy)ethyl acrylate, the latter acting as an initiator for atom
transfer radical polymerization.77

The ﬁnal general membrane preparation technique discussed here is the
one used in this dissertation, the physiochemical modiﬁcation of membrane
surfaces. One such method involves the plasma treatment of polymeric
membranes. This process can form groups that increase permeability or,
depending on the type of plasma used, induce cross-linking to increase

stability.78

Membranes can also be chemically treated to increase performance,
i.e. sulfonation to increase water ﬂux and ion rejection.79 Another surface
modiﬁcation technique involves the direct attachment of polymers to the
membrane surfaceao'86 Ulbricht and Yang demonstrated this method when they
grew acrylic acid from initiators trapped at the surface of polypropylene
membranes.86

The speciﬁc surface modiﬁcation method utilized here involves the layer-
by-layer adsorption of anionic and cationic polyelectrolytes on a ceramic alumina

membrane.5°"‘37'89

More than 40 years, ago, Michaels demonstrated that
precipitated polycation/polyanion complexes are capable of selective
separations,90 and a number of studies examined adsorption of single

91'” However,

polyelectrolytes on the surface of separation membranes.
deposition of multilayer polyelectrolyte ﬁlms differs from adsorption of single
polyelectrolytes in that the multilayer ﬁlms do more than just modify the

membrane, they become the selective layer. Moreover, in comparison to

13

precipitated polyelectrolyte complexes, the layer-by-layer process should provide
much better control over both membrane thickness and permeability. The next
section more thoroughly discusses the large body of research concerning the

properties and structure of multilayer polyelectrolyte ﬁlms.

1.2 Multilayer Polyelectrolyte Films as Skin Layers in NF Membranes
Multilayer assemblies formed via layer-by-layer deposition have been
extensively explored in recent literature.“105 A variety of interactions can be
used to assemble such ﬁlms, including both covalent and hydrogen bonding‘oz'104
as well as donor-acceptor coupling.105 However, the most popular method for
forming multilayer ﬁlms employs electrostatic interactions between polycations
and polyanions.106 Assembly of such ﬁlms can occur using the simple “dip-and-
rinse” procedure illustrated in Figure 1.4, where charged substrates are
immersed in a polyelectrolyte solution, rinsed with water, immersed in a solution

of oppositely charged polyelectrolyte, and rinsed again.107

This process is
repeated until a ﬁlm of desired thickness is produced. These ﬁlms are attractive
as skin layers of membranes because their thickness can be controlled simply by
varying the number of deposited layers, and the use of a variety of constituent
polyelectrolytes should allow control over ﬁlm permeability.

PEM Assembly and Structure. Knowledge of the structure of PEMs will be vital

to understanding their permeability, and a number

14

<3th Co 9.3 9: a2 Esocm 6: 9m BoEboq Co mEcSEﬂE

ucm mcoioEzoo .mEE 8‘8:ng 2>_o:oo_o>_oa B anmmmm Em>m_->n¢o>m_ 9: ho cozatomov oszocom “3. 9:9“.

I

+

:o_cm>_0n_ +

$er

+

+

 

cosmozon.
\omEm

- @w 6'
- -67 0V

i l a" a"
-ewge"

'
6" @V

 

15

of insightful studies have been performed in this regard. 100408421 Some of the
earliest reports of PEMs by Decher and coworkers state that smooth, ordered
films can be built up with over 100 layers with little to no change in adsorption

characteristics.108

Additionally, these ﬁlms do not exist in a discrete, layered
structure, as interpenetration between the layers is estimated to be at least 4
polymer layers thick,100'109 though this value can vary depending on ﬁlm type and
assembly conditions.

The structural properties of PEM ﬁlms made with strong polyelectrolytes
(polyelectrolyte repeat units that are fully ionized in solution) are heavily
inﬂuenced by the salt concentration present during deposition."°'112 High
concentrations of supporting electrolyte result in screening of the charges on the
polyelectrolyte and lead to coiled polymer chains that form thick ﬁlms."°'111 For
example, Dubas and Schlenoff demonstrated that there is a linear relationship
between salt concentration in deposition solutions and the thickness of
poly(styrene sulfonate) (PSS)/poly(diaIlyldimethylammonium chloride)
(PDADMAC) ﬁlms.110 The structure of ﬁlms made with weak polyelectrolytes
(polyelectrolyte repeat units that are not completely ionized in solution, such as
primary amines and carboxylic acids) is especially sensitive to deposition pH,
which controls the degree of polyelectrolyte ionizationm'115 Yoo et al. showed
that bilayer composition, surface wettability, layer interpenetration, and layer
thickness are all controlled by simply varying the polyelectrolyte deposition pH.114

Interestingly, some polyelectrolyte systems grow exponentially (as

opposed to linearly) with the number of added polymer layers.“€"121 This rapid

16

growth may occur because at least one of the polyelectrolytes is capable of
penetrating and diffusing through the bulk of the ﬁlm, as opposed to linearly
growing ﬁlms that interpenetrate only over a couple of bilayers. When the ﬁlm is
rinsed, polyelectrolytes that diffuse throughout the ﬁlm are not readily removed,
and when the membrane is brought into contact with a polyelectrolyte with
opposing charge, the previously deposited polymer chains diffuse back towards
the solution/ﬁlm interface, precipitate with the new polyelectrolyte and form an

120,122

extremely thick layer. Films constructed from hyaluronic acid (HA) and

chitosan exhibit this non-linear growth behaviorm"121

and possess very
interesting permselectivity and swelling properties, as discussed in Chapters 2
and 3 of this dissertation.

Separations with PEMs. Careful consideration of properties such as
polyelectrolyte type (strong or weak polyelectrolytes, linearly or exponentially
growing) and deposition conditions (pH and salt concentration) should allow

33,35,44.49,53,123

tailoring of the permeability of PEMs. Many studies of transport

through PEMs have been published, and include techniques like

3334324427 NF,“"'53'123'128 and gas separation.“9'129 Krasemann and

pervaporation,
Tieke reported that 60-bilayer ﬁlms exhibit a diffusional selectivity of over 100 for
a mixture of Na+/Mg2+ and 45 for a mixture of Cl' and S042”.130 However, these
results were obtained with thick ﬁlms that limited ﬂux. Tieke’s group also

produced PEMs for NF that exhibited 8042‘ rejections of up to 98.5%, but

solution ﬂux was only 0.003 m3/(m2 day bar).123 Stanton et al. used much thinner

17

PEM ﬁlms to obtain a sulfate rejection of 95% with a solution ﬂux of
0.4 m3/(m2 day bar), a stark improvement in hydraulic permeability.53

Despite their charged nature, PEMs are not limited to ionic separations.
They recently found utility in pervaporation separations and are attractive for
removing water from organic/water mixtures.33'3“‘124125"127 Very recent work by
Schwarz and Malsch demonstrated that PEMs can separate cyclohexane and
benzene via pervaporation.126 However, these separations are likely based upon
solute solubility in the membrane material, so fractionation of molecules with
similar polarities may be challenging.

This work explores the use of PEMs in NF of neutral molecules and in the
separation of salts from neutral molecules. Although a few papers examined NF
of salts using PEMs, only one study, published by our group, examined NF of
neutral molecules with PEMs.“ That work showed that high selectivities
between glucose and sucrose are possible, but the high rejection of organic
solutes (the rejection of even methanol is 70%) may limit the use of these
membranes in applications where high solute recovery is desired. This work
shows that both control over MWCO and increased NF ﬂuxes are possible with
appropriate selection of the polyelectrolytes in PEMs. Moreover, ellipsometric
data demonstrate that ﬁlm swelling in aqueous solutions increases with
decreasing charge density of the constituent polyelectrolytes and correlates well
with MWCO.

This dissertation also shows the promise of multilayer polyelectrolyte ﬁlms

for two speciﬁc applications: the fractionation of oligosaccharides and the

18

separation of salt from sugar. Process streams in sugar reﬁning consist of a
variety of sugar oligomers, such as glucose, sucrose, rafﬁnose, and stachyose,
that all fulﬁll various chemical and biological roles.‘3"133 Ideally, membranes with
carefully controlled MWCOs could fractionate these sugar oliogomers, which vary
in molecular weight by 160-180 g/mol per additional ring unit. One commercial
membrane utilized for this task, the DS-5-DL from Osmonics, exhibits a
glucose/sucrose selectivity of about 40, but the solution ﬂux through these ﬁlms
is less than 0.2 m3/(m2 day bar).131 The PEM NF membranes detailed in this
dissertation are capable of achieving similar selectivities, but with twice the ﬂux.
Another possible application of PEM membranes involves the separation
of salt and sugar. During the sugar reﬁning process, the feed is passed through
a bed of anion exchange resin to remove color bodies and other impurities.
Once the anion exchange resin reaches its total exchange capacity, it is

subsequently regenerated with NaCl.132

The regeneration efﬂuent contains a
signiﬁcant amount of sugar as well as excess salt not used in the regeneration
process. This solution is usually considered waste, but NF membranes are
capable of recovering the salt as well as the sugar for future use. NF45
membranes from Dow Chemical/Filmtec have been investigated for this
separation, but they exhibit a water flux of only 0.1 m3/(m2 day bar) and have a
relatively high NaCl rejection of ~40°/o.133 The membranes in this dissertation

can increase ﬂux by a factor of 4 over commercial products as well as produce

sucrose rejections greater than 99% with 75% NaCl recovery.

19

The above applications, in concert with several others discussed in this
dissertation, demonstrate the power of PEMs as NF membranes. Interestingly,
simple variation of constituent polyelectrolytes, deposition conditions, and
capping layer results in ﬁlms with diverse retention and ﬂux properties. Utilizing
all of these parameters, the PEMs presented in Chapters 24 are capable of
effecting the selective separation of species ranging in size from salts to albumin

(MN 70000).

1.3 Outline of this Dissertation

Chapter 2 of this dissertation shows how PEMs can be tailored for speciﬁc
separations by varying ﬁlm composition. Liu and Bruening previously
demonstrated that in NF, PSS/poly(allylamine hydrochloride) (PAH) ﬁlms exhibit
glucose/sucrose selectivities in excess of 100. However, glucose passage in
those experiments was <3%, limiting the potential for saccharide separations.“
The data in Chapter 2 show that changing the polycation from PAH to the less
densely charged PDADMAC results in membranes that exhibit at least twice the
ﬂux of PSS/PAH films with glucose recoveries in excess of 40%.
PSS/PDADMAC ﬁlms also separate NaCl and sucrose with high recoveries of
salt and ﬂuxes 2—3 times greater than commercial membranes (Typical water
fluxes through commercial membranes are about 0.9 m3/(m2 day) at 5 ban“).
Additional experiments show that these transport properties are highly dependent
on top-layer choice, as PSS/PDADMAC ﬁlms capped with PDADMAC exhibit
rafﬁnose dialysis ﬂuxes 300 times greater than ﬁlms with PSS as the terminating

layer (the reason why this occurs is discussed in Chapter 3). Finally, the use of

20

exponentially growing PEMs with even lower charge densities demonstrates the
potential for high-resolution separations of the proteins myoglobin (MN 17000)
and bovine serum albumin (MN 70000).

The results in Chapter 2 clearly show that variables such as constituent
polyelectrolytes, top-layer charge, and deposition conditions directly affect
transport properties. In an effort to better understand the permeability of
polyelectrolyte ﬁlms, Chapter 3 aims at correlating transport through PEMs with
their swelling in aqueous and ethanolic solutions. PEM swelling has been
previously investigated in several published reports.‘“'145 Wong et al. showed
that swelling of PSS/PAH ﬁlms is proportional to relative humidity.138 Hiller and
Rubner reported that PSS/PAH ﬁlms exhibit unique swelling behavior as a
function of swellant pH,137 and Burke and Barrett observed that HA/PAH ﬁlms
swell as much as 8-times their dry thicknesses,145 one of the highest expansions
reported. Despite the various reports describing PEM swelling, however, very
few data correlate solvent uptake with transport.135'146

The in-situ ellipsometry experiments discussed in Chapter 3 demonstrate
that in water, permeability increases as ﬁlm swelling increases. For example,
HA/chitosan ﬁlms swell 4 times more than PSS/PAH coatings, and in NF
experiments, the HA/chitosan membranes permit a 250-fold greater fractional
passage of sucrose. Similar results are seen for diffusion dialysis experiments.
PEMs also display diverse swelling properties in ethanol, but transport rates do
not correlate with ethanol uptake, most likely due to a complex interplay between

hydrophobicity and ionic crosslinking.

21

Despite these advances in membrane performance by utilizing PEMs,
further improvement is always desirable. Previously reported PSS/PAH ﬁlms
exhibit glucose/sucrose selectivities 6 times greater than the PSS/PDADMAC
films described in Chapter 2,“ but PSS/PDADMAC ﬁlms have solution ﬂuxes 2-3
times greater than those through PSS/PAH ﬁlms and 10-fold greater glucose
recoveries. An “ideal” PEM for NF of sugars would combine the best features of
both PSS/PAH and PSS/PDADMAC ﬁlms to give high selectivity, ﬂux, and
glucose recovery. Chapter 4 describes two types of ﬁlm modiﬁcations that are
intended to improve performance: adsorption of the capping layer of PSS/PAH
ﬁlms from high ionic strength and deposition of highly selective PSS/PAH layers
on high ﬂux PSS/PDADMAC “gutter layers.” Hybrid PSS/PAH/PSS/PDADMAC
ﬁlms are capable of improving selectivities in sugar and NaCl/sucrose
separations compared to pure PSS/PDADMAC ﬁlms. Interestingly, simply
increasing the solution ionic strength during the deposition of the ﬁnal PSS-layer
in PSS/PAH ﬁlms results in high selectivities coupled with greater ﬂuxes than
pure PSS/PAH assemblies. This increased performance is likely because the
high ionic strength induces reorganization throughout the ﬁlm. In addition to
revisiting sugar/sugar and sugar/salt separations from Chapter 2, Chapter 4
presents results from the puriﬁcation of an idealized fermentation broth mimic to
show an example of a salt/neutral molecule separation where salts are the
rejected species.

Finally, Chapter 5 brings together several conclusions about this work.

Overall, improvements in NF are possible through the use of novel polymeric

22

films. Polyelectrolyte multilayers can serve as highly tunable, selective skin
layers for NF membranes and are capable of separating multiple types of

analytes. Chapter 5 also brieﬂy discusses possible future paths of research for

PEMs in the ﬁeld of separation science.

23

1.4 Works Cited

(1) Wu, 8.; Sun, A.; Liu, R. J. Chromatogr., A 2005, 1066, 243-247.

(2) Page, M. M.; Page, C. L.; Shaw, S. J.; Sawada, S. J. Sep. Sci. 2005, 28,
471-476.

(3) Chevron U.S.A, Reﬁning Process,
http://www.chevron.com/products/about/pascagoula/refininqprocess/proccrudes
html, Accessed 5-25-05

(4) Liu, K.; Tong, Z.; Liu, L.; Feng, X. J. Membr. Sci. 2005, 256, 193-201.

(5) Singh, A.; Hiwale, R.; Mahajani, S. M.; Gudi, R. D.; Gangadwala, J.;
Kienle, A. Ind. Eng. Chem. Res. 2005, 44, 3042-3052.

(6) Skouras, S.; Skogestad, S.; Kiva, V. AIChE Joumal 2005, 51, 1144-1157.

(7) Alsop, L.; Cowley, A. R.; Dilworth, J. R.; Donnelly, P. S.; Peach, J. M.;
Rider, J. T. Inorg. Chim. Acta 2005, 358, 2770-2780.

(8) Machura, B.; Kruszynski, R.; Jaworska, M. Polyhedron 2005, 24, 419-426.

(9) Bronfman, M.; Loyola, G.; Koenig, C. S. Anal. Biochem. 1998, 255, 252-
256.

(10) Lohaus, G.; Pennewiss, K.; Sattelmacher, B.; Hussmann, M.; Muehling, K.
H. Physiol. Plant. 2001, 111, 457-465.

(11) Wiig, H.; Aukland, K.; Tenstad, O. Am. J. Physiol. 2003, 284, H416-H424.
(12) Glavin, D. P.; Bada, J. L. Anal. Chem 1998, 70, 3119-3122.
(13) Glavin, D. P.; Schubert, M.; Bada, J. L. Anal. Chem 2002, 74, 6408-6412.

(14) Goryunkov, A. A.; Kuvychko, I. V.; Ioffe, l. N.; Dick, D. L.; Sidorov, L. N.;
Strauss, S. H.; Boltalina, O. V. J. Fluorine Chem. 2003, 124, 61 -64.

(15) Haitzer, M.; Aiken, G. R.; Ryan, J. N. Environ. Sci. Technol. 2002, 36,
3564-3570.

(16) Johnson, M. T.; Komane, L. L.; N'Da, D. D.; Neuse, E. W. J. Appl. Polym.
Sci. 2005, 96, 10-19.

(17) Malygin, A.; Parakhnevitch, N.; Karpova, G. Biochim. Biophys. Acta 2005,
1747, 93-97.

(18) McHowat, J.; Jones, J. H.; Creer, M. H. J. Chromatogr, B 1997, 702, 21-
32.

24

(19) Kasche, V. J. Biochem. Biophys. Methods 2001, 49, 49-62.

(20) Grancher, D.; Jaussaud, P.; Durix, A.; Berthod, A.; Fenet, B.; Moulard, Y.;
Bonnaire, Y.; Bony, S. J. Chromatogr., A 2004, 1059, 73-81.

(21) Mohsen, M. S.; Jaber, J. O.; Afonso, M. D. Desalination 2003, 157, 167.

(22) Matsubara, Y.; Iwasaki, K.-i.; Nakajima, M.; Nabetani, H.; Nakao, S.-i.
Biosci. Biotech. Bloch. 1996, 60, 421-428.

(23) Hutson, T. Int. Dairy Fed. 1998, 9804, 88-92.

(24) Vellenga, E.; Tragardh, G. Desalination 1998, 120, 211-220.

(25) Wedian, F.; Atkinson, D. B. Environ. Sci. Technol. 2002, 36, 4152-4155.
(26) Wedian, F .; Atkinson, D. B. Environ. Sci. Technol. 2003, 37, 4425-4428.

(27) Goulas, A. K.; Grandison, A. S.; Rastall, R. A. J. Sci. Food Agr. 2003, 83,
675-680.

(28) van Staden, J. F; Hattingh, C. J. Anal. Chim. Acta 1995, 308, 222-230.
(29) Shaposhnik, V. A.; Kesore, K. J. Membr. Sci. 1997, 136, 35-39.

(30) Slesarenko, V. V. Desalination 2003, 158, 303-311.

(31) Bazinet, L. Crit. Rev. Food Sci. Nutr. 2005, 45, 307-326.

(32) Bohdziewicz, J.; Bodzek, M.; Wasik, E. Desalination 1999, 121, 139-147.
(33) Krasemann, L.; Tieke, B. J. Membr. Sci. 1998, 150, 23-30.

(34) Meier-Haack, J.; Lenk, W.; Lehmann, D.; Lunkwitz, K. J. Membr. Sci.
2001, 184, 233-243.

(35) Toutianoush, A.; Tieke, B. Mat. Sci. Eng. C 2002, C22, 459-463.
(36) Pinnau, |.; Koros, W. J. J. Appl. Polym. Sci. 1991, 43, 1491-1502.
(37) Pinnau, |.; Koros, W. J. Ind. Eng. Chem. Res. 1991, 30, 1837-1840.

(38) Xiao, K. P.; Harris, J. J.; Park, A.; Martin, C. M; Pradeep, V.; Bruening, M.
L. Langmuir 2001, 17, 8236-8241.

(39) Yaroshchuk, A.; Staude, E. Desalination 1992, 86, 115-134.

25

(40) Bhattacharyya, D.; Williams, M. E.; Ray, R. J.; McCray, S. B., In
Membrane Handbook; Ho W.S.W, Sirkar, K.K., Eds. Van Nostrand Reinhold:
New York, 1992.

(41) Baker, R. W. Membrane Technology and Applications; 2nd ed.; John
Wiley: Chichester, England, 2004.

(42) Schafer, A. l.; Fane, A. G.; Waite, T. D. In Nanoﬁltration: Principles and
Applications; 1st ed.; Waite, T. D., Ed.; Elsevier: Oxford, UK, 2005.

(43) Osmonics Filtration Spectrum, http://www.osmonics.com/Iibrarv/zoom-
spoldhtm, Accessed 06-03-05

(44) Liu, X.; Bruening, M. L. Chem. Mater. 2004, 16, 351-357.

(45) Causserand, C.; Aimar, P.; Cravedi, J. P.; Singlande, E. Water Res. 2005,
39, 1594-1600.

(46) Boussahel, R.; Bouland, S.; Moussaoui, K. M.; Montiel, A. Desalination
2000, 132, 205-209.

(47) Koyuncu, I. J. Chem. Technol. Biotechnol. 2003, 78, 1219-1224.

(48) Kiefer, C. A.; Brinson, F.; Suratt, B. Maximizing the Resources: Solutions
for Pure and Plentiful Water, Membrane Technology Conference Proceedings,
Atlanta, GA, United States, Mar. 2-5, 2003 2003, 325-334.

(49) Krasemann, L.; Tieke, B. Mat. Sci. Eng. C-Bio S 1999, 08-09, 513-518.

(50) Harris, J. J.; Stair, J. L.; Bruening, M. L. Chem. Mater. 2000, 12, 1941-
1946.

(51) Stair, J. L.; Harris, J. J.; Bruening, M. L. Chem. Mater. 2001, 13, 2641-
2648.

(52) Dai, J.; Jensen, A. W.; Mohanty, D. K.; Erndt, J.; Bruening, M. L. Langmuir
2001, 17, 931-937.

(53) Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003,
19, 7038-7042.

(54) Mulder, M. Basic Principles of Membrane Technology; 2nd ed.; Kluwer
Academic Publishers: Dordrecht, The Netherlands, 1997.

(55) Bowen, W. R.; Welfoot, J. S. In Nanoﬁltration: Principles and Applications;
1st ed.; Waite, T. D., Ed.; Elsevier: Oxford, UK, 2005.

(56) Bowen, W. R.; Mukhtar, H. J. Membr. Sci. 1996, 112, 263-274.

26

(57) Bowen, W. R.; Mohammad, A. W.; Hilal, N. J. Membr. Sci. 1997, 126, 91-
105.

(58) Siegel, J.; Suhling, K.; Leveque-Fort, S.; Webb, S. E. D.; Davis, D. M.;
Phillips, D.; Sabhanrval, Y.; French, P. M. W. Rev. Sci. Instrum. 2003, 74, 182-
192.

(59) Tedeschi, C.; Mbhwald, H.; Kirstein, S. J. Am. Chem. Soc. 2001, 123,
954-960.

