as.“ . , .Lhﬁ
: “31.: , ‘ r -‘ Gui-ﬁleg e .
aiwﬁﬁ; . V . wﬂywmuﬁ .ymﬂwwwigmg
_ . 3.5mm, ﬁwmﬂi «gig.
. 9 .Q . w! s: 3‘ ‘1
. ﬂakym ,. Lg.
3 1...! 11:; I

5

Sn

AL 33:.
hit: i6.
thrill lawn
«mint.- .3 to

. .3

v2.2.2
2.. J.” r532»...

{51
x
I. J . I.

3 3.

‘
V; . .2. H... t . n
y 1 .19 it$3vur¥hwvo .

3:: Mn. .

, vi... .1 v

3.3!. «hammsizgﬁ.

0. E“ til l9 0.
Emmi... .a g? km...

r59? x. 4

. U'Iﬁ‘I'l! .
r». ‘sum‘ﬂ‘k SK“. — Q 9.» I in.
3 .

331-..... $63 . a
z in... 5.32.. a . i
. l ‘KL.

9. I: I Zunw a

saint... tilt Jr
'35. :I—‘ﬁ‘ ‘11 4.?-

! i4

.3
f'1‘l';
I‘.IV‘1‘UI
.u it: 071),
3 4...! 3.1!. x .23:
313.21... 3.1).; ttl!
. 1:: iv: t v 1.
a; sA A. ,
zf Ssnvéﬂm at
,ﬂiulh! 2-.
3‘: 61:

at. . .‘1. .
5:925:7193
\ﬂsi"1..l
. Vila

.
!.-4hdn~:.’xll- Q
:25 O . .I
o .. f..3.r:1:‘lv‘(.v u
l: :3‘1245) 3.:S:
921 VI. in
«Ru ,

. .

‘ .32. lsvvlr:
.. 1...»? 3
34.

I.
i.mku«m~rs.§s .ﬁmEaﬁuhii. , amggkgﬁmﬂ.
.. ‘ . . z .. hr .7

i.
4 .x 2.2” 5. S
.3» 53.61;}. 1

.c :z A

 

l 71435.:

200?

This is to certify that the
dissertation entitled

POLYMER BRUSH MEMBRANES FOR
PERVAPORATION AND HIGH CAPACITY PROTEIN
BINDING

presented by

LEI SUN

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemistry

 

 

       
    

51%

M

ajor Professor’s S' ‘

Adm aklj aogg

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

..--- n...- .---0- u

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 DA'I'E DUE

tug? '51; 30MB

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/CIRC/DateDue.indd-p.1

 

POLYMER BRUSH MEMBRANES FOR PERVAPORATION
AND HIGH CAPACITY PROTEIN BINDING
By

Lei Sun

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2006

ABSTRACT
POLYMER BRUSH MEMBRANES FOR PERVAPORATION AND HIGH
CAPACITY PROTEIN BINDING
By

Lei Sun

Atom transfer radical polymerization (ATRP) Of 2-hydroxyethyl methacrylate
from the surface of porous alumina supports yields polymer-brush membranes that can be
readily modiﬁed to control their properties. Derivatization of poly(2-hydroxyethyl
methacrylate) (PHEMA) coatings with octyl, hexadecyl, or pentadecaﬂuorooctyl side
chains provides ﬁlms that are sufﬁciently hydrophobic to allow highly selective
pervaporation of volatile organic compounds from water. Moreover, the polymerization
and derivatization reactions readily yield defect—free ﬁlms with thicknesses of 200 nm or
less, and this minimal thickness results in ﬂuxes of ~1.4 kg/mzh (0.05 wt% VOCs at 22
°C) through ﬂuorinated ﬁlms. These ﬂuxes are generally an order of magnitude greater
than those through high-performance poly(dimethyl siloxane) pervaporation membranes.
The use of different monomers may further enhance the ﬂux and selectivity of
pervaporation membranes prepared by ATRP.

In related work, ATRP of PHEMA brushes inside the pores of alumina supports
followed by functionalization of PHEMA with nitrilotriacetate-Cu2+ complexes yields
membranes that adsorb proteins via coordination of Cu2+ to histidine residues.
Adsorption isotherms show that these membranes have binding capacities as high as 0.9

mg of bovine serum albumin (BSA) per cm2 of external membrane surface area (150

mg/cm3 Of membrane), and breakthrough curves indicate that saturation of the
membranes with BSA or myoglobin occurs in less than 15 min. Nine cycles of protein
loading, elution, and membrane regeneration result in no detectable loss of binding
capacity. The unusually high capacity Of these membranes for rapid protein binding
makes them attractive for applications such as puriﬁcation of his tagged proteins.
Expansion on the protein adsorption work demonstrates that polymer brushes
derivatized with Ni2+ complexes are very attractive for puriﬁcation of his tagged proteins.
Protein binding experiments with ﬁlms on gold substrates show that PHEMA-NTA-Ni2+
brushes are capable Of selectively binding large amounts (3.1 ug/cmz) of his tagged
ubiquitin (HisU). Porous alumina membranes modiﬁed with derivatized brushes also
show remarkable capacities (0.72 mg/cmz) for the binding of HisU. Most importantly,
gel electrophoresis studies show that these high-capacity membranes are highly selective
for puriﬁcation Of his tagged proteins. Even in a 20-fold excess of BSA, membranes

gave 99% HisU.

To My Husband

iv

ACKNOWLEDGMENTS

The ﬁrst person I would like to thank is my advisor Merlin Bruening. I joined his
group in 2002 and have been working with him for 4 years. During these four years, I
have known him as a pleasant person, a great advisor as well as an excellent scientist. His
understanding, encouraging and personal guidance have provided a good basis for the

present thesis. He is the most considerate advisor a student could ever ask for.

I also acknowledge Professor Baker who as my second supervisor provided

tremendous help tO me.

Dr. Jinhua Dai has been a great mentor to me. He, as a true scientist, has taught me

how to do great research. I dedicate my greatest acknowledgement to him.

Thank my committee members Dr Blanchard, Dr Borhan and Dr Swain for their kind

help and great suggestions.

I would like to give my thanks to all Bruening group members. The last four years has
been a wonderful experience for me to be with talented people like Malai, Dan, Matt, Sri,

Jamie, Christin, Lu, Maneesha, Parul, Fei and David.

Last but not least, I would like to express my deepest gratitude to my husband. I am

very grateful to him for his long-distance support.

TABLE OF CONTENTS

Page

List Of Tables ........................................................................................ ix

List of Schemes ...................................................................................... x
List of Figures ........................................................................................ xi
List Of Abbreviations .............................................................................. xiv
Chapter 1. Introduction and Background ................................................... 1
1.1. Polymer Brushes ........................................................................... 1
1.1.1. Deﬁnition and Assets of Brushes ..................................................... 1
1.1.2. Applications of Brushes ............................................................... 2
1.1.2-a. Polymer Brushes as Lithographic Resists .................................... 2
1.1.2-b. Polymer Brushes as Adhesion Promoters .................................... 9
1.1.2-c. Polymer Brushes as Lubricants ............................................... ll
1.1.2—d. Polymer Brushes as Stabilizers of Colloidal Particles .................... 12
1.1.2-e. Polymer Brushes as Nonfouling Coatings .................................. 15

1.1.2-f. Polymer Brushes as Responsive Materials ................................. 17

1.2. Growth of Polymer Brushes from Surfaces ........................................... 19
1.2.1. Surface Initiated Polymerization .................................................... 19
1.2.2. Surface Initiated Free Radical Polymerization .................................... 20
1.2.3. Surface Initiated Anionic Polymerization ......................................... 25
1.2.4. Surface Initiated Cationic Polymerization ......................................... 26
1.2.5. Surface Initiated Ring-opening Metathesis Polymerization ..................... 28
1.2.6. Surface Initiated Controlled Radical Polymerization ............................. 31
1.2.7. Surface Initiated Atom Transfer Radical Polymerization ....................... 34
1.3. Introduction to Pervaporation Membranes ............................................ 39
1.4. Importance of Membrane Absorbers ................................................... 42
1.5. Outline of the Dissertation ............................................................... 43
1.6. References ................................................................................. 45
Chapter 2. Polymer Brush Membranes for Pervaporation ............................ 53
2.1. Introduction ................................................................................ 53
2.2. Experimental Section ..................................................................... 55
2.2.1. Materials ................................................................................ 55
2.2.2. Polymerization of HEMA and Subsequent Derivatization ...................... 56
2.2.3. Characterization of Membranes ...................................................... 57

vi

2.2.4. Pervaporation Experiments ............................................................. 57

2.2.5. Sorption Measurements ................................................................. 59
2.3. Results and Discussion ...................................................................... 61
2.3.1. FTIR and SEM Characterization of Derivatized PHEMA Membranes ......... 61
2.3.2. Pervaporation of Ethyl acetate-water Mixtures Using Derivatized
PHEMA .................................................................................... 65
2.3.2-a. Effect of Feed Concentration ................................................... 65
2.3.2-b. Effect of Feed Temperature .................................................... 70
2.3.3. Pervaporation of Different Organic Compounds through Derivatized
PHEMA Membranes............ 73

2.3.4. Sorption Behavior and the Solution-diffusion Model.................................74
2.3.5. Comparison with Related Membrane Systems76

2.4. Conclusions ............................................................................... 78
2.5. References ................................................................................. 80

Chapter 3. High-Capacity, Protein-Binding Membranes Based on Polymer

Brushes Grown in Porous Substrates ........................................ 82

3. 1. Introduction ................................................................................ 82

3.2. Experimental Section ..................................................................... 84

3.2.1. Materials .................................................................................... 84
3.2.2. Polymerization of HEMA in Porous Alumina Membranes and

on Au Substrates ........................................................................ 85

3.2.3. PHEMA Derivatization and Protein Immobilization ............................ 85

3.2.4. Determination of Protein Concentrations .......................................... 88

3.2.5. Film Characterization Methods ..................................................... 88

3.3. Results and Discussion ................................................................... 89

3.3.1. FTIR Characterization Of PHEMA Derivatization and Protein Binding. . .....89
3.3.2. SEM and EDS Characterization of PHEMA inside Alumina Membranes....93

3.3.3. BSA Binding to PHEMA-NTA-Cu2+ Brushes .................................... 96
3.3.4. Elution of BSA and Reuse of Alumina-PHEMA-NTA-Cu2+ Membranes...102
3.3.5. Adsorption of Myoglobin .......................................................... 104
3.4. Conclusions .................................................................................. 105
3.5. References .................................................................................... 106
Chapter 4. Puriﬁcation of Histidinex-Tagged Proteins Using Membranes
Modiﬁed with Polymer Brushes ................................................ 109
4. 1 . Introduction .................................................................................. 109
4.2. Experimental Section ................................................................... 112
4.2.1. Materials ................................................................................. 1 12
4.2.2. Polymerization of HEMA in Porous Alumina Membranes
and on Au Substrates ............................................................... 112

vii

4.2.3. PHEMA Derivatization and Protein Immobilization .......................... 1 14

4.2.4. Determination of the Amount of Coordinated Ni2+ in the Membrane ...... 115
4.2.5. Determination of Protein Concentrations ......................................... 115
4.2.6. Determination Of Protein Purity by SDS-PAGE ............................... 115
4.2.7. Film Characterization Methods ................................................... 116
4.3. Results and Discussion ................................................................. 116
4.3.1. Characterization of PHEMA Derivatization ..................................... 116
4.3.2. HisU Binding to PHEMA-NTA-Ni2+ Brushes on Gold Substrate ............ 121
4.3.3. HisU Binding to PHEMA-NTA—Ni2+ Brushes in Membranes. . . . . . . . 123
4.3.4. Separation of Protein Mixtures. . . . . . . .. ......................................................... 127
4.4. Conclusions... .. . .. . ........................................................................................... 129
4.5. References... .. . . . . ............................................................................................. 130
Chapter 5. Summary and Future Work .................................................. 132

viii

Table 2.1.

Table 2.2.

Table 2.3.

Table 2.4.

Table 3.1.

LIST OF TABLES

Molecular properties of several VOCS and separation factors
for pervaporation (22 °C) of 0.05% solutions through C8-PHEMA,

C16-PHEMA, and ﬂuorinated PHEMA membranes. 73
Degrees of sorption and sorption selectivities for ﬁve VOCS (0.05%) in
C8-PHEMA, Cl6-PHEMA, and fluorinated PHEMA ﬁlms. 74
Separation factors (22 °C) for pervaporation (apv), sorption (as),

and diffusion (0.1)) of VOCS (0.05% aqueous solutions) in C8-PHEMA,
C16-PHEMA, and ﬂuorinated PHEMA membranes. 75
Comparison of the pervaporation performance Of ﬂuorinated PHEMA

and several PDMS membranes. ........................................................ 77
Comparison of the protein—binding capacity of

alurnina—PHEMA-NTA-Cu2+ membranes with other

membrane absorbers. ........................................................................ 101

ix

Scheme 1.].

Scheme 1.2.

Scheme 1.3.

Scheme 1.4.

Scheme 1.5.

Scheme 1.6.

Scheme 1.7.

Scheme 1.8.

Scheme 1.9.

Scheme 1.10.

Scheme 1.11.

Scheme 1.12.

Scheme 2.1.

Scheme 3.1.

Scheme 4.1.

LIST OF SCHEMES

Diagram of the reversible formation of intermolecular hydrogen

bonds between PNIPAM chains and water molecules and

intramolecular hydrogen bonding between C=O and N—H groups

in PNIPAM chains below and above the LCST. 18

Photochromic transformation of spiropyran. . ........................................ 19

Schematic description of the preparation of polymer brushes by
radical polymerization from covalently immobilized initiators. ............ 23

Free radical polymerization from initiator layers attached to

cross-linked thiol monolayers on Au. ............................................ 24
Schematic illustration of anionic SIP from clay surfaces. 25
Surface initiated cationic polymerization of 2-Oxazolines. 27

Stepwise fabrication of polymeric nanostructures by surface initiated
ROMP on SiOz nanopattems generated by anodization lithography ....... 30

Synthesis of polystyrene brushes by TEMPO-mediated radical
polymerization ........................................................................................ 32

General process of surface-initiated RAFT polymerizations using
2-phenylprop-2-y1 dithiobenzoate as the chain transfer agent and a

substrate-immobilized azo initiator. .............................................. 33
Mechanism of ATRP. ................................................................... 35
Ligands used for copper-mediated ATRP. .................................... 36

Schematic illustration of the immobilization of ATRP initiators using
the LB technique. .......................................................................... 37

Attachment of a trichlorosilane initiator to a porous alumina
support, ATRP from the immobilized initiator, and derivatization
of PHEMA with acid chlorides ................................................. 54

Derivatization of PHEMA for protein immobilization ...................... 87

Derivatization Of PHEMA with NTA-Ni2+ before protein absorption... 1 13

Figure 1.1.
Figure 1.2.

Figure 1.3.

Figure 1.4.

Figure 1.5.

Figure 1.6.

Figure 1.7.

Figure 1.8.

Figure 2.1.

Figure 2.2.

Figure 2.3.

Figure 2.4.

Figure 2.5.

Figure 2.6.

LIST OF FIGURES

Polymer brushes on a ﬂat or spherical surface. .................................... l
The Microcontact printing process ................................................................ 4
Schematic diagram showing the formation of patterned

surfaces by capillary force lithography and SIP. .................................. 8
Chemical structure Of PMMA-b-PSGMA copolymer and schematic
representation of Fshca, and Emma. between the brush-bearing surfaces ......... 12
Swelling-based method for preparing stable, functionalized

polymer colloids. ............................................................................... l4
Schematic representation of protein molecules in contact with a
polymerbrush. ................................................................................... 15
Schematic diagrams of (a) a pervaporation system and (b) a

cross-sectional view of the membrane cell Of the apparatus... . ... .......4.....()
Schematic diagram of a solution-diffusion model. 40
Reﬂectance FT IR spectra of PHEMA ﬁlms on gold before and after
derivatization with acid chlorides. ..................................................... 63
Cross-sectional FESEM images of alumina supports coated with

PHEMA, C8-PHEMA, C16-PHEMA, and ﬂuorinated PHEMA. ..............64
Permeate composition, separation factor, and ﬂux in pervaporation

of several ethyl acetate/water mixtures through C8-PHEMA

membranes at 50 °C. .......................................................................... 66
Permeate composition, separation factor, and ﬂux in pervaporation

of several ethyl acetate/water mixtures through C16-PHEMA

membranes at50 °C. .......................................................................... 67
Permeate composition, separation factor, and ﬂux in pervaporation

Of several ethyl acetate/water mixtures through ﬂuorinated-PHEMA
membranes at 50°C. ............................................................................ 68
Flux and separation factor for pervaporation of water/ethyl acetate

mixtures through C8-PHEMA membranes at several temperatures. ...........72

xi

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 3.4.

Figure 3.5.

Figure 3.6.

Figure 3.7.

Figure 3.8.

Figure 3.9.

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.

Figure 4.6.

Reﬂectance FTIR spectra of a PHEMA ﬁlm before and after

several derivatization steps. ............................................................... 89
Reﬂectance FTIR spectra of a PHEMA ﬁlm (3750 to 1250 cm")

on Au before and after derivatization with succinic anhydride. . 9.0
FESEM images of PHEMA nanotubes synthesized by ATRP

in porous alumina substrates. ............................................................. 93
Energy dispersive X-ray Spectrum of PHEMA nanotubes grown

by ATRP in porous alumina membranes. . . . .. ........................................... 95
UV-vis spectra of the permeate collected at certain time intervals

when 0.56 mg/mL BSA solution was pumped through porous

alumina membranes coated with PHEMA-NTA- Cu ....96
Breakthrough curve for BSA adsorption Inzporous alumina

membranes coated with PHEMA-NTA- Cu .............................................. 97
Equilibrium binding capacity of PHEMA-NTA-Cu2+-coated

alumina membranes as a function of BSA concentration. 98
Binding amounts and elution efﬁciencies for several cycles of BSA
adsorption and elution followed by Cu2+ reloading of an alumina

membrane modiﬁed with PHEMA-NTA .................................................. 102
Breakthrough curves of 0.3 mg/mL BSA and 0.35 mg/mL myoglobin

on the same alumina-PHEMA-NTA- Cuz” membrane” ... .. ...104
Selective binding of PHEMA-NTA-Ni2+ to a his tagged protein ................. 111
Transmission IR of PHEMA ﬁlms inside an alumina membrane

before and after several derivatization steps. . . . . . . .. ................................... 117
XPS spectrum of a PHEMA-NTA-Ni2+ ﬁlm on a gold substrate ................. 119
Determination of the amount of immobilized N i2+ in a

PHEMA—NTA-Ni2+ ﬁlm in the membrane ................................................... 120
Subtracted reﬂectance FT IR spectra of protein immobilized on
PHEMA-NTA-Ni2+ ﬁlms after exposure of the ﬁlms to

(a) 0.01 mg/mL HisU or (b) 1 mg/mL BSA ................................................. 122

Breakthrough curves for adsorption Of 0. 3 mg/mL HisU,
0.3 mg/mL BSA, and 0.35 mg/mL myoglobin In membranes
modiﬁed with PHEMA-NTA-Ni2+ (HisU) or

xii

Figure 4.7.

Figure 4.8.

Figure 5.1.

PHEMA—NTA-Cu2+ (BSA and myoglobin). .................................... 124

UV-vis spectra Of (a) a 0.2 mg/mL BSA solution; (b) eluent

(imidazole solution) from an alumina-PHEMA-NTA—Ni2+membrane;

(c) a 0.2 mg/mL myoglobin solution; ((1) Eluent (imidazole solution)

from an alumina-PHEMA-NTA-Ni2+membrane. 126

SDS-PAGE analysis (Silver staining) of protein solutions and eluents
from alumina-PI-IEMA-NTA-Ni2+ membranes loaded with proteins. ...... 128

FESEM image of a PVDF membrane with a nominal cutoff of
0.45 pm and an alumina membrane with pore size of 0.25 pm ............... 132

xiii

AFM
AIBN
ATRP
BSA
DMAP
DMF
DPE
EDC
EDTA
ENB

FT IR
GPS
IMAC
LCST
NHS

N MP
ODMS
ODT
OEG
PAAm
PDMS
PE

PEG
PHEMA
PMMA
PNIPAM
PPEI

PS

PSA
PSGMA
PVDF
RAFT
RGD
ROMP
SA
SAM
SIP
TEMPO
VOCS

LIST OF ABBREVIATIONS

Atomic Force Microscopy
Azobisisobutyronitrile

Atom transfer radical polymerization
Bovine serum albumin
4-dimethylaminopyridine
Dimethylformamide

1,1-Diphenylethylene
1-[3—(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride
Ethylenediamine tetraacetic acid
5-Ethy1idene-2-norbomene

Fourier transform infrared spectroscopy
3-glycidoxypropyltrimethoxysilane
Immobilized metal-afﬁnity chromatography
Liquid Critical Solution Temperature
N-hydroxysuccinimide

Nitroxide-mediated polymerization
Octadecylmethyldiethoxysilane
Octadecanethiol

Oligo(ethylene glycol)

Poly(acrylamide)

Polydimethylsiloxane

Polyelectrolyte

Poly(ethylene glycol)

Poly(2-hydroxyethyl) methacrylate
Poly(methyl methacrylate)
Poly(N-isopropylacrylamide)
Poly(N-propionylethylenimine)

Polystyrene

Pressure-sensitive adhesive

Poly(sodium sulphonated glycidyl methacrylate)
Poly(vinylidene ﬂuoride)

Reversible addition fragmentation chain transfer
Arginine-glycine—aspartic acids
Ring-opening metathesis polymerization
Succinic anhydride

Self assembled monolayer

Surface initiated polymerization

2, 2, 6, 6-Tetramethyl—1-piperidinyloxy
Volatile organic compounds

xiv

Chapter 1 Introduction and Background

This dissertation describes the growth of polymer brushes on and in porous
supports to form high—ﬂux pervaporation membranes as well as membrane absorbers
capable of purifying proteins modified with histidinex tags. The research builds on
previous studies of the synthesis and application of polymer brushes, so to put this work
in perspective, this introduction ﬁrst describes polymer brushes and several of their
extensively investigated applications. Subsequently, I discuss a number of recent
techniques for synthesizing polymer brushes in a controlled fashion. Sections 1.3 and 1.4
of the introduction then present brief background on the importance of pervaporation
membranes and membrane absorbers, respectively, and show why polymer brushes might
be attractive for these applications. Finally, I present an outline of the dissertation.

1.1. Polymer Brushes
1.1.1. Deﬁnition and Assets of Brushes
Polymer brushes are assemblies of macromolecules bound to a substrate with a

chain density high enough to force the polymers to extend from the surface.1 The brushes

 

Figure 1.1. Polymer brushes on a ﬂat or spherical surface.

can be held on the surface by either physisorption or covalent bonding, usually near one
end of the polymer chain.2 Because of their surface conﬁnement, in appropriate solvents
polymer brushes are extended away from the rigid substrate boundary and can be highly
swollen, which makes them attractive platforms for capture and puriﬁcation of proteins,
as demonstrated herein}5 The high swelling of polymer brushes should also allow
molecules to rapidly diffuse through them, which is important for both rapidly capturing
biomolecules and allowing fast transport of small molecules in pervaporation separations.
An additional asset of polymer brushes is that they contain a high density of functional
groups that can be readily derivatized to tailor brush properties for speciﬁc applications.
1.1.2. Applications of Brushes

Polymer brushes ﬁrst found application in the early 19505, when Clayﬁeld and
coworkers grafted poly(ethylene oxide) onto colloidal particles to effectively prevent
ﬂocculation.6 More recently, brushes have been employed as materials for lithographic
masks, adhesion promoters, lubrication, nonfouling coatings, and responsive surfaces, as
described below.2
1.1.2-a. Polymer Brushes as Lithographic Resists

Lithography is a process used in semiconductor device fabrication to transfer a
pattern from a photo mask to the surface of a wafer. Patterned polymer brushes have
drawn signiﬁcant attention for lithography because of the diversity of chemical structures
suitable for brush formation and the physical/mechanical robustness of the grafted ﬁlms.
Micropatteming of functional polymer brushes on substrates is of crucial importance to
the development of biochips, microarrays, and microdevices for cell-growth, regulation

of protein adsorption, and drug delivery.7 Ober and coworkers prepared Oligo(ethylene

glycol) (OEGn)-containing styrene polymer brushes and compared the biocompatibility
of these brushes with that of assemblies of surface-bound OEGn-terminated silanes with
selected chain lengths.8 The monomer (4-(Oligoethyleneoxy)oxymethylstyrene) was
prepared by reaction of 4-chloromethylstyrene and monomethoxy OEG derivatives of
varying lengths in the presence of sodium hydride. Surface initiated nitroxide-mediated
polymerization was used to polymerize this monomer to form polymer brushes. Polymer
brushes possessed a superior ability to inhibit protein adsorption and cell adhesion when
compared to surface assemblies with the same OEGn length. They attributed this result to
the greater thickness and surface coverage of polymer brushes with respect to monolayers
of OEGn—terminated silanes.

Various techniques have been reported for fabricating patterned polymer brushes,
including microcontact printing, UV/electron—beam lithography, scanning-probe
lithography, and imprint lithography.9 Microcontact patterning (also called soft
lithography) was principally introduced by Whitesides and coworkers and is a
remarkably simple technique that allows generation of patterned surfaces in any

laboratory.IO In this procedure (Figure 1.2),

 

Figure 1.2. The Microcontact printing process. (a) a patterned PDMS
stamp is placed in ethanol containing octadecanethiol (ODT); (b) ODT from
the solution lightly adsorbs to the PDMS stamp; (c) The PDMS stamp with
the CDT is placed on the gold substrate; ((1) The stamp is removed and the
CDT remains adsorbed to the gold. (Redrawn from Angew. Chem. Int. Ed.

1998, 37, 550—575.)

an elastomeric polydimethylsiloxane (PDMS) stamp is used to transfer a monolayer of
molecules of an ink to a surface at contacted regions. After printing, a different self
assembled monolayer (SAM) can be formed on the unmodiﬁed regions of the surface by
exposing the patterned substrate to a dilute solution containing a second adsorbing
molecule. If one of the two adsorbed molecules is an initiator of polymerization, polymer
brushes can be grown from these initiators to amplify the surface pattern.”I3 Hawker
and coworkers prepared patterned brushes using atom transfer radical polymerization
(ATRP) and speculated that the brushes would provide a signiﬁcantly more robust resist
than a simple monolayer.ll Huck and coworkers successfully grew thick poly(N-
isopropylacrylamide) (PNIPAM) brushes using surface initiated polymerization (SIP)
from patterned monolayers to prepare smart micro-pattemed domains.l4 PNIPAM is a
very interesting material that is widely investigated due to its inverse solubility-
temperature relationship in water, i.e., PNIPAM dissolves in water at low temperature,
but becomes insoluble when the solution temperature reaches the liquid critical solution
temperature (LCST), 32 °C. Atomic force microscopy (AFM) experiments showed that
patterned PNIPAM chains grown from a surface display a reversible phase transition
accompanied by measurable changes in adhesion between the surface and the AFM tip. A
hydrophilic to hydrophobic change occurs when raising the temperature from 25 °C to 35
°C.

Zauscher and coworkers reported a “top-down/bottom-up” technique to fabricate
patterned polymer brush arrays on the micrometer and nanometer length scales.‘5 Their
approach combined the traditional electron beam lithography technique (top down) with

SIP (bottom up). The primary advantage of electron beam lithography is that it provides a

way to overcome the diffraction limit of light to prepare features in the sub-um regime.
Electron beam widths were projected to be on the order of nanometers as of the year 2005.
The electron beam can also be replaced with UV irradiation.'6 The combination of SIP
and electron beam lithography gives control over polymer functionality, shape, feature
dimension and inter-feature spacing on the nanometer length scale. This technique
yielded patterned polymer brushes with a height of about 300 nm and width of 1.8 pm.

