THESls

53001

 

__

LIBRARY
Ichigan State
University

This is to certify that the
dissertation entitled

THE STUDY OF SINGLE NANOPARTICLE AND MOLECULE
PHYSICS

presented by

TIFFANY EVA BOHNSACK

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemical Engineering and
Materials Science

 

 

 

 

l”

W/7/g/

Major Professor’s Signa ure
/ 2 J MN? é? W

Date

 

MSU is an affirmative-action, equal-opportunity employer

.—.—.—--.-.—.—.-.-.—--.—.-.---u-

..—n-I-o-o---o-o-o-c-o-I-t-I-u-n-u-I-o-I-o-n-o-I-n-IC

....-o—----- - -

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.

 

DAIEDUE

DAIEDUE

DATEDUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/C|RC/DaleDue.indd-p.1

 

THE STUDY OF SINGLE NANOPARTICLE AND MOLECULE PHYSICS
By

Tiffany Eva Bohnsack

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemical Engineering and Materials Science

2007

ABSTRACT
THE STUDY OF SINGLE NAN OPARTICLE AND MOLECULE PHYSICS
By
Tiffany Eva Bohnsack

We intend to use cross-linked, polymeric nanoparticles as a device to store
information when they are deformed (1) or in their native undeformed (0) state. To do
this, information about the interaction between the nanoparticles and different surfaces
must be determined. The substrates tested include a high energy mica surface and a low
energy silanized silicon wafer. The nanoparticles collapse on the mica substrate, but
remain robust and structured on the silanized wafer, yet an extreme amount of cross-
linking is required for the nanoparticles to retain their original spherical shape regardless
of the substrate surface energy. The nanoparticle behavior was also observed at elevated
temperatures to reveal that the height of the extremely cross-linked nanoparticles slowly
decreases. The temperature where a rapid size change occurs was well below the bulk
glass transition temperature, suggesting unique phenomena at the nanoscale. The
formation of ordered nanoparticle arrays is another essential aspect of molecular
technology and can be produced by using single-wall carbon nanotubes as a template.
Single wall carbon nanotubes serve as nucleation sites to focus nanoparticles toward them
through strong van der Waals forces that are enhanced from geometrical effects. This
interaction drives the nanoparticles to collect onto the nanotubes, which creates an
alignment of nanoparticles onto carbon nanotubes. In final studies the nanoparticles were

robustly attached to the surface through polymer film embedment. Embedding the

nanoparticles into a cross-linked thin polymer film locks the nanoparticles in place to

prevent disruption of the nanoparticles during deformation.

DEDICATION

This dissertation is dedicated to my parents, Larry and Diane Dukette, for their love,

guidance, and help through all of my school years.

iv

ACKNOWLEDGEMENTS

There are many people that I would like to thank for their help and support during
my graduate school studies. First, I would like to thank my advisor Professor Michael
Mackay for his help and guidance during the research process. I am grateful to the
members of the NIRT/CINCO team, specifically Karen Wooley and Brooke Van Horn
for educating me in the details of the polymer synthesis involved in this research and
Craig Hawker for providing the polystyrene nanoparticles used in my research. I would
like to acknowledge Professor Phil Duxbury and Dr. Erin McGarrity for their discussions
and help with the theory and physics entailed in this research. I would also like to thank
my group members David Bohnsack, Melissa Holmes, Anish Tuteja, R.S. Krishnan, and
Leslie Passeno for their help with laboratory experiments and research questions.

I would finally like to thank my family for all of their love and support throughout
these years. I am profoundly grateful to my parents, Larry and Diane Dukette, and my
brother and sister, Matthew and Kimberly Dukette, for helping me through all these years

of college and supporting me and being there every step of the way.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES .......................................................................................................... viii
Chapter 1: Introduction ....................................................................................................... 1
1.1 MOTIVATION FOR RESEARCH ............................................................................................................... 1
1.2 RESEARCH BACKGROUND .................................................................................................................... 3
Chapter 2: Conformation of Intramoleculary Cross-linked Polymer Nanoparticles on
Solid Substrates ................................................................................................................... 6
2.1 INTRODUCTION ..................................................................................................................................... 6
2.2 EXPERIMENTAL .................................................................................................................................... 8
2.3 RESULTS AND DISCUSSION ................................................................................................................. 12
2.4 CONCLUSION ...................................................................................................................................... 21
Chapter 3: The behavior of individual polystyrene nanoparticles and macromolecules at
elevated temperatures ........................................................................................................ 22
3.1 INTRODUCTION ................................................................................................................................... 22
3.2 BACKGROUND .................................................................................................................................... 25
3.3 EXPERIMENTAL .................................................................................................................................. 27
3. 3.1 A Softening Eflect Exhibited in Single Polystyrene Nanoparticles and Macromolecules .......... 27
3.3.2 Thermal Conformation Changes of Polystyrene Macromolecules ............................................. 29
3.4 RESULTS AND DISCUSSION ................................................................................................................. 31
3.4.1 A Softening Eflect Exhibited in Single Polystyrene Nanoparticles and Macromolecules .......... 31
3.4.1 Thermal Conformation Changes of Polystyrene Macromolecules ............................................. 46
Chapter 4: Directional Self Assembly: Nanoparticles Drawn to Carbon Nanotubes ....... 56
4.1 INTRODUCTION ................................................................................................................................... 56
4.2 EXPERIMENTAL .................................................................................................................................. 58
4.3 RESULTS AND DISCUSSION ................................................................................................................. 61
4.4 CONCLUSION ...................................................................................................................................... 70
Chapter 5: Embedding Nanoparticles into a Cross-linked Network ................................. 71
5. I INTRODUCTION ................................................................................................................................... 7 1
5.2. EXPERIMENTAL ................................................................................................................................. 73
5.3. RESULTS AND DISCUSSION ................................................................................................................ 76
5.4. CONCLUSION ..................................................................................................................................... 83
Chapter 6: Conclusions ..................................................................................................... 84
Appendix A: Atomic Force Microscopy ........................................................................... 86
Appendix B: Height vs. Temperature Data for Polystyrene Macromolecules and
N anoparticles ..................................................................................................................... 91
References ......................................................................................................................... 97

vi

LIST OF TABLES

Table 2.1 Sample, Molecular Mass, Polydispersity Index, Degree of Cross-Linking,
Nature (i.e., linear precursor or cross-linked nanoparticle), and Diameter if the molecules
collapse to the bulk density of polystyrene for the samples ................................................ 9

Table 3.1. Sample, Molecular Mass, Polydispersity Index, Degree of Cross-linking,
Nature (i.e. cross—linked or linear), and Diameter if the Macromolecules or Nanoparticles
Collapse to the Bulk Density of Polystyrene (Eq.3.2) for the Samples. ........................... 27

Table 4.1. Sample Code, Number Average Molecular Weight (kD), Solvent, Diameter if
the Macromolecules or N anoparticles Collapse to the Bulk Density of Polystyrene and
the AFM Height (Diameter) of Au and CdSe NPs ............................................................ 59

Table 4.2. Hamaker constants for typical substances used in our experiments ............... 67

Table 4.3. Hamaker constants and energies for typical substances and geometries used in
our experiments. ................................................................................................................ 68

Table 5.1. Sample, Molecular Mass, Polydispersity Index (PDI), Degree of Cross-
linking, Nature (i.e. cross-linked or linear), and Diameter if the Linear Precursors or
Nanoparticles Collapse to the Bulk Density of Polystyrene (Eq.5.l) for the Samples. 74

vii

LIST OF FIGURES

Figure 1.1. Using nanoparticles as data storage for a nano-chemical reactor. Each
nanoparticle can represent a bit of intormation. In its deformed state it represents a “1”
and in its undeformed state it signifies a “0”. ..................................................................... 2

Figure 2.1. (a) Intramolecular cross4linking of polystyrene macromolecules is
accomplished by opening a pendent cyclobutene (BCB) group and subsequent reaction
cross-linking with a similarly activated group. The nanoparticles are produced by
dripping a solution of linear-chain precursors into a hot solvent (benzyl ether) to activate
the cross-linking process. (b) Example of the nanoparticle. The molecule is representative
of a tightly cross-linked nanoparticle with every fifth monomer unit potentially cross-
linked. .................................................................................................................................. 9

Figure 2.2. (a) Atomic force microscopy height images of T-PSZl 1k polystyrene
nanoparticles on the mica substrate have an average height of 2.9 t 0.1 nm. The image
lateral dimensions are 3 x 3 am. (b) Atomic force microscopy height images of T-PSZl 1k
polystyrene nanoparticles on the Sigmacotem substrate have an average height of 9.3 i
0.2 nm. The image lateral dimensions are 3.5 x 3.5 am. (0) Should the nanoparticles
adopt a spherical conformation on the substrate, then their height (H) should equal their
diameter (D) which is 8.6 nm for the T-PSZl 1k system (see Table 2.1). The nanoparticles
may adopt a deformed shape on the substrate of equal volume to the sphere with a contact
diameter d as described by the J KR theory. ...................................................................... 12

Figure 2.3. Height variation Of lightly and tightly cross-linked nanoparticles (with
different molecular weights) on substrates of high (a) and low (b) surface energy. The
height values measured on the mica surface are much less than the predicted value, eq 2.1
(curve labeled H ~ Mm), whereas the values measured on the Si gmacoteTM surface
approach the prediction of a spherical object for the tightly cross-linked nanoparticles.
Height variation for three different moduli on the low-energy substrates is also Shown (eq
2.6) ..................................................................................................................................... 14

