_ as
Rea

L 1.5
. 5..
A3

 

i:

.‘ xucx, .w.

13...“.“13
5B,.

.35.

I I
:. z?
I. x . :

 

i217“.

 

 

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

HYPERBRANCHED POLYMER FILMS AS
ULTRATHIN MEMBRANE SYSTEMS

presented by

80 Young Kim

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Chemistry

J" 7,4 (
ﬂ g/km
Major professU *

 

Date 0 O a"

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65-p.15

HYPERBRANCHED POLYMER FILMS AS ULTRATHIN

MEMBRANE SKINS AND ADHESIVES

By

Bo Young Kim

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

2002

ABSTRACT
HYPERBRANCHED POLYMER FILMS AS ULTRATHIN
MEMBRANE SKINS AND ADHESIVES
By

Bo Young Kim

Hyperbranched polymers are attractive materials for forming membranes and thin
ﬁlms because their tree-like structure allows them to cover underlying substrates. Field-
emission scanning electron microscopy and atomic force microscopy images clearly
show that hyperbranched poly(acry1ic acid) (PAA) ﬁlms can completely cover permeable
substrates without ﬁlling underlying pores, thus creating ultrathin membranes on porous
supports. The hyperbranched PAA provides an ultrathin, discriminating membrane,
while the porous support gives mechanical strength. Gas-transport studies indicate that
3-layer PAA ﬁlms do not show a high selectivity by themselves, but selectivity improves
signiﬁcantly after covalent derivatization of PAA with H2NCH2(CF2)6CF3.

Hyperbranched poly(amidoamine) (PAMAM) dendrimers are another material
used for synthesis of hyperbranched membranes. These polymers are well-deﬁned
polycations that contain a high charge density at their surface due to multiple amine
groups. Thus, dendrimers provide a unique material for depositing ﬁlms by alternating
electrostatic adsorption of polycations and polyanions. The thickness of PAMAM
dendrimer/linear poly(acry1ic acid) ﬁlms varies dramatically with deposition pH, e.g.,
the thickness of an adsorbed generation 4 (G4) PAMAM dendrimer/PAA “bilayer” can

be tuned from <10 A to >2,000 A. Thicknesses of GS PAMAM dendrimer/PAA bilayers

can be as high as 4,000 A. The thickest ﬁlms result when the dendrimer is deposited at
pH 8 and PAA is deposited at pH 4 because both polymers are only partially ionized at
these pH values. The relatively low charge density on both PAA and dendrimers requires
more adsorption to compensate the surface charge. Alternating deposition of 5-bilayer
G4 PAMAM dendrimer/PAA ﬁlms on porous alumina supports yields highly gas
permeable membranes, presumably due to the open interior of these molecules.
Dendrimer/PAA ﬁlms can also form microporous polyelectrolyte ﬁlms through structural
changes induced by a reduction in ambient pH.

The high density of functional groups at the exterior of dendrimers also makes
these materials attractive as possible adhesives. This dissertation shows that a dendrimer
interlayer is capable of bonding strongly two PAA-coated surfaces. Speciﬁcally, two
gold-coated Si wafers were coated with PAA-terminated dendrimer ﬁlms, and a drop of
10'5 M dendrimer solution was placed between the wafers. Pressing the wafers together
then yielded a bond strength of >1.6 MPa. Cross-sectional ﬁeld-emission scanning
electron microscope (FESEM) images show that the ﬁlms on the two wafers are in
intimate contact, suggesting that the thin dendrimer interlayer electrostatically binds the
surfaces together. This room-temperature, aqueous-based adhesion may prove useful in
applications such as bonding of wafers to form micro-electro-mechanical devices or lab-
on-a-chip systems.

Overall, these studies demonstrate the versatility and unique character of
hyperbranched polymer ﬁlms. The dense functionality and unusual structure of these

systems provide unique opportunities for surface modiﬁcation, separation, and adhesion.

T 0 My Family

iv

ACKNOWLEDGEMENTS

I would like to express my sincere appreciation to Prof. Merlin L. Bruening for
his guidance, support, encouragement and mentoring throughout the work presented here.
It has been a real pleasure to work with him, and I have certainly learned a great deal
during the research in the Bruening lab.

I also wish to thank my guidance committee members - Dr. Blanchard, Dr.
Borhan, Dr. Geiger for their guidance and constructive comments. I am thankful to my
former advisor Dr. Eick for his encouragement and help throughout graduate study. Also
a sincere thank-you to all the people in the Bruening group for being “Bruening group
family” at some point: Dr. Nagale, Dr. Xiao, Dr. Huang, Jeremy, Anika, Skanth, Sandra,
Brian, Dan, Matt, Anagi, Jinhua, Shaoyun, Yingda and Dr. Hong. In the absence of Dr.
Nagale’s help for the gas permeability measurements, I would probably still have
crosstalk on the ultrathin, hyperbranched Poly(acrylic acid) (PAA) on porous alumina
supports. Without Dr. Bi’s wonderfully managed Field-Emission Electron Microscope
(FESEM) and Atomic Force Microscope (AF M), none of the micrographs presented here
would have been possible.

I give thanks to my father, mother, sisters and brother’s family for listening,
giving advice, and providing me with perspective. They always encouraged me to
continue in the pursuit of my scientiﬁc endeavors. I also give sincere thanks to J ineun,

Yuna, and Dale for their support and understanding me.

TABLE OF CONTENTS

List of Tables
List of Figures

CHAPTER 1. Introduction

1.1 Synthetic Separation Membranes
1.2 Gas Separation Membranes

1.3 Hyperbranched Polymers

1.4 Structure of this Dissertation

1 .5 References

CHAPTER 2. Ultrathin Hyperbranched Poly(acrylic acid)
Membranes on Porous Alumina Supports
2. 1 Introduction

2.2 Experimental

2.3 Results and Discussion
2.4 Conclusions

2.5 References

CHAPTER 3. pH-Dependent Growth and Morphology of
Multilayer Dendrimer/Poly(acrylic acid) Films
3. 1 Introduction

3.2 Experimental

3.3 Results and Discussion
3.4 Conclusions

3.5 References

CHAPTER 4. The Use of Dendrimer/Poly(acrylic acid) Films
as Adhesives
4. 1 Introduction

4.2 Experimental

4.3 Results and Discussion
4.4 Conclusions

4.5 References

CHAPTER 5. Conclusions and Future Work

vi

vii

viii

l 1
20
22

27
28
32
35
70
71

74
75
76
80
93
94

97
98
99
101
110
111

112

LIST OF TABLES

Table 2.1 Ratios of the ﬂuxes of different gases through 3-
layer PAA ﬁlms and ﬂuorinated 3-1ayer PAA ﬁlms. 68

Table 3.1 Ellipsometric thicknesses of G4 and G8 PAMAM

dendrimer (pH=8)/PAA (pH=4) ﬁlms on gold as a
function of the number of bilayers. 84

vii

Figure 1.1
Figure 1.2
Figure 1.3
Figure 1.4

Figure 1.5

Figure 1.6

Figure 1.7

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5(a)

LIST OF FIGURES

Membrane-based separation.

Schematic representation of various membranes.
Bisphenol A polysulfone.

Polycarbonate.

Gas transport through a dense membrane according to the
solution-diffusion model.

Synthesis of hyperbranched PAA ﬁlms.
Synthesis of PAMAM Dendrimers.

Idealized cartoon of a hyperbranched PAA ﬁlm graﬁed
onto a porous alumina substrate. Substrate thickness is not
drawn to scale.

Schematic diagram of the synthesis of 1 layer of PAA on a
surface. Additional layers were prepared by grafting onto
previously deposited PAA.

External reﬂection FTIR spectra of 1, 2, 3, 4, 5, and 6
layers of PAA on porous alumina (0.02 mm pore diameter)
that was sputter coated with 63 nm thick gold.

Transmission F TIR spectra of a 3-layer PAA ﬁlm before
and aﬁer ﬂuorination. The substrate was gold-coated (5
nm) porous alumina (0.02 um pore diameter).

Field-emission scanning electron micrograph (cross-
section) of porous alumina (0.02 um pore diameter) before
deposition of PAA ﬁlms. In this image, 5 nm of gold was
deposited before PAA growth.

viii

13

15

30

31

36

38

42

 

 

Figure 2.5(b)

Figure 2.5(e)

Figure 2.5(d)

Figure 2.5(e)

Figure 2.5(f)

Figure 2.6

Figure 2.7

Field-emission scanning electron micrograph (cross-
section) of porous alumina (0.02 pm pore diameter) after
deposition of 4-layer PAA ﬁlms. In this image, 5 nm of
gold was deposited before PAA growth. After cleaving
the membrane, sample was sputter-coated with 5 nm of
gold imaging.

F ield-emission scanning electron micrograph (cross-
section) of porous alumina (0.02 pm pore diameter) after
deposition of 6-layer PAA ﬁlms. In this image, 5 nm of
gold was deposited before PAA growth. After cleaving
the membrane, sample was sputter-coated with 5 nm of
gold imaging.

F ield-emission scanning electron micrograph (cross-
section) of porous alumina (0.02 um pore diameter) before
deposition of PAA ﬁhns. In this image, 63 nm of gold
was deposited before PAA growth.

F ield-emission scanning electron micrograph (cross-
section) of porous alumina (0.02 pm pore diameter) after
deposition of 4-layer PAA ﬁlms. In this image, 63 nm of
gold was deposited before PAA growth. After cleaving
the membrane, sample was sputter-coated with 5 nm of
gold imaging.

Field-emission scanning electron micrograph (cross-
section) of porous alumina (0.02 um pore diameter) after
deposition of 6-layer PAA ﬁhns. In this image, 63 nm of
gold was deposited before PAA growth. After cleaving
the membrane, sample was sputter-coated with 5 nm of
gold imaging.

Field-emission scanning electron micrographs (cross-
section) of porous alumina (0.2 pm pore diameter) before
and after deposition of 6—PAA layers. 5-nm of gold was
deposited before PAA growth, and samples were sputter-
coated with 5 nm of gold for imaging after cleaving.

Field-emission scanning electron micrograph (cross-
section) of a 4-layer PAA ﬁlms grown of a gold-coated
(63 nm) porous alumina membrane (0.02 pm pore
diameter). The membrane was imaged without coating
with gold after cleavage.

ix

43

44

45

46

47

48

49

Figure 2.8(a)

Figure 2.8(b)

Figure 2.8(c)

Figure 2.8(d)

Figure 2.8(e)

Figure 2.8(f)

Figure 2.9(a)

Figure 2.9(b)

Tapping mode atomic force micrograph of

gold-coated porous alumina before grafting of PAA ﬁlms.
The gold coating on 0.02 pm surface pore-diameter-
Anopore membranes was 5 nm thick in this image.

Tapping mode atomic force micrograph of

gold-coated porous alumina after grafting of 4-layer PAA
ﬁlms. The gold coating on 0.02 pm surface pore-
diameter-Anopore membranes was 5 nm thick in this
image.

Tapping mode atomic force micrograph of

gold-coated porous alumina after grafting of 6-layer PAA
ﬁlms. The gold coating on 0.02 pm surface pore-
diameter—Anopore membranes was 5 nm thick in this
image.

Tapping mode atomic force micrograph of

gold-coated porous alumina before grafting of PAA ﬁlms.
The gold coating on 0.02 um surface pore-diameter-
Anopore membranes was 63 nm thick in this image.

Tapping mode atomic force micrograph of

gold-coated porous alumina after grafting of 4-layer PAA
ﬁlms. The gold coating on 0.02 pm surface pore-
diameter-Anopore membranes was 63 nm thick in this
image.

Tapping mode atomic force micrograph of

gold-coated porous alumina after grafting of 6-layer PAA
ﬁlms. The gold coating on 0.02 pm surface pore-
diameter-Anopore membranes was 63 nm thick in this
image.

Tapping mode atomic force micrographs of gold-coated
porous alumina before grafting of hyperbranched PAA.
The gold coating on 0.2 pm pore-diameter-Anopore
membranes was 5 nm thick in this image.

Tapping mode atomic force micrographs of gold-coated
porous alumina after grafting of 6-layers hyperbranched
PAA. The gold coating on 0.2 um pore-diameter-Anopore
membranes was 5 nm thick in this image.

52

53

54

55

56

57

58

59

 

Figure 2.9(c)

Figure 2.9(d)

Figure 2.10

Figure 2.11

Figure 2.12

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Tapping mode atomic force micrographs of gold-coated
porous alumina before grafting of hyperbranched PAA.
The gold coating on 0.2 pm pore-diameter-Anopore
membranes was 63 nm thick in this image.

Tapping mode atomic force micrographs of gold-coated
porous alumina after grafting of 6-1ayers of hyperbranched
PAA. The gold coating on 0.2 um pore-diameter-Anopore
membranes was 63 nm thick in this image.

Optical micrograph of a defect on a porous alumina
substrate coated with 63 nm of gold.

Flow rates of several gases through a 3-layer PAA ﬁlm on
gold-coated (5 nm) porous alumina (0.02 um pore
diameter) as a function of inlet pressure. The membrane
area was 2.8 cm2. Permeability values with standard
deviations are listed in the text.

Flow rates of several gases through a ﬂuorinated 3-layer
ﬁlm on gold—coated (5 nm) porous alumina (0.02 um pore
diameter) as a function of inlet pressure. The membrane
area was 2.8 cm2. Permeability values with standard
deviations are listed in the text.

Schematic diagram of the deposition of a dendrimer/PAA
bilayer on a gold surface. Repetition of steps 2 and 3
yields multiplayer ﬁlms.

Reﬂection FTIR spectra of a MFA-modiﬁed gold
substrate coated with (a) 0.5 (b) 1.5 (c) 2 and (d) 2.5 G4
dendrimer/PAA bilayers. The half bilayer results in the
dendrimer being the top layer in the ﬁlm. Deposition pH
values were 8 for G4 dendrimers and 4 for PAA.

Ellipsometric thicknesses of 2.5-bilayer G4
dendrimer/PAA ﬁlms as a function of the pH used for
dendrimer and PAA deposition. Standard deviations of
thicknesses measured on three different samples ranged
from 3 — 35 % of the measured values.

Ellipsometric thicknesses of 2.5—bilayer G8
dendrimer/PAA ﬁlms as a ﬁmction of the pH used for
dendrimer and PAA deposition. Standard deviations of
thicknesses measured on three different samples ranged
from 2-29 % of the measured values.

xi

60

61

63

65

66

79

81

82

86

Figure 3.5

Figure 3.6

Figure 3.7

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5

Optical microscope images of (a) 2.5-bilayer G8 dendrimer
(pH=4)/PAA (pH=4) and (b) G8 dendrimer (pH=8)/ PAA
(pH=4) ﬁlms. The ellipsometric thicknesses of the ﬁlms
were 80 A and 4,000 A, respectively.

AF M images of 5-bilayer G4 dendrimer (pH=8)/PAA
(pH=4) ﬁlms before (a) and alter (b, c) immersion in a pH
2.5 aqueous solution. For images (a) and (b,
polyelectrolytes were deposited in the presence of 0.2 M
NaCl, and for image (c), no salt was used during ﬁlm
deposition.

Cross-sectional F ESEM images of porous alumina before
(a) and after (b) coating with a 4.5-bilayer G4 dendrimer
(pH=8)/PAA (pH=4) ﬁlm.

Strategy for wafer bonding using dendrimer/PAA ﬁhns.
The inset on the right shows intertwining of dendrimer/PAA
bilayers that should occur throughout the ﬁlm.

Ellipsometric thicknesses of dendrimer/PAA ﬁlms on a base
composed of 5 bilayers of PAH/PAA. Deposition pH values
were 8 for G4 dendrimers and 4 for PAA. The PAH/PAA
base was deposited at pH 4.

Cross-sectional ﬁeld-emission scanning electron microscopy
images of (a) 5 dendrimer/PAA bilayers on 5 PAH/PAA
bilayers on a gold coated wafer and (b) two wafers coated
with similar ﬁlms that were pressed together with a drop of
dendrimer solution between them.

Cross—sectional ﬁeld-emission scanning electron microscopy
image of two ﬁlm-coated wafers that were pressed together
with a drop of water between them. The ﬁlm on the top
gold-coated wafer was composed of 5.5 dendrimer/PAA
bilayers on 5 PAH/PAA bilayers, while the ﬁlm on the
bottom gold-coated contained 5 dendrimer/PAA bilayers on
5 PAH/PAA bilayers.

The tensile strength measurement conﬁguration. A tensile
tester is used to pull the bonded wafers apart.

xii

88

90

92

103

104

107

108

109

Chapter 1

INTRODUCTION

1.1 Synthetic Separation Membranes

Membrane—based separations are important in a wide range of industrial processes
including microﬁltration,"2 ultraﬁltrationf‘4 reverse osmosis,5’6 gas separation7'9 and
pervaporationm'l3 Microﬁltration and ultraﬁltration rely on porous membranes for

”’13 while reverse

separation of species such as proteins, viruses, and colloidal particles,
osmosis, gas separation and pervaporation utilize dense, nonporous membranes. Perhaps
the most well-known application of dense membrane separation is desalination of
seawater.14 Despite such successful applications of membrane technology, many
membrane-based separations are limited by insufﬁcient selectivity and/or permeability.15 '
'8 Because the permeability of many membrane materials is inversely related to
selectivity,7’8 a system with acceptable selectivity often exhibits low ﬂux and vice versa.
Thus, synthesis of new membranes with both increased selectivity and high permeability
presents a challenge in materials research.16

In its operation, a membrane acts as a selective barrier, or interphase, between two
phases that are often homogeneous. Separation occurs when some of the species from a
multi-component mixture are transported across the membrane via driving forces such as
gradients in concentration (AC), temperature (AT), electric potential (AE), and pressure
(AP), while other components are retained by the membrane (Figure 1.1). The success of

the separation depends on the selectivity and the ﬂux of the transported species, and these

quantities are, in turn, dependent on the physical and chemical properties of the

interphase.l4 This dissertation focuses on developing new polymeric materials that are
capable of forming ultrathin membranes, because decreasing membrane thickness

provides one means of allowing high ﬂux with a high-selectivity material.'9

1 H l“ 1

 

 

AC. AP. AT. AE

Figure 1.1 Membrane-based separation.l4

Membrane systems are often classiﬁed into two categories: symmetric and
asymmetric, as shown in Figure 1.2. The symmetric membranes may be further
subdivided into three groups: 1) cylindrical porous membranes which are often used for
enzyme and DNA separations from dilute solutions,20 2) porous membranes that can be
used for microﬁltration,"2 and 3) nonporous, homogeneous dense membranes that are
normally used for separation of small molecules such as gases.7'8 However, the ﬂux
through nonporous, symmetric systems is usually very low because the membrane must
be relatively thick to be mechanically stable.

Asymmetric membranes can also be subdivided into three groups: 1) porous
membranes that have a pore size gradient over the membrane thickness, 2) porous
membranes with a dense top layer, and 3) composite membranes that consist of a dense

layer or skin on a support prepared from a different material that contains 50 to 150 um-

 

 

/ _ ,I'V“ a"! W.
. ,’,' _ .i‘, u '1 a".

‘9" " .‘ ~‘er . .

A. '.' .4"

q .
Porous 5' '~ .- : ' 2.

I... 'P I I.‘:__ -»:.o.ﬁ:“‘

.o :L ‘~ “->‘ . ’

 

 

 

Symmetric

/ \.

Homogeneous

 

 

 

 

(nonporous)

Membranes

 

Porous

\ /'

Asymmetric _, Porous with
tOPIayer (Integrally

skinned)
Dense
+toplayer
<— Porous
suppon

   

- , . .