Scanning probe lithography is an emerging area of research in which the scanning
tunneling microscope or AFM is used to pattern nanometer-scale features. Zauscher
presented a strategy to fabricate stimulus—responsive, surface-conﬁned PNIPAM brush
nanopattems prepared in an approach that combines “nanoshaving”, a scanning probe
lithography method, with SIP.”

Another mechanism for formation of polymer-brush nanopattems is a new
nonconventional lithographic method called nanoimprint lithography.l8 This method is
regarded as one of the “10 emerging technologies that will change the world” by MIT
Technology Review.'9 The key advantage of this lithographic technique is the ability to
pattern sub—25 nm structures over a large area with a high-throughput and low—cost.20 In
the ﬁrst step of nanoimprint lithography, a mold with nanostructures on its surface is
pressed into a thin resist cast on a substrate, followed by removal of the mold. This step
duplicates the nanostructures on the mold in the resist ﬁlm. The second step is the pattern
transfer where an anisotropic etching process, such as reactive ion etching is used to
remove the residual resist in the compressed area. Carter and his coworkers patterned
polystyrene (PS), poly(methyl methacrylate) (PMMA), and poly(2-hydroxyethyl)

methacrylate (PHEMA) into submicrometer sized features on a silicon wafer using a

combination of nanoimprint lithography and living free radical polymerization.”
(lnitiators were attached to exposed areas of the pattern and polymerization from these
initiators followed.) They also grew polydihexylfluorene brushes from a nanoimprint-
pattemed surface using metal-catalyzed step-growth condensation polymerization.19
Polymer brushes of well-deﬁned nanoscopic structure, molecular weight, and
controllable molecular architecture were obtained and have potential application in a
variety of different areas such as molecular-scale electronics, magnetic storage,
optoelectronics, and biotechnology.21

In one ﬁnal approach to the use of polymer brushes to amplify surface patterns,
Luzinov and coworkers developed the use of capillary force lithography and SIP to
generate patterned binary brushes containing poly(ethylene glycol) methyl ether
methacrylate and PNIPAM.9 They found that when an ultrathin PS pattern was deposited,
over a layer of polymerization initiator already anchored to a surface (Figure 1.3), the
pattern could withstand subsequent polymerization conditions and stay intact during
brush synthesis, provided that no solvent for PS was involved. Thus, no monomer and
catalyst were delivered to the initiator located under the PS-protected areas. They utilized
this interesting phenomenon and achieved a variety of combinations of binary patterned

polymer brushes.

— PDMS mold
_ .— PS thin film

Macroinitiator

 

1 Heating, T>Tg

   

1 Cooling and mold removal

 

1 First SIP
/ PNIPAM

 

1 PS removal

 

1 Second SIP

 

"tote: .

Figure 1.3. Schematic diagram showing the formation of patterned
surfaces by capillary force lithography and SIP. (Redrawn from J. Am. Chem.

Soc. 2006, 128, 8106-8107.)

1.1.2-b. Polymer Brushes as Adhesion Promoters

Adhesion at polymer/polymer and polymer/solid interfaces is of great importance
for numerous applications from microelectronics to the aircraft industry.22 For both
permanent and reversible adhesion, it is essential that the chemical composition and
morphology of the interfaces to be joined are well controlled. Polymer brushes provide a
controlled and tailorable surface, and because they are often covalently bonded to
surfaces, they can promote adhesion to the substrates on which they reside.

Creton and coworkers investigated the adhesive properties of binary
heterogeneous polymer brushes made from a mixture of end-functionalized PS and
poly(2-vinylpyridine) chains.22 The adhesive properties were tested using a soft
hydrophobic pressure-sensitive adhesive (PSA), which contains a symmetric (PS-
polyisoprene-PS) triblock copolymer and a low molecular weight, high-Tg resin that is
completely miscible with the isoprene phase. To test the adhesive properties of the
binary polymer brushes, a silicon wafer coated with the binary brush ﬁlm is glued on the
end of a cylindrical stainless steel probe. The probe then approaches and comes in contact
with the PSA layer deposited on a glass microscope slide. By observing the events
occurring at the adhesive/brush interface with a CCD camera and recording the
stress/strain curve, the adhesive properties of the binary brushes were investigated. The
adhesive properties of the binary brush structure can be reversibly switched by exposing
the brush to solvents that bring PS or poly(2-vinylpryridine) to the surface. When PS is at
the surface, adhesion to the hydrophobic PSA is strong, while little adhesion occurs when
poly(2-vinylpyridine) is at the surface. This class of adhesives is gaining increasing

interest in industry and medicine for its low toxicity and ease of use. However, more

work is needed on model systems to better understand the subtle interactions taking place
at the interface between a brush with mixed composition and an adhesive.

Polymer brushes can also be modiﬁed to control cellular adhesion. The ability to
spatially control such adhesion on a biocompatible substrate is an important factor in
designing new biomaterials for use in wound healing and tissue engineering applications.
Since arginine-glycine-aspartic acids (RGD) were found to promote cell adhesion in 1984,
numerous materials have been functionalized with RGD for academic studies or medical
applications.23‘ 24 Current efforts to precisely engineer cell-adhesive or non-adhesive
surfaces often focus on including the ROD peptide sequence in a SAM.” 26 Metter’s
group recently developed a novel method for creating RGD-ligand density gradients in
anionic polymer brushes (poly(methacrylic acid) salts) to control the spatial attachment
of ﬁbroblasts across patterned surfaces.” By utilizing a controlled photopolymerization
technique, polymer brushes exhibiting spatially deﬁned gradients in chain density are
created. Subsequently these brushes are functionalized with RGD to create ligand density
gradients capable of inducing cell adhesion on an otherwise weakly adhesive substrate.
Their studies of cell adhesion indicated that these ligand-functionalized surfaces are
noncytotoxic, with cellular adhesion increasing with ROD-ligand density across the
gradient of polymer brush density.

Polymer brushes are much more ﬂexible than SAMs because of their long chain
lengths. The ﬂexibility and high density of functional groups in polymer brushes, as well
as the ability to modify the surfaces of both inorganic and polymeric materials, give these
materials signiﬁcant advantages over SAMs for the preparation of ligand-functionalized

materials.26

10

1.1.2-c. Polymer Brushes as Lubricants

The use of lubricants to reduce friction between rubbing surfaces has been
documented since antiquity.28 More recently, research has focused on boundary
lubrication by surfactant-like species coating the surfaces. Klein and coworkers ﬁrst
discovered that polymer brushes can facilitate the sliding of one surface over another in
1994.29 They determined the normal (Fnormal) and shear (Fshcar) forces between curved,
molecularly smooth mica surfaces using a surface forces apparatus, and then attached PS
brushes to the mica surfaces and measured the shear forces again.30 The shear forces with
polymer brushes were orders of magnitude lower than the forces needed for sliding the
bare surfaces. The attachment of polymer brushes resulted in an entropy-driven long-
range equilibrium separation of the surfaces, which gave decreased friction. Theoretical
studies suggested only a limited interpenetration of the brushes, so as the surfaces slide
past each other, there is rapid relaxation of the interpenetrating chains and low friction.

After proving that the presence of neutral polymer brushes may lead to a massive
reduction in sliding friction between the surfaces to which they are attached, Klein further
showed that hydrated ionic brushes can also act as extremely efﬁcient lubricants between
sliding charged surfaces.30 They created a diblock copolymer brush of poly(methyl
methacrylate) and poly(sodium sulphonated glycidyl methacrylate) (PMMA-b-PSGMA)
on mica. The copolymer contains hydrophobic (PMMA) and hydrophilic polyelectrolyte
(PSGMA) blocks (Figure 1.4). Effective friction coefﬁcients with these polyelectrolyte
(PE) brushes are lower than about 0.0006 — 0.001 even at low sliding velocities and at
pressures of up to several atmospheres. These low values of the effective friction

coefﬁcient reveal a remarkable capacity for lubrication by the PE, di—block brushes. The

11

Hydrophobic Hydrophilic
f—JTr—‘A—ﬂ

ragge— 3+
115

 

OC/ \
i ‘t’ 0
CH3 CH2 \ A
a HiH—OH b I Fshear
2 |:normal
SOa'Na“

Figure 1.4. (a) Chemical structure of PMMA-b-PSGMA copolymer. (b)
Schematic representation of Fshcar and Fnonm. between the brush-bearing
surfaces. (Redrawn from J. Polym. Sci., Part B: Polym. Phys. 2005, 43, 193-

204.)

authors attributed this to the exceptional resistance to mutual interpenetration displayed
by the compressed, counterion-swollen brushes, together with the ﬂuidity of the
hydration layers surrounding the charged, rubbing polymer segments. They also
mentioned that these charged polymer brushes may have applications as biolubricants in
artiﬁcial implants.
1.1.2-d. Polymer Brushes as Stabilizers of Colloidal Particles

Interactions between polymer-coated colloidal particles have been an important
issue for both industrial applications and scientiﬁc investigations.31 These interactions
form one of the bases of steric stabilization of colloidal dispersions that are widely used

in industry.32 Since Clayﬁeld and coworkers found that polymer brushes can effectively

prevent ﬂocculation and maintain the stability of colloids,6 much research has been
performed in this area.”37
In the absence of a macromolecular covering, ﬂocculation of colloids occurs
rapidly because attractive van der Waals forces between the colloids cause them to
aggregate. However, when two polymer-coated particles approach, a reduction in the
number of conﬁgurations available to the ﬂexible polymer chains gives rise to an
“entropic” repulsive force between the particles, which may keep colliding particles
separated such that the van der Waals interaction energy is insufﬁcient for coherence.38
The Birshtein group found that colloidal stability is signiﬁcantly inﬂuenced by a
mechanism called capillary condensation in a binary solvent system. Capillary
condensation occurs in porous particles when multilayer adsorption from a vapor
proceeds to the point at which pore spaces are ﬁlled with liquid separated from the gas
phase. In particular, when the particles are somewhat hydrophilic, there will be a
molecularly thin ﬁlm of water adsorbed onto their surfaces. When two particles come
near to each other, there will be a thick water ﬁlm between the two particles.
Consequently, the particles cannot separate without breaking the water ﬁlm, and such
ﬁlms can hold particles together. Birshtein’s group referred to this process as capillary
condensation mediated ﬂocculation and showed theoretically that one way to protect
particles from capillary condensation ﬂocculation is to graft hydrophobic polymer

brushes onto their surfaces.37

13

  
   
 
  

PEG 6PEG +

PPG

 

F-108 Pluronic Pal/Styrene particle

Swell
with toluene

    
 

#

Strip toluene
(heat to 96 °C)

 

Figure 1.5. Swelling—based method for preparing stable, functionalized

polymer colloids. (Redrawn from J. Am. Chem. Soc. 2005, 127, 1592-1593.)

Very recently, Crocker and coworkers used a swelling-based method to prepare
stable, brush-functionalized polymer colloids (Figure 1.5). They ﬁrst adsorbed block
copolymer surfactants, poly(ethylene glycol)—b-poly(propylene g1ycol)-b-poly(ethylene
glycol), from deionized water onto the surfaces of the PS particles and then swelled the
particles with a small amount of toluene, allowing the hydrophobic block of the
surfactant to penetrate the surface. Finally, they deswelled the particles by stripping the
solvent. Optical microscope images show the high stability of these particles. Unlike
physical adsorption, the swelling method permanently anchors the molecules to the
polymer surfaces, and the poly(ethylene glycol) chains of the anchored polymer provide a

steric barrier to aggregation, even at high ionic strength.33

1.1.2-e. Polymer Brushes as Nonfouling Coatings

Biofouling is a ubiquitous phenomenon that can occur on surfaces ranging from
living marine organisms (known as epibiosis) to synthetic surfaces such as separation
membranes. Most frequently biofouling is an undesirable process that needs to be
prevented. For example the buildup of biofoulant ﬁlms in groundwater wells can limit
the rate of water recovery. Elimination of biofouling will require biocompatible materials
that eradicate, or largely reduce, the nonspeciﬁc adsorption of proteins and other
biomolecules. In recent years special attention has been given to the modiﬁcation of
protein—surface interactions using grafted polymer brushes.39 It is believed that the role

of the polymers tethered to the surface is to present a steric barrier to approaching

Protein

    

Grafted
O‘— polymer

Surface

Figure 1.6. Schematic representation of protein molecules in contact with
a polymer brush. (Redrawn from Curr. Opin. Solid State Mater. Sci. 1997, 2,

337—344.)

proteins (Figure 1.6).40 To be speciﬁc, the protein approaches the polymer surface by
diffusion and is affected by the van der Waals attraction between the polymer brush and
protein through water. Further approach of the protein initiates the compression of
polymer chains, which induces a steric repulsion effect; an additional van der Waals
attraction becomes important between the substrate and protein through the water
solvated polymer brush layer. The van der Waals component with the substrate decreases
with increasing surface density and chain length of terminally attached polymer chains.
Thus, high density of polymer brushes tends to prevent more protein adsorption.

Surfaces coated by terminally grafted, water soluble chains are highly resistant to
protein adsorption in aqueous media.“ In 1997, Szleifer wrote a review article discussing
the understanding of the ability of tethered polymer chains to prevent protein adsorption
along with the up-to—date experimental status of these systems. Some theoretical studies
that enable calculation of the interactions of proteins with tethered polymer layers are
also included in the review. In the last 10 years, experimental and theoretical studies that
utilize polymer brushes to prevent nonspeciﬁc protein adsorption appeared
continuously.” 4' 46

Among all the polymer brushes used for antifouling, poly(ethylene glycol) (PEG)
is the most widely used. The ability of PEG ﬁlms to reduce the adsorption of proteins
onto surfaces has opened a broad ﬁeld of applications in biotechnology.40 Both the ability
of PEG to prevent protein adsorption and its low toxicity have resulted in various
attempts to create surfaces presenting PEG to the environment.45

Chilkoti et al synthesized PEG brushes from gold using SIP and examined the

adsorption of different proteins onto these modiﬁed gold slides.42 They observed no

16

protein adsorption from pure solutions of ﬁbronectin or 10% or 100% fetal bovine serum.
Wang and coworkers extended the formation of brushes using SIP to poly(vinylidene
ﬂuoride) (PVDF) substrates.43 The ﬂuorine atoms on the polymer surface served as direct
initiation sites. Then they used ATRP to prepare poly(2—(N, N-dimethylamino) ethyl
methacrylate) and poly(ethylene glycol) methacrylate brushes from PVDF. Protein
adsorption experiments revealed substantial resistance to fouling for both grafted ﬁlms in
comparison with the native PVDF surface.
1.1.2-f. Polymer Brushes as Responsive Materials

Some polymer brushes can act as materials that respond to environmental
conditions such as pH, temperature or photo-irradiation, and PE brushes are particularly
notable in this regrard.47 The conformation and swelling of PE brushes strongly depends
on the ionic strength, valence of counter ions, and the pH of the solutions to which they
are exposed. Stamm et a1. synthesized a mixed PE brush of carboxy terminated
poly(acrylic acid) and poly(2-vinylpyridine)) on silicon wafers. They ﬁrst absorbed 3-
glycidoxypropyltrimethoxysilane (GPS) on the silicone wafer. Then a grafted layer of
carboxy terminated poly(tert—butyl acrylate) was prepared via spin coating of this
polymer on the GPS modiﬁed Si wafer and subsequent annealing in a vacuum oven at
150 °C for 20 min. Carboxy terminated poly(2-vinylpyridine) was then grafted on the
same surface using a similar procedure. Finally, hydrolysis of poly(tert—butyl acrylate)
yielded poly(acrylic acid). The charge of the mixed PE brush can switch rapidly with a
change in pH, and the swelling of the brushes also changes dramatically. The

responsive/switching behavior of the PE brush can be explored to tune surface properties

17

in aqueous environments by applying a pH signal, which is of potential interest for
microﬂuidic technologies, smart nanodevices, and drug delivery systems.48
PNIPAM exhibits large swelling changes in aqueous media in response to small

changes in temperature (Scheme 1.1). As mentioned before, the LCST of PNIPAM is

0". 0"
u H u
} vii __ rir_< p N _ H-<
Heating up ‘ Ca 0
V Cooling down P N ' P N

b ‘4-

 

 

 

 

> LCST

< LCST

Scheme 1.1. Diagram of the reversible formation of intermolecular
hydrogen bonds between PNIPAM chains and water molecules (left) and
intramolecular hydrogen bonding between C=O and N—H groups in PNIPAM
chains (right) below and above the LCST. (Redrawn from Angew. Chem. Int.

Ed. 2004, 43, 357-360.)

near 32 °C. PNIPAM chains hydrate to form expanded structures in water when the
solution temperature is below its LCST but become compact structures by dehydration
when heated up above the LCST.49 Temperature-responsive properties of PNIPAM
brushes have been utilized in a variety of applications including controlled drug

delivery50 and solute separation.5 '

l8

 

 

_ UV A
O N O 0 N02 Vis, heat
I
R
Spiropyran form Merocyanine form

Scheme 1.2. Photochromic transformation of spiropyran. (Redrawn from

Macromolecules 1998, 31, 2606-2610.)

To realize a photo response in polymer brushes, Imanishi et al. grafted a spiropyran-
containing methacrylate-MMA block copolymer onto a glass ﬁlter. The
photoisomerization of the Spiropyran groups (Scheme 1.2) might induce structural
variations in the macromolecular chain, and it also yields changes in ﬁlm color.52 The
appearance of a blue color upon ultraviolet light irradiation indicated isomerization of the
pendant Spiropyran groups to merocyanine groups, and this color change occurred
reversibly upon ultraviolet and visible light irradiations. Toluene permeation through the
glass ﬁlter modiﬁed with the grafted copolymer increased upon ultraviolet-light

irradiation, and decreased after visible-light irradiation.

1.2. Growth of Polymer Brushes from Surfaces
1.2.1. Surface Initiated Polymerization

Because polymer brushes have many potential applications, the preparation of
these materials on solid substrates has been investigated extensively. Successful
synthetic techniques for forming brushes can be divided into two large categories:

“ raftin to” methods53 and “ raftin from” methods.“ 55 In “ raftin to” a roaches,
‘8 g g 8 g g PP

l9

preformed end functionalized polymer molecules react with functional groups on a
substrate to form polymer brushes. The “grafting to” technique results in relatively low
densities of polymer chains on the surface because steric hindrance prevents incoming
polymer chains from diffusing through the ﬁlm to reactive sites on the substrate.

When higher grafting densities are needed, the “grafting from” approach is the
method of choice for brush formation. “Grafting from” is accomplished by covalently
anchoring initiators on the substrate and then activating these initiators to start
polymerization. The polymer brushes grown from the surface by this technique possess
high chain densities (~1 chain per 2 nmz) because monomers, rather than polymer chains
must diffuse to the surface for ﬁlm growth to occur. 3

“Grafting to” can be accomplished with a number of polymerization techniques
including free-radical, cationic, and anionic polymerization, as well as ring-opening
metathesis polymerization (ROMP), ATRP, nitroxide-mediated polymerization NMP),
and reversible addition fragmentation chain transfer (RAFT) polymerization. The
following sections will focus on the synthesis of polymer brushes using these various SIP
techniques.

1.2.2. Surface Initiated Free Radical Polymerization

Free radical polymerization is a relatively versatile tool to grow polymer brushes
on a surface because it allows the use of a variety of monomers, including those with
polar and unprotected functional groups.56 Immobilized azo compounds are usually used
as initiators in surface initiated free radical polymerization because they can be readily
activated at low temperatures or with UV light. To immobilize the initiator, a

difunctional anchor molecule is frequently linked to the surface of the substrate and the

20

initiating azo species is linked to this anchor molecule in one or several additional steps.57
However, if the surface reactions are not quantitative, the stepwise generation of the
initiator monolayers can lead to low initiator densities that consequently reduce the
density of brushes.

To avoid these problems, Riihe and coworkers developed a system in which the
complete initiator is attached to the substrate in one step as shown in Scheme 1.3.55' 58’ 59
The initiator system contains an anchoring group (A) that connects the molecule to the
surface, the initiator itself (I-I*), and a cleavable group (CD) that allows for the
detachment of the polymer brushes after polymerization. (Detachment of the brushes is
useful for analysis of the polymer composition using methods such as gel-permeation
chromatography.) The complete initiator is ﬁrst self assembled on the substrate, and in a
subsequent reaction the initiator is activated and polymer is grown in situ from the
surface of the substrate. Riihe and coworkers reported the free radical polymerization of
styrene using SAMs of azo initiators covalently bound to high surface area silica gels54
and studied the kinetics and mechanism of this free radical polymerization.” Very
recently, they grew mixed polymer brushes consisting of PS and PMMA on a silicon
surface using this polymerization technique. These mixed polymer brushes show
nanophase separation into deﬁned patterns, depending on the molecular parameters of the
brushes.60‘ 6' This nanopattem has a domain memory effect, which is deﬁned as the
ability of single patterns of the brush topography to recover after erasure and regeneration
of the nanopattem by exposure to selective and nonselective solvents, respectively. To be

speciﬁc, the size, shape, and position of the PS-PMMA domains are found to vary with

the nature of the solvent to which the brushes are exposed. The brushes are treated ﬁrst

21

with toluene (good solvent for both blocks), which leads to strong swelling and structure
erasure. Subsequent exposure to acetone (a selective solvent for PMMA) leads to a ﬁlm
that contains PMMA chains at the surface and PS chains in its interior.

Bruening and Huang extended this method of polymerization to include siloxane-
based initiators attached to thin oxide surfaces formed on SAMs on Au.56 That study was
performed because Au surfaces are homogeneous and compatible with a broad variety of
surface analytical techniques including ellipsometry, contact angle measurements, AFM,
Fourier transform infrared spectroscopy (FTIR), surface plasmon resonance, and quartz
crystal microbalance gravimetry. In addition, contact-printing schemes allow patterning
of layers on Au surfaces.l3 Bruening and Huang also applied free-radical polymerization
to grow PS from azo—initiator thiol monolayers attached directly to a planar Au substrate.
However, the stability of alkanethiol monolayers appears to decrease in the presence of
free radicals in solution. Desorbed alkanethiols may serve as efﬁcient chain-transfer
reagents that hinder polymerization. To overcome this problem, they utilized a simple
cross-linking procedure (Scheme 1.4) to enhance the stability of SAMs and make thermal

radical polymerization from Au surfaces more effective.

22

anchoring group cleavable group

 

 

immobilization

_A_(D—-|—I*

 

 

 

 

polymerization
4.92%
cleavage

WW

Scheme 1.3. Schematic description of the preparation of polymer brushes
by radical polymerization from covalently immobilized initiators. (Redrawn

from Macromolecules 1998, 31, 592-601.)

23

/OCH3
s/\/‘SI—OCH3

Au OCH3
/OCH3
s/\/‘Si—OCH3
OCH3
H3O“
l
/o
S/\/\Si\-OH
Au /0
SMSi\-OH
)0
/ '\/\/ NN’ \‘QCN
CH30 0 CH3
l I
/o
SMSi-O-SK-
0
Au /\/\ O / ﬁ' H3C CN CH3
0 0 Ha
/
/
A
/(l) I
SMSi-O-SK-
0
Au /\/\ 0 / if H3C CN H
s SI—O'Sl\\/\/ CH2_C+
O n
/

Scheme 1.4. Free radical polymerization from initiator layers attached to
cross—linked thiol monolayers on Au. (Figure redrawn from Langmuir 2001,

17, 1731-1736.)

24

1.2.3. Surface Initiated Anionic Polymerization

Among the surface initiated polymerization techniques, living anionic
polymerizations are very attractive for possible control of polymer architecture.62
Anionic polymerization has been used to grow polymer brushes from a variety of
substrates such as silica,"3 clay nanoparticles64 and ﬂat surfaces.62‘ 65' 66

Jordan et a1. presented the ﬁrst report of surface initiated anionic polymerization,
with the growth of styrene from a SAM on gold.65 A SAM of 4’-bromo-4-
mercaptobiphenyl was self assembled on gold and reacted with s-BuLi to obtain 4’-lithio-
4-mercaptobiphenyl, which served as the initiator for the anionic polymerization of

styrene. This synthetic approach results in homogeneous “brush type” PS layers with a

very high grafting density as well as surprisingly high stability.

 

Q '56“) O n-BuLi/Benzene o | 69 Q
g 7 \(CH2)120 O V .2 7N\(CH2)120 0 6L9
Bu

 

 

Scheme 1.5. Schematic illustration of anionic SIP from clay surfaces.

(Redrawn from Langmuir 2002, I8, 451 1—4518.)

However, controlled anionic polymerization requires the proper choice of initiator,
monomer, and solvent as well as high purity reagents and an inert atmosphere.67
Recently, Advincula et al. reported the synthesis of polymer brushes on clay

nanoparticles using 1,1-diphenylethylene (DPE) initiators (Scheme 1.5).64‘68 Anionic SIP

25

directly from clay particles is difﬁcult, because the hydrophilic larnellar hosts inevitably
contain moisture, which can terminate a propagating living anion. To overcome this
problem, they demonstrated that by using a DPE derivative as the initiator and carefully
controlling the polymerization conditions (high temperature and high vacuum to remove
excess water), SIP from silicate clay surfaces is feasible. They inferred a “living” nature
for anionic polymerization of PS from the DPE initiators on nanoparticles by
demonstrating a linear relationship between monomer concentration and molecular mass
and observing the appearance/disappearance of a red-colored Li-DPE anion complex.
They then investigated the feasibility of anionic SIP of homopolymers and block
copolymers using SAMs of DPE derivatives on Si, SiOx and Au surfaces. The initiator
precursor DPE was functionalized with alkylsilane or alkylthiol groups and grafted onto
planar Si wafers and Au-coated surfaces, respectively. n-BuLi was used to activate the
DPE initiator, which allowed the anionic polymerization of PS or PS-b-polyisoprene to
proceed in benzene solution with both types of substrates.
1.2.4. Surface Initiated Cationic Polymerization

Compared to anionic polymerizations, cationic polymerizations are not as well
studied and have nOt been made suitable for complex macromolecular synthesis.69

Factors that inﬂuence surface initiated cationic polymerization include solvent polarity

and additives.70

26

 

 

 

 

 

 

 

 

 

OH OH OHOHOH
SOZCFa
O
+Au 0(802CF3)2
= t R:
EtOH, rt, 16h 1 day.
vapor phase
R 8
HS 3 S S S
O
CQHs C H
Csz 0 4K 2 5253'“)
A 0sz :1 n N 0 on
N 0 ON x
LJ ‘ OTt C2Hs>_N
0°C, reflux 7 days, 0 "-1
CHCI3,reﬂux
Fl
Fl
.1 ...I
§ §
3:1 if terminal
NFALKYL 8L (9] functionality
LALKYL on N”
1day, v C2H5 ’§
CHCI3, n 02"“ m'yme'
n-1
// // // W
R Au

Scheme 1.6. Surface initiated cationic polymerization of 2-oxazolines.

(Redrawn from J. Am. Chem. Soc. 1998, 120, 243-247.)

Jordan and Ulman presented the surface initiated cationic ring opening

polymerization of a 2-substituted 2-oxazoline to give poly(N-acylethylenimine).71 They
ﬁrst formed a SAM of triﬂate groups attached to n-alkyl chains as the initiator (Scheme
1.6). After 7 days of polymerization, the resulting polymer brush layer (10 nm thick) of

linear poly(N-propionylethylenimine) (PPEI) was of uniform thickness and was found to

27

be very stable. PPEI modiﬁed surfaces are of interest for their low toxicity and
denaturing potential toward numerous proteins as well as their resistance to protein
adsorption. Later, the Ulman group extended the polymerization of PPEI to the surface of
gold nanoparticles.72 3D (D-fUIICIIOI‘laIIZCd SAMs of thiolates on gold nanoparticles were
used to initiate the polymerization, and the resulting gold/polymer nanocomposite was
found to be thermally stable.