Figure 2.4. Height variation of two different cross-linked PS nanoparticles (20% and
60%), corresponding to the T-PS and E-PS systems, on high- and low-energy substrates.
It can be seen that the E-PS system faintly changes shape on the hi gh-energy (a) and low-
energy (b) substrate. Height variation for three different moduli on the high and low free
energy substrates is also shown (eq 2.6). A modulus of 3.2 GPa is the calculated modulus
for the extremely cross-linked system using eq 2.8 as described in the text. Note the open
triangle represents the height of the extremely cross-linked nanoparticle under the
assumption that dimers are present within the system. ..................................................... 16

Figure 3.1. (a) A repeat scan of height versus temperature for the 393kD polystyrene
macromolecules on Si gmacoteTM. Cycle 1 is the initial heights Obtained at 25°C and 75°C
and cycle 2 is the second observation of height versus temperature for the 393kD PS. In
cycle 2 the macromolecules were heated from room temperature to 75°C in increments of
10°C. The AFM height of the macromolecules was measured at each temperature. A
decrease in height is Observed with a transition at approximately 40°C. Note that the large

viii

error bars can be attributed to polydispersity and AFM error of 0.1 nm. (b) Distribution of
height values from 25°C to 75°C for the cycle 2 values graphed in (a). ........................... 32

Figure 3.2. Surface tension vs temperature for polystyrene.25 The surface tension of
benzyl ether at 200°C is also shown and is much less than the surface tension of
polystyrene in this temperature range (180°C — 240°C) to suggest that the polystyrene
does not wet the substrate (Si gmacotem) at 200°C since the contact angle of benzyl ether
on SigmacoteTM from 25°C to 200°C remains unchanged. ............................................... 34

Figure 3.4. Should the nanoparticles adopt a spherical conformation on the substrate,
then their height (H) should equal their diameter (D) which is 4.7 nm for the E-PS33kD
system. The nanoparticles may adopt a deformed shape on the substrate of equal volume
to the sphere with contact diameter d as described by the J KR theory. ............................ 38

Figure 3.5. Modulus vs. temperature plot Of 393kD polystyrene macromolecules on the
silanized substrate and 33kD E-PS nanoparticles plotted from Eq. 3.8. The modulus for
both systems was found for the second cycle values found on the silanized substrate. 41

Figure 3.6. For small values of a, the volume of the outer shell of monomers becomes a
large fraction of a spherical particle’s total volume fraction. ........................................... 42

Figure 3.7. AFM heights of 33kD E-PS NPS measured at room temperature after 5 heat
cycles to 75°C. Cycle 0 represents the initial height of the nanoparticles prior to heat.

The nanoparticle height continuously decreases in height as the cycle number increases
suggesting an increase in monomer adsorption to the substrate with each cycle ............. 43

Figure 3.8. (a) Height of 393kD polystyrene macromolecules at room temperature after
heating to the corresponding temperatures found on the graph. At 180°C the
macromolecule’s height increases significantly. (b) Height of 393kD PS macromolecules
at room temperature after being heated to the corresponding temperatures on the graph.
The height of the macromolecules fluctuates as the temperature is cycled between 75°C
and 200°C. The raw data is Shown as points to indicate the measurement errors caused by
polydispersity. All heights were obtained by AFM. ......................................................... 48

Figure 3.9. (a) Average Rg2 versus temperature for chains of length N = 124. (b) Average
macromolecule height from the surface as a function of temperature for N = 124. From
the inset, one can see the height begins to increase rapidly near T9. The scaled
experimental data is shown with error bars for comparison. ............................................ 51

Figure 3.10. <R02>/N versus temperature for different chain lengths (N: 64, 90, 124).
The crossing point for the length 90 and 124 chains gives a rough estimate of the 0
temperature for the model, T9 3 5.15. ............................................................................... 53

Figure 4.1. AFM height images of (a) SWNTS spin coated onto a silicon wafer, (b) E-
P833kD NPS deposited from chloroform on mica (c) T-PS211kD NPS deposited from
benzene on a Silicon wafer ((1) Au NPS deposited from toluene on mica. Note that for the
AFM images in b,c, and d the SWNTS were deposited onto the selected substrate prior to
nanoparticle deposition. .................................................................................................... 62

Figure 4.2. (a) and (b) TEM images of Au NPS aligned along SWNTS. (c) CdSe
nanoparticles aligned along SWNTS and (d) T-211kD PS NPS agglomerated along
SWNTS. ............................................................................................................................. 63

ix

Figure 4.3. (a) Spherical nanoparticle (NP) diffusing in solvent near a nanotube (NT) on
a substrate. (b-d) Coordinate systems for the geometries used in our systems: (b) '
sphere/Sphere (c) sphere/plate and (d) sphere/cylinder. .................................................... 65

Figure 5.1. (a) AFM height image of PSZSkD film on SigmacoteTM after being heated to
250°C for 30 min. (b) Line analysis of the scratch made in the film. The film appears to
be 4-5 nm thick. ................................................................................................................. 77

Figure 5.2. (3) Diagram of nanoparticle embedded in thin polymer film. (b) Graph of
the height of a T-PSZl 1k nanoparticle on Si gmacoteTM before embedment (black line)
and the height of a nanoparticle after embedment (dashed line). The average difference
in height between these particles is ~ 5—6 nm consistent with the film thickness
determined in Fig. 5.1. ...................................................................................................... 78

Figure 5.3. AFM height images of T-PS211k NPS embedded in a thin polymer film
(PS25kD linear precursor). (a) Height image prior to cure (50 ug/mL). (b) Height image
after heating to 250°C for 30 minutes (50 pig/mL). (c) Height image after the sample in
bwas dipped in benzene for 1 minute (SOug/mL). ((1) Height image after heating to 250°C
for 30 minutes (100 ug/mL). (e) Height image after heating to 250°C for 30 minutes
(200ug/mL). ...................................................................................................................... 79

Figure A.1. Atomic force microscopy diagram that illustrates the laser being emitted
from the diode to the cantilever and reflected back to the detector. The feedback unit uses
the input signal to produce the voltage that drives the piezoelectric ceramic. .................. 86

Figure A.2. (a) Example of a good AFM phase image of 33kD E-PS NPS. (b) Example of
a bad phase image of 33kD E-PS NPS. Some of the particles in the image appear
triangular which indicates the tip is broken. The streaks that appear in the phase image
indicate that either the surface is contaminated (e. g. finger print) or something is stuck to

the tip such as a dust particle or nanoparticle .................................................................... 87
Figure A.3. (a) AFM tip prior to tip approach. (b) AFM tip after tip approach. (c) AFM
tip after scanning a mica substrate. ((1) Damaged AFM tip. ............................................. 88

Figure A.4. AFM heater design. Samples can be heated to a maximum temperature of
100°C without damaging the AFM. .................................................................................. 89

Figure A.5. (a)AFM height image of nanoparticles aligned on single-walled nanotubes
before tip approaches sample (b) AFM height image of nanoparticles aligned on single-
walled nanotubes after tip approaches sample. In both (a) and (b) the black circle
represents the targeted tip approach area and in (b) the dotted circle represents where the
tip actually struck the sample. ........................................................................................... 90

Figure B.l. Height vs. temperature for 44kD polystyrene macromolecules heated from
25°C to 75°C. The sample was spray deposited from a solvent/non-solvent solution onto
a silanized wafer. ............................................................................................................... 92

Figure B.2. Height vs. temperature for 1.5MD polystyrene macromolecules heated from
25°C to 75°C. The sample was spray deposited from a solvent/non-solvent solution onto a
silanized wafer ................................................................................................................... 92

Figure 8.3. Height vs. temperature for 78kD T-PS nanoparticles heated from 25°C to
75°C. The sample was drop deposited onto a silanized wafer from benzene. .................. 93

Figure 8.4. Height vs. temperature for 211kD T-PS nanoparticles heated from 25°C to
75°C. The sample was drop deposited onto a silanized wafer from benzene. .................. 93

Figure 8.5. Height of 393kD PS macromolecules at room temperature after being heated
to the corresponding temperatures on the graph. The height Of the macromolecules
fluctuates as the temperature is cycled between 75°C and 200°C. The graph shows the
heights obtained after cycles 20 and 30. ........................................................................... 94

Figure B.6. Height vs temperature graphs of the 211kD T-PS NP system. (a) NPS were
deposited in pure benzene and (b) NPS were deposited in benzene/Z-MEA. The
nanoparticles do not undergo the same two phase behavior that is observed in the
macromolecules. ................................................................................................................ 94

Figure B.7. FTIR spectrum of 393kD PS macromolecules directly after preparation. A
0.1 ug/mL solution of 70% benzene/30% 2-MEA was sprayed onto a silanized wafer.
Only polystyrene peaks are present in the scan, there is no trace of residual solvent or
non-solvent ........................................................................................... 95

Figure B.8. XPS spectrum of 393kD PS macromolecules directly after preparation. A 0.1
ug/mL solution Of 70% benzene/30% 2-MEA was Sprayed onto a silanized wafer. The
silanized wafer was used for the background. There is no trace of any residual solvent or
non-solvent. ....................................................................................................................... 96

xi

Chapter 1: Introduction
1.1 Motivation for Research

Extensive research has been performed in the advancement of data storage
systems at the nanoscale. The motivation for this research stems from the eventual
application to the millipede."5 The millipede is a recent innovation in nanomechanical
data storage systems and consists of an array of cantilevers (64 x 64) and a micro-
mechanical scanner that moves the storage medium (polymer film) relative to the array of
cantilevers. With this system it is possible to read, write, and erase information within the
polymer film. The millipede stores information by making an indentation into a polymer
film with a heated atomic force microscopy (AFM) tip. Each indentation made into the
polymer film creates a bit of information on the surface. In order to erase the bit the
polymer film is reheated causing the polymer to soften and the stored elastic strain
together with surface forces to relax. The array is supported by a 10 x 10 mm silicon
wafer and can store greater than 1 terabit of information. To further advance this
technology one could replace the polymer film with individual nanoparticles possibly
allowing more information to be stored on a smaller area and decreasing fabrication
problems, such as erasure of a single bit of information.