Composite

 

Figure 1.2 Schematic representation of various membrane types.l4

diameter pores. In the second category of asymmetric membranes, the system is usually
composed of only one material that has an open porous bottom layer, a second layer with
smaller pores in the range of 5 - 500 nm and a dense, continuous phase as a top layer.
These integrally-skinned, asymmetric membranes are generally prepared by a phase-
inversion method that occurs when casting the membrane.21 In both integrally skinned
and composite systems, the porous underlying material stabilizes the weak, ultrathin
surface ﬁlms without presenting a substantial barrier to transport. The thin, dense top
layer, in contrast, offers high selectivity and some mass transfer resistance, but this
resistance is limited due to the minimal thickness of the ﬁlm. In order to obtain high
performance separation membranes with the desired high selectivity and high
permeability, the top layer in the composite membranes must be extremely thin because
ﬂux is inversely proportional to thickness. Additionally, this layer must be free of defects
that decrease membrane selectivity.

In chapter 2, we present the development of composite membranes through
deposition of hyperbranched poly(acry1ic acid)(PAA) on porous alumina supports. The
hyperbranched structure of these ﬁhns allows an ultrathin (< 50 nm) ﬁlm to cover
underlying pores without ﬁlling them and thus, provides 3 hi gh-ﬂux membrane.
Additionally, these ﬁlms can be derivatized widely to modulate selectivity. To provide
background for that work, we ﬁrst discuss research and commercial applications of gas
separation membranes, then review the solution-diffusion mechanism of gas transport
through dense membranes. Subsequently, we review previous work with hyperbranched

PAA ﬁlms as well as dendrimers. The background information on dendrimers is relevant

to chapters 3 and 4, which discuss formation of dendrimer/PAA ﬁlms that are highly gas

permeable and the possible use of these ﬁlms for promoting adhesion.

1.2 Gas-Separation Membranes

In 1866, Graham conducted the ﬁrst studies of membrane-based gas separations
and showed that porous membranes can partially separate gas mixtures according to their
molecular masses.22 More recently, development of synthetic membranes resulted in
industrial gas separations such as removal of C02 from natural gas,23 separation of 02
and N; from air,24 and hydrogen recovery from purge gas in ammonia synthesis.25

One of the most important breakthroughs that allowed practical gas separation
actually occurred in the area of reverse osmosis membranes. In the early 1960’s Loeb
and Sourirajan developed a process to produce cellulose acetate such that it contained a
thin, selective skin on a porous base of the same material.26 The ultrathin skin was
absolutely necessary for achieving a reasonable gas ﬂux, and similar membranes were
eventually utilized in gas separation.27

Many other membranes based on phase inversion have been developed since the
seminal work of Loeb and Souriraj an.”30 Pinnau and Koros developed a dry-wet phase-
inversion method for formation of integrally skinned asymmetric membranes that consist
of ultrathin skin layers supported by microporous substrates.29 Studies with these
membranes along with calculations showed that the gas ﬂux through the substrate
(without the skin layer) should be ~10 times higher than that after addition of the skin

layer to achieve ~90 % of the gas selectivity of the skin-layer material. In many cases,

the resistance of the skin layer can predominantly determine gas transport properties of

. 9,
asymmetric membranes.2 3‘

Many types of asymmetric membranes have been prepared
with skin layers ranging from molecular sieves to polyimides.“32 Suda et al. reported
fabrication of asymmetric molecular sieve membranes prepared from polyamic acid
membranes via imidization and pyrolysis. The pyrolysis at high temperatures and low
heating rates produced the promising gas-selective molecular sieve membranes.”

A large number of studies also demonstrate the formation of composite
membranes with ultrathin, selective skins.”35 These systems are attractive because only
small amounts of the skin material are employed, and thus relatively expensive skin
materials can be used. Several studies show that skins with thicknesses of about 50 nm
can indeed form the selective skin of a composite membrane. For example, Martin and
coworkers employed interfacial polymerization to form poly(aniline) ﬁlms on the surface
of microporous supports. Even though the ﬁhns were only 50 nm thick, they exhibited
an Oz/Nz selectivity coefﬁcient of 8.34 Regen et al. even utilized a single monolayer to
form the selective skin of a composite membrane. These membranes were fabricated
using poly[1-(trimethylsilyl)-1~propyne] (PTMSP) as a support material, a-
cyclodextrin/sodium pheonoxide as a gutter layer, and a single monolayer of a modiﬁed
calix[6]-arene as the primary barrier for transport. This preparation resulted in He/Nz
selectivity values as high as 70.35

Industrial applications of polymer membranes in gas separation began in the
1970s at Monsanto. Researchers at Monsanto overcoated asymmetric polysulfone
membranes with silicon rubber and showed that the coating could plug defects in the
ultrathin, selective layer without decreasing the permeability or selectivity of the skin.

This discovery enabled

installation of the ﬁrst large scale gas-separation membrane (Prism®) for the recovery of
H2 from ammonia streams.

In the 19805, Perrnea introduced their Prism Alpha® membrane for air separation,
which was based on asymmetric membranes of bisphenol-A polysulfone (Figure 1.3).
General Electric later used ultrathin (< 500 A) silicon/polycarbonate (Figure 1.4)

membrane skins for treatment of natural gas to remove impurities such as carbon dioxide,

............

Figure 1.3 Bisphenol A polysulfone

1
I

 

 

hydrogen sulﬁde and water. Carbon dioxide and hydrogen sulﬁde are both corrosive to
pipelines, and hydrogen sulﬁde is very toxic. Moisture must be removed to prevent
corrosion and line blackage due to its freezing. At about the same time, Generon used
poly(4-methyl-1-pentene) membranes with an Oz/Nz selectivity of about 4.

In the 19903, Praxair

O
l |
and Medal produced polyimide OQQ C H 3)2 «QC ‘ C }
n

membranes with Oz/Nz
selectivities of 6 to 8. These Figure 1'4 Polycarbonate
membranes could produce 99 % pure nitrogen and gave a cost-competitive alternative to
delivered liquid nitrogen for many small users.36

As the above applications show, many successes have been achieved in

commercial gas separations. However, on both the commercial and research scale,

synthesis of practical membrane skins with thicknesses less than 50 nm is still a

challenge. Development of such membranes should enhance ﬂux and increase the
versatility of membrane-based separations. We felt that hyperbranched polymers might
address this challenge and provide a versatile, derivatizable material for specialized gas
or liquid separations.
1.2.1 The Solution-Diffusion Mechanism of Gas Transport

Gas separation using nonporous, dense polymeric membranes depends on
differences in permeation among the gases of interest, and the mechanism of permeation
through such membranes is generally described by the solution-diffusion model.37“40
According to this mechanism (Figure 1.5), permeation of gases proceeds in a three-stage
sequence: 1) sorption of the gas at the feed-side membrane surface, 2) diffusion through
the membrane, and 3) desorption at the permeate-side membrane surface. Separation of
gases is thus based on both solubility and diffusivity factors.
The gas-transport properties of a membrane are expressed by two characteristic
parameters: 1) the permeability coefﬁcient, P, which is equal to the product of the
solubility coefﬁcient (S) and the diffusivity (D) for a speciﬁc gas, and 2) the separation

factor, a, which is a measure of selectivity of the membrane for one gas over another.14

An understanding of these parameters begins with Fick’s ﬁrst law, which describes gas
. . . . AC . . 41
diffusron due to a concentration gradient, A—X_ , as shown 1n equation 1.1.

J=—Dég 1.1
AX

In this equation, J is the ﬂux of a speciﬁc molecule through the membrane, D is the

diffusion coefﬁcient of that molecule in the membrane, AC is the concentration

Feed phase Dense Permeate phase
membrane

   
  

 

1 Desorption

Diffusion

4—->
AX

Figure 1.6 Gas transport through a dense membrane according to the solution-
diffusion model

difference across the membrane, and AX is the membrane thickness. This equation can
often be restated as equation 1.2, where C, is the concentration just inside the membrane
at the feed side and C, is the concentration in the membrane at the permeate side. The
equilibrium concentration C of a speciﬁc component at the feed or permeate side of the
polymer membrane is often related to its partial pressure, p, and solubility coefﬁcient, S,

according to Henry’s Law as shown in equation 1.3.

 

J:D(Co—C,) 1.2
AX
Substitution of equation 1.3 into equation 1.2 gives equation 1.4,
C=Sxp 13
J=DS(p°_”’) 1.4
AX

where p0 is the partial pressure of the gas on the feed side of the membrane and p, is the
partial pressure of the gas on the permeate side of the membrane. The product DS is
called the permeability, P, and can be expressed for a speciﬁc gas according to equation

1.5.

JAX
po_pl

 

P=DxS= 1.5

The ideal selectivity or of the membrane is deﬁned as the ratio of the individual
perrneabilities of a gas pair (a and b) as shown in equation 1.6.

P D S
crab: “z ”x “ 1.6
Pb Db Sb

 

 

 

10

(I (I

and
b 5.

 

The ratios

 

are known as the “diffusivity selectivity” and the “solubility

selectivity”, respectively. “Diffusivity selectivity” favors transport of the smallest

molecule while “solubility selectivity” generally favors the most condensable gas.

1.3 Hyperbranched Polymers
Hyperbranched polymers are highly branched macromolecules that provide new
3-dimensional molecular structures that may yield interesting properties for applications

4 ,4 - 4
2 3 enhancement of adhesron, 4 entrapment of

such as surface modiﬁcation,
enzymes,45 and chemical sensing.46 These polymers contain an architecture where each
subsequent layer contains more polymer chains than the preceding layers, and thus they
generally contain an open interior and a relatively crowded periphery. The unique tree-
like structure of these molecules could provide at least two potential advantages in
membrane formation. First, ﬁlms may be more capable of covering a porous structure
than traditional linear polymers. Second, the open interior of hyperbranched molecules
such as dendrimers should provide unique adsorption and difﬁision properties. Below,
We describe below the speciﬁc properties of the hyperbranched polymers employed in
this thesis, hyperbranched poly(acry1ic acid) and poly(amidoamine) dendrimers.
1.3.1 Hyperbranched Poly(acrylic acid) Films

Hyperbranched PAA ﬁhns are unique candidate for forming ultrathin membranes
because their thickness can be controlled over a wide range (2 to 100 nm) by varying the
number of surface-grafted PAA layers, and because the carboxylate groups of PAA can

be readily derivatized. Using the graﬁ-on-graft method, Zhou and co-workers ﬁrst

synthesized hyperbranched PAA ﬁlms on a gold surface.42 In this synthetic procedure,

11

gold slides are ﬁrst immersed in 0.001 M mercaptoundecanoic acid (MUA) in ethanol
(EtOH) to form a self-assembled monolayer (SAM). Activation of the carboxylic acid
groups to a mixed anhydride followed by reaction with an u,m-diamino poly(tert-
butylacrylate) (HzNR(PTBA)RNH2, R = ((CH2)2NHCO(CH2)2C(CN)(CH3)) yields a
grafted layer of PTBA that is speciﬁcally synthesized to produce hyperbranched PAA.
Hydrolysis of the tert—butyl ester groups of PTBA then yields one grafted PAA layer as
shown in Figure 1.6.47 Repeated grafting at multiple —COOH sites on each previous graft
leads to hyperbranched PAA ﬁlms that possess a high density of reactive -COOH
functional groups. These functional groups can be derivatized subsequently to introduce
speciﬁc functionalities into the ﬁlm. The thickness of these ﬁlms grows nonlinearly with
the number of layers, presumably because each subsequent layer contains more grafts.
This procedure allows a 50 mm ﬁlm to be synthesized with about 5 grafting steps.

Although this dissertation focuses on the use of hyperbranched PAA for forming
ultrathin, selective membranes, several studies show a variety of potential applications
for these ﬁlms. Lackowski and Crooks reported micron-scale patterning of surfaces
using a combination of micro-contact printing of self-assembled monolayers and polymer
grafting.48 The three-dimensional nature of the pattern and the ﬁmctionality of the PAA
ﬁlms yielded segregated, chemically sensitive interfaces and corrals for isolating cells.48

Further demonstrating the versatility of hyperbranched PAA, Dermody and co-
workers46 prepared gold electrodes coated with hyperbranched PAA that was derivatized
with B-cyclodextn'n receptors, and capped with chemically grafted, ultrathin polyamine
ﬁlter layers. Such ﬁlms can serve as unique, selective electrochemical sensors.

Speciﬁcally, a thin, grafted poly(amidoamine) or poly-D-lysine surface layer served as a

12

Number of
PAA
Layers

 

 

1) 000,3
a—co, H N methy' m°m“°""e *Z—CONHR-PTBA-RNHR) p’TSOH'HZO>

2) HZNR-PTBA-RNH2

1 j»CONHR-PAA-RNH,

Step 1), then steps 2) and 3)

2 ﬁ—CONHR—{CH2CI3HJn—ECH2CfH)m—zRNHCOZCHZCHg
2

//c\ co,H

Nl-lR-PAA-RNH,

Step 1), then steps 2) and 3)

n2 R2 "‘2' 2

co,H o/ \
NHR PAA-RNH,
O\//C

\,NHR-[CH ,ﬁJH]n—[CH, ZCfH}—RNHCOZ CH,CH3

co, H
O/ C\NHR-PAA-RNH,

Figure 1.6 Synthesis of hyperbranched PAA ﬁlms.47

13

pH-sensitive molecular ﬁlter that allowed selective passage of charged analytes. These
studies demonstrate the versatility of hyperbranched PAA, and this versatility could
certainly prove useful in membrane development. For example, deposition of speciﬁc
surface layers could decrease fouling in reverse osmosis or nanoﬁltration membranes.
1.3.2 Hyperbranched Poly(amidoamine) Dendrimers

In contrast to hyperbranched PAA, dendrimers are nearly monodisperse polymers
that posess a relatively well—deﬁned molecular architecture.43’49’55 Due to this unique
geometry and functionality, many studies have investigated the synthesis, structure, and
physical properties of dendrimers. Generally, the synthesis of dendrimers occurs in a
step-wise fashion from simple, branched monomer units. There are two synthetic
strategies: divergent and convergent. In divergent approaches, the dendritic molecule is
built up layer-by-layer from the core to the periphery, while in convergent syntheses,
dendrimer components are synthesized initially and then coupled to the dendrimer core.
Poly(amidoamine) (PAMAM) dendrimers are perhaps the most common dendritic
family, and these molecules are prepared by a divergent scheme as shown in Figure 1.7.
In this synthetic approach, each cycle of Michael addition of methyl acrylate (MA) to
amines and amidation of the resulting methyl esters with ethylenediamine (EDA) leads to
the addition of one more layer of branches, called a generation. Therefore, the generation
number of the dendrimer is equal to the number of cycles of Michael addition and
amidation performed.49'51’53

One of the unique properties of dendrimers is their ability to serve as molecular
“boxes”. Hydrophobic binding,56 hydrogen bonding,5 7metal-1i gand coordination,58 and

phys1cal encapsualtron5 can occur in the dendrimer interior, whereas electrostatic

l4

+EDA CH3OH
HZNTQN O o HNHz
«A fj—NH
NXN
HN-(J ‘—>_H
H NH O O NH
2 NH2
1-
CH30H
+EDA
CH3OH
NH2 H2N
HN /
<0 0 NH
HszNm 3 H_/-NH2
N NM
0 ‘ﬁ 0 N 0
”MK 0 N’i-T
NXNJ
AM: Wit
Nit/N ° “ °
Hsz D ﬂxNHz
H? o 9 NH
NH, tun.

Figure 1.7 Synthesis of PAMAM Dendrimers 77

15

52'60'63 and multidentate coordination 6" can take place at the dendrimer

interactions
surfaces. These properties may dramatically affect the performance of membranes
containing these materials.

Formation of membranes with dendrimers will require deposition of dendrimer-
containing ﬁlms, and several methods exist for doing this. Wells and Crooks modiﬁed
MUA-SAMs with amine-terminated poly(amidoamine) PAMAM dendrimers through
amide linkages."5 The thickness of these dendrimer monolayers was considerably smaller
than the diameter of the dendritic macromolecules in solution, suggesting that dendrimers
are compressed along the surface normal and ﬂattened."“”52 Tsukruk reported fabrication
of self-assembled dendrimer monolayers by electrostatic deposition, and the thicknesses
of these monolayers were also much smaller than the diameter of dendrimers in
solution.52 Regen and Watanabe employed amine-terminated PAMAM dendrimers and
KthCl4 to construct multilayers by repeated coordination bonding."6
1.3.3 Dendrimers in Alternating Polyelectrolyte Deposition

The technique for formation of dendrimer-containing ﬁhns of most relevance to
this study is alternating polyelectroyte deposition. This method is a layer-by-layer
process that begins with dipping of a charged substrate into a solution containing an
oppositely charged polyelectrolyte. After rinsing with deionized water to remove
polyelectroyte that is not strongly bound to the substrate, immersion in a solution
containing an oppositely charged polyelectrolyte results in adsorption of a second layer

on the surface. Upon adsorption of each polyelectroyte, there is charge

overcompensation on the surface so that subsequent polymer layers of opposite charge

16

will electrostatically bind to the surface."7 Repetition of the cycle of polyanion and
polycation adsorption can continue until the desired numbers of layers are deposited.

The structure and composition of multilayered polyelectrolyte ﬁlms will have a
large effect on their properties. Interestingly, although alternating polyelectrolyte
deposition occurs via sequential adsorption, multilayer ﬁlms are not stratiﬁed into well-
deﬁned layers. Except at the surface of the ﬁlm, polycations and polyanions are
interdigitated to maximize electrostatic interactions,"8 and charges are nearly completely
compensated by neighboring polyelctrolytes in these polycation/polyanion complexes.
Recent experiments on polyelectrolyte pairs, such as poly(diallydimethylammonium
chloride) (PDADMA)/poly(styrene sulfonate) (PSS) ﬁlms showed essentially complete
charge compensation within the bulk of the ﬁhns, i.e., no external ions were present in
the ﬁlms to maintain electrical neutrality. However, excess charge is present at the ﬁlm
surface, and this is essential for ﬁlm formation.69

In chapter 3 of this dissertation, we examine the pH-dependent grth and
morphology of multilayer dendrimer/PAA ﬁlms prepared by alternating polyelectrolyte
deposition. Because dendrimers provide polycations with a unique geometry, ﬁlms
formed from these macromolecules have unusual properties and thicknesses. Dendrimer-
containing ﬁlms can be signiﬁcantly thicker than ﬁlms prepared from linear polymers,
and the permeability of dendrimer/PAA membranes to gases is extremely high.
Additionally, ﬁlm thickness can be controlled by varying the number of generations in

the constituent dendrimers.

l7

1.3.4 Dendrimers in Adhesion

While working with dendrimer ﬁlms and membranes, we began to consider their
potential for use as adhesives. The multidentate structure of these ﬁlms may make them
particularly well-suited for binding two substrates together. Below, we give some
background on organic adhesives and the use of dendrimers in this area.