Because the surface area per polymer chain is generally greater than the surface
area per initiator bound to a surface, only a fraction of the immobilized initiators can
initiate polymerization. Zhao et al. described the cationic polymerization of tethered PS
brushes on silicate substrates and presented a discussion of initiator efﬁciency. They
used nondeuterated (2-(4-(1l-triethoxysilylundecyl) phenyl-2—methoxypropane) and
deuterated (2-(4-trichlorosilylphenyl)-2-methoxy-d3-propane) initiators to form the SAM
on the silicate substrate. FT IR-ATR was used to estimate the initiator efﬁciency, which
was about 7%.

1.2.5. Surface Initiated Ring-opening Metathesis Polymerization

Surface initiated ring opening polymerization is an attractive tool for coating
surfaces with thin layers of polycaprolactone, polylactide and other polymers. ROMP of
strained cyclic monomers, in particular functionalized norbomenes, has attracted recent
attention for the synthesis of polymers with useful electrical properties.73 Whitesides and
coworkers formed a variety of patterned polymer ﬁlms using the ROMP catalyst
developed by Schwab et a1. "-75 [(Cy3P)2C12Ru=CHPh, Cy=cyclohexyl]. The surface
bound catalytic sites were produced by forming a trichlorosilane-derived SAM containing

norbomene groups, and then exposing the SAM to a solution of the above catalyst.

28

Subsequent exposure of the substrate to norbomene-based monomers produced a
polymeric ﬁlm on the surface. Typical polymerization times were 30 min and ﬁlm
thicknesses were about 90 nm.76

Zauscher et al. combined AFM anodization lithography with ROMP to produce
nanopattemed polymer brushes.77 They showed for the ﬁrst time that AFM anodization
lithography, a form of ﬁeld-induced scanning probe lithography, offers a powerful means
to nanopattem semiconductor surfaces. First, anodic oxide patterns were generated on
octadecylmethyldiethoxysilane (ODMS)-SAM coated silicon substrates (Scheme 1.7).
Next, a 5-(bicycloheptenyl)triethoxysilane linker was covalently tethered to the patterned
area, allowing for the subsequent attachment of a Ru-based metathesis catalyst. Finally, a
nanopattem was ampliﬁed by immersing the patterned substrate in a monomer solution

containing 5—ethylidene-2-norbomene (ENE).

29

ODMS
SAM

 

 

 

(e)

 

/ Cl,,,.F|‘1
T IU‘CI
2
L,=PCy3 or IMesH2
L2=Pcw
Ph
/ H H
n
—Sl— \ Me
polyENB

Scheme 1.7. Stepwise fabrication of polymeric nanostructures by surface

initiated ROMP on SiOz nanopattems generated by anodization lithography.

a) ODMS resist SAMs were pre-coated on silicon substrates; b) anodization

lithography was performed by applying a positive bias voltage to the p-type

Si/SiOz substrate; c) an-Si(OEt)3 linkers were then covalently attached to

the patterned SiOz features; (1) A Ru-based metathesis catalyst (ﬁrst- or

second—generation Grubbs catalyst) was then covalently tethered to the

linkers; e) ROMP of ENB was performed in solution or vapor phase.

(Scheme and caption taken from Small 2006, 2, 848 — 853.)

30

1.2.6. Surface Initiated Controlled Radical Polymerization

Controlled radical polymerization has attracted much attention over the past
decade because it provides simple and robust synthetic routes to well-deﬁned, low-
polydispersity polymers. Among the controlled radical polymerization techniques, NMP,
ATRP, and RAFT polymerizations have already been applied to SIP by immobilizing
either a dormant species or a conventional radical initiator on the surface.78

The living character of NMP depends on the reversible capping of the active
chain-end radical with a nitroxide leaving group. The ﬁrst application. of NMP to surface
initiated graft polymerization was reported by Hussemann et al. in 1999.79 They
succeeded in densely grafting PS using a surface-bound dormant species with a 2, 2, 6, 6-
tetramethyl-l-piperidinyloxy (TEMPO) moiety (Scheme 1.8). A free dormant compound
with TEMPO was also added to the solution to help control the polymerization. The
stable TEMPO will cleave during the initiating process and subsequently reversibly cap
the chain-end radicals to control the concentration of radicals at a given time. PS brushes
over 100 nm thick were produced in 16 hours using this method. The living nature of the

grafting was suggested by the formation of a block copolymer brush.

31

 

 

 

 

I W O/N ﬁ—OH
Cl—Si O + —OH
I —OH
__r0H
R=Me, CI Silicon wafer
R
'-/\/\/\/\/\ O/N
—O—S|I O
R
_O_l_
“0+
R /N

 

 

_o:r\/\:C\§/l\ﬂog

 

 

i CFO/:3
hO—SIiWOO
—-o—l—R

 

 

01%;“?

Scheme 1.8. Synthesis of polystyrene brushes by TEMPO-mediated

radical polymerization. (Redrawn from Macromolecules 1999, 32,

1424-1431.)

32

RAFT polymerization is another important technique for controlled radical
polymerization. In RAFT, chain growth is initiated using a conventional free radical
initiator such as azobisisobutyronitn'le (AIBN) and mediated by a dithioester chain
transfer agent. Baum and Brittain prepared styrene, methyl methacrylate, and N,N-
dimethylacrylamide brushes under RAFT conditions using silicate surfaces that were
modiﬁed with surface-immobilized azo initiators (Scheme 1.9).80 2-Phenylprop-2-y1

dithiobenzoate was added as a free (unbound) RAFT agent to control the graft

R O
I CN
R' 8 CN

 

N

Chain Transfer Agent
(CTA)

AIBN,
solvent

0 CN R S .. “
H'KZHQH—OWS + free polymer
RI n

Scheme 1.9. General process of surface-initiated RAFT polymerizations using 2-
phenylprop-2-y1 dithiobenzoate as the chain transfer agent and a substrate-

immobilized azo initiator. (Redrawn from Macromolecules 2002, 35, 610-615.)
polymerization. At temperatures up to 90 °C and times of up to 48 hours, PMMA brushes

with a thickness of 28 nm and poly(N,N-dimethylacrylamide) and PS brushes up to 11

nm thick were grown. Although RAFT is relatively slow compared to ATRP and NMP,

33

the fact that this technique supported the easy re-initiation.of the polymer chains suggests
it is a highly controlled process.
1.2.7. Surface Initiated Atom Transfer Radical Polymerization
Since Matyjaszewski and Sawamoto ﬁrst reported the use of ATRP in 1995,81 ‘ 82
this technique has become one of the most widely employed techniques for the formation
of polymer brushes. Among all the controlled radical polymerization techniques, it is
particularly successful and has attracted commercial interest because it utilizes a simple
experimental setup and employs readily accessible and inexpensive catalysts (usually
copper complexes formed with aliphatic amines, imines, or pyridines, many of which are
commercially available), as well as simple initiators (often alkyl halides).83 ATRP is
probably the most robust and efﬁcient controlled radical polymerization technique for
preparing well-deﬁned polymers with controlled topology, composition and functionality.
This technique is compatible with a variety of functionalized monomers and the living/
controlled character of the ATRP process yields polymers with a low polydispersity
(MW/Mn)“

The controlled nature of ATRP is due to the reversible activation—deactivation

reaction between the growing polymer chains and a copper—ligand species. Scheme 1.10

shows the mechanism of ATRP.

34

kact
RX + Cu (I) X/bpy.:_ F10 + Cu (ll) X2/bpy
k kt

deact
+M k Termination

Scheme 1.10. Mechanism of ATRP. (Adapted from J. Am. Chem. Soc. 1998,

120, 243-247.)

Typically, ATRP utilizes a transition metal complex as a catalyst (e.g.,
Cu(I)X/bpy) and an initiator with a transferable halogen. The transition metal complex
acts as a halogen atom transfer agent between active and dormant chain ends. Fast
deactivation by the Cu(IDXZ/ligand complex, where X is the transferred halogen atom,
leads to a low concentration of propagating radicals, thus ensuring that chain termination
and radical transfer reactions are minimized so that polymer chains are able to grow at
slow, but nearly constant rates to yield a narrow molecular weight distribution.

Monomers which have been successfully polymerized by ATRP include
styrenes,85 acrylates,86 and acrylonitrile.87 All of these compounds contain substituents
capable of stabilizing propagating radicals (e. g., phenyl or carbonyl groups).88 In contrast,
ATRP is not compatible with monomers containing carboxylic acid groups (because the
carboxylic acid side groups react with the metal complexes to form carboxylate salts
which are inefﬁcient catalysts for ATRP) and halogenated or alkylsubstituted alkenes
(due to their low intrinsic reactivity toward radical polymerization).

A variety of transition metal catalysts have been successfully used in ATRP,

including Cu,89’ 90 Fe,” Ru,92 Ni,93 Mo, 94and Rh95 complexes. Among these transition

35

metal catalyst systems, the Cu halide system is the most popular due to its compatibility

with various monomers. To increase the solubility of Cu salts in organic media and to
i) bidentate ligands
R R
_ _ _ _ _ 6—
W M Ch ,1 y, /
ii) tridentate ligands

«I» —~
\ K/

/

iii) tetradentate ligands
\N Ni> \N N/ (\NT
N [N N] ~—
'— / \N ——
\ l\/l \ / N\

 

(

—N
/

Scheme 1.11. Ligands used for copper-mediated ATRP. (Redrawn from

Curr. Org. Chem, 2002, 6, 67-82.)

adjust the redox potential of the metal center for appropriate reactivity, various

polydentate nitrogen based ligands, such as phenanthroline and its derivatives, and

pyridineimines, have been used for copper-mediated ATRP (Scheme 1.11).96

Ejaz et a1. ﬁrst succeeded in using ATRP to synthesize dense polymer brushes

7 They deposited the initiator by the

(low-polydispersity PMMA) on silicon wafers.9

36

SO2 Initiating group for ATRP
__ Spread Hydrolysis
CH2 —— ————> ————>
I / Si on the water surface / Si
”3320/
' 3
H3CO/PI\OCH
H3CO Monolayer formation

 
 
  

Compression
—

     

Condensation Water

 

LB Trough
Deposition . .‘

     

Graft Polymn

‘

of MMA

Fixation

 

 

 

 

 

Scheme 1.12. Schematic illustration of the immobilization of ATRP
initiators using the LB technique. (Redrawn from Macromolecules 1998, 31,

5934-5936.)

Langmuir—Blodgett technique and subsequently exposed this substrate to monomer and
the ATRP catalyst (Scheme 1.12). PMMA ﬁlm thicknesses of up to 80 nm were
achieved in less than 12 hours at 90 °C, as measured by ellipsometry.

Au-coated substrates provide a model surface that is readily characterized and
complimentary to Si, but due to the limited thermal stability of the S — Au bond in
initiator monolayers, ATRP can only be carried out at a relatively low temperature. Kim

and coworkers reported surface-conﬁned ATRP of MMA on gold surfaces using room-

37

temperature ATRP.98 The use of very active Cu-MebTREN complexes as catalysts
allowed for successful room-temperature polymerization of dense PMMA brushes on
gold. Based on solution studies by Arrnes and coworkers,99 Huang and coworkers found
that room-temperature, water-accelerated ATRP allows for growth of thick polymer ﬁlms
in short reaction times. They obtained 700 nm thick PHEMA ﬁlms in just 12 hours using
water as the solvent.3 Huck and coworkers observed a similar effect in the aqueous-
based polymerization of poly(glycidyl methacrylate).100’ 10' They were able to grow 125
nm-thick poly(glycidyl methacrylate) brushes in 100 min.

Surface initiated ATRP was applied not only on planar substrates but also on
various types of ﬁne particles. Patten extended the technique to silica nanoparticlesloz‘ .
'03 In his procedure, (2-(4-chloromethylphenyl)ethyl) dimethylethoxysilane initiator was
deposited on the surface of the nanoparticles, and well-deﬁned PS chains were grown
from the modiﬁed nanoparticle surfaces. Besides silica nanoparticles, ATRP of polymers
on gold,'04’ '05 iron oxide106 and magnetic particles has been demonstrated.107

Porous materials with polymer brushes grafted on them are also fascinating
targets for chromatographic and other applications such as the membrane absorbers
described herein. Wirth et al. reported the grafting of poly(acrylamide) (PAAm) on a
porous silica gel and showed the protein separation ability of these materials.” A living
PAAm thin ﬁlm is grown from a self assembled benzyl chloride monolayer on silica gel
using Cu(bpy)2Cl to control the radical polymerization. To test for intact pore structure of
the PAAm-modiﬁed silica gel, a size-exclusion separation was performed for sodium
azide and four proteins: thyroglobulin (MW, 669,000), ovalburnin (MW, 44,000),

ribonuclease A (MW, 13,700), and aprotinin (MW, 6,500). The ﬁve species eluted in

38

order of decreasing molecular weight, as expected for a size exclusion separation. This
proved that the polymerization is sufﬁciently well controlled that ﬁlms grown on
nanoporous silica leave the pores intact. Later on, they applied ATRP of acrylamide to
coat the channels of PDMS microchips and studied the electrophoretic behavior of
proteins in these channels.'09 Bare PDMS is adsorptive toward proteins due to its
hydrophobicity, and it would be valuable to be able to modify the surface of PDMS to
make it permanently hydrophilic and resistant to protein adsorption. By growing PAAm

in the PDMS channels, adsorption of bovine serum albumin was largely reduced.

1.3. Introduction to Pervaporation Membranes

Membrane separation techniques such as microﬁltration, ultraﬁltration, reverse
osmosis, dialysis, electrodialysis, gas separation, and pervaporation have been applied to
industrial, medical, and biological ﬁelds.”°‘ I" Pervaporation is an especially promising
membrane—based technique in which a liquid mixture contacts the surface of a membrane,
and selective transport of one component through the membrane to a vapor phase (often a
modest vacuum) affords separation (Figure 1.7).”2 It is attractive because it allows for
the separation of azeotropic mixtures and often requires less energy than conventional

distillation.l '3

Many of the successes with pervaporation involve the removal of water
from organic solvents, but the reverse separation has also been demonstrated.
Hydrophilic membranes allow selective pervaporation of water for dehydration of
organic solvents, while hydrophobic ﬁlms often permit selective passage of volatile

organic compounds (VOCS) from aqueous solutions for water remediation or VOC

analysis by membrane- introduction mass spectrometry.

39

 

 

 

 

vacuun out

 

Figure 1.7. Schematic diagrams of (a) a pervaporation system and (b) a cross-
sectional view of the membrane cell of the apparatus. (Figure taken from J. Membr.

Sci. 2005, 248, 161 — l 70.)

In the pervaporation process, the separation effectiveness of a membrane is
quantiﬁed by two parameters, ﬂux and selectivity.”2 The membrane selectivity comes
from the different solubilities and diffusivities of the components in the membrane. A

solution-diffusion model is generally used to describe the transport of molecules in

Sample stream CAm (vacuum)

 

 

 

Analyte Concentration

 

Figure 1.8. Schematic diagram of a solution-diffusion model.

40

membranes. In this model (Figure 1.8), the permeation process can be divided into three
steps: (1) sorption into the membrane at the feed side (often assumed to be at
equilibrium), (2) diffusion through the membrane, and (3) desorption at the permeate side.
Solubility is a thermodynamic property, and diffusivity is a kinetic property, but they
both affect the selectivity.‘ '4

Selective removal of organics from water is very important in chemical analysis
and puriﬁcation of waste streams.”5 The solution-diffusion mechanism suggests that
pervaporation of organic compounds from water will require a material with high
sorption selectivity for the organic molecule and a minimal diffusion selectivity for
water.”6'”8 Hydrophobic polymers with high free volumes, especially PDMS, provide
these characteristics, and there are a number of reports of selective pervaporation of
volatile organic compounds through such materials.”9"2'

In addition to selectivity, both analytical and preparative pervaporation
applications require high ﬂuxes that necessitate the use of membranes that are as thin as
possiblem’ 123 However, due to the mechanical weakness of ultrathin polymer ﬁlms,
high-ﬂux membranes generally consist of a thin, selective skin layer on a nonselective,

highly porous support.124

Although PDMS membranes provide high selectivity towards
removal of organics from water, it’s difﬁcult to achieve defect-free skin layers that are
less than 500 nm thick using simple techniques for forming PDMS membranes.125

To overcome this problem, polymer brushes can be used as selective layers on
porous supports. As mentioned previously in this chapter, there are a lot of controlled

polymerization techniques available to grow polymer brushes on a porous support. ATRP

is especially useful for preparing well-deﬁned polymers with controlled topology,

41

composition and functionality, and polymer brushes prepared by this technique will be

well suited to cover the pores of porous membrane but not block their interior pores.

1.4. Importance of Membrane Absorbers

Potential therapeutic applications of proteins along with rapid developments in
biotechnology and genetic engineering have 'greatly increased the rate of production of
various proteins and enzymes.'26‘ 127 Sales of therapeutic proteins are projected to reach a
remarkable value of $90 Billion in 2009.128 Accompanying this increase in the scale of
protein production is a need for fast and convenient puriﬁcation techniques. Afﬁnity
chromatography is probably the most convenient way to purify proteins due to its ability
to separate biomolecules based on their biological interactions)” '29 but the rate of
separations with packed—bead afﬁnity columns is limited by slow diffusion of
biomacromolecules within bead poresm’13 1

Compared to column chromatography, membrane chromatography has a lower
pressure drop, higher flow rate, and higher productivity as a result of the
microporous/macroporous structure of the thin membrane.132 Flow through membrane
pores greatly enhances mass transport to binding sites. Easy packing and scale-up of
membranes is an additional asset of membrane chromatography. Consequently, this
technique is a promising large-scale separation process for the isolation, puriﬁcation, and
recovery of proteins and enzymes. Many publications have reported the performance of
adsorptive membranes (ion-exchange membranes, afﬁnity membranes, and hydrophobic

interaction membranes), as well as their theoretical description and optimal designm' 129'

133

42

Despite their potential, the major disadvantage of membrane absorbers is their
low binding capacity relative to beads. The speciﬁc surface area of membranes is simply
not as great as that of beads. This can be overcome to some extent through the use of
polymer chains grafted in membrane pores and binding of multilayers of proteins to these
polymers. For example, Ulbricht and Yang achieved a lysozyme-binding capacity of 20
mg/cm3 by grafting poly(acrylic acid) to polypropylene membranes modiﬁed with
adsorbed photoinitiators.4

Our research aims at utilizing the controlled growth of polymer brushes in porous
membranes to develop afﬁnity membranes with remarkably high protein-binding
capacities. The use of ATRP to grow PHEMA from initiators bound to a porous alumina
surface affords relatively ﬁne control over polymer molecular weight, which allows large
increases in capacity without clogging of membrane pores. PHEMA brushes are
particularly attractive because they can be functionalized to exploit a number of afﬁnity
interactionsm’ ‘34' '35 Speciﬁcally, functionalization of PHEMA with nitrilotriacetate-
Cu2+ complexes results in protein binding via metal-ion afﬁnity interactions, and the
microporous alumina support provides a ~500-fold increase in surface area relative to

two-dimensional supports.

1.5. Outline of the Dissertation

In Chapter 2 of this dissertation, I report preparation of composite pervaporation
membranes via ATRP of 2-hydroxyethyl methacrylate from the surface of porous
alumina membranes and subsequent derivatization of PHEMA using a series of acid

chlorides. The pervaporation performance of a series of derivatized PHEMA membranes

43

is compared to provide correlations between membrane chemistry and transport
properties. Sorption studies, which helped us elucidate the factors behind the selectivities
of different ﬁlms, will also be presented in this chapter. Last but not the least, I compare
the performance of our membranes with related membrane systems that have been
described in the literature.

Chapter 3 of this dissertation ﬁrst describes the growth, characterization, and
derivatization of polymer brushes to yield protein-binding materials. To quantify the
binding capacity of these polymer brushes on gold substrates, we developed a simple
technique which utilizes both FTIR and ellipsometry for calibration of IR absorbances.
Subsequently, chapter 3 presents synthesis and derivatization of polymer brushes inside
the pores of porous alumina membranes. Simple surface area calculations clarify why we
need to grow polymer brushes inside of highly porous membranes. Finally, protein
binding studies in membranes and a comparison with related research illustrate the
extremely high protein-binding capacity of our polymer brushes.

In chapter 4, I show that derivatized polymer brushes are very attractive for
puriﬁcation of histindinex—tagged proteins. Brushes formed on a gold substrate were
capable of selectively binding the equivalent of many monolayers of his tagged ubiquitin.
Porous alumina membranes modiﬁed with derivatized brushes also show remarkable
capacities for the binding of his tagged ubiquitin. Gel electrophoresis studies show that
these high-capacity membranes are highly selective for puriﬁcation of his tagged proteins.

In the last chapter, I will present the conclusions drawn from my research and

some proposed future work.

1.6.

10.

11.

l2.

l3.

14.

15.

l6.

17.

References
Milner, S. T., Science 1991, 251, 905-914.

Bhat, R. R.; Tomlinson, M. R.; Wu, T.; Genzer, J ., Adv. Polym. Sci. 2006, 198,
51-124.

Huang, W.; Kim, J .-B.; Bruening, M. L.; Baker, G. L., Macromolecules 2002, 35,
1175-1179.

Ulbricht, M.; Yang, H., Chem. Mater. 2005, I 7, 2622-2631.

Brown, H. R.; Char, K.; Deline, V. R., Macromolecules 1990, 23, 3383-3385.
Ash, 8. G.; Clayﬁeld, E. J ., J. Colloid Interface Sci. 1976, 55, 645-657.

Xu, F. J .; Kang, E. T.; Neoh, K. G., J. Mater. Chem. 2006, 16, 2948-2952.

Andruzzi, L.; Senaratne, W.; Hexemer, A.; Sheets, E. D.; llic, B.; Kramer, E. J .;
Baird, B.; Ober, C. K., Langmuir 2005, 21, 2495-2504.

Liu, Y.; Klep, V.; Luzinov, 1., J. Am. Chem. Soc. 2006, 128, 8106-8107.
Xia, Y.; Whitesides, G. M., Angew. Chem. Int. Ed. 1998, 37, 550-575.

Shah, R. R.; Merreceyes, D.; Husemann, M.; Rees, 1.; Abbott, N. L.; Hawker, C.
J .; Hedrick, J. L., Macromolecules 2000, 33, 597-605.

Schmelmer, U.; Jordan, R.; Geyer, W.; Eck, W.; Golzhauser, A.; Grunze, M.;
Ulman, A., Angew. Chem. Int. Ed. 2003, 42, 559-563.

Husemann, M.; Mecerreyes, D.; Hawker, C. J .; Hedrick, J. L.; Shah, R.; Abbott,
N. L., Angew. Chem. Int. Ed. 1999, 38, 647-649.

Jones, D. M.; Smith, J. R.; Huck, W. T. 8.; Alexander, C., Adv. Mater. 2002, 14,
1130-1134.

Ahn, S. J .; Kaholek, M.; Lee, W.-K.; LaMattina, B.; LaBean, T. H.; Zauscher, 8.,
Adv. Mater. 2004, 16, 2141-2145.

Padeste, C.; Solak, H. H.; Brack, H.-P.; Slaski, M.; Gursel, S. A.; Scherer, G. G.,
J. Vac. Sci. Technol, B 2004, 22, 3191-3195.

Kaholek, M.; Lee, W.-K.; LaMattina, B.; Caster, K. C.; Zauscher, S., Nano Lett.
2004, 4, 373-376.

45

l8.

19.

20.

21.

22.

23.
24.
25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

. 35.

Krauss, P. R.; Chou, S. Y., J. Vac. Sci. Technol, B 1995, 13, 2850-2852.

Beinhoff, M.; Appapillai, A. T.; Underwood, L. D.; Frommer, J. E.; Carter, K. R.,
Langmuir 2006, 22, 2411-2414.

Chou, S. Y.; Krauss, P. R.; Renstrom, P. J ., J. Vac. Sci. Technol, B 1996, 14,
4129-4133.

von Weme, T. A.; Germack, D. S.; Hagberg, E. C.; Sheares, V. V.; Hawker, C. J .;
Carter, K. R., J. Am. Chem. Soc. 2003, 125, 3831-3838.

Retsos, H.; Gorodyska, G.; Kiriy, A.; Stamm, M.; Creton, C., Langmuir 2005, 21 ,
7722-7725.

Pierschbacher, M. D.; Ruoslahti, E., Nature 1984, 309, 30-33.
Hersel, U.; Dahmen, C.; Kessler, H., Biomaterials 2003, 24, 4385-4415.
Houseman, B. T.; Mrksich, M., Biomaterials 2001, 22, 943-955.

Maheshwari, G.; Brown, G.; Lauffenburger, D. A.; Wells, A.; Grifﬁth, L. G., J.
Cell Sci. 2000, 113, 1677-1686.

Harris, B. P.; Kutty, J. K.; Fritz, E. W.; Webb, C. K.; Burg, K. J. L.; Metters, A.
T., Langmuir 2006, 22, 4467-4471.

Israelachvili, J. N.; Tabor, D., Nature 1973, 241, 148-149.

Klein, J .; Kumacheva, E.; Mahalu, D.; Perahla, D.; Fetters, L. J ., Nature 1994,
370, 634—636.

Raviv, U.; Giasson, S.; Kampf, N.; Gohy, J .-F.; Jerome, R.; Klein, J ., Nature 2003,
425, 163-165.

Singh, G; Pickett, G. T.; Zhulina, E.; Balazs, A. C., J. Phys. Chem. B 1997, 101,
10614-10624.

Roan, J .-R.; Kawakatsu, T., J. Chem. Phys. 2002, 116, 7295-7310.

Kim, A. J.; Manoharan, V. N.; Crocker, J. C., J. Am. Chem. Soc. 2005, 127, 1592-
1593.

Zaman, A. A.; Bjelopavlic, M.; Moudgil, B. M., J. Colloid Inteiface Sci. 2000,
226, 290-298.

Walker, H. W.; Grant, 8. B., Langmuir 1996, 12, 3151-3156.

46

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

Yan, Y. D.; Glover, S. M.; Jameson, G. J.; Biggs, 8., Int. J. Miner. Process. 2004,
73, 161-175.

Leermakers, F. A. M.; Zhulina, E. B.; van Male, J .; Mercurieva, A. A.; Fleer, G. J .;
Birshtein, T. M., Langmuir 2001, 17, 4459—4466.

Gerber, P. R.; Moore, M. A., Macromolecules 1977, 10, 476-481.

Satulovsky, J.; Carignano, M. A.; Szleifer, 1., Proc. Natl. Acad. Sci. U. S. A. 2000,
97, 9037-9041.

Szleifer, 1., Curr. Opin. Solid State Mater. Sci. 1997, 2, 337-344.
Halperin, A., Langmuir 1999, 15, 2525-2533.
Ma, H.; Hyun, J.; Stiller, P.; Chilkoti, A., Adv. Mater. 2004, 16, 338-341.

Chen, Y.; Liu, D.; Deng, 0.; He, X.; Wang, X., J. Polym. Sci, Part A: Polym.
Chem. 2006, 44, 3434-3443.

Drobek, T.; Spencer, N. D.; Heuberger, M., Macromolecules 2005, 38, 5254-5259.

Brown, A. A.; Khan, N. S.; Steinbock, L.; Huck, W. T. 8., Eur. Polym. J. 2005,
41, 1757-1765.