The concept of replacing the polymer film with an array of nanoparticles is
investigated by studying the conformation of various nanoparticles and molecules. It is
hypothesized that each individual nanoparticle represents a bit of information; i.e., “1”
when deformed and a “0” in its initial state. An example of how this advancement of the
millipede can be used is shown in Figure 1.1. The picture represents a nano-chemical

reactor surrounded by lines of nanoparticles. These nanoparticles will then be used to

store information about the nano—chemical reactor by using “1”s and “0”s. A nanoparticle
in its deformed state (“1”) is Shown as a red particle and in its undeformed state (“0”) it is

represented as a green particle.

‘ ~100-200nm .

00100100

N ano

Chemical
Reactor

o—Y

Figure 1.1. Using nanoparticles as data storage for a nano-chemical reactor. Each
nanoparticle can represent a bit of intorrnation. In its deformed state it represents a “1”
and in its undeformed State it signifies a “0”.

OOOO-‘OOO

 

00100100

1.2 Research Background

In this work, we studied single polymer nanoparticle and molecule physics in
order to accomplish use of this potential technology. First, we studied the substrate
specific conformation of the nanoparticles. We recognize the initial conformation of
nanoparticles on solid substrates, prior to deformation, is dictated by their intrinsic
rigidity and interaction with a substrate which represents a critical aspect of this
technology. The substrate free energy is changed by choosing a high energy surface,
freshly cleaved mica, and a low energy surface, silanized silicon wafer.

Studying the nanoparticles’ behavior at high temperature is also critical to this
research. By observing an individual nanoparticle’s or molecule’s response to high
temperatures we can understand thermal transitions that occur in a single molecule or
nanoparticle. Quantifying a thermal transition of an individual molecule has been
scarcely researched and rarely observed. Currently the only known way to determine a
thermal transition of a single polymer molecule is by freeze drying an ultradilute solution
with the molecule in its expanded state and then examining a differential scanning
calorimetry (DSC) trace of the final product.6 The first scan shows that an individual
polystyrene molecule with molecular weight of 1.8 x 10‘5 Da has a glass transition
temperature (Tg) Of 41°C and 60°C on the second and third scans. These temperatures are
drastically lower than the Tg of bulk polystyrene which is 105°C. In the present study we
provide information on the behavior of individual nanoparticles at high temperatures
which is essential during the eventual deformation of the nanoparticles we will study on a

substrate.

Atomic force microscopy (AFM) was used to study the conformation of
nanoparticles on solid substrates at both ambient and elevated temperatures. AFM is a
scanning-probe system that uses a sharp tip (probe) to scan over a sample while
measuring local properties and creating a topographical image of the sample. Appendix A
contains a detailed description of AFM and the methods used in all experiments
described throughout this work.

Forming nanoparticle arrays is also an essential aspect in molecular technology.
In past research, arrays of particles have been formed by using focused ion beam milling
(FIB) to make patterns within a polymer film. Solutions of silicon dioxide particles were
then deposited onto the patterned wafer and the particles arranged in forms similar to the
patterned wafer.7 In other studies, nanoparticles were aligned along nanoscale ridge-and-
valley structured substrates.8'9 In this research nanoparticle arrays are formed by aligning
them against carbon nanotubes. Strong van der Waals interactions between a sphere
(nanoparticle) and cylinder (nanotube) cause the nanoparticles to be drawn to the
nanotubes. The nanoparticles collect onto the nanotubes creating a line of nanoparticles.
By aligning the nanotubes on the surface one could potentially use this technique to
create an array of nanoparticles.

In the final task, the nanoparticles were robustly attached to the surface to prevent
tip forces from disordering the nanoparticle array during deformation. At the present time
embedding the nanoparticles within a polymer film is the best way to achieve this.

Embedding nanoparticles into a cross-linked network of polymer film locks the
particles into position on the substrate. The polymer film layer is thin enough so that the

top of the nanoparticle is visible above the film. Researchers are currently exploring the

optical properties that are produced by embedding Ag nanoparticles into polymer films
for metal/semiconducting polymer systems,l°’ 1' but there is no current research being

done on trapping a portion of the nanoparticles into a polymer film for molecular memory

application.

Chapter 2: Conformation of Intramoleculary Cross-linked Polymer Nanoparticles
on Solid Substrates

2.1 Introduction

The first item addressed in this research is the substrate specific conformation of
the nanoparticles. We recognize the initial conformation of nanoparticles on solid
substrates, prior to deformation, is dictated by their intrinsic rigidity and interaction with
a substrate which represents a critical aspect of this technology. The substrate free energy
is changed by choosing a high energy surface, freshly cleaved mica, and a low energy
surface, silanized Silicon wafer.

The conformation of cross-linked, monomolecular, polystyrene nanoparticles on a
solid substrate is studied here as a function of cross—linking degree and substrate surface
energy.12 It is found that an extreme degree of cross-linking is required for the ca. 5-10
nm diameter polystyrene nanoparticles to retain their original spherical shape, regardless
of surface energy. The nanoparticles’ rigidity is studied on a mica and silanized silicon
wafer substrate. Both substrates are smooth at the Angstrom level thereby minimizing
roughness effects on measurements.

It should be noted that the conformation of larger organic nanoparticles on solid
substrates has been studied before by determining how they change size upon heating.13
Also, in 1965, prior to present day atomic force microscopy, Richardson used scanning
electron microscopy (SEM) to determine the molecular weight of individual molecules.l4
Polystyrene solution was sprayed onto a carbon-backed mica substrate and shadowed
with deposits of gold and palladium. The shadow length for each molecule could then be
determined through SEM images. To calculate the molecular weight of the individual

molecules, it was assumed that the particles were spherical and that the diameter of the

particles was equal to their height. Due to limitations, there are two major difficulties in
using this technique: defining the shadow limit cast by the spheres and the assumption of

a spherical conformation on the substrate.

2.2 Experimental

Polystyrene nanoparticles were synthesized at IBM by Hawker, et al. 15as
illustrated in Figure 2.1. A linear random copolymer consisting of styrene monomer and
benzocyclobutene (BCB) was first synthesized. The degree of cross-linking is dictated by
the BCB content. We use three BCB levels: 2.5, 20, and 60 mol%, denoted as lightly (L),
tightly (T), and extremely (E) cross-linked. A dilute solution of the linear precursor is
then dripped into a hot solvent activating the intramolecular cross-linking reaction to
create nanoparticles containing a single macromolecule whose size is dictated by the
initial precursor molecular weight and the degree of cross-linking. The sample codes,
molecular weights (Mn, number average molecular weight), polydispersity index (PDI,
ratio of weight to number average molecular weight), and degree of cross-linking for the
systems studied here are given in Table 2.1. The polystyrene nanoparticles’ molecular
weights were determined with a Wyatt “Dawn E08” 18 angle static light scattering

detector, and so are absolutely measured and not relative to calibration standards.

Crossllnklng group /'

Linear ChaIn Precursor

flA
Q

Nanopanicle

 

Figure 2.1. (a) Intramolecular cross-linking of polystyrene macromolecules is
accomplished by opening a pendent cyclobutene (BCB) group and subsequent reaction
cross-linking with a similarly activated group. The nanoparticles are produced by
dripping a solution of linear—chain precursors into a hot solvent (benzyl ether) to activate
the cross-linking process. (b) Example of the nanoparticle. The molecule is representative
of a tightly cross-linked nanoparticle with every fifth monomer unit potentially cross-

linked.

Table 2.1 Sample, Molecular Mass, Polydispersity Index, Degree of Cross—Linking,
Nature (i.e., linear precursor or cross-linked nanoparticle), and Diameter if the molecules
collapse to the bulk density of polystyrene for the samples.

 

 

mol %
sample Mn pm cross-linking nature diameter
agent (nm)
P833k 33.0 1.05 20 linear 4.7
P358k 58.0 1.14 20 linear 5.6
PSt 93k 193 1.28 20 linear 8.4
L-P825k 24.5 1.14 2.5 cross-linked 4.2
L-PSGOk 60.1 1.16 2.5 cross-linked 5.7
L-P8158k 158 1.4 2.5 cross-linked 7.8
T-PS41 k 41.0 1.04 20 cross-linked 5.0
T-PS78k 78.0 1 .14 20 cross-linked 6.2
T-P8211k 211 1.32 20 cross-linked 8.6
E-P833k 33.0 1 .91 60 cross-linked 4.7

Surface profile measurements were performed with a Pacific Nanotechnology
Nano-R atomic force microscope in close contact (oscillating) mode to generate height
images that were not altered other than a simple leveling procedure. Silicon tips with a
spring constant of 36 N/m, tip curvature of 10-20 nm, and a resonance frequency Of 286-
339 kHz were used for all experiments.

Sample preparation for AFM samples depended on the surface energy of the
substrate. Freshly cleaved mica and silanized silicon wafer substrates were used for all
experiments described below. The mica was cleaved and used immediately to avoid
contamination from dust, atmospheric contaminants, and ionic crystals that are attracted
to the surface due to mica’s hydrophilic nature.16 Because mica is a high free energy
substrate, which most solvents will wet, solutions of sample were Spin coated onto the
mica at 5000 rpm for 40 seconds. The surface energy of the mica is dependent on its
environment, with a value of ~4500 mN/m in high vacuum and ~300 mN/m in humid
laboratory air.17 The linear polymer precursors and nanoparticles were dissolved in
benzene, and all solutions were filtered with a 0.2 pm Teflon filter to reduce the amount
of large dust particles and atmospheric contaminants prior to deposition.