(60) Vankelecom, I. F. J.; De Smet, K.; Gevers, L. E. M.; Jacobs, P. A. In
Nanoﬁltration: Principles and Applications; 1st ed.; Waite, T. D., Ed.; Elsevier:
Oxford, UK, 2005.

(61) Ahmad, A. L.; Ooi, B. S. J. Membr. Sci. 2005, 255, 67-77.

(62) Song, Y.; Sun, P.; Henry, L. L.; Sun, B. J. Membr. Sci. 2005, 251, 67-79.

(63) Kim, S. H.; Kwak, S.-Y.; Suzuki, T. Environ. Sci. Technol. 2005, 39, 1764-
1770.

(64) Kwak, S.-Y.; Kim, C. K.; Kim, J.-J. J. Polym. Sci, Part B: Polym. Phys.
1996, 34, 2201-2208.

(65) Tomaschke, J. E., Amine monomers and their use in preparing
interfacially synthesized polyamide membranes for reverse osmosis and
nanoﬁltration, US. Patent 5,922,203, 1999

(66) Tomaschke, J. E., lnterfacial polymerization of bipiperidine-polyamide
membranes for reverse osmosis, nanoﬁltration, and seawater desalination, Eur.
Pat. Appl. 1060785. 2000

(67) Loeb, S.; Sourirajan, S. Advances in Chemistry Series 1963, 38, 117-132.

(68) Fan, S.-C.; Wang, Y.-C.; Li, C.-L.; Lee, K.-R.; Liaw, D.-J.; Huang, H.-P.;
Lai, J.-Y. J. Membr. Sci. 2002, 204, 67-79.

(69) Huang, J.; Cranford, R. J.; Matsuura, T.; Roy, C. J. Membr. Sci. 2004,
241, 187-196.

(70) Kinzer, K. E.; Lloyd, D. R.; Gay, M. S.; Wightman, J. P.; Johnson, B. C.;
McGrath, J. E. J. Membr. Sci. 1985, 22, 1-29.

(71) Tang, 8.; Xu, T.; Gong, M.; Yang, W. J. Membr. Sci. 2005, 248, 119-125.

(72) Liu, X.; Neoh, K. G.; Kang, E. T. Surf. Interface Anal. 2004, 36, 1037-
1040.

(73) Feng, C.; Shi, B.; Li, G.; Wu, Y. J. Membr. Sci. 2004, 237, 15-24.

27

(74) Stephan, A. M.; Renganathan, N. G.; Gopukumar, 8.; Dale, T. Mater.
Chem. Phys. 2004, 85, 6-11.

(75) Zhai, G.-Q.; Lei, Y.; Kang, E. T.; Neoh, K. G. Surf. Interface Anal. 2004,
36, 1048-1051.

(76) Jeon, J.-D.; Cho, B.-W.; Kwak, S.-Y. J. Power Sources 2005, 143, 219-
226.

(77) Zhai, G.; Kang, E. T.; Neoh, K. G. Macromolecules 2004, 37, 7240-7249.
(78) Lai, J. Y.; Chao, Y. C. J. Appl. Polym. Sci. 1990, 39, 2293-2303.

(79) Byun, H. S.; Burford, R. P.; Fane, A. G. J. Appl. Polym. Sci. 1994, 52, 825-
835.

(80) Ulbricht, M.; Oechel, A.; Lehmann, C.; Tomaschewski, G.; Hicke, H.-G. J.
Appl. Polym. Sci. 1995, 55, 1707-1723.

(81) Ulbricht, M.; Schwarz, H.-H. J. Membr. Sci. 1997, 136, 25-33.

(82) Frahn, J.; Malsch, G.; Schwarz, H.-H. J. Mater. Proc. Tech. 2003, 143-
144, 277-280.

(83) Hwang, T.-s.; Park, J.-w. Macromol. Res. 2003, 11, 495-500.

(84) Hilal, N.; Kochkodan, V.; Al-Khatib, l_.; Levadna, T. Desalination 2004,
167, 293-300.

(85) Geismann, C.; Ulbricht, M. Macromol. Chem. Phys. 2005, 206, 268-281.
(86) Ulbricht, M.; Yang, H. Chem. Mater. 2005, 17, 2622-2631.
(87) Esumi, K.; Masuda, A.; Otsuka, H. Langmuir 1993, 9, 284-287.

(88) Rignenbach, E.; Chauveteau, G.; Pefferkorn, E. Colloids Surf, A 1995, 99,
161-173.

(89) Esumi, K.; Goino, M. Langmuir1998, 14, 4466-4470.
(90) Michaels, A. S.; Miekka, R. G. J. Phys. Chem. 1961, 65, 1765-1773.

(91) Dauginet, L.; Duwez, A. S.; Legras, R.; Demoustier—Champagne, S.
Langmuir 2001, 17, 3952-3957.

(92) Murata, T.; Tanioka, A. J. Colloid Interface Sci. 1997, 192, 26-36.
(93) Nystrom, M. J. Membr. Sci. 1989, 44, 183-196.

28

(94) Lvov, Y.; Decher, G.; Mdhwald, H. Langmuir1993, 9, 481-486.
(95) Decher, G.; Lvov, Y.; Schmitt, J. Thin Solid Films 1994, 244, 772-777.

(96) Ferreira, M.; Cheung, J. H.; Rubner, M. F. Thin Solid Films 1994, 244,
806-809.

(97) Keller, S. W.; Johnson, S. A.; Brigham, E. S.; Yonemoto, E. H.; Mallouk, T.
E. J. Am. Chem. Soc. 1995, 117, 12879-12880.

(98) Caruso, F.; Niikura, K.; Furlong, D. N.; Okahata, Y. Langmuir 1997, 13,
3427-3433.

(99) Ageev, E. P.; Vikhoreva, G. A.; Gal’braikh, L. S.; Matushkina, N. N.;
Chaika, E. M.; Jaminsky, l. V. Vysokomol. Soedin., Ser. A Ser. B 1998, 40, 1198-
1204.

(100) Cassier, T.; Lowack, K.; Decher, G. Supramol. Sci. 1998, 5, 309-315.
(101) Kotov, N. A.; Magonov, S.; Tropsha, E. Chem. Mater. 1998, 10, 886-895.
(102) Cho, J.; Caruso, F. Macromolecules 2003, 36, 2845-2851.

(103) DeLongchamp, D. M.; Hammond, P. T. Langmuir 2004, 20, 5403-5411.

(104) Zhang, H.; Wang, Z.; Zhang, Y.; Zhang, X. Langmuir 2004, 20, 9366-
9370.

(105) Mwaura, J. K.; Pinto, M. R.; Witker, D.; Ananthakrishnan, N.; Schanze, K.
S.; Reynolds, J. R. Langmuir, ACS ASAP.

(106) Decher, G. In Multilayer Thin Films; Schlendff, J. 8., Ed.; Wiley-VCH:
Weinheim, 2003.

(107) Decher, G. Science 1997, 277, 1232-1237.

(108) Decher, 6.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210/211, 831-
835.

(109) Baur, J. W.; Rubner, M. F.; Reynolds, J. R.; Kim, S. Langmuir 1999, 15,
6460-6469.

(110) Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153-8160.

(111) Schoeler, B.; Kumaraswamy, G.; Caruso, F. Macromolecules 2002, 35,
889-897.

(112) Ladam, G.; Schaad, P.; Voegel, J. C.; Schaaf, P.; Decher, G.; Cuisinier, F.
Langmuir 2000, 16, 1249-1255.

29

(113) Shiratori, S. S.; Rubner, M. F. Macromolecules 2000, 33, 4213-4219.

(114) Yoo, D.; Shiratori, S. S.; Rubner, M. F. Macromolecules 1998, 31, 4309-
4318.

(115) Choi, J.; Rubner, M. F. Macromolecules 2005, 38, 116-124.

(116) Lavalle, P.; Gergely, C.; Cuisinier, F. J. G.; Decher, G.; Schaaf, P.;
Voegel, J. C.; Picart, C. Macromolecules 2002, 35, 4458-4465.

(117) McAloney, R. A.; Sinyor, M.; Dudnik, V.; Goh, M. C. Langmuir 2001, 17,
6655-6663.

(118) Richert, L.; Lavalle, P.; Payan, E.; Shu, X. Z.; Prestwich, G. D.; Stoltz, J.-
F .; Schaaf, P.; Voegel, J.-C.; Picart, C. Langmuir 2004, 20, 448-458.

(119) Lavalle, P.; Picart, C.; Mutterer, J.; Gergely, C.; Reiss, H.; Voegel, J.-C.;
Senger, B.; Schaaf, P. J. Phys. Chem. B 2004, 108, 635-648.

(120) Picart, C.; Lavalle, P.; Hubert, P.; Cuisinier, F. J. G.; Decher, G.; Schaaf,
P.; Voegel, J. C. Langmuir 2001, 17, 7414-7424.

(121) Kujawa, P.; Badia, A.; Winnik, F. M. Polym. Mater. Sci. Eng. 2004, 90,
192-193.

(122) Picart, C.; Mutterer, J.; Richert, L.; Luo, Y.; Prestwich, G. D.; Schaaf, P.;
Voegel, J. C.; Lavalle, P. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 12531-12535.

(123) Jin, W.; Toutianoush, A.; Tieke, B. Langmuir 2003, 19, 2550-2553.
(124) Lenk, W.; Meier-Haack, J. Desalination 2002, 148, 11-16.

(125) Toutianoush, A.; Krasemann, L.; Tieke, B. Colloids Surf, A 2002, 198-
200, 881-889.

(126) Schwarz, H. H.; Malsch, G. J. Membr. Sci. 2005, 247, 143-152.
(127) Sullivan, D. M.; Bruening, M. L. J. Membr. Sci. 2005, 248, 161-170.

(128) Lajimi, R. H.; Ben Abdallah, A.; Ferjani, E.; Roudesli, M. S.; Deratani, A.
Desalination 2004, 163, 193-202.

(129) Sullivan, D. M.; Bruening, M. L. Chem. Mater. 2003, 15, 281-287.
(130) Krasemann, L.; Tieke, B. Langmuir 2000, 16, 287-290.

(131) Goulas, A. K.; Kapasakalidis, P. G.; Sinclair, H. R.; Rastall, R. A.;
Grandison, A. S. J. Membr. Sci. 2002, 209, 321-335.

30

(132)

Hinkova, A.; Bubnik, Z.; Kadlec, P.; Pridal, J. Sep. Purif. Technol. 2002,

26,101-110.

(133)
(134)

Wang, X.-L.; Zhang, C.; Ouyang, P. J. Membr. Sci. 2002, 204, 271-281.
Sukhorukov, G. B.; Schmitt, J.; Decher, G. Berichte der Bunsen-

Gesellschaft 1996, 100, 948-953.

(135)
(136)
(137)
(138)

Harris, J. J.; Bruening, M. L. Langmuir 2000, 16, 2006-2013.

Kim, B.-S.; Vinogradova, O. l. J. Phys. Chem. B 2004, 108, 8161-8165.
Hiller, J. A.; Rubner, M. F. Macromolecules 2003, 36, 4078-4083.
Wong, J. E.; Rehfeldt, F.; Haenni, P.; Tanaka, M.; Klitzing, R. v.

Macromolecules 2004, 37, 7285-7289.

(139)

8878.

(140)
(141)
(142)
(143)
(144)

Ruths, J.; Essler, F.; Decher, G.; Riegler, H. Langmuir 2000, 16, 8871-

Carriere, D.; Krastev, R.; Schonhoff, M. Langmuir 2004, 20, 11465-11472.
Poptoshev, E.; Schoeler, B.; Caruso, F. Langmuir 2004, 20, 829-834.
Dubas, S. T.; Schlenoff, J. B. Langmuir 2001, 17, 7725-7727.

Tanchak, O. M.; Barrett, C. J. Chem. Mater. 2004, 16, 2734-2739.
Mendelsohn, J. D.; Yang, S. Y.; l-liller, J.; Hockbaum, A. l.; Rubner, M. F.

Biomacromolecules 2003, 4, 96-106.

(145)
(146)

Burke, 8. E.; Barrett, C. J. Biomacromolecules 2005, 6, 1419-1428.
Wang, L.; Schéenhoff, M.; Mohwald, H. J. Phys. Chem. B 2004, 108,

4767—4774.

31

Chapter 2

CONTROLLING THE NANOFILTRATION PROPERTIES OF MULTILAYER
POLYELECTROLYTE MEMBRANES THROUGH VARIATION OF FILM

COMPOSITION

SUMMARY

This chapter describes the use of a variety of polyelectrolyte multilayers
(PEMs) as selective skins in composite membranes for nanoﬁltration (NF) and
diffusion dialysis. Deposition of PEMs occurs through simple alternating
adsorption of polycations and polyanions, and separations can be optimized by
varying the constituent polyelectrolytes as well as deposition conditions. In
general, the use of polycations and polyanions with lower charge densities allows
separation of larger analytes. Depending on the polyelectrolytes employed, PEM
membranes can remove salt from sugar solutions, separate proteins, or allow
size-selective passage of speciﬁc sugars. Additionally, because of the minimal
thickness of PEMs, NF pure water ﬂuxes through these membranes typically
range from 1.5 to 3 m3/(m2 day) at 4.8 bar. Speciﬁcally, to separate sugars, we
employed poly(styrene sulfonate) (PSS)/poly(diallyldimethylammonium chloride)
(PDADMAC) ﬁlms, which allow 42% passage of glucose along with a 98%
rejection of rafﬁnose and a pure water ﬂux of 2.4 m3/(m2 day). PSS/PDADMAC
membranes are also capable of separating NaCl and sucrose (selectivity of ~10),
while high-ﬂux hyaluronic acid/chitosan membranes (pure water ﬂux of

5 m3/(m2 day) at 4.8 bar) may prove useful in protein separations.

32

2.1 Introduction

Nanoﬁltration (NF) is an important membrane-based separation technique
that is similar to reverse osmosis (R0), but the relatively high permeability of NF
membranes allows high-ﬂux separations at operating pressures that are much
lower than those used in R0.” The economic advantages of lower operating
pressures have recently led to applications of NF in several areas.“'12 Water
softening is probably the biggest NF operation, and plants have been designed to
treat as much as 40 million gallons of water per day.13 Other NF applications
include recovery of ammonium lactate from a fermentation solution,4 recycling of
NaCl from textile dyeing wastewater,9 and reduction of the salinity of seawater for
its use as a body washing solution.11 Development of stable membranes with
even higher fluxes and selectivities, as well as resistance to fouling could further
expand the utility of NF.

This chapter examines the potential of a new class of NF membranes
(polyelectrolyte multilayers) for the separation of different saccharides and
isolation of sugar from salt solutions. Because of the industrial importance of
such separations, several groups have investigated the performance of
commercial membranes in this area.”18 Wang et al. used NF45 membranes
from Dow Chemical to separate glycerol and several saccharides and obtained
rejections of 20%, 81%, 95%, and near 100% for glycerol, glucose, sucrose, and
rafﬁnose, respectively.14 Though this membrane could recover saccharides, the
high glucose rejection would pose a problem for sugar separations, and pure

water ﬂux was only 0.5 m3/(m2 day) at 4.8 bar.”19 Wang also succeeded in

33

separating NaCl from sugar but reported a 0.01 M NaCl rejection of ~40%, which
may lead to difﬁculties in recovering salt from a process stream.14 Similarly,
Vellenga and Tragardh used a D85 membrane from Desalination Systems to
separate NaCl and sucrose, but NaCl rejection was above 60%.18 Another
commercial membrane, the DS-5-DL from Osmonics, successfully separated
glucose from higher saccharides, but it allowed a water ﬂux of only
~1 m3/(m2 day) at 5 bar.15 Hence, membranes with ﬂuxes >1 m3/(m2 day) at a
pressure <5 bar and the ability to provide low (<40%) rejection of salts or glucose
while rejecting larger saccharides should provide performance improvements
over commercial systems. Of course, in addition to rejections and ﬂuxes, fouling
resistance and stability are vital to the application of any membrane.

To achieve high ﬂuxes, separation membranes generally contain a dense,
thin layer on a porous support. The “skin” layer provides selectivity, but its
minimal thickness still allows high ﬂux. In contrast, the porous support supplies
mechanical stability while adding little resistance to mass transport. Several
recent studies suggest that polyelectrolyte multilayer (PEM) ﬁlms are promising
candidates for “skin” layers in composite membranes.”29 PEM ﬁlms are
attractive for this role because of their deposition procedure, which simply
involves alternating adsorption of polycations and polyanions.30 This layer-by-
layer technique affords control over thickness through variation of the number of
adsorbed layers and allows formation of “skins” with thicknesses less than 50

26,31

nm Of equal importance, 3 wide range of polyelectrolytes are capable of

34

forming PEM ﬁlms,‘°"°"35

and judicious selection of constituent polyelectrolytes
should permit tailoring of ﬂux, selectivity, and possibly fouling rates.27'2"”'35'38

In spite of the versatility of PEM ﬁlms, NF studies of membranes
containing these materials have thus far focused primarily on
poly(vinylamine)/poly(vinyl sulfate) and poly(styrene sulfonate)/poly(allylamine
hydrochloride) (PSS/PAH) systems. Tieke and coworkers showed that 60-bilayer
poly(vinylamine)/poly(vinyl sulfate) ﬁlms exhibit sulfate rejections greater than
95%, but ﬂux through these membranes was relatively low due to the large
number of bilayers.26 Our group examined 4.5-bilayer PSS/PAH ﬁlms (the extra
0.5 bilayer indicates that P88 is the top layer in the ﬁlm) deposited on porous
alumina and also achieved 95% Na2804 rejection with appropriate deposition
conditions.27 Moreover, the use of 4.5-bilayer films affords ﬂuxes that are
comparable to or higher than those of state of the art commercial NF
membranes.‘°"27"?8'39 More recently, Liu and Bruening examined NF of methanol,
glycerol, glucose, and sucrose to probe the size-based selectivities of PSS/PAH-
containing membranes.28 While these membranes show glucose/sucrose
selectivities in excess of 100, the rejection of both sugars is large enough to
preclude the use of PSS/PAH ﬁlms for realistic saccharide separations.28 I

Building on previous work, this chapter demonstrates the control over NF
fluxes and rejections that is possible through varying constituent polyelectrolytes
in PEM-containing membranes. In an effort to lower the rejection of glucose

while still separating it from sucrose or rafﬁnose, we examined

PSS/poly(diallyldimethylammonium chloride) (PDADMAC) ﬁlms because

35

literature reports show that PSS/PDADMAC is much more permeable than
PSS/PAH in pervaporation and diffusion dialysis (DD) applications?“35
Optimization of PSS/PDADMAC films permits high-ﬂux (>2 m3/(m2 day))
glycerol/sucrose and glucose/rafﬁnose separations that would not be possible
with PSS/PAH. ln stark contrast to PSS/PDADMAC and PSS/PAH, PEM ﬁlms
prepared from hyaluronic acid (HA) and chitosan allow essentially quantitative
passage of glucose, sucrose, and rafﬁnose along with a pure water ﬂux of
5m3/(m2 day) at 4.8 bar. Even for myoglobin (MW 17000), HA/chitosan
membranes exhibit rejections <15%. These data are consistent with the
mechanism of formation of HA/chitosan ﬁlms which likely involves diffusion of
chitosan throughout the ﬁlm.40 However, HA/chitosan ﬁlms do show 97%
rejection of bovine serum albumin (Mw 67000). Thus, the use of different
polyelectrolytes should allow separation of molecules with molecular weights
ranging from 100 to ~30,000 along with pure water ﬂuxes from 2 to 5 m3/(m2 day)

at only 4.8 bar.

2.2 Experimental

Materials. Poly(styrene sulfonic acid) sodium salt (PSS, Mw 125000, Alfa
Aesar), poly(diallyldimethylammonium chloride) (PDADMAC, MW 100000-
200000, 20 wt% in water, Aldrich), NaCl (CCI), glycerol (anhydrous, CCI),
glucose (Aldrich), sucrose (Aldrich), rafﬁnose (Aldrich), myoglobin (Horse, Mw
17000, Aldrich), bovine serum albumin (BSA, MN 67000, Aldrich), hydrogenated

dextran (Mw 4000 - 6000, Polysciences), chitosan (“medium molecular weight”

36

(Mw 190000-310000 based on viscosity measurements by Aldrich) , 75-85%
deacetylated, Aldrich), hyaluronic acid (HA, MW 1.5 x 106 - 1.8 x 10“, sodium salt,
Fluka), and 3-mercaptopropanoic acid (MPA, Aldrich) were used as received.
The porous alumina supports (0.02 pm Whatman Anodisc ﬁlters) were UV/03
cleaned with the filtrate side up (Boekel UV-Clean model 135500) for 15 min
before ﬁlm deposition. Deionized water (Milli-Q, 18.2 M!) cm) was used for
membrane rinsing and preparation of polyelectrolyte solutions.

Film Deposition. A UV/Og-cleaned bare alumina support was oriented in
an O-ring holder so that only the feed side of the alumina contacted the
polyelectrolyte solutions. PSS/PDADMAC deposition started with immersion of
the support in an aqueous solution containing 0.02 M P88 in 0.1 or 0.5 M NaCl
for 3 min (molarities of polyelectrolytes are given with respect to the repeating
unit). The alumina support was rinsed with deionized water for 1 min before
exposure to 0.02 M PDADMAC in 0.1 or 0.5 M NaCl for 3 min, followed by
another water rinse for 1 min. This process was repeated until the target number
of bilayers was produced. To make PSS/chitosan films, the substrate was
immersed in 0.02 M P88 in 0.5 M NaCl for 3 min and then 0.005 M chitosan at
pH 2.2 for 5 minutes, with 1 min water rinses after deposition of each
polyelectrolyte. Hyaluronic acid (HA) and chitosan ﬁlms were prepared using a
literature procedure.40 Brieﬂy, we exposed the alumina support to alternating
solutions of 1 mg/mL HA and chitosan in 0.15 M NaCl adjusted to pH 5 with 0.1

M acetic acid with water rinsing between each deposition step.

37

Film Thickness Determinations. Ellipsometric thickness determinations
(J.A. Woollam model M-44 rotating analyzer ellipsometer) for PSS/PDADMAC
films were performed under ambient conditions (40-55% relative humidity) on Al-
coated Si wafers (200 nm Al on Si(100) wafers) using a previously reported
procedure.27'3"41 Formation of polyelectrolyte ﬁlms on Al wafers took place under
the same conditions as on alumina supports, and reported uncertainties in
thicknesses are the standard deviations of measurements on at least three
substrates. To estimate thicknesses of ﬁlms on porous alumina, images of
membrane cross sections were obtained with a Hitachi S4700 ll ﬁeld-emission
scanning electron microscope (SEM). Prior to imaging, membranes were
fractured under liquid nitrogen and sputter-coated (Pelco model SC-7 auto
sputter coater) on both sides with 5 nm of gold.