Chilkoti, A.; Ma, H. USA 2005-US4947, 20050217, 2005.
Ito, Y.; Park, Y. S., Polym Adv. Technol. 2000, 11, 136-144.
Houbenov, N.; Minko, S.; Stamm, M., Macromolecules 2003, 36, 5897-5901.

Takei, Y. G.; Aoki, T.; Sanui, K.; Ogata, N.; Sakurai, Y.; Okano, T.,
Macromolecules 1994, 27, 6163-6166.

Kost, J .; Langer, R., Adv. Drug Delivery Rev. 2001, 46, 125-148.
Feil, H.; Bae, Y. H.; Feijen, J .; Kim, S. W., J. Membr. Sci. 1991, 64, 283-294.
Park, Y. S.; Ito, Y.; Imanishi, Y., Macromolecules 1998, 31, 2606-2610.

Luzinov, 1.; Julthongpiput, D.; Malz, H.; Pionteck, J .; Tsukruk, V. V.,
Macromolecules 2000, 33, 1043-1048.

Matyjaszewski, K.; Miller, P. J .; Shukla, N.; Irnmarapom, B.; Gelman, A.;

Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T., Macromolecules 1999, 32, 8716-8724.

47

55.

56.

57.

58.

59.

60.

61.

62.

63.

65.

66.

67.

68.

69.

70.

71.

72.

Prucker, 0.; Riihe, J ., Langmuir 1998, 14, 6893-6898.

Huang, W.; Skanth, G.; Baker, G. L.; Bruening, M. L., Langmuir 2001, 17, 1731-
1736.

Sugawara, T.; Matsuda, T., Macromolecules 1994, 2 7, 7809-7814.
Prucker, 0.; Riihe, J ., Macromolecules 1998, 31, 592-601.
Prucker, 0.; Riihe, J ., Macromolecules 1998, 31 , 602-613.

Santer, S.; Kopyshev, A.; Donges, J .; Yang, H.-K.; Riihe, J ., Langmuir 2006, 22,
4660—4667.

Santer, S.; Kopyshev, A.; Yang, H.-K.; Riihe, J ., Macromolecules 2006, 39, 3056-
3064.

Advincula, R.; Zhou, Q.; Park, M.; Wang, S.; Mays, J .; Sakellariou, G.; Pispas, S.;
Hadjichristidis, N., Langmuir 2002, 18, 8672-8684.

Zhou, Q.; Nakamura, Y.; Inaoka, S.; Park, M.-K.; Wang, Y.; Mays, J .; Advincula,
R., Polym. Mater. Sci. Eng. 2000, 82, 290-291.

Fan, X.; Zhou, Q.; Xia, C.; Cristofoli, W.; Mays, J .; Advincula, R., Langmuir
2002, 18, 4511-4518.

Jordan, R.; Ulman, A.; Kang, J. F.; Rafailovich, M. H.; Sokolov, J ., J. Am. Chem.
Soc. 1999, 121, 1016-1022.

Ingall, M. D. K.; Honeyman, C. H.; Mercure, J. V.; Bianconi, P. A.; Kunz, R. R.,
J. Am. Chem. Soc. 1999, 121, 3607-3613.

Quirk, R. P.; Mathers, R. T.; Cregger, T.; Foster, M. D., Macromolecules 2002,
35, 9964-9974.

Zhou, Q.; Fan, X.; Xia, C.; Mays, J.; Advincula, R., Chem. Mater. 2001, 13,
2465-2467.

Advincula, R., Adv. Polym. Sci. 2006, 197, 107-136.
Zhao, B.; Brittain, W. J ., Macromolecules 2000, 33, 342-348.
Jordan, R.; Ulman, A., J. Am. Chem. Soc. 1998, 120, 243-247.

Jordan, R.; West, N.; Ulman, A.; Chou, Y.-M.; Nuyken, 0., Macromolecules
2001, 34, 1606-1611.

48

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

Edmondson, S.; Osborne, V. L.; Huck, W. T. S., Chem. Soc. Rev. 2004, 33, 14-22.
Jeon, N. L.; Choi, I. S.; Whitesides, G. M.; Kim, N. Y.; Laibinis, P. E.; Harada, Y.;
Finnie, K. R.; Girolami, G. S.; Nuzzo, R. G., Appl. Phys. Lett. 1999, 75, 4201-
4203.
Schwab, P.; Grubbs, R. H.; Ziller, J. W., J. Am. Chem. Soc. 1996, 118, 100-110.
Kim, N. Y.; Jeon, N. L.; Choi, I. S.; Takami, S.; Harada, Y.; Finnie, K. R.;
Girolami, G. S.; Nuzzo, R. G.; Whitesides, G. M.; Laibinis, P. E.,
Macromolecules 2000, 33 , 2793-2795.
Lee, W.-K.; Caster, K. C.; Kim, J .; Zauscher, S., Small 2006, 2, 848-853.

Tsujii, Y.; Ohno, K.; Yamamoto, S.; Goto, A.; Fukuda, T., Adv. Polym. Sci. 2006,
197, 1-45.

Husseman, M.; Malmstroem, E. E.; McNamara, M.; Mate, M.; Mecerreyes, D.;
Benoit, D. G.; Hedrick, J. L.; Mansky, P.; Huang, E.; Russell, T. P.; Hawker, C. J .,
Macromolecules 1999, 32, 1424-1431.

Baum, M.; Brittain, W. J ., Macromolecules 2002, 35, 610-615.

Wang, J.-S.; Matyjaszewski, K., J. Am. Chem. Soc. 1995, 117, 5614-5615.

Kato, M.; Kamigaito, M.; Sawamoto, M.; Higashimura, T., Macromolecules 1995,
28, 1721-1723.

Matyjaszewski, K.; Spanswick, J., Mater. Today 2005, 8, 26-33.
Matyjaszewski, K.; Xia, J., Chem. Rev. 2001, 101 , 2921-2990.

Teare, D. 0. H.; Barwick, D. C.; Schoﬁeld, W. C. E.; Garrod, R. P.; Ward, L. J .;
Badyal, J. P. S., Langmuir 2005, 21, 11425-11430.

Brar, A. S.; Saini, T., J. Polym. Sci., Part A: Polym. Chem. 2006, 44, 1975-1984.
Hou, C.; Ying, L.; Wang, C., J. Appl. Polym. Sci. 2006, 99, 1050-1054.
Matyjaszewski, K., Curr. Org. Chem. 2002, 6, 67-82.

Wang, J .-S.; Matyjaszewski, K., Macromolecules 1995, 28, 7901-7910.

Wang, J .-S.; Matyjaszewski, K., Macromolecules 1995, 28, 7572-7573.

49

91. Gibson, V. C.; O'Reilly, R. K.; Wass, D. F.; White, A. J. P.; Williams, D. J.,
Macromolecules 2003, 36, 2591-2593.

92. Saenz-Galindo, A.; Textle, H. M.; Jasso, A. R.; Torres-Lubian, J. R., J. Polym.
Sci., Part A: Polym. Chem. 2005, 44, 676-680.

93. Duquesne, E.; Habimana, J .; Degee, P.; Dubois, P., Macromolecules 2005, 38,
9999-10006.

94. Maria, S.; Stoffelbach, F.; Mata, J .; Daran, J .-C.; Richard, P.; Poli, R., J. Am.
Chem. Soc. 2005, 127, 5946-5956.

95. Mecerreyes, D.; Moineau, G.; Dubois, P.; Jerome, R.; Hedrick, J. L.; Hawker, C.
J .; Malmstrom, E. E.; Trollsas, M., Angew. Chem. Int. Ed. 1998, 37, 1274-1276.

96. Tang, W.; Matyjaszewski, K., Macromolecules 2006, 39, 4953-4959.

97. Ejaz, M.; Yamamoto, S.; Ohno, K.; Tsujii, Y.; Fukuda, T., Macromolecules 1998,
31 , 5934-5936.

98. Kim, J.-B.; Bruening, M. L.; Baker, G. L., J. Am. Chem. Soc. 2000, 122, 7616-
7617.

99. Wang, X. S.; Jackson, R. A.; Arrnes, S. P., Macromolecules 2000, 33, 255-257.
100. Jones, D. M.; Brown, A. A.; Huck, W. T. S., Langmuir 2002, 18, 1265-1269.
101. Jones, D. M.; Huck, W. T. 5., Adv. Mater. 2001, 13, 1256-1259.

102. von Weme, T.; Patten, T. E., J. Am. Chem. Soc. 1999, 121, 7409-7410.

103. von Weme, T.; Patten, T. E., J. Am. Chem. Soc. 2001, 123, 7497-7505.

104. Nuss, S.; Bottcher, H.; Wurm, H.; Hallensleben, M. L., Angew. Chem. Int. Ed.
2001, 40, 4016-4018.

105. Ohno, K.; Koh, K.-m.; Tsujii, Y.; Fukuda, T., Macromolecules 2002, 35, 8989-
8993.

106. Li, G.; Fan, J.; Jiang, R.; Gao, Y., Chem. Mater. 2004, 16, 1835-1837.

107. Marutani, E.; Yamamoto, S.; Ninjbadgar, T.; Tsujii, Y.; Fukuda, T.; Takano, M.,
Polymer 2004, 45, 2231-2235.

108. Huang, X.; Wirth, M. J ., Anal. Chem. 1997, 69, 4577-4580.

50

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.

Xiao, D.; Van Le, T.; Wirth, M. J ., Anal. Chem. 2004, 76, 2055-2061.

0hshima, T.; Miyata, T.; Uragami, T.; Berghmens, H., J. Mol. Struct. 2005, 739,
47-55.

Vane, L. M., J. Chem. Technol. Biotechnol. 2006, 81, 1328.

Sun, L.; Baker, G. L.; Bruening, M. L., Macromolecules 2005, 38, 2307-2314.
Sullivan, D. M.; Bruening, M. L., J. Membr. Sci. 2005, 248, 161-170.

Doong, S. J .; Ho, W. S.; Mastondrea, R. P., J. Membr. Sci. 1995, 107, 129-146.
Kujawski, W., Sep. Sci. Technol. 2000, 35, 89-108.

Mishima, S.; Kaneoka, H.; Nakagawa, T., J. Appl. Polym. Sci. 1999, 71, 273-287.

Chandak, M. V.; Lin, Y. S.; Ji, W.; Higgins, R. J ., J. Appl. Polym. Sci. 1998, 67,
165-175.

Samdani, A. R.; Manda], S.; Pangarkar, V. G., Sep. Sci. Technol. 2003, 38, 1069-
1092.

Kujawski, W.; Ostrowska-Gumkowska, B., Sep. Sci. Technol. 2003, 38, 3669-
3687.

Uragami, T.; 0hshima, T.; Miyata, T., Macromolecules 2003, 36, 9430-9436.

Trong Nguyen, Q.; Bendjama, Z.; Clement, R.; Ping, Z., Phys. Chem. Chem. Phys.

2000, 2, 395-400.
Olsson, J .; Tragardh, G.; Lipnizki, F., Sep. Sci. Technol. 2002, 3 7, 1199-1223.

Nijhuis, H. H.; Mulder, M. H. V.; Smolders, C. A., J. Membr. Sci. 1991, 61, 99-
111.

Yoshida, W.; Cohen, Y., J. Membr. Sci. 2004, 229, 27-32.

Blume, 1.; Wijmans, J. G.; Baker, R. W., J. Membr. Sci. 1990, 49, 253-286.
Arica, M. Y.; Testereci, H. N.; Denizli, A., J. Chromatogr. A 1998, 799, 83-91.
Kawai, T.; Saito, K.; Lee, W., J. Chromatogr. B 2003, 790, 131-142.

Therapeutic Proteins, h. w. m. c. p. d. a. p. x. r., accessed March 2006.

51

129.

130.

131.

132.

133.

134.

135.

Zou, H.; Luo, Q.; Zhou, D., J. Biochem. Bioph. Methods 2001, 49, 199-240.

Che, S.; Liu, Z.; 0hsuna, T.; Sakamoto, K.; Terasaki, 0.; Tatsumi, T., Nature
2004, 429, 281-284.

Clairbois, A. S.; Letoumeur, D.; Muller, D.; Jozefonvicz, J ., J. Chromatogr. B
1998, 706, 55-62.

Zeng, X.; Ruckenstein, E., Biotechnol. Prog. 1999, 15, 1003-1019.

Reichert, U.; Linden, T.; Belfort, G.; Kula, M.-R.; Thommes, J., J. Membr. Sci.
2002, 199,161-166.

Miller, M. D.; Baker, G. L.; Bruening, M. L., J. Chromatogr. A 2004, 1044, 323-
330.

Bayramoglu, G.; Yilmaz, M.; Arica, M. Y., Colloids Surf, A 2004, 243, 11-21.

52

Chapter 2

Polymer Brush Membranes for Pervaporation

2.1. Introduction

This chapter describes our efforts to utilize polymer brushes as the skin layer of
pervaporation membranes, where polymer brushes are grown only from the surface of the
substrate so that they cover underlying pores without ﬁlling them.

Our hypothesis in this area is that polymer brushes grown on top of porous
supports may provide a platform for preparing ultrathin skin layers and enhancing flux.
In support of this hypothesis, this chapter reports preparation of composite pervaporation
membranes via ATRP“1 of HEMAS'7 from the surface of a porous substrate (Scheme 2.1).
Although this work employs porous alumina as a support, formation of brush membrane
skins should be possible from a wide range of porous materials such as cellulose,8
PDMS,9 PS'0 and PEs adsorbed on various surfaces.ll To make PHEMA membranes
sufﬁciently hydrophobic for pervaporation of organic analytes from water, we derivatize
them with hydrophobic acid chlorides, i.e., octanoyl chloride, palmitoyl chloride, and
pentadecafluorooctanoyl chloride. Jennings previously showed that derivatization of
PHEMA can yield hydrophobic ﬁlms,” '3 and we reported that ﬂuorinated PHEMA

'1 However, derivatized

membranes show modest selectivities in gas separations.
PHEMA ﬁlms are much more attractive for pervaporation than gas separations, as they
allow unusually high pervaporation ﬂuxes (>1 kg/mzh). The ability to derivatize

PHEMA with a variety of hydrophobic molecules permits comparison of different

53

 

 

 

 

 

f—OH cu—salcrazm
U -—OH , J$<Br
"'4 L—-OH 18h BrI/O: ’Si-{CH2}(‘)10
N _.
A. OH Initiator Film
HEMA
CuCl/CuBrg \ o
”W 08ch 1L0 CH3”
0’ 2 11 '1
PHEMA 0“
\ o c
CH3(CH2)GCOCI
/ >

0 Ct) H3 Br
‘Si CH O

0’ -{ 2171 n
\ O

o
bojéCHJgHa
o

 

CB-PHEMA

CF3(CF2)5COCI \o 0 CH3
’ O‘s'ﬂCsz—o 8'
0' 1 1 n

Fluorinated PHEMA “0130339
0

 

\ 0 C

0 H3
0. I Br
Si CH O
O’ -{ 21:1 W n
\ O

o
K/OjrlCHJ-aHa
0

CH CH COC
K 3( 2114 >

 

— [:3 — —v

C16-PHEMA
Scheme 2.1. Attachment of a trichlorosilane initiator to a porous alumina
support, ATRP from the immobilized initiator, and derivatization of PHEMA

with acid chlorides.

54

membrane materials without the need to develop new polymerization processes, and this
chapter compares the pervaporation properties of three different elaborations of PHEMA.

Sorption studies help elucidate the factors behind the selectivities of different ﬁlms.

2.2. Experimental Section
2.2.1. Materials

Octanoyl chloride (99%), palmitoyl chloride (98%), pentadecaﬂuorooctanoyl
chloride (97%), dimethylformamide (DMF, anhydrous, 99.8%), tetrahydrofuran (T HP,
anhydrous, inhibitor free, 99.8%), ll-mercaptoundecanol (97%), 2-bromopropionyl
bromide (97%), ethyl 2-bromoisobutyrate (98%), CuCl (99.999%), CuBrz (99%), and
2,2’-bipyridine (bpy, 99%) were used as received from Aldrich. 2-hydroxyethyl
methacrylate (HEMA, Aldrich, 97%, inhibited with 300 ppm hydroquinone monomethyl
ether) was puriﬁed by passing it through a column of activated basic alumina (Spectrum),
and the trichlorosilane initiator (1 1-(2-bromo-2-
methyl)propionyloxy)undecy1trichlorosilane) was synthesized according to a literature
procedure.3 Sorption and pervaporation tests employed deionized water (Milli-Q, 18.2
Mﬂcm) and one of the following solvents: ethyl alcohol (EtOH, 100%, Pharmco), ethyl
acetate (EtOAc), dichloromethane (CH2C12), trichloroethylene (TCE), or benzene. Other
than ethanol, solvents were analytical grade and purchased from Aldrich. AnodiscTM
porous alumina membranes (Fisher) with 0.02 um-diameter surface pores were used as
supports for membrane formation, while gold-coated wafers (200 nm of sputtered Au on
20 nm Cr on a Si (100) wafer) were used as substrates for ellipsometry, contact angle

measurements, and reﬂectance Fourier Transform Infrared (FT IR) spectroscopy.

55

2.2.2. Polymerization of HEMA and Subsequent Derivatization

The alumina substrates were ﬁrst cleaned in a UV/ozone cleaner (Boekel model
135500) for 15 minutes, inserted into a glove box, and immersed in a 2 uM solution of
trichlorosilane initiator in anhydrous THF for ~12 h. After removal from the glove box,
samples were rinsed with acetone, sonicated in DMF for 5 minutes, rinsed with water and
acetone, and ﬁnally dried with a ﬂow of nitrogen. Deposition of initiators on Au-coated
wafers occurred through formation of a mercaptoundecanol monolayer and subsequent
derivatization of this layer with 2-bromopropiony1 bromide as described in the literature.7

Polymerization of HEMA occurred by immersion of the initiator-coated substrate
in an aqueous HEMA solution containing a Cu catalyst system (Scheme 2.1).5 To
prepare this polymerization solution, 15 mL of puriﬁed monomer was added to 15 mL of
deionized water, and this solution was degassed via three freeze-pump-thaw cycles. Next,
82.5 mg (0.825 mmol) of CuCl, 54 mg (0.24 mmol) of CuBrz, and 320 mg (2.04 mmol)
of bpy were quickly added to the mixture, which was then subjected to another freeze-
pump-thaw cycle and stirred until a homogeneous, dark brown solution formed. Initiator-
coated substrates and the sealed ﬂask containing the polymerization solution were then
transferred to a glove bag that was purged with nitrogen for about 1 h. The
polymerization solution was ﬁnally transferred to the container holding the substrates,
and polymerization was carried out for 2 h. After polymerization, substrates were
removed from the container, sonicated in DMF for 10 minutes, rinsed with water
followed by acetone, and dried with a ﬂow of nitrogen. Subsequent derivatizations of
PHEMA were carried out by immersing the PHEMA-coated substrates in DMF solutions

containing 0.2 M octanoyl chloride, pentadecaﬂuorooctanoyl chloride, or palmitoyl

56

chloride and 0.1 M pyridine. After a 2 h immersion in the solution of acid chloride,
substrates were rinsed with DMF followed by ethanol, and dried with a ﬂow of nitrogen.
2.2.3. Characterization of Membranes

Film growth on alumina supports was veriﬁed by ﬁeld-emission scanning electron
microscopy (FESEM, Hitachi S-470011, acceleration voltage of 10 kV). Membranes
were freeze-fractured under liquid N2 and sputter-coated (Pelco model SC-7) on both
sides with 5 nm of gold prior to FESEM. Ellipsometric thickness determinations were
performed using a rotating analyzer ellipsometer (model M-44, J.A. Woollam). We
assumed a ﬁlm refractive index of 1.5, except in the case of the ﬂuorinated ﬁlm, where
both refractive index and thickness were ﬁt to ellipsometric data. These ellipsometric
studies were carried out with ﬁlms grown from gold-coated wafers. For each polymer
ﬁlm, thicknesses were measured at three different spots and averaged. Reﬂectance FT IR
spectroscopy of ﬁlms on gold-coated wafers was performed with a Nicolet Magna-IR 560
instrument using a Pike grazing angle (80°) accessory. Static water contact-angle
measurements (Firsttenangstroms contact angle analyzer) on alumina membranes were
also performed before and after derivatization.
2.2.4. Pervaporation Experiments

0ur home-built pervaporation apparatus was described previously.M Brieﬂy, the
solution was pumped across the surface of a membrane (effective area of 3.1 cm2) that
was supported by a stainless steel frit. The membrane cell was connected to a coiled
stainless steel feed tube, and both the membrane cell and coiled tube were immersed into
a thermostated water bath to control the temperature of the feed solution. 0n the

permeate side of the membrane, a vacuum pressure of 0.06 mbar was applied, and the

57

permeate was collected in one of two liquid nitrogen-cooled traps. Feed solution was
usually pumped across the membrane at a ﬂow rate of 5 mUmin, but in a few
experiments, we did vary ﬂow rate to determine if there were large boundary layer effects
on transport. In pervaporation of 0.05 wt% EtOAc, CHzClz, or TCE through C8-PHEMA
at 65 °C, variation of feed ﬂow rates from 1.4 mUmin to 11 mUmin yielded minimal
changes in ﬂux (<3% variation) and selectivity (<10% variation). Pervaporation of 0.05
wt% CH2C12 through ﬂuorinated PHEMA membranes at 22 °C also showed little
dependence on ﬂow rate, suggesting that boundary layer effects are negligible. Tubing
on the permeate side was warmed with heating tape to avoid condensation. Finally, prior
to sample collection, pervaporation was performed for at least 1 h to achieve stable,
steady-state ﬂuxes, and then the samples were collected in a second trap. The masses of
the collected samples were measured using an electronic balance, and the composition of
the permeate was determined by gas chromatography (Shimadzu GC-17A equipped with
a Restek RTx-BACl column), using methanol as an internal standard. In cases where the
permeate separated into organic and aqueous phases, the sample was diluted sufﬁciently

with water to achieve a homogeneous mixture.

The separation effectiveness of pervaporation is quantiﬁed by two parameters,

ﬂux and selectivity. Flux (J) can be calculated using equation 1,

J Q

=— 1
At ()

where Q is the mass of the permeate collected in time t, and A is the area of the

membrane exposed to the feed solution. In the case of a binary feed of components A

and B, selectivity or separation factor, a; , can be expressed by equation 2,

58

A _ yA/YB
aB -
XA/XB

 

(2)

where x and y represent mass fractions in the feed and permeate, respectively. Reported

selectivities and ﬂuxes are averages of results from three different membranes.

2.2.5. Sorption Measurements

Sorption experiments were performed with bulk polymers because of the inherent
difﬁculties in measuring sorption in ultrathin ﬁlms. Derivatized HEMA monomers were
ﬁrst synthesized by allowing acid chlorides to react with HEMA, and then these
monomers were polymerized in solution. Speciﬁcally, HEMA (10 mL, 0.08 mol) was
placed in a 250 mL round-bottom ﬂask and dissolved in 100 mL of anhydrous
dichloromethane. The ﬂask was sealed with a septum, and pyridine (8 mL, 0.1 mol) was
added to this solution followed by 0.09 mol of the desired acid chloride. The solution
was purged with nitrogen for 1 h and stirred overnight at room temperature. The mixture
was extracted with 1 M aqueous HCl, dried over Na2804, and the solvent was removed
by rotary evaporation. The product was puriﬁed by column chromatography (20%
EtOAc in hexanes, basic alumina as stationary phase) and dried under vacuum to obtain
monomers as colorless oils. C8-HEMA (18 g, 90% yield), 1H NMR (300 MHz, CDCl3) 5
6.17-6.04 (1H, d), 5.65-5.49 (1H, d), 4.47—4.19 (4H, m), 2.45-2.19 (2H, m), 2.06-1.84
(3H, s), 1.76-1.49 (2H, m), 1.42-1.23 (8H, m), 1.01-0.75 (3H, m). C16-HEMA (25 g,
85% yield), 1H NMR (300 MHz, CDC13) 8 6.19-6.03 (1H, d), 5.65-5.48 (1H, d), 4.44-
4.22 (4H, m), 2.47-2.19 (2H, m), 2.03-1.85 (3H, s), 1.72-1.47 (2H, m), 1.44-1.10 (24H,

m), 0.98-0.76 (3H, m). Fluorinated-HEMA (30 g, 70% yield), 1H NMR (300 MHz,

59

CDCl3) 5 6.23-6.02 ( 1H, d), 5.70-5.49 (1H, d), 4.39-4.16 (2H, m), 3.94-3.67 (2H, m),
2.07-1.83 (3H, s).

For polymerization, 0.05 mol of monomer was added to a 100 mL Schlenk ﬂask
and degassed by three freeze-pump-thaw cycles. The ﬂask was then inserted into a glove
box, and 40 mL of anhydrous dichloromethane was added to the monomer. Next CuBrz,
CuCl, and bpy (0.82 mmol, 0.24 mmol, and 2.0 mmol, respectively) were added to the
solution followed by the initiator, 0.50 mmol of ethyl 2-bromoisobutyrate. After 10 h,
the ﬂask was taken out of the glove box, and the brown solution was extracted with
aqueous 1 M HCl to remove the copper complex. Evaporation of the solvent yielded
colorless, viscous polymers. Those polymers were then dissolved in acetone (3 g/ 10 mL),
and 2 mL of this polymer solution was cast on a 2 x 1 cm gold-coated Si wafer. After
letting the acetone evaporate for 4 h at room temperature, the cast ﬁlms were dried
overnight in an oven at 60 °C, and ﬁnally removed from the substrate.

To measure sorption selectivities and polymer swelling by solvents, the
derivatized bulk PHEMA ﬁlms were immersed for 24 h in room-temperature aqueous
solutions containing 0.05 wt % of a volatile organic compound (VOC). After that, the
ﬁlms were taken out of the vessel, wiped quickly with ﬁlter paper, and weighed. Degree
of sorption of the VOC solution into the membranes was determined using equation 3,
where m and m; denote the weights of the dried membrane and the swollen membrane,
respectively.

Degree of sorption = (m1- m0) / m x 100 (3)
To measure sorption selectivity, we employed a literature procedure. Brieﬂy,

solvent-swollen membranes were dried with ﬁlter paper and placed in a trap that was

60

quickly cooled with liquid nitrogen and subsequently evacuated.l5 The sample was then
heated with a blow drier, and the vapor was collected in a second, cooled ﬂask that was
connected to the ﬁrst by a valve. The mass of the bulk polymer was then measured again
to make sure that we had removed essentially all of the water and solvent. The
concentration of the VOC in the collected solution was determined by gas
chromatography, and the separation factor was calculated using equation 2 with y

representing the composition of the adsorbed liquid that was eventually collected.

2.3. Results and Discussion

2.3.1. FTIR and SEM Characterization of Derivatized PHEMA Membranes

PHEMA is relatively hydrophilic and, hence, would not be expected to allow
selective pervaporation of VOCs from water. However, each repeat unit of PHEMA has
a hydroxyl group that can be easily esteriﬁed with a variety of acid chlorides. We let
PHEMA ﬁlms react with octanoyl chloride, palmitoyl chloride and
pentadecaﬂuorooctanoyl chloride in the presence of a base to obtain polymers with
hydrophobic side chains (Scheme 2.1). The static water contact angle on native PHEMA
ﬁlms is 50°, but contact angles increase to 75°, 95°, and 98° for porous alumina-supported
ﬁlms derivatized with octanoyl chloride (CS-PHEMA), palmitoyl chloride (C 16-
PHEMA), and pentadecaﬂuorooctanoyl chloride (ﬂuorinated PHEMA), respectively.
These values fall between the advancing and receding contact angles reported by
Jennings and coworkers.” '6 To verify derivatization of PHEMA using reflectance FTIR

5,13

spectroscopy, we grew ﬁlms on gold-coated wafers as described previously. Figure

» 2.1 shows the reﬂectance FT IR spectra of PHEMA ﬁlms before and after esteriﬁcation.