SigmacoteTM was used to silanize the silicon wafers by spin coating the
SigmacoteTM solution onto a silicon wafer at 5000 rpm for 40 seconds. SigrnacoteTM
(Sigma-Aldrich) is a solution consisting of 2.5% chlorosiloxane ((SiClzC4H9)20) and
97.5% heptane that functionalizes the surface with short alkane chains. The wafer was
then rinsed with Millipore water to eliminate the excess, and then pure benzene was spin
coated directly onto the wafer. The silanized wafer was then checked by AFM to ensure

the coating provided a smooth surface containing no precipitates or dust particles. Due to

10

Sigmacote’sTM low free energy surface (29 mN/m), the solutions could not be spin coated
directly onto the surface and therefore a drop of solution (concentration, 0.01 ug/mL) was

placed on the surface and exposed to air until the benzene had completely evaporated.

ll

2.3 Results and Discussion

The nanoparticle heights were determined with AFM by taking the average of 50
particle heights; however, the lateral size could not be found due to convolution effects
created by the tip."“ ‘9 In Figure 2.2 an example of a three-dimensional AFM height
image of the T-PS211k nanoparticles on the mica and SigmacoteTM substrates is shown.
The image clearly demonstrates uniformity of nanoparticles on the surfaces and the
narrow height distribution provides evidence that only individual polystyrene

nanoparticles are present and no agglomeration of them occurs.

 

Figure 2.2. (a) Atomic force microscopy height images of T-PSZl 1k polystyrene
nanoparticles on the mica substrate have an average height of 2.9 i 0.1 nm. The image
lateral dimensions are 3 x 3 am. (b) Atomic force microscopy height images ofT-PS211k
polystyrene nanoparticles on the SigmacoteTM substrate have an average height of 9.3 i
0.2 nm. The image lateral dimensions are 3.5 x 3.5 urn. (c) Should the nanoparticles
adopt a spherical conformation on the substrate, then their height (H) should equal their
diameter (D) which is 8.6 nm for the T-PS21 1k system (see Table 2.1). The nanoparticles
may adopt a deformed shape on the substrate of equal volume to the Sphere with a contact
diameter d as described by the JKR theory.

To ascertain the nanoparticle conformation on the substrate shown in Figure 2.2,
we determine the diameter (D) assuming the macromolecule collapses to a sphere with

the bulk density of polystyrene (p ~ 1.04 g/cm3) through

12

 

)6
D=[6M" ] (2.1)
WAX)

where NA is Avogadro’s number. The values are given in Table 2.1, and for sample T-
PS211k, Shown in Figure 2.2, the value of D (8.6nm) is much greater than its height on
the high-energy substrate (the average height is 2.9 :1: 0.1 nm). Thus, on the mica surface
the nanoparticles adopt a pancake-like conformation. This conformation is also observed
in dendrimer Structures where studies have shown that both charged and uncharged
dendrimers adsorb to a mica surface forming a flat disk structure.20

To ensure that the height was not an AFM artifact caused by the cantilever drive
amplitude (force), the amplitude was changed from 900 to 1500 mV when examining the
T-PSZl 1k nanoparticle sample. There was no change in the average nanoparticle height
due to the equivalent pressure change, and therefore, it is suspected the high surface
energy mica (~300 mN/m) is the most likely cause for the nanoparticles’ collapse onto
the surface. So, it is believed the set point or cantilever distance from sample did not play
a role in these observations since the lowest possible setpoint was used for all AFM
scans.

The effect of surface energy on the nanoparticle conformation is shown in Figure
2.3 and it is clear that the nanoparticles’ height is affected. The solid line in each graph,
labeled H ~ Mm, is a plot of eq 2.1 assuming the density is equal to that for bulk
polystyrene. The height of the linear precursor macromolecules and lightly cross-linked
nanoparticles are affected little by a change in the surface energy and fall below the value

for a robust sphere. However, the tightly cross-linked nanoparticles Show a drastic

difference in the height profile with molecular mass and approach the scaling suggested

13

by eq 2.1 on the low-energy substrate, SigmacoteTM. So, there is a clear interaction
between the substrate surface energy and nanoparticle stiffness implied through the

degree of cross—linking.

   

12 20%
. 20% 20% Unear
. - 2.5%
10 2 2%; Lmear 2.5%Llnear
A 8 2.5% Linear A
E E
5 5
E 6 E
.9 .9
or a)
:I: 4 :I:
2
0
0 40 80 120 160 200 0 40 80 120 160 200
Molecular Mass (kDa) Molecular Mass (kDa)

Figure 2.3. Height variation of lightly and tightly cross—linked nanoparticles (with
different molecular weights) on substrates of high (a) and low (b) surface energy. The
height values measured on the mica surface are much less than the predicted value, eq 2.1
(curve labeled H ~ Mm), whereas the values measured on the SigmacoteTM surface
approach the prediction of a spherical object for the tightly cross-linked nanoparticles.
Height variation for three different moduli on the low—energy substrates is also shown (eq
2.6).

Since the T-PS series of nanoparticles collapses on the high energy mica
substrate, a system was designed to ensure minimal collapse on any energy surface. This

was done by recognizing the work of Chubynsky and Thorpe21 who Show a rigid network
system is developed when the average coordination number (r) is 2.4 in a cross-linked
system. This value is determined by

(r) = 2x, + 3[1— xL] (2.2)

where, XL is the mole fraction of linear segments with coordination number 2 and 1 — XL,
the mole fraction of cross-linked sites with coordination number 3. According to this
theory, at least 40 mol % cross-linking, or two of every five monomer units, must be

cross-linked to ensure rigidity. The extreme cross-linked system was designed to surpass

this value with (r) = 2.6 where three of every five monomer units are cross-linked.

The extremely cross-linked nanoparticles were observed on both the high and low
surface energy substrates and as shown in Figure 2.4 the nanoparticles slightly change
shape on the hi gh-energy substrate. For reference, the T-PS series is also shown in Figure
2.4 and it is clear that these nanoparticles are distorted on the high energy substrate.The
open triangle is a representation of dimerization that could have occurred during
synthesis of the extremely cross-linked nanoparticles. The experimental error Observed in
both Figures 2.3 and 2.4 could be due to the nanoparticle’s polydispersity index (PDI)
effect on the height. For the E-PS33k polystyrene nanopartricle series, it is found that the
change in height based on PDI is 1.5 nm. The AFM height’s experimental error is only
0.1 nm, which suggests the experimental error is mostly due to the nanoparticles’
polydispersity. For a polydisperse system the molecular weight standard deviation (SM)
is given by MD x \/(PDI-1) which can be used to calculate the error in the diameter (6D);
SD/D=6M/3M, through propagation of error. Since the height data are influenced by
small and discrete random effects, such as polydispersity index (PDI) and AFM error
(~0.l nm), the data can be normally distributed using the Central Limit Theorem (CLT).
Using a sample Size of 50, the CLT can be used to recognize that a distribution of an
average approaches a normal distribution; thus, the normally distributed standard

deviation (ox) becomes

15

 

a, = r (2.3)

’21

where, <5x is the non-normal standard deviation and N is the sample size. Thus, the error
in the mean value (0,) is the measured error (0,.) divided by the square root of the sample

size (N).

 

12 _l.. ‘J W'Sigrnacote .l '. lb]

1

 

      

10 _ e 20% Nanoparticle i i 7
A 60% Nanoparticle =

OOMPa - - - E = 100 MPa ;
.— E=1 a .4......_._. .'. .. I—
—-E=3.2 GPa :

 

 

.2 GPa

 

Height (nm)
Height (nm)
0)

 

 

 

 

 

O 40 80 120 1&3 200 0 4O 80 120 160 200

Molecular Mass (kDa) Molecular Mass (kDa)

 

Figure 2.4. Height variation of two different cross-linked PS nanoparticles (20% and
60%), corresponding to the T-PS and E-PS systems, on high- and low-energy substrates.
It can be seen that the E-PS system faintly changes shape on the high-energy (a) and low—
energy (b) substrate. Height variation for three different moduli on the high and low free
energy substrates is also shown (eq 2.6). A modulus of 3.2 GPa is the calculated modulus
for the extremely cross-linked system using eq 2.8 as described in the text. Note the open
triangle represents the height of the extremely cross-linked nanoparticle under the
assumption that dimers are present within the system.

These results demonstrate that an extremely large degree of cross-linking is
necessary to stabilize polymeric nanoparticles on the high-energy substrate suggesting a
high modulus is required. Modeling of the linear polymer and nanoparticle conformation
on the substrate can be achieved using the classic Johnson, Kendall, and Roberts (JKR)

theory22 and the assumption that the adsorbate geometry adopts a spherical cap shape

16

(Figure 2.2c). The JKR theory is an extension Of the Hertz theory of elastic contact and is
used to explain the adhesion between two elastic bodies under a compressive force. In the
JKR theory it is assumed that elastic deformation must occur; however, studies
performed by Kendall and Padget have shown that the JKR theory applies to the
coalescence behavior of latex particles even though large deformations are present and
the material is not truly elastic.23 Given this observation, we use the JKR theory for a
sphere in contact with a flat surface, experiencing zero load, where the contact diameter

((1) becomes

d3 = 97r(1— 1201115402 (2.4)

with D being the predicted nanoparticle diameter (eq 2.1), E is Young’s modulus,v is

Poisson’s ratio, and WA is the work of adhesion given by
WA =7A +75 “7% ”2090'?de (2-5)

with 7 being the surface energy of the adsorbate-vapor (A), substrate-vapor (S), and
adsorbate-substrate (A/S) interfaces. The approximation is due to Fowke’s relation24 that
is applicable to lower energy substrates that interact through dispersive forces (the

superscript d represents the dispersive part of the surface energy or tension). In our case,
polystyrene25 interacts via dispersive forces and has yAd = 40 mN/m while the silanized
substrate has ysd z 28.6 i 0.6 mN/m at room temperature.