Transport Studies. DD through polyelectrolyte ﬁlms was studied using a
home-built apparatus with a membrane surface area of 2.3 cm2 that was
described previously.28 For sugar separations, the permeate side of the dialysis
cell was ﬁlled with deionized water, while the composition of the feed solution
varied slightly with the membrane type in order to achieve detectable amounts of
sugar in the receiving phase. Feed solutions contained 0.001 M glycerol,
glucose, sucrose and rafﬁnose when using bare alumina supports and alumina
coated with 3- and 4-bilayer PSS/PDADMAC ﬁlms deposited from 0.1 M NaCl, as
well as 4- and 5-bilayer ﬁlms deposited from 0.5 M NaCl. For all other
PSS/PDADMAC ﬁlms, the feed solution contained 0.005 M glycerol, glucose, and

sucrose and 0.015 M rafﬁnose. For sucrose/NaCl separations, the feed was 0.01

38

M NaCl, 0.001 M sucrose when using bare alumina supports and alumina coated
with 3- and 4—bilayer PSS/PDADMAC ﬁlms deposited from 0.1 M NaCl and 4-
and 5-bilayer PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl. For all other
PSS/PDADMAC ﬁlms, the feed solution was 0.01 M NaCl, 0.005 M sucrose.

A 2-mL sample was taken from the permeate dialysis cell every 10
minutes, and an equal volume was withdrawn from the feed side to ensure that
differing ﬂuid levels on each side of the membrane would not contribute to ﬂux.
To determine NaCl concentration, a conductivity measurement (Oakton CON 100
or Orion 115 conductivity meters) was taken at the same time as sampling, and
the feed cell conductivity was also measured after completion of the dialysis.
Conductivities were converted to concentrations using a calibration curve. The
sugar and glycerol concentrations were determined by liquid chromatography
(Dionex, DX-600, CarboPac PA-10 column, 100 mM NaOH mobile phase)
coupled with integrated amperometric detection (Dionex, ED-50). Flux values
were normalized by dividing them by the concentration of the probe molecule in
the feed at the end of the experiment. Both myoglobin and dextran were also
detected via integrated amperometric detection using the Dionex DX-600.
Bovine serum albumin was detected by UV absorption using a Perkin Elmer,
Lambda 40 spectrophotometer set to 198 nm.

NF was performed at a pressure of 4.8 bar‘27 with the cross-ﬂow apparatus
shown in Figures 2.1 and 2.2. Despite the bench scale of this apparatus, it

possesses the same components found in nearly all cross-ﬂow NF units.

39

._m. E :9:me Eo: noﬁmum m. 920: m_E. .cozmtommﬁ £5 5 now: ES ”2 05 9.36% 232.28 ”mm 9sz
KN .

x o

 

 

 

 

m=oo
cozmoncmz
Em... .14
soon
\/> 95:. r L

 

 

 

 

 

 

 

 

 

 

 

39% 23mm...”— LopoEBoE

C

O

 

4O

 

4 I) :nnxzu: 5
6 —

2:: 7

 

 

Figure 2.2: Schematic diagram of a “nanoﬁltration cell" from Figure 2.1 showing
the assembly of the cell components. From top to bottom: (1) and (2) are the
inlet and outlet ports, (3) are the threads, (4) is the rubber O-ring, (5) is the
membrane, (6) is the stainless steel frit that the membrane is placed on, (7) is the
cap that holds the frit, and (8) is the bottom part of the cell that screws into the
threads (3). This ﬁgure is adapted from Stanton et al.27 The arrow indicates ﬂow

direction through the cell.

41

Pressure is applied from a tank of Ar, and the pump provides the cross-ﬂow. A
ﬂow meter ensures that the cross-ﬂow rate is 18 mL/min, which is ~100 times the
permeate ﬂow rate and sufﬁcient to minimize concentration polarization,27 and
the pressure gauges verify that there is minimal pressure drop between
membranes ordered in series. Flow passes parallel to the membrane surface
(membrane surface area 1.5 cm2) while solution that passes through the
membrane exits the system where it is collected for analysis.

The feed solution for NF of sugars contained 0.001 M glycerol, glucose,
sucrose, and rafﬁnose. Salt/sugar NF separations employed a feed solution
containing 0.01 M NaCl and 0.001 M sucrose. After an 18 h equilibration time,
four samples were collected for times ranging from 15 to 40 min each, depending
on the ﬂux through the membrane, and the feed was sampled at the conclusion
of the experiment. Solution analysis occurred as described above for DD. Flux
measurements reported are for pure water passage through the membranes.
When sugars or salts were present, the solution ﬂux decreased by 5-25% for
PSS/PDADMAC and PSS/chitosan ﬁlms. All reported transport results are the

average of experiments with at least 3 different membranes.

2.3 Results and Discussion

To examine how polyelectrolyte structure affects separation properties, we
investigated three polyelectrolyte systems: PSS/PDADMAC, PSS/chitosan, and
hyaluronic acid (HA)/chitosan. (Figure 2.3 shows polyelectrolyte structures.)

The polyelectrolytes in these ﬁlms have a wide range of charge densities, which

42

5 w

N+ Cl-
/ \
'N + Poly(diallyldimethylammonium chloride)
503 a PDADMAC
Poly(styrene sulfonate)
PSS
OH
0“ coo-Na+
O O
0 H0 ’I/
o’I’n ’% O n
HO O
Chitosan Hyaluronic Acid
HA

Figure 2.3: Structures of polyelectrolytes used in this study.

should lead to varying degrees of ionic cross-linking. Based on previous ion-

dialysis and pervaporation studies by Tieke and coworkers?“35 ﬁ

lms with high
densities of ionic cross-links should resist swelling and provide high NF rejections
and selectivities. In contrast, lower cross-linking densities should result in
swollen membranes capable of separating larger analytes. Below, we briefly
discuss film characterization and then present separations that employ a series
of hydrophilic molecules (Table 2.1) to probe size-based selectivities in each of
the PEM systems. We most fully studied PSS/PDADMAC, as this is one of the

prototypical polyelectrolyte pairs. Additionally, we examined separation of NaCl

from sucrose to illustrate a potential application of PEM membranes.

43

Table 2.1: Molecular weights, aqueous diffusion coefﬁcients (D), and Stokes’

radii (rs) of the neutral molecules used in transport studies.“44

 

Solute Molecular Weight (g/mol) 0(10'9m'zs") rs(nm)

 

Glycerol 92 0.95 0.26
Glucose 180 0.69 0.36
Sucrose 342 0.52 0.47
Rafﬁnose 504 0.42 0.56

44

Film Characterization

To determine approximate film thicknesses, we initially deposited PEMs
on Al-coated Si wafers. The aluminum oxide that forms on the surface of the
coated wafers should be similar to the chemical structure of the porous alumina
supports used in NF and DD. In accord with literature results, ellipsometric
studies of PSS/PDADMAC ﬁlms showed that thickness increases approximately
linearly with the number of bilayers deposited (see Table 2.2 for thickness
values)”46

Cross-sectional SEM images corroborate ellipsometric thickness
measurements. Figure 2.4a shows the SEM image of a 5-bilayer
PSS/PDADMAC ﬁlm (deposited from 0.5 M NaCl) on alumina. The thickness of
the ﬁlm in the ﬁgure is ~30 nm, which is in good agreement with the ellipsometric
thickness of 34 nm (Table 2.2). SEM-determined thicknesses of 8.5-bilayer
HA/chitosan ﬁlms (~40 nm) also agree well with ellipsometric results (38 i 6 nm).
The agreement between ellipsometric data and SEM images suggests that there
is little effect of the SEM vacuum on ﬁlm thickness. However, thicknesses of

films in water may be substantially higher than in air.“7'“8

The deposition
conditions we employed for PSS/chitosan ﬁlms (pH 2.2 for chitosan) corroded the
AI-coated wafers, precluding the use of ellipsometry, but the SEM-determined
thicknesses of 4 and 4.5-bilayer ﬁlms were ~35 and ~45 nm, respectively. In
addition, top-down SEM images such as the one shown in Figure 2.4b were

taken for each type of membrane. These images demonstrate that all of the

polyelectrolyte ﬁlms used in this study are thick enough to cover the 20-nm pores

45

.mcoEucoo EmBEm um cosmmmE E:c_E:_m 85:33 Co .95 026 m5 :0 2.568% mEE Co 39:25. oEoEomazmo
6.928 3.6355 .8 mm:_m> x2: 9: ho 0:9 9: mm amazon 9m $52623 can

.cozmzcmocoo 329858 65 3 9.636 >2 822.6: mm; xzim

 

8308 «.3 8.343 336 :3 238 832 NHS mm,
:33 3033 to“? $8 23: 233 m3? 3% S
8:58 ma? 8.3m: 33d mi 2H8 832 :3 3t
Bonk: 8.28s 8.3m? $8 23$ 232 238 EN 4
8.58 NHS 8.3m: toned Ni SHE 83S moan: mm
8.3%.? 8.843 3.3m? 23.2 833 838 omnomm $2 2%.“wa
885% $205 $820 Es 2.2295?
$88.0 $88.0 >985 885% 898m $820 6820 Dawn“; 8963

 

nb_>=om_mw AL). rm N.Eo .085 «SSE um~=mE52

._Omz .2 md E0: Umzwoamn wEE O<§D<0Ewma w30tm> 53>

nmﬁoo mEEBm .393 3:95 mwocES cam .3903 .8836 ._o._mo>_m *0 GD .8 mm:_m> @2623 cam SSE ”NM 2an

46

 

Figure 2.4: SEM images of porous alumina coated with polyelectrolyte ﬁlms. (a)
cross section of a ~30 nm thick, 5-bilayer PSS/PDADMAC ﬁlm deposited from

0.5 M NaCl. (b) top-down view of a 4.5-bilayer PSS/chitosan ﬁlm.

47

on the alumina substrate, as the pores in the underlying support were not visible.
Diffusion Dialysis with PSS/PDADMAC Membranes

We initially performed DD of glycerol, glucose, sucrose, and rafﬁnose to
rapidly screen the size-based selectivity of PSS/PDADMAC ﬁlms as a function of
the number of deposited layers and deposition conditions. In these experiments,
polyelectrolyte-coated alumina membranes are positioned between equal
volumes of a feed solution containing the analytes of interest and a receiving
phase that is initially deionized water. The rate of transport across the
membrane is then determined by observing solute concentrations in the receiving
phase as a function of time. At short dialysis times, the receiving-phase
concentration is negligible when compared to the feed solution, so there is a
constant concentration gradient across the membrane that results in a linear
increase in receiving-phase concentration with time.

Table 2.2 shows probe ﬂux values and selectivities for DD with several
PSS/PDADMAC membranes deposited from 0.5 M NaCl. Films capped with
PSS (3.5, 4.5, and 5.5 bilayers) allow much lower ﬂuxes of sucrose and rafﬁnose
than ﬁlms capped with PDADMAC, and this results in a ~100-fold greater
glucose/rafﬁnose selectivity and a ~10-fold greater glucose/sucrose selectivity for
the PSS-terminated ﬁlms. The ﬂuxes of glycerol and glucose are high and much
less affected by the choice of capping layer. In fact, glycerol and glucose ﬂuxes
through PDADMAC-capped ﬁlms are only 20% less than those through bare

alumina, even with ﬁlm thicknesses as high as 34 nm. The lower selectivity of

48

films terminated in PDADMAC suggests that these coatings are more swollen
than those capped with PSS. This is consistent with swelling studies detailed in
Chapter 3 of this dissertation as well as NMR studies that indicate that water
molecules are more mobile when PSS/PDADMAC ﬁlms terminate with
PDADMAC rather than Pss.49

When we deposited PSS-capped ﬁlms from 0.1 M NaCl, we observed ﬂux
and selectivity values similar to those of the PSS-capped systems deposited from
0.5 M NaCl. The decrease in glucose/sucrose and glucose/rafﬁnose selectivity
due to PDADMAC capping was not as dramatic for ﬁlms prepared from 0.1 M
NaCl, but selectivities did decrease by a factor of 2 to 4 with PDADMAC as the
outer layer. Salt concentration in deposition solutions probably has more effect
on PDADMAC-capped ﬁlms than PSS-capped systems because swelling is
much larger in the former case. DD data for ﬁlms deposited from 0.1 M NaCl are
available in Table 2.3.
Nanoﬁltration with PSS/PDADMAC Membranes

Table 2.4 contains percent rejection values, selectivities, and water ﬂuxes
from NF experiments with several PSS/PDADMAC membranes and feed
solutions containing glycerol, glucose, sucrose, and raffinose. Percent rejection,
R, is deﬁned by Equation (2.1) where Cpem. and Creed are the solute
concentrations in the permeate and feed, respectively. Selectivity for solute A
over B is deﬁned by Equation (2.2), which can conveniently be expressed in
terms of rejections as shown. Percent rejection and selectivity were determined

after allowing the system to equilibrate for 18 h to achieve steady-state permeate

49

8:026:00 EmﬁEm Cm 00.88:. 8:52.20 00.0338 :0 :03. 02x0 05 :0 008008 mEE v6 88x25 0_=mEoma___mo
.858 _m:0_>_u:_ :9 828 x3: 05 00 0:0: 05 8 08:00 08 80533028 $9.06 58220“.

.:o=m::8:00 80:98:38 05 >0 0:636 >0 08:08.0: 8; lem

 

wave tonne foam: mdave whom mama mam? Nonﬁow n
0269 :5 8.33:. fouvd ram mHNn o :53 008.: Do
mﬁmw NdHNm modﬂmoe foam; Navm o 250 0 :53 003.9 o
oo Euoom Nam modﬁmne mdﬁmd NH: whom 959 908.3 0m
mavm deNm mod“ Fm; mdaﬁm 33 $50. 083 NOHVNF m
Am:_E:_<
madame; modﬁvme 3.03m; 0 Snow; 0N3: omaoww 838 .32
9mm: o
805:3. 80.0.5 80020 80:_ 0 80:0: 800: 28> mmwﬂEnW. 0.4—20450an
\80020 \80020 3:820 .t m w .0 _ .0 o E_v__n_ .f. 90225

 

ab_>:00_0w b-27wﬁE0L0E5 axe“. 08:05:02
.502 _2 to E0: 00:80:00 mEE O<§Q<on=mmd 80:9 55, 000.80 05820 0:0 m:_E:_m 80:00

0:00 330:: 80:50: 0:0 .8908 .8820 ._0:8>_m Lo m_m>_m_0 88.56 :00 828 33:00.8 0cm xsi ”0N 0.00.:

50

cor 0:0 .3 .5 ”0.03 33.80.00 8850:8820 :0.— 0020> .0.:0E:0me.U
00.. 0:0 .8 Km .2. .3 ”0:02, 33000.00 88:00:88.3 .02 003.8 0.820%?

:00 0.10 :2 002, “.2.m

 

 

 

mum? :6 0.0039 :5 mama 000.0 000 0.0004 0.0
0.42.3 0.03.0 F000; 0000 $50 0.0mm mam Youmd a. 2:0 006 n.
omﬂom mam deoN 0.03.00 mama :00 000? N089 0.0
000000 E": mdamd 5000.00 0.03.5 030 E"? 0.0“ F .N 0.0
.002 .2 md
700 me No a N... 3.000: 080 mawm E": 000 3.08:0. v 80.: 0020800
. . . . . . . 022.290.00.wa
900000 N00 m 000 F :00 n cam mm 0000 mam? m on? N m m
.300 E: E.
80:50m 80:05 00820 N 0
\ 0 0 0 0 2 O \ 0 0 0 02 0 220020 80:500. 8900.0. 8820 .8820 ax2n. 000.5 00...... E_.n.
L020>> 05n.
0.2.88 8 8.00.00

.08.: sin. 00 30.09 0 5.2, 00.08 0:.E20

00800 0:03 00300.9: .0030: .0 00:8 0 .0 8:02:80: E0: 00.03.200.00 0:0 .0003: .0003 08.80.01 .vN 0.00..

51

Cerm
R = [1——p—]x100% (2.1)

feed

CA,perm CB,feed =1OOOA’ _ RA
CA,feed CB,penn 100% _ RB

 

Selectivity = (2.2)

concentrations, and the feed volume was sufﬁcient that its concentration varied
only slightly during the experiment.

NF selectivities of PSS/PDADMAC ﬁlms are similar to those found in DD,
with the exception of glucose/rafﬁnose, which is 3- to 8-fold lower in NF with
PSS-capped ﬁlms. The lower glucose/rafﬁnose selectivity may reﬂect the fact
that transport in NF occurs primarily by convection, rather than diffusion.
Concentration polarization could also reduce the selectivity in NF, but higher
cross-ﬂow rates did not change the NF results, suggesting that diffusion layers
are not a major issue. The agreement between DD and NF selectivities for
glucose/sucrose and the high sucrose rejections also suggest that the effect of
concentration polarization is minimal. In any case, the NF glucose/rafﬁnose
selectivity of ~50 is still sufﬁcient for high quality separations.

There are large uncertainties in the glucose/rafﬁnose selectivities in both
NF and DD experiments. This error could be sourced from small differences in
the structure of individual membranes, as changes in pore size will strongly affect
selectivities when highly rejected species are involved. Slight variations in
rinsing pattern or in polyelectrolyte deposition conditions could result in pore size
irregularity. These inconsistencies in membrane manufacturing could be
mitigated by employing a “dipping robot” or other automated mechanical device

to make the ﬁlms, removing human error from this part of the experiment. It is

52

also possible that unseen defects in the alumina support translate to areas where
the PEM ﬁlm does not cover the pores completely. Depositing additional PEM
bilayers may mitigate these effects.

The percent rejection values reinforce the fact that PDADMAC-capped
films are more open than PSS-capped systems. The 4-bilayer ﬁlms deposited
from 0.5 M NaCl exhibited a rafﬁnose rejection of 36%, while PSS-capped ﬁlms,
regardless of deposition conditions, had rafﬁnose rejections ranging from 98-
99%. In spite of the fact that ﬁlms terminated with PDADMAC show low
rejections, water ﬂux through these ﬁlms is essentially the same as that through
comparable ﬁlms terminated with PSS. Perhaps this is a reﬂection of the already
high water ﬂuxes with these systems. Typical ﬂuxes through commercial NF
membranes are about 50% of the ﬂuxes through PSS/PDADMAC?"39
Membranes deposited from 0.1 M NaCl perform similarly to membranes
prepared from 0.5 M NaCl, except that the glucose/sucrose and glucose/rafﬁnose
selectivities of PDADMAC-capped ﬁlms are 3- and 5-fold higher for ﬁlms
deposited from 0.1 M NaCl (Table 2.5). This trend is consistent with DD data.
Additionally, pure water ﬂuxes are greater for ﬁlms deposited from 0.1 M NaCl,
probably because these coatings are half as thick as corresponding ﬁlms
deposited from 0.5 M NaCl.

The NF data in Table 2.4 demonstrate both the potential and the
limitations of PSS/PDADMAC ﬁlms capped with PSS for the separation of small
molecules. Glycerol rejection is less than 20% for all ﬁlms and, thus, high

glycerol recoveries can be achieved. In contrast, sucrose and rafﬁnose

53

:00 0.0 :0 :2 003 “.20

 

 

00000 0: N.000.N 0000.00 0.00000 0000 000 0.000.N 0.0
00 00 N00? N.000.N 0.00000 N000 0000 0000 0.000.N 0.0
0.0000 00050 00.00009 0.000. 5 N000 0000 0000 0.0000 0
80:500. 82030 80020 .300 NEXMEV 0:0>0..0

000:.t0m. 820:0 000020 .2820

000020 000020 22820 0x20220>> 0.:n. E_.u.

 

0.20000 .0... 8.00.00
.002 s. to 59. 80.880 0&0 0.088280 5.; 808 05520

0:28 0203 00.80.02 .080: :0 00:00 0 :0 8:02:80: Eo: 00:3:00.8 0:0 .0003: .202, 08:00.00. 0N 0.000

54

rejections are greater than 95%, so separation of glycerol from molecules larger
than sucrose is possible, but recovery of sucrose from solutions containing even
larger molecules is not. In the case of glucose, selectivity over rafﬁnose (~50) is
impressive, and selectivity over sucrose is as high as 14, but the ~60% rejection
of glucose may prohibit practical separations. To achieve higher recoveries
(lower rejections) of glucose, we investigated the polyelectrolyte systems
described below.
Nanoﬁltration with PSSIChitosan Membranes

Chitosan, the deacetylated form of the natural polymer chitin (Figure 2.3),
contains free amine groups and can therefore serve as a polycation in PEM ﬁlms.
Because the charge density in chitosan is slightly lower than in PDADMAC, we
thought that PSS/chitosan ﬁlms would have fewer ionic cross-links and, hence,
show lower rejections than PSS-capped PSS/PDADMAC. (PDADMAC and
chitosan contain one charge per 9 and 11 non-hydrogen atoms, respectively.)
For PSS—terminated ﬁlms, there is a lower rejection of 'all molecules by
PSS/chitosan than by PSS/PDADMAC (Table 2.4), but the difference between
the two types of ﬁlms is not large. For PSS/chitosan ﬁlms terminated with
chitosan (4 bilayers), glucose rejection is only 33%, so these ﬁlms could allow
relatively high glucose recoveries in glucose/rafﬁnose separations. In spite of
lower rejections, water ﬂuxes through PSS/chitosan membranes are similar to or
slightly lower than those through PSS/PDADMAC ﬁlms deposited from 0.5 M
NaCl, suggesting that the relatively high thickness of PSS/chitosan ﬁlms (~40

nm) may restrict ﬂux.

55

Nanoﬁltration with chitosan/hyaluronic acid ﬁlms

We also performed NF experiments with 8.5-bilayer hyaluronic acid
(HA)/chitosan polyelectrolyte membranes. This polyelectrolyte system is
intriguing because a number of literature reports show that HA/chitosan ﬁlms
grow exponentially with the number of deposited layers using the conditions we

employed.“°"”°‘52

Our ellipsometric measurements also conﬁrmed exponential
growth as ﬁlm thickness increased from ~40 to ~80 nm on going from an 8.5-
bilayer to a 10.5-bilayer ﬁlm. Such rapid increases in thickness occur because
polycations can diffuse readily through HA-containing ﬁlms,40 and thus, these
ﬁlms should be highly permeable to small analytes. Although HA/chitosan ﬁlms
are known to be somewhat heterogeneous,40 top-down SEM images of 4.5 and
8.5-bilayer ﬁlms indicate that deposition of 8.5 bilayers is more than sufﬁcient to
completely cover the porous supports.