61

Nearly complete disappearance of the CH stretch at 3500 cm‘l and a doubling of the
carbonyl (CO) stretch at 1700 cm'1 suggest >90% conversion of the hydroxyl groups to
esters for C8- and C16-PHEMA ﬁlms.5 In the case of ﬂuorinated PHEMA, the CH
stretch also disappears, but the peak due to the newly formed ester carbonyl appears at
1790 cm‘l, rather than at the position of the initial ester of PHEMA. (The

electronegativity of the F atoms results in the shifted position of the carbonyl stretch.)

62

Absorbance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a b
l 0-01 I 0.05
3
C8- 5
PHEMA %
3 Fluorinated
“ PHEMA
35'00 3000 25100 201m 1500 1000 :500 3&0 25100 20"” 15‘00 1000
Wavenumbers (cm'1) Wavenumbers (cmd)
C
0.02
8
3 C16-
8 PHEMA
.n
<
PHEMA
asoo 3060 asho 2800 who who

Wavenumbers (cm")

Figure 2.1. Reﬂectance FTIR spectra of PHEMA ﬁlms on gold before and

after derivatization with (a) octanoyl chloride, (b) pentadecaﬂuorooctanoyl

chloride, and (c) palmitoyl chloride.

63

Figure 2.2 shows cross-sectional FESEM images of PHEMA, C8-PHEMA, C16-
PHEMA, and ﬂuorinated PHEMA ﬁlms grown from porous alumina supports. Before
derivatization, the thickness of PHEMA ﬁlms is about 50 nm, as shown in Figure 2.2(a),
and the ﬁlm fully covers substrate pores. Ellipsometric measurements of PHEMA ﬁlms
grown on gold-coated wafers under the same conditions also give a ﬁlm thickness of 50

nm. After derivatization with octanoyl chloride, the molecular weight of the polymer

~ 5 ~

ii-‘x
i _

i 1
IV A
., a
‘lSO‘t-V 16 0mm xiii} SEW} 51520456 17’5‘. . . .

Fluorinated-PH EMA

 

Figure 2.2. Cross-sectional FESEM images of alumina supports coated with

(a) PHEMA, (b) C8-PHEMA, (c) Cl6-PHEMA, and (d) ﬂuorinated PHEMA.

repeat unit should approximately double, so we would expect that the thickness of C8-

PHEMA ﬁlms would be about twice that for PHEMA, assuming a similar density for

both ﬁlms.5 Both ellipsometry results and SEM images such as that in Figure 2.2(b)
indicate that ﬁlm thickness does indeed double upon reaction with octanoyl chloride.

For C16-PHEMA, we would expect to see a 180% increase in ﬁlm thickness
based solely on the molecular weight of the new repeat unit. Ellipsometric measurements
show a thickness increase of 125%. The SEM image in Figure 2.2(c) suggests a ~200%
increase in thickness, but it is difﬁcult to determine exactly where the ﬁlm begins in this
image. In the case of ﬂuorinated ﬁlms, the polymer density should be higher than that of

hydrogenated aliphatic polymers,” '8

so it is not difﬁcult to understand why we saw only
a 170% increase in ﬁlm thickness in ellipsometry in spite of a 300% increase in the
molecular weight of the monomer unit. The FESEM image suggests a 240% increase in
thickness. We should note that the high density of ﬂuorinated PHEMA reﬂects the high
atomic mass of ﬂuorine relative to H and does not imply a low free volume for the
ﬂuorinated ﬁlm. Most importantly, the PHEMA membranes and their derivatives are an
order of magnitude or more thinner than typical PDMS membranes and other types of
hydrophobic membranes currently used in pervaporation.'9'2' Moreover, the ATRP
process (Scheme 2.1) allows effective control over the thickness of the membranes by
changing the polymerization time.
2.3.2. Pervaporation of Ethyl acetate-water Mixtures Using Derivatized PHEMA
2.3.2-a. Eﬁ'ect of Feed Concentration

We initially employed EtOAc/water mixtures to probe the pervaporation
properties of derivatized PHEMA membranes as a function of feed concentration and

temperature. Figures 2.3, 2.4, and 2.5 show that trends in ﬂux and selectivity related to

feed composition vary dramatically among the different membranes. For C8-PHEMA

65

 

 

 

300/0 "' 30
£1
8
§ 25% --25
0
D.
C a.
'3 20% - -- 20 .3
< a
9 IL
E 15%1 «- 15 ,5
° ‘6
S E
'3 10%- -- 10 g
E
g 5% a a. 5
O
3

00/0 1 I I I O

 

0.0% 2.0% 4.0% 6.0%
Weight Fraction of EtOAc in Feed

 

 

 

 

 

 

 

1.2
a
1
,3? 0.8
E
g 06 , +Total Flux
x + Water Flux
3 0 4 -
u. ° + EtOAc Flux
0.2 -
0 I I I
0.0% 2.0% 4.0% 6.0%

Weight Fraction of EtOAc in Feed

Figure 2.3. Permeate composition (top), separation factor (top), and ﬂux
(bottom) in pervaporation of several ethyl acetate/water mixtures through C8-
PHEMA membranes at 50 °C. The standard deviations of ﬂux values and

ethyl acetate weight fractions in the permeate were less than 15%.

66

 

100% - -- 180
3 w 160
8
g 800/0 ‘_ 140
83
.E ‘P 120 3
2 600/0 " E
g -- 100 “=-
5 .2
o .. 30 ‘5
g 400/0 '1 i
g "60 8
IL
35, 20% " 4°
0
B -1- 20

00/0 I ﬂ T I 0

 

 

 

0.00% 2.00% 4.00% 6.00%
Weight Fraction of EtOAc in Feed

 

   
 

 

 

 

 

 

0.60 -
0.50 ._ + Total Flux
+ Water Flux
a? 0.40 - + EtOAc Flux
E
g 0.30 1
x
:3
E 0.20 -
0.10
0.00 F I I U
0.0% 2.0% 4.0% 6.0%

Figure 2.4. Permeate composition (top), separation factor (top), and ﬂux
(bottom) in pervaporation of several ethyl acetate/water mixtures through

C16—PHEMA membranes at 50 0C. The standard deviations of ﬂux values

and ethyl acetate weight fractions in the permeate were less than 15 %.

67

35%

 

 

 

 

2
8 30% all up 30
E
0
Q 250/0 "' "' 25
C u.
n; g
<
9 20% - -r 20 if
I" 5
35 as
r; 15% - -- 15 E
.2 §
‘g m
.z 10% - ._ 10
.‘c‘
.9
g 5% - " 5

0% . . a . 0

0.00% 2.00% 4.00% 6.00%

Weight Fraction of EtOAc in Feed

   
   
  

 

+ Total Flux
+ Water Flux
+ EtOAc Flux

 

 

 

 

Flux (kg/man)

  

 

   

I

0.0% 2.0% 4.0% 6.0%
Weight Fraction of EtOAc in Feed

I I

Figure 2.5. Permeate composition (top), separation factor (top), and ﬂux
(bottom) in pervaporation of several ethyl acetate/water mixtures through
ﬂuorinated-PHEMA membranes at 50°C. The standard deviations of ﬂux

values and ethyl acetate weight fractions in the permeate were less than 15%.

68

membranes (Figure 2.3), the total ﬂux increases modestly from 0.8 to 1.1 kg/mzh as the
EtOAc concentration in the feed goes from 0.1% to 6%. In contrast, total ﬂux increases
4—fold on going from 0.1 to 6% EtOAc with Cl6—PHEMA membranes (Figure 2.4),
although even with 6% EtOAc, ﬂuxes through C16-PHEMA are half of those through
C8-PHEMA. Fluorinated PHEMA membranes allow the highest ﬂuxes (3- to 6—fold
higher than with C8-PHEMA), with a doubling in ﬂux as EtOAc concentration goes from
0.1 to 6% (Figure 2.5).

The trends in ﬂux values are likely a reﬂection of differences in side chain
packing in the different membrane systems. Hexadecyl side chains should pack in a
more crystalline arrangement than shorter octyl side chains, so one would expect lower
ﬂuxes through C16-PHEMA membranes. (C16-PHEMA ﬁlms are also 50% thicker than
C8-PHEMA membranes, but this should result in only a ~33% lower ﬂux for these
membranes.) The more hydrophobic C16-PHEMA should also be more susceptible to
plasticization than C8-PHEMA, which is consistent with the large increase in ﬂux
through C16-PHEMA on going from 2% to 6% EtOAc. In the case of ﬂuorinated
PHEMA, the perﬂuorooctyl chain is bulkier than typical octyl chains, and this likely
leads to higher free volume and, hence, higher permeability for the ﬂuorinated PHEMA
relative to C8-PHEMA.22’ 23 Although the ﬂuorinated ﬁlms have a slightly higher water
contact angle than Cl6-PHEMA, the ﬂuorinated systems appear to be less prone to
plasticization, as suggested from a smaller relative increase in ﬂux on going from 0.1 to
6% EtOAc. This may reﬂect the fact that typical hydrocarbons are less soluble in

ﬂuorinated than hydrogenated aliphatic chains.24

69

Separation factors also seem to depend on polymer packing. At low EtOAc
concentrations (0.1%), separation factors are similar for all three membranes (25, 45, and
30, for C8-PHEMA, C16-PHEMA, and ﬂuorinated PHEMA, respectively). However,
separation factors for C8-PHEMA and ﬂuorinated PHEMA decrease with increasing
EtOAc concentration, while selectivities of C16-PHEMA membranes show a 4-fold
increase on going from 0.1 to 6% EtOAc (Figures 2.3, 2.4, and 2.5). Perhaps due to the
tight packing in C16-PHEMA ﬁlms, absorbed EtOAc molecules block the transport of
water. Such an effect was seen previously in the separation of TCE from water using a

ﬂuoroalkyl methacrylate-grafted PDMS membrane.25

2.3.2-b. Effect of Feed Temperature

Figure 2.6 shows the ﬂuxes and selectivities for pervaporation of EtOAc solutions
through C8-PHEMA membranes at several temperatures. In agreement with other
studies,”28 ﬂux increases with increasing feed temperature, while selectivity remains
essentially constant. Generally, as temperature increases, the thermal motion of polymer
chains and permeate molecules intensiﬁes, leading to more rapid diffusion of analytes
through the membrane and, hence, higher ﬂuxes. In contrast, temperature changes do not
greatly alter separation factors because solubility (sorption) selectivity among permeating
molecules is not a strong function of temperature.26

High pervaporation temperatures could also affect membrane performance by
inducing polymer degradation. To test the stability of C8-PHEMA membranes, we
performed pervaporation of 0.05 wt% EtOAC at 65 °C before and after storage of the

membrane in the hot pervaporation solution for one week. The week-long exposure to 65

. °C 0.05 wt% EtOAc in water had no signiﬁcant effect on pervaporation. Additionally,

70

we examined the stability of Cl6-PHEMA and ﬂuorinated PHEMA ﬁlms on gold wafers
using FTIR spectroscopy. Storage of these ﬁlms in 65 °C 0.05 wt% EtOAc in water for
one week yielded no change in their IR spectra other than a uniform ~10% decrease in
intensity, which is probably due to desorption of a small quantity of polymer from the
gold surface. The desorption could result from a small amount of physisorbed polymer

or the instability of the Au-thiol linkage at high temperatures.29

71

1.35
l

+ 65°C + 50°C + 22°C I

1.15

 

 

 

H
E
NE 0.95
E:
’3‘ 0.751
E
0.55 ~
0.35 “if 1 T f
0.0% 2.0% 4.0% 6.0%

Weight Fraction of EtOAc in Feed

Separation Factor

 

 

 

0.0% 2.0% 4.0% 6.0%
Weight Fraction of EtOAc in Feed

Figure 2.6. Flux (top) and separation factor (bottom) for pervaporation of
water/ethyl acetate mixtures through C8-PHEMA membranes at several

temperatures. The standard deviations of ﬂux values and separation factors

were less than 15%.

72

2.3.3. Pervaporation of Different Organic Compounds through Derivatized
PHEMA Membranes

To further compare the properties of the derivatized PHEMA membranes, we
examined pervaporation of a series of VOCs in water. We utilized a relatively low VOC
concentration (0.05 %) in these experiments to avoid plasticization and ensure that the
VOCs remained soluble. For all VOCs, ﬂux decreased in the order ﬂuorinated PHEMA
(1.2-1.5 kg/mzh) >C8-PHEMA (0.3—0.5 kg/mzh) >C16-PHEMA (0.1 — 0.2 kg/mzh).
However, VOC/water separation factors did not vary greatly among the three types of
membranes (Table 2.1). Selectivity did increase with the hydrophobicity of the VOC
(decreasing water solubility) from EtOH to TCE, with the exception of CHzClz. The

higher than expected CH2C12/water selectivity may occur in part because the small molar

Table 2.1. Molecular properties22 of several VOCs and separation factors
(VOC/Water) for pervaporation (22 0C) of 0.05% solutions through C8-

PHEMA, Cl6-PHEMA, and ﬂuorinated PHEMA membranes.

 

 

 

73

Molar Solubility Separation Factor
Compound Volume 38' in water
(mL/mol) (W170) C8 C16 Fluorinated

Water 18.0 100 —— — — __

EtOH 58.5 78 Miscible 20:05 3.0103 1.0 10.5
EtOAc 98.5 77 8.24 48:3 56:5 49:1
CHzCl2 63.9 40 2.00 330:30 430160 430160
benzene 89.0 80 0. l 8 190130 250120 240120

TCE 90.2 87 0.1 l 430:50 540150 510190

volume of CHzClz relative to EtOAc, benzene, and TCE results in more rapid diffusion of
CHzClz.30 The low boiling point of CHzClz may also increase its pervaporative transport.
2.3.4. Sorption Behavior and the Solution-diffusion Model

In the solution-diffusion model, transport through membranes includes both
sorption and diffusion. To decouple the effects of diffusion and sorption on selectivity,
we measured sorption directly for all three PHEMA derivatives. To attain sufﬁcient
sensitivity for these measurements, we employed bulk polymers rather than thin polymer
brushes grown from a surface, but sorption properties of bulk polymers and thin ﬁlms

should be similar. Table 2.2 shows that both the degrees of sorption and VOC/water

Table 2.2. Degrees of sorption and sorption selectivities for ﬁve VOCs
(0.05% aqueous solutions) in C8-PHEMA, C16-PHEMA, and ﬂuorinated
PHEMA ﬁlms. Measurements were performed at 22 °C, and solubilities of the

VOCs in water are given for reference.

 

 

 

Degree of Sorption Selectivity
Solubility
Compound in water
”1%) cs C16 Fluorinated C8 C16 Fluorinated
EtOH Miscible 4.2:0.7 37:04 64:07 8.1:1.2 12310.5 3.5:1 .2
EtOAc 8.24 48:5 10:1 59:7 203:30 369:5 120:17
C HzCl2 2.00 50:8 20:2 68:7 520:5 645:6 450:23
benzene 0.18 55:12 27:1 88:9 580:56 1600:20 310:30
TCE 0. l 1 68:9 32:2 99:15 1 540:90 3700:50 570:80

sorption selectivities for the three derivatized PHEMA membranes generally increase

- with decreasing water solubility of the VOCs, as would be expected. The ﬂuorinated

74

membrane provides the only exception to this trend with the sorption selectivity of
benzene being 30% lower than that of CHzClz.

Among the three membranes, C16-PHEMA exhibits both the highest sorption
selectivity and the lowest degree of sorption. This is consistent with the high
hydrophobicity of this polymer and a low free volume. The ﬂuorinated polymer has the
highest degree of sorption of the three derivatives, suggesting again that this material has
a high free volume. However, with the exception of CH2C12, sorption selectivities for
C8-PHEMA are 70% to 170% greater than those for ﬂuorinated PHEMA. Although the
ﬂuorinated membrane has a higher water contact angle than C8-PHEMA, the low

solubility of organics in ﬂuorous phases likely reduces VOC/water sorption selectivity.25

Table 2.3. Separation factors (22 °C) for pervaporation (apv), sorption (as),
and diffusion ((1.1)) of VOCS (0.05% aqueous solutions) in C8-PHEMA, C16-

PHEMA, and ﬂuorinated PHEMA membranes.

 

 

 

Molar capmamn Cl6-PHEMA Fluorinated PHEMA
Volume
Compound
(mUmO') “9v as 0In any as an aw as an
EtOH 58.5 2.0 8.1 0.25 3.0 12.3 0.24 1.0 3.5 0.26
EtOAc 98.5 48 203 0.24 56 369 0.15 49 120 0.41
CH2C12 63.9 330 520 0.63 430 645 0.67 430 460 0.93
benzene 89.0 190 580 0.33 250 1600 0.16 240 310 0.77
TCE 90.2 430 1540 0.28 540 3700 0.14 510 570 0.89

75

The pervaporation separation factor, apv, in the solution-diffusion model is simply the
product of the sorption selectivity, as, and the diffusion selectivity, (in. Table 2.3
contains values of apv and as along with calculated values of (19 for all three membranes.
Because of the small molar volume of water (18 mUmol), an is less than 1 for all of the
membranes. The higher diffusion selectivity for the ﬂuorinated membranes (lower
water/VOC selectivity) is probably another reﬂection of a high free volume. Among the
VOCS, CH2C12 shows the largest on, which would be expected for this relatively small
molecule. However, ethanol has an even lower molar volume than CHzClz, but its an is
around 0.25 for all of the membranes. Perhaps ethanol is diffusing through the
membrane in a hydrated state.3| Trends in up for EtOAc, benzene, CHzClz, and TCE
correlate reasonably well with molar volume, with the exception of TCE in the
ﬂuorinated membrane.

2.3.5. Comparison with Related Membrane Systems

Table 2.4 compares the pervaporation performance of ﬂuorinated PHEMA
membranes with several PDMS membranes. Blume and coworkers coated microporous
polymer supports with PDMS to achieve 3.5-um thick PDMS skins (row 1, Table 2.4).32
Those membranes show TCE/water pervaporation selectivities that are similar to those of
ﬂuorinated PHEMA, but the ﬂuorinated PHEMA allows a 7-fold higher ﬂux because it is
20-times thinner. When normalized to ﬁlm thickness, the PDMS membranes are more
permeable than ﬂuorinated PHEMA, but Blume and coworkers reported that their coating

method could not produce defect-free ﬁlms with thicknesses less than 0.5 um.

In the case of commercial PERVAP 1060 and 1070 membranes, EtOAc/water

selectivity is 10- to 20-fold higher for the commercial membranes than for ﬂuorinated

76

PHEMA. However, EtOAc ﬂux is 2.5 to lO-fold higher for ﬂuorinated PHEMA, again

because of the minimal thickness of the polymer brushes (Table 2.4). (Both PERVAP

Table 2.4. Comparison of the pervaporation performance of ﬂuorinated

PHEMA and several PDMS membranes.

 

Membrane

 

F dC trat' “r t lFl '
Membrane VOC ee oncen lon T Thickness o a ux Separation
(wt%) (°C) (kg/mzh) Factor
(um)
PDMS 32 TCE 0.05 30 3.5 0.2 440
Fluorinated
PHEMA TCE 0.05 22 0.17 1.5 510
PDMS 2°
EtOAc l 30 8 0.4 a 300
(PERVAP 1050)
PDMS 20 ,
EtOAc 1 30 10 0.18 450
(PERVAP 1070)
Fluorinated a
PHEMA EtOAc l 22 0.17 l 25
PDMSDMMA33 benzene 0.05 40 270 0.0514 1853
Fluorinated ,
PHEM A benzene 0.05 22 0.17 1.5 240
a EtOAc ﬂux only.

1060 and PERVAP 1070 are composite membranes with a PDMS skin, but PERVAP
1070 is ﬁlled with zeolites.) In a recent paper, Uragami and coworkers polymerized
PDMS dimethyl methacrylate macromonomers to prepare extremely selective
membranes for pervaporation of benzene from water.33 These membranes have 8-fold
higher benzene/water selectivities than ﬂuorinated PHEMA, but again the polymer brush
membranes allow much higher ﬂuxes. All of these comparisons neglect the fact that the
transport through PHEMA membranes was measured at lower temperatures than those

used with the comparison membranes. Higher pervaporation temperatures could result in

77

as much as a doubling of ﬂux, so transport through ﬂuorinated PHEMA is generally an
order of magnitude or more greater than that reported for high-performance membranes.
Additionally, decreasing the thickness of ﬂuorinated PHEMA ﬁlms may also allow
further increases in ﬂux. Although these comparisons demonstrate the high ﬂuxes that
can be achieved by using ultrathin polymer brushes as the skin layers in composite
membranes, they also show that it may be possible to achieve even higher ﬂuxes and
selectivities if polymer brushes with different compositions can be prepared. Our future

work will focus on this area.

2.4. Conclusions

Membranes for selective pervaporation of VOCs were prepared by ATRP of
HEMA from the surface of porous alumina followed by derivatization of the resulting
PHEMA ﬁlms with hydrophobic acid chlorides. Reﬂectance FT IR spectroscopy and
FESEM measurements conﬁrm the derivatization process and show that the derivatized
PHEMA skin layers fully cover the porous alumina substrates. The VOC/water
pervaporation selectivities of these membranes reach values as high as 500, and the
minimal thickness of the PHEMA ﬁlms allows ﬂuxes that are an order of magnitude
higher than those of currently used pervaporation membranes. Flux through ﬂuorinated
PHEMA is higher than through C8- or C16-PHEMA, presumably because the bulky
ﬂuorinated side chains pack less tightly than alkyl chains. C16-PHEMA exhibits the
highest sorption selectivities, but the tight packing of the C16 chains likely results in
signiﬁcant water/organic diffusion selectivity. Overall, tradeoffs between diffusion and
sorption selectivities result in similar pervaporation selectivities for ﬂuorinated PHEMA,

, C8-PHEMA and C16-PHEMA. Given the similar selectivities, the ﬂuorinated system

78

appears to be most promising because of its high ﬂux. ATRP of other monomers from

porous surfaces may yield membranes with even higher ﬂuxes and selectivities.

79

2.5.

10.

11.

12.

13.

14.

15.

16.

17.

References

Huang, W.; Baker, G. L.; Bruening, M. L., Angew. Chem, Int. Ed. 2001, 40,
1510-1512.

Kim, J .-B.; Bruening, M. L.; Baker, G. L., J. Am. Chem. Soc. 2000, 122, 7616-
7617.

Matyjaszewski, K.; Xia, J., Chem. Rev. 2001, 101, 2921-2990.
Wang, J .-S.; Matyjaszewski, K., J. Am. Chem. Soc. 1995, 117, 5614-5615.

Huang, W.; Kim, J .-B.; Bruening, M. L.; Baker, G. L., Macromolecules 2002, 35,
1 175-1 179.

Jones, D. M.; Huck, W. T. 8., Adv. Mater. 2001, 13, 1256-1259.

Kim, J .-B.; Huang, W.; Bruening, M. L.; Baker, G. L., Macromolecules 2002, 35,
5410-5416.

Carlmark, A.; Malmstroem, E. E., Biomacromolecules 2003, 4, 1740-1745.
Xiao, D.; Zhang, H.; Wirth, M., Langmuir 2002, 18, 9971-9976.

Luo, N.; Husson, S. M.; Hirt, D. E.; Schwark, D. W., J. Appl. Polym. Sci. 2004,
92, 1589-1595.

Balachandra, A. M.; Baker, G. L.; Bruening, M. L., J. Membr. Sci. 2003, 227, 1-
14.

Bantz, M. R.; Brantley, E. L.; Weinstein, R. D.; Moriarty, J .; Jennings, G. K., J.
Phys. Chem. B 2004, 108, 9787-9794.

Brantley, E. L.; Holmes, T. C.; Jennings, G. K., J. Phys. Chem. B 2004, 108,
16077-16084.

Sullivan, D. M.; Bruening, M. L., J. Membr. Sci. 2005, 248, 161-170.

Chang, Y.-H.; Kim, J.-H.; Lee, S.-B.; Rhee, H.-W., J. Appl. Polym. Sci. 2000, 77,
2691-2702.

Brantley, E. L.; Jennings, G. K., Macromolecules 2004, 3 7, 1476-1483.

Alentiev, A. Y.; Shantarovich, V. P.; Merkel, T. C.; Bondar, V. 1.; Freeman, B. D.;

Yampolskii, Y. P., Macromolecules 2002, 35, 9513-9522.

80

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

De Angelis, M. G.; Merkel, T. C.; Bondar, V. 1.; Freeman, B. D.; Doghieri, F.;
Sarti, G. C., Macromolecules 2002, 35, 1276-1288.

Jonquieres, A.; Clement, R.; Lochon, P.; Neel, J .; Dresch, M.; Chretien, B., J.
Membr. Sci. 2002, 206, 87-1 17.

Kujawski, W., Sep. Sci. Technol. 2000, 35, 89-108.
Okumus, E.; Gurkan, T.; Yilmaz, L., J. Membr. Sci. 2003, 223, 23-38.

Kim, J. H.; Chang, B. J.; Lee, S. B.; Kim, S. Y., J. Membr. Sci. 2000, 169, 185-
196.

Nagel, C.; Guenther-Schade, K.; Fritsch, D.; Strunskus, T.; Faupel, F.,
Macromolecules 2002, 35, 2071-2077.

Nakamura, M.; Samejima, S.; Kawasaki, T., J. Membr. Sci. 1988, 36, 343-351.
Mishima, S.; Kaneoka, H.; Nakagawa, T., J. Appl. Polym. Sci. 1999, 71 , 273-287.
Ki Hong, Y.; Hi Hong, W., J. Membr. Sci. 1999, 159, 29-39.

Matsuda, H.; Yanagishita, H.; Negishi, H.; Kitamoto, D.; Ikegami, T.; Haraya, K.;
Nakane, T.; Idemoto, Y.; Koura, N.; Sano, T., J. Membr. Sci. 2002, 210, 433-437.

Sano, T.; Yanagishita, H.; Kiyozumi, Y.; Mizukami, F.; Haraya, K., J. Membr.
Sci. 1994, 95, 221-228.

Huang, W.; Skanth, G.; Baker, G. L.; Bruening, M. L., Langmuir 2001, 17, 1731-
1736.

Johnson, T.; Thomas, S., J. Appl. Polym. Sci. 1999, 71 , 2365-2379.
Radovanovic, P.; Thiel, S. W.; Hwang, S. T., J. Membr. Sci. 1990, 48, 55-65.
Blume, 1.; Wijmans, J. G.; Baker, R. W., J. Membr. Sci. 1990, 49, 253-286.

Uragami, T.; 0hshima, T.; Miyata, T., Macromolecules 2003, 36, 9430-9436.

81

Chapter 3
High-Capacity, Protein-Binding Membranes Based on Polymer

Brushes Grown in Porous Substrates

3.1. Introduction

While the previous chapter discussed the synthesis of ultrathin polymer ﬁlms at
the surface of porous substrates to prepare high-ﬂux membranes, this chapter explores the
growth of brushes in the interior of porous substrates to prepare membranes with
unusually high adsorption capacities. Such systems may prove useful for rapid
puriﬁcation of proteins, which is important because potential therapeutic applications of
proteins along with rapid developments in biotechnology and genetic engineering have
greatly increased protein production and the need for fast, convenient puriﬁcation
technologies.“ 2 Afﬁnity chromatography is currently the most convenient way to purify
proteins due to its ability to separate biomolecules based on their biological interactions}
3 but the rate of separations with packed-bead afﬁnity columns is limited by slow
diffusion of biomacromolecules within bead pores.3’5

Afﬁnity membrane chromatography, which was initially introduced by Henis and
coworkers in 1987,6 has the potential to overcome this disadvantage of packed-bead
columns because the protein solutions being treated must ﬂow through the membrane

pores.7‘9

Additionally, membrane chromatography avoids challenges in packing of
columns, which makes it an attractive technique for large-scale separation processes. A

. . . 0 .
number of adsorptlve membranes are now avallable commerClally,l and a wrde range of

82

2. 3 - 4 15
' ' chltosan,l and nylon have been used as

porous materials such as cellulose,ll chitin,
substrates that can be modiﬁed for protein immobilization.