As stated above, the adsorbate geometry is assumed to be either a spherical cap or
a disk with equal (constant) density between the sphere (see eq 2.1) and the adsorbed

shape (see Figure 2.2c). For the spherical cap geometry one finds

17

d3: 8 (Da-H3)% (26)
s 373 [1% .

 

where H is the adsorbate’s height on the substrate. Combining the JKR theory and the
Spherical cap, one can determine the interrelation between the nanoparticle height and
surface and molecular properties

H3 +kH —1)3 =0 (2.7)

where

 

E

k :[277rs/3(1_V2)W %

_AD2
8

By solving for H, with varying D, we were able to generate height-molecular weight
relations for different modulus values as shown in Figure 2.4. Note that this model only
corresponds to the nanoparticles and linear precursors that have not collapsed.

The E-PS series can be further investigated since these nanoparticles exhibit
minimal collapse on the low- and hi gh-energy surfaces. Due to this phenomenon we can

assume D 2 H and therefore [D — H l —> 0. Using Eq. 2.7 and the assumed limit, H can be

expressed as

 

C
H=D l— 2.8
where
7
r: E02 adC:[%-)3
WA(1—v)

18

The term C, in eq 2.8 refers to the crumble number, which was developed by Kendall and
Padget, to serve as a guideline for the coalescence behavior of dispersed elastic particles.
The crumble number can be used to determine whether a latex film will become porous
and opaque (C, > 10) or tough, transparent, and nonporous (C, < 1).23 Using this

expression for the two substrates (labeled 1 and 2 for mica and Sigmacotem), one finds
Assuming U = 0.5 then the modulus is the only unknown parameter and we find E z 3.2
GPa. It is Shown in Fig. 2.4 that a model curve with a modulus of 3.2 GPa corresponds
well with the height value for the E-PS series on either substrate, especially for the case
of possible dimerization. Further, the modulus for the E-PS series proves to be
approximately equal to that for bulk polystyrene (:3 3 GPa), traditionally determined with
tensile testing or with an AFM.26 This may be expected due to the extreme degree of
cross-linking, yet, the system is a Single molecule and hence may be subject to finite size
effects.

We can also estimate the modulus for T-PS nanoparticles by noting that they
exhibit minimal collapse on the low energy substrate. Assuming C, > 10 for this system
(eq. 2.8) we estimate E > 100-150 MPa, where in fact we find E = 1 GPa agrees fairly
well with the height data (Fig. 2.3). Thus, increasing the degree of cross-linking 3-fold
produces a change in the modulus by about a half-order of magnitude.

The L-PS and T-PS, as well as the linear precursor polymers collapse on the
substrates, particularly on the high energy mica. To understand why their height becomes
so small (:3 nm) and essentially independent of molecular weight we consider the work

of Rubinstein and co-workers who studied single chain adsorption on surfaces,27 with an

19

emphasis on the adsorption of polyelectrolytes on charged surfaces.” 29 If the surface is
weakly adsorbing then the number of monomer units in contact with the surface will
increase in order to gain adsorption energy; therefore, the chain height on the substrate
will be smaller than its unperturbed size and lose conformational entropy. This
phenomenon is observed on the mica substrate for the L-PS and T-PS nanoparticle
systems where the number Of monomer units within the nanoparticle that are in contact
with the surface must increase causing the nanoparticles to spread out onto the surface
into a pancake-like conformation. The size of the adsorption blob (fiads), which is
equivalent to the chain height on the substrate, can be estimated by knowing the number
of monomers in each adsorption blob that are in contact with the surface and each

individual adsorption blob’s energy gain estimated through

'3'
&BT[-§id—=‘-) z kBT (2.10)
where &BT is the thermal energy gain of each individual monomer in contact with the
surface. Here b represents the Kuhn monomer length, which is 1.8 nm for polystyrene.
Assuming éads : H, one can find the adsorption energy per monomer for a linear polymer
or collapsed nanoparticle (H z 3 nm, see Fig. 2.3) to be 2 0.5kBT. Since dispersion forces

are 7: kBT in strength this estimate of the adsorption blob-substrate interaction energy is

expected as is the adsorbate conformation.

20

2.4 Conclusion

After investigating the interaction between nanoparticles on a low and high energy
substrate, it was determined that the conformation of nanoparticles changes significantly.
It was discovered that the L-PS and T-PS series collapse forming a pancake-like
conformation on a high energy substrate with an adsorption energy adequately described
by a theory developed for linear polymer systems. Conversely, on the low energy
substrate, the T-PS and E—PS series form robust spherical objects with a height that
corresponds to their molecular weight. Yet, the modulus is estimated to be a half-order Of
magnitude larger for the E-PS system. Further, only the E-PS molecule does not
significantly collapse on either substrate demonstrating an extremely rigid network is
required for the nanoparticle to retain its Shape regardless of substrate surface energy.
Since the motivation for this work requires a deformable nanoparticle, it is clear that
extreme cross-linking cannot be used and so the interplay of substrate surface energy and

nanoparticle modulus must be carefully considered.

21

Chapter 3: The behavior of individual polystyrene nanoparticles and
macromolecules at elevated temperatures

3.1 Introduction

The behavior of polystyrene nanoparticles and macromolecules was observed at
elevated temperatures to reveal that the height of the particles slowly decreases. The
temperature where a rapid size change occurs was well below the bulk glass transition
temperature suggesting unique phenomena at the nanoscale. A softening effect that
occurs in individual nanoparticles and macromolecules can be determined via AFM by
observing the height of the nanoparticles at different temperatures using an AFM hot
stage. In this work we found a softening point that occurs at ~40°C for all polystyrene
nanoparticles and macromolecules. This softening point is attributed to the scale of a
Single nanoparticle and macromolecule from that of the bulk. The JKR theory is used to
study the change in modulus of the particles as the surface tension of polystyrene changes
with temperature.

Thermal transitions of individual macromolecules are scarcely researched
experimentally because of the difficulty imposed by the length scales involved.
Currently, the only known way to determine the thermal transition of a single polymer
macromolecule is by freeze drying an ultra-dilute solution with the macromolecule in its
expanded state and then examining a differential scanning calorimetry (DSC) trace Of the
final product. Xue, et al.,30 used this technique to suggest that an individual polystyrene
macromolecule with molecular weight of 1.8 x 106 D has a glass transition temperature
(Tg) somewhere between 40°C and 60°C. These temperatures are drastically lower than

the Tg of bulk polystyrene which is 105°C.

22

A lower glass transition temperature has also been observed in thin film studies.
The Tg Of thin polymer films has been extensively studied using x-ray reflectivity,
ellipsometry, reflection absorption FTIR, and dielectric and x-ray photoelectron

spectroscopy.”’ 3"37

In these studies it is found that as the polymer film thickness
decreases the glass transition temperature becomes less than that of the bulk polymer. In
recent work, Torkelson et al. used intrinsic fluorescence to study the Tg of thin polymer
films in their equilibrated state and showed that nano-confinement effects cause Tg of the
polymer to decrease from the bulk Tg in both equilibrium and non-equilibrium states.” In
this work low temperature experiments in the range 25°C - 75°C will be discussed to
reveal a softening effect that occurs within macromolecules and nanoparticles well below
the bulk Tg Of the polymer.

High temperature experiments (~200°C) were also conducted. In the high
temperature experiments it was observed that a single macromolecule exhibits a two
phase behavior between the temperatures of 75°C and 200°C. At 200°C a single
macromolecule can be thermally stimulated to explore the half-space above the substrate
and is experimentally shown to be in an extended conformational state. The
macromolecule can fluctuate between this extended state at 200°C and collapsed state at
75°C and so it may be possible to develop a data storage device by thermally stimulating
single macromolecules between compact and extended states.

Property measurement Of single macromolecules at the nanometer scale has only
become possible in recent years.6' 33 These macromolecules exhibit two conformational

states: a collapsed state at low temperatures and an expanded state at high temperatures.

In this work we use a combination of experiment and Monte Carlo molecular simulation

23

to examine and explain this unusual and never before demonstrated behavior. The height
was independent of the molecular weight (size) of the nanoparticles. Similar behavior is
expected for single linear (uncross-linked) chains on the surface. Flory-type calculations
and scaling theory predict that single chains adsorbed onto subtrates from solution will
have essentially two-dimensional conformations whose height is also found to be

independent of molecular weight. 27

24

3.2 Background

In a previous study12 it was discovered that nanoparticles with a low degree of
cross—linking spin coated onto a substrate will adopt a pancake shape due to van der
Waals force exerted by the surface (see Chapter 2). Single polystyrene macromolecules
will also adopt this adsorbed shape when spin coated in pure benzene. To explain this
adsorbed height, we considered Rubinstein and co-workers’ work on the study of single
chain adsorption to understand why a polystyrene nanoparticle or macromolecule height
becomes SO small (~ 3nm) and essentially independent of molecular weightzmg’29 The
adsorption theory can be used to Show that the chain will exchange part of its
conformational entropy for the energy it gains by adsorbing to the surface thereby
collapsing onto the substrate. The conformation of larger organic particles on solid
substrates has been studied before by determining how they change size upon heating.13
These larger core-shell nanoparticles reversibly change shape upon thermal treatment
from a disk to a sphere conformation. At temperatures below the melting temperature of
the crystalline core the particles adopt a disk-like structure and above this melting
temperature the particles adopt a spherical structure.

Richardson used scanning electron microscopy (SEM) to determine the molecular
weight of single macromolecules.l4 Polystyrene solution was sprayed onto a carbon-
backed mica substrate and shadowed with deposits of gold and palladium. The shadow
length for each macromolecule could then be determined through SEM and this length
was then assumed to be related to the height of the macromolecule on the substrate. To
calculate the molecular weight Of the individual macromolecules, it was also assumed

that the particles were spherical and that the diameter of the particles was equal to their

25

height. Due to limitations, there are two major difficulties in using this technique:
defining the shadow limit cast by the spheres and the assumption of a spherical
conformation on the substrate.