The most commonly used organic adhesives include cyanoacrylates, anaerobic
adhesives such as methacrylate esters, and epoxies. a-cyanoacrylate esters polymerize
rapidly at room temperature in the presence of surface moisture and then cure to form
strong bonds with the substrate.70 Anaerobic adhesives are based on methacrylate esters
that are part of a curing system that remains liquid in a plastic bottle that transmits
oxygen. The oxygen acts as a radical scavenger and prevents polymerization. In the
assembly of metal parts, oxygen is eliminated from the joint, and polymerization
occurs.70 Epoxy resins are formed from low molar mass prepolymers containing epoxide
end-groups. The diglycidyl ether prepolymers, prepared by reaction of excess
epicholrohydrin with bisphenol-A in the presence of a base, usually cured through the use
of multi-functional amines that undergo a polyaddition reaction with the terminal epoxide
groups.71

The possible mechanisms72 for adhesion include mechanical bonding, chemical
bonding and thermodynamic bonding. In mechanical bonding, the two adhering surfaces
have surface irregularities that can act as mechanical anchors. Most substrates are rough
on the microscopic scale, and the extent of mechanical interlocking is dependent on the
rheological properties of the adhesive. Strong mechanical bonding can be obtained on

substrates such as wood, paper, and textiles. In chemical bonding, covalent bonds are the

18

primary form of chemical interactions between adhesives and substrates. Secondary
interactions involve nonpolar dispersion forces W an der Waals forces), polar dipole
interactions, and polar Lewis acid/base interactions. For thermodynamic bonding
interaction of the adhesive and substrate should result in a low-energy interface.
Generally, a low surface energy adhesive will spread on a higher surface energy substrate
to reduce the total surface energy of the system.

In addition to the commercially available organic adhesives mentioned above,
amine-terminated dendrimer monolayers are useful as adhesion promoters between
vapor—deposited Au ﬁlms and Si-based substrates.44 A recent peel-test study evaluated
the strength of adhesion between gold and glass using G2, G4, G6, and G8 PAMAM
dendrimer adhesive layers. The G8 adhesion layer gave the strongest connection
between the substrate and the Au ﬁlm because of its large diameter and many surface
functionalities. Scanning tunneling microscope images showed that the presence of the
dendrimer adhesion layer had no signiﬁcant impact on the nanostructure of the Au ﬁlm.

In our work, we bond two gold-coated silicon wafers using a G4 dendrimer
adhesion layer. We pursued these studies because such ﬁlms may eventually provide a
gentle method of wafer bonding, which is important in the fabrication of micro-electro-
mechanical system and lab-on-a-chip devices.”74

There are currently a number of different methods available for silicon wafer-to-
wafer bonding, but most of them require relatively harsh conditions.”75 Bonding
techniques performed in microstructure fabrication can be divided into three categories.
1) direct bonding: the silicon wafers are directly contacted without an interrnediating

layer. The contacted wafers are heated to 800 °C, and Si-O-Si bridging bonds are

19

formed between the surfaces as water molecules are liberated. To obtain bond strengths
similar to those in silicon crystals, the heating temperature is increased to 1,000 °C. This
method relies on forces that naturally attract surfaces together when they are very smooth
and ﬂat. Application of this method is often challenging because of voids between
wafers that are created by particles, trapped air or contamination”75 2) Anodic bonding:
the bond is formed by applying an electric potential of ZOO-1,000 V at a temperature of
300 - 450 °C. The bonding is typically performed between a sodium-bearing glass wafer
and a silicon wafer. One has to ensure that the glass has a good thermal coefﬁcient of
expansion ﬁt to silicon.76 3) Intermediate-layer bonding: eutectic bonding involves the
deposition of intermediate metallic ﬁhns prior to formation of the bond. An alloy is
formed between the wafers by solid-liquid inter-diffusion at the contact interface. Upon
cooling the alloy solidiﬁes.

In chapter 4 of this dissertation, we present preliminary results that show very
strong (>1.6 MPa) adhesion between wafers bonded with dendrimer/PAA layers. This
binding method occurs at room temperature and does not require the use of any organic

solvents.

1.4 Structure of this Dissertation

This chapter of the dissertation describes relevant previous research with
membranes and hyperbranched polymeric materials. In the ﬁrst part, various membranes
are presented and compared on the basis of synthetic types. Asymmetric or composite
membranes for gas separation, and the mechanism of gas permeation are described. In

the second part, the chemical and physical properties of hyperbranched PAA and

20

poly(amidoamine) dendrimers are discussed. Additionally, adhesion is discussed with
respect to dendrimer layers.

In Chapter 2, work with composite membranes composed of hyperbranched
poly(acry1ic acid) grafted on porous alumina substrates is presented. Using F ESEM and
AFM, we found that hyperbranched PAA forms defect-free membranes and effectively
covers alumina supports without ﬁlling underlying pores. Gas-transport measurements
with PAA before and after derivatization with H2NCH2(CF2)6CF3 are described and
discussed.

Chapter 3 details our results on the pH-dependent growth of multilayer ﬁlms
prepared from PAMAM dendrimers and PAA by alternating polyelectrolyte deposition.
We show that by systematically varying the pH of deposition solutions and the dendrimer
generation, ﬁlm thickness can be varied over several orders of magnitude. The structures
of generation 4 and generation 8 dendrimers as well as solution pH have a profound
effect on electrostatic adsorption. Gas permeability measurements with these dendrimer
ﬁlms show that they have a very open structure.

Chapter 4 discusses the use of these dendrimer/PAA ﬁlms in wafer-to-wafer
bonding. We inserted a positively charged G4 dendrimer layer between negatively
charged PAA layers on each of the two wafers. Mechanical testing revealed that the
tensile strength of wafers bonded using the dendrimer/PAA system was >1.6 MPa.
Finally, chapter 5 presents general conclusions and future directions in using
hyperbranched PAA and dendrimers as membranes, as well as the continued

characterization of adhesion resulting from dendrimer/PAA ﬁlms.

21

1.5 References

(1)

(2)
(3)

(4)

(5)

(6)
(7)

(8)
(9)
(10)

(11)

(12)

(13)

(14)

(15)
(16)
(17)

(18)

Lee, J. D.; Lee, S. H.; Jo, H.; Park, P. K.; Lee, C. H.; Kwak, J. W. Environ. Sci.
Techno]. 2000, 34, 3780-3788.

Zhang, M.; Song, L. Environ. Sci. T echnol. 2000, 34, 3767-3773.

Merchese, J.; Ochoa, N. A.; Pagliero, C.; Almandoz, C. Environ. Sci. T echnol.
2000, 34, 2990-2996.

Schuster, B.; Pum, D.; Sara, M.; Braha, O.; Baylay, H.; Sleytr, U. B. Langmuir
2001, 34, 499-503.

Alvarez, S.; Riera, F. A.; Alvarez, R.; Coca, J. Ind. Eng. Chem. Res. 2001, 40,
4925-4934.

Kwak, S. Y.; Kim, S. S.; Kim, S. H. Environ. Sci. Technol. 2001, 35, 2388-2394.

Robeson, L. M.; Burgoyne, W. F.; Langsam, M.; SaVoca, A. C.; Tien, C. F.
Polymer 1994, 35, 4970-4978.

Robeson, L. M. J. Membr. Sci. 1991, 62, 165-185.
Freeman, B. D. Macromolecules 1999, 32, 375-380.
Neel, J.; Aptel, P.; Clement, R. Desalination 1985, 53, 297-326.

Itoh, T.; Toys, H.; Ishihara, K.; Shimihara, I. J. Appl. Polym. Sci., 1985, 30, 179-
188.

Brun, J. P.; Larchet, C.; Melet, M.; Bulvestre, G. J. Membr. Sci. 1985, 23, 257-
283.

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

Mulder, M. Basic Principles of Membrane Technology; lst ed.; Kluwer Academic
Publishers: Dordrecht, 1996, pp 7-21.

Henis, J. M. S.; Tripodi, M. K. Science 1983, 220, 11-17.
Haggin, J. Chem. Eng. News 1990, 68, 22-23.
Abelson, P. H. Science 1989, 244, 1421.

Spillman, R. W. Chem. Eng. News 1989, 68, 11-17.

22

(19)

(20)

(21)

(22)
(23)

(24)

(25)
(26)
(27)

(28)

(29)

(30)

(31)

(32)

(33)
(34)

(35)

(36)

Nagale, M.; Kim, B. Y.; Bruening, M. L. J. Am. Chem. Soc. 2000, 122, 11670-
11678.

Lvov, Y.; Decher, G.; Sukhorukov, G. Macromolecules 1993, 26, 5396-5399.
Suda, H.; Haraya, K. In Membrane Formation and Modiﬁcation; Pinnau, I.,
Freeman, B. D., Eds; ACS Symposium Series 744, American Chemical Society:
Washington, DC, 2000, pp 295-313.

Graham, T. Philos. Mag. 1866, 32, 401-420.

Mazur, W. H.; Chan, M. C. Chem. Eng. Prog. 1982, 78, 38-40.

Matson, S. L.; Ward, W. J .; Kimura, S. G.; Browall, W. R. J. Membr. Sci. 1986,
29, 79-96.

Spillman, R. W. Chem. Eng. Prog. 1989, 85, 41-62.
Loeb, S.; Sourirajan, S. Adv. Chem. Ser. 1962, 38, 117.
Gantzel, P. K.; Merten, U. 1&EC. Proc. Des. Dev. 1970, 9, 331.

Reuvers, A. J .; van den Berg, J. W. A.; Smith, C. A. J. Membr. Sci. 1987, 34, 45-
86.

Pinnau, I.; Koros, W. J. Ind. Eng. Chem. Res. 1991, 30, 1837-1840.

Kesting, R. E. In Materials Science of Synthetic Membranes; Lloyd, D. R., Eds.;
ACS Symposium Series 269, American Chemical Society: Washington, DC,
1985, pp 131-163.

Strathmann, H.; Scheible, R; Baker, R. W. J. Appl. Polym. Sci, 1971, 15, 811-
828.

Kim, T. H.; Koros, W. J .; Husk, G. R.; O'Brien, K. C. J. Membr. Sci. 1988, 37,
45-62.

Henis, J. M. S.; Tripodi, M. K. J. Membr. Sci. 1981, 8, 233-246.
Liu, C.; Martin, C. R. Nature 1991, 352, 50-52.

Zhang, L. H.; Hendel, R. A.; Cozzi, P. G.; Regen, S. L. J. Am. Chem. Soc. 1999,
121,1621-1622.

Baker, R. W. Ind. Eng. Chem. Res. 2002, 41, 1393-1411.

23

(37)

(38)
(39)
(40)
(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

(50)

(51)
(52)

(53)

(54)

Frisch, H. L.; Stem, S. A. CRC Crit. Rev. Solid State Mater. Sci. 1983, 11, 123-
187.

Stem, S. A.; Frisch, H. L. Annu. Rev. Mater. Sci. 1981, 11, 523-550.
Crank, J. The Mathematics of Diffusion; 2nd ed.; Clarendon: Oxford, 1975.
Matson, S. L.; Lopez, J .; Quin, J. A. Chem. Eng. Sci. 1983, 38, 503-524.
Pick, A. Ann. Phys. Chem. 1955, 94, 59.

Bruening, M. L.; Zhou, Y.; Aguilar, G.; Agee, R.; Bergbreiter, D. E.; Crooks, R.
M. Langmuir 1997, 13, 770-778.

Fréchet, J. M. J. Science 1994, 263, 1710-1715.

Baker, L. A.; Zamborini, F. P.; Sun, L.; Crooks, R. M. Anal. Chem. 1999, 71,
4403-4406.

F ranchina, J. G.; Lackowski, W. M.; Dermody, D. L.; Crooks, R. M.; Bergbreiter,
D. E. Anal. Chem. 1999, 71, 3133-3139.

Dermody, D. L.; Peez, R. F.; Bergbreiter, D. E.; Crooks, R. M. Langmuir 1999,
15, 885-890.

Zhou, Y.; Bruening, M. L.; Bergbreiter, D. E.; Crooks, R. M.; Wells, M. J. Am.
Chem. Soc. 1996, 118, 3773-3774.

Lackowski, W.; Ghosh, P.; Crooks, R. M. J. Am. Chem. Soc. 1999, 121 , 1419-
1420.

Tomalia, D. A.; Naylor, A. M.; Goddard, III. W. A. Angew. Chem. Int. Ed. Engl
1990, 29, 138-140.

Newkome, G. R.; Moorﬁeld, C. N.; Baker, G. R.; Saunders. M. J .; Grossman, S.
H. Angew. Chem, Int. Ed. Eng. 1991, 30, 1178-1180.

Zeng, F.; Zimmerman, S. C. Chem. Rev. 1997, 97, 1681-1712.
Tsukruk, V. V. Adv. Mater 1998, 10, 253-257.

Tomalia, D. A.; Hedstrand, D. M.; Ferritte, M. S. Macromolecules 1991, 24,
1435-1438.

Saville, P. M.; Reynolds, P. A.; White, W. W.; Hawker, C. J .; Fréchet, J. M. J .;
Wooley, K. L.; Penfold, J.; Webster, J. R. P. J. Phys. Chem. 1995, 99, 8283-8289.

24

(55)

(56)

(57)

(58)

(59)

(60)

(61)
(62)
(63)
(64)
(65)
(66)
(67)
(68)
(69)

(70)

(71)

(72)

Alonso, B.; Moran, M.; Casado, C. M.; Lobete, F.; Losada, J.; Cuadurado, I.
Chem. Mater. 1995, 7, 1440-1442.

Naylor, A. M.; Goddard, 111. W. A.; Kiefer, G. E.; Tomalia, D. A. J. Am. Chem.
Soc. 1989, 111, 2339-2341.

Newkome, G. R.; Guther, R.; Mooreﬁeld, C. N.; Cardullo, F .; Echegoyen, L.;
Perezcordero, E.; Luftmann, H. Angew. Chem, Int. Ed. Eng. 1995, 34, 2023-
2026.

Sttodart, J. F.; Philip, D. Angew. Chem, Int. Ed. Engl. 1996, 35, 1154-1156.

Jansen, J. F . G. A.; de Brabander-Van Den Berg, E. M. M.; Meijer, E. W. Science
1994, 266, 1226-1229.

He, J. A.; Valluzzi, R.; Yang, K.; Dolukhanyan, T.; Sung, C. M.; Kumar, J.;
Tripathy, S. K.; Samuelson, L.; Balogh, L.; Tomalia, D. A. Chem. Mater. 1999,
11, 3268-3274.

Khopade, A. J.; Caruso, F. Nano. Lett. 2002, 2, 415-418.

Cheng, L.; Cox, J. A. Electrochem. Commun. 2001, 3, 285-289.

Yoon, H. C.; Kim, H. S. Anal. Chem. 2000, 72, 922-926.

Posnett, D. N.; Mcgrath, H.; Tam, J. P. J. Biol. Chem. 1988, 263, 1719-1725.
Wells, M.; Crooks, R. M. J. Am. Chem. Soc. 1996, 118, 3998-3999.
Watanabe, S.; Regen, S. L. J. Am. Chem. Soc. 1994, 116, 8855-8856.
Schlenoff, J. B.; Dubas, S. T. Macromolecules 2001, 34, 592-598.

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

Schlenoff, J. B.; Ly, H.; Li, M. J. Am. Chem. Soc. 1998, 120, 7626-7634.

Lee, L. H. Adhesion Science and Technology; Plenum Press: New York and
London, 1975; Vol. 98; pp 467-475.

Young, R. J.; Lovell, P. A. Introduction to Polymers; 2nd Edition ed.; Chapman
& Hall: London, 1992.

Kalantar, J .; MS Thesis; Michigan State University: East Lansing, MI, 1988.

25

(73)

(74)
(75)

(76)

(77)

Baldwin, R. P.; Roussel, T. J .; Crain, M. M.; Bathlagunda, V.; Jackson, D. J .;
Gullapalli, J.; Conklin, J. A.; Rekha, P.; Naber, J. F.; Walsh, K. M.; Keynton, R.
S. Anal. Chem. 2002, 74, 3690-3697.

Schmidt, M. A. Proceedings of the IEEE. 1998, 86, 1575-1585.

Bengtsson, S.; Amirfeiz, P. J. Electronic Mater. 2000, 29, 909-915.

Choi, W. B.; Ju, B. K.; Lee, Y. H.; Haskard, M. R.; Oh, M. H. J. Vac. Sci.
Techno]. B 1997, 15, 477-481.

Uppuluri, S.; Keinath, S. E.; Tomalia, D. A.; Dvomic, P. R. Macromolecules
1998, 31, 4498-4510

26

Chapter 2

ULTRATHIN, HYPERBRANCHED POLY(ACRYLIC ACID)
MEMBRANES ON POROUS ALUMINA SUPPORTS“

Summary

The synthesis of hi gh-ﬂux composite membranes requires deposition of an
ultrathin, discriminating layer on a highly permeable support. This chapter describes
synthesis, derivatization, and characterization of hyperbranched poly(acry1ic acid) (PAA)
membranes on porous alumina supports. PAA ﬁlms as thin as ~ 40 nm effectively cover
underlying pores without ﬁlling them, presumably because of their hyperbranched
structure. Synthesis of these ﬁlms begins by sputtering a thin gold layer on the alumina
support and then grafting a layer of PAA to a self-assembled monolayer of
mercaptoundecanoic acid on the gold. Graft-on-graft deposition of PAA yields the
hyperbranched membrane. FESEM (ﬁeld-emission scanning electron microscopy) and
AFM (atomic force microscopy) images clearly show that hyperbranched PAA ﬁlms can
cover the substrate surface completely without ﬁlling underlying pores, thus creating an
ultrathin membrane on a porous support. PAA membranes are especially attractive
because derivatization permits control over transport properties. Gas-transport studies
indicate that 3-layer PAA ﬁlms do not show a high selectivity by themselves, but
selectivity improves signiﬁcantly after covalent derivatization of PAA with

HzNCH2(CFz)6CF3.

 

” This chapter is taken from a publication by Milind Nagale, Bo Young Kim, and Merlin L. Bruening (J.
Am. Chem. Soc. 2000, 122, 11670-11678), and is a joint work of the three authors.

27

2.1 Introduction

Although membrane separations are attractive because of low energy costs and
simple operation, low perrneabilities and/or selectivities often limit membrane
applicationsl'4 Successful commercial applications of membrane separations already
include production of nitrogen from air and recovery of hydrogen from mixtures with
other larger components like nitrogen, methane, and carbon monoxide.5 Improvements
in throughput and selectivity, however, could greatly increase the impact of membrane
separations.6

One major limitation of gas-separation membranes is that selective materials are
generally not highly permeable.7'9 This tradeoff between selectivity and permeability
makes decreasing membrane thickness vital for increasing ﬂux without sacriﬁcing
selectivity-1,10,“ As shown in Equation 2.1, gas ﬂux through a membrane (F) is
inversely proportional to membrane thickness (1) and directly proportional to the pressure

gradient across the membrane (Ap) and the permeability coefﬁcient (P) for a speciﬁc

I?

a.

gas.
F = —— 2.1

The mechanical weakness of ultrathin ﬁlms necessitates their deposition on highly

permeable supports that provide strength. Unfortunately, synthesis of defect-free,

ultrathin (<50 nm) membrane “skins” is challenging.“ 1'13

There are several approaches to preparing membranes with ultrathin skins. One
strategy involves phase-inversion to prepare an asymmetric membrane from a single

material. The ﬁrst asymmetric gas separation membrane was a Loeb-Sourirajan-type

28

cellulose acetate (CA) material formed by drying CA membranes using quick-freezing

. . l4 . . .
and vacuum sublrmatron at —10 °C. More recent efforts include synthesrs of integrally

. . . . r3 ,
skinned membranes usmg a dry/wet-phase 1nversron process. Asymmetric carbon
molecular sieve membranes can even be prepared by formation of a capillary type

polyamic acid membrane followed by imidization and pyrolysis to form a dense selective

15
layer supported by a porous layer.