This research aims at utilizing the controlled growth of polymer brushes in porous
substrates to develop afﬁnity membranes with remarkably high protein-binding
capacities. The use of ATRP to grow PHEMA from initiators bound to a porous alumina
surface affords relatively ﬁne control over polymer molecular weight, which allows large
increases in capacity without clogging of membrane pores. PHEMA brushes are
particularly attractive because they can be functionalized to exploit a number of afﬁnity

interactions. '6' l 8

Speciﬁcally, functionalization of PHEMA with nitrilotriacetate-Cu2+
complexes results in protein binding via metal-ion afﬁnity interactions, and the
microporous alumina support provides a ~500-fold increase in surface area relative to
two-dimensional supports. Overall, the combination of the alumina supports and
functionalized polymer brushes described here yields a remarkable binding capacity of
0.9 mg of bovine serum albumin (BSA) per cm2 of external membrane surface (150
mg/cm3 of membrane). Typical membrane absorbers have protein capacities of 4 - 20
mycmsw—zl

A few recent studies demonstrated the possibility of using ATRP and other
polymerization techniques to prepare protein-adsorbing polymer brushesz’ '9’ ”'24
Matyjaszewski and coworkers reported electrostatic adsorption of 10 to 15 monolayers of
BSA to poly(dimethylamino ethylmethacrylate) brushes grown on ﬂat surfaces.22

Luzinov and coworkers used ATRP to prepare poly(2-vinylpyridine) brushes in porous

PVDF. Although these brushes were reported to enhance lysozyme adsorption, no

83

capacity was provided.25 The PHEMA-modiﬁed membranes described here clearly

demonstrate the advantages of polymer brushes for membrane-based afﬁnity separations.

3.2. Experimental Section

3.2.1. Materials

AnodiscTM porous alumina membranes with 0.2 um-diameter surface pores were
obtained from Fisher Scientiﬁc. SEM images suggest that pore diameters in the bulk of
these membranes are about 0.25 um. Dimethylforrnamide (DMF, anhydrous, 99.8%),
ll-mercaptoundecanol (97%), 2-bromoisobutyryl bromide (98%), CuCl (99.999%),
CuBrz (99%), 2,2’-bipyridyl (bpy, 99%), l-[3-(dimethylamino)propyl]-3-
ethylcarbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), 4-
dimethylaminopyridine (DMAP), ethylenediamine tetraacetic acid (EDTA), bovine
serum albumin (BSA), and myoglobin were used as received from Sigma Aldrich.
CuSOa-SHZO (Columbus Chemical), NaNa—bis(carboxymethyl)-L-lysine hydrate (Fluka,
aminobutyl-NTA), succinic anhydride (SA, Matheson Coleman & Bell), and Coomassie
protein assay reagent (Pierce) were also used as received. 2-Hydroxyethyl methacrylate
(HEMA, Aldrich, 97%, inhibited with 300 ppm hydroquinone monomethyl ether) was
puriﬁed by passing it through a column of activated basic alumina (Spectrum), and the
trichlorosilane initiator (11-(2-bromo-2-methyl)propionyloxy)undecyltrichlorosilane) was
synthesized according to a literature procedure.26 Buffers were prepared using analytical

grade chemicals and deionized (Milli-Q, 18.2 MQ cm) water.

84

3.2.2. Polymerization of HEMA in Porous Alumina Membranes and on Au
Substrates

The alumina membrane was ﬁrst cleaned in a UV/ozone cleaner (Boekel model
135500) for 12 minutes and sandwiched inside a membrane cell (Millipore, Swinnex 25).
A Masterﬂex 7518-60 peristaltic pump was used to circulate 1 mM trichlorosilane
initiator in DMF solution through the membrane for 3 h at a rate of 10 mUmin, and
subsequent rinsing with ethanol (20 mL), water (20 mL) and acetone (20 mL) at a ﬂow
rate of 10 mL/min yielded an initiator-modiﬁed membrane. Polymerization of HEMA
occurred by circulating a solution containing 15 mL of puriﬁed HEMA, 15 mL methanol,
82.5 mg (0.825 mmol) of CuCl, 54 mg (0.24 mmol) of CuBrz, and 320 mg (2.04 mmol)
of bpy through the initiator-modiﬁed membrane using a peristaltic pump (7 mUmin) for
1 h. This HEMA solution was initially degassed using 3 freeze-pump-thaw cycles and
prepared as described in Chapter 2.16 After polymerization, the membrane was cleaned
by ﬂowing ethanol (20 mL), water (20 mL) and acetone (20 mL) through the membrane
at 7 mUmin.

To prepare ﬁlms for reﬂectance FT IR characterization, Au-coated silicon wafers
were modiﬁed with a mercaptoundecanol monolayer that was subsequently derivatized

with 2-bromoisobutyryl bromide as described previously.27‘ 28

Polymerization of HEMA
from these substrates occurred as described above, except the substrate was simply
immersed in a degassed polymerization solution that was kept in a glove bag.

3.2.3. PHEMA Derivatization and Protein Immobilization

To derivatize PHEMA-coated alumina membranes, a 55 °C DMF solution

containing SA (10 mg/mL) and DMAP (15 mg/mL) was passed through the membrane

85

for 3 h at 7 mL/min followed by rinsing with DMF (20 mL), water (20 mL) and ethanol
(20 mL) at 7 mUmin (Scheme 3.1). Next, a room-temperature solution containing
EDC/NHS (0.1 M of each in pure water) was circulated for 30 min (6.8 mL/min) through
the membrane, which was then brieﬂy washed with water and ethanol (20 mL each). The
EDC/NHS-activated membrane was allowed to react with a ﬂowing (6.5 mUmin) 0.1 M
aminobutyl-NTA solution (adjusted to pH 10.2) for 1 h, rinsed with 20 mL water, and
ﬁnally exposed to a 0.1 M CuSO4 solution (6.5 mUmin) for 30 min and rinsed with 20
mL of water. The membrane was then taken out of the Swinnex 25 cell, rinsed
sequentially with 5 mL of water, ethanol and acetone using a pipette, and dried with a
ﬂow of nitrogen. The water remaining in the pump tubing was next pumped out, and the
membrane was put back into the cell. A solution containing BSA or myoglobin (in pH
7.2 phosphate buffer) was then forced through the membrane, and the permeate was
collected for analysis at speciﬁc time intervals. Permeate volume was estimated using a
graduated cylinder. Subsequently, the membrane was rinsed with pH 7.2 phosphate
buffer, and protein was eluted using a 50 mM EDTA solution (pH adjusted to 7.2). To
prepare ﬁlms for reﬂectance FT IR characterization, a PHEMA ﬁlm on a gold substrate
was treated in a similar procedure by immersing the substrate in appropriate solutions and

rinsing with solutions from a pipette.

86

11 CIH3 9H3
o—c-ctcHz-crner

CH3 9:0
9
0:1 7:0 55 °C. 3 h
DMF
CH3
CH2"‘C\')‘an
9:0 0 0
0 II II

CHz-CHz-O—C-CHz-CHg-C-OH

 

 

OH
111 N=C=N
Ciro >114
EDC
NHS 11
H3
CH2—$)r_18r
9:0 o
O
9 ll (11
CH2_CH2—O'C—CH2"CH2—C—O_N
CHZCOO' O
N—CHZCOO'
H2N 000‘ 0
H102 "_ -
l p s 0
H2 (‘3‘
CHz—C-iﬁBr 9‘0
(i=0 0
9 9 ..
CHz-CH2_O_C_CH2_CH2’C—NH
1C02+ O\\ 40.-
- " OH
\ On- ‘. ,v' 2
O‘c’ CHiZCugf
1,. ‘. 0H2
CH3 CH2”N ‘\
l ﬁ—O'
CHz-(lliaBr o
C=O
Q
CH2’CH2‘O-C’CH2’CH2'C‘NH
_ HN . ......
Protein O\\ ,Q‘ Q \ 25"" .-
o o—.-- x“ 50190
\‘c’ CHixc'u’sE ““2“"
1'.’ I\ N \
$143 CH2'N . L
——o- it
HzCir-‘Br ﬁ
_ O

c—o
o
CH2'CH2‘O_C—CH2'CH2_C—NH

Scheme 3.1. Derivatization of PHEMA for protein immobilization.

87

3.2.4. Determination of Protein Concentrations

To determine the concentration of protein in permeate and eluent solutions, 50 uL
of the sample was added to 2.95 mL of a solution of Coomassie reagent, and the mixture
was shaken a few times and allowed to react for 5 min at room temperature. The UV/vis
absorbance spectra of these solutions were then obtained with a Perkin-Elmer UVN is
(model Lambda 40) spectrophotometer. Calibration curves for the absorbance of BSA
and myoglobin/Coomassie solutions at 595 nm were obtained using a series of protein
solutions (concentration range of 100 u g to 2 mg of protein per mL) that were mixed with
Coomassie reagent. All spectra were measured against a Coomassie reagent background.
3.2.5. Film Characterization Methods

Film growth inside alumina membranes was veriﬁed using ﬁeld-emission
scanning electron microscopy (FESEM, Hitachi S-47OOII, acceleration voltage of 15 kV)
and energy dispersive X-ray spectroscopy (EDS, Phoenix EDAX instrument). For SEM
and EDS observations, the alumina membrane was attached to a Si wafer by double-sided
tape, and the alumina was dissolved by immersing it into 1 M NaOH at 25 °C for two
hours followed by careful washing several times with distilled water and ethanol.
Subsequently, 7 nm of gold was sputtered onto the remaining polymer prior to FESEM
and EDS examination. Reﬂectance FTIR spectra of ﬁlms on gold-coated wafers were
obtained with a Nicolet Magna-IR 560 instrument using a Pike grazing angle (80°)
accessory. The spectrum of reﬂection from a UV/ozone-cleaned gold-coated wafer was

used as a background.

88

3.3. Results and Discussion

3.3.1. FTIR Characterization of PHEMA Derivatization and Protein Binding
Reﬂectance FT IR spectra of PHEMA ﬁlms and their derivatives on Au-coated Si

conﬁrm the steps of the derivatization and protein-adsorption procedure shown in

Scheme 3.1. To prepare brushes capable of protein binding, we ﬁrst reacted PHEMA

 

I 0-1 x2 9

Absorbance
F

 

WJ c

b

 

 

1850 1750 1650 1550 1450
Wavenumbers (cm")
Figure 3.1. Reﬂectance FTIR spectra of a PHEMA ﬁlm before (a) and after the
following sequential steps: (b) reaction with SA, immersion in a pH 9.9 buffer, and
rinsing with ethanol; (c) activation with EDC/NHS; ((1) reaction with aminobutyl
NT A followed by immersion in pH 9.9 buffer; (e) exposure to 0.1 M Cu2+ to form
N'I‘A-Cu2+ complexes; (f) exposure to 1 mg/mL BSA followed by immersion in pH
9.9 buffer. Spectrum (g) is the difference spectrum resulting from subtraction of (d)

from (f) and multiplication by a factor of 2.

89

with SA to create free —COOH groups in the ﬁlm. Prior to obtaining the reﬂectance IR
spectrum of SA~derivatized ﬁlms, we immersed the ﬁlm-coated gold slide in pH 9.9
buffer for 15 min and rinsed with ethanol. At pH 9.9, the newly introduced carboxylic
acid groups should be deprotonated,29 and consistent with derivatization and
deprotonation, the FT IR spectrum of these ﬁlms contains a new peak around 1594 cm’1

(Figure 3.1b), which is likely due to the —C00’ symmetric stretch. The absorbance at

 

Absorbance

r_,__/\ b

-OH stretch T

T
Na_

] T

 

 

 

 

 

3750 3250 2750 2250 1 750 1 250
Wavenumbers (cm")

Figure 3.2. Reﬂectance FTIR spectra of a PHEMA ﬁlm (3750 to 1250
cm") on Au (a) before and (b) after derivatization with succinic anhydride,

'1‘

immersion in pH 9.9 buffer, and rinsing in ethanol.

90

~1740 cm", which is assigned to ester carbonyl groups, approximately doubled after SA
derivatization, suggesting a high degree of conversion of the —OH groups of PHEMA to
ester groups. The disappearance of the —OH stretch of PHEMA (3650 - 3100 cm",
Figure 3.2) also indicates nearly 100% derivatization.

After activation of the —COOH groups of modiﬁed PHEMA using EDC/NHS,
peaks due to the succinimide ester appeared at 1817 and 1786 cm'1 (Figure 3.1c).30'32
The asymmetric stretch of succinimide (~1753 cm") overlaps with the C=O stretch (1740
cm") of the previously formed esters to yield a broad peak with an absorbance that is
about double that for the ester carbonyl after SA derivatization. The new peak at 1817
cm'1 is due to the carbonyl stretch of the active ester formed by reaction with NHS, while
the absorbance at 1786 cm'1 results from the symmetric succinimide stretch.

The EDC/NHS activated PHEMA was then reacted with aminobutyl—NTA
(Na,Na—bis(carboxymethyl)-L-lysine hydrate) and immersed in pH 9.9 buffer for 15 min
and rinsed with ethanol. Reaction of the active ester with both aminobutyl NTA and
water resulted in the loss of the active ester absorbances, and the shift of the 1753 cm'1
peak back to 1740 cm'1 (Figure 3.1d). The new peak at 1680 cm’1 provides evidence for
NT A immobilization and is probably due to the COO’ groups from NTA as well as the
amide bond formed between SA and NTA. The broad absorbance centered at 1600 cm’1
could result from carboxylate groups in either NTA or the hydrolyzed active esters.
Exposure of NTA-derivatized PHEMA to 0.1 M CuSO4 followed by rinsing with water
yielded immobilized NTA-Cu2+ that is capable of binding proteins through their histidine
groups. Unfortunately, there was no signiﬁcant change in the IR spectrum (Figure 3.1e)

of the ﬁlm after Cu2+ coordination except that the peak at 1600 cm'I appeared to shift to

91

1630 cm'l and become sharper. However, this shift could be due to the fact that spectrum
((1) was measured after treatment with pH 9.9 buffer, while the sample for spectrum (e)
was not treated with this buffer because Cu(OH); might precipitate in the ﬁlm at high pH.
XPS data conﬁrmed the presence of Cu2+ in the ﬁlm and showed a Cu : N ratio of 0.4 : l,
which is consistent with our previous results and indicative of nearly one Cu2+ per NTA
moiety.33

After an overnight exposure of the PHEMA-NTA-Cu2+ ﬁlm to a 1 mg/mL BSA
solution in pH 7.2 buffer followed by immersion in pH 9.9 buffer for 15 min and rinsing
in EtOH, we saw growth in the absorbance at 1680 cm"l (amide I band) and the
appearance of a small peak at 1545 cm'I (amide H band) (Figure 3.1f). Subtracting the
spectrum of PHEMA-NTA (Figure 3.1d) from that of PHEMA-NTA-Cu2+-BSA (Figure
3.10 yielded a bound-BSA spectrum, which is dominated by amide absorbances (Figure
3.1g). We chose to subtract the PHEMA-NTA rather than the PHEMA-NTA-Cu2+
spectrum because the former was taken after immersion in a pH 9.9 buffer as was the
spectrum of PI-IEMA-NTA-Cu2+-BSA. Comparison of the amide intensities in Figure
3.1g with those of spin—coated ﬁlms of pure BSA suggests that about 40 nm of BSA
adsorbed to this ﬁlm.33 Assuming a monolayer height of 4.0 nm,34 this thickness
corresponds to 10 monolayers of bound BSA. The PHEMA ﬁlm was initially 50 nm

thick.

92

3.3.2. SEM and EDS Characterization of PHEMA inside Alumina Membranes

 

Figure 3.3. FESEM images of PHEMA nanotubes synthesized by ATRP
in porous alumina substrates. (a) Top view of the nanotubes. (b) Cross-
sectional image showing the length of the nanotubes. (c) Highly ﬂexible
PHEMA nanotubes. (d) Top view of the nanotubes obtained by removing the
surface layer by polishing with sand paper. The arrow in (b) shows the

approximate length of the tubes.

In chapter 2, we used ATRP to graft PHEMA from the surface of porous alumina

'6 In contrast,

substrates and form composite membranes with ultrathin PHEMA skins.
this work aimed at grafting PHEMA within the pores of the membrane to achieve a high

surface area for protein capture. By pumping the initiator and monomer solutions

through the membrane, we were able to achieve polymerization throughout the support

93

and increase the surface area from which polymerization occurred by 500 fold. One
challenge in such polymerizations is to grow polymer brushes without clogging the
membrane. To overcome this difﬁculty, we used methanol instead of water as the solvent
for brush growth to decrease the rate of polymerization.35

The SEM images in Figure 3.3 clearly demonstrate the formation of hollow
PHEMA tubes in the membrane pores.” 37 To obtain these images, we dissolved the
alumina template in 1 M NaOH and collected the polymer tubes before taking SEM
images. (PHEMA likely does not dissolve in 1 M NaOH because it is lightly cross-
linked.38) The right side of Figure 3.3a shows an image of the top of the brushes, which
is similar to images of bare porous alumina. Figure 3.3b indicates that the length of the
resulting nanotubes is about 45 um, which is slightly less than the thickness of the
alumina support (60 um), perhaps because the tubes curled up. Figure 3.3c shows the
ﬂexibility of the tubes. The image in Figure 3.3d was obtained after removing the skin
layer of the membrane by polishing with sandpaper to reveal that the interior of the pores
is open. (Although Figure 3.3a shows that the pores are open at the surface of the
membrane, it does not necessarily imply that the interior of the pores is open.) Figure
3.3d shows open, interior pores, but the thickness of the polymers is much higher than
what we calculated from flow experiments (see below). Because the sanding process
may cause the pores to appear thicker than they are, we do not think that Figure 3.3d can

be employed to determine the inner diameter of the tubes.

94

The energy dispersive X-ray spectrum of the nanotubes (Figure 3.4) reveals C and

O and not Al, demonstrating that these images are indeed those of PHEMA.

 

 

1.00 2.00 3.00 0.00 3.00 t... 1.00

Figure 3.4. Energy dispersive X-ray spectrum of PHEMA nanotubes grown
by ATRP in porous alumina membranes. The nanotubes were obtained by
attachment of the membrane onto Si wafers and dissolution of the alumina in
1.0 M NaOH. These samples were sputtered with 7 nm of Au prior to
imaging. Peaks due to Si and Au are due to the support and sputtering,

respectively.

95

3.3.3. BSA Binding to PHEMA-NTA-Cu2+ Brushes
To test the binding capacity of PHEMA-NTA-Cu2+ ﬁlms, we initially pumped a

0.56 mg/mL BSA solution through the membrane, collected the permeate over speciﬁc

 

 

 

Absorbance (a.u.)
5': .6

 

 

 

-O.3 . u u u

400 500 600 700 800
Wavelength (nm)

Figure 3.5. UV-vis spectra of the permeate collected at certain time
intervals when 0.56 mg/mL BSA solution was pumped through porous
alumina membranes coated with PHEMA-NTA-Cu2+. The permeate flow
rate was initially 2.4 mL/min and 0.9 mUmin at the end of the experiment.
Prior to measurement of the UV-vis spectra, 50 uL of permeate was mixed
with 2.95 mL of Coomassie reagent. Pure Coomassie reagent was used as the

background.

96

time intervals, and analyzed these solutions using the absorbance at 595 nm in a Bradford
assay. Figure 3.5 presents typical UV-vis spectra from the Bradford assays.

The breakthrough curve (Figure 3.6) obtained from these measurements shows
that essentially all of the BSA was adsorbed until breakthrough occurred, and that
saturation of the membrane took less than 15 min.” 39’ 40 Integration of the difference

between the feed concentration and the concentrations given in Figure 3.6 yielded a

.0
o?

    
  
  

.0
0.1

p
f:

.0
"3

BSA Concentration (mg/mL)
_o o
—s 0)

 

 

I I U I I

O 2 4 6 8 1O 12
Volume (mL)

.0
0

Figure 3.6. Breakthrough curve for BSA adsorption in porous alumina
membranes coated with PHEMA-NTA-Cu2+. The BSA concentration in the
feed solution was 0.56 mg/mL and the initial ﬂow rate was 2.4 mUmin,

decreasing to 0.9 mUmin at the end of the experiment.

remarkable 0.8 mg of bound BSA per cm2 of external membrane surface area. Repetition

of this experiment on 4 different membranes yielded an average BSA binding of 0.82 t

97

0.07 mg/cmz. This very high capacity is due in part to the high internal surface area of
the alumina, but even taking into account the surface area of the pores (assuming 50%
porosity, pore diameters of 0.25 pm, and a membrane thickness of 60 um), the capacity is
1.6 ug of BSA per cm2 of pore area or about 4 monolayers of BSA (assuming a BSA
density of 1 g/cm3 and a monolayer thickness of 4.0 nm) adsorbed throughout the pores
of these membranes.

Figure 3.7 shows how the amount of equilibrium binding of BSA to PHEMA-

_L
O
I

.0
m
l

.0
o:
I

.0
A
n

 

Binding Amount (mg/cmz)
O
ix:

0'0 r I I I I I
0.0 0.8 1.6 2.4 3.2 4.0

BSA Concentration (mg/mL)

 

Figure 3.7. Equilibrium binding capacity of PHEMA-NTA-Cu2+-coated
alumina membranes as a function of BSA concentration. The solid line

represents the ﬁt of the data to the Langmuir isotherm.

NTA—Cu2+ ﬁlms in alumina membranes varies with the concentration of BSA. In this

experiment, binding was performed at a speciﬁc BSA concentration, and the membrane

98

was regenerated with 0.05 M EDTA (pH 7.2) followed by 0.1 M Cu2+ prior to examining
adsorption at another BSA concentration. As can be seen, BSA binding increases with its
concentration in the feed solution until approaching saturation at concentrations around 1
mg/mL. The data are reasonably described by the Langmuir isotherm,9“““i3 q = qu/(Kd
+ C), where q is the amount of protein bound, qm represents the saturation capacity, C is
the concentration of protein in solution, and Kd is the dissociation equilibrium constant.
The value of qm determined from the Langmuir plot is 0.9 mg per cm2 of external
membrane surface area, and the dissociation constant is 0.09 mg/mL, which also
corresponds to 1.3 uM. This Kd value is reasonably consistent with the previously
reported value of 2.3 uM for Chitosan-Cu2+-BSA,44 and indicates a strong binding
afﬁnity between BSA and immobilized NTA-Cu2+ complexes.

The high protein-binding capacity of these membranes is also consistent with
decreases in ﬂow rate after protein binding. As noted in Figure 3.6, flow rate decreased
from 2.4 to 0.9 mUmin during binding of 0.8 mg of BSA per cm2 of external membrane
surface area (130 mg/cm3). We achieved a similar decrease in ﬂow rate when we loaded
protein in the membrane using a dead-end ﬁltration cell (Millipore, Model 8010) with a
controlled pressure of 6.9 x 104 pascals (10 psig). Assuming Poiseuille ﬂow in the
membranes (flow rate is proportional to pore radius raised to the 4th power), the drop in
ﬂow rate suggests that the radius of the membrane pores decreased by a factor of 1.28
after protein binding. Using an open pore radius of 0.112 um after PHEMA-NTA-Cu2+
deposition (see below) and assuming a 50% porosity prior to ﬁlm deposition, the
decrease in radius by a factor of 1.28 upon protein binding would yield a decrease in pore

volume of 0.16 cm3/cm3 of total membrane. With a protein density of 1 g/cm3, this

99

decrease in pore volume agrees well with both the BSA binding capacity of 130 mg/cm3
for the membrane used to obtain Figure 3.6 and the saturated binding capacity of 150
mg/cm3 (Figure 3.7). Thus, the high protein-binding capacity is consistent with the flow
rates through the membrane.

Flow rates before polymerization and after deposition of PHEMA-NTA-Cu2+
were 10 mljmin and 6.5 mUmin, respectively, suggesting that the pore radius decreased
from 0.125 mm before polymerization (radius estimated from FESEM images) to 0.112
um after deposition of the PHEMA-NTA-Cu2+ ﬁlm. This implies a ﬁlm thickness of 10
nm, which appears to be less than that in the SEM images of PHEMA tubes in Figure 3.3,
but as mentioned previously, the polymers might be much thicker at their sanded surface
than in the bulk of the membrane. We should note that ﬂow rate dropped from 6.5
mUmin to 2.5 mUmin almost instantaneously when we began ﬁltering the BSA solution.
We think that a thin layer of contaminants may quickly adsorb at the membrane surface.
Elution with EDTA and/or regeneration with 0.1 M Cu2+ did not return the flow rate to

6.5 mL/min.

100

Table 3.1. Comparison of the protein-binding capacity of alumina-PHEMA-

NTA-Cu2+ membranes with other membrane absorbers.

 

 

Th’ kn ‘ Surface Pore Binding Binding
Membrane ('C m )6 58 Area Size Porosity Li gate capacity capacity
ll (cm?) (um) (mg/cmz) (mg/cm3)

Glass ﬁber20 8000 17.3 3 87% BSA 6.9 8.6

Poly (acrylic acid)
(Polypropylene 150 3.14 0.4 ----- Lysozy me 0.3 20
base polymer)”

Cellulose” 64 125 0.2 ----- BSA 0.027 4.2

PHEMA (This work)

. 60 3.14 0.2 50% BSA 0.9 150
(Alumina support)

Table 3.1 compares the BSA-binding capacity of PHEMA-NTA-Cu2+-coated
alumina membranes with several other membrane systems. Ruckenstein and Guo
modiﬁed glass ﬁber ﬁlters with trypsin or papain to bind proteins.” 45 They then packed
20 modiﬁed glass membrane ﬁlters (about 8 mm thick) into a cartridge and achieved an
extremely high binding capacity of 6.9 mg of BSA per cm2 of external membrane surface
area. However, if one takes the membrane thickness into account and converts binding
capacity into mg/cm3, the modiﬁed alumina membranes described here show a l7-fold
higher capacity (compare rows 1 and 4 in Table 3.1). Very recently, Ulbricht and Yang
used UV-initiated graft polymerization to grow poly (acrylic acid) from porous
polypropylene microﬁltration membranes.19 The resulting membranes showed a
promising performance in adsorption of lysozyme using ion-exchange interactions (Table
3.1, row 2). Still the capacity of those membranes was 7-fold lower than that described
here. Kubota and coworkers modiﬁed porous cellulose membranes with iminodiacetate-

Cu2+ complexes,” but those membranes (Table 3.1, row 3) gave only 1/36 of the binding

101

capacity described here . The growth of brushes from alumina using ATRP provides an
unusually productive method for introducing a high concentration of binding groups into
membranes.
3.3.4. Elution of BSA and Reuse of Alumina-PHEMA-NTA-Cu2+ Membranes

In addition to rapid, high-capacity adsorption, elution and regeneration are also

48

vital in the use of membrane absorbers.“ Figure 3.8 shows results from multiple

 

 

a? 1.2 . 1000/0
I
E W >
E» 90% 3
v1.0 '3
g ~80°/o E
g c
<1: 0 8 A A ‘ ‘4 70% 2%
L—*—————*'

.E’ " i e m
'2 ”600/0
5

0.6 50%

 

 

 

1 2 3 4 5 6 7 8 9
Number of Cycles
Figure 3.8. Binding amounts (triangles) and elution efﬁciencies (squares) for
several cycles of BSA adsorption and elution followed by Cu2+ reloading of an

alumina membrane modiﬁed with PHEMA-NT A. (Flow rate - 0.9 mL/min;

BSA concentration - 0.6 mg/mL; Elution buffer — 50 mM EDTA, pH 7.2)

experiments comprising cycles of absorption of BSA, elution with 50 mM EDTA, and
regeneration with 0.1 M Cu2+ on a single alumina-PHEMA-NTA-Cu2+ membrane.
Within experimental error (<15%), there was no loss in binding amount over 9 cycles,

and >93% recovery of protein occurred in each cycle. The consistency in capacity over 9

102

loadings implies that the elution buffer removed essentially all of the adsorbed protein
from the membranes, even though the average calculated elution efﬁciency was only
94%. There likely was a systematic error in the analyses that resulted in an
underestimation of elution efﬁciencies. Given the small volumes involved in the
breakthrough experiments along with inaccuracies in the volume measurements and the
need to integrate data such as those in Figure 3.5 to obtain adsorption amounts, this is
certainly possible.