In order to prevent the macromolecules from collapsing on the substrate,
Richardson used a precipitant/solvent solution. Adsorption of the uncross-linked
macromolecules was minimized here using Richardson’s method14 by spraying a solution
consisting of a solvent and non-solvent (precipitant) where the non-solvent has a higher
surface tension than the substrate as well as a low vapor pressure. Using this technique
the macromolecules still collapse substantially on the substrate, but they do not collapse
to their adsorbed height when spin coating from a good solvent alone. This method
increases the macromolecule height by ~2 nm and we use it in the present study to
understand how a macromolecule, which is not strongly adsorbed onto a substrate,

changes conformation.

26

3.3 Experimental

3.3.1 A Softening Effect Exhibited in Single Polystyrene Nanoparticles and
Macromolecules

The polystyrene macromolecules were standards purchased from Scientific
Polymer Products, Inc. The weight average molecular weights used in the experiments
were 44 kD, 393 kD and 1.5 MD and the polydispersity indices (PDIs) of the polystyrene
standards are 1.07, 1.16, and 1.03, respectively. We prepared ultra-dilute solutions of the
polystyrene (0.001ug/mL-20ug/mL) in pure benzene or a mixture of benzene and 2-
methoxyethylacetate (2-MEA).

Polystyrene nanoparticles were synthesized according to the procedure of Harth,
et al.15 and kindly provided by Prof. Craig Hawker. A detailed description of the
nanoparticles characteristics, including molecular weights and polydispersity index can
be found in previous work12 (see Chapter 2) and are given in Table 3.1.

Table 3.1. Sample, Molecular Mass, Polydispersity Index, Degree of Cross-linking,

Nature (i.e. cross-linked or linear), and Diameter if the Macromolecules or Nanoparticles
Collapse to the Bulk Density of Polystyrene (Eq.3.2) for the Samples.

 

 

mol %

Mn cross-linking diameter
sample (kDa) PDI agent nature (nm)
P841 k 41.0 1.07 0 linear 5.0
P8339k 339.0 1.16 0 linear 10.1

PS1530k 1530.0 1.03 0 linear 16.7

T-PS78k 78.0 1 .14 20 cross-linked 6.2
T-P821 1k 21 1 .0 1.32 20 cross-linked 8.6

E-PS33k 33.0 1 .91 60 cross-linked 4.7

Surface profile measurements were performed with a Pacific Nanotechnology
Nano-R atomic force microscope in close contact (oscillating) mode to generate height
images that were not altered other than a simple leveling procedure. Silicon tips with a

spring constant of 25-75 N/m, tip curvature Of <10 nm, and a resonance frequency of

27

200-400 kHz were used for all experiments. The nanoparticle heights were determined by
taking the average of 50 nanoparticle heights; however, the lateral Size could not be
found due to convolution effects created by the AFM tip.'8'19 A hot stage was designed
for the Nano-R AFM so that individual macromolecules and nanoparticles could be
observed at elevated temperatures. The hot stage consisted of a solid metal cylindrical
sample puck with an alumina insulator, heater element, and thermocouple connected to a
PD) controller to monitor the temperature (see Appendix A, Figure A.4). AFM scans at
temperatures above room temperature require that the hot stage be turned off during
scanning due to electrical interference effects; therefore, the high temperature
experiments have a temperature range of approximately 1- 5°C as the heater cools slightly
during AFM scans. In all elevated temperature experiments the sample and AFM
components were kept at the desired temperature for ~3 hours to ensure that all
components, specifically tip and sample, had reached the desired temperature, which
aided in the prevention of thermal drift.

The polystyrene macromolecules were prepared in a solution of 30% 2-MEA and
70% benzene and sprayed onto a silanized substrate using a Meinhard nebulizer (model
TR-30-A1). In this case the benzene is expected to quickly evaporate leaving a Single
polystyrene chain in a drop of 2-MEA which is a non—solvent for polystyrene. The 2-
MEA drop will deposit on the substrate and slowly evaporate, leaving a macromolecule
that does not completely collapse.

The silanized substrate (Sigmacotem) was chosen for these experiments because
it has a lower surface energy (:29 mN/m) than the surface tension energy of 2-MEA

(234mN/m) and therefore the 2-MEA will not wet the substrate. If the 2-MEA had wetted

28

the substrate it would have caused the macromolecule to collapse to an adsorbed state of
3nm as we confirmed by other non-solvents such as n-butanol, which has a surface
tension energy of ~25 mN/m.

SigmacoteTM was used to (silanize the silicon wafers by spin coating the
SigmacoteTM solution onto a silicon wafer at 5000 rpm for 40 seconds. SigmacoteTM
(Sigma-Aldrich) is a solution consisting of 2.5% chlorosiloxane ((SiC12C4H9)20) and
97.5% heptane that functionalizes the surface with short alkane chains. The wafer was
then rinsed with Millipore water to eliminate the excess. The silanized wafer was checked
by AFM to ensure the coating provided a smooth surface with an RMS roughness below
0.5 nm.

Solutions of polystyrene nanoparticles were prepared in benzene at concentrations
of 1 —— 10 ug/ml and either deposited by allowing a drop of solution to evaporate onto a
silanized substrate or spin coated onto a freshly cleaved mica substrate. Mica is a high
surface energy substrate (3 300 mN/m). The nanoparticles were not sprayed onto the
substrates because a better dispersion of particles can be obtained with spin-coating and
drop deposition. Also, since Richardson’s method does not affect the nanoparticles
Spraying them onto the substrate was not necessary.

3.3.2 Thermal Conformation Changes of Polystyrene Macromolecules

The polystyrene (PS) macromolecules were standards purchased from Scientific
Polymer Products, Inc. The weight average molecular weight used in the experiments
was 393kD with a polydispersity index (PDI) of 1.16. We prepared ultra-dilute solutions

of the polystyrene (0.00 Lug/mL - 20pg/mL) in pure benzene or a mixture of benzene and

29

2-methoxyethylacetate (2-MEA). All solutions were filtered with a 0.2pm Teflon filter to
reduce the amount of atmospheric contaminants prior to deposition.
Surface profile measurements were performed with a Pacific Nanotechnology

NanO-R atomic force microscope in close contact (see section 3.3.1 for details).

30

3.4 Results and Discussion
3.4.1 A Softening Effect Exhibited in Single Polystyrene Nanoparticles and
Macromolecules

In this section we will look at two distinct phenomena that occur for the
polystyrene macromolecules and nanoparticles. Both systems appear to become less rigid
at higher temperatures by exhibiting a softening effect at ~ 40°C. These systems also
undergo a partial adsorption to the substrate with each heat cycle.

Low Temperature Experiments for Polystyrene Macromolecules. The heights
of individual polystyrene macromolecules were measured at temperatures ranging from
25°C to 75°C. Scans were repeated for three heating cycles. In Figure 3.1a the measured
heights Of 393kD PS macromolecules at room temperature and at higher temperatures by
increments of 10°C up to 75°C is shown. After being heated to 75°C, the sample was
slowly cooled down on the AFM hot stage to room temperature and then re-measured to
determine if they recover to their initial deposited height.

A height decrease and a transitional temperature are Obvious in all the samples.
After heating to 75°C and measuring the height at room temperature it can be seen that
they do not completely recover their original height. This same phenomenon was
observed in the 44kD and 1.5MD polystyrene macromolecules (see Appendix B, Figures
BI and B2). Note that the large error bars in Figure 3.1 can be attributed to the
polydispersity of the macromolecules to yield a height standard deviation (8h) of :l: lnm.

For a polydisperse system the molecular weight standard deviation (SM) is given by

an =M, x,/(PDI—1) (3.1)

31

Height (nm)

which can be used to calculate the error in the diameter of an assumed spherical
morphology (8D); OD/D = 8M/3M, through propagation of error. This effect comes from
measuring different sets of particles on each repeat scan since the AFM probe can not be
precisely positioned at the same place for each temperature. The per-particle error is
limited by the AFM instrument and is about 0.1nm as we confirmed by analyzing a single
macromolecule repeatedly. The results in Figure 3.1b show the distribution of the data
where the mean (Figure 3.1a) and the highest and lowest values all suggest a height

decrease with increasing temperature.

 

 

 

 

1

 

 

 

 

 

 

.............. I

3 f 3 '
I 393kD PS (cycle 1
O 393kD PS (cycle 2)

 

 

 

 

Height (nm)
.h

 

 

 

 

# 01

+4

01
00.0.0“. 0.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 A - 393kD PS (after cycle 2) j 2 'L 393kD PS.(Cyc'e. 2)”
30 40 50 60 70 30 40 50 60 70
Temperature (9C) Temperature (QC)
(a) (b)

Figure 3.1. (a) A repeat scan of height versus temperature for the 393kD polystyrene
macromolecules on Sigmacotem. Cycle 1 is the initial heights Obtained at 25°C and 75°C
and cycle 2 is the second observation of height versus temperature for the 393kD PS. In
cycle 2 the macromolecules were heated from room temperature to 75°C in increments of
10°C. The AFM height of the macromolecules was measured at each temperature. A
decrease in height is Observed with a transition at approximately 40°C. Note that the large
error bars can be attributed to polydispersity and AFM error of 0.1 nm. (b) Distribution of
height values from 25°C to 75°C for the cycle 2 values graphed in (a).