The pure water ﬂux through 8.5-bilayer HA/chitosan membranes is
5210.6 m3/(m2 day), about twice that through PSS/PDADMAC and PSS/chitosan
films. Along with a high water ﬂux, the ﬁlms show minimal rejections <12% of
any of the previously mentioned neutral molecules. To probe the molecular
weight cutoff (MWCO, molecular weight at which rejection reaches 90%) of these
membranes, we performed NF with much larger solutes. The rejections of 4000-
6000 MW hydrogenated dextran (0.5 g/L) and myoglobin (125 mg/L, MN 17000)
were still less than 15%, but rejection of bovine serum albumin (250 mg/L, MN

67000) was 97%, indicating that the molecular weight cutoff (MWCO) for these

HA/chitosan membranes is between 17000 and 67000. Although HA/chitosan

56

films are much too permeable for small molecule separations, they might prove
useful in fractionating proteins. We should note that when performing NF with
myoglobin, the solution ﬂux through HA/chitosan membranes decreased by 80%,
and when using BSA the ﬂux decreased by 90%. Protein adsorption likely
reduces the ﬂux through the membranes.
Salt-sugar separations- Diffusion Dialysis

Salt/sugar separations are important for recovering the NaCl used to
regenerate ion-exchange columns.53 To quickly screen a large number of ﬁlms
for such separations, we examined DD with source-phase solutions containing
NaCl and sucrose. Our ultimate goal is to maximize the ﬂux of the NaCl through
the membrane while rejecting the neutral sucrose. Figure 2.5 shows how NaCl
flux and NaCl/sucrose selectivity vary with the number of bilayers for
PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl. Again, there is an obvious
dependence of membrane performance on the top layer composition, and
PDADMAC-terminated ﬁlms (4- and 5-bilayers) show selectivities similar to that
of a bare alumina support. The 3.5- and 4.5-bilayer systems gave a
NaCl/Sucrose selectivity of 40, 16 times better than the bare alumina value of 2.5
and an improvement of 60% over ﬁlms deposited from 0.1 M NaCl (see Figure
2.6 for data for ﬁlms deposited from 0.1 M NaCl). The 5.5 bilayer
PSS/PDADMAC ﬁlm deposited from 0.5 M NaCl suffered from lower NaCl ﬂuxes,

which reduced selectivity.

57

300 ~

250

200 ,

150

 

100 -

NaCl Flux (nmol*cm‘2*s'1IM)

50,

o - .M, , -,- 0 -,___, o
3.5 4 4.5 5 5.5

Bilayers PSS/PDADMAC

Figure 2.5: NaCl ﬂux and NaCl/Sucrose selectivity from DD through porous
alumina coated with PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl. The
arrows indicate which y-axis pertains to each data set. Flux was normalized by

dividing by the source-phase concentration.

58

 

NaCl/Sucrose Selectivity

300

 

 

- 250
g a
'7‘” Z;
«3' 200 - 8
E 2
.2 8
g 150 3
5 2
0
X 3
2 m
u. 100 3
- 0
% . z
z 50 ~ ‘
- 5
0 ~ --~~~-——- , ,- ~ - .-_ - -- - ”—0 o
5 5.5 6 6.5 7

Bilayers PSS/PDADMAC

Figure 2.6: NaCl ﬂux and NaCl/Sucrose selectivity from diffusion dialysis through
porous alumina coated with PSS/PDADMAC ﬁlms deposited from 0.1 M NaCl.
The arrows indicate which y-axis pertains to each data set. Flux was normalized

by dividing by the source-phase concentration.

59

Salt-sugar separations- Nanoﬁltration

Table 2.6 shows the percent rejection, water ﬂux, and selectivity for NF of
a NaCl/sucrose solution using PSS/PDADMAC ﬁlms. The results show trends
similar to those seen with DD. Films capped with PDADMAC and deposited from
0.5 M NaCl were so open that they provided a NaCl/sucrose selectivity of only
1.3, which is too low to attempt any reasonable separation. In contrast, the
selectivity of 10, NaCl rejection of 30%, and water ﬂux of 2.4 m3/(m2 day) at
4.8 bar for the 3.5-bilayer PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl may
be attractive for NaCl/sucrose separations. PSS-terminated ﬁlms deposited from
0.1 M NaCl show even lower NaCl rejections along with high ﬂux and >90%
sucrose rejections, and such ﬁlms are even more attractive for salt/sugar
separations.

Salt-sugar separations are complicated by charge-exclusion of cations or
anions. Because most cations and anions are relatively small, salt rejection is
likely to be influenced more by surface charge than bulk ﬁlm density. Decreasing
surface charge by lowering the salt concentration in deposition solutions appears
to be one way to reduce NaCl rejections.54
Comparison of PEM Systems

All of the rejection data suggest that the use of polyelectrolytes with high
charge densities results in heavily ionically cross-linked PEM membranes that
exhibit high rejections of neutral molecules. Glycerol, glucose, and sucrose
rejections decrease in the order PSS/PAH28>PSS/PDADMAC

>PSS/chitosan>HA/chitosan, which is also the decreasing order of

60

Table 2.6:

Rejections, water ﬂuxes, and selectivities from nanoﬁltration of

sucrose and NaCl by porous alumina coated with PSS/PDADMAC ﬁlms.

 

 

 

Rejection (%L Selectivity
Pure Water NaCl/
Film Type Bilayers Flux“ NaCl Sucrose Sucrose
(m3/(m2 day))
3.5 24:05 28:4 92.5:0.8 10:1
PSS/PDADMAC
Deposited from 4 2.5:0.7 14:5 32:4 1.27:0.07
0.5 M NaCl
4.5 21:03 40:4 93.8:0.8 10:1
5.5 16:02 41 :3 92:2 7:2
5 33:08 21:2 81:4 4.3:0.7
PSS/PDADMAC
Deposited from 5.5 26:05 20:2 94:1 15:5
0'1 M ”30' 6.5 2300.4 22:2 9101 9:1

aNF was run at 4.8 bar.

61

polyelectrolyte charge densities. Previous DD data for poly(acrylic acid)/PAH
ﬁlms suggest that this system would have higher rejections than even
PSS/PAH,28 further conﬁrming the trend of rejection versus polyelectrolyte
charge density. This chapter demonstrates the utilization of this trend to prepare
polyelectrolytes capable of separating molecules with different size ranges. For
example, the high rejections of PSS/PAH are not practical for sugar separation,
but the use of PSS/PDADMAC may be. In the case of protein separations,
HA/chitosan may prove useful while the other polyelectrolyte systems we have
tested would not.

Although the PEM films presented in this chapter are quite attractive for
sugar, salt/sugar, and even protein separations, many issues in the utilization of
polyelectrolyte membranes for practical separations must still be addressed.
Commercially employed NF membranes are often spiral-wound cartridges, where
the membrane is wrapped with a flow spacer in a cylindrical conﬁguration.55 This
allows a high degree of membrane surface area to be conﬁned in a small
volume. Unfortunately, the PEMs in this dissertation are deposited on alumina,
which is not compatible with spiral-wound technology. To counter this problem,
recent research has focused on exploring NF with PEMs deposited on polymeric
membrane supports.56

Questions about the effects of both extreme solution conditions and
fouling on PEMs also need to be answered. The highly charged surface of PEMs
could attract oppositely charged species that may affix to the ﬁlm, obstructing the

pores over time (fouling). Additionally, since electrostatic interactions bind the

62

film together, exposure to high ionic strength and pH extremes may decompose
the ﬁlm, reducing performance. Long term fouling and stability studies are both
important steps that must be performed before PEM assemblies are scaled up to

the industrial level.

2.4 Conclusions

Variation of the constituent polyelectrolytes in PEM membranes allows
tailoring of NF properties. Rejection of neutral molecules increases with an
increasing charge density on the polyelectrolytes that constitute the PEM
membrane, and this effect is large enough to allow synthesis of polyelectrolyte
membranes with MWCOs ranging from 100 (PSS/PAH) to >20,000
(HA/chitosan). Moreover, the minimal thickness of polyelectrolyte membranes
allows NF to occur at ﬂuxes of more than 2 m3/(m2 day). DD and NF data with
PSS/PDADMAC membranes also demonstrate that rejection, ﬂux, and selectivity
strongly depend on deposition conditions and which polyelectrolyte terminates

the membrane.

2.5 Acknowledgment
We thank the Division of Chemical Sciences in the Ofﬁce of Basic Energy
Sciences of the Department of Energy for ﬁnancial support of this work. We also

thank Pall Corporation for ﬁnancial support.

63

2.6 Works Cited

(9)
(10)

(11)

(14)
(15)

(16)

(17)

Yaroshchuk, A.; Staude, E. Desalination 1992, 86, 115-134.
Bohdziewicz, J.; Bodzek, M.; Wasik, E. Desalination 1999, 121, 139-147.

Bhattacharyya, D., et al, In Membrane Handbook; Ho W.S.W, Sirkar, K.K.,
Eds.; Van Nostrand Reinhold; New York, 1992.

Kang, S. H.; Chang, Y. K.; Chang, H. N. Biotechnol. Progr. 2004, 20, 764-
770.

Aydogan, N.; Gurkan, T.; Yilmaz, L. Separ. Sci. Technol. 2004, 39, 1059-
1072.

Li, S.-L.; Li, C.; Liu, Y.-S.; Wang, X.-L.; Cao, Z.-A. J. Membr. Sci. 2003,
222, 191-201.

Van der Bruggen, B.; Daems, B.; Wilms, D.; Vandecasteele, C. Sep. Purif.
Technol. 2001, 22/23, 519-528.

Akbari, A.; Remigy, J. C.; Aptel, P. Chem. Eng. Process. 2002, 41, 601-
609.

Koyuncu, l.; Topacik, D.; Yuksel, E. Sep. Pun'f. Technol. 2004, 36, 77-85.

Sungpet, A.; Jiraratananon, R.; Luangsowan, P. Desalination 2004, 160,
75-81.

Pontie, M.; Diawara, C.; Rumeau, M.; Aureau, D.; Hemmery, P.
Desalination 2003, 158, 277-280.

Favre-Reguillon, A.; Lebuzit, G.; Foos, J.; Guy, A.; Draye, M.; Lemaire, M.
Ind. Eng. Chem. Res. 2003, 42, 5900-5904.

Kiefer, C. A.; Brinson, F.; Suratt, B. Maximizing the Resources: Solutions
for Pure and Plentiful Water, Membrane Technology Conference
Proceedings, Atlanta, GA, United States, Mar. 2-5, 2003 2003, 325-334.

Wang, X.-L.; Zhang, C.; Ouyang, P. J. Membr. Sci. 2002, 204, 271-281.

Goulas, A. K.; Kapasakalidis, P. G.; Sinclair, H. R.; Rastall, R. A.;
Grandison, A. S. J. Membr. Sci. 2002, 209, 321-335.

Goulas, A. K.; Grandison, A. S.; Rastall, R. A. J. Sci. Food Agr. 2003, 83,
675-680.

Gyura, J.; Seres, Z.; Vatai, G.; Molnar, E. B. Desalination 2002, 148, 49-
56.

64

(18)
(19)

(20)

(21)
(22)
(23)
(24)
(25)

(26)
(27)

(28)
(29)

(30)
(31)

(32)

(33)

(34)
(35)
(35)

(37)

Vellenga, E.; Tragardh, G. Desalination 1998, 120, 211-220.

Literature value for water ﬂux is reported as a function of pressure. This
value is based on the same pressure used in our experiments, 70 psi.

Stroeve, P.; Vasquez, V.; Coelho, M. A. N.; Rabolt, J. F. Thin Solid Films
1996, 284-285, 708-712.

Levaesalmi, J.-M.; McCarthy, T. J. Macromolecules 1997, 30, 1752-1757.
Kotov, N. A.; Magonov, S.; Tropsha, E. Chem. Mater. 1998, 10, 886-895.
Schlenoff, J. B.; Ly, H.; Li, M. J. Am. Chem. Soc. 1998, 120, 7626-7634.
Krasemann, L.; Tieke, B. Langmuir 2000, 16, 287-290.

Dubas, S. T.; Farhat, T. R.; Schlenoff, J. B. J. Am. Chem. Soc. 2001, 123,
5368-5369.

Jin, W.; Toutianoush, A.; Tieke, B. Langmuir 2003, 19, 2550-2553.

Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003,
19, 7038-7042.

Liu, X.; Bruening, M. L. Chem. Mater. 2004, 16, 351-357.

Meier-Haack, J.; Lenk, W.; Lehmann, D.; Lunkwitz, K. J. Membr. Sci.
2001, 184, 233-243.

Decher, G. Science 1997, 277, 1232-1237.

Harris, J. J.; Stair, J. L.; Bruening, M. L. Chem. Mater. 2000, 12, 1941-
1946.

Lvov, Y.; Ariga, K.; lchinose, |.; Kunitake, T. J. Am. Chem. Soc. 1995, 117,
6117-6123. ,

Liu, R. C. W.; Morishima, Y.; Winnik, F. M. Macromolecules 2001, 34,
9117-9124.

Hollman, A. M.; Bhattacharyya, D. Langmuir 2004, 20, 5418-5424.
Krasemann, L.; Tieke, B. Mat. Sci. Eng. C—Bio S 1999, C8-CQ, 513-518.

Thierry, B.; Winnik, F. M.; Merhi, Y.; Silver, J.; Tabrizian, M.
Biomacromolecules 2003, 4, 1564-1571.

Berg, M. C.; Yang, S. Y.; Hammond, P. T.; Rubner, M. F. Langmuir 2004,
20, 1362-1368.

65

(38)
(39)

(40)

(41)
(42)

(43)
(44)
(45)

(46)

(47)
(48)
(49)

(50)

(51)

(52)

(53)

(54)
(55)

(55)

Carroll, T.; Booker, N. A.; Meier-Haack, J. J. Membr. Sci. 2002, 203, 3-13.

Hutchinson, R. A.; Beuermann, S.; Paquet, D. A., Jr.; McMinn, J. H.;
Jackson, C. Macromolecules 1998, 31 , 1542-1547.

Richert, L.; Lavalle, P.; Payan, E.; Shu, X. Z.; Prestwich, G. D.; Stoltz, J.-
F .; Schaaf, P.; Voegel, J.-C.; Picart, C. Langmuir 2004, 20, 448-458.

Harris, J. J.; Bruening, M. L. Langmuir 2000, 16, 2006-2013.

Derlacki, Z. J.; Easteal, A. J.; Edge, A. V. J.; Woolf, L. A.; Roksandic, Z. J.
Phys. Chem. 1985, 89, 5318-5322.

Creeth, J. M.; Woolf, L. A. J. Phys. Chem. 1963, 67, 2777-2780.
Boyle, M. D. P.; Gee, A. P.; Borsos, T. J. Immunol. 1979, 123, 77-82.

Lavalle, P.; Gergely, C.; Cuisinier, F. J. G.; Decher, G.; Schaaf, P.;
Voegel, J. C.; Picart, C. Macromolecules 2002, 35, 4458-4465.

Ramsden, J. J.; Lvov, Y. M.; Decher, G. Thin Solid Films 1995, 254, 246-
251.

Tanchak, O. M.; Barrett, C. J. Chem. Mater. 2004, 16, 2734-2739.
Hiller, J. A.; Rubner, M. F. Macromolecules 2003, 36, 4078-4083.

McCormick, M.; Smith, R. N.; Graf, R.; Barrett, C. J.; Reven, L.; Spiess, H.
W. Macromolecules 2003, 36, 3616-3625.

Lavalle, P.; Picart, C.; Mutterer, J.; Gergely, C.; Reiss, H.; Voegel, J.-C.;
Senger, B.; Schaaf, P. J. Phys. Chem. B 2004, 108, 635-648.

Picart, C.; Lavalle, P.; Hubert, P.; Cuisinier, F. J. G.; Decher, G.; Schaaf,
P.; Voegel, J. C. Langmuir 2001, 17, 7414-7424.

Kujawa, P.; Badia, A.; Winnik, F. M. Polym. Mater. Sci. Eng. 2004, 90,
192-193.

Hinkova, A.; Bubnik, Z.; Kadlec, P.; Pridal, J. Sep. Purif. Technol. 2002,
26, 101-110.

Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153-8160.

Baker, R. W. Membrane Technology and Applications; 2nd ed.; John
Wiley: Chichester, England, 2004.

Malaisamy, R.; Bruening, M. L. Langmuir 2005, Accepted for Publication.

66

Chapter 3

CORRELATION OF THE SWELLING AND PERMEABILITY OF

POLYELECTROLYTE MULTILAYER FILMS

SUMMARY

Alternating adsorption of polycations and polyanions on porous supports
yields a variety of size-selective membranes whose swelling and transport
properties depend on constituent polyelectrolytes, capping layer choice
(polycation or polyanion), and deposition conditions. This chapter shows that in
aqueous experiments, ellipsometrically determined swelling percentages
correlate well with nanoﬁltration (NF) rejections and diffusion dialysis ﬂuxes. For
example, hyaluronic acid (HA)/chitosan ﬁlms swell 4 times more than
poly(styrene sulfonate) (PSS)/poly(allylamine hydrochloride) coatings, and in NF
experiments, the HA/chitosan membranes permit a 250-fold greater fractional
passage of sucrose. In general, ﬁlms prepared from polyelectrolytes with a high
charge density show low swelling and slow solute transport, presumably because
of a high degree of ionic cross-linking. In the case of PSS/poly(diallyldimethyl-
ammonium chloride) (PDADMAC), PDADMAC-capped ﬁlms can swell 4-fold
more than their PSS-terminated counterparts, and as would be expected,
glucose and sucrose transport rates in diffusion dialysis are about 1.7- and 17-
fold more, respectively, when these ﬁlms end in PDADMAC. Polyelectrolyte
multilayers also exhibit wide-ranging swelling properties in ethanol, but transport

rates do not correlate with ethanol uptake. In this solvent, the density of ionic

67

cross-links and ﬁlm hydrophobicity likely exert opposite effects on swelling, which

could complicate the correlation between swelling and transport.

3.1 Introduction

Polyelectrolyte multilayers (PEMs) are attractive as selective ﬁlms in
applications such as pervaporation,”3 nanoﬁltration (NF),7'11 and
encapsulation.”21 Their synthesis, which simply involves alternating adsorption
of polycations and polyanions,22 yields ultra-thin (<50 nm) coatings capable of

9,11,23

allowing high ﬂuxes. Additionally, many materials can be used to form

PEMs,‘°"2“‘26 and judicious selection of component polyelectrolytes and deposition
conditions results in ﬁlms with a wide range of permeation properties.2'3'9'1“2527'28
In the above applications, PEMs are in contact with solvent, reducing the
relevance of physiochemical measurements in the “dry” state. To better
understand the permeability of polyelectrolyte ﬁlms, this chapter aims at
correlating transport through PEMs with their swelling in both aqueous and
ethanolic solutions.

Several groups have already examined the swelling of individual PEMs in
solvents. Neutron reﬂectometry studies suggest that poly(styrene sulfonate)
(PSS)/protonated poly(allylamine) (PAH) ﬁlm swelling is a function of the capping
layer, as ﬁlms capped with PAH swell 25% less than those capped with PSS
(40% versus 30% D20 as a function of capping layer).29 Wong et al. observed

the same outer-layer dependence for PSS/PAH ﬁlms when they performed

ellipsometric swelling experiments in 99% relative humidity.30 Harris and

68

Bruening found that immersion of [PSS/PAHlto ﬁlms in pH 3.2- and 6.3-buffered
water solutions results in a thickness increase of 40% relative to ambient
humidity conditions, while exposure of these ﬁlms to pH 10 buffers yields even
greater swelling followed by ﬁlm delamination.31 Other studies showed that
PSS/PAH swelling is affected by deposition pH and ionic strength as well as
swellant pH.32'33 Though most swelling research has been performed on
PSS/PAH ﬁlms,”36 Schlenoff and Dubas demonstrated that water uptake in
poly(acrylic acid) (PAA)/poly(diallyldimethylammonium chloride) (PDADMAC),
PSS/PDADAMC, and PSS/PAH ﬁlms is a strong function of the swellant ionic
strength and that PAA/PDADMAC and PSS/PDADMAC ﬁlms swell more in water
than PSS/PAH coatings.37 Two recent papers showed that the swelling of
PAA/PAH ﬁlms depends on both deposition conditions and pretreatments.38‘39
Burke and Barrett also found that in some cases, PAH/hyaluronic acid (HA) ﬁlms
are capable of 800% swelling.40 Excepting Schlenoff and Dubas’ work,37
however, there have been no systematic studies of how “dry” versus water-
swollen PEM thicknesses differ with variables such as constituent
polyelectrolytes and capping-layer choice.

In contrast, several studies show that the permeability of PEMs varies
dramatically with their composition. Tieke and coworkers found that transport
rates in pervaporation and diffusion dialysis (DD) through PEMs generally
decrease as the charge density in the ﬁlm increases.2 Presumably, greater
charge density on the polyelectrolytes results in more ionic cross-linking, less

2,8.9,11

swelling, and lower permeabilities. As not all prospective applications of

69

PEMs are in water, swelling in other solvents is important as well. PEM swelling
should depend on both the solvent and the hydrophobicity of the polyelectrolytes.
Poptoshev et al. demonstrated that exposure of poly(ethyleneimine)
(PE|)/PSS/PAH ﬁlms to a solution of >40% ethanol in water collapses these ﬁlms
to essentially their dry thickness,36 but another study suggests that thicker
PSS/PAH ﬁlms undergo only a 5% thickness reduction when immersed in
ethanol rather than water.41 Regardless, the permeability of solutes through
PEMs will likely be signiﬁcantly different in ethanol than in water.

Few publications link polyelectrolyte swelling and transport,31"”“4 and
many specialize in PEM capsules41 or materials formed by precipitation of

polyanion/polycation complexes,”44

rather than layer-by-layer adsorption. In this
work, we attempt to directly correlate swelling with permeability data for three
PEM systems: PSS/PAH, PSS/PDADMAC, and HA/chitosan. These systems
were selected in part because they exhibit a wide range of transport properties,

as reported previously.9'11

For example HA/chitosan membranes have a
molecular weight cutoff (MWCO, solute molecular weight required to achieve
90% rejection in NF) of >17,ooo, while PSS/PAH ﬁlms have a cho of 200.811
Consistent with these transport data, this work shows that the percent sWelIing of
HA/chitosan ﬁlms in water is 4-fold greater than that for PSS/PAH. Remarkably,
similarly striking differences in swelling and transport properties occur on going
from a PSS-capped to a PDADMAC-capped PSS/PDADMAC ﬁlm. Below, we

examine swelling and transport as a function of ionic strength, capping layer

composition, and swelling solvent.