To more accurately assess elution efﬁciency, we passed 10 mL of 0.56 mg/mL
BSA in pH 7.2 buffer through the membrane and analyzed this entire solution to
determine how much BSA was adsorbed in the membrane. We also passed 5-10 mL of
rinsing buffer through the membrane to rinse the system and determined the amount of
protein in this efﬂuent. We subtracted the total amount of protein that came through the
membrane from the amount of protein added to the initial 10 mL BSA solution to
determine the amount of bound BSA. (In these experiments we used weight rather than
volume to more accurately determine the amount of solution passing through the
membrane.) We eluted the protein with 10-15 mL of 0.1 M EDTA and analyzed the
protein concentration in the entire solution. For three cycles of adsorption, elution, and
Cu2+ binding, the average amount bound to the membrane was 0.781002 mg/cmz, which
is in reasonable agreement with the previous measurements. The elution efﬁciency was
101 i 1 %. Hence, essentially all of the protein is eluted.

Control experiments with membranes that were derivatized with PHEMA-NTA
ﬁlms but not exposed to Cu2+ showed elution of <0.02 mg/cm2 BSA from a membrane

loaded by a 0.56 mg/mL BSA solution and rinsed with buffer. These control experiments

103

help to validate the experimental procedure and also indicate that the PHEMA-NTA ﬁlms

are not prone to high levels of non-speciﬁc BSA adsorption under these conditions.

3.3.5. Adsorption of Myoglobin

We also examined the adsorption of myoglobin to alumina-PHEMA-NTA-Cu2+
membranes to see if molecular weight affects protein adsorption. BSA has a molecular
weight of 67 kDa, an isoelectric point of 4.9, 17 histidine residues, and dimensions of 4.0
nm x 4.0 nm x 11.5 nm.34 In contrast, myoglobin has a molecular weight of 17 kDa, an
9

isoelectric point of 7.0, 11 histidine residues and a size of 4.4 nm x 4.4 nm x 2.5 nm.4

Figure 3.9 shows the breakthrough curves for the two proteins on the same membrane.

.0
h
I

+ Myoglobin
1 -0- BSA

    

.0
co
1

.0
m

.0
...-L

 

Protein Concentration (mg/mL)

 

   

.0
o

I I I I

0 2 4 6 10 12 14 16
Volume(mL)

Figure 3.9. Breakthrough curves for adsorption of 0.3 mg/mL BSA and 0.35
mg/mL myoglobin on the same alumina-PHEMA-NTA-Cu2+ membrane. After
adsorption of BSA, the membrane was regenerated with 0.05 M EDTA and 0.1

M Cu“.

104

The membrane showed high binding capacities for both BSA (0.67 mg/cmz, feed
concentration: 0.30 mg/mL) and myoglobin (0.50 mg/cmz, feed concentration: 0.35
mg/mL), but the breakthrough curve is sharper for myoglobin. This could be due to more
rapid diffusion of the smaller myoglobin in the brushes, but this needs further
investigation. Future studies examining how the rate of binding varies with brush
thickness and density should be helpful in this regard. (Brush thickness can be controlled
by varying the polymerization time, while density can be controlled by changing the

38. 50

denisyt of initiators. ) The fraction of histidine residues in myoglobin is also higher,

and this could result in more rapid binding.“

3.4. Conclusions

Immobilized metal-afﬁnity membranes were prepared by ATRP of HEMA inside
the pores of alumina substrates followed by a series of derivatization steps to immobilize
NTA-Cu” complexes. FESEM measurements conﬁrmed the growth of PHEMA brushes
inside the pores of alumina membranes, and reﬂectance FTIR spectroscopy veriﬁed the
efﬁcacy of the derivatization procedures. Equilibrium binding essentially followed the
Langmuir isotherm, and the saturation binding capacity of PHEMA-NTA—Cu2+ for BSA
was 0.9 mg per cm2 of membrane. Moreover, saturation of the membrane with protein
occurred in less than 15 min. After 9 cycles of adsorption, elution, and regeneration, the
membranes showed no detectable loss of capacity, and elution efﬁciencies were
essentially 100%. The derivatized PHEMA brushes also showed rapid, high capacity
binding of myoglobin. Hence, the use of polymer brushes in the pores of membranes is

an effective means for creating high-capacity protein adsorbers.

105

3.5.

10.

ll.

12.

13.

14.

15.

16.

17.

18.

19.

References

Arica, M. Y.; Testereci, H. N .; Denizli, A., J. Chromatogr. A 1998, 799, 83-91.
Kawai, T.; Saito, K.; Lee, W., J. Chromatogr. B 2003, 790, 131-142.

Zou, H.; Luo, Q.; Zhou, D., J. Biochem. Biophys. Methods 2001, 49, 199-240.

Clairbois, A. S.; Letoumeur, D.; Muller, D.; Jozefonvicz, J., J. Chromatogr. B
1998, 706, 55-62.

Sun, G.-Y.; Shi, Q.-H.; Sun, Y., J. Chromatogr. A 2004, 1061, 159-165.

Henis, J. M. S.; Tripodi, M. K.; Stimpson, D. I. European Patent 0,221,046B1,
October 30, 1987.

Charcosset, C., J. Chem. Technol. Biotechnol. 1998, 71 , 95-110.

Urmenyi, A. M.; Poot, A. A.; Wessling, M.; Mulder, M. H. V., J. Membr. Sci.
2005, 259, 91-102.

Wu, C.-Y.; Suen, S.-Y.; Chen, S.—C.; Tzeng, J.-H., J. Chromatogr. A 2003, 996,
53-70.

Roper, D. K.; Lightfoot, E. N ., J. Chromatogr. A 1995, 702, 3-26.
Sarfert, F. T.; Etzel, M. R., J. Chromatogr. A 1997, 764, 3-20.
Ruckenstein, E.; Zeng, X., Biotechnol. Bioeng. 1997, 56, 610-617.
Zeng, X.; Ruckenstein, E., J. Membr. Sci. 1999, 156, 97—107.
Zeng, X.; Ruckenstein, E., J. Membr. Sci. 1996, 117, 271-278.

Castilho, L. R.; Deckwer, W. D.; Anspach, F. B., J. Membr. Sci. 2000, 172, 269-
277.

Sun, L.; Baker, G. L.; Bruening, M. L., Macromolecules 2005, 38, 2307-2314.

Miller, M. D.; Baker, G. L.; Bruening, M. L., J. Chromatogr. A 2004, 1044, 323-
330.

Bayramoglu, G.; Yilmaz, M.; Arica, M. Y., Colloids Surf, A 2004, 243, 11-21.

Ulbricht, M.; Yang, H., Chem. Mater. 2005, I 7, 2622-2631.

106

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

Guo, W.; Ruckenstein, E., J. Membr. Sci. 2003, 215, 141-155.

Kubota, N.; Nakagawa, Y.; Eguchi, Y., J. Appl. Polym. Sci. 1996, 62, 1153-1160.
Kusumo, A.; Bombalski, L.; Qiao, L.; Kowalewski, T.; Matyjaszewski, K.;
Schneider, J. W.; Tilton, R. D., AIChe Annual Meeting, Cincinnati, OH, 2005,
01 C14, 123d.

Ito, H.; Nakamura, M.; Saito, K.; Sugita, K.; Sugo, T., J. Chromatogr. A 2001,
925, 41-47.

Okamura, D.; Sait, K.; Sugita, K.; Tamada, M.; Sugo, T., J. Chromatogr. A 2002,
953,101-109.

Singh, N.; Husson, S. M.; Zdyrko, B.; Luzinov, I., J. Membr. Sci. 2005, 262, 81-
90.

Matyjaszewski, K.; Miller, P. J .; Shukla, N.; Irnmarapom, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T., Macromolecules 1999, 32, 8716-8724.

Kim, J.-B.; Bruening, M. L.; Baker, G. L., J. Am. Chem. Soc. 2000, 122, 7616-
7617.

Kim, J.-B.; Huang, W.; Bruening, M. L.; Baker, G. L., Macromolecules 2002, 35,
5410-5416.

Hong, Y. K.; Hong, W. H., Sep. Purif. Technol. 2005, 42, 151-157.

Dordi, B.; Schoenherr, H.; Vancso, G. J ., Langmuir 2003, 19, 5780-5786.

Lahiri, J.; Isaacs, L.; Tien, J .; Whitesides, G. M., Anal. Chem. 1999, 71, 777-790.
Xiao, S.-J.; Brunner, S.; Wieland, M., J. Phys. Chem. B 2004, 108, 16508-16517.

Dai, J .; Bao, 2.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L., Langmuir
2006, 22, 4274-4281.

Tsuneda, S.; Saito, K.; Furusaki, S.; Sugo, T., J. Chromatogr. A 1995, 689, 211-
218.

Robinson, K. L.; Khan, M. A.; de Banez, M. V.; Wang, X. S.; Armes, S. P.,
Macromolecules 2001, 34, 3155-3158.

Hou, S.; Wang, J .; Martin, C. R., J. Am. Chem. Soc. 2005, 127, 8586-8587.

107

37.

39.

40.

41.

42.

43.

45.

46.

47.

48.

49.

50.

Martin, C. R., Science 1994, 266, 1961-1966.

Huang, W.; Kim, J .-B.; Bruening, M. L.; Baker, G. L., Macromolecules 2002, 35,
1175-1179.

Shi, W.; Zhang, F.; Zhang, G., J. Chromatogr. A 2005, 1081, 156-162.
Hao, W.; Wang, J ., Chromatographia 2005, 62, 55-62.
Iwats, H.; Saito, K.; Furusaki, S., Biotechnol. Progr. 1991, 7, 412-418.

Arica, M. Y.; Denizli, A.; Salih, B.; Piskin, E.; Hasirci, V., J. Membr. Sci. 1997,
129, 65-76.

Bayramoglu, G., J. Appl. Polym. Sci. 2003, 88, 1843-1853.

Shi, Q.-H.; Tian, Y.; Dong, X.-Y.; Bai, S.; Sun, Y., Biochem. Eng. J 2003, 16,
317-322.

Guo, W.; Ruckenstein, E., J. Chromatogr. B 2003, 795, 61-72.

Avrarnescu, M.—E.; Girones, M.; Bomeman, Z.; Wessling, M., J. Membr. Sci.
2003, 218, 219-233.

Garipcan, B.; Andac, M.; Uzun, L.; Denizli, A., React. F unct. Polym. 2004, 59,
119-128.

Kubota, N.; Kounosu, M.; Saito, K.; Sugita, K.; Watanabe, K.; Sugo, T., J.
Membr. Sci. 1997, 134, 67-73.

Kent, M. S.; Yim, H.; Sasaki, D. Y.; Satija, S.; Seo, Y.—S.; Majewski, J.,
Langmuir 2005, 21 , 6815-6824.

Bao, Z.; Bruening, M. L.; Baker, G. L., Macromolecules 2006, 39, 5251-5258.

108

Chapter 4
Puriﬁcation of Histidinex-Tagged Proteins Using Membranes

Modiﬁed with Polymer Brushes

4.1. Introduction

Advances in genetic engineering have generated an ongoing need for efﬁcient
puriﬁcation of recombinant proteins.l One of the most powerful methods for isolating
such proteins, afﬁnity chromatography, is based on speciﬁc interactions between
immobilized ligands and an afﬁnity tag (e.g., glutathione-S-transferrase (GST)2,
Streptavidin,3 or polyhistidine4) that is appended to the protein of interest. Polyhistidine,
the most frequently employed tag, allows puriﬁcation by immobilized metal-afﬁnity
chromatography (IMAC),5'8 where selectivity is usually based on the interaction of the
polyhistidine with an immobilized Ni2+ complex. Proteins containing a polyhistidine tag
(his tag) are selectively bound to the chromatographic matrix, while other cellular
proteins are washed away.9

The advantages of IMAC, which include ligand stability, high protein loading,
mild elution conditions, simple regeneration and low cost,'0 are very important when
developing puriﬁcation procedures. Nevertheless, this technique has some serious
limitations, including long separation times, difﬁculties in packing large columns,
relatively high pressure drops, and slow intra-bead diffusion of solutes.”13 These

limitations will be particularly important for large scale separations.

109

 

In an effort to solve some of these problems that are general to afﬁnity
chromatography, membrane chromatography was ﬁrst introduced in 1987.'4 In this
technique ﬂow of solution through membrane pores containing immobilized ligands
enhances the rate of mass transport to the ligands, and scale up is in principle relatively
simple.9 Despite their potential, however, the major disadvantage of membrane
absorbers is that their low speciﬁc surface area (when compared to porous beads) yields a
relatively low binding capacity. To overcome this problem, we and others are developing
membranes modiﬁed with polymer brushes that have multiple protein—binding sites.”18
The controlled growth of polymers in membrane pores using ATRP is particularly
attractive because the thickness of the resulting polymer brushes can be readily controlled
to optimize capacity without completely ﬁlling pores. Our previous study (Chapter 3 of
this thesis) showed that growth of PHEMA from initiators bound to a porous alumina
membrane followed by functionalization of the PHEMA with nitrilotriacetate-Cu2+
(NTA-Cu2+) complexes resulted in membranes that bind large amounts of BSA and
myoglobin, presumably via formation of a complex between the histidine groups of the
proteins and the immobilized NTA-Cu“. Overall, the combination of a porous alumina
support and functionalized polymer brushes yielded a remarkable binding capacity of 0.9
mg of BSA per cm2 of external membrane surface (150 mg/cm3 of membrane).

However, the NTA-Cu2+ complex, as suggested from the high BSA-binding
capacity of membranes containing PHEMA-NTA-Cu2+ brushes, is prone to binding of
many proteins and is not sufﬁciently selective to effectively purify his tagged proteins.

This chapter describes the use of PHEMA-NTA—Ni2+, rather than PHEMA-NTA-Cuzi',

brushes to modify membranes and selectively purify his tagged proteins. It is well

110

known that NTA-Ni2+ is highly selective for binding of his tagged proteins because the

 

 

Surface

 

 

 

 

Figure 4.1. Selective binding of PHEMA-NTA-Ni2+ to a his tagged

protein. (a: his tagged protein; b-d: non his tagged proteins).

interaction of NTA-Ni2+ with histidine is much weaker than the interaction of NTA-Cu2+
with histidine (Figure 4.1).4’ '9 Thus, polyhistidine is required for efﬁcient binding to the
NTA-Ni2+ complex. This work demonstrates that membranes modiﬁed with PHEMA-
NTA-Ni2+ selectively bind his tagged ubiquitin (HisU) from solutions containing equal
amount of HisU, myoglobin, and BSA or a 20—fold excess of BSA relative to HisU. Gel
electrophoresis revealed that the purity of HisU recovered from such solutions is >99%,
and that the binding capacity of these membranes is as high as 120 mg HisU/cm3 of

membrane (for a 0.3 mg/mL HisU solution).

111

4.2. Experimental Section
4.2.]. Materials

AnodiscTM porous alumina membranes with 0.2 um-diameter surface pores were
obtained from Fisher Scientiﬁc. Dimethylformamide (DMF, anhydrous, 99.8%), 11-
mercaptoundecanol (97%), 2-bromoisobutyryl bromide (98%), CuCl (99.999%), CuBrz
(99%), 2,2’-bipyridy1 (bpy, 99%), 1-[3-(dimethy1amino)propyl]-3-ethylcarbodiimide
hydrochloride (EDC), N—hydroxysuccinimide (NHS), 4-dimethylamin0pyridine (DMAP),
ethylenediamine tetraacetic acid (EDTA), imidazole (99%), TWEEN-20 surfactant,
bovine serum albumin (BSA), ubiquitin (N-terminal histidine6 tagged) and myoglobin
were used as received from Sigma Aldrich. NiSO4-5HZO (Columbus Chemical),
NaHzPO4 (CCI), NazHPO4 (Aldrich), Na,Nar—bis(carboxymethyl)-L-lysine hydrate
(Fluka, aminobutyl-NTA), succinic anhydride (SA, Matheson Coleman & Bell), and
Coomassie protein assay reagent (Pierce) were also used as received. 2—Hydroxyethyl
methacrylate (HEMA, Aldrich, 97%, inhibited with 300 ppm hydroquinone monomethyl
ether) was puriﬁed by passing it through a column of activated basic alumina (Aldrich),
and trichlorosilane initiator (11-(2-bromo-2-methyl)propionyloxy)undecy1trichlorosilane)
was synthesized according to a literature procedure.20 Buffers were prepared using
analytical grade chemicals and deionized (Milli-Q, 18.2 M!) cm) water.
4.2.2. Polymerization of HEMA in Porous Alumina Membranes and on Au
Substrates

The procedure for polymerizing HEMA inside alumina membranes was reported
in the last chapter. Brieﬂy, the alumina membrane was sandwiched inside a membrane

cell (Millipore, Swinnex 25), and the trichlorosilane initiator solution was ﬁrst passed

112

.. ‘3”3 ?”3
O-C-Cr‘tCHz-CrnBr

CH3 9:0
9
0:1 7:0 55 °C.3h
DMF
CH3
CHg—Ci-nBr
83:0
o o o

. ll ll
CH2"CH2‘O"C‘CH2'CH2_C'OH

 

 

OH
'1‘ NZCZN
020:0 >fo
EDC
NHS ll
“3
CHz-Ci-r-‘Br
=0
3 ‘i ‘3 °
CHZ-CH2_O—C_CH2-CHz—C—O—N
CHZCOO' o
N—CH2COO'
H2N COO. 0
pH 10.2 ”_ .
v 9 O
i“? ii
9H3 N—Eg-g-C—O.
CHz—q—WI' 9
qzo O
O O

I II II
CHz—CHQ-O—C_CH2—CH2‘C—NH

Ni2+ O\\ ,0."
- \. OH
°V°¢Hé9~itoé
CH3 GHQ-N" 2
c—o-
CHz-giﬁgr ('5
b ‘3 o

CHZ-CHz—o—c-CHz—CHz—c—NH

Scheme 4.1. Derivatization of PHEMA with NTA-Ni2+ before protein

immobilization.

113

through the membrane followed by subsequent rinsing. Polymerization of HEMA
occurred by circulating a degassed solution containing 15 mL of puriﬁed HEMA, 15 mL
methanol, 82.5 mg (0.825 mmol) of CuCl, 54 mg (0.24 mmol) of CuBrz, and 320 mg
(2.04 mmol) of bpy through the initiator-modiﬁed membrane for 1 hour. After
polymerization, the membrane was cleaned with ﬂowing ethanol (20 mL), water (20 mL)
and acetone (20 mL).

To prepare ﬁlms for reﬂectance FTIR characterization, Au-coated silicon wafers
were coated with a mercaptoundecanol monolayer that was subsequently derivatized with
2-bromoisobutyryl bromide as described previously?" 22 Polymerization of HEMA from
these substrates occurred as described above, except that the substrate was simply
immersed in a polymerization solution that was kept in a glove bag.

4.2.3. PHEMA Derivatization and Protein Immobilization

The derivatization procedure was also described in the last chapter and is shown
in Scheme 4.1. However, in this case the NTA-derivatized membrane was exposed to a
0.1 M NiSO4, rather than 0.1 M CuSO4, solution and rinsed with solvents. A solution
containing pure protein or a mixture of proteins (in 20 mM phosphate buffer, pH 7.2) was
then pumped through the membrane using a peristaltic pump, and the permeate was
collected for analysis at speciﬁc time intervals. Subsequently, the membrane was rinsed
with 20 mL pH 7.2 washing buffer (20 mM phosphate buffer containing 0.1% Tween-20
surfactant and 0.15 M NaCl) and 20 mL phosphate buffer, and protein was eluted using
5-10 mL elution buffer (20 mM sodium phosphate, 0.5 M NaCl, 0.5 M imidazole, pH
7.4). Ni2+ was later eluted using a 50 mM EDTA solution (pH adjusted to 7.2), and the

PHEMA-NTA ﬁlm was recharged with Ni2+ prior to reuse. To prepare derivatized ﬁlms

114

for reﬂectance FF IR characterization, a PHEMA ﬁlm on a gold substrate was treated in a
similar procedure by immersing the substrate in appropriate solutions and rinsing with
solutions from a pipette.
4.2.4. Determination of the Amount of Coordinated Ni2+ in the Membrane

A calibration curve was obtained by measuring the UV-vis spectra of N i2+-EDTA
standard solutions (5, 10, 20, 40 mM NiSOa in 50 mM EDTA at pH 7.2), and a sample
solution was prepared by using 5.0 mL of 50 mM EDTA (pH 7.2) to remove Ni2+ from a
PHEMA-NTA-Ni2+-coated membrane. The UV-vis spectrum of the stripping solution
was acquired, and the amount of Ni2+ in the solution was calculated using the calibration
curve.
4.2.5. Determination of Protein Concentrations

To determine the concentration of protein in permeate and eluent solutions, 50 uL
of the sample was added to 2.95 mL of a solution of Coomassie reagent, and the mixture
was shaken a few times and allowed to react for 5 min at room temperature. The UV/vis
absorbance spectra of these solutions were then obtained with a Perkin-Elmer UVN is
(model Lambda 40) spectrophotometer. Calibration curves for the absorbance of BSA,
HisU and myoglobin solutions at 595 nm were obtained using a series of protein
solutions (concentration range of 100 u g to 1 mg of protein per mL) that were mixed with
Coomassie reagent. All spectra were measured against a Coomassie reagent background.
4.2.6. Determination of Protein Purity by SDS-PAGE

The protein soutions were analyzed by SDS-PAGE with a 15% cross-linked
separating gel and a 4% cross-linked stacking gel (acrylamide). Protein bands were

. . . . . . 13
Visualized usmg a standard Silver staining procedure.

115

4.2.7. Film Characterization Methods

Reﬂectance FTIR spectra of ﬁlms on gold-coated wafers were obtained with a
Nicolet Magna—IR 560 instrument using a Pike grazing angle (80°) accessory. The
spectrum of reﬂection from a UV/ozone-cleaned gold-coated wafer was used as a

background. Film growth inside alumina membranes was veriﬁed using Transmission

FTIR (Mattson Galaxy Series 3000) with an air background.

4.3. Results and Discussion
4.3.1. Characterization of PHEMA Derivatization

In the last chapter, we examined PHEMA formation inside the pores of alumina
membrane using SEM and EDS, and veriﬁed derivatization of PHEMA only for ﬁlms on
Au—coated Si. Here I present transmission FTIR spectroscopy that demonstrates both the
formation and derivatization of PHEMA within alumina membranes. Figure 4.2 Shows
the transmission FT IR spectra of PHEMA ﬁlms and their derivatives inside the alumina
membrane. To ensure that derivatization happens throughout the pores, we passed the
reactant solutions through the membrane using a peristaltic pump. After each
derivatization step, the membrane was taken out of the membrane cell, immersed in
appropriate buffers, rinsed with ethanol and acetone and dried under a ﬂow of nitrogen.
We then placed the membrane in a holder for taking IR spectra, and a spectrum of air was
used as a background. The transmission IR data are consistent with the spectra of ﬁlms

on gold surface and prove successful derivatization inside the alumina.

116

 

0.1

Absorbance

 

 

 

 

 

 

 

 

1 850 1 750 1 650 1 550 1450
Wavenumbers (cm")

Figure 4.2. Transmission 1R of (a) PHEMA ﬁlms inside an alumina
membrane before (a) and after the following sequential steps: (b) reaction
with SA, immersion in a pH 9.9 buffer. (c) activation with EDC/NHS; ((1)
reaction with aminobutyl NTA followed by immersion in a pH 9.9 buffer;

(e) exposure to 0.1 M Ni2+ to form NTA-Ni2+ complexes.

To be speciﬁc, we ﬁrst passed a solution of SA in DMF through a PHEMA-
derivatized membrane to create free —COOH groups in the ﬁlm. We then immersed the
membrane in pH 9.9 buffer for 15 min and rinsed with ethanol and acetone before taking

a transmission IR spectrum. The absorbance at ~1740 cm", which is assigned to ester

117

carbonyl groups, approximately doubled after SA derivatization (Figure 4.2b), suggesting
a high degree of conversion of the —OH groups of PHEMA to ester groups. A new peak
also appeared at ~1594 cm", which is consistent with the deprotonation of the newly
introduced carboxylic acid groups.

Passing a 0.1 M mixture of EDC and NHS in water (pH~4.9) through the
membrane converted —COOH groups to succinimidyl esters. The membrane was then
rinsed with water, ethanol and acetone. Peaks due to the succinimide ester appeared at
1817 and 1786 cm'1 G=igure 4.2c). The asymmetric stretch of succinimide (~1753 cm")
overlaps with the C=O stretch (1740 cm") of the previously formed esters to yield a
broad peak with an absorbance that is about double that for the ester carbonyl after SA
derivatization. The new peak at 1817 cm'1 is due to the carbonyl stretch of the active ester
formed by reaction with NHS, while the absorbance at 1786 cm’1 results from the
symmetric succinimide stretch.

The EDC/NHS-activated PHEMA was then reacted with aminobutyl-NTA
(NaNa—bis(carboxymethyl)-L-lysine hydrate) and immersed in pH 9.9 buffer for 15 min
and rinsed with ethanol and acetone. Reaction of the active ester with both aminobutyl
NTA and water resulted in the loss of the active ester absorbances, and the shift of the
1753 cm" peak back to 1740 cm1 (Figure 4.2d). The new peak at 1680 cm1 provides
evidence for NTA immobilization and is probably due to the C00 groups from NT A as
well as the amide bond formed between SA and NTA. The broad absorbance centered at
1600 cm'I could result from carboxylate groups in either NTA or the hydrolyzed active

CSICI‘ S.

118

Exposure of NTA-derivatized PHEMA to 0.1 M NiSO4 followed by rinsing with
water yielded immobilized NTA-Ni2+ that is capable of binding proteins through their
histidine groups. Unfortunately, there was no large change in the IR spectrum (Figure
4.2e) of the ﬁlm after Ni2+ coordination except that the peak at 1600 cm'1 appeared to
shift to 1630 cm"1 and become sharper. The Shift could be due to the fact that spectrum
(d) was measured after treatment with pH 9.9 buffer, while the sample for spectrum (e)
was not treated with this buffer because N i(OH)2 might precipitate in the ﬁlm at high pH.
The similarity is probably due to the high spectral Similarity of the Na+ and Ni“ salts of
NTA. XPS data for a PHEMA-NTA-Ni2+ ﬁlm on a gold substrate (Figure 4.3) conﬁrmed
the presence of Ni2+ in these ﬁlms and showed a Ni : N ratio of 0.5 : 1, which is

consistent with our previous results and indicative of one Ni2+ per NT A moiety.” '7

 

 

10 -
...]
_l
N 5
_ E .
8 j E
x o. :1
N
s ‘9 4'2:
2
6 .—
4 l l l l
1100 1000 900 800 700

Binding Energy (EV)

Figure 4.3. XPS Spectrum of a PHEMA-NTA-Ni2+ ﬁlm on a gold substrate.