32

The unusual phenomenon exhibited in Figure 3.1a was carefully investigated.
Initial concerns were centered on the possibility of residual solvent (esp. 2-MEA)
remaining in the sample and influencing the height behavior. Fourier transform infrared
spectroscopy (FT IR) and X-ray photoelectron spectroscopy (XPS) were performed to
prove that 2-MEA was not present in the sample after initial preparation (see Appendix
B, Figures B7 and 8.8). The possible role of surface energy dependence on temperature
was also investigated. Advancing and receding contact angle measurements were
performed on a Si gmacoteTM surface at 25°C and 200°C using benzyl ether (boiling point
~298°C) and it was found that the contact angle remained unchanged. The surface tension
of benzyl ether39 from 25°C to 200°C changes from 36 mN/m — 23 mN/m and since the
surface tension of polystyrene at these temperatures (40mN/m - 27mN/m) is larger than
the surface tension of the benzyl ether the polystyrene will not wet the substrate (see Fig.
3.2). So, the height decrease appears to be the result of a transition within the

macromolecule itself.

33

 

r

 

30 Tl I 60kDa Polystyrene (Sauer)

I O 8.5kDa Polystyrene (Sauer)

E - - - Benzyl Ether (y at 2009C)
. = a z

E,

c 5 . i A

.9 3 a. I

'- E -

cu _ ...-..._.

‘t 24 .

3 2 é : i

(D 0 0 I .. ....... '0 ...... 0 000000 d: 0 O 0

 

 

 

Temperature (QC)

Figure 3.2. Surface tension vs temperature for polystyrene.25 The surface tension of
benzyl ether at 200°C is also shown and is much less than the surface tension of
polystyrene in this temperature range (180°C — 240°C) to suggest that the polystyrene
does not wet the substrate (Sigmacotem) at 200°C since the contact angle of benzyl ether
on SigrnacoteTM from 25°C to 200°C remains unchanged.

Low Temperature Experiments for Polystyrene Nanoparticles. The
nanoparticles that were investigated for these experiments included the 78kD T-PS,
211kD T-PS, and 33kD E-PS nanoparticles (Table 3.1). The L-PS series was not studied
because this system collapses on both the high and low energy substrates regardless of
the solvent treatment during spin coating. The solvent/non-solvent technique
(Richardson’s method) was performed on the cross-linked systems, specifically the L-PS
series, and it was discovered that the height of the particles changed minimally from the
heights discussed below. The nanoparticle systems contain loops of monomer units that

are connected by cross-links. These loops are thought to be more easily adsorbed to the

surface than the freely-jointed chain in a macromolecule. This suggests that Richardson’s

34

technique only affects the uncross—linked polystyrene macromolecules described above
and not the cross-linked systems.

In Fig. 3.3a a graph of the thermal behavior for the 33kD E-PS system is
displayed for both the high surface energy (mica) and low surface energy substrates
(SigmacoteTM). See Appendix B, Figures B3 and B4 for temperature graphs of the 78kD
T-PS NP and 211kD T-PS NP systems. The 33kD E-PS system was studied in more
detail due to its high degree of cross-linking, which prevents the nanoparticle from
collapsing on both low and high free energy substrates when deposited from solution. In
both cases there is a softening temperature at approximately 40°C proving that surface
energy of the substrate does not play a role in the height decrease, which suggests that
this decrease in height as temperature increases is solely due to a temperature effect
within the polystyrene system. Note that in thin film studies it was found that the
interfacial energy between the polymer and substrate altered the glass transition
temperature of the polymer thin film.37 Contrary to our results, they found polymer films
on lower surface free energies exhibit a Tg below the bulk Tg, whereas films on higher
surface free energies exhibit a Tg above the bulk Tg. Note that after being heated to 75°C
and cooling down to room temperature the nanoparticle does not completely regain its
original shape. The other nanoparticles in the T-PS series also exhibit a similiar behavior;
therefore, all systems studied for both nanoparticles and macromolecules have a softening
temperature around 40°C regardless of molecular weight or degree of cross-linking.

One element to take into consideration despite the results of our previous XPS
and FTIR studies is the possibility that residual benzene is present within the

macromolecules and evaporates as temperature is increased, thus causing the nanoparticle

35

to shrink. In order to study this behavior a second cycle was performed on the
nanoparticles to ensure that the nanoparticle will undergo the same behavior and with the
assumption that any residual benzene is no longer present and the system is in
equilibrium. In cycle 2 the height of the nanoparticle decreases as temperature increases
and again has a thermal transition at approximately 40°C. This behavior is graphed in
Figure 3.3b. After the second cycle the nanoparticle iS slowly cooled down to room
temperature and re-imaged. The height appears to be only slightly less than the initial
height of the nanoparticle at room temperature for the second cycle. An additional
experiment was performed to determine if the nanoparticle will continuously decrease
from its initial height. The sample was heated for five cycles to 75°C for one hour and
then measured via AFM once the sample had completely cooled down. The sample
gradually decreased in height with each cycle suggesting that the nanoparticle is slowly
adsorbing on the substrate and gaining molecular contact despite its high degree of cross-

linking.

36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7 M 4: 7 ’[SigmacFe]
6 _._. 6 _ ......... . .. .
E, 4 E I T T . ._'
E r 4 I I I
g, 3 “Q ‘- ‘5’ 3 T -
. . ._ a)
I . . . I '
2 o E-NP (heated) mica 2 .. ..
E E-NP (after heat) mica - -
1 - 0 E-NP (heated) sigmacote 1 _ ......... : 33:3 :2: 3:: Egg: {3
; I 5N? Wer “930 Sigmacme‘ - A 33kD E-PS NPs (after cycle 2)
0 7 l' ' "" Predicted Height 0 .. ......... . - Predicted Height
30 40 50 60 70 80 30 4o 50 60 70
Temperature (9C) Temperature (9C)

(a) (b)

Figure 3.3. (a) Height vs temperature graph for 33kD E—PS system on mica and
SigmacoteTM. A softening temperature of approximately 40°C is observed for the
nanoparticles on both substrates, suggesting that surface energy does not play a role in
this transitional temperature. (b) Height vs temperature graph for 33kD E-PS system on
SigmacoteTM for 2 cycles. For both (a) and (b) the macromolecules were heated from
room temperature to 75°C in increments of 10°C. The height of the macromolecules was
measured at each temperature. A decrease in height is observed with a transition at
approximately 40°C.

Polystyrene nanoparticles and macromolecules exhibit the same softening effect
at temperatures slightly above 40°C. The graphs in Figures 3.1 and 3.3 reveal that
molecular weight, cross-linking degree, and substrate surface energy do not affect the
softening effect at ~40°C. Since these factors can be eliminated as a cause for this we are
lead to further examine the effect of the surface tension of polystyrene.

As temperature is increased the surface tension of the polystyrene is decreased,
which will cause the polystyrene to maximize its contact with the silanized wafer with a

concomitant decrease in height. In Dee and Sauer’s work25 they showed that the surface

tension of polystyrene decreases as temperature increases, but the surface tension of

37

polystyrene does not fall below the surface energy of the silanized wafer until
approximately 200°C; thus, suggesting that below 200°C the polystyrene will not wet the
substrate. Yet, we can consider the small decrease in the surface tension of polystyrene
between 25°C and 75°C and see how it can cause a height decrease using the JKR model.
The JKR theory22 can be utilized to understand how surface tension affect the
nanoparticles and macromolecules adsorption. The JKR theory is a model that describes
the adhesion between two bodies under a compressive force and accounts for the short-
ranged surface forces acting inside the contact area. The theory is based on the Hertzian
theory yet takes into consideration the adhesion in the contact zone. We can examine this
softening effect by observing the modulus change of the adsorbate through the use of the
JKR theory and the assumption that the adsorbate geometry adopts a spherical cap shape
(see Fig. 3.4). Note that for the JKR model the contact geometry is allowed to deform,
which accounts for the nanoparticle and macromolecule collapse or deformation as

temperature increases.

 

Figure 3.4. Should the nanoparticles adopt a spherical conformation on the substrate,
then their height (H) should equal their diameter (D) which is 4.7 nm for the E—PS33kD
system. The nanoparticles may adopt a deformed shape on the substrate of equal volume
to the sphere with contact diameter d as described by the J KR theory.

38

To ascertain the nanoparticle conformation on the substrate shown in Figure 3.4,
we determine the diameter (D) assuming the macromolecule collapses to a sphere with
the bulk density of polystyrene (p ~ 1.04 g/cm3) through

M V.
=[6 "]
nNAp

 

(3.2)
where N A is Avogadro’s number.
We use the JKR theory for a sphere in contact with a flat surface, experiencing

zero load, where the contact diameter ((1) becomes

d3 =97r(1—-v"')%D2

(3.3)
with D being the predicted nanoparticle and macromolecule diameter (eq 3.2), E is

Young’s modulus, v is Poisson’s ratio, and WA is the work of adhesion given by

WA=7A+75'7%=2(7Adl’sd)}/2 (34)

with 7 being the surface energy of the adsorbate-vapor (A), substrate-vapor (S), and
adsorbate-substrate (A/S) interfaces. The approximation is due to Fowke’s relation24 that
is applicable to lower energy substrates that interact through dispersive forces (the

superscript d represents the dispersive part of the surface energy or tension). In our case,

d —
polystyrene25 interacts via dispersive forces and has 74 _ 40 mN/m while the silanized

d ~
substrate has 78 "' 28-61'0-6

mN/m at room temperature. As mentioned earlier the
surface energy of the silanized substrate was found to be constant for temperatures up to

200°C.