Composite membranes provide a versatile alternative to asymmetric skinned
systems. In this case a polymer “skin” is deposited on a separate, highly permeable
support. Composite membranes have the advantage that only a small amount of the

possibly expensive “skin” material needs to be used. Strategies for forming the “skins”
. . 16.17 . . . 18 . . .
of these membranes include casting, 1n srtu castmg, 1n Situ condensatron of

polymers and/or monomers at the porous support,19 and plasma polymerization.20 In all
of these methods, synthesis of defect-free “skins” with thicknesses < 50 nm is difﬁcult.
A few recent reports demonstrate that composite membranes with ultrathin
“skins” are capable of selective separations. With the thinnest synthetic gas-separation
membranes to date, Regen and co-workers showed that Langmuir-Blodgett ﬁhns as thin

as a single monolayer can serve as selective membranes when deposited on a continuous
21-23 , . . . . .
glassy polymer support. L1u and Martin reported fabrication of ultrathin, conducting

r
polymer membranes (~50 nm thick) with an Oz/Nz selectivity of 8.l In this case
interfacial photopolymerization at the alumina surface resulted in an ultrathin conducting

polymer membrane. Similar selectivities can be achieved by electrochemical synthesis of

. . 24 . .
ultrathin ﬁlms on porous alumina membranes. Thin ﬁlm composrte membranes can

also be formed on hollow ﬁbers. Paul et a]. formed multilayered composite hollow-ﬁber

29

membranes by dip-coating hollow-ﬁber polysulfone supports with a selective poly(4-
vinylpyridine) layer (50-150 mm thick) and sealing with a silicone rubber layer. This
system has a selectivity of 7 for Oz/Nz.25

We report fabrication of surface-grafted, hyperbranched poly(acry1ic acid) (PAA)
membranes. These ﬁlms are unique in that their highly branched structure allows them to

cover underlying pores without ﬁlling them, as shown schematically in Figure 2.1.215

<— 3rd layer

<— 2nd layer
<— 1st layer

0 is an amide bond

 

Solid Substrate

Figure 2.1 Idealized cartoon of a hyperbranched PAA ﬁlm grafted onto
a porous alumina substrate. Substrate thickness is not drawn to scale.

Additionally, in situ derivatization of these materials allows synthesis of a variety of
membranes without the need to prepare new polymers. The synthesis (Figure 2.2)
involves ﬁrst forming a self-assembled monolayer of a carboxylic acid-terminated thiol
on a gold-coated porous substrate followed by the conversion of carboxylates to mixed
anhydrides. These mixed anhydrides are then used as attachment points for the grafting
of amine-terminated poly(tert-butyl acrylate) (PTBA) onto the surface. The hydrolysis of

tert-butyl groups to carboxylic acid groups forms a PAA layer, and the whole process is

30

~2‘~Z‘~Z‘~Z‘

O O O O
CICOZCHZCH3
N-methyl morpholine

NH2R-PTBA-RNH2

 

 

 

 

w w is» Irv
s s e a
r t s 5
<5 .3 r 5
Q“ Q”
0 $0 5 050 gr p-TsOH-HZO, 55°C g or 5 gr
PAA= tQH-CHzi-n ‘ PTBA: -[-|CH-CH2-]—n
(3:0 C|=O
OH 0+

Figure 2.2 Schematic diagram of the synthesis of 1 layer of PAA on a surface.
Additional layers were prepared by grafting onto previously deposited PAA.

31

then repeated to graft additional layers to the immobilized PAA. The layer-by-layer
synthesis of hyperbranched PAA affords control over ﬁlm thickness, while derivatization
of the ﬁlms provides a way to control transport properties chemically.26 Derivatization
occurs through amidation or esteriﬁcation of the —COOH groups of PAA with functional
amines or alcohols, and a prior study shows that PAA ﬁlms can be modiﬁed to give
ﬂuorescent, hydrophobic, ion-binding, biocompatible, and electroactive ﬁlms.27 Thus,
derivatization can provide a wide variety of hyperbranched membranes for investigating
relationships between transport and membrane chemistry. This paper focuses on
derivatization of hyperbranched PAA with HzNCH2(CF2)6CF3 because this procedure
produces selective gas-separation membranes. The occurrence of gas selectivity shows

that these membranes are free of defects.

2.2 Experimental

Materials. Porous alumina (0.02 and 0.2 um-pore diameter AnoporeTM ﬁlters)
substrates were purchased from Thomas Scientiﬁc. Ethyl chlorofonnate (97%), N-
methyl morpholine (99%), p-toluene sulfonic acid monohydrate (98%), 4,4’-Azobis(4-
cyanovaleric acid) (75%+), 1,4-dioxane (99%), and tert-butyl acrylate (98%) were
purchased from Aldrich. H2NCH2(CF2)6CF3 (97%) was purchased from Lancaster.
Acetonitrile (Spectrum, 99.5%), and ethanol (Pharmco, 100%) were used as received.
DMF (Aldrich, 99.9%) was dried with molecular sieves for 24 h before use, and
deionized water (18.2 MQmm) was obtained using a Milli-Q system.

Mercaptoundecanoic acid (MUA) was initially synthesized with a modiﬁed literature

procedure,28 but most of the MUA used in this research was purchased from Aldrich.

32

a,u)-diamino-terminated poly(tert-butyl acrylate) (H2NR-PTBA-RNH2), R =

(CH2)2NHCO(CH2)2C(CN)(CH3), was prepared according to a literature procedure?“27

He (99.995%), N2 (99.99%), 02 (99.99%), CO2 (99.8%), CH, (99.3%), and SF6 (99.99%)
were purchased from AGA.

Synthesis of PAA ﬁlms. Alumina substrates were cleaned for 10 min in a
UV/ozone cleaner (Boekel Industries, model 135500) after a 10 min immersion in boiling
methanol. Subsequently, gold (63 nm or 5 nm) was sputtered (Ted Pella, Pelco SC-7)
onto the substrates. Sputtering was performed with a current of 20 mA and a pressure of
~ 0.08 mbar (Ar plasma) at a rate of 1 A/s. Gold-coated porous alumina substrates were
cleaned in the UV/ozone cleaner for 12 min before beginning monolayer deposition.

To synthesize PAA ﬁlms, we grafted H2NR-PTBA-RNH2 via amide formation onto a
mercaptoundecanoic acid (MUA) monolayer attached to gold-coated porous alumina.
Hydrolysis of tert-butyl ester groups with p-toluenesulfonic acid results in a grafted PAA
ﬁlm. Repeating these steps using additional grafting at multiple sites on each prior graft

produced layered, hyperbranched polymer membranes. This procedure was described

. 26
prevrously.

Derivatization of 3-layer PAA ﬁlms. After synthesis of 3-layer PAA
membranes and measurement of their gas permeability, these ﬁlms were derivatized with
H2NCH2(CF2)6CF3. To derivatize PAA membranes, the ﬁlm was ﬁrst activated with
ethyl chloroforrnate in the presence of N—methylmorpholine to form mixed anhydrides.
The ﬁlm was rinsed with ethyl acetate, dried with N2, soaked in a 0.1 H2NCH2(CF2)6CF3

solution in DMF for 1 hr, rinsed with ethanol, and dried with nitrogen.29

33

Fourier Transform Infrared Spectroscopy. External reﬂection F TIR spectra
of PAA ﬁlms on substrates coated with 63 nm of gold were acquired using a Nicolet
FTIR spectrometer (MAGNA 560) equipped with a MCT detector and a PIKE grazing
angle accessory (incident angle of 80°, 256 scans at 4 cm'1 resolution). Transmission
FTIR spectroscopy (Mattson Instruments, Inﬁnity Gold) was employed to monitor the
growth of hyperbranched PAA on alumina supports coated with only 5 nm of gold
because these substrates are not sufﬁciently reﬂective for external reﬂection F TIR.

Field-Emission Scanning Electron Microscopy (FESEM). Electron
microscope images of membranes were obtained using a Hitachi S-47OOII ﬁeld-emission
scanning electron microsc0pe. To prepare membrane cross-sections for imaging, samples
were cleaved using tweezers, sputter-coated with 5 nm of gold, rinsed with methanol, and
dried using nitrogen unless otherwise noted. In the cleaving process, a membrane was
held by one pair of tweezers and bent with another pair to break the alumina support.

The polymer support ring around the edge of the alumina was then cut away using a pair

of scissors. During imaging, a low accelerating voltage (4 kV; beam current, 10 uA) was

. . . . . . 30.31
used to minimize the effects of sample chargrng and to provrde better surface detaﬂs.

Atomic Force Microscopy (AFM). AF M images were obtained with a
Nanoscope IIIa (Digital instruments) instrument using the tapping mode (scan rate = 1
Hz). A cantilever having a nominal spring constant of 20-100 N/m was used along with
etched silicon tips. The tips have a nominal radius of curvature of 20-60 nm.
Membranes could be directly imaged by AF M without any special preparation.

Gas-Transport Measurements. Gas-transport measurements were performed

using a permeation cell (MKS Instruments, cell 02910-40) equipped with a pressure relief

34

valve. Permeate ﬂux was measured as a function of inlet pressure (10-40 psig) using a
digital soap bubble ﬂowmeter (Fisher scientiﬁc, Model 420). The area of the membrane
exposed to the gas stream was 2.8 cm2. The perrneabilities of He, N2, O2, CO2, CH4 and
SF (2 were determined for each sample. Three different membranes of each type were
tested with at least three steady state measurements of ﬂux for each gas at a given
pressure. The order in which the permeability of different gases was determined was
deliberately varied, and each sample was tested at least twice for the entire set of gases to
check the stability of the membrane. The cell was purged several times with the gas of
interest at a pressure of 50 psig using the pressure relief valve before performing any
measurements with a particular gas. All measurements were obtained after the ﬂux
reached a steady-state value to ensure complete purging of the cell. After permeability
measurements of membranes coated with PAA layers, samples were removed, analyzed
by transmission FTIR, and ﬂuorinated as described in the synthesis section. Permeability

of ﬂuorinated PAA layers was then measured.

2.3 Results and Discussion

Spectroscopic Characterization of PAA Membranes. Hyperbranched PAA
ﬁlms with up to 6 layers were synthesized on gold-coated porous alumina as shown in
Figure 2.2, and the synthetic procedure was monitored by FTIR spectroscopy. In

agreement with previous syntheses of hyperbranched PAA on gold-coated silicon

wafers,26 external reﬂection FTIR spectra conﬁrmed the steps of the synthetic process,
including formation of a mixed anhydride, attachment of H2NR-PTBA-RNH2 to the

surface, and hydrolysis of tert-butyl ester groups. External reﬂection F TIR spectra of l

35

 

 

 

 

 

 

 

 

 

 

 

Ions I coo"
a:
O
C
a
-E
o J
m
.a
< 6PAA
5PAA
3““ A ‘—
._12AA \: \
l I \
2000 1800 1600 1400 1200
Wavenumber(cm1)

Figure 2.3 External reﬂection F TIR spectra of l, 2, 3, 4, 5, and 6 layers of PAA on
porous alumina (0.02 pm pore diameter) that was sputter coated with 63 nm of gold.

36

to 6 layers of PAA grafted onto gold-coated (63 nm) porous alumina indicated that the
absorbance due to the C=O band of PAA increases nonlinearly as a function of the
number of layers, which suggests hyperbranched growth of these ﬁlms (Figure

2.3). We estimated the thickness of PAA ﬁlms on porous alumina by comparing

their IR absorbances with absorbances due to ﬁlms of known ellipsometric thickness on
gold wafers. Using this method, the estimated thickness of a 6-layer PAA ﬁhn on porous
alumina is about 940 A.32 This compares reasonably well with FESEM images, which
show 6-1ayer PAA ﬁlms to be about 850 A thick. Because alumina substrates are highly
porous, direct ellipsometric measurements are not possible on these membranes.

We utilized transmission FTIR spectroscopy (Figure 2.4) in an effort to
investigate the synthesis of PAA ﬁlms on porous alumina coated with 5 nm of gold
because these surfaces are not sufﬁciently reﬂective for external reﬂection FTIR
spectroscopy. Large carbonyl absorbances in the transmission FTIR spectra of PAA-
coated supports (absorbance of ~1 .0 at 1,712 cm'1 for 6-layer ﬁlms) suggest that
considerable PAA is adsorbed inside the alumina pores. We conﬁrmed this by control
experiments where PAA ﬁlms were grown with and without MUA monolayers on gold-
coated (5 nm) alumina membranes. Absorbance values are comparable for the two cases,
which indicates that growth inside the pores is the primary reason for the large IR
absorbances.

It is not surprising that some PAA would be present in substrate pores because a
minimal amount of adsorption will be ampliﬁed by hyperbranched growth. During the
deposition of PTBA, some physisorption likely occurs. After hydrolysis, interaction of

the —COO‘ groups of PAA with alumina may result in even stronger physisorption. The

37

 

   
 

9-05 Amide u

/

  

 

 

 

 

 

(D
O
C
B
5 COOH
U)
.0
<1:
3—PAA
l l l
1800 1600 1400
Energy (cm'1)

Figure 2.4 Transmission FTIR spectra of a 3-layer PAA ﬁlm before and alter
ﬂuorination. The substrate was gold-coated (5 nm) porous alumina (0.02 um pore
diameter).

38

high internal surface area of the alumina supports leads to large carbonyl absorbances
even when only a small fraction of the pores is ﬁlled with PAA. Comparison of IR
absorbances of PAA ﬁlms grown on alumina supports to absorbances of known amounts
of PAA physisorbed on alumina supports indicates that there is 1.1 i 0.4 mg of PAA
distributed through alumina supports after the synthesis of a 6-layer PAA ﬁlm. The
number of ~0.2 rim-diameter pores in the bulk of porous alumina is 2 x 109/cm2
(determined from SEM images), giving a total internal surface area of 1800 cm2 for these
50 um-thick membranes. If we assume that the PAA is distributed evenly across the
internal surface area, there would be a 6 i 2 nm-thick ﬁlm along the pore walls. This
would reduce the inside diameter of pores by < 10%. For 3-layer PAA ﬁlms, the pore
diameter would hardly be affected. FESEM images and gas-permeability studies (vide
infra) clearly show that PAA is not ﬁlling a substantial fraction of membrane pores.

Transmission F TIR spectroscopy conﬁrms modiﬁcation of 3-layer PAA ﬁlms
with H2NCH2(CF2)6CF3 (Figure 2.4). After derivatization, the acid carbonyl peak (1,712
cm") diminishes and amide I (1,669 cm") and amide 11 (1,537 cm") bands appear.
Additionally, a CFx stretching peak appears at 1,254 cm“l . Other CFx stretches cannot be
observed because of the large absorbance of alumina below 1,210 cm’l. Although these
spectra represent material both within the alumina and on its surface, similar chemistry
should occur in both places.

Field-Emission Scanning Electron Microscopy (FESEM). Cross-sectional
FESEM images of alumina supports before and after PAA growth indicate that PAA
ﬁlms cover substrate pores without ﬁlling them. Figure 2.5 (a-f) shows FESEM images

of porous alumina before and after coating with 4- and 6-layer PAA ﬁlms. In the case of

39

6-layer deposition, the PAA is easily observed, and yet, pores are unclogged. Because of
their small thickness, 4-1ayer PAA ﬁlms are not easy to see. The large increase in ﬁlm
thickness on going from four to six layers conﬁrms the non-linear PAA growth shown by
FTIR. Prior to ﬁlm synthesis, we sputter-coated the alumina with either 5 nm (a-c) or 63
nm (d-f) of gold. As seen by comparing images with no PAA ﬁlms (a, d), the 63 nm-
gold layer is clearly distinguishable from the membrane and has a porous structure that
does not form an impermeable barrier.

All of the images in Figure 2.5 (a-f) were taken on substrates that have a nominal
0.02 pm pore diameter. These substrates have a very thin cake layer on their surface with
0.02 rim-diameter pores, while the bulk of the membrane has 0.2 um-diameter pores.
This substrate geometry makes it easier to cover pores without ﬁlling them. However,
images of substrates with 0.2 um-diameter pores (no cake layer) show that PAA ﬁlms
cover even these large pores (Figure 2.6). A PAA ﬁlm is clearly visible after deposition
of 6 layers on alumina coated with 5 nm of gold. Despite the larger pores, deposition is
restricted primarily to the substrate surface.

In an effort to ensure that gold coating after cleavage does not affect the electron
micrographs, we also imaged a sample covered with 4-layers of PAA (0.02 um-diameter
pores coated with 63 nm gold) without coating with gold alter cleavage (Figure 2.7). The
image shows the presence of PAA and conﬁrms that the ﬁlm is, in fact, PAA and not an
artifact of sample preparation. FESEM images (cross—sections) of other types of control
samples on which we attempted to graft 6-layer PAA and ﬂuorinated 6-layer PAA ﬁlms
on gold-coated (5 nm) alumina without depositing a MUA monolayer do not show

signiﬁcant ﬁlm growth on the alumina surface. This is consistent with gas permeability

40

data that show minimal inhibition of ﬂux for membranes prepared without ﬁrst

depositing a layer of MUA.

41

llllrlvllllll

4.0kvig&7mm x70.01< SE(U) 1.3/28199 12:17 429nm

 

Figure 2.5 (a). Field-emission scanning electron micrograph (cross-section) of
porous alumina (0.02 pm surface pore diameter) before deposition of PAA ﬁlms.
In this image, 5 nm of gold was deposited before PAA growth.

42

PAA

 

Figure 2.5 (b). Field-emission scanning electron micrograph (cross-section) of
porous alumina (0.02 pm surface pore diameter) after deposition of 4-layer
PAA ﬁlms. In this image, 5 nm of gold was deposited before PAA growth.
After cleaving the membrane, sample was sputter-coated with 5 nm of gold for
imaging.

43

PAA

:1

“ 1%

.1
9
1111111111
42

‘1 1
mm x70.0k SE[U) 811/99 21 :09 1

I

 

I

Figure 2.5 (c). Field-emission scanning electron micrograph (cross—section) of
porous alumina (0.02 pm surface pore diameter) after deposition of 6-layer
PAA ﬁlms. In this image, 5 nm of gold was deposited before PAA growth.
Alter cleaving the membrane, sample was sputter-coated with 5 nm of gold for
imaging.

lllllll’llll

4.0kV 12.3mm x70.Dk seru) 5125299 16:21

 

Figure 2.5 (d). Field-emission scanning electron micrograph (cross-section) of
porous alumina (0.02 pm surface pore diameter) before deposition of PAA
ﬁlms. In this image, 63 nm of gold was deposited before PAA growth.

45

PAA
Gold

32'.

1.

a
u ‘2
- ti

3

111111.111!
»428nm

 

Figure 2.5 (e). F ield-emission scanning electron micrograph (cross-section) of
porous alumina (0.02 pm surface pore diameter) after deposition of 4-layer PAA
ﬁlms. In this image, 63 nm of gold was deposited before PAA growth. After
cleaving the membrane, sample was sputter-coated with 5 nm of gold for
imaging.

46

 

Figure 2.5 (f). Field-emission scanning electron micrograph (cross-section) of
porous alumina (0.02 pm surface pore diameter) after deposition of 6-layer
PAA ﬁlms. In this image, 63 nm of gold was deposited before PAA growth.
After cleaving the membrane, sample was sputter-coated with 5 nm of gold
for imaging.

47

 

Figure 2.6. Field-emission scanning electron micrographs (cross-section) of
porous alumina (0.2 pm pore diameter) before and after deposition of 6-PAA
layers. 5-nm of gold was deposited before PAA growth, and samples were
sputter-coated with 5 nm of gold for imaging alter cleaving.

48

 

Figure 2.7. Field-emission scanning electron micrograph (cross-section) of a 4-
layer PAA ﬁlm grown on a gold-coated (63 nm) porous alumina membrane (0.02
pm surface pore diameter). The membrane was imaged without coating with gold
after cleavage.