119

To determine how much Ni2+ was bound to the PHEMA-NTA ﬁlms inside

alumina membranes, we passed 5 mL of 50 mM EDTA through a PHEMA-NTA-Ni2+-

   
  
  

 
 
  
 

0.6 -
o 0.5 -
o 0.4 -
$0.4 -
0 (D
s o 20 3 .
8 f2“ '
0.24 (0.2 ‘
0.1 ~

 

 

 

 

0 o . r i .
300 350 400 0 0.01 0.02 0.03 0.04
Wavelength (nm) Ni2+-EDTA (M)

Figure 4.4. Determination of the amount of immobilized Ni2+ in a
PHEMA-NTA-Ni2+ ﬁlm in the membrane. The ﬁgure on the left shows the
calibration Spectra for Ni2+-EDTA standards (solid lines) as well as the
spectrum of the sample (circle), which was prepared by passing 5.0 mL of 50
mM EDTA (pH 7.2) through a PHEMA-NTA-Ni2+-derivatized alumina
membrane. The plot on the right shows the calibration curve for absorbance

at 384 nm and indicates the absorbance of the sample.

120

coated membrane and subsequently analyzed the EDTA solution for Ni2+ using UV-vis
spectroscopy (Figure 4.4). According to this procedure, the amount of Ni2+ bound in the
membrane is 29.5 umol. Considering the effective surface area of the membrane is ~1500
cm2 (pore diameter of 0.25 um, 50% porosity, and a membrane thickness of 60 pm), this
corresponds to a Ni2+ coverage of 18.4 nmol/cmz. Bruening and coworkers reported Cu2+
coverage to be 70 nmol/cm2 for a 55 nm PAA ﬁlm derivatized with NTA-Cu2+ on a gold
substrate.15 Our Ni2+ coverage is ~Mt of the previously reported Cu2+ coverage,
presumably because of a thinner PHEMA ﬁlm (10 nm according to ﬂow experiments).l7
4.3.2. HisU Binding to PHEMA-NTA-Ni2+ Brushes on Gold Substrates

A gold substrate modiﬁed with PHEMA-NTA-Ni2+ was immersed in a solution
containing 0.01 mg/mL HisU in pH 7.2 phosphate buffer for 2 hours followed by
immersion in pH 7.2 phosphate buffer (no HisU) for 15 min and rinsing in EtOH. The
spectrum of the ﬁlm Showed growth in the absorbance at 1680 cm'1 (amide 1 band) and
the appearance of a small peak at 1545 cm'1 (amide H band), similar to the BSA binding
spectrum shown in last chapter (Figure 3.1). Subtracting the spectrum of PHEMA-NTA-
Ni2+ from that of PHEMA-NTA—Niztiiisu yielded a bound-HisU spectrum, which is
dominated by amide absorbances (Figure 4.5a). (The negative peak at 1740 cm'1 likely
result from deprotonation of some —COOH groups.)

Comparison of the amide intensities in Figure 4.5a with those of spin-coated ﬁlms
of pure HisU with known thicknesses suggests that about 31 nm of HisU adsorbed to this
ﬁlm. This corresponds to a binding capacity of 3.1 (lg/cm2 or about 8 monolayers of

HisU (assuming a HisU density of 1 g/cm3 and a monolayer thickness of 4.0 nm).17

121

 

0.02

 

 

Absorbance

 

b -/\_ _,\__

‘—

 

 

 

 

1 850 1 750 1 650 1 550 1450

Wavenumbers (cm")

Figure 4.5. Subtracted reﬂectance FT IR spectra of protein immobilized on
PHEMA-NTA-Ni2+ ﬁlms after exposure of the ﬁlms to (a) 0.01 mg/mL HisU or (b)
1 mg/mL BSA. The spectra were obtained by subtracting the spectrum of PHEMA-

NTA—Ni2+ from that of PHEMA-NTA-Ni2+-protein (both ﬁlms were rinsed with

buffer and ethanol prior to the measurement).

To Show that PHEMA-NTA-Ni2+ ﬁlms bind his tagged species selectively over other
proteins, we also immersed these coatings in 1 mg/mL, stirred BSA solutions for ~15 h.

Theamide intensities in the bound BSA spectrum (Figure 4.5b) Show that only 3.2 nm of

122

BSA was absorbed to the ﬁlm, or <10% of the amount of HisU that bound to these
coatings. Moreover, the concentration of HisU in solution was 0.01 mg/mL, while that of
BSA was 100-fold higher. At concentrations of 0.01 mg/mL, no BSA binding to

PHEMA-NTA-Ni2+ was detected.

4.3.3. HisU Binding to PHEMA-NTA-Ni2+ Brushes in Membranes

To test the HisU binding capacity of PHEMA-NTA-Ni2+ ﬁlms in membranes, we
pumped 10 mL of a 0.3 mg/mL HisU solution in pH 7.2 phosphate buffer through the
membrane, collected the permeate over speciﬁc time intervals and analyzed the permeate
samples using a Bradford assay. The breakthrough curve for HisU binding to PHEMA-
NTA-Ni2+ in membranes and similar curves for BSA and myoglobin binding to PHEMA
—NTA-Cu2+-modiﬁed membranes are shown in Figure 4.6. PHEMA-NTA-Cu2+ modiﬁed
membranes showed high binding capacities for both BSA (0.67 mg/cmz) and myoglobin
(0.50 mg/cmz), but the PHEMA-NTA-Ni2+-modiﬁed membrane Showed an even higher
binding capacity for HisU (0.72 mg/cmz). Moreover, the breakthrough curve was much
Sharper for HisU. The sharpness of the breakthrough curves increases with decreasing
molecular mass of the absorbing species, and suggests more rapid diffusion of smaller
proteins into the polymer brushes. Full breakthrough curves for BSA and myoglobin
binding to membranes with PHEMA-NTA-Ni2+ brushes could not be obtained because of

the minimal binding in this case.

123

0.4 -

 

 

 

-£i-My0 ; A
“03‘ -B-HisU '- ‘r: ‘9
g -e-BSA ,. ', C
\o)
E, .
u
E 02 ' 0
E!
‘E
o
2 ll
00.1 - I
o I;
O :i'” ‘/".'T‘5T. . . r
0 5 10 15

Volume (m L)
Figure 4.6. Breakthrough curves for adsorption of 0.3 mg/mL HisU, 0.3
mg/mL BSA, and 0.35 mg/mL myoglobin in membranes modiﬁed with
PIIEMA-NTA-Ni2+ (HisU) or PHEMA-NTA-Cu2+ (BSA and myoglobin).
The permeate ﬂow rate was initially 2.4 mL/min and 0.9 mL/min at the end

of the experiment.

After measuring the breakthrough curve of HisU, the PHEMA-NTA-Ni2+-HisU
membrane was washed with 20 mL washing buffer followed by 20 mL phosphate buffer,
and HisU was then eluted with 5-10 mL elution buffer (imidazole). Analysis of the

eluent using a Bradford assay showed that 99% of the bound HisU was recovered.

124

To prove that PHEMA-NTA-Ni2+ ﬁlms selectively bind HisU, we also
investigated the binding of BSA and myoglobin to PHEMA-NTA-Ni2+ ﬁlms in
membranes. A PHEMA-NTA-Ni2+-derivatized membrane was ﬁrst loaded with 10 mL
of 0.2 mg/mL BSA or 10 mL of 0.2 mg/mL myoglobin in buffer. After washing the
membrane with washing buffer, we passed 6 mL of elution buffer through the membrane
and analyzed both the feed solutions before passing through the membrane and eluents.
Figure 4.7 a is the UV-vis spectrum of the 0.2 mg/mL BSA feed solution, while Figure
4.7 b is the spectrum of the eluent from the membrane loaded with BSA. The
absorbance of the eluent at 595 nm is <0.001, showing that an undetectable amount of
BSA was bound to and eluted from the membrane. Similar spectra for myoglobin (spectra
4.7 c and (1 show that the membrane binds a small amount (0.05 mg) myoglobin.
However, the amount of myoglobin bound is only 3% of the amount of HisU bound

under similar conditions.

125

0.3 «

 

Absorbance (a.u.)

 

 

Wavelength (nm)

 
 
 
 
 

0.4 -

.0 .o
N (D
l l

O
1.;
l

 

Absorbance (a.u.)

 

-0 .2 I I I I 1
400 500 600 700 800

Wavelength (nm)
Figure 4.7. UV-vis spectra of (a) a 0.2 mg/mL BSA solution; (b) eluent (imidazole
solution) from an alumina-PHEMA-NTA-Ni2+membrane treated with 10 mL of 0.2
mg/mL BSA solution and washing buffer; (c) a 0.2 mg/mL myoglobin solution; ((1)
Eluent (imidazole solution) from an alumina-PHEMA-NTA-Ni”membrane treated
with 10 mL of 0.2 mg/mL myoglobin solution and washing buffer. All solutions
were mixed with Coomassie reagent prior to measurement of the spectra, and blank

Coomassie reagent was used as a background.

126

 

4.3.4. Separation of Protein Mixtures

Two sets of experiment were carried out to test the ability of alumina-PHEMA-
NTA-Ni2+ membranes to purify his tagged proteins. In the ﬁrst experiment, 6 mL of
phosphate buffer containing BSA, myoglobin and HisU (0.05 mg/mL each) was passed
through the PHEMA-NTA-Ni2+-derivatized membrane, and the membrane was then
rinsed with 20 mL washing buffer followed by 20 mL phosphate buffer. Finally, we used
5-10 mL elution buffer to recover the bound protein and then analyzed the eluent by
electrophoresis. We then washed the membrane with 10 mL of 50 mM EDTA and rinsed
it with water. After recharging the membrane with 0.1 M Ni2+ and rinsing with 20 mL
water and 20 mL phosphate buffer, in a second experiment, 10 mL of phosphate buffer
containing 1 mg/mL BSA and 0.05 mg/mL HisU was passed through the membrane, and
the membrane was rinsed, treated with elution buffer, and analyzed using the procedure
described above.

Figure 4.8 shows SDS-PAGE analyses of the protein samples. Lanes 2 and lane 6
are the feed and eluent of the ﬁrst experiment. Bands for all three proteins are clearly
visible for the feed solution, but only a HisU band appears in the eluent, showing that the
membrane selectively binds HisU. The gel suggests that the purity of the HisU is >99%.
Lanes 7 and 8 demonstrate the results for the second experiment with a solution
containing 1 mg/mL BSA and 0.05 mg/mL HisU. In this case, only HisU appears in the
electropherogram of the eluent, even though BSA was in 20-fold excess in the feed.
Purity of the HisU relative to BSA is at least 99%. Analysis of the protein concentration

in the eluent using a Bradford assay showed that essentially all of the bound HisU

127

(>99%) was eluted. These experiments clearly demonstrate the ability of PHEMA-NTA-

Ni2+-modiﬁed membranes to purify his tagged proteins.

5%?

 

- <— Myoglobin

HisU

 

‘ I ‘1

12345678

Figure 4.8. SDS-PAGE analysis (silver staining) of protein solutions and
eluents from alumina—PHEMA-NTA-Ni2+ membranes loaded with proteins.
Samples are: lane 1, standard broad range ladder; lane 2, mixture of BSA,
myoglobin and HisU (0.05 mg/mL each); lane 3, pure BSA; lane 4, pure
myoglobin; lane5, pure HisU; lane 6, eluent from a membrane loaded with a
mixture of BSA, myoglobin and HisU (0.05 mg/mL each); lane 7, mixture of
BSA and HisU; lane 8, eluent from a membrane loaded with 10 mL of a

solution containing 1 mg/mL BSA and 0.05 mg/mL HisU.

128

4.4. Conclusions

PHEMA-NTA-Ni2+ brushes immobilized on both gold substrates and alumina
membranes were successfully prepared according to a procedure similar to that given in
Chapter 3. Transmission IR conﬁrmed the growth and successful derivatization of
PHEMA inside the alumina membrane, and XPS data suggested that the ratio of Ni2+ to
NTA in these ﬁlms is approximately 1:1. Protein binding experiments with ﬁlms on gold
substrates Showed that PHEMA-NTA—Ni2+ brushes are capable of selectively binding
large amounts (3.] ug/cmz) of HisU, and porous alumina membranes modiﬁed with
derivatized brushes also showed remarkable capacities (120 mg/cm3) for the binding of
HisU. Most importantly, gel electrophoresis studies Showed that these high-capacity
membranes are highly selective for puriﬁcation of his tagged proteins. Even in a 20-fold

excess of BSA, membranes gave 99% HisU.

129

4.5.

10.

ll.

12.

13.

14.

15.

l6.

17.

18.

References

Zhen, G.; Falconnet, D.; Kuennemann, E.; Voros, J .; Spencer, N. D.; Textor, M.;
Zurcher, S., Adv. F unct. Mater. 2006, 16, 243-251.

Draveling, C.; Ren, L.; Haney, P.; Zeisse, D.; Qoronﬂeh, M. W., Protein
Expression Purif. 2001, 22, 359-366.

Skerra, A.; Schmidt, T. G., Biomol. Eng. 1999, 16, 79-86.

Stiborova, H.; Kostal, J .; Mulchandani, A.; Chen, W., Biotechnol. Bioeng. 2003,
82, 605-611.

Gaberc-Porekar, V.; Menart, V., J. Biochem. Biophys. Methods 2001, 49, 335-
360.

Everson, R. J .; Parker, H. E., Bioinorg. Chem. Appl. 1974, 4, 15-20.

Porath, J .; Olin, B., Biochemistry 1983, 22, 1621-1630.

Porath, J.; Carlsson, J .; Olsson, 1.; Belfrage, 6., Nature 1975, 258, 598-599.
Suen, S.-Y.; Liu, Y.-C.; Chang, C.-S., J. Chromatogr. B 2003, 797, 305-319.
Ozkara, S.; Yavuz, H.; Denizli, A., J. Appl. Polym. Sci. 2003, 89, 1567-1572.
Zou, H.; Luo, Q.; Zhou, D., J. Biochem. Biophys. Methods 2001, 49, 199-240.

Che, S.; Liu, Z.; Ohsuna, T.; Sakamoto, K.; Terasaki, O.; Tatsumi, T., Nature
2004, 429, 281-284.

Clairbois, A. S.; Letoumeur, D.; Muller, D.; Jozefonvicz, J., J. Chromatogr. B
1998, 706, 55-62.

Henis, J. M. S.; Tripodi, M. K.; Stimpson, D. 1. European Patent 0,221,046B1,
October 30, 1987.

Dai, J .; Bao, Z.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L., Langmuir
2006, 22, 4274-4281.

Kawai, T.; Saito, K.; Lee, W., J. Chromatogr. B 2003, 790, 131-142.

Sun, L.; Dai, J .; Baker, G. L.; Bruening, M. L., Chem. Mater. 2006, 18, 4033-
4039.

Ulbricht, M.; Yang, H., Chem. Mater. 2005, 17, 2622—2631.

130

19.

20.

21.

22.

Johnson, R. D.; Arnold, F. H., Biotechnol. Bioeng. 1995, 48, 437-443.
Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Irnmaraporn, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T., Macromolecules 1999, 32, 8716-8724.

Kim, J .-B.; Bruening, M. L.; Baker, G. L., J. Am. Chem. Soc. 2000, 122, 7616-
7617.

Kim, J.-B.; Huang, W.; Bruening, M. L.; Baker, G. L., Macromolecules 2002, 35,
5410-5416.

131

Chapter 5

Summary and Future Work

The research in this dissertation investigated the application of polymer brushes in
two areas: high ﬂux pervaporation and high capacity protein binding. In chapter 2, I
discussed the synthesis of ultrathin polymer ﬁlms at the surface of porous substrates to
prepare high-ﬂux pervaporation membranes. Chapter 3 explored the growth of brushes in
the interior of porous substrates to prepare membranes with unusually high protein
adsorption capacities. Chapter 4 expanded on the protein adsorption work and
demonstrated that polymer brushes derivatized with Ni2+ complexes are very attractive
for puriﬁcation of his tag proteins.

In applications of polymer brush-containing membranes as adsorbers, although

brush-modiﬁed alumina membranes are capable of high—capacity protein binding, they

 

Figure 5.1. FESEM image of a PVDF membrane (left) with a nominal

cutoff of 0.45 um and an alumina membrane (right) with pore size of 0.25

urn.

132

suffer from low ﬂux, fragility, and high cost. The largest pores available in commercial
alumina membranes are 0.25 um in diameter (Figure 5.1), and this greatly limits the ﬂux
available with these systems. To overcome this challenge, our future research aims at
extending the very promising results with BSA and HisU binding to derivatized PHEMA
brushes in porous alumina to the preparation of polymer brushes in porous polymer
supports (Figure 5.1). The PVDF membrane shown in Figure 5.1 has a nominal ﬁltration
cutoff of 0.45 urn, but as can be seen pore sizes are much larger than this. Most polymer
supports are both more robust and less expensive than the alumina substrates, and more
importantly, careful control over pore geometry in the polymeric substrates will allow the
increased permeabilities needed for high-ﬂux, high-capacity membrane separations.
Nevertheless, formation of high-capacity absorbers in polymer supporters will require
new chemistries for initiator attachement, and could be quite difﬁcult because of the
incompatibilitiy of many polymeric materials with organic solvents. Relatively thick,
highly adsorbing brushes will also be needed to maintain high capacities in large pores.
PHEMA-NTA-Nizi—modiﬁed membranes showed excellent performance in
selectively recovering HisU from protein mixtures. However, much more research needs
to be done with complex systems such as whole cell extracts and sera. Additionally, after
demonstrating separations from complex mixtures, we will need to remove imidazole
from the isolated protein and cleave the his tag to obtain the desired product. However,
literature procedures including Size exclusion chromatography and dialysis are already
available for these ﬁnal puriﬁcation steps." 2

In the case of pervaporation membranes, we have demonstrated that ultrathin,

brush skins allow high ﬂux and reasonable selectivity. However, further research into

133

polymerization of monomers which are more suitable for pervaporation should give
membranes with an order of magnitude improvement in both ﬂux and selectivity. The
appendix describes some unﬁnished research in this direction. Brush membranes
prepared from high performance monomers will provide far superior performance to that

of commercial membrane systems.

134

References

l. Colangeli, R.; Heijbel, A.; Williams, A. M.; Manca, C.; Chan, J .; Lyashchenko,
K.; Laura Gennaro, M., J. Chromatogr. B 1998, 714, 223-235.

2. Casey, J. L.; Keep, P. A.; Chester, K. A.; Robson, L.; Hawkins, R. E.; Begent, R.
H. J., J. Immunol. Methods 1995, 179, 105-116.

135

 

Appendix

Preliminary Work in Separating Organic Solvent Mixtures Using

Other Ultrathin Polymer Membranes

Introduction

Ethyl acetate—hexane mixtures are used as eluents in liquid chromatography in
academic laboratories, as well as in the pharmaceutical and chemical industry.
Conventional distillation-based separations of ethyl acetate and hexane for recycling are
not possible, however, due to the formation of an azeotropic mixture (Figure A1). In this

study, I examined the possibility of using a polymer membrane to recover hexane from

 

1

0.8 “

0.6 ‘

YHexane

0.4 “

0.2 “

 

 

 

o l l I l
0 0.2 0.4 0.6 0.8 1

xHexane

Figure Al. Vapor liquid equilibrium of ethyl acetate-hexane
mixtures at 101.32 kPa. (The X and Y axes are the fractions of hexane

in the liquid and vapor phases, respectively.)

136

ethyl acetate-hexane mixture. This appendix also describes efforts aimed at the
polymerization of dimethyldi(methacryloyloxy-1-ethoxy)silane (DMDMAES) from a
surfac to prepare ultrathin membranes with pervaporation properties similar to those of
PDMS.
Experiments and Results
Materials

Poly(allylamine hydrochloride) (PAH) (Mw = 70,000), sodium poly(styrenesulfonate)
(PSS) (Mw = 70,000), 3-mercaptopropionic acid (MPA), dimethylformamide (DMF,
anhydrous, 99.8%), tetrahydrofuran ('1‘ HF, anhydrous, inhibitor free, 99.8%), acetone
(anhydrous, 99%), 2-bromopropionylbromide (2-BPB), and CuBr (99.999%) were used
as received from Aldrich. MnClz (Acros) and NaBr (Spectrum) were also used as
received. HEMA (Aldrich, 97%, inhibited with 300 ppm hydroquinone monomethyl
ether) was puriﬁed by passing it through a column of activated basic alumina (Spectrum).
Dichlorodimethylsilane (99%), triethylamine (Et3N, 99.5%), and N, N, N’, N’, N”-
Pentamethyldiethylenetriamine (PMDETA,99%) were also purchased from Aldrich and
used as received. Pervaporation experiments were performed using analytical grade ethyl
acetate and hexane (Aldrich). AnodiscTM porous alumina membranes (Fisher) with 0.02
um-diameter surface pores were used as supports for membrane formation, while gold-
coated wafers (200 nm of sputtered Au on 20 nm Cr on a Si (100) wafer) were employed
as substrates for ellipsometry and reﬂectance Fourier Transform Infrared (FT IR)

spectroscopy.

137

Monomer Synthesis

Dimethyldi(methacryloyloxy-l-ethoxy)silane (DMDMAES, Figure A2), was prepared
following a literature by the etheriﬁcation reaction of HEMA with dichlorodimethylsilane
in anhydrous THF in the presence of Et3N (base).l Speciﬁcally, THF (100 mL), Et3N
(24.7 mL, 17.9 g, 0.177 mol), and HEMA (18.3 mL, 19.6 g, 0.151 mol) were added to a
250 mL round-bottom ﬂask and kept under a dry nitrogen atmosphere. The contents were
stirred and cooled to 0-5 °C in an ice bath.. Next, dichlorodimethylsilane (7.3 mL, 7.8 g,
0.060 mol) was added slowly for 5 minutes to the mixture while stirring, giving an
exotherm (3.2-10.8 °C), and the reaction mixture was stored in a refrigerator overnight.
Filtering of the reaction mixture removed the precipitated triethylamine hydrochloride,
and THF (50 mL) was then added to wash the ﬁltrate. THF and unreacted Et3N were
subsequently removed under reduced pressure using a rotary evaporator at room
temperature. The resulting oily residue was puriﬁed by column chromatography (silica
gel/CHzClzz hexane 75:25) to give an oily liquid, which was then distilled under vacuum.
The purity (>95%) of this cross-linkable monomer was conﬁrmed by 1H NMR
spectroscopy, and the relevant spectrum is displayed in Figure A2.
Polymerization of DMDMAES

Before polymerization, we ﬁrst deposited a multilayer polyelectrolyte ﬁlm on the
substrate and subsequently attached initiators to these modiﬁed surfaces. The initiator
anchoring was’done according to the literature procedure outlined in scheme A1.2
Polymerization of DMDMAES occurred by immersion of the initiator-coated substrate in
a degassed acetone solution containing DMDMAES and a CuBr/PMDETA catalyst

system (Scheme A1). To be speciﬁc, 43.2 mg CuBr, 10 mL

138

b
r—*—\ c

d
l H -_ --l

r 1 f 7 l I l l I l

 

8 7 6 5 4 3 2 ‘I 0
a b b a
H3C\C/CH2 e HZC\\C/CH3
l CH3 I
O¢C\O—C:12—C:l2—O—Sli—O—C:2—C:12—O/C§O
CH3

Figure A2. ‘H NMR spectrum of DMDMAES in 0303.

DMDMAES and 10 mL acetone were ﬁrst mixed in a Schlenk ﬂask and degassed via
three freeze-pump-thaw cycles. Then, 62.7 uL of PMDETA was quickly added to the
mixture, which was then subjected to another freeze-pump-thaw cycle and stirred until a
clear blue solution formed. Initiator-coated substrates and the sealed ﬂask containing the
polymerization solution were then transferred to a glove bag that was purged with
nitrogen for about 1 h. The polymerization solution was ﬁnally transferred to the

container holding the substrates, and polymerization was carried out for 20 h.

139

«69 1G) + +
ﬁe) ‘9
ﬁg) ‘6)
d6) 2P$SIPAH ‘9 _+ + _+ +
=® ‘9 + +
..a :e + +
BrCOCH(CH3)Br,
N(C2H5)3 l THF

 

Cu Brl PM DETA,
DMDMAES

Acetone, 22°C,
20h

 

Scheme Al. Schematic diagram showing polymerization of DMDMAES
from a polyelectrolyte surface modiﬁed with an initiator. PSS stands for

poly(styrenesulfonate), and PAH represents protonated poly(allylamine).

140

FTIR Characterization of Membranes

The FTIR spectrum (Spectrum a, Fig. A3) of initiator attached to 5 PSS/PAH bilayers
has a strong peak intensity in the amide region (1650—1560 cm"), indicating initiator
attachment. After growth of a PDMDMAES ﬁlm, a strong carbonyl peak appeared at
1730 cm“1 in the spectrum (Spectrum b, Fig. A3) as well as a C=C stretching band (1637
cm") from unpolymerized vinyl groups. Ellipsometry results Show that the polymerized
ﬁlm has a thickness of about 25 nm, including the polyelectrolyte layer which was 15 nm

thick.

 

0.05

C=O & C=C
stretch

Absorbance

 

b
:: Va A Amide bondj

I I

4000 3000 2000 1 000
Wavenumbers (cm")

 

 

 

 

Figure A3. Reﬂectance FTIR spectra of (a) the initiator attached

onto 5 bilayers of PSS/PAH ﬁlm, (b) PDMDMAES ﬁlm after 20 h

growth from (a).

141

Pervaporation Experiments

Pervaporation was performed at room temperature with mixtures of ethyl acetate and
hexane. Figure A4 shows the effect of feed concentration on the performance of the
membrane. As the weight fraction of hexane in the feed solution increases from 20% to
80%, the weight fraction of hexane in the permeate rapidly increase until it reaches a
limiting value. Our membrane greatly outperformed the vapor—liquid equilibrium
separation achievable by vacuum distillation, and permeate compositions were not
affected by the formation of the hexane/ethyl acetate azeotrope, as shown in Figure A4
(a). The hydrophobic nature of the membrane allows preferential adsorption of hexane,
which likely swells the membrane due to its hydrophobicity. The total ﬂux of
hexane/ethyl acetate mixture increases with the hexane concentration in the feed, while
selectivity, initially increases and then decreases as the hexane concentration in the feed
increases. This is presumably because at the initial stage, the membrane is swollen by
hexane and preferentially absorbs hexane molecules. At extremely high swelling,
however, both hexane and ethyl acetate can diffuse through the membrane at Similar rates,
and selectivity declines.
Future Work

Growth of a series of PDMS-like polymers from porous alumina supports using ATRP
should be investigated. The pervaporation performance of these polymers also needs to

be examined.

142

 

 

 

XHexane in Permeate

 

 

0 0.2 0.4 0.6 0.8 1
XHexane in Feed

 

 

 

 

5o 5
(b)

401 4
£30 .___ 3 :5
5320* 2 as
w ’5

10- 1d r-1L—L

o o

 

o 0.2 0.4 0.6 0.8 1
XHexane in Feed

Figure A4. (a) Hexane/ethyl acetate vapor—liquid equilibrium (VLE)
relative to separation achieved by PDMDMAES membrane(Exp); (b)

Effect of hexane concentration in the feed on separation factor and ﬂux.

143

References
l. Themistou, E.; Patrickios, C. S., Macromolecules 2004, 3 7, 6734-6743.

2. Balachandra, A. M.; Baker, G. L.; Bruening, M. L., J. Membr. Sci. 2003, 227, l-
14.

144

MICHIGAN STATE UNl E STY LIBRARIES

\llllllllllll llllllllllllll11111211

3 1293 0284