70

3.2 Experimental

Materials. Poly(styrene sulfonic acid) sodium salt (Mw 125000, Alfa
Aesar), poly(diallyldimethylammonium chloride) (MW 100000-200000, 20 wt% in
water, Aldrich), poly(allylamine hydrochloride) (MN 70000, Aldrich), chitosan
(“medium molecular weight” (Mw 190000-310000 based on viscosity
measurements by Aldrich), 75-85% deacetylated, Aldrich), hyaluronic acid (Mw
1500000-1800000, sodium salt, Fluka), polyethyleneimine (MN 25000, Aldrich),
NaCl (CCI), glycerol (anhydrous, CCI), glucose (Aldrich), sucrose (Aldrich),
rafﬁnose (Aldrich), hydrogen peroxide (30%, Jade Scientiﬁc), sulfuric acid
(concentrated, CCI) and ethanol (Pharmco) were used as received. Deionized
water (Milli-Q, 18.2 MO cm) was used for membrane rinsing, preparation of
polyelectrolyte solutions, and aqueous swelling experiments.

Film Deposition. For swelling experiments, ﬁlms were prepared on
pieces of silicon wafers (Si(100), Silicon Quest International) that were ﬁrst
cleaned in a 3:1 solution of concentrated sulfuric acid and hydrogen peroxide.
(Caution! This solution reacts violently with organic compounds and should be
stored in slightly open containers!) Following copious rinsing with water, the
wafers were dried in a stream of N2 and then cleaned with UV/03 (Boekel UV-
Clean model 135500) for 15 minutes. The cleaned silicon was immersed in 1
mg/mL PEI at pH 9 for 15 minutes to establish a dense, positively charged

layer,45

and ﬁlms were then built on this precursor layer. For NF and DD
experiments, porous alumina supports (0.02 pm Whatman Anodisc ﬁlters) were

also UV/03 cleaned for 15 minutes, but because these supports are positively

71

charged below pH 9,46 no precursor PEI layer was necessary. The cleaned
alumina membranes were subsequently placed in an o-ring holder so that the
concentrate side of the alumina membrane contacted the deposition solution.

Synthesis of PSS/PDADMAC. ﬁlms began with a 3-min immersion of the
substrate in an aqueous solution containing 0.02 M PSS (concentrations of
polyelectrolytes are given with respect to the repeat unit) and 0.1 M or 0.5 M
NaCl. The substrate was then rinsed with deionized water for 1 min and dipped
in a 0.02 M solution of PDADMAC in 0.1 M or 0.5 M NaCl for 3 min. The sample
was then rinsed again with deionized water for 1 min, and this entire process was
repeated until the desired number of bilayers was deposited. PSS/PAH ﬁlms
were deposited using the same polyelectrolyte concentrations (no pH
adjustment) and deposition times, except that ﬁlms were only deposited from 0.5
M NaCl. HA/chitosan ﬁlms were deposited using 5 min immersions of the PEI-
coated silicon slides or bare alumina supports in pH 5 solutions containing 0.15
M NaCl and 1 mg/mL polyelectrolyte, with 1 min rinses with 0.15 M NaCl at pH 5
after polycation and polyanion adsorption. Films were rinsed with pure water and
dried with N2 only after all layers were deposited.

Ellipsometry. Ellipsometric thicknesses of the SiOz layers on Si wafers
were ﬁrst determined assuming literature values for the refractive indices of Si
and Si02 at the 44 wavelengths of the ellipsometer (J.A. Woollam model M-44
rotating analyzer ellipsometer, 75° angle of incidence) between 414.0 nm and
736.1 nm. After coating of these wafers, ﬁlm thicknesses under nitrogen (<5%

relative humidity (RH)), water, or ethanol were obtained using a home-built cell

72

with glass windows. The ellipsometric thicknesses and refractive indices of each
type of ﬁlm were determined at three different points per wafer on three different
wafers, and the reported results are the averages and standard deviations of
these values. Film thicknesses were also obtained in 55% RH (ambient) for
comparison to AFM data. Optical constants of water as a function of wavelength
were calculated using the Cauchy equation in coordination with constants in the
Iiterature.“7'48 For ethanol, literature optical constants were interpolated to obtain
data at the ellipsometer wavelengths, but due to limited ethanol literature data,
fitting of ellipsometric measurements in this solvent was performed only between
476.5 nm and 632.8 nm.49 Swelling percentages were subsequently determined
using Equation 3.1.

Swollen Thickness - Dry Thickness

Percent Swelling = Dry Thickness

 

*100% (3.1)

Atomic Force Microscopy. AFM experiments (Digital Instruments
Dimension 3100, Nanoscope llI controller in Tapping Mode, TappingModeTM
etched silicon probe tip, spring constant 20-100 N/m) were performed to validate
the ‘dry’ ellipsometric PEM thicknesses. Thicknesses were determined by
scratching a ﬁlm-coated Si wafer with Techni-ToolTM tweezers and scanning a 3
x 25 pm area over the scratch to produce an average line scan. A ‘step height’
measurement subtracted the average height of the bare silicon wafer from the
average height of the ﬁlm on top of the wafer. Three PEM-coated wafers were
each scanned three times (in different places) for every type of ﬁlm examined in
this study. The RMS roughness values were equal to or less than the standard

deviation of the thickness measurements for all coatings except 4.5-bilayer

73

PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl, which had an RMS roughness
of 11% of the film thickness. Due to experimental constraints, the AFM
thicknesses were measured at 54% RH and are compared to ellipsometric
thicknesses measured at 55% RH.

Transport Experiments. Film permeation properties were investigated
by DD and cross-ﬂow NF experiments, some of which were reported previously.9
New data presented here include all DD in ethanol, aqueous dialysis with the
PSS/PAH and HA/chitosan systems, and NF data for PSS/PAH. Diffusion
dialysis was performed using a glass apparatus in which the membrane
separated a source phase from a receiving phase that was initially deionized
water or pure ethanol.9'11 In water, source-phase solutions for all PSS/PAH ﬁlms,
PDADMAC-capped PSS/PDADMAC films deposited from 0.5 M NaCl, and
HA/chitosan ﬁlms contained 5 mM glycerol, glucose, sucrose, and rafﬁnose,
while DD solutions for PSS/PDADMAC ﬁlms deposited from 0.1 M NaCl and
PSS-capped PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl contained 5 mM
glycerol, glucose, and sucrose and 15 mM rafﬁnose. For ethanol-based diffusion
dialysis, the feed contained only 140 uM glucose, sucrose, and rafﬁnose because
of the low solubility of these compounds. Because glycerol was not present in all
solutions (it co-elutes with ethanol during analysis), its transport rates are not
reported. When present, glycerol was always the fastest transporting solute.
Samples were collected every 10-30 minutes and subsequently analyzed by
liquid chromatography (Dionex, DX-600, CarboPac PA-10 column, 100 mM

NaOH mobile phase) with integrated amperometric detection (Dionex, ED-50).

74

Nanoﬁltration experiments occurred at a pressure of 4.8 bar (70 psi), and
feed solutions were ﬂowed across the membrane at a rate of 18 mL/min.”11 NF
rejection is deﬁned by Equation 3.2, where R is the percent solute rejection and
Cperm and Cfeed are the concentrations of the solute in the permeate and the feed,
respectively. The NF feed solutions contained 1 mM glycerol, glucose, sucrose,
and rafﬁnose for PSS/PDADMAC as well as HA/chitosan ﬁlms, and 1 mM
glycerol and glucose with 5 mM sucrose and rafﬁnose for PSS/PAH ﬁlms. The
system was equilibrated for 18 h before permeate samples were acquired.

R = [1—9ﬂ1]x100% (3.2)
feed
3.3 Results and Discussion

Swelling. Ellipsometry, which involves the measurement of the ratios of
the complex reﬂection coefﬁcients for p- and s-polarized light, served as the
primary tool for ascertaining the extent of ﬁlm swelling. From the phase
difference, A, and the ratio of amplitudes, tan W, of the two reﬂection coefﬁcients,
one can calculate ﬁlm thickness and refractive index using a model that sums the
many individual reﬂections in the system (Figure 3.1). In the particular case of
coatings on Si wafers, this model includes both ﬁlm and Si02 layers on the
substrate, so oxide layer thicknesses were determined prior to deposition of
ﬁlms. To examine the reliability of ellipsometric data, we calculated how the
ellipsometric parameters A and LI) vary with coating refractive index and
thickness in nitrogen, water, and ethanol. Figures 3.2, 3.3 and 3.4 show
examples of simulations where the ﬁlm is immersed in water, nitrogen (<5%

relative humidity), and ethanol. With a possible error of 03° in A and LI) due to

75

Incident

Light (linearly

polarized)
Ambient, n1

Film, n2 d2 4 '1 }d2

Silicon Oxide, n3 k3 d3 }d3

Silicon, n4 k4

Reﬂected Light (sum,
elliptically polarized)

 
 

Figure 3.1: Diagram depicting the model of light reﬂection used to ﬁt
ellipsometric data. The n and k terms describe the real and imaginary
parts of the complex refractive indices, while d represents the layer

thickness. The layers are not drawn to scale.

76

Refractive

87 Index
86.5 *
86 '
85.5 ,.
85 ~

 

A (degrees)

 

52.5 nm 55 nm 57.5 nm

10 12 14 16 18 20
‘1’ (degrees)

Figure 3.2: Calculated W and A values for a Si/SiOzlﬁlm system (water as
the ambient medium and a 2.0 nm silicon oxide layer) as a function of the
refractive index and thickness of the ﬁlm. The simulation was performed at
a wavelength of 450.5 nm where the optical constants are: water -
n=1.3395, silicon oxide - n=1.4644, k=0, and silicon - n=4.7108, and
k=0.0963. The point with a thickness of 55 nm and a refractive index of
1.44 represents a 9.5-bilayer PSS/PDADAMC ﬁlm deposited from 0.1 M
NaCl, and the black lines through that point show the :0.3° uncertainty in W
and A measurements. Enclosed data points correspond to identical

thicknesses at different refractive indices.

77

 
  
  

 

87 _, Refractive lust
gal Index l1.52l
* “.531
g z: . 22.5 nm X154
1.5
g, 83 , 23.75 nm :15:
z 82 i 171-25]:
81 4
80 ~
79
78 -- - — - , ~— z.,-+,ﬂ._,____
20.5 21.5 22.5 23.5 24.5

‘1’ (degrees)

Figure 3.3: Calculated LlJ and A values for the system depicted in Figure 3.1
(nitrogen with <5% relative humidity as the ambient medium and a 2.0 nm silicon
oxide layer) as a function of the refractive index, n2, and thickness, d2, of the ﬁlm.
The simulation was performed at a wavelength of 450.5 nm where the optical
constants are n1=1.000, n3=1.4644, k3=0, n4=4.7108, and k4 = 0.0963. The point
with a thickness of 25 nm and a refractive index of 1.54 represents a 9.5-bilayer
PSS/PDADAMC ﬁlm deposited from 0.1 M NaCl, and the black lines through that
point show the i0.3° uncertainty in LP and A measurements. Enclosed data

points correspond to identical thicknesses at different refractive indices.

78

75

   
 
 
 

55nm
74 52.5 nm
73 50 nm
.... + 47.5 nm
§ 72 .
28,-'1 51.44
g 71 . jl1.45j
4 : _ A1.46|
70 , Refractive x1471
Index :*1.48'
69 .0149
. +1.50
10 11 12 13 14 15

‘I’ (degrees)

Figure 3.4: Calculated W and A values for the system depicted in Figure 3.1
(ethanol as the ambient medium and a 2.0 nm silicon oxide layer) as a function of
the refractive index, n2, and thickness, d2, of the ﬁlm. The simulation was
performed at a wavelength of 480.2 nm where the optical constants are
n1=1.3645, n3=1.4636, k3=0, n4=4.412, and k4 = 0.0629. The point with a
thickness of 50 nm and a refractive index of 1.47 represents a 9.5-bilayer
PSS/PDADAMC ﬁlm deposited from 0.1 M NaCl, and the black lines through that
point show the 203° uncertainty in W and A measurements. Enclosed data

points correspond to identical thicknesses at different refractive indices.

79

both window effects and measurement uncertainty, the simulations show that ﬁlm
thicknesses and refractive indices can be determined to at least :l:5% and 10.01,
respectively, in both water and ethanol. For “dry” ﬁlms, thicknesses and
refractive indices can be determined to 525% and 10.04, respectively.

To further validate the ellipsometric results, we determined PEM
thicknesses using atomic force microscopy (AFM) images of intentionally
scratched PEM-coated Si wafers. A typical AFM image can be found in
Figure 3.5. AFM-derived thicknesses for PSS/PDADMAC deposited from 0.1 M
NaCl, [PSS/PDADMAC]4PSS deposited from 0.5 M NaCl, and HA/chitosan ﬁlms
were consistently 10-15% greater than the ellipsometric thicknesses acquired at
a similar relative humidity. AFM-derived thicknesses for [PSS/PDADMAC]4
deposited from 0.5 M NaCl and PSS/PAH ﬁlms were not signiﬁcantly different
from ellipsometrically determined thicknesses. The positive deviation of AFM
thicknesses from ellipsometric thicknesses has been reported before50 and could
result from scratching of the underlying Si02 layer or deposition of the removed
material on the nearby ﬁlm. In any case, the ellipsometric measurements are
validated by the reasonable agreement between ellipsometric and AFM methods.
One assumption in most ellipsometric thickness determinations is that ﬁlms are
smooth and uniform. The RMS roughness values of PSS/PAH, PSS/PDADMAC,
and HA/chitosan ﬁlms were always less than 15% of ﬁlm thickness.

Table 3.1 shows the ellipsometric thicknesses of PSS/PAH,
PSS/PDADMAC, and HA/chitosan ﬁlms under nitrogen (<5% RH), water, and

ethanol (See Figures 3.6, 3.7, and 3.8 for some typical experimental and ﬁtting

80

 

 

 

 

 

-50~ , . ,
0 10 20 Mm

Figure 3.5: A 3 x 25 um AFM topographical image and the derived average

 

 

 

 

line scan for thickness analysis of a [PSS/PDADMAC]10 ﬁlm deposited from
0.1 M NaCl. The region between the two black arrows on the left determines
the average ﬁlm + wafer height, and the area between the two dark grey
arrows on the right determines the height of the scratched Si wafer. The

average PEM thickness for this particular scan is 28.3 nm.

81

Table 3.1: Ellipsometric thicknesses of PEMs under nitrogen (<5% RH), water,

or ethanol, and percent swelling in the two solvents.

 

. Film Film
Dry Film . Percent . Percent
Film Type Thickness Thickness Swelling Thickness Swelling
(nm) '” H20 in H20 '” Ethan” in Ethanol
(nm) (nm)

[HA/Chitosan]a 24:4 1 18:9 390:50 55:1 130:20
[HA/Chitosan17HA 24:3 1 18:6 390:40 55:3 130:20
[PSS/PDADMAC]4

from 0.5 M N aCI 20:1 95:9 380:60 30:1 54:10
[PSS/PDADMAC]4PSS
from 0.5 M N aCl 24.4:0.4 50:2 106:9 33:2 37:8
[PSS/PDADMACLO
from 0.1 M N aCl 24.7:0.9 57:1 129:7 52:1 110:7
[PSS/PDADMAC]9PSS
from 0.1 M NaCl 24:1 54:1 124:8 51 2:05 112:5
[PSS/PAij 25.2:0.7 51 :1 101:6 49:2 94:8
[PSS/PAH19PSS 25. 1 :0. 5 49:2 95:9 49:3 96:13

82

Generated and Experimental Data

 

 

 

 

 

27 I l I I U l l w
24- 1
_ —84
“’21— d
i - 42>
E ' “80%
15—

3' _ 1 §
12_ —78
— 1
g . l . l . l . 75
401) 5000 am 7000 801)

Wavelength inA
Figure 3.6: The experimental (dashed line) and generated (ﬁt, solid line) LP and
A ellipsometric data for a dry (under <5% RH nitrogen) 10-bilayer
PSS/PDADMAC ﬁlm deposited from 0.1 M NaCl. This ﬁt corresponds to. a ﬁlm

that is 24 nm thick and has a refractive index of 1.541 at 450.5 nm.

83

 

Generated and Experimental Data
l

    

 

 

 

20 . . , . I . 140
l -
18— ——ModelFit
_ ----- ExpPsiE75° ‘120
m 16 —ModelFit _ D
— ----- DeltaE75°
§ _ Exp -100
O, _.
g 14.. - g
.E - —80 g
12— c0
9 _ 1 8
10- 76°
, 1
8 l . l . l . 40
4000 5000 6000 7000 8000

V\bvelengthinA
Figure 3.7: The experimental (dashed line) and generated (ﬁt, solid line) LlJ and
A ellipsometric data for a water-submerged 8-bilayer HA/chitosan ﬁlm. This ﬁt

corresponds to a ﬁlm that is 124 nm thick and has a refractive index of 1.389 at

450.5 nm.

84

Generated and Experimental Data

 

 

 

 

125‘ l I l I r I r I i 75
’ ——Model Flt — 70
12.0- ----- 510F957? ‘

x —Modeii=it _65

g _ ----- ExpDeltaE75° , l>
gus ""\3 -mg
.0 . '— ‘\\ \\\\ﬂ ‘ o-
E ‘ ~558
11.0.. 750‘”
_ “ ~ - ‘ _45

10.5 . I 1 l . l 1 I "In“ 40
4800 51(1) 54(1) 57(1) 6W0 63(1)

vaelengthinA

Figure 3.8: The experimental (dashed line) and generated (ﬁt, solid line) 4’ and
A ellipsometric data for an ethanol-submerged 9.5-bilayer PSS/PDADMAC ﬁlm
deposited from 0.1 M NaCl. This ﬁt corresponds to a ﬁlm that is 52 nm thick and

has a refractive index of 1.453 at 480.2 nm.

85

data for “dry,” water-submerged, and ethanol-submerged ﬁlms). The number of
bilayers in these ﬁlms was chosen such that all coatings would have similar
thicknesses. There are clear variations in percent swelling as a function of both
the constituent polyelectrolytes and the swelling solvent. Figure 3.9, which
presents the structures of the constituent polyelectrolytes, shows that chitosan,
PDADMAC, and PAH contain one positive charge per 11, 9, and 4 non-hydrogen
atoms, respectively. In the case of the polyanions, PSS contains one negative
charge per 12 non-hydrogen atoms, while HA has only one charge per 26 non-
hydrogen atoms. Thus, if swelling increases with decreasing charge density on
the polyelectrolytes (due to a lower density of ionic cross-links), HA/chitosan
should swell much more than PSS/PDADMAC, which should swell more than
PSS/PAH, and this is generally the case in water, though swelling in ethanol is
complicated by other factors. The especially low charge density on HA and
chitosan results in ﬁlms that are nearly 80% water. In accord with such a high
water content, refractive indices at 603.1 nm for swollen [HA/chitosan]7HA ﬁlms
are only 1.382. (At the same wavelength, dry [HA/chitosan]7HA ﬁlms have a
refractive index of 1.53, while the refractive index of water is 1.333.) For all types
of water-swollen or ethanol-swollen ﬁlms, the refractive index is essentially a
linear combination of the refractive indices of water and the dry ﬁlm, as described
by Equation 3.3 where Tsf and Tdf are the thickness of the swollen ﬁlm and the
dry ﬁlm, respectively, and nsf, ndf, and n. are the refractive indices of the swollen

film, the dry ﬁlm, and the swelling solvent (ethanol or water), respectively.

86

. /N+\Cl-
oly(allylamine hydrochloride)
PAH
Poly(diaIlyldimethylammonium
SO3 Na chloride)
Poly(styrene PDADMAC
sulfonate)
PSS
COO Na
4/ O HO
O
OH 0 NHCOCH
HO Ni—i3 Cl 3
Chitosan Hyaluronic Acid
HA

Figure 3.9: Structures of the polyelectrolytes used in this work.

87

Estimations of nsf calculated with Equation 3.3 vary from ellipsometrically
determined nsf values by less than 1.5% in water and less than 3% in ethanol.
Ts,nsf = (Tsf -Td,)ns + T,,,ndf (3.3)
In most cases, the difference in swelling between ﬁlms terminated with a
polycation and a polyanion is not statistically signiﬁcant. This is consistent with
previous studies that suggest that although water uptake can depend on the
composition of the terminating layer in PSS/PAH ﬁlms,29'3°'51 the solvent fraction
in PAH-capped ﬁlms is only 25% less than in PSS-capped ﬁlms (40% versus
30% water as a function of capping layer)”30 PSS/PDADMAC ﬁlms deposited
from 0.5 M NaCl are a notable exception to the phenomenon of capping layer
choice not affecting swelling. The water uptake in [PSS/PDADMAC}; ﬁlms
prepared in 0.5 M NaCl is almost 4-fold greater than that in the corresponding
[PSS/PDADMACjaPSS ﬁlms. We speculate that this occurs because the
PDADMAC penetrates the entire ﬁlm and disrupts ionic cross-linking This is
similar to the explanation for the rapid, exponential (as a function of the number

of adsorption steps) growth of some PEMs.52'53

Indeed, we observed that
PSS/PDADMAC ﬁlms grown in >03 M NaCl do show exponential growth, while
films grown in 0.1 M NaCl do not.54 Consistent with PDADMAC penetrating the
entire ﬁlm, Smith et al. used 13C solid-state NMR to show that PDADMAC is more
mobile than PSS in PSS/PDADMAC ﬁlms.55 This high mobility was attributed to

the low glass transition temperature of PDADMAC, which is below room

temperature when there is >20% water content.”56 McCormick et al. also

88

reported an increase in both water and PDADMAC mobility in PSS/PDADMAC
films when PDADMAC is the top layer.57

Interestingly, in ethanol, there is only a small difference between the
swelling of [PSS/PDADMAC]4 and [PSS/PDADMACj4PSS prepared in 0.5 M
NaCl (54 versus 37%), and the swelling of both of these ﬁlms is only 1/3 to 1/2 of
that for corresponding coatings prepared in 0.1 M NaCl. Films deposited from
0.1 M NaCl likely swell more in ethanol because they contain fewer ion-exchange
sites and are less hydrophilic than ﬁlms deposited from 0.5 M NaCl. The
hydrophilic, non-polyelectrolyte-paired (ion-exchange) charged groups in
PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl likely make them less
susceptible to swelling in ethanol.37 Moreover, ethanol may not lower the glass
transition temperature of PDADMAC the same way that water does, which could
decrease chain mobility and reduce swelling. NMR studies of PEMs in ethanol
could reveal if the lack of swelling in PSS/PDADMAC depositedfrom 0.5 M NaCl
correlates with a lack of PDADMAC mobility.57 HA/chitosan ﬁlms also swell less
in ethanol than in water, and this likely reﬂects the fact that ethanol is a poorer
solvent for these hydrophilic polymers. Additionally, increased ion pairing (cross-
linking) may occur in the presence of ethanol. In the case of ﬁlms that are
already heavily cross-linked in water (PSS/PAH and PSS/PDADMAC deposited
from 0.1 M NaCl), swelling is similar in ethanol and water.