49

Atomic Force Microscopy (AF M). AF M images show that 6-layer PAA ﬁlms
cover the porous alumina surface completely. Figure 2.8 contains images of gold-coated
porous alumina substrates before and after grafting of 4- and 6-layers of hyperbranched
PAA to the surface. With the addition of successive PAA layers, surface features of the
porous alumina become less distinct as the ﬁlm covers the substrate. Prior to PAA
coating, images (a and (I) show the pointed surface features of gold-coated alumina
substrates. These images are similar to naked porous alumina. Interestingly, the
sputtered gold appears to follow the pattern of the underlying substrate even when 63 nm
of gold is deposited. This explains why the cross-sectional FESEM images show that
thick gold coatings are porous. The pores in the alumina substrates most likely reside
between the peaks on the surface. This topology allows hyperbranched PAA to
effectively cover pores because the polymer can grow above the pore from all sides.

As the images show, covering the underlying alumina with PAA ﬁlms greatly
reduces surface roughness. To quantitate the changes in roughness, we used the
instrument software to calculate the average surface roughness. This average roughness,
R,, is deﬁned by equation 2.2 where Z, is the deviation between the surface height at any

Ra : {7 211211 2.2
point and the average surface height, and N is the number of points within the measured
area. The surface of the alumina coated with 5 nm of gold has an average roughness of

40 i 10 run, while the surface of the alumina coated with 63 nm of gold has an average

roughness of 19 i 2 nm. In contrast, 6-layer PAA surfaces on 5-nm gold-coated supports

50

have average roughnesses of 7 i 1 nm, and similar ﬁlms on substrates coated with 63 nm
of gold have average roughnesses of 6 i 2 nm. These measurements show that 6-layer
PAA ﬁlms form a smooth surface that covers underlying sharp features.

Figure 2.9 shows AF M images of PAA ﬁlms that were prepared on gold-coated
substrates with 0.2 rim-diameter surface pores. Pores in the substrate before coating with
PAA (a, c) are an order of magnitude larger than those in Figure 2.9, as would be
expected. After deposition of PAA, surface roughness decreased ﬁom 20 i 2 nm to 9 i 2
nm on the substrate coated with 5 nm of gold and from 18.0 :t 0.4 nm to 8 :1: 1 nm on the
surface coated with 63 nm of gold. The ﬁgure clearly shows that six layers of PAA

completely cover underlying pores.

51

X 0.500 tin/div
2 100.000 nM/diu

 

Figure 2.8 (a). Tapping-mode atomic force micrograph of gold-coated porous
alumina before grafting of PAA ﬁlms. The gold coating on 0.02-um surface
pore-diameter-Anopore membrane was 5 nm thick in this image.

52

X 0.500 LIN/dIU
2 100.000 nn/diu

 

Figure 2.8 (b). Tapping-mode atomic force micrograph of gold-coated porous
alumina after grafting of 4-layer PAA ﬁlms. The gold coating on 0.02-um surface
pore-diameter-Anopore membrane was 5 nm thick in this image.

53

X 0.500 un/diu
2 100.000 nM/diu

 

Figure 2.8 (c). Tapping-mode atomic force micrograph of gold-coated porous
alumina after grafting of 6-layer PAA ﬁlms. The gold coating on 0.02-um surface
pore-diameter-Anopore membrane was 5 nm thick in this image.

54

X 0.500 lJM/diu
2 100.000 hM/dlU

 

Figure 2.8 (d). Tapping-mode atomic force micrograph of gold-coated porous
alumina before grafting of PAA ﬁlms. The gold coating on 0.02-um surface
pore-diameter-Anopore membrane was 63 nm thick in this image.

55

X 0.500 un/cliu
2 100.000 nH/diu

 

Figure 2.8 (e). Tapping-mode atomic force micrograph of gold-coated porous
alumina after grafting of 4-layer PAA ﬁlms. The gold coating on 0.02—um
surface pore-diameter-Anopore membrane was 63 nm thick in this image.

56

X 0.500 uM/diu
2 100.000 nM/diu

 

Figure 2.8 (f). Tapping-mode atomic force micrograph of gold-coated porous
alumina after grafting of 6-layer PAA ﬁlms. The gold coating on 0.02-um
surface pore-diameter-Anopore membrane was 63 nm thick in this image.

57

X 0.500 uM/diu
2 100.000 nM/diu

 

Figure 2.9 (a). Tapping-mode atomic force micrograph of gold-coated alumina
before grafting of hyperbranched PAA. The gold coating on 0.2 um pore-
diameter-Anopore membrane was 5 nm thick in this image.

58

X 0.500 un/diu
2 100.000 hM/dlU

 

Figure 2.9 (b). Tapping-mode atomic force micrograph of gold-coated alumina after
grafting of 6-layers of hyperbranched PAA. The gold coating on 0.2 um pore-
diameter-Anopore membrane was 5 nm thick in this image.

59

X 0.500 lm/diu
2 100.122 nn/diu

 

Figure 2.9 (c). Tapping-mode atomic force micrograph of gold-coated alumina
before grafting of hyperbranched PAA. The gold coating on 0.2 um pore-diameter-
Anopore membrane was 63 nm thick in this image.

60

X 0.500 LIN/div
2 100.000 nn/diu

 

Figure 2.9 (d). Tapping-mode atomic force micrograph of gold-coated alumina
after grafting of 6-layers of hyperbranched PAA. The gold coating on 0.2 pm
pore-diameter-Anopore membrane was 63 nm thick in this image.

61

Defects in Gold Coatings. Selective transport requires defect-free ﬁlms
because ﬂux through open channels will negate the selectivity of the membrane.20
Unfortunately, during initial permeation studies we noticed that the 63 nm-thick gold
coatings had small, but visible holes. Careful inspection under a microscope revealed
that there were at least 20 small (1-100 um) holes in the gold layer on a substrate with an
area of 3 cm2. We inspected substrates under a microscope immediately after metal
deposition, and most of the substrates had holes in the gold layer. Rinsing these samples
with methanol and drying them with nitrogen produced additional holes. Optical
microscope images (Figure 2.10) often show that the small piece of gold removed from
the surface after rinsing sits next to the newly formed hole.

Our attempts to improve adhesion of gold by cleaning with hot solvents, dilute
ammonium hydroxide, UV/ozone, or Ar plasma did not provide a defect-free metal layer.
The use of an adhesion layer (Cr or Ti) to prevent peeling was also unsuccessful. The
only method that appeared to give defect-free gold coating was deposition of a much
thinner layer of gold (5 nm). We note, however, that it is more difﬁcult to see defects in
these thin gold coatings. Selectivities observed in gas-transport measurements (described
below) give corroborating evidence that the 5-nm layer was free of defects.

Gas-Transport Measurements. In the solution-diffusion model of ﬂux

through a membrane, the permeability coefﬁcient of a given analyte is related to its

. . . . 5
solubility and diffusion coefﬁcient in the membrane material as shown in Equauon 2.3,
P = S x D 2.3
where P is the permeability coefﬁcient, S is the apparent solubility constant for the given

gas—membrane pair, and D is the apparent diffusion coefﬁcient. Selectivity is deﬁned as

62

 

Figure 2.10 Optical micrograph of a defect on a gold-coated (63 nm) porous alumina
(0.02 um-diameter surface pores) substrate.

63

the ratio of the permeability coefﬁcients of two different gases. Thus, selectivities
between gases can result from differences in either solubility or diffusivity.

Using 5 nm gold-coated substrates, we began investigating the gas-permeability
of 3-layer hyperbranched PAA ﬁlms. Figure 2.11 shows how gas ﬂow through a 3-layer
PAA ﬁlm varies with inlet pressure for various gases. Total ﬂux through these PAA
membranes depends primarily on the molar mass of the gases; thus, He has the highest
ﬂux and SF, the lowest. The selectivity ratios are not very high and fall around or below
calculated Knudsen-diffusion values. Knudsen-diffusion of gases results from diffusion

through pores with radii that are much smaller than the mean free path of the gases, and

ﬂux is inversely proportional to the square root of molar mass.20 These results suggest
that 3-layer PAA ﬁlms are rather porous. We note, however, that the permeability of
these membranes is orders of magnitude lower than the permeability of gold-coated
porous alumina.

When PAA ﬁlms were modiﬁed by ﬂuorination with H2NCH2(CF2)6CF3,
selectivities improved signiﬁcantly, and the relative ﬂuxes did not depend solely on
molar mass (Figure 2.12). The enhancement in selectivity occurred because the
permeability of some gases (He, O2, and CO2) increased upon ﬂuorination, while other
gases were relatively unaffected by derivatization. The selectivity ratios for several of
the gas pairs are 3-4 times higher than those obtained with PAA ﬁlms. Table 2.1 shows
selectivity ratios for 3-layer PAA and 3-layer ﬂuorinated PAA membranes, along with
calculated values for Knudsen diffusion-based selectivities. The increase in selectivity
upon ﬂuorination is not simply due to ﬁlling of defects in the PAA ﬁlms. If that were the

case, ﬂux would decrease for all gases and increases in ﬂux for some of the gases upon

64

 

 

 

 

l I l l
0 He
4 _ A CH4 _
O
I N2
[:1 02
o
A A co2
:
ug 3 I— 0 SF6 . —
:1
E. o A
9
co 2 - A r
h o
g A I
_ I Q
0 A
L I 151 O
1 - A I 5 O -
O A g g Q
0 l l l I
0 10 20 30 4O

Inlet Pressure (psig)

Figure 2.11 Flow rates of several gases through a three-layer PAA ﬁlm on gold-coated (5
nm) porous alumina (0.02 pm surface pore diameter) as a function of inlet pressure. The
membrane area was 2.8 cm2. Permeability values with standard deviations are listed in
the text.

65

 

 

 

 

I I I I
0 He
A CH4
30 " I N2 A "
El 02
c A CO2
-- A
g 0 SF6 .
E 20 - A . "
45‘
s A -
a . -
E _ O _
10 A D
O D [:1
0 D E]
I
[1 D I I l I
0 5 H n O Q 0 Q
0 1O 20 30 40

Inlet pressure, psig

Figure 2.12 Flow rates of several gases through a ﬂuorinated three-layer PAA ﬁlm on
gold-coated (5 nm ) porous alumina (0.02-um pore diameter) as a function of inlet
pressure. The membrane area was 2.8 cm2. Permeability values with standard deviation
are listed in the text.

66

ﬂuorination certainly would not occur. Also, the addition of subsequent PAA layers,
which should increase coverage, does not improve selectivity. We tested gas
permeabilities of 4-, 5-, and 6-layer PAA ﬁlms. 4- and 5-layer PAA ﬁlms still showed
Knudsen-like selectivities, and 6-layer PAA ﬁhn permeabilities were below the limit of
detection of our ﬂow-meter. The fact that additional PAA layers do not improve
selectivity suggests that the lack of selectivity in these ﬁlms is likely a property of
hyperbranched PAA and not a manifestation of defects in substrate coverage by these
ﬁlms.

Derivatization of PAA ﬁlms with H2NCH2(CF2)6CF3 probably increases
selectivities because of differing solubilities of gases in these ﬁlms. Formation of a more
crystalline membrane may also increase selectivity. The O2/N 2 selectivities we observe
with ﬂuorinated PAA agree well with selectivities of other ﬂuorinated membranes.
Several studies report higher 02 permeability values and improved O2/N2 selectivities (2
— 6) in a variety of ﬂuorinated polymers such as a ﬂuorine-containing
polydimethoxysilane (PDMS)-tn'methylsilylpropyne (PTMSP) block copolymer,
polyimides containing triﬂuoromethyl groups, and a ﬂuorine-containing PDMS-ethyl
cellulose graft copolymer.”38 Aoki et al. also showed that adding a small amount

(1 wt %) of poly (triﬂuoromethyl substituted arylacetylene) to PDMS and PTMSP ﬁlms

3
enhances the oxygen permeability and selectivity of O2/N 2. 9 The increased oxygen

permeability of ﬂuorinated membranes is most likely due to the high solubility of oxygen

in ﬂuorinated hydrocarbons.39'40 COz/CH4 selectivities of ﬂuorinated PAA are also
similar to those in a recent report on gas permeation through the ﬂuorinated polymer

poly(2,2-bis(triﬂuoromethyl)-4,5-diﬂuoro- 1 ,3-dioxole-co-tetraﬂuoroethylene

67

Table 2.1 Ratios of the ﬂuxes of different gases through 3-layer PAA ﬁlms and

ﬂuorinated 3-layer PAA ﬁlms.“

 

 

 

 

 

 

 

 

H6 CH4 N2 02 C02 SF6
1.3 (2.0) 1.9 (2.6) 2.0 (2.8) 2.0 (3.3) 2.4 (6.1)
He —
[6.7] [6.3] [2.8] [0.81] [24.4]
1.3 (2.0) 1.4 (1.3) 1.5 (1.4) 1.5 (1.7) 1.8 (3.0)
CH4 —
[6.7] [0.94] [0.42] [0.12] [3.6]
1.9 (2.6) 1.4 (1.3) 1.1(1.1) 1.1(1.3) 1.3 (2.3)
N .—
2 [6.3] [0.94] [0.45] [0.13] [3.9]
2.0 (2.8) 1.5 (1.4) 1.1(1.1) 1.0 (1.2) 1.2 (2.1)
O _
2 [2.8] [0.42] [0.45] [0.29] [8.6]
2.0 (3.3) 1.5 (1.7) 1.1(1.3) 1.0 (1.2) 1.2 (1.8)
(:0 —
2 [0.81] [0.12] [0.13] [0.29] [30.0]
2.4 (6.1) 1.8 (3.0) 1.3 (2.3) 1.2 (2.1) 1.2 (1.8)
SF _
6 [24.4] [3.6] [3.9] [8.6] [30.0]

 

 

 

 

 

 

 

a Values for ﬂuorinated ﬁlms are in brackets. The ratios for Knudsen-diffusion selectivity are
given in parentheses. All value represent the ﬂux of the lighter gas divided by the ﬂux of the
heavier gas.

68

 

(TFE/BDD87).4l Of most relevance to this chapter, selectivities are consistent with those
of poly(1,1 ’-dihydroperﬂuorooctyl acrylate), which is very similar in chemical
composition to PAA derivatized with H2NCH2(CF2)6CF3.42

Control experiments show that the selectivity shown by ﬂuorinated PAA
membranes is due to the ﬁlm on the surface of the membrane and not to grth inside the
pores. After attempted growth of a 3-layer PAA ﬁhn on a control sample (no MUA was
deposited prior to PAA grafting) ﬂux was too high to be measured by our ﬂowmeter (>>
50 mL/min). After deposition of a 6-layer PAA ﬁlms on a control sample, ﬂux was still
higher than that for a 3-layer PAA ﬁlm grown with a MUA monolayer. These 6-layer
PAA ﬁlms did not show selectivity (not even Knudsen selectivity), even after
ﬂuorination. As mentioned earlier, FESEM analysis of control samples did not show
blocking of the pores or the presence of a ﬁlm on the surface. This control experiments
show that the presence of a monolayer is required for ﬁlm growth on the surface and
effective coverage of pores. Additionally, any growth inside the pores is not affecting
permeability results.

We calculated permeability coefﬁcients for 3-layer PAA membranes using
Equation 1 and assuming a ﬁhn thickness of 20 nm. The ﬁlm thickness was estimated
based on the thickness of 3-layer PAA ﬁlms on gold-coated silicon wafers and FESEM
images of PAA ﬁlms on gold-coated porous alumina. The permeability coefﬁcients for
various gases are He, 2.6 i 0.7 barrer; N2, 1.4 i 0.6 barrer; SFb, 1.1 i 0.5 barrer; 02, 1.3 i
0.6 barrer; CO2, 1.3 i 0.7 barrer; and CH4 2.0 i 0.8 barrer, where 1 barrer is [1010 cm3
(STP) cm/(cm2 s cm(Hg)]. In calculating the permeability coefﬁcients of ﬂuorinated

ﬁlms, we used a thickness of 40 run because ﬁlm thickness doubles upon ﬂuorination

69

with H2NCH2(CF2)(,CF3.29 After ﬂuorination, the permeability coefﬁcients are higher for
CO2 (24 i 3 barter), He (20 i 4 barrer), O2 (7 i 2 barrer), and N2 (3.1i 0.7 barrer),
whereas coefﬁcients for CH4 (2.9 i 0.4 barrer) and SF6 (0.8 i 0.1 barrer) are relatively

unchanged. The permeability coefﬁcients are 10-50 times higher than those for highly
selective, ultrathin, conducting polymer membranes,ll but about 2 orders of magnitude

lower than those reported by Merkel et al. for TFE/BDD87.“ The relatively low ﬂuxes

of ﬂuorinated PAA may be due to crystallinity resulting from interactions between

perﬂuorooctyl groups. We note that teﬂon has a low permeability.4l Low ﬂuxes may also

result from hydrogen bonding interactions in residual —COOH groups in PAA.” 4"

2.4 Conclusions

The graft-on-graft synthesis and derivatization of PAA on gold-coated porous
alumina yields selective, ultrathin membranes. Film thickness increases nonlinearly with
the number of deposited layers and these ﬁlms effectively cover porous substrates,
presumably because of their hyperbranched structure. FESEM and AF M images clearly
show that six-layer PAA ﬁhns cover the surface without ﬁlling underlying pores. Gas-
permeability measurements indicate that three-layer PAA ﬁlms show modest Knudsen
diffusion-based selectivity, but ﬂuorination of these ﬁlms provides selective, ultrathin
membranes for gas separations. Although the complex synthesis of hyperbranched PAA
ﬁlms will likely prohibit their use in commercial gas separations, these permeability data
do show that PAA membranes can be defect-free and easily derivatized. Thus, PAA

membranes might prove valuable in small-scale sensing or separations applications.

70

2.5 References

(1)
(2)
(3)
(4)
(5)

(6)
(7)
(8)
(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)
(18)

Abelson, P. H. Science 1989, 244, 1421.

Henis, J. M. S.; Tripodi, M. K. Science 1983, 220, 11-17.

Haggin, J. Chem. Eng. News 1990, 68, 22-26.

Spillman, R. W. Chem. Eng. Prog. 1989, 85, 41-62.

Freeman, B. D.; Pinnau, I. in Polymer Membranes for Gas and Vapor Separation
Freeman, B. D. and Pinnau, 1., Ed.; American Chemical Society: Washington, DC,
1999; pp1-27.

Maier, G. Angew. Chem. Int. Ed. 1998, 37, 2960-2974.

Freeman, B. D. Macromolecules 1999, 32, 375-380.

Robeson, L. M. J. Membr. Sci. 1991, 62, 165-185.

Robeson, L. M.; Burgoyne, W. F.; Langsam, M.; Savoca, A. C.; Tien, C. F.
Polymer 1994, 35, 4970-4978.

Rezac, M. E.; Koros, W. J. J. Appl. Polym. Sci. 1992, 46, 1927-1938.
Liu, C.; Martin, C. R. Nature 1991, 352, 50-52.

Prasad, R.; Shaner, R. L.; Doshi, K. J. in Polymeric Gas Separation Membranes;
Paul, D. R. and Yampol'skii, Y. P., Ed.; CRC: Boca Raton, 1994; pp 513-614.

Pinnau, I.; Koros, W. J. Ind. Eng. Chem. Res. 1991, 30, 1837-1840.

Gantzel, P. K.; Merten, U. 1&EC Proc. Des. Dev. 1970, 9, 331.

Suda, H.; Haraya, K. in Membrane Formation and Modification ; Pinnau, I. and
Freeman, B. D., Ed.; American Chemical Society: Washington, DC, 2000; pp 295-
3 l 3.

Francis, P. Fabrication and Evaluation of New Ultrathin Reverse Osmosis
Membranes; NTIS Report, Pb-l77083, Springﬁeld, VA, 1966.

Le Roux, J. D.; Paul, D. R. J. Membr. Sci. 1992, 74, 233-252.

Kesting, R. Synthetic Polymeric Membranes: A Structural Perspective; 2nd ed.;
John Wiley & Sons, New York, 1985; pp 224-236.

71

(19)

(20)

(21)

(22)

(23)

(24)
(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

Cadotte, J. in Materials Science of Synthetic Membranes ; Lloyd, D. R., Ed.;
American Chemical Society: Washington, DC, 1985; pp 273-294.

Kesting, R. E.; Fritzsche, A. K. Polymeric Gas Separation Membranes; John Wiley
& Sons, New York, 1993; pp 284-318.

Conner, M. D.; Janout, V.; Kudelka, I.; Dedek, P.; Zhu, J .; Regen, S. L. Langmuir
1993, 9, 2389-2397.

Hendel, R. A.; Nomura, E.; Janout, V.; Regen, S. L. J. Am. Chem. Soc. 1997, 119,
6909-6918.

Zhang, L.-H.; Hendel, R. A.; Cozzi, P. G.; Regen, S. L. J. Am. Chem. Soc. 1999,
121,1621-1622.

Liu, (3.; Chen, w. J.; Martin, c. R. J. Membr. Sci. 1992,65, 113-128.
Shieh, J. J.; Chung, T. 8.; Paul, D. R. Chem. Eng. Sci. 1999, 54, 675-684.

Zhou, Y.; Bruening, M. L.; Bergbreiter, D. E.; Crooks, R. M.; Wells, M. J. Am.
Chem. Soc. 1996, 118, 3773-3774.

Bruening, M. L.; Zhou, Y.; Aguilar, G.; Agee, R.; Bergbreiter, D. E.; Crooks, R. M.
Langmuir 1997, 13, 770-778.

Odukale, A. A.; MS thesis; Michigan State University: East Lansing, MI, 1999.

Zhou, Y.; Bruening, M. L.; Liu, Y.; Crooks, R. M.; Bergbreiter, D. E. Langmuir
1996, 12, 5519-5521.

Kim, K. J .; Dickson, M. R.; Chen, V.; Fane, A. G. Micron and Microscopia Acta
1992, 23, 259-271.

Coates, V. J. Proc. 40th Ann. Electron Microscopy Soc. Am. 1982, 752-753.

A hyperbranched PAA ﬁlm on a gold-coated wafer with an ellipsometric thickness
of 970 A, has a carbonyl absorbance of 0.1 13. Using this data point and the fact
that the absorbance due to a PAA membrane on alumina was 0.109, we estimated
that the membrane thickness was 940 A. Ellipsometric measurements on gold
wafers were performed with a Woollam M-44 rotating analyzer ellipsometer.
Substrate parameters (11 and k) were determined prior to ﬁlm deposition.

Aoki, T. Prog. Polym. Sci. 1999, 24, 951-993.

Coleman, M. R.; Koros, W. J. J. Membr. Sci. 1990, 50, 285-297.

72

(35)
(36)

(37)

(38)

(39)
(40)

(41)

(42)

(43)

(44)

Nagase, Y.; Ochiai, J.; Matsui, K.; Uchikura, M. Polym. Commun. 1988, 29, 10-13.
Nagase, Y.; Ochiai, J.; Matsui, K.; Uchikura, M. Polymer 1988, 29, 740-745.

Kim, T. H.; Koros, W. J .; Husk, G. R.; O'Brien, K. C. J. Membr. Sci. 1988, 37, 45-
62.

Stem, S. A.; Mi, Y.; Yamamoto, H. J. Polym. Sci, Polym. Phys. Ed. 1989, 27,
1887-1909.

Aoki, T.; Oikawa, E.; Hayakawa, Y.; Nishida, M. J. Membr. Sci. 1991, 57, 207-216.
Inagaki, N.; Kawai, H. J. Polym. Sci, Polym. Chem. Ed. 1986, 24, 3381-3391.

Merkel, T. C.; Bondar, V.; Nagai, K.; Freeman, B. D.; Yarnpolskii, Y. P
Macromolecules 1999, 32, 8427-8440.

Arnold, M. E.; Nagai, K.; Freeman, B. D.; Spontak, R. J .; Betts, D. E.; DeSimone, J.
M.; Pinnau, I. Macromolecules 2001, 34, 5611-5619.

Bergbreiter, D. E.; Tao, G. J. Polym. Sci. Part A: Polym. Chem. 2000, 38, 3944-
3953.

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

73

Chapter 3

pH-DEPENDEN T GROWTH AND MORPHOLOGY OF
MULTILAYER DENDRIMER/POLY(ACRYLIC ACID) FILMS

Summary

Because hyperbranched poly(amidoamine) dendrimers are well-deﬁned
polycations with a high density of surface charge, they provide unique materials for
controlled synthesis of ﬁlms by alternating electrostatic adsorption of polycations and
polyanions. Thicknesses of generation 4 (G4) dendrimer/poly(acry1ic acid) (PAA)
“bilayers” can be tuned from <10 A to >2,000 A simply by varying deposition pH, while
generation 8 (G8) dendrimer/PAA bilayers can be as thick as 4,000 A. The highest
thicknesses occur when using deposition pH values at which PAA and dendrimers are
only partially charged, and the thickest ﬁlms show high surface roughnesses, which may
account in part for the rapid ﬁlm growth. Deposition of dendrimer/PAA ﬁlms on porous
alumina supports yields highly gas-penneable membranes, in sharp contrast to several
other multilayer polyelectrolyte systems, which show little gas permeability.
Dendrimer/PAA ﬁlms can also be induced to form microporous coatings through pH and
salt-induced structural changes, and pore diameters in these structures can be varied from

0.02 11m to 0.4 um.

74

3.1 Introduction

Alternating electrostatic adsorption of polycations and polyanions is a convenient,
versatile technique for forming ultrathin ﬁlms."2 Synthesis of these ﬁlms simply involves
altemating immersions of a substrate in solutions containing polycations and polyanions,
and possible ﬁlm constituents include materials as diverse as polymers}4 inorganic
nanoparticles,5 and proteins.° Selection of these constituents as well as deposition
conditions (i.e. pH,7 polyelectrolyte concentration,3 immersion time,8 and supporting
electrolyte concentrationg'”) allows tailoring of ﬁhn properties for possible applications

16,17

in sensing,”’13 separations,”’15 surface protection, controlled release,18 and

electronics. 1 9

Because of their unusual hyperbranched structure, dendrimers present unique
polycations that can be included in multilayer polyelectrolyte ﬁlms.”22
Poly(amidoamine) (PAMAM) dendrimers are especially attractive constituents for such
ﬁlms because they have a relatively well-deﬁned shape, a narrow molecular weight
distribution, and a high concentration of amine groups at their periphery.”’24 The open

21 ,22,25

interior and dense shell of dendrimers may also impart unique adsorption and

26-28

transport properties to ﬁlms containing these materials. We should note, however,

that the structures of dendrimers depend on pH,29 dendrimer type,30 and whether the

- - 25,31
dendrimer 1s bound to a surface.

In some cases, backfolding of the dendrimer
periphery may occur.30
Several studies demonstrated deposition of dendrimer-containing ﬁlms in a layer-

by-layer fashion.20'22’28’29’32'36 Regen and Watanabe constructed multilayered ﬁlms using

metal-ligand interactions between layers of dendrimers and organometallic complexes of

75

Ptz’L,22 while Liu and co-workers employed covalent bond formation to form ﬁlms by
alternating deposition of dendrimers and a poly(anhydride).28 Examples of alternating
electrostatic adsorption with dendrimers includes dendrimer/p0ly(styrenesulfonate)
(PSS),35‘3°, dendrimer/ glucose oxidase, dendrimer/p0lyoxometalates, and carboxylate-
terminated dendrimer/amino-terminated dendrimer ﬁlmszo’z‘ In these cases, ﬁlm
thickness increases with the number of bilayers, but ﬁlm formation was not studied as a
function of deposition conditions.

This paper reports the formation of multilayer ﬁlms by alternating adsorption of
protonated PAMAM dendrimers and ionized poly(acrylic acid) (PAA). Because these
polyelectrolytes are weak acids and bases, their deposition should be very sensitive to
pH.37 Remarkably, the thickness of dendrimer/PAA bilayers in such ﬁlms can be varied
from <10 A to as high as 4000 A by controlling pH and selecting the generation of
dendrimer. Rubner and coworkers previously showed that deposition pH affects the
thickness of poly(allylamine hydrochloride) (PAH)/PAA ﬁlms, but the maximum
thickness per bilayer in those studies was <150 A.37 The hyperbranched architecture of
the dendrimers evidently results in much greater thicknesses. Like PAH/PAA systems,
dendrimer/PAA ﬁlms are capable of forming microporous structures upon exposure to
low pH solutions.38 Unlike PAH/PAA, however, dendrimer/PAA ﬁlms are highly
permeable to gases.

Experimental Section
Materials. G4 amine-terminated PAMAM dendrimer (10 wt % solution in

methanol, MW = 14215), poly(acrylic acid) (25 wt % aqueous solution, Mw = 90,000 or

76

powder with Mw = 2000), and 3-mercaptopropionic acid (MPA) were used as received
from Aldrich Chemical.

G8 amine-terminated PAMAM dendrimer (Mw = 233,383) was obtained from Dendritech
(Midland, MI) as a 6.9 wt % aqueous solution. Gold-coated silicon wafers (200 nm of
sputtered gold on 20 nm chromium on Si (100)) were used as substrates for ellipsometric

measurements and external reﬂection FTIR spectroscopic studies. AnodiscTM porous

alumina membranes with 0.02 [rm-diameter surface pores (Whatman) were used as
supports for the preparation of composite membranes.

Film preparation. Gold-coated silicon substrates were immersed in piranha
solution (3:1 v/v mixture of concentrated H2SO4 and 30 % H202; caution, piranha reacts
violently with organic compounds and should not be stored in closed containers) for 1-3
min, rinsed thoroughly with deionized water (Milli-Q, 18 MQcm), and dried with N2.
The clean substrates were then immersed in a 1 mM mercaptopropionic acid (MPA)
solution in ethanol for 30 min, rinsed with ethanol and deionized water, and dried with
N2. This produces a carboxylic acid-containing monolayer that will be negatively charged
upon deprotonation. Deposition of one dendrimer layer occurred by immersion of a
monolayer-coated substrate in a solution of 10'5 M dendrimer for 10 min, followed by
washing with water and drying with N2. The dendrimer-coated substrate was then
immersed in 0.02 M PAA (molarity is given with respect to the repeating unit) for 5 min,
washed with water, and dried with N2. This procedure was repeated until the desired
number of dendrimer/PAA bilayers was deposited. The pH of dendrimer and PAA

solutions was adjusted by adding a few drops of dilute HCl or NaOH solutions.

77

Film characterization. Ellipsometry, external reﬂection F TIR spectroscopy,
ﬁeld-emission scanning electron microscopy (FESEM), atomic force microscopy (AF M),
optical microscopy, and gas transport measurements were performed as described in

- - . 14.39
prevrous publicatlons.

78

 

W IAU

Mercaptopropionic acid (1)
(MPA)

qoppppgpp.9920,09 MAPA
u

 

 

 

 

 

C-D carboxylate
g} protonated multiply charged PAMAM Dendrimer

Figure 3.1 Schematic diagram of the deposition of a dendrimer/PAA
bilayer on a Au surface. Repetition of steps 2 and 3 yields multilayer
ﬁlms.

79

3.2 Results and Discussion

Film formation and characterization. Figure 1 shows schematically the
preparation of multilayer dendrimer/PAA ﬁlms on gold. External reﬂection F TIR spectra
conﬁrm the formation of these ﬁlms as shown in Figure 2. The spectra contain strong
absorbances at 1550 cm'1 and 1650 cm‘1 that are due to the amide bonds in the dendrimer
internal structure, and these bands increase in intensity with the addition of each
dendrimer layer. After deposition of PAA layers, an absorbance at 17 30 cm'1 appears
due to the carbonyl stretch of the —COOH groups of PAA (spectrum c). Upon subsequent
deposition of the dendrimer, the acid carbonyl peak decreases to < 20 % of its initial
value (spectrum d), suggesting that most of the —COOH groups are deprotonated during
dendrimer adsorption. This is reasonable as the dendrimer was deposited at pH 8. The —
C 00' symmetric stretch (1400 cm") appears after deposition of the dendrimer, while the
—COO' asymmetric stretch (~1575 cm") is probably masked by the amide II band.
Spectra are similar for G8 dendrimer/PAA systems, and these data are consistent with
results for PAH/PAA ﬁlms.7’l7

Effect of deposition pH on the thickness of G4 dendrimer/PAA ﬁlms. Rubner
and coworkers showed that deposition pH has a large effect on the thickness of
PAH/PAA ﬁlms.37 To see if similar effects occur in dendrimer/PAA ﬁlms, we
independently varied the pH of dendrimer and PAA deposition solutions from 2 to 8.
Figure 3 shows ellipsometric thicknesses of 2.5-bilayer G4 dendrimer/PAA ﬁlms as a
function of the deposition pH for both dendrimers and PAA. (In this terminology, ﬁlms
with an integer number of bilayers have PAA as the top layer, whereas ﬁlms with an

additional 0.5 bilayer are terminated with dendrimers.) When dendrimers are deposited

80

Absorbance

 

 

Amide I
I

 

 

 

 

 

 

 

 

 

 

 

 

0-05 11 Amide 11
Acid ‘1
c=o
A .__..._ C
M v. p
1800 1600 1400 1200

Wavenumbers(cm '1)

Figure 3.2 Reﬂection FTIR spectra of a MPA-modiﬁed gold substrate coated
with (a) 0.5 (b) 1.5 (c) 2 and (d) 2.5 G4 dendrimer/PAA bilayers. The half

bilayer results in the dendrimer being the top layer in the ﬁlm. Deposition pH
values were 8 for G4 dendrimers and 4 for PAA

81

Thickness (A)

 

Figure 3.3 Ellipsometric thicknesses of 2.5-bilayer G4 dendrimer/PAA ﬁlms as a
function of the pH used for dendrimer and PAA deposition. Standard deviations of

thicknesses measured on three different samples ranged from 3 to 35 % of the
measured values.

82

at pH 2, 4, or 6, and PAA is deposited at pH 8, little ﬁlm growth occurs (thickness <5 A).
At these pH values, both dendrimers and PAA are highly charged, and minimal ﬁlm
growth takes place probably because PAA is in a thin, extended conformation parallel to
the substrate surface.37 In most other cases, thickness increases when decreasing the
deposition pH for PAA or increasing the deposition pH for dendrimers. Higher
thicknesses likely result from a lower charge density on both PAA and dendrimers in
these solutions and the resultant need for more adsorption to compensate the surface
charge.37‘40’4| At very low PAA deposition pH values (pH 2), PAA may desorb from the
ﬁlm,37 and thus some of these coatings are not as thick as ﬁlms in which PAA was
deposited at pH 4. The maximum thickness of 2.5-bilayer G4 dendrimer/PAA ﬁlms was
1200 A, while reported maximum bilayer thicknesses for PAH/PAA ﬁlms are < 150 A.37
Thus the architecture of the dendrimer must greatly affect ﬁlm growth.

The trends in Figure 3 also apply as more layers are added to the ﬁlm. After
deposition of 4.5 dendrimer/PAA bilayers, ﬁlms deposited at a dendrimer pH of 4 and a
PAA pH of 6 still have thicknesses less than 10 A. In contrast, Table 1 shows that the
thicknesses of G4 dendrimer (pH=8)/PAA (pH=4) ﬁhns grow by 4,300 A on going from
a 2.5-bilayer to a 4.5-bilayer ﬁlm. Table 1 also indicates that ﬁlm thicknesses initially
increase nonlinearly with the number of bilayers, but bilayer thickness becomes relatively
constant after deposition of 2.5 dendrimer/PAA bilayers. Such a grth proﬁle is often
seen in the build-up of polyelectrolyte multilayers because a consistent surface charge

distribution does not fully develop until after adsorption of a few bilayersfl’42

83

Table 3.1 Ellipsometric thicknesses of G4 and G8 PAMAM Dendrimer
(pH=8)/PAA (pH=4) ﬁhns on gold as a function of the number of bilayers.

 

 

 

G4 Dendrimer G8 Dendrimer
iigngnfigiﬁig/Sliﬂa Thickness/A Thickness/A Thickness/A
bilayers) MW of PAA = 90000 Mw of PAA = 90000 MW of PAA= 2000
0.5 9 i 1 14 i 2 10 i- l
1 28i4 36:7 32.11.02
1.5 110:30 310i10 140:10
2 300i30 670i30 3003210
2.5 ll70i90 3990:60 1020i 10
3 2410i 190 4950:80 l930i30
3.5 3060 i 100 7000 i 10 2780 j: 20
4 4540 i 370 8940 i 210 3990 j: 60
4.5 5470 i 600 11390 i 550 Fil’ggjggme

 

“ Half bilayers correspond to dendrimers being the last layer deposited

84

Effect of deposition pH and PAA chain length on the thickness of G8
dendrimer/PAA ﬁlms. Thicknesses of G8 dendrimer/PAA ﬁlms provide further
evidence that ﬁlm thickness depends on dendrimer structure. G8 dendrimers have 16
times as many peripheral amine groups as G4 dendrimers and about double the diameter,
and this increased size and crowded exterior functionality is clearly reﬂected in the
thicknesses of G8 dendrimer/PAA ﬁlms (Figure 4). As with G4 PAMAM dendrimers,
thickness usually increases with increasing dendrimer deposition pH and decreasing PAA
deposition pH. However, 2.5-bilayer G8 dendrimer (pH=8)/PAA (pH=4) ﬁlms have
thicknesses of 4000 A, more than three times that for similar G4 dendrimer/PAA ﬁlms.

As shown in Table l, the thicknesses of G8 dendrimer/PAA ﬁhns grow by as
much as 3300 A upon dendrimer deposition. This thickness increase would represent an
addition of 30 dendrimer monolayers (assuming a dendrimer diameter of 92 A43). Most
likely, the PAA already in the ﬁlm elongates or rearranges to provide adsorption sites for
these dendrimers. Extension of PAA by 3000 A would require at least 1000 monomer
units, or a PAA molecular weight of 72000, assuming an extended length of 3 A per
monomer. The PAA employed in this study had an average molecular weight of 90000,
so such an extension is possible. When we prepared ﬁlms from PAA with a molecular
weight of only 2000, however, we observed much smaller growth as seen in Table l.44
The lower growth would be consistent with the inability of PAA to extend over large
distances, but dendrimer deposition still resulted in thickness increases of 700 - 900 A.
These thicknesses are ten times longer than the extended length of PAA chains with a
molecular weight of 2000. Evidently, the PAA can also redistribute in the ﬁlms.

Previous studies of multilayer polyelectrolyte ﬁlms such as PAH/PSS and alternating

85

Thickness (A)

 

Figure 3.4 Ellipsometric thicknesses of 2.5-bilayer G8 dendrimer/PAA ﬁlms
as a function of the pH used for dendrimer and PAA deposition. Standard
deviations of thicknesses measured on three different samples ranged from 2 to
29 % of the measured values.