Below, we compare swelling and transport results. For HA/Chitosan and
PSS/PDADMAC deposited from 0.5 M NaCl, ﬁlms used in transport and swelling

experiments had essentially the same number of bilayers, but in the case of

89

PSS/PDADMAC deposited from 0.1 M NaCl and PSS/PAH, more bilayers were
used in ellipsometric than in transport studies. This was necessary because
ellipsometric measurements require relatively thick ﬁlms for accurate refractive
index and thickness determinations, but overly thick ﬁlms severely retard ﬂux in
transport experiments. PSS/PDADMAC deposited from 0.1 M NaCl and
PSS/PAH are less permeable than the other systems, so the minimal thickness
required for accurate transport experiments with these ﬁlms was not sufﬁcient for
ellipsometric thickness determinations in solvents. A previous study showed that
the permeability of PSS/PDADMAC is constant, after the deposition of 5 to 6
bilayers,58 so we expect that the swelling of 9.5 and 10-bilayer ﬁlms should still
be relevant to transport through 5 and 5.5-bilayer systems. Our previous NF
studies also indicated that solute rejections by PSS/PDADMAC ﬁlms deposited
from 0.1 M NaCl did not change on going from 5.5 to 6.5 bilayer ﬁlms.9
Transport Experiments in Water. Large changes in swelling by water
correlate well with both NF and DD data. Table 3.2 shows that the rejections of
glucose, sucrose, and rafﬁnose in NF generally increase as ﬁlm swelling
decreases, as would be expected. l-lighly swollen [HA/chitosan]8HA ﬁlms show
essentially no rejection of any of the sugars, while the least swelling system,
PSS/PAH, rejects >99.6% of sucrose. DD data (Table 3.3) conﬁrm the trends
seen in NF. With the possible exception of the comparison of PSS/PAH ﬁlms
with PSS-capped PSS/PDADMAC deposited from 0.5 M NaCl (the swelling is
similar between the ﬁlms), ﬂuxes of glucose, sucrose, and rafﬁnose all increase

with increasing ﬁlm swelling, even though the more swollen HA/chitosan and

90

Table 3.2: Percent rejection in nanoﬁltration of glucose, sucrose, and rafﬁnose
dissolved in water. PSS/PDADMAC and HA/chitosan data are from Chapter 2.9

The swelling values are from analogous ﬁlms (Table 3.1).

Percent Rejection

 

 

Percent
Film Type Swelling Glucose Sucrose Rafﬁnose
in Water
[HA/Chitosan]8HA 390:40 <12% <12% <12%
[PSS/PDADMACL‘ 380:60 17:4 28:9 36:8

from 0.5 M NaCl

[PSS/PDADMAC]4PSS
from 0.5 M NaCl 106:9 64:6 97.2:O.9 98.9:O.7

[PSS/PDADMAst
from 0.1 M NaCl 129:7 44:5 85:2 91.3:0.8
[PSS/PDADMAC]5PSS
from 0.1 M NaCl 124:8 58:8 96:2 98.8:0.6
[PSS/PAHjs 101:6 91.3:06 99.6:0.1 99.8:0.1
[PSS/PAH]5PSS 95:9 88:1 99.70:0.09 99.90:0.07

91

Table 3.3:

Fluxes in diffusion dialysis of glucose, sucrose, and rafﬁnose

(dissolved in water) through porous alumina coated with various polyelectrolyte

ﬁlms. The bare alumina and PSS/PDADMAC data are from Chapter 2.9 All ﬁlm

thicknesses were measured in 3 <5% RH nitrogen atmosphere except for that of

[HA/chitosan]3HA, which is an estimate based upon the thickness of 8-bilayer

HA/chitosan ﬁlms.

Normalized Fluxa

(nmol cm'2 s'1 M")

 

 

Film Percent
Film Type Thickness Swelling Glucose Sucrose Rafﬁnose
(nm) in szD
Bare Alumina N/A N/A 180:20 140:20 120:10
[HA/ChitosankHA ~25 390:40 180:30 140120 1 10:20
[HA/Chitosanh; 24:4 390:50 180:20 131:9 1 10:10
[PSS/PDADMAC]4
from 0.5 M NaCl 20:1 380:60 150:10 120:10 95:9
[PSS/PDADMAC]4PSS
from 0.5 M NaCl 24.4:0.4 106:9 90:10 7:2 03:01
[iii/2%,? “(12816 15:1 129:7 90:10 24:2 19:01
[Psff’éz‘og‘E’nAﬂgss 12:1 124:8 90:4 1 1:2 05:03
[PSS/PAH]5 12.5:0.6 101:6 43:3 12:04 06:04
[PSS/PAH]5PSS 13.1:0.4 95:9 56:9 1.1:0.7 0.5:0.5

aFluxes were normalized by dividing by the source-phase concentration at the

end of the experiment.

bValues are taken from Table 3.1 and were measured with ﬁlms that in some
cases had different thicknesses than those used here.

92

PSS/PDADAMC films prepared in 0.5 M NaCl are about 1.5 to 2-fold thicker than
the other ﬁlms. Fluxes through both HA-terminated and chitosan-terminated
HA/chitosan ﬁlms are essentially the same as those through bare alumina. This
is not surprising considering the low (<12%) rejection of sugars in NF and the
rapid passage of molecules as large as myoglobin through HA/chitosan ﬁlms.9
For the PSS/PDADMAC systems, the transport data demonstrate that
variations in ﬁlm permeability can occur upon changing the top layer in the ﬁlm
from a polycation to a polyanion. On going from PDADMAC-capped to P88-
capped PSS/PDADMAC ﬁlms made in 0.5 M NaCl, sucrose rejection in NF
increases from 28 to 97%, and ﬂuxes in DD decrease by factors of 1.7, 17, and
300 for glucose, sucrose, and rafﬁnose, respectively. This correlates well with
the almost 4-fold decrease in swelling that occurs upon addition of a PS8
capping layer. A similar, though smaller, effect occurs for PSS/PDADMAC ﬁlms
prepared in 0.1 M NaCl, but in this instance the differences in the rejections and
fluxes exhibited by PSS-terminated and.PDADMAC-terminated ﬁlms are larger
than what one might expect, given the insigniﬁcant differences in their water
uptake. Other factors such as polymer intermingling and chain mobility probably
affect the permeability of these ﬁlms. As discussed earlier, NMR experiments by
McCormick and Smith indicate that the PDADMAC portions of PSS/PDADMAC
films are more mobile when the entire ﬁlm is terminated end in PDADMAC.55'57
The aqueous NF and DD data show that PEMs can exhibit extraordinarily
diverse permeability properties ranging from nearly complete rejection to nearly

complete passage of sugar molecules, depending on polyelectrolyte type, top-

93

layer choice, and deposition conditions.9'11

However, many processes are
incompatible with water, and aqueous data may not describe PEM behavior in
organic solvents.41 Thus, we investigated diffusion dialysis through PSS/PAH,
PSS/PDADAMC, and HA/chitosan ﬁlms in ethanol.

Diffusion dialysis in ethanol. Table 3.1 shows that there is signiﬁcantly
less swelling of PEMs in ethanol than in water, particularly for the highly swollen
films [HA/chitosanjg, [HA/chitosan]7HA, and [PSS/PDADMAC]3 prepared in 0.5 M
NaCl. Still, there is a signiﬁcant variation in ethanol swelling among the
polyelectrolyte systems, ranging from 37% for [PSS/PDADMAC]4PSS deposited
from 0.5 M NaCl to 130% for HA/chitosan ﬁlms. Nevertheless, DD data in
ethanol exhibit minimal correlation between ﬂuxes and ﬁlm swelling (Table 3.4).
For example, polycation-capped [PSS/PDADMACja ﬁlms swell slightly more in
ethanol than [PSS/PDADMAC]4PSS ﬁlms (both prepared in 0.5 M NaCl), and
glucose, sucrose, and rafﬁnose DD ﬂuxes through [PSS/PDADMAC]5PSS and
[PSS/PDADMAst films differ by factors of 2, 6, and 4, respectively. However,
the more highly swollen PSS/PAH and PSS/PDADMAC ﬁlms deposited from 0.1
M NaCl exhibit less sugar ﬂux than either of the less swollen PSS/PDADMAC
films deposited from higher ionic strength. Moreover, ﬂuxes through
[HA/chitosan]8 ﬁlms are 70-75% lower than those through [PSS/PDADMAC15

deposited from 0.5 M NaCl, even though the HA/chitosan swells over twice as

much, and the thicknesses of these ﬁlms differ by only 15%.

94

Table 3.4: Fluxes in diffusion dialysis of 140 uM glucose, sucrose, and rafﬁnose
in ethanol through porous alumina coated with various polyelectrolyte ﬁlms. All
film thicknesses were measured in a <5% RH nitrogen atmosphere except for
that of [HA/chitosan]3HA, which is an estimate based upon the thickness of 8-
bilayer HA/chitosan ﬁlms.

Normalized Fluxa

(nmol cm'2 s'1 M")

 

 

. Percent
F"m Swelling
Film Type Thickness in Glucose Sucrose Rafﬁnose
(”r") Ethanolb
Bare Alumina N/A N/A 210:20 180:40 190:20
[HA/Chitosan]3HA ~25 130:20 100:30 40:20 30:10
[HA/Chitosanja 24:4 130:20 40:20 30:20 30:20
[PSS/PDADMAC]5
from 0.5 M NaCl 27.3:09 54:10 150:10 110:10 100:20
[PSS/PDADMAC]5PSS
from 05 M N aCl 31.8:0.6 37:8 74:6 19:2 28:2
[PSS/PDADMACja c c
from 0.1 M NaCl 15:1 110:7 50:10 <4 <4
[PSS/PDADMAC15PSS c c
from 0.1 M NaCl 12:1 112:5 40:10 <4 <4
[PSS/PAHjs 12.5:0.6 94:8 13.1:0.1 <4c <4c
[PSS/PAH]5PSS 13.1:0.4 96:13 10.4:0.1 <4c <4c

aFluxes were normalized by dividing by the source-phase concentration at the
end of the experiment.

bValues are taken from Table 3.1 and were measured with ﬁlms that in some
cases had different thicknesses than those used here.

cLower limits of measurable, normalized ﬂux are higher in ethanol than in water
because of limited sugar solubility in ethanol.

95

In contrast to PSS/PDADMAC prepared in 0.5 M NaCl, HA/chitosan ﬁlms
show signiﬁcant (2-fold) decreases in glucose ﬂuxes in ethanol when the outer
layer of the ﬁlm is a polycation rather than a polyanion (Table 3.4). Again,
however, these trends do not correlate with swelling, as the solvent uptake of
both types of ﬁlms does not depend signiﬁcantly on the composition of the top
layer. PSS/PAH ﬁlms deposited from 0.5 M NaCl show a slightly higher glucose
flux when capped by the polycation, but the differences in swelling are negligible.
We think that the ethanol swelling reﬂects a tradeoff between ionic cross-linking
and ﬁlm hydrophilicity. Highly cross-linked ﬁlms may be more hydrophobic and
soluble in ethanol, and this may oppose the reduction of swelling due to ionic
cross-links, even though cross-linking could slow transport by limiting chain
mobility. In contrast, such an effect would amplify decreases in aqueous swelling
that result from ionic cross-linking, and in that case, we do see a strong

correlation between swelling and ﬂux.

3.4 Conclusions

Swelling of PEMs in water increases as the density of charge on
constituent polyelectrolytes decreases, and increased water uptake generally
leads to decreased sugar rejections in NF and higher solute ﬂuxes in diffusion
dialysis. Presumably, higher swelling occurs when fewer ionic cross-links and/or
many hydrophilic ion-exchange sites are present in the ﬁlm. With

PSS/PDADMAC ﬁlms prepared in 0.5 M NaCl, water uptake can vary 4-fold

96

depending on whether the capping layer is a polycation or a polyanion, and the
higher water content leads to minimal sugar rejections in NF. Solvent uptake is
generally smaller with ethanol than water, and there is no clear correlation

between ethanolic DD data and swelling.

3.5 Acknowledgements
We thank the Department of Energy Office of Basic Energy Sciences for
financial support of this work. We also thank Dr. Baokang Bi of Michigan State

University for his assistance in the AF M studies.

97

3.6 Works Cited

(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)

(9)
(10)

(11)
(12)

(13)

(14)

(15)

(15)
(17)

(18)

(19)

Krasemann, L.; Tieke, B. J. Membr. Sci. 1998, 150, 23-30.

Krasemann, L.; Tieke, B. Mat. Sci. Eng. C-Bio S1999, 08-09, 513-518.
Toutianoush, A.; Tieke, B. Mat. Sci. Eng. C 2002, 022, 459-463.

Lenk, W.; Meier-Haack, J. Desalination 2002, 148, 11-16.

Sullivan, D. M.; Bruening, M. L. J. Membr. Sci. 2005, 248, 161-170.
Schwarz, H. H.; Malsch, G. J. Membr. Sci. 2005, 247, 143-152.

Jin, W.; Toutianoush, A.; Tieke, B. Langmuir 2003, 19, 2550-2553.

Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003,
19, 7038-7042.

Miller, M. D.; Bruening, M. L. Langmuir 2004, 20, 11545-11551.

Lajimi, R. H.; Ben Abdallah, A.; Ferjani, E.; Roudesli, M. S.; Deratani, A.
Desalination 2004, 163, 193-202.

Liu, X.; Bruening, M. L. Chem. Mater. 2004, 16, 351-357.

Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.; Mohwald, H.
Angew. Chem, Int. Ed. 1998, 37, 2202-2205.

Caruso, F.; Trau, D.; Mtihwald, H.; Renneberg, R. Langmuir 2000, 16,
1485-1488.

Diaspro, A.; Silvano, D.; Krol, S.; Cavalleri, O.; Gliozzi, A. Langmuir 2002,
18, 5047-5050.

Trau, D.; Yang, W.; Seydack, M.; Caruso, F.; Yu, N.-T.; Renneberg, R.
Anal. Chem. 2002, 74, 5480-5486.

Yu, A.; Caruso, F. Anal. Chem. 2003, 75, 3031-3037.

Kidambi, 8.; Dai, J.; Li, J.; Bruening, M. L. J. Am. Chem. Soc. 2004, 126,
2658-2659.

Shchukin, D. G.; Shutava, T.; Shchukina, E.; Sukhorukov, G. B.; Lvov, Y.
M. Chem. Mat. 2004, 16, 3446-3451.

Pargaonkar, N.; Lvov, Y. M.; Li, N.; Steenekamp, J. H.; de Villiers, M. M.
Pharm. Res. 2005, 22, 826-835.

98

(20)
(21)
(22)
(23)

(24)

(25)
(26)

(27)

(28)

(29)

(30)

(31)
(32)
(33)

(34)
(35)

(36)
(37)
(38)

(39)

Zahr, A. S.; de Villiers, M.; Pishko, M. V. Langmuir 2005, 21, 403-410.
Zhu, H.; McShane, M. J. Langmuir 2005, 21, 424-430.
Decher, G. Science 1997, 277, 1232-1237.

Harris, J. J.; Stair, J. L.; Bruening, M. L. Chem. Mater. 2000, 12, 1941-
1946.

Lvov, Y.; Ariga, K.; lchinose, l.; Kunitake, T. J. Am. Chem. Soc. 1995, 117,
6117-6123.

Hollman, A. M.; Bhattacharyya, D. Langmuir 2004, 20, 5418-5424.

Caruso, F.; Niikura, K.; Furlong, D. N.; Okahata, Y. Langmuir1997, 13,
3427-3433.

Garza, J. M.; Schaaf, P.; Muller, 8.; Ball, V.; Stoltz, J.-F.; Voegel, J.-C.;
Lavalle, P. Langmuir 2004, 20, 7298-7302.

Rivin, D.; Meermeier, G.; Schneider, N. S.; Vishnyakov, A.; Neimark, A. V.
J. Phys. Chem. B 2004, 108, 8900-8909.

Carriere, D.; Krastev, R.; Schoenhoff, M. Langmuir 2004, 20, 11465-
11472.

Wong, J. E.; Rehfeldt, F.; Haenni, P.; Tanaka, M.; Klitzing, R. v.
Macromolecules 2004, 37, 7285-7289.

Harris, J. J.; Bruening, M. L. Langmuir 2000, 16, 2006-2013.
Hiller, J. A.; Rubner, M. F. Macromolecules 2003, 36, 4078-4083.

Sukhorukov, G. B.; Schmitt, J.; Decher, G. Berichte der Bunsen-
Gesellschaft1996, 100, 948-953.

Kim, B.-S.; Vinogradova, O. l. J. Phys. Chem. B 2004, 108, 8161-8165.

Ruths, J.; Essler, F.; Decher, G.; Riegler, H. Langmuir 2000, 16, 8871-
8878.

Poptoshev, E.; Schoeler, B.; Caruso, F. Langmuir 2004, 20, 829-834.
Dubas, S. T.; Schlenoff, J. B. Langmuir 2001, 17, 7725-7727.

Mendelsohn, J. D.; Yang, S. Y.; Hiller, J.; Hockbaum, A. l.; Rubner, M. F.
Biomacromolecules 2003, 4, 96-106.

Tanchak, O. M.; Barrett, C. J. Chem. Mater. 2004, 16, 2734-2739.

99

(40)
(41)

(42)

(43)
(44)
(45)

(45)
(47)

(48)

(49)
(50)

(51)
(52)

(53)

(54)

(55)

(55)
(57)

(53)

Burke, S. E.; Barrett, C. J. Biomacromolecules 2005, 6, 1419-1428.

Kim, B.-S.; Lebedeva, O. V.; Koynov, K.; Gong, H.; Glasser, G.;
Liebenlvilth, l.; Vinogradova, O. I. Macromolecules 2005, 38, 5214-5222.

Cardenas, A.; Argueelles-Monal, W.; Goycoolea, F. M.; Higuera-Ciapara,
|.; Peniche, C. Macromol. Biosci. 2003, 3, 535-539.

Burke, S. E.; Barrett, C. J. Macromolecules 2004, 37, 5375-5384.
Kim, S. J.; Lee, K. J.; Kim, S. l. J. Appl. Polym. Sci. 2004, 93, 1097-1101.

Tedeschi, C.; Méhwald, H.; Kirstein, S. J. Am. Chem. Soc. 2001, 123,
954-960.

Rezwan, K.; Meier, L. P; Gauckler, L. J. Langmuir 2005, 21, 3493-3497.

Schiebener, P.; Straub, J.; Sengers, J. M. H. L.; Gallagher, J. S. J. Phys.
Chem. Ref. Data 1990, 19, 677-717.

Harvey, A. H.; Gallagher, J. S.; Sengers, J. M. H. L. J. Phys. Chem. Ref.
Data 1998, 27, 761-774.

Moreels, E.; De Greef, C.; Finsy, R. Appl. Opt. 1984, 23, 3010-3013.

Porcel, C. H.; lzquierdo, A.; Ball, V.; Decher, G.; Voegel, J. C.; Schaaf, P.
Langmuir 2005, 21, 800-802.

Schwarz, B.; Schonhoff, M. Langmuir 2002, 18, 2964-2966.

Lavalle, P.; Picart, C.; Mutterer, J.; Gergely, C.; Reiss, H.; Voegel, J.-C.;
Senger, B.; Schaaf, P. J. Phys. Chem. B 2004, 108, 635-648.

Picart, C.; Mutterer, J.; Richert, L.; Luo, Y.; Prestwich, G. D.; Schaaf, P.;
Voegel, J. C.; Lavalle, P. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 12531-
12535.

McAloney, R. A.; Sinyor, M.; Dudnik, V.; Goh, M. C. Langmuir 2001, 17,
6655-6663.

Smith, R. N.; Reven, L.; Barrett, C. J. Macromolecules 2003, 36, 1876-
1881.

Yeo, S. C.; Eisenberg, A. J. J. Macromol. Sci, Phys. 1977, B13, 441.

McCormick, M.; Smith, R. N.; Graf, R.; Barrett, C. J.; Reven, L.; Spiess, H.
W. Macromolecules 2003, 36, 3616-3625.

Farhat, T. R.; Schlenoff, J. B. Langmuir 2001, 17, 1184-1192.

100

Chapter 4

MODIFIED POLYELECTROLYTE MULTILAYER FILMS FOR ENHANCED

NANOFILTRATION OF NEUTRAL AND CHARGED MOLECULES

SUMMARY

Polyelectrolyte multilayer ﬁlms are promising materials for nanoﬁltration
(NF) membranes because their minimal thickness affords high ﬂux, and control
over ﬁlm composition allows tailoring of solute rejections. This chapter describes
two modiﬁcations that attempt to improve ﬂuxes and selectivities through these -
ﬁlms. The ﬁrst alteration involves hybrid multilayer polyelectrolyte membranes
composed of poly(styrene sulfonate) (PSS)/protonated poly(allylamine) (PAH)
bilayers deposited on PSS/poly(diallyldimethylammonium chloride) (PDADMAC)
“gutter layers.” Ideally, these ﬁlms would combine the high permeability of
PSS/PDADMAC ﬁlms with the selectivity of PSS/PAH, leading to higher NF
solution ﬂuxes than pure PSS/PAH ﬁlms. A more successful modiﬁcation is the
deposition of the ﬁnal layer of PSS/PAH assemblies from high ionic strength
solutions ([PSS/PAH]4P8S*). This alters the ﬁlm structure, enhancing both NF
solution ﬂuxes and recoveries while maintaining the high selectivities of pure
PSS/PAH ﬁlms. Both types of modiﬁed ﬁlms exhibit NaCl/sucrose NF
selectivities at least ten-fold greater than those of pure PSS/PDADMAC
membranes, though only [PSS/PAH]4PSS* films allows ﬂuxes superior to pure
PSS/PAH ﬁlms. These systems also exhibit glucose/sucrose selectivities that are

twice those of pure PSS/PDADMAC films and glucose recoveries that are almost

101

double those of pure PSS/PAH assemblies. Additionally, both
[PSS/PDADMAC]...[PSS/PAH]nPSS* and [PSS/PAH]4PSS* ﬁlms can reject
multivalent salts from solutions containing butanetriol while maintaining ﬂuxes
greater than those of commercial NF membranes. Such separations could prove
valuable in the puriﬁcation of butanetriol, which is a precursor to a high explosive,

produced by fermentation.

4.1 Introduction

Alternating adsorption of polycations and polyanions on porous supports is
an attractive method for forming the skins of composite membranes because it
can yield extremely thin (<50 nm), defect-free layers.“3 The minimal thickness of
these coatings results in high ﬂuxes, and careful selection of constituent
polyelectrolytes or deposition parameters such as salt concentration allows
tuning of permselectivity."'8 Recently, several studies examined the use of

multilayer polyelectrolyte ﬁlms in encapsulationf"14 nanoﬁltration (NF),3'5'7'15

16-23 42425

pervaporation, and gas separations.