86

carboxylate- and amino-terminated dendrimers showed little dependence of thickness on
the molecular weights of polycations and polyanions.4'2"45 However in those cases, ﬁlms
were relatively thin and there would have been little need from chains to extend to
accommodate incoming polyelectrolytes.

Surface Roughness. One might expect that very thick polyelectrolyte ﬁlms
would be quite rough. Figure 5 contains optical microscope images of 2.5-bilayer G8
dendrimer (pH=4)/PAA (pH=4) and 2.5-bilayer G8 dendrimer (pH=8)/PAA (pH=4)
ﬁlms. As the images show, ﬁlms deposited with the dendrimer solution at pH 4
(thickness=80 A) appear smooth, whereas a dendrimer deposition pH of 8
(thickness=4,000 A) produces a rough surface.

Quantitative measurements of surface morphology using AF M also show the
roughness of thick ﬁlms. Differences in the surface morphology can be expressed in
terms of a roughness parameter, Rq, which is the standard deviation of the Z values
(heights) within a given area.

R. : Jan—2.32

Np

 

 

(1)

Values of R, were calculated using equation 1, where Z.- is a speciﬁc Z value, Zavg is the
average of the Z values within the given area, and Np is the number of points within a
given area.46 The surfaces of 2.5-bilayer G8 dendrimer (pH=4)/PAA (pH=4) and 2.5-
bilayer G8 dendrimer (pH=8)/PAA (pH=4) ﬁlms had roughnesses of 2.9 i 0.5 nm and
12.5 i 2.6 nm, respectively for 1 pm x 1 11m spots. This is consistent with the optical
microscopy data and suggests that in the case of thicker ﬁlms, there is some aggregation

of dendrimer during ﬁlm growth.

87

 

Figure 3.5 Optical microscope images of (a) 2.5-bilayer G8 dendrimer
(pH=4)/PAA (pH%) and (b) 2.5-bilayer G8 dendrimer (pH=8)/PAA
(pH=4) ﬁlms. The ellipsometric thicknesses of the ﬁlms were 80 A and
4000 A, respectively.

88

Microporous ﬁlm formation. In addition to their pH-dependent growth
properties, PAH/PAA ﬁhns are rather unique because they can form microporous
structures upon exposure to solutions with different pH values38 or ionic strengths.47
Dendrimer/PAA ﬁlms are also capable of forming such microporous ﬁlms. Figures 6(a)
and 6(b) show AF M images of 5-bilayer G4 dendrimer (pH=8)/PAA (pH=4) ﬁlms
prepared in the presence of 0.2 M NaCl before and after immersion in a pH 2.5 solution
for 1 min and washing with pure water. Exposure to acidic solution results in micropores

(0.02 - 0.07 pm in diameter) in the initially homogeneous ﬁlm. Figure 6(c) shows that

larger pores (0.1 - 0.4 um diameter) form when ﬁhns are prepared in the absence of salt.
Rubner suggested that these changes in morphology occur because the low pH results in
cleavage of interchain ionic bonds through protonation of the carboxylate groups of
PAA.”47 This, in turn, allows large scale reorganization and phase separation from the
acidic water solvent, yielding micrometer-sized pores. This reorganization appears to be
independent of whether the cation is PAH or a dendrimer, but it does depend on the ionic
strength used in ﬁlm deposition. The addition of supporting electrolyte should screen
charges on the ﬁlms and result in a structure with more loops and tails.4 Such a loopy
structure could result in the smaller pores seen for ﬁlms deposited in 0.2 M NaCl and

exposed to pH 2.5 solution.

89

10.0 n-

 

 

Figure 3.6 AFM images of 5-bilayer G4 dendrimer (pH=8)/PAA (pH=4) ﬁhns
before (a) and after (b, c) immersion in a pH 2.5 aqueous solution. For images
(a) and (b), polyelectrolytes were deposited in the presence of 0.2 M NaCl, and
for image (c), no salt was used during ﬁlm deposition.

90

Permeability of Dendrimer/PAA Layers. Because dendrimers have a
relatively open interior, we thought that dendrimer/PAA membranes might be highly
permeable to gases. To test the permeability of these ﬁlms, we deposited 4.5-bilayer G4
dendrimer (pH=8)/PAA (pH=4) ﬁlms on porous alumina substrates. Cross-sectional
ﬁeld-emission scanning electron microscope (FESEM) images (Figure 7) show that these
thick ﬁlms cover the substrate effectively without ﬁlling pores.48 Interestingly, ﬂuxes of
all the gases that we tested (He, O2, N2, H2, CO2, and CH4) were larger than the detection
limit of our gas ﬂow meter (>030 mL/cm2 sec), even at a gas pressure of 10 psig. This is
in stark contrast to 2-bilayer PAH/PAA ﬁlms (thickness of 100 A), which allow very low
ﬂuxes at 10 psig gas inlet pressures (0.003 to 0.04 mL/cm2 sec for He, 02, N2, H2, CO2,
and CH4). The dendrimer/PAA ﬁhn is about 35 times thicker than the PAH/PAA and yet
it allows ﬂuxes that are several orders of magnitude higher than those through PAH/PAA.
We also tested gas permeation through 5, 4, and 3-bilayer PAH/PAA ﬁlms for this
comparison, but ﬂuxes through these ﬁlms were below the limit of detection of our ﬂow
meter (<6 x 10 “5 mL/cm2 sec). At this point, we do not know if the high ﬂux through
dendrimer/PAA ﬁlms is due to an open structure of the dendrimer,” loose packing of the
ﬁlm or the presence of defects. However, FESEM images suggest that the ﬁhns are
defect-free. The high ﬂux through dendrimer/PAA ﬁlms suggests that they may prove
useful in providing a continuous substrate on which to deposit highly selective, ultrathin
polyelectrolyte gas-separation layers.50 The minimal thickness of the selective layer

would allow high ﬂux in addition to high selectivity.

91

jf~ tetﬁymm x7008 sew) 6/2319912 17

 

Figure 3.7. Cross-sectional FESEM images of porous alumina

before (a) and alter (b) coating with a 4.5-bilayer G4 dendrimer
(pH=8)/PAA (pH=4) ﬁlm.

92

3.5 Conclusions

The thicknesses of multilayer dendrimer/PAA ﬁlms depend on both dendrimer
generation and pH. For a given dendrimer, the thickest ﬁlms result from deposition of
dendrimers at high pH and PAA at pH 4, and ﬁlms prepared with G8 dendrimers are
generally 2 - 4-fold thicker than ﬁlms made with G4 dendrimers. Because multilayers of
dendrimer adsorb to the surface in a single deposition step, PAA likely reorients in the
ﬁlm to accommodate the dendrimers. Lower molecular weight PAA is less capable of
accommodating dendrimers and thus yields thinner ﬁlms. Although dendrimer/PAA
ﬁlms can be very thick, they are also highly permeable to gases, perhaps because of their
open structure. Additionally, if desired, microporous structures can be fabricated by brief
exposure of ﬁlms to pH 2.5 aqueous solutions followed by rinsing with water. The high
permeability and thicknesses of dendrimer/PAA ﬁlms show that the structure of these
materials is signiﬁcantly different from that of PAH/PAA. Thus dendrimer-containing

ﬁlms should provide expanded versatility for multilayer polyelectrolyte deposition.

93

3.6 References

(1)
(2)
(3)
(4)

(5)

(6)

(7)

(8)
(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

Decher, G. Science 1997, 277, 1232-1237.
Iler, R. K. J. Colloid Interface Sci. 1966, 21, 569-594.
Ferreira, M.; Rubner, M. F. Macromolecules 1995, 28, 7107-7114.

Ldsche, M.; Schmitt, J .; Decher, G.; Bouwman, W. G.; Kjaer, K. Macromolecules
1998, 31, 8893-8906.

Ostrander, J. W.; Mamedov, A. A.; Kotov, N. A. J. Am. Chem. Soc. 2001, 123,
1 101-1110.

Lvov, Y.; Ariga, K.; Ichinose, 1.; Kunitake, T. J. Am. Chem. Soc. 1995, 11 7,
6117-6123.

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

Decher, G.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210, 831-835.
Schlenoff, J. B.; Ly, H.; Li, M. J. Am. Chem. Soc. 1998, 120, 7626-7634.
Sukhishvili, S. A.; Granick, S. J. Chem. Phys. 1998, 109, 6861-6868.

Sukhorukov, G. B.; Schmitt, J .; Decher, G. Ber. Bunsen-Ges. Phys. Chem. 1996,
100, 948-953.

Yang, X.; Johnson, 8.; Shi, J.; Holesinger, T.; Swanson, B. Sensors and
Actuators, B: Chemical 1997, 45, 87-92.

Monterel, M. M.; Sukhorukov, G. B.; Petrov, A. 1.; Shabarchina, I.; Sukhorukov,
B. I. Sensors and Actuators B: Chemical 1997, 42, 225-231.

Harris, J. J.; Stair, J. L.; Bruening, M. L. Chem. Mater. 2000, 12, 1941-1946.
Krasemann, L.; Tieke, B. Langmuir 2000, 16, 287-290.

Farhat, T. R.; Schlenoff, J. B. Electrochemical and Solid-State Lett. 2002, 5, B13-
B15.

Dai, J.; Sullivan, D. M.; Bruening, M. L. Ind. Eng. Chem. Res. 2000, 39, 3528-
3535.

Caruso, F.; Yang, W.; Trau, D.; Renneberg, R. Langmuir 2000, 16, 8932-8936.

94

(19)
(20)

(21)

(22)
(23)

(24)

(25)
(26)
(27)

(28)

(29)

(30)

(31)

(32)

(33)
(34)

(35)

(36)

(37)

Fou, A. C.; Rubner, M. F. Macromolecules 1995, 28, 7115-7120.
Tsukruk, V. V. Adv. Mater. 1998, 10, 253-257.

Tsukruk, V. V.; Rinderspacher, F.; Bliznyuk, V. N. Langmuir 1997, 13, 2171-
2176.

Watanabe, S.; Regen, S. L. J. Am. Chem. Soc. 1994, 116, 8855-8856.
Tomalia, D. A.; Durst, H. D. Top. Curr. Chem, 1993, 165, 193-313.

Li, J.; Piehler, L. T.; Qin, D.; Baker, J. R., Jr.; Tomalia, D. A.; Meijer, D. J.
Langmuir 2000, 16, 5613-5616.

Bliznyuk, V. N.; Rinderspacher, F.; Tsukruk, V. V. Polymer 1998, 39, 5249-5252.
Kovvali, A. S.; Sirkar, K. K. Ind. Eng. Chem. Res. 2001, 40, 2502-2511.
Dillon, R. E. A.; Shriver, D. F. Chem. Mater. 2001, 13, 1369-1373.

Liu, Y.; Bruening, M. L.; Bergbreiter, D. E.; Crooks, R. M. Angew. Chem. Int.
Ed. Engl. 1997, 36, 2114-2116.

Wang, J.; Chen, J.; Jia, X.; Cao, W.; Li, M. Chem. Commun. 2000, 6, 511-512.

Bosman, A. W.; Bruining, M. J .; Kooijman, H.; Spek, A. L.; Janssen, R. A. J .;
Meijer, E. W. J. Am. Chem. Soc. 1998, 120, 8547-8548.

Tokuhisa, H.; Zhao, M.; Baker, L. A.; Phan, V. T.; Dermody, D. L.; Garcia, M.
E.; Peez, R. F.; Crooks, R. M.; Mayer, T. M. J. Am. Chem. Soc. 1998, 120, 4492-
4501.

Anzai, J. I.; Kobayashi, Y.; Nakamura, N.; Nishimura, M.; Hoshi, T. Langmuir
1999, 15, 221-226.

Yoon, H. C.; Kim, H. S. Anal. Chem. 2000, 72, 922-926.

Cheng, L.; Cox, J. A. Electrochem. Commun. 2001, 3, 285-289.

He, J. A.; Valluzzi, R.; Yang, K.; Dolukhanyan, T.; Sung, C. M.; Kumar, J .;
Tripathy, S. K.; Samuelson, L.; Balogh, L.; Tomalia, D. A. Chem. Mater. 1999,
11, 3268-3274.

Khopade, A. J .; Caruso, F. Nano Lett. 2002, 2, 415-418.

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

95

(38)

(39)

(40)
(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

(50)

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

Nagale, M.; Kim, B. Y.; Bruening, M. L. J. Am. Chem. Soc. 2000, 122, 11670-
11678.

Schlenoff, J. B.; Dubas, S. T. Macromolecules 2001, 34, 592-598.
Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153-8160.

Hoogeveen, N. G.; Cohen Stuart, M. A.; Fleer, G. J. Langmuir 1996, 12, 3675-
3682.

Tomalia, D. A.; Naylor, A. M.; Goddard III, W. A. Angew. Chem. Int. Ed. Engl.
1990, 29, 138-175.

Growth of G4 dendrimer (pH=8)/PAA (pH=4) ﬁlms using PAA with a molecular
weight of 2000 resulted in cloudy ﬁlms after deposition of only 2 bilayers. Thus
we could not determine whether PAA molecular weight affects the thicknesses of
these ﬁlms.

Cheung, J. H.; Stockton, W. B.; Rubner, M. F. Macromolecules 1997, 30, 2712-
2716.

Singh, S.; Khulbe, K. C.; Matsuura, T.; Rarnamurthy, P. J. Membr. Sci. 1998,
142,111-127.

Fery, A.; Scholer, B.; Cassagneau, T.; Caruso, F. Langmuir 2001, 17, 3779-3783.

Thicknesses of these ﬁlms are reasonably consistent with thicknesses of ﬁlms on
gold-coated wafers.

Jansen, J. F. G. A.; de Brabander-van den Berg, E. M. M.; Meijer, E. W. Science
1994, 266, 1226-1229.

Sullivan, D. M.; Bruening, M. L. Chem. Mater 2002, in press.

96

Chapter 4

THE USE OF DENDRIMER/POLY(ACRYLIC ACID) FILMS AS
ADHESIVES

Summary

Dendrimer-mediated adhesion is a convenient method for bonding substrates
because it provides a thin adhesive layer that can be deposited from aqueous solution.
This study shows the use of amine-terminated poly(amidoamine) (PAMAM) dendrimers
as an adhesive interlayer between poly(acrylic acid) (PAA)-tenninated surfaces.
Speciﬁcally, gold-coated silicon wafers modiﬁed with 5 G4 dendrimer/PAA bilayers on 5
precursor PAH/PAA bilayers are bonded with a G4 dendrimer interlayer. Cross-sectional
ﬁeld-emission scanning electron microscope (FESEM) images show that the dendrimer
interlayer bonds the surfaces together without leaving gaps between the substrates. The
adhesive forces between the substrates are most likely the strong electrostatic interactions
that exist between the exterior amine groups of protonated G4 dendrimers and the —COO'
groups of ionized PAA. Tensile tests show that the bonding strength is >1.6 MPa. The
thin adhesive layer along with the high bonding strength may allow use of variations of
this technique for applications such as wafer bonding to form lab-on-a-chip or micro-

electro-mechanical-systems (MEMS).

97

4.1 Introduction

Bonding of intrinsically nonadherent materials is generally achieved using a
variety of adhesives such as epoxies "2 and cyanoacrylates.3 However, for some
application such as wafer bonding to form three-dimensional micromachined devices,
deposition of a thick layer of adhesive is not acceptable, and thin adhesion layers may
prove useful.4'7 It is, however, difﬁcult to utilize organic monolayers for adhesion due to a
lack of contact between opposite monolayers on even slightly rough surfaces.8’9 We
surmised that the deposition of slightly thicker PAA/PAMAM dendrimer adhesion layers
might be useful for wafer-to-wafer bonding because of the possibility of interrningling of
layers and strong electrostatic interactions between protonated dendrimers and ionized
PAA. Such a process would require no heating and no organic solvents.

Traditionally, wafer-bonding has been achieved in high-temperature processes
using direct bonding, ﬁeld-assisted silicon-to-glass bonding (anodic bonding) or eutectic
bonding between silicon wafers via a thin metal interlayerfuo Because of the difﬁculties
associated with these procedures (e. g., the need for strict control over ambient conditions,
rigorous chemical cleaning and conditioning of the surfaces, and high temperatures),
simple bonding using organic adhesion layers is highly desirable. With respect to
traditional wafer-bonding methods, the main advantages of organic adhesives are the
relatively low bonding temperatures, the absence of electric voltage or current, and the
ability to join practically any kind of wafer materials.5

A few recent studies demonstrated that the multidentate nature of dendrimers can

be useful in promoting adhesion. Zamborini and co-workers recently described the use of

dendrimer monolayers as adhesion promoters between vapor-deposited gold ﬁlms and

98

glass or silicon wafers.11 Interaction of the PAMAM dendrimers with both the substrate
and the gold ﬁhns should increase with the number of multidentate interactions, and thus
the high density of amine groups at the densely packed surface of G8 dendrimers (1,024
primary amine groups per dendrimer) provides strong adhesion. Street et al. also noted
that the presence of a self-assembled monolayer of amine-terminated G8 dendrimers on
glass results in decreased surface roughness and increased hardness of an evaporated gold
ﬁlm. 12,13

Because dendrimer/silicon and dendrimer/ glass interactions are relatively weak,ll
we thought that the use of dendrimer/poly(acrylic acid) (PAA) ﬁlms might enhance the
adhesion properties of dendrimer-containing ﬁlms and allow wafer-to-wafer bonding.
Poly(acrylic) acid adheres well to charged surfaces, and we hoped to use electrostatic
interactions (between dendrimers and PAA as well as PAA and the substrate), rather than
coordination, to enhance adhesion. In this chapter, we present a preliminary study of
adhesion promoted by a protonated G4 dendrimer layer between two surfaces coated with
partially ionized PAA-terminated dendrimer/PAA ﬁlms. Gold-coated silicon wafers are
bonded using this dendrimer adhesion interlayer after coating the surface with S-bilayer
dendrimer/PAA ﬁlms using alternating electrostatic adsorption. The ~l .3 um-thick
dendrimer/PAA multilayer ﬁlm and the electrostatic attraction between dendrimers and
PAA yields bonding strengths as high as ~1.6 MPa. In fact, epoxied joints between the
wafer and the testing apparatus fracture before the adhesion layer does.
4.2 Experimental

Materials. Poly(acrylic acid) (MW = 90,000, 25 wt % aqueous solution),

poly(allylamine hydrochloride) (Mw=70,000), 3-mercaptopropioinc acid (MPA) and a 10

99

wt % solution of G4 amine-terminated PAMAM dendrimer (Mw = 14,215) in methanol
were purchased from Aldrich and were used as received. 18 MQ-cm Milli-Q water was
used in solution preparation. Gold-coated silicon wafers (200 nm of sputtered gold on 20
nm chromium on Si (100)) were used as substrates for ellipsometric measurements,
external reﬂection F TIR spectroscopy, and wafer-bonding studies. PAH (0.02 M,
pH=4.0) and PAA (0.02 M, pH=4.0) solutions contained 0.5 M NaBr and 0.5 M NaCl,
respectively, and solution pH values were adjusted by adding NaOH or HCl. Dendrimer
solutions (10'5 M) were adjusted to pH 8.0 and contained no supporting electrolyte. (The
molarities of PAA and PAH are given with respect to the repeating unit, while dendrimer
molarity is given with respect to polymer molecular weight.)

Film Deposition. Gold-coated silicon wafers were initially cleaned in piranha
solution (3:1 H2804:H202-caution-strong oxidizer, store waste piranha only in loosely
capped bottles) for 1 minute, rinsed with deionized water, and dried with N2. Cleaned
gold-coated silicon wafers were ﬁrst exposed to a solution of 1 mM mercaptopropionic
acid (MPA) in ethanol for 30 min, rinsed with ethanol and deionized water, and dried
with nitrogen to produce a —COOH-terminated surface. The substrates were then
alternately immersed in solutions containing PAH and PAA for 5 min with rinsing (l min
with a stream of deionized water) between adsorption of each polyelectrolyte. After
formation of S-bilayers of PAH/PAA, the precursor ﬁlm-coated wafers were immersed in
an aqueous
10'5 M G4 dendrimer solution for 10 min, washed with water, and dried with N2. The
substrates were then immersed in a PAA solution for 5 min, rinsed, and dried with N2.

This procedure was repeated until 5 G4 dendrimer (pH = 8)/PAA (pH=4) bilayers were

100

deposited on the precursor 5-bilayer PAH/PAA ﬁlm. For the wafer-to-wafer bonding,
ﬁlms were prepared on two gold-coated wafers with areas of 1.6 cm2, and then one drop
of 10'5 M G4 dendrimer solution (pH=8, no salt) was placed on one of the wafers, after
which the two wafers were simply pressed together by hand for about 20 min. Shorter
amount of pressing time may work to make wafer-to-wafer bonding.

Characterization. External reﬂectance FTIR spectra of polyelectrolyte ﬁlms
on gold-coated silicon wafers were obtained using a Nicolet Magna-IR 560 spectrometer.
Ellipsometric thickness determinations were made with a rotating analyzer ellipsometer
(Model M-44: J. A. Woollam) using WVASE32 software, and thicknesses were
determined at three separate areas on each sample. Field-emission scanning electron
microscopy (FESEM) (Hitachi S-4700) was used to take cross-sectional images of
bonded wafers, and testing of bonding strength was performed using a United SFM-20
testing machine, manufactured by the United Calibration Corporation. A strain-gauge
type load cell measures the load applied to the sample, and an extensiometer measures
the strain imposed on the sample. Increasing load was applied to the sample at a rate of
158 lbs/min. The bonded wafers were attached to two aluminum posts with epoxy so that

a load could be applied to the system.

4.3 Results and Discussion

Film formation and characterization. In chapter 3, we showed that the
highest thicknesses of dendrimer/PAA ﬁlms occur when using a deposition pH of 4 for
PAA and 8 for dendrimers.l4 These high thicknesses likely result from a lower charge

density on both PAA and dendrimers at these pH values and the resultant need for more

101

adsorption to compensate the surface charge.”’” We chose these two pH values for this
work so that dendrimer/PAA ﬁlms would be sufﬁciently thick for wafer bonding. Figure
4.1 presents schematically multilayer dendrimer/PAA ﬁlms on two gold-coated silicon
wafers prior to bonding. Addition of the dendrimer interlayer presumably zips the ﬁlms
together as shown in Figure 4.1. We used 5-bilayer PAH/PAA ﬁlms as a precursor
because this allows formation of slightly thicker dendrimer/PAA coatings. However, the
use of this precursor is probably not necessary.

Figure 4.2 shows ellipsometric thicknesses of G4 dendrimer (pH = 8)/PAA (pH =
4) ﬁlms on a base of 5 PAH (pH = 4)/PAA (pH = 4) bilayers as a function of the number
of dendrimer/PAA bilayers. The formation of dendrimer/PAA ﬁlms on the base layer
occurs largely due to the electrostatic attractions between negatively charged PAA and
positively charged G4 dendrimers. Ellipsometry demonstrates that thickness increases
rapidly and linearly with the addition of each new bilayer after deposition of the base 5-
bilayer PAH/PAA ﬁlm, and suggests that a consistent surface charge distribution has
already fully developed after adsorption of one dendrimer/PAA bilayer."”18 The average
thickness of dendrimer/PAA bilayers (excluding the ﬁrst bilayer) was ~ 280 nm. This
value is about ~1.3 times that of bilayer thicknesses for similar dendrimer/PAA ﬁlms
deposited directly on gold modiﬁed with mercaptopropionic acid. The extra thickness
with the base PAH/PAA ﬁlms may occur because the presence of salt during base layer
deposition results in a loopy structure that allows excess polymer to be accommodated
over several polymer layers.15 Regardless of the reason, deposition on a base layer easily

allows ~l .3 rim-thick ﬁlms to adsorb onto a gold-coated silicon wafer.

102

  
  

 

 

5dendrimer/PAA< .1“: ‘M‘Tf'
bilayers

 

 

 

5 dendrimer/PAAy
bilayers

 

 

 

Si Wafer

 

,. ‘ '+, 11+;

base /
5 PAH/PAA bilayers

Figure 4.1 Strategy for wafer bonding using dendrimer/PAA ﬁlms. The inset on the
right shows intertwining of dendrimer/PAA bilayers that should occur throughout the
ﬁlm.

103

1400

 

1200 -

A 1000

800

600

Thickness(nm
+

400 — 9

O
O L = 1 .
Base 1 2 3 4 5

Bilayers(dendrimer/PAA)

 

 

 

Figure 4.2 Ellipsometric thicknesses of dendrimer/PAA ﬁlms deposited on a base
composed of 5 bilayers of PAH/PAA. Deposition pH values were 8 for G4
dendrimers and 4 for PAA. The PAH/PAA base was deposited at pH 4.

104

FESEM Characterization of Bonded Wafers. After deposition of 5
dendrimer/PAA bilayers on a modiﬁed substrate, wafer-to-wafer bonding can be effected
by adding a drop of 10'5 M dendrimer solution to one wafer and pressing a second wafer
onto this surface using ﬁnger pressure for 20 min. Figure 4.3 shows cross-sectional
FESEM images of dendrimer/PAA ﬁlms on gold-coated wafers before and after bonding
two wafers together using the insertion of a dendrimer adhesion layer. These images
suggest that wafers bond with coatings in intimate contact. Using these FESEM images
we estimated ﬁlm thickness to be ~1.3 um, which is consistent with ellipsometric
measurements.

In addition to using insertion of a dendrimer interlayer to bond coated wafers, we
attempted to directly bond a wafer coated with 5 dendrimer/PAA bilayers to one that was
coated with 5.5 dendrimer/PAA bilayers (both dendrimer/PAA ﬁlms were deposited on a
base of 5 PAH/PAA bilayers). We thought that the interaction of a dendrimer-
terminated ﬁlm with a PAA-terminated ﬁlm might bond the wafers together. We used 1
drop of deionized water to promote interaction between the two ﬁlms. We found,
however, that the two wafers could be separated by hand, suggesting that little interaction
between the two ﬁlms occurred. Figure 4.4 shows a cross-sectional FESEM image of
this system before separating the two wafers. This image indicates that the dendrimer-
terminated surface and the PAA-terminated surface are not in intimate contact. In fact,
the gap between the two wafers coated wafers is ~10 um. Evidently, pressing these
films together was not sufﬁcient to initiate full contact between the wafers.

Bonding Strength Measurements. The schematic diagram in Figure 4.5

illustrates the conﬁguration used for testing the strength of bonding between wafers.

105

Aluminum rods (2.5 cm diameter, 2 cm height) were attached to each of the bonded
wafers with epoxy adhesive, and then force was slowly applied to pull the rods apart.
This test yields a value for the tensile strength, 0 ,5, which is expressed by 0 (5 = F max /A,
where F max is the maximum tensile load before fracture and A is the bonded area. The
dimensions of the bonded area were 1.25 X 1.25 cm. However, when breakage did
ﬁnally happen, it didn’t occur between the wafers, but rather between the epoxy and the
aluminum rod. Thus we could measure only a minimum tensile strength of >1 .6 MPa.
However, this value is still very high. The experiment was repeated with at least 2
samples, and they never broke at the junction between the wafers. This strong wafer
bond using a dendrimer adhesion layer is likely due to dendrimers that interact
electrostatically with both outer PAA layers. Multiple electrostatic interactions should
hold the layers together.

One question in this bonding is just how much dendrimer was added to the
surface, and would this amount be compatible with a ﬁll] intermingling of dendrimers and
outer PAA layers. Approximately 0.03 mL of 10’5 M dendrimer solution was applied to
the 1.6 cm2 wafer area. Assuming a dendrimer density of 1 g/cm3, this would result in
deposition of ~260 A of dendrimer. This is about 20 % of the amount of dendrimer
added during deposition of a dendrimer/PAA layer, and thus intermingling of the layers is

possible, because a thick interlayer composed solely of dendrimer need not form.

106

 

Figure 4.3 Cross-sectional ﬁeld-emission scanning electron
microscopy images of (a) 5 dendrimer/PAA bilayers on 5 PAH/PAA
bilayers on a gold-coated wafer and (b) two wafers coated with similar
ﬁlms that were pressed together with ~ 0.025 ml of dendrimer solution
between them.

107

   

5.5(dendn'mer/PAA) o?
5(PAH/PAA)

5(dendn’mer/PAA) oﬁ' '- ’-
5(PAH/PAA)
— E
10.0 u m

 

 

Figure 4.4 Cross-sectional ﬁeld-emission scanning electron microscopy image
of two ﬁlm-coated wafers that were pressed together with a drop of water
between them. The ﬁlm on the top gold-coated wafer was composed of 5.5
dendrimer/PAA bilayers on 5 PAH/PAA bilayers, while the ﬁlm on the bottom
gold-coated wafer contained 5 dendrimer/PAA bilayers on 5 PAH/PAA bilayers.

108

Applied Force

 

Aluminum Rod
/

Epoxy Adhesive

 

all”

Bonded Wafer Pair { . 7 -. _ .- 3<— Derxlrimer Adhesion Layer

.\

/ Aluminum Rod

 

Epoxy Adhesive

 

 

 

App lied Force

Figure 4.5 The tensile strength measurement conﬁguration. A tensile
tester is used to pull the bonded wafer apart.10

109

4.4 Conclusions

Deposition of G4 dendrimer between dendrimer/PAA ﬁlms capped with PAA
allows preparation of a 3 um-thick adhesive ﬁlm between wafers. Electrostatic
adsorption of protonated dendrimer in the two ionized PAA layers probably results in
electrostatic interactions between the ﬁlms. The adhesion strength of this bonding is
higher than that of the epoxy joint used to attach the wafers to the testing apparatus (>1.6
MPa). These results demonstrate the usefulness of dendrimer/PAA interfaces for
promoting adhesion.

Future work with these systems will include examination of how thick the
dendrimer/PAA layers on the two wafers need to allow wafer bonding. Thinner layers
will be desirable from many bonding applications. More work also needs to be done to
examine the mechanism of bonding in these systems. Some information about this could
be gained by changing the functionalities at the dendrimer surface and examining
whether adhesion still occurs. Investigation of the use of other polycations to promote
adhesion at these interfaces would also be illustrative. Additionally, although there are
technical challenges, a spectroscopic investigation of the interface would provide insight

into the chemistry that is occurring.

110

4.5 References

(1)
(2)
(3)
(4)

(5)

(6)
(7)

(8)
(9)

(10)

(11)

(12)

(13)

(14)
(15)
(16)
(17)

(18)

Amott, D. R.; Kindermann, M. R. J. Adhesion 1995, 48, 101-119.
de Neve, B.; Shanahan, M. E. R. J. Adhesion 1995, 49, 165-176.
Okamoto, Y.; Philip, K. J. Adhesion 1993, 40, 81-91.

Choi, W. B.; Ju, B. K.; Lee, Y. H.; Haskard, M. R.; Oh, M. H. J. Vac. Sci.
T echnol. B 1997, 15, 477-481.

Niklaus, B; Anderson, H.; Enoksson, P.; Stemme, G. Sensors and Actuators
A 2001, 92, 235-241.

Schmidt, M. A. Proceedings of the IEEE. 1998, 86, 1575-1585.

Tong, Q. Y.; Cha, G.; Gaﬁteanu, R.; Gosele, U. J. Microelectromech. Syst.
1994, 3, 29-30.

Kim, T.; Chan, K. C.; Crooks, R. M. J. Am. Chem. Soc. 1997, 119, 189-193.
Goss, C. A.; Charych, D. H.; Majda, M. Anal. Chem. 1991, 63, 85-83.

Tong, Q. Y. Semiconductor Wafer Bonding: Science and Technology;
John Wiley & Sons, Inc.: New York, 1998.

Baker, L.; Zamborini, F. P.; Sun, L.; Crooks, R. M. Anal. Chem. 1999, 71,
4403-4406.

Rar, A.; Zhow, J. N.; Liu, W. J .; Barnard, J. A.; Bennett, A.; Street, C. S.
Appl. Surface Sci. 2001, 175, 134-136.

Street, C. 8.; Rar, A.; Zhou, J. N.; Liu, W. J.; Barnard, J. A. Chem. Mater.
2001, 13, 3669-3677.

Kim, B. Y.; Bruening, M. L. Langmuir 2002, in press.

Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153-8160.
Schlenoff, J. B.; Dubas, S. T. Macromolecules 2001, 34, 592-598.
Shiratori, S. S.; Rubner, M. F. Macromolecules 1998, 33, 4213-4219.

Hoogeveen, N. G.; Cohen Stuart, M. A.; Fleer, G. J. Langmuir 1996, 12,
3675-3682.

111

Chapter 5

CONCLUSIONS AND FUTURE WORK

In this dissertation, we have developed of hyperbranched poly(acry1ic acid)
(PAA) and layered hyperbranched dendrimer ﬁlms as membrane skins and adhesives.
The work performed in chapter 2 shows that synthesis of hyperbranched PAA ﬁlms by a
graﬁ-on-graﬁ strategy yields a selective, ultrathin membrane skin. F ield-emission
scanning electron microscope and atomic force microscope images demonstrate that
hyperbranched PAA effectively covers porous alumina to form an ultrathin, selective
layer. Derivatization of PAA ﬁlms with H2NCH2(CF2)6CF3 results in increased Oz/Nz
and COz/CH4 selectivities for these composite membranes. Upon ﬂuorination, the ﬂuxes
of Oz and C02 increase, while the ﬂuxes of other gases such as CH4 and N2 are
comparatively unaffected. The selectivity of ﬂuorinated PAA membranes shows that
they are free of defects, as even a small percentage of defects would negate selectivity.
Because hyperbranched polymer membranes can be derivatized to contain a wide variety
of functional groups, they will be attractive for applications where functional, defect-free
membranes are required.

Although deposition of PAA ﬁlms on porous alumina substrates yields defect-free
materials, the alumina support is, in general, too fragile for practical applications. In the
future, it may be possible to construct composite ﬁlms on polymeric substrates that have
increased mechanical stability and durability. Possible supports include porous
polysulfone and polycarbonate,"5 but the surface of these materials would ﬁrst have to be
functionalized.°’7 Both Crooks and Bergbreiter showed that hyperbranched ﬁlms can be

formed on functionalized polymeric surfaces.8'9 Should a more continuous surface be

112

needed, ﬁlms could be deposited on poly[1-trimethylsilyl-1-propyne] (PTMSP) as
PTMSP is a glassy, highly permeable support that presents a continuous surface.lO
Coverage of underlying pores will not be a problem in this system. However, when
working with some polymeric supports, the synthetic procedure will need some
modiﬁcation because the substrates may be sensitive to the organic solvents used in
processing.

To continue to examine the transport properties of hyperbranched PAA, transport
of neutral molecules such as ortho-xylene, meta-xylene and para-xylene could be studied
using a non-aqueous solvent. Using an organic solvent such as cyclohexane may allow
the xylene isomers to be separated by steric effects, showing that hyperbranched PAA
membranes are selective to an organic system.

In chapter 3 of this dissertation, the thicknesses and morphology of multilayer
dendrimer/PAA ﬁlms showed the strong effect of deposition pH and dendrimer
generation on ﬁlm formation. The use of a deposition pH of 8 for dendrimers and 4 for
PAA resulted in G4 dendrimer/PAA bilayer thickness values as high as 2,000 A, while
G8 dendrimer/PAA bilayers were as thick as 4,000 A. At these solution pH values, the
relatively low charge densities of both polymers are important factors contributing to the
rapid grth of these ﬁlms.

Although both dendrimer/PAA and hyperbranched dendrimer/PAA ﬁlms contain
hyperbranched polymers, the structures of these two types of ﬁlms are very different.
Hyperbranched PAA does not have the controlled architecture that dendrimers do, and
thus gas separation membranes prepared from PAA behave similarly to conventional

polymers.” In contrast, dendrimer/PAA ﬁlms are extremely permeable to gases. Thus

113

these ﬁlms may prove to be very useful supporting layers for even thinner selective
layers. For example, Sullivan and Bruening recently demonstrated the formation of
ultrathin (35-40 nm) polyimide membranes using alternating polyelectrolyte
deposition.12 If such ﬁlms can be deposited on uniform, dendrimer/PAA ﬁlms, perhaps
defect-free polyimide ﬁlms as thin as 10 nm could be produced.

The wafer bonding studies in chapter 4 suggest that dendrimer mediated adhesion
is stronger than that obtained with a commercial epoxy glue. However, much more work
needs to be done in this area. First, we would like to limit the number of dendrimer/PAA
layers that need to be deposited. This would allow fewer processing steps as well as a
thinner adhesive layer. Second, we do not yet know the mechanism of adhesion. Do
dendrimers penetrate into each of the PAA surface layers? Further work using XPS
results and IR studies is needed in this area. In summary, this dissertation shows the
promise of hyperbranched ﬁlms for both separation and adhesion studies. However,
further work is needed to better understand these materials and bring applications to

fruition.

114

5.1 References

(1)
(2)

(3)

(4)

(5)

(6)

(7)

(8)
(9)

(10)

(11)

(12)

Hopkins, J .; Badyal, J. P. S. Langmuir 1996, 12, 4205-4210.

Schonenberger, C.; van der Zande, B. M.; Fokkink, L. G. J .; Henny, M.; Schmid,
C.; Kruger, M.; Bachtold, A.; Huber, R.; Birk, H.; Staufer, U. J. Phys. Chem. B.
1997, 101, 5497-5505.

Ramamoorthy, M.; Raju, M. D. Ind. Eng. Chem. Res. 2001, 40, 4815-4820.

Reid, B. D.; Ruiz-Trevino, F. A.; Musselman, I. H.; Balkus, K. J ., Jr.; Ferraris, J.
P. Chem. Mater. 2001, 13, 2366-2373.

Ghosh, K.; Chem, R. T.; Freeman, B. D.; Daly, W. H.; Negulescu, I. I.
Macromolecules 1996, 29, 4360-4369.

Dauginet, L.; Duwez, A. S.; Legras, R.; Demoustier-Champaign, S. Langmuir
2001, 1 7, 3952-3957.

Kim, I. W.; Lee, K. J.; Jho, J. Y.; Park, H. C.; Won, J.; Kang, Y. S.; Guiver, M.
D.; Robertson, G. P.; Dai, Y. Macromolecules 2001, 2908-2913.

Dermody, D. L.; Peez, R. F.; Bergbreiter, D. E. Langmuir 1999, 15, 885-890.

Zhao, M.; Liu, Y.; Crooks, R. M.; Bergbreiter, D. E. J. Am. Chem. Soc. 1999,
121, 923-930.

Parthasarathy, A.; Brumlik, c. J .; Martin, C. R.; Collins, G. E. J. Membr. Sci.
1994, 94, 249-254.

Toy, L. G.; Freeman, B. D.; Spontak, R. J .; Morisato, A.; Pinnau, I.
Macromolecules 1997, 30, 4766-4769.

Sullivan, D.; Bruening, M. L. Chem. Mater. 2002, in press.

115

I lllllllll lllllllllllllllllllllll'lll‘lllllllll’l

 

1293 02328 84