Within the realm of NF alone, polyelectrolyte-modiﬁed porous substrates

3'5'15 sugars and other small neutral molecules,6'7

are capable of separating ions,
and even proteins.6 Jin et al. demonstrated that polyvinylamine/polyvinylsulfate
films can achieve 94% rejection of NaCl at 40 bar, although ﬂuxes through those
ﬁlms were low because they contained 60 bilayers.3 Much thinner (4.5 bilayers)

poly(styrene sulfonate) (PSS)/poly(allylamine hydrochloride) (PAH) ﬁlms can

separate chloride and sulfate with a selectivity of 33 and a solution ﬂux of 1.8

102

m3/(m2 day) at a pressure of only 4.8 bar.5'26 [PSS/PAH]5PSS ﬁlms also exhibit
glucose/sucrose selectivities as high as 150, but low glucose recoveries limit the
applicability of these membranes in sugar separations.7 As shown in chapter 2,
simply changing the polycation employed to form such ﬁlms from PAH to
poly(diallyldimethylammonium chloride) (PDADMAC) results in a 4-10 fold
increase in glucose recovery and a ﬂux of more than 2 m3/(m2 day), albeit with
lower glucose/sucrose selectivities (~15).6'7 Additionally, PSS/PDADMAC ﬁlms
can separate NaCl from sucrose with a selectivity of 105'27 Though these results
are impressive ﬁrst steps, greater selectivities matched with higher glucose or
NaCl passages and high ﬂuxes are desirable for practical separations such as
the reﬁning of oligosaccharides in corn syrup or the recovery of brine from sugar
decolorization processesze"29

This chapter describes attempts to improve the NF solution ﬂux and
selectivity of polyelectrolyte multilayer (PEM) membranes through two structure
modiﬁcations. The ﬁrst involves hybrid ﬁlms, i.e., PEMs that contain regions with
different polyelectrolyte pairs as shown in Figure 4.1. Speciﬁcally, selective
PSS/PAH layers are deposited on top of PSS/PDADMAC “gutter layers”. A
previous report stated that the NF ﬂux through 30 nm-thick (at ambient
conditions) PSS/PAH ﬁlms is about 0.9 m3/(m2 day),7 while the solution ﬂux
through 30 nm PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl is 1.5-
1.9 m3/(m2 day).30 The PSS/PDADMAC portion of the ﬁlm is presumably more
permeable than PSS/PAH, as the lower charge density on PDADMAC than on

17,31-33

PAH results in fewer ionic cross-links and more swelling. (As shown in

103

Figure 4.1: Schematic illustration of the concept of hybrid polyelectrolyte
membranes containing PSS, PDADMAC, and PAH deposited onto a porous
alumina support. The color-coded chemical structures of the individual
polyelectrolytes are also shown. (I' he porous support is actually several orders

of magnitude thicker than the polyelectrolyte ﬁlm.) *

N‘CI-
/ \
Poly(dlallyldimethylammonium
chloride)
PDADMAC
so ~Na‘
Poly(styrene sulfonate) n
PSS

NH3"CI‘

“' 4 ' -
I l I ll: lwlfllglﬂlglilliglalnlnlﬁl I Poly(allylamigilli'ydrochloride)

 

*Images in this dissertation are presented in color.

104

Figure 4.1, PAH has 4 non-hydrogen atoms per unit charge, while PDADMAC
has 9). We hypothesized that a hybrid ﬁlm would be advantageous because a
few PSS/PDADMAC bilayers would allow for coverage of a porous support with
fewer PSS/PAH bilayers than if PSS/PAH were used alone - producing ﬁlms with
high ﬂux and high selectivity.

A few recent papers demonstrated the formation of hybrid polyelectrolyte
films?“7 Garza et al. showed that deposition of PSS/PAH between hyaluronic
acid/poly(L-Iysine) bilayers results in a ﬁlm with highly permeable compartments
separated by relatively impermeable PSS/PAH barriers.37 We previously
employed membranes consisting of either poly(acrylic acid)/PAH34 layers or
poly(pyromellitic dianhydride-phenylenediamine)/PAH3’5 layers on top of
PSS/PAH ﬁlms to enhance selectivity between Cl‘ and 8042' in diffusion dialysis.
However, this study represents the ﬁrst investigation of the potential beneﬁts of
hybrid ﬁlms in a practical separation technique such as nanoﬁltration.

The second film modiﬁcation described here entails a slight change in the
deposition of recently described PSS/PAH ﬁlms.7 Previously, the PSS layers of
[PSS/PAH]6PSS ﬁlms were deposited from pH 2.1 0.5 M MDCiz, and the ﬁlms
exhibited a molecular weight cutoff (MWCO, the molecular weight at which 90%
of a solute is rejected) of about 100 g/mol, as glycerol (MW of 92) was ~90%
rejected. However, this chapter shows that depositing the ﬁnal PSS layer of a
PSS/PAH ﬁlm from 2.5 M MnClz at a pH of 2.1 (these ﬁlms are referred to as
[PSS/PAH]4PSS* ﬁlms from this point on) results in a higher MWCO (>200 g/mol)

while still maintaining high 8042‘ rejection (>96%). Previously, the use of high

105

ionic strength solutions for depositing PSS capping layers enhanced SO42'
rejection,5 but this chapter shows that it can also be advantageous for obtaining
higher passage of small, neutral molecules. NF solution ﬂuxes through
[PSS/PAH]4PSS* membranes rival those through PSS/PDADMAC ﬁlms,6 and are
about twice the ﬂux through commercial NF membranes.”39 The positive effect
of high ionic strength during deposition likely results from ﬁlm rearrangement, as
several studies showed that PEMs can rearrange, sometimes dramatically, in
response to changes in ionic strength or pH.32'40

This chapter shows that the hybrid and modiﬁed [PSS/PAHlaPSS* ﬁlms
yield double to triple the glucose/sucrose selectivity and at least ten times the
NaCl/sucrose selectivity of pure PSS/PDADMAC ﬁlms. Additionally, solution
ﬂuxes with these ﬁlms are comparable to those through PSS/PDADMAC
membranes (2 m3/(m2 day) at 4.8 bar), though only [PSS/PAH]4PSS* ﬁlms
possess statistically enhanced ﬂux compared to pure PSS/PAH ﬁlms. Also, both
membrane systems remove sulfate and phosphate anions from a 1,2,4-
butanetriol (BT) solution while maintaining a high (>60%) recovery of BT. Such
separations may be relevant to the production of BT from renewable
feedstocks,41 which is important because BT is a precursor for the high explosive

124-butanetriol trinitrate.
4.2 Experimental

Materials. Poly(styrene sulfonic acid) sodium salt (PSS, MW 125000, Alfa

Aesar), poly(diallyldimethylammonium chloride) (PDADMAC, Mw 100000-200000,

106

20 wt. % in water, Aldrich), poly(allylamine hydrochloride) (PAH, MW 70000,
Aldrich), polyethyleneimine (MW 25000, Aldrich), MnClz (Acros), NaBr (Spectrum),
NaCl (CCI), Na2804 (CCI), 1,2,4-butanetriol (Aldrich), glycerol (anhydrous, CCI),
glucose (Aldrich), sucrose (Aldrich), and rafﬁnose (Aldrich) were used as
received. Deionized water (Milli-Q, 18.2 MOcm) was used for membrane rinsing
and preparation of polyelectrolyte solutions.

Film Preparation. For swelling experiments, [PSS/PDADMAC]4PSS*
films (Film 1 from Table 4.1) were prepared on pieces of silicon wafers (Si(100),
Silicon Quest International) that were ﬁrst cleaned with UV/O3 (Boekel UV—Clean
model 135500) for 15 minutes. The cleaned silicon was immersed in 1 mg/mL
polyethyleneimine at pH 9 for 15 minutes to establish a dense, positively charged
layer, and ﬁlms were then built on this precursor layer.

Porous alumina substrates (0.02 pm Whatman Anodisc ﬁlters) were
initially cleaned with UV/Og for 15 min and then placed concentrate-side up in an
O-ring holder to limit ﬁlm deposition to the feed side of the membrane. To make
hybrid ﬁlms, the alumina supports were ﬁrst exposed to an aqueous solution of
0.02 M P88 in 0.1 or 0.5 M NaCl for 3 minutes and rinsed with deionized water
for 1 minute. (Polymer concentrations are always given with respect to the
repeating unit.) Next, the samples were exposed to a 0.02 M PDADMAC
solution in 0.1 M or 0.5 M NaCl for 3 min and rinsed again. This process was
repeated until the desired number of bilayers for the gutter-layer was formed. To

form the selective capping layer(s), the PDADMAC-terminated ﬁlms were

107

aces. s. oN FN :5 so: ooﬁoooo as; moo so 5:: ass.

.03 can .3 .5. ”2o; $3589: 229% 895 Go moz_>=oo_omo

saga/Beam: N 5820 so: on saga
.ch 31m :2 mm; mza

 

 

ooX oNuoo eonoo ooox vouooo :5 EN Non: woof/49mm: o
ooNA one. ooneo ooow woneoo NnoN one N.onN.N hoof/scams. o
N
oeuos Zoo Enos Noncoo N.ouN.oo out o snow Nono. F amwwmmdwnuwmwm: e
oouos ouoN moaNN nouNoo mouNoo oueo oNnoF N.ono.N amwmmwﬁmwwﬂ_ o
. . . . . . . . ammo N
.oonoo E”: o oao N N ouo oo o oaN so oueo sum: o ouo N saga/Beam:
. . . . .mmo
o Enos Nae e one a memo sumo onso oonoe o ouo a 40:20/Beams: F
omocfmm omocoaw 3820 A33 E: Ev
$8020 58020 3.820 $253. $226 $820 .9820 axe”. mozzwom was»... E_E E__n_
328.com 3: 8:88:

0r: *0 COEwOn—OU E 0mm: mm>> _OmZ S— md

.co_mm:om_.o 5 2m 2 98:5: :95 9m wEE .296. o<20<on=wmd

N.omEE 82968628 653.2: 36 38:? m ES, 858 mEEEm

moocoa 9.6: 83029: 5.58: :6 8:3 m a6 cosmozcocm: 80: 852623 Ucm .855 8:28 magnum—om ”:4 £33.

108

exposed to an aqueous solution of pH 2.1 0.02 M PSS in 0.5 M MnClz for 2
minutes, rinsed with deionized water for 1 minute, immersed in a solution of pH
2.3 0.02 M PAH with 0.5 M NaBr for 5 minutes, and again rinsed with water. This
procedure was repeated as necessary to form more capping bilayers. The ﬁnal
PSS layer of the PSS/PAH capping layers was deposited from 2.5 M MnCIz at pH
2.1 instead of 0.5 M MnCIz at pH 2.1 (thus, these ﬁlms are designated
[PSS/PDADMAC]m[PSS/PAH],.PSS*). Deposition of [PSS/PAH]4PSS* ﬁlms
utilized the same procedure as the synthesis of capping PSS/PAH layers of
hybrid ﬁlms. The ﬁnal PSS layer was again deposited from 2.5 M MI‘ICiz at pH
2.1 for 2 minutes, followed by a ﬁnal rinse.

Ellipsometry. Ellipsometric thicknesses of the Si02 layers on Si wafers
were ﬁrst determined assuming literature values for the refractive indices of Si
and SiOz at the 44 wavelengths of the ellipsometer (J.A. Woollam model M-44
rotating analyzer ellipsometer, 75° angle of incidence) between 414.0 nm and
736.1 nm. After coating of these wafers, ﬁlm thicknesses were obtained either in
ambient conditions (40% relative humidity (RH)), or under water using a home-
built cell with glass windows for swelling experiments. The ellipsometric
thicknesses and refractive indices of [PSS/PDADMAC]4PSS* (Film 1) ﬁlms were
determined at three different points per wafer on three different wafers, and the
reported results are the average and standard deviation of these values. Optical
constants of water as a function of wavelength were calculated using the Cauchy

equation in coordination with constants in the literature.“""43

109

Nanoﬁltration. NF was performed as previously reported (see Chapter
2).5'7 The feed for sugar/NaCl separation experiments with hybrid,
[PSS/PAH]4PSS*, and [PSS/PAH]4PSS ﬁlms contained 1 mM glycerol and
glucose, 5 mM sucrose and rafﬁnose, and 10 mM NaCl. For sugar separations
with PSS/PDADAMAC ﬁlms, the feed contained 1 mM glycerol, glucose, sucrose,
and rafﬁnose. For salt/sugar separations with pure PSS/PDADMAC the feed
was 1 mM sucrose and 10 mM NaCl, and for BT/salt separations, the feed was
either 1 mM BT and 2.5 mM Na2804 or 1 mM BT and 21.5 mM K2HPOa. The pH
of the BT/phosphate solution was adjusted to 7 or 8 by addition of 1 M HCI.

For all experiments, the feed was passed over the membrane at 18
mL/min, and the system was pressurized to 4.8 bar (70 psi). After 18 h of
equilibration time, membrane permeate was acquired for intervals of 15-30
minutes. Glycerol, glucose, sucrose, rafﬁnose, and BT concentrations were
determined via chromatography (Dionex, DX-600, CarbonPac PA-10 column,
mobile phase: 0.1 M NaOH at 1.0 mL/min) coupled with pulsed amperometric
detection (Dionex, ED-50). Single salts in solution (i.e. NaZSO4/BT separations)
were analyzed with conductivity measurements (Orion model 115).

Equation 4.1 deﬁnes solute rejection, where Creed is the concentration of
the solute in the feed, Cperm is the concentration of the solute in the permeate,
and R is the percent rejection of the solute. Equation 4.2 describes the
membrane selectivity between two different solutes, where RA is the percent

rejection of solute A, and R3 is the percent rejection of solute B.

110

Cperm 0

feed

CA,perm CB,feed =1OOO/O _ RA

Selectivity = o
CA,feed CB,perm 100 /° _ RB

 

(4.2)

4.3 Results and Discussion

This section ﬁrst presents the enhancement in selectivity that hybrid
[PSS/PDADMAC]m/[PSS/PAH],.PSS* and [PSS/PAH]4PSS* membranes provide
in NF of small neutral molecules such as sugars. Subsequent results show that
these ﬁlms are very promising for NaCl/sucrose separations, and ﬁnally, we
examine the performance of such ﬁlms in a speciﬁc potential application, removal
of salts from a solution containing BT.
Sugar Separations

Figure 4.2 shows a scanning electron microscope (SEM) image of a
[PSS/PDADMACMPSS/PAH]PSS* ﬁlm deposited on an alumina support. The
20 nm-diameter surface pores and 200 nm-diameter bulk pores of the alumina as
well as the 30 nm-thick polyelectrolyte skin layer are clearly visible. The image
suggests that hybrid ﬁlms fully cover underlying pores without ﬁlling them, and
the high rejections of rafﬁnose and sucrose in Table 4.1 are indicative of a ﬁlm
with very few defects.

Consistent with a size-exclusion mechanism, Table 4.1 shows that
PSS/PAH membranes (Film 6) show larger rejections of glucose, sucrose, and
rafﬁnose than do the lower charge density pure PSS/PDADMAC ﬁlms (Film 2).

Our goal in using hybrid and [PSS/PAH]4PSS* ﬁlms to separate neutral

111

Figure 4.2: Cross-sectional SEM image of a [PSS/PDADMAC]3[PSS/PAH]PSS*

ﬁlm deposited on porous alumina. The ﬁnal PSS layer was deposited from 2.5 M

MnClz.“

J
i
w
I
’5
f
‘ii'
¥

500 nm

 

112

molecules was to create a membrane that had the high rejections of pure
PSS/PAH ﬁlms along with the high ﬂuxes typical of pure PSS/PDADMAC
membranes. The data in Table 4.1 illustrate that [PSS/PAH]4PSS* ﬁlms barely
achieve this goal, but not hybrid ﬁlms, as their ﬂuxes are not statistically greater
than those through pure PSS/PAH ﬁlms. (The salt concentration in
PSS/PDADMAC deposition solutions did not have a signiﬁcant effect on neutral
molecule transport through these ﬁlms, so only the data with gutter layers
deposited from 0.5 M NaCl are presented here.)

There are small increases in glucose, sucrose, and rafﬁnose rejection with
[PSS/PDADMACMPSS/PAH]PSS*, [PSS/PDADMAC]3-[PSS/PAH]2PSS*, and
[PSS/PAHjaPSS* ﬁlms (Films 3, 4, and 5, respectively) compared to pure
PSS/PDADMAC ﬁlms (Film 2), although in some cases the differences are not
statistically signiﬁcant. For example, [PSS/PDADMAC]4PSS ﬁlms'(Film 2) reject
97% of sucrose, [PSS/PDADMACMPSSIPAH]PSS* membranes (Film 3) reject
98% of sucrose, and [PSS/PDADMAC]3[PSSIPAH]2PSS* as well as
[PSS/PAH]4PSS* assemblies (Films 4 and 5, respectively) reject 99% of sucrose.
Beneﬁcially, [PSS/PAH]4PSS* ﬁlms with 99+% sucrose rejection exhibit glucose
rejections of only 76%, leading to glucose/sucrose selectivities 3-times greater
than pure PSS/PDADMAC assemblies. Note that glucose rejection through
[PSS/PAH]4PSS* membranes (Film 5) is about 11% lower compared to typical
PSS/PAH ﬁlms (Film 6), almost doubling glucose recovery. [PSS/PAH]4PSS*

films also show increased rafﬁnose rejection relative to PSS/PDADMAC ﬁlms

113

(Film 2), resulting in glucose/rafﬁnose selectivities in excess of 250. Glycerol
rejection is below 30% for all modiﬁed ﬁlms.

Notably, the solution ﬂuxes through hybrid ﬁlms (Films 3 and 4) are
essentially unaffected by the addition of a second PSS/PAH capping layer. The
PSS/PDADMAC ﬁlms with no capping layers (Film 2) have a solution ﬂux of
2 m3/(m2 day), and the hybrid ﬁlms show at most a 10% loss in ﬂux. Though the
sucrose rejection increases due to capping layers, the cumulative transport
resistance due to both PSS/PDADMAC and the additional PSS/PAH layers
reduces the ﬂux to the point that it is indistinguishable from pure [PSS/PAHj4PSS
films. Thus, the beneﬁt of hybrid ﬁlms over pure PSS/PAH ﬁlms in sugar
separations is limited to a slight increase in glucose recovery (compare ﬁlms 3
and 5).

Considering the above results, we wondered if just depositing a single
layer of PSS from 2.5 M MnClz on top of a gutter layer is enough to create these
improved selectivities compared to pure PSS/PDADMAC ﬁlms. We deposited 4
bilayers of PSS/PDADMAC from 0.5 M NaCl and then capped the ﬁlms with a
single layer of PSS from 2.5 M MnClz at pH 2.1 (Film 1, [PSS/PDADMAC]4PSS*).
Ellipsometry experiments suggest that increasing the ionic strength of the ﬁnal
PSS deposition solution signiﬁcantly alters the composition of the ﬁlm. As shown
previously, depositing the ﬁnal PSS layer on top of a 4-bilayer PSS/PDADMAC
membrane (assembled in 0.5 M NaCl) from 0.5 M NaCl results in only a 20%
thickness increase (Chapter 2, Table 2.2).6 Surprisingly, when that PSS ﬁnal

layer is added from 2.5 M MnClz instead, ﬁlm thicknesses double from 21.6:06

114

nm to 44:3 nm. Ellipsometric swelling experiments suggest that
[PSS/PDADMAC]4PSS* ﬁlms (Film 1) swell almost 5 times as much in water as
4.5-bilayer PSS/PDADMAC assemblies where every layer is deposited from 0.5
M NaCl (Film 2). (Film 1 swells 380:70% while a version of Film 2 with more
bilayers swells 106:9%, see Chapter 3 Table 3.1.45)

In conjunction with this high swelling, the NF results for
[PSS/PDADMAClaPSS* from Table 4.1 (Film 1) depict lower rejections for all
sugar molecules. Interestingly, though initial solution ﬂuxes through these ﬁlms
were extremely high (>5 m3/(m2 day)), after 18 h of operation their ﬂux was 35%
less than [PSS/PDADMAC]4PSS ﬁlms deposited from 0.5 M NaCl (Film 2).
These ﬁlms likely undergo signiﬁcant compaction under pressure. Apparently,
PAH plays a critical role in establishing a hybrid system.

Salt-Sugar Separations

The data in Table 4.1, in conjunction with previous research,5'6 suggest that
modiﬁed ﬁlms should allow passage of NaCl while retaining sucrose and
rafﬁnose, which would be ideal for an application like salt recovery from ion-
exchange regeneration wastes.27 Table 4.2 shows the ﬂux and rejection data in
NF of solutions containing NaCl and sucrose by modiﬁed ﬁlms as well as pure
PSS/PAH and PSS/PDADMAC membranes. In the case of the pure ﬁlms,
[PSS/PAHj4PSS (Film 6) shows a higher sucrose rejection than
[PSS/PDADMAC]4PSS (Film 2), but the latter ﬁlms exhibit slightly less NaCl
rejection that might allow better salt recovery. Both types of modiﬁed ﬁlms

exhibit less salt rejection than PSS/PAH ﬁlms and higher selectivities than

115

Table 4.2: Rejections, solution ﬂuxes, and selectivities from nanoﬁltration of

sucrose and NaCl by porous alumina coated with a variety of PEM ﬁlms arranged

in order of increasing NaCl/sucrose selectivity.

deposition of the PSS/PDADMAC layers.

0.5 M NaCl was used in

Films are numbered to aid in

 

discussion.
. 3 NaCl Sucrose NaCI/
Film Film Type 8&3:ng 5:1?) Rejection Rejection Sucrose
y (%) (%) Selectivity
1 [PSS/PDAQMACL‘ 13:03 32:4 88:4 6:3
PSS
2 [PSS/PDAEBMACL‘ 20:03 40:4 93.8:08 10:1
PSS
[PSS/PDADMAClg
3 [PSS/PAH]PSS* 20:02 24:2 98.2:0.5 40:10
[PSS/PDADMAC]3
4 [PSS/PAH]2PSS* 18:02 23:3 99.2:0.2 100:20
5 [PSS/PAH]4PSS* 22:02 29:3 99.4:0.1 130:30
6 [PSS/PAH]4PSS 17:02 50:9 99.6:0.4 170:70

aNF was run at 4.8 bar.

bThese PSS/PDADMAC data are from Chapter 2.6

*Final layer of PSS was deposited from pH 2.1 2.5 M MnClz.

116

PSS/PDADMAC ﬁlms, as [PSS/PDADMAC]3[PSS/PAH]2PSS* and
[PSS/PAH]4PSS* (Films 4 and 5, respectively) show a NaCl rejection of only 25-
30% and a NaCl/sucrose selectivity of 100 or more. The high NaCl passage
makes them more attractive than [PSS/PAH]4PSS for NaCl-sugar separations.
Selectivity increases 2.5-fold on going from [PSS/PDADMAC]3[PSS/PAH]PSS* to
[PSS/PDADMAC]3[PSS/PAH]2PSS* (Films 3 and 4, respectively). For all
modiﬁed ﬁlms, flux is ~2 m3/(m2 day), about twice the ﬂux through commercial NF

membranes?"39

The remarkable ﬂux, NaCl/sucrose selectivity, and NaCl
recovery possible with these ﬁlms make them attractive for salt recovery
applications.
Removal of sulfate and phosphate salts from solutions containing
butanetriol

The ﬁnal application explored here involves purification of BT from
solutions containing sulfate and phosphate salts. This separation is relevant to
collection of BT from a fermentation broth“ in which magnesium sulfate, sulfuric
acid and potassium phosphate are added for pH adjustment but are
contaminants of the end product. Frost and coworkers recently showed that
production of this explosive precursor from renewable feedstocks may be
preferable to the current synthetic method involving the reduction of esteriﬁed
D,L-malic acid.“ In this separation, instead of rejecting an organic molecule and

allowing salt to pass as described above, we want to design a ﬁlm to reject

divalent salts and allow a small organic molecule (BT) to pass.

117

Table 4.3 shows the rejections, solution ﬂuxes, and BT/sulfate selectivities
for a variety of PEM ﬁlms. Pure PSS/PAH ﬁlms were not explored for this
application as previous data show that divalent anions are less than 60%
rejected with these assemblies.5 Initial experiments with 4.5-bilayer membranes
of PSS/PDADMAC deposited from 0.5 M NaCl (Film 2) showed high ﬂuxes and
low BT rejection (2.3 m3/(m2 day) and 20%, respectively), but, unfortunately, only
75°/o sulfate rejection. Deposition of the terminating PSS layer from 2.5 M MnCIz
at pH 2.1 (instead of 0.5 M NaCl) (Film 1) gave a signiﬁcantly lower ﬂux of
1.5 m3/(m2 day), the same BT rejection, and 91% sulfate rejection. The higher
sulfate rejection was encouraging, but previous reports using PSS/PAH ﬁlms with
terminating PSS layers deposited from 2.5 M MnClz suggested that modiﬁed
ﬁlms would lead to considerably higher sulfate rejection.5

The sulfate rejection with [PSS/PDADMAC]3[PSS/PAH]PSS* ﬁlms (Film 3)
was 96% while the BT rejection was 40%, and ﬂuxes were an adequate 1.45
m3/(m2 day). [PSS/PAH]4PSS* ﬁlms (Film 5) produced nearly identical rejection
data, but with surprisingly higher ﬂuxes than the hybrid ﬁlm systems. Previous
data suggest that gutter layer thickness does affect ﬂux (solution ﬂux through
[PSS/PDADMAC]5[PSS/PAH]PSS* ﬁlms is only 33% that of
[PSS/PDADMAC]2[PSSIPAH]PSS* assemblies),44 but thinner PSS/PDADMAC
gutter layers may not cover the alumina pores completely, which could lead to

sulfate leakage.

118

Table 4.3: Rejections, solution ﬂuxes, and selectivities from nanoﬁltration of BT

and sodium sulfate using porous alumina coated with a variety of PEM ﬁlms. All

PSS/PDADMAC layers were deposited from 0.5 M NaCl unless indicated

otherwise. Films are numbered to aid in discussion.

 

Solution BT Sulfate
Film Film Type Fluxa Rejection Rejection 2:123:33
(m3/(m2 day» 1%) <%>
1 [PSS/PP%’:9MAC]4 15:05 20:20 91:4 12:7
2 [PSS/PEQEMACL‘ 23:03 20:10 75:4 34:09
3 [Tgsséfgpﬁrggﬁ 1.45:0.07 40:20 96.2:0.9 20:10
5 [PSS/PAHlaPSS* 24:03 30:10 96.5:0.8 22:6

aNF was run at 4.8 bar.

*Final layer of PSS was deposited from pH 2.1 2.5 M MnClz

119

 

A second aspect of this separation, removal of phosphate salts, was more
difﬁcult. Because phosphate is a trivalent base, its charge is a function of pH,
and ion-exclusion separations can be challenging. Schlenoff and coworkers
previously reported that control over pH can have a dramatic effect on the
transport of weak acids.46 Initial work with PEM ﬁlms and pH 7 solutions resulted
in only 50% rejection of the total phosphate species because at pH 7 the
phosphate is roughly half HPo42' and half H2POa'.

Table 4.4 shows the phosphate rejection, BT rejection, and solution ﬂux
for hybrid and [PSS/PAHlaPSS* ﬁlms after adjusting the solution pH to 8. At this
point the phosphate is approximately 90% HPOaz‘, and rejection is about 90%
with hybrid ﬁlms. Similar phosphate rejections were observed for
[PSS/PDADMAC]3[PSS/PAH]PSS* and [PSS/PDADMAC]3[PSS/PAH]2PSS*
membranes (Films 3 and 4), and the flux was 1.6-1.7 m3/(m2 day) for both ﬁlms.
[PSS/PAH]4PSS* ﬁlms (Film 5) also exhibit low BT rejection, but unfortunately
they reject 15% less phosphate than hybrid ﬁlms. The lower phosphate rejection
of [PSS/PAH]4PSS* ﬁlms could result from slow ﬁlm decomposition at pH 8, as
individual [PSS/PAH]4PSS* membranes with higher ﬂux possessed lower
phosphate rejection. For example, a membrane with an average ﬂux of
1.9 m3/(m2 day) exhibited a phosphate rejection of 86%, while a ﬁlm with an
average ﬂux of 2.4 m3/(m2 day) possessed a 69% phosphate rejection. The
gutter layers may supply additional stability to the hybrid systems, as the
PSS/PDADMAC layers are composed of strong polyelectrolytes and thus are pH

tolerant. Separations at even higher pH values where nearly all phosphate

120

Table 4.4: Rejections, solution ﬂuxes, and selectivities from nanoﬁltration of BT
and potassium phosphate at pH 8 using porous alumina coated with PEM ﬁlms.
All PSS/PDADMAC layers were deposited from 0.5 M NaCl. Films are numbered

to aid in discussion.

 

Solution Fluxa BT Phosphate BT/
Film Film Type (m3/(m2 da )) Rejection Rejection Phosphate
y (%) (%) Selectivity
[PSS/PDADMAC]3
3 [PSS/PAH]PSS* 17:02 30:10 89:1 7:1
[PSS/PDADMAC13
4 [PSS/PAH]2PSS* 16:01 10:10 90:1 9:1
5 [PSS/PAH14PSS* 20:04 32:8 75:11 4:2

aNF was run at 4.8 bar.

*Final layer of PSS was deposited from pH 2.1 2.5 M MnClz

121

species would be highly charged are challenging, as complete decomposition of
the PEMS or dissolution of the alumina support may occur.

The best choice of PEM for the removal of sulfate and/or phosphate from
a BT solution depends on a number of factors. If sulfate anions are the
contaminant of interest, then [PSS/PAH]4PSS* assemblies are the best choice,
as they possess higher ﬂuxes than hybrid ﬁlms (Table 4.3). However, if removal
of phosphate is paramount, then hybrid ﬁlms would be better choices, as they
reject the most phosphate. A case where both salts are present is more
problematic. If high ﬂuxes are more critical than permeate quality, then
[PSS/PAH]4PSS* ﬁlms would sufﬁce. However, if ﬂux was not as important as
obtaining low permeate salt concentrations, then hybrid ﬁlms are superior as they
reject 96% of sulfate and 90% of phosphate anions while rejecting only 40% of

BT.

4.4 Conclusion

This chapter describes two types of modiﬁed ﬁlms that attempt to improve
NF ﬂuxes and selectivities in three diverse, model applications. In both sugar
and sugar/salt separations, PSS/PAH ﬁlms where the terminating PSS layer is
deposited from 2.5 M MnClz exhibit greater ﬂuxes than pure PSS/PAH ﬁlms while
maintaining similar selectivities. Hybrid ﬁlms, where selective PSS/PAH layers
are deposited on presumably more permeable PSS/PDADMAC, are capable of
higher selectivities than pure PSS/PDADMAC ﬁlms, but the ﬂuxes are not

statistically greater than those through pure PSS/PAH ﬁlms. In the puriﬁcation of

122

BT, both types of modiﬁed ﬁlms exhibit BT/sulfate selectivities of twenty, though
[PSS/PAHI4PSS* membranes have higher ﬂuxes. However, in BT/phosphate
separations, hybrid ﬁlms Show double the selectivity of [PSS/PAH]4PSS* ﬁlms, as
the latter is likely not as tolerant of the slightly basic pH. Overall, hybrid ﬁlms are
better suited to remove salts from BT/salt solutions, while [PSS/PAH]4PSS*

assemblies excel at sugar and sugar/salt separations.

4.5 Acknowledgements

We thank the Division of Chemical Sciences in the Ofﬁce of Basic Energy

Sciences of the Department of Energy for ﬁnancial support of this work.

123

4.6 Works Cited

(1)
(2)

(3)
(4)
(5)

(10)

(11)

(12)

(13)
(14)

(15)

Decher, G. Science 1997, 277, 1232-1237.

Harris, J. J.; Stair, J. L.; Bruening, M. L. Chem. Mater. 2000, 12, 1941-
1946.

Jin, W.; Toutianoush, A.; Tieke, B. Langmuir 2003, 19, 2550-2553.
Krasemann, L.; Tieke, B. Mat. Sci. Eng. C-Bio S1999, C8-C9, 513-518.

Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003,
19, 7038-7042.

Miller, M. D.; Bruening, M. L. Langmuir 2004, 20, 11545-11551.
Liu, X.; Bruening, M. L. Chem. Mater. 2004, 16, 351-357.
Hollman, A. M.; Bhattacharyya, D. Langmuir 2004, 20, 5418-5424.

Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, 8. A.; Méhwald, H.
Angew. Chem, Int. Ed. 1998, 37, 2202-2205.

Caruso, F .; Trau, D.; Mtihwald, H.; Renneberg, R. Langmuir 2000, 16,
1485-1488.

Diaspro, A.; Silvano, D.; Krol, 8.; Cavalleri, O.; Gliozzi, A. Langmuir 2002,
18, 5047-5050.

Trau, D.; Yang, W.; Seydack, M.; Caruso, F.; Yu, N.-T.; Renneberg, R.
Anal. Chem. 2002, 74, 5480-5486.

Yu, A.; Caruso, F. Anal. Chem. 2003, 75, 3031-3037.

Shchukin, D. G.; Shutava, T.; Shchukina, E.; Sukhorukov, G. B.; Lvov, Y.
M. Chem. Mater. 2004, 16, 3446-3451.

Lajimi, R. H.; Ben Abdallah, A.; Ferjani, E.; Roudesli, M. S.; Deratani, A.
Desalination 2004, 163, 193-202.

Krasemann, L.; Tieke, B. J. Membr. Sci. 1998, 150, 23-30.

Krasemann, L.; Toutianoush, A.; Tieke, B. J. Membr. Sci. 2001, 181, 221-
228.

Meier-Haack, J.; Lenk, W.; Lehmann, D.; Lunkwitz, K. J. Membr. Sci.
2001, 184. 233-243.

Toutianoush, A.; Tieke, B. Mat. Sci. Eng. C 2002, C22, 459-463.

124

(20)
(21)

(22)
(23)
(24)
(25)
(25)

(27)

(28)

(29)

(30)

(31)
(32)
(33)
(34)

(35)

(36)

Lenk, W.; Meier-Haack, J. Desalination 2002, 148, 11-16.

Budd, P. M.; Ricardo, N. M. P. S.; Jafar, J. J.; Stephenson, 8.; Hughes, R.
Ind. Eng. Chem. Res. 2004, 43, 1863-1867.

Sullivan, D. M.; Bruening, M. L. J. Membr. Sci. 2005, 248, 161-170.
Schwarz, H. H.; Malsch, G. J. Membr. Sci. 2005, 247, 143-152.
Quinn, R. J. Membr. Sci. 1998, 139, 97-102.

Sullivan, D. M.; Bruening, M. L. Chem. Mater. 2003, 15, 281-287.

Dai, J.; Jensen, A. W.; Mohanty, D. K.; Erndt, J.; Bruening, M. L. Langmuir
2001, 17, 931-937.

Hinkova, A.; Bubnik, Z.; Kadlec, P.; Pridal, J. Sep. Purif. Technol. 2002,
26, 101-110.

Gyura, J.; Seres, Z.; Vatai, G.; Molnar, E. B. Desalination 2002, 148, 49-
56.

Goulas, A. K.; Grandison, A. S.; Rastall, R. A. J. Sci. Food Agr. 2003, 83,
675-680.

This estimate is based upon the ﬂux of sugar solutions through 4.5-bilayer
and 5.5-bilayer PSS/PDADMAC ﬁlms deposited from 0.5 M NaCl. These
ﬁlms have thicknesses of 25 and 35 nm under ambient conditions,
respectively, according to Table 2.2 in Chapter 2. However, the data in
this chapter suggest that hybrid ﬁlms have no ﬂux advantage over the
thinner 4.5-bilayer PSS/PAH ﬁlms. This is because 4.5-bilayer PSS/PAH
ﬁlms have almost double the ﬂux of previously reported 6.5-bilayer
assembiles.,

Krasemann, L.; Tieke, B. Chem. Eng. Technol. 2000, 23, 211-213.
Dubas, S. T.; Schlenoff, J. B. Langmuir 2001, 17, 7725-7727.
Miller, M. D.; Bruening, M. L. Chem. Mater., Submitted for Publication.

Stair, J. L.; Harris, J. J.; Bruening, M. L. Chem. Mater. 2001, 13, 2641-
2648.

Sullivan, D. M.; Bruening, M. L. J. Am. Chem. Soc. 2001, 123, 11805-
11806.

Kharlampieva, E.; Sukhishvili, S. A. Langmuir 2004, 20, 10712-10717.

125

(37)

(38)

(39)

(40)

(41)

(42)

(43)

<44)

(45)

(45)

Garza, J. M.; Schaaf, P.; Muller, 8.; Ball, V.; Stoltz, J.-F.; Voegel, J.-C.;
Lavalle, P. Langmuir 2004, 20, 7298-7302.

Bhattacharyya, D., et al, In Membrane Handbook; Ho W.S.W, Sirkar, K.K.,
Eds.; Van Nostrand Reinhold; New York, 1992.

Hutchinson, R. A.; Beuermann, S.; Paquet, D. A., Jr.; McMinn, J. H.;
Jackson, C. Macromolecules 1998, 31 , 1542-1547.

Mendelsohn, J. D.; Barrett, C. J.; Chan, V. V.; Pal, A. J.; Mayes, A. M.;
Rubner, M. F. Langmuir 2000, 16, 5017-5023.

Niu, W.; Molefe, M. N.; Frost, J. W. J. Am. Chem. Soc. 2003, 125, 12998-
12999.

Schiebener, P.; Straub, J.; Sengers, J. M. H. L.; Gallagher, J. S. J. Phys.
Chem. Ref. Data 1990, 19, 677-717.

Harvey, A. H.; Gallagher, J. S.; Sengers, J. M. H. L. J. Phys. Chem. Ref.
Data 1998, 27, 761-774.

Stanton, B. W., Ultrathin, Multilayered Polyelectrolyte Films as
Nanoﬁltration Membranes, MS. Thesis, Department of Chemistry,
Michigan State University, 2003

Please note that the swelling for Film 2 is relative to dry ﬁlms in <5% RH,
while the swelling for Film 1 is relative to dry ﬁlms in 40% RH.). However,
ﬁlm thicknesses increase in the presence of water vapor, as ellipsometry
experiments with 4-bilayer PSS/PDADMAC ﬁlms show that they swell
20% on going from <5% to 55% RH. Thus, Film 1 likely possesses greater
than 380% swelling,

Rmaile, H. H.; Farhat, T. R.; Schlenoff, J. B. J. Phys. Chem. B 2003, 107,
14401-14406.

126

Chapter 5

CONCLUSIONS AND FUTURE WORK

5.1 Conclusions

This body of work demonstrates that novel polymeric ﬁlms are capable of
effecting separations in nanoﬁltration (NF). Polyelectrolyte multilayers (PEMS)
assembled by a Simple layer-by-layer technique can cover porous alumina
supports with just a few bilayers. In addition, Simple adjustments to synthesis
variables such as the constituent polyelectrolytes, deposition conditions, and the
capping-layer choice (polyanion or polycation) result in ﬁlms with molecular
weight cutoffs (MWCOS) ranging from 100 to >20000. These PEMS possess
higher NF ﬂuxes than commercially available membranes and can separate
sugars, salts from sugars, and even proteins.

This dissertation also investigates how swelling of PEMS relates to
polyelectrolyte structure and solute transport. In situ ellipsometry experiments
reveal that the swelling of PEMS in water is directly related to polyelectrolyte
charge density. Assembling PEM ﬁlms from polyelectrolytes that have a low
charge density results in ﬁlms that swell more, which in turn leads to higher
MWCOs. PEM swelling can also be top-layer dependent, as poly(styrene
sulfonate) (PSS)/poly(diallyldimethylammonium chloride) (PDADMAC) ﬁlms swell
4 times more when the polycation is the capping layer as opposed to the

polyanion. Additionally, swelling correlates to permeability, as ﬁlms with greater

127

water uptake have lower NF rejections and higher sugar ﬂuxes than PEMS that
swell less.

Moreover, one can tailor membranes for speciﬁc applications. The
selectivities and overall ﬂuxes in oligosaccharide fractionation and the separation
of salt from sugar can be optimized by simple adjustments to the PEM deposition
system. For example, the NaCl/sucrose selectivity of PSS/PDADMAC ﬁlms can
be increased by a factor of 10 by simply capping these coatings with 1.5-bilayers

of PSS/poly(allylamine hydrochloride) (PAH).

5.2 Future Work

There are a number of research paths available with PEM ﬁlms. One of
the most interesting directions involves optimization of the previously described
hyaluronic acid (HA)/chitosan ﬁlms. Chapter 2 describes them as possessing a
MWCO between 20 and 70 kDa with ﬂuxes much higher than any other
membrane system studied. As the functional groups of these polyelectrolytes
are primary amines and carboxylic acids, such ﬁlms may be susceptible to
crosslinking. Formation of covalent amide bonds would likely limit the mobility of
the polymer chains, reducing swelling and decreasing the MWCO. Thus,
stoichiometric or temporal control of the crosslinking reaction could result in
various transport properties. The crosslinking could be performed chemically via
an 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC)/N-hydroxysuccinimide
(NHS) coupling as shown in Figure 5.1.1 Reaction progress could be monitored

by changes in ﬁlm swelling or by FT-IR studies of crosslinked ﬁlms deposited

128

 

O OH
0 0171—9 EDC 5" N=(
HO 0 o OH
HO OH O NHCOCH

Hyaluronic Acid HO OH O NHCOCH3
HA
o
NHS o 0’ 8 OH
__> o
0H0 o 0.];

HO OH O NHCOCH3

Chitosan HO NH
——'> O OH

HO OH O NHCOCH3

Figure 5.1: Simpliﬁed mechanism for the crosslinking reaction of HA and

chitosan by EDC/NHS coupling.

129

on Al or Au-coated silicon wafers. Heating the PEM may be another way to
produce a crosslinked system,2 though membrane decomposition may be a
concern.

Another interesting application of PEMS would be their use as size-
exclusion stationary phases in electrochromatography.3'4 Capillary
electrophoresis cannot separate neutral solutes without buffer additives. Even
with the addition of a stationary phase, neutral solutes with Similar polarities may
still be challenging to resolve. However, utilizing the Size-selective properties of
PEMS could facilitate these separations by excluding larger solutes while
allowing small molecules to partition into the stationary phase, increasing their
retention time. Additionally, PEMS are charged so electroosmosis could still be
employed as a driving force for ﬂow.5'6 PSS/PDADMAC ﬁlms could be ideal for
this application, as they do not reject small organics like glycerol, but exclude
sucrose and rafﬁnose (see Tables 2.2 and 2.3). However, thick ﬁlms with a
sufﬁcient phase ratio to effect separations may take considerable time to deposit.
To combat this possible limitation, exponentially growing PSS/PDADMAC ﬁlms
deposited from high ionic strength could be used,7 though the transport
properties of these ﬁlms are currently unknown. Additionally, non-polar solutes
could be driven into the stationary phase by modifying the PEM ﬁlm with
hydrophobic groups.8

One ﬁnal project involving PEMS would be a study of the effect of ionic
strength on the transport of neutral molecules. Several previous studies report

that PEM swelling is directly related to the ionic strength of the surrounding

130

solutionf"11 and Chapter 3 of this thesis suggests that transport properties are a
strong function of swelling. Thus, the ionic strength of the feed solution should
directly affect the Sieving properties of a PEM ﬁlm. In commercial applications a
wide range of solution conditions may be encountered, thus knowing how
rejection and ﬂux depend on changes in ionic strength is vital.

Overall, there are numerous avenues for the future exploration of PEMS.
The straightforward manipulation of PEM permselectivities as well as their ease
of deposition makes them highly attractive for practical applications. With
continued research and development polymeric ﬁlm systems have the capacity to

beneﬁt separations of all sizes.

131

5.3 Works Cited

(1) Sehgal, D.; Vijay, l. K. Anal. Biochem. 1994, 218, 87-91.

(2) Harris, J. J.; DeRose, P. M.; Bruening, M. L. J. Am. Chem. Soc. 1999,
121, 1978-1979.

(3) Ding, F.; Stol, R.; Kok, W. T.; Poppe, H. J. Chromatogr., A 2001, 924, 239-
249.

(4) Stol, R.; Pedersoli, J. L., Jr.; Poppe, H.; Kok, W. T. Anal. Chem 2002, 74,
2314-2320.

(5) Sui, Z.; Schlenoff, J. B. Langmuir 2003, 19, 7829-7831.

(6) Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003,
19, 7038-7042.

(7) McAloney, R. A.; Sinyor, M.; Dudnik, V.; Goh, M. C. Langmuir 2001, 17,
6655-6663.

(8) Dai, J.; Balachandra, A. M.; Lee, J. I.; Bruening, M. L. Macromolecules
2002, 35, 3164-3170.

(9) Sukhorukov, G. B.; Schmitt, J.; Decher, G. Berichte der Bunsen-
Gesellschaft 1996, 100, 948-953.

(10) Gao, C.; Leporatti, S.; Moya, S.; Donath, E.; Méhwald, H. Chem.--Eur. J.
2003, 9, 915-920.

(11) Burke, S. E.; Barrett, C. J. Biomacromolecules 2005, 6, 1419-1428.

132

 

 

 

 

   

ljlljjjiijjijjjjji