39

For higher temperatures the surface tension of polystyrene can be determined

from the equation‘25

yPS = 41.5 — 0.068T (3.5)

where surface tension (7) is measured in mN/m and T is measured in °C.
As stated above and in previous work”, the adsorbate geometry is assumed to be
a spherical cap with equal (constant) density between the Sphere (see eq. 3.2) and the

adsorbed shape (see Fi g. 3.1). The volume of the spherical cap geometry is expressed as

77 drop 2 2
Vmp =-6—H 3 2 +H

Analyzing the spherical cap geometry with equal (constant) density between the sphere

 

(3.6)

(see eq. 2) and the adsorbed shape (see Figure 3.1) we find

 

3
d,_ 8 (03—113)?
373 3
H’ (3.7)

where H is the adsorbate’s height on the substrate. Combining the JKR theory and the
Spherical cap geometry (eq. 3.3 and 3.7), we can determine the interrelation between the

nanoparticle height and surface energy

  

3
H3 +: 27:J§(1—v2)v—:?i02:| H —D3 =0

 

(3.8)
Taking the surface tension change of the polystyrene with temperature (eq. 3.5) we
examine how modulus of the macromolecules and nanoparticles changes with
temperature Shown in Fig. 3.5. In Fig. 3.5 we see a decrease in modulus as temperature

increases suggesting a softening effect is occurring in both macromolecules and

40

nanoparticles. The modulus values shown in Fig. 3.5 were measured for the second cycle
values of each system (Fig. 3.1a and 3.3b) to ensure the systems had reached equilibrium.
Note the modulus of the macromolecules is less than that of the nanoparticles. This
difference in modulus is due to the rigidity of the nanoparticles, which is caused by the
large degree of cross-linking (60%). 12’ 21 The modulus of the E-PS NP system for cycle 2
on SigmacoteTM at 25°C is z 1.5 GPa, which is approximately half the modulus of bulk
polystyrene (E z 3GPa). This is due to partial adsorption to the substrate after the initial

cycle.

 

K «a» 393kD PS
\\ -o— 33kD E-PS NP :

\r-x

 

 

 

 

1 000
8
6

h
firrrrrr
Lljll

 

 

 

 

 

 

 

 

TI?
LIJ 100 . .
8: I
6' :
4L -
"m ”no...“m '3
. ..-..-.,........,, m. J
2 I M... “j

 

30 40 50 60 70
Temperature (9C)

Figure 3.5. Modulus vs. temperature plot of 393kD polystyrene macromolecules on the
silanized substrate and 33kD E-PS nanoparticles plotted from Eq. 3.8. The modulus for
both systems was found for the second cycle values found on the silanized substrate.

41

 

Figure 3.6. For small values of a, the volume of the outer shell of monomers becomes a
large fraction of a spherical particle’s total volume fraction.

To understand in more detail why the macromolecules and nanoparticles soften
and collapse from their kinetically trapped metastable conformations below the glass
transition, T = 40°C << Tg, we need to examine the scale of the problem. Consider a
simple model where the surface elements of a single polymer chain collapsed into a
spherical globule (see Fig. 3.6). We simplify this further by ignoring the chain bonding
elements and examine only the van der Waals forces between Kuhn monomer units. The

volume of a layer of monomer units on the surface is given by

47(613 0' 3
SS— 3 1—[I_Z) (3.9)

 

where o is the Kuhn monomer diameter and a is the radius of the globule. Dividing this
by the volume of the sphere (41ta3/3) we find the fraction of monomers on the surface.

For a nanoparticle of radius a = Snm and thermal blob size s = 1 nm we have,

42

§i=1—os3z05
V

S

(3.10)

or half of the monomer units are on the surface. These surface monomer units experience
less van der Waals force than the ones in the center, because they have less surrounding
neighbors than internal monomers, and so one may expect the energy required to remove
one of these external monomer units is quite a bit less, say one-half the energy of one
from the bulk of the globule. We hypothesize this has the effect of reducing the

temperature at which the particle can be deformed plastically by an external force below

its bulk Tg.

 

 

o t 33kD alps NPs (Heightj ' ' 20
-I- 33kD E-PS NPs (Contact Area) i

 

 

 

' r C I
g s a
I § E i
1 i i
"' 5 3

. .
. = 2

 

 

, i
4 0 H...i-.....--......a ..........-..,..
. I
i
1 I.

Height (nm)
‘
(mu) [2er toetuoo

   

A .
. v
. .
. . .
, Q -
D , u
2 u
, '
I I O I
. u
- r .
u -

 

 

 

Cycle #

Figure 3.7. AFM heights of 33kD E-PS NPS measured at room temperature after 5 heat
cycles to 75°C. Cycle 0 represents the initial height of the nanoparticles prior to heat.
The nanoparticle height continuously decreases in height as the cycle number increases
suggesting an increase in monomer adsorption to the substrate with each cycle.

43

We also need to consider that the macromolecules and nanoparticles gradually
adsorb to the substrate after being heated to 75°C. As mentioned earlier, the E-PS NPS
continuously collapse after 5 cycles to a height of approximately 3.5 nm, suggesting that
the nanoparticles will eventually collapse to their adsorbed height of 3 nm.27 Fig. 3.7
contains a graph of the AFM height values of the 33kD E—PS NPS measured after being
heated to 75°C for 5 cycles. The contact area after each temperature cycle is also graphed
in Fi g. 3.7 and is determined by assuming constant density and a spherical cap geometry
(eq. 3.7). After each temperature cycle more monomer units are in contact with the
substrate which prevents the nanoparticle from returning to its initial conformation as the
chain cools down to room temperature.

To understand this collapse transition, we study the equilibrium state of the
nanoparticles in solvent and in air. The nanoparticles are initially deposited onto the
substrate by weak adsorption from solvent. While in the solvent, they adopt ideal
configurations, from which we may estimate their height off the substrate with the help of
a scaling theory.27 When the solvent evaporates, the particles become trapped in non-
equilibrium conformation by their glassy character and connectivity constraints. This
results in chain configurations that extend off the substrate higher than their equilibrium
adsorption blob height in air. To collapse completely requires the particle to undergo
massive changes in their conformations, which occurs during repeat heat cycles.

We begin by calculating the height for an adsorbed nanoparticle in air using the
de Gennes method.” 40 In this case, the nanoparticle should be completely collapsed,
since air is a non-solvent for PS. The average volume fraction of adsorption blobs with

gads monomers is given by

b3
¢z_&£.~.1 (3.11)

where ems is the adsorption blob size, and b is the Kuhn monomer length. In the collapsed
case all the monomers in the adsorption blob are in the attractive well of the surface so
gads =( sad, lb)3. The number of Kuhn monomers in contact with the surface can be

determined from

¢ 2 82m
The macromolecule exchanges conformational entropy for adsorption energy by being in

contact with the surface. The energy gain is written as

82
flcBT-f’zi 2: kBT, (3.13)

where 8 is the adsorption strength (for weak adsorption 5 < 1). Solving for 5 we find
b2
5 z 22;. (3.14)
From our previous work12 the adsorption blob size of a polystyrene macromolecule or
nanoparticle in air is, Eads ~ 3nm. Using b ~ l.8nm as the Kuhn length for polystyrene, we
find §~ 0.36. Using 5 to compute the height, 8, of the nanoparticles in solvent for ideal

and real chain statistics,27 we find

8 z z Snm (ideal) and 6‘ = — z 8nm (real), (3.15)

b
6
which can be used for lower and upper estimates of height immediately after solvent

evaporation. This is in very good agreement with the AFM height on the substrate

measured directly after deposition of the 33kD E-PS nanoparticles onto the substrate.

45

After each heating cycle, the nanoparticles gradually adsorb until they reach their
equilibrium state in air. This is graphed in Figure 3.7, where we see that after each heat

cycle the nanoparticles are gradually decreasing in height or adsorbing to the substrate.

3.4.1 Thermal Conformation Changes of Polystyrene Macromolecules

A two phase behavior in single polystyrene macromolecules examined on a solid
substrate at elevated temperatures was also studied. A fluctuation between an extended
state at 200°C and a collapsed state at 75°C is observed using atomic force microscopy
measurements and Monte Carlo simulations.

In our first experiment, the substrates were heated in an oven for 30 minutes to six
different temperatures ranging from 100°C to 200°C and then cooled for measurement at
room temperature. As the macromolecules cool they collapse partially, but a signature of
their high temperature extended state remains. This phenomenon was further investigated
to determine if the macromolecules could be thermally annealed into completely
adsorbed states. The macromolecules’ heights were measured after repeated heating
cycles between 75°C and 200°C.

The results from these experiments are shown in Fig. 3.83 and b. Each point on
these graphs is an average of 50 measured particles. The error bars were computed using
the standard deviation of the mean since this accounts for the polydispersity in our
sample. In the first experiment (Fig. 3.8a), the macromolecules seem to have the same
height upon cooling until the heating temperature reaches 180°C, at which point, they
freeze into a taller configuration. We believe this extended height is an indication the

particles are undergoing a molecular reordering.

46

The heights cycled between the two temperatures are shown in Fig. 3.8b, where it
is seen that in the initial cycles the fluctuations between the two states are large, until
cycle 4, after which they appear to stabalize into a two state system. Subsequent
measurements beyond cycle 7 showed similar oscillations (the macromolecules were
taken through 30 cycles total- see Appendix B, Fig. B.5).

The observed height change in these single macromolecules is remarkable. In
order to explain our findings it is important to illuminate the differences between our
experiments and the null experiment (macromolecules deposited with pure solvent). The
main difference is the starting state, which, in the case of our experiments is non-
equilibrium. When the chains are deposited, they are in a partially pinned state where

41

trains of monomers are strongly adsorbed to the surface. This combined with the chain

constraints results in a high energetic barrier to relaxation and subsequent adsorption.

47

Height (nm)

 

 

 

 

 

 

 

 

 

7'0 e 393kD PS Molecules] "‘ ' ‘a‘i‘ - 393kDa PS
i , . . ‘1' , f .
i i E
. ; 61 9 é
4.5 -?~l »%-l , -l l l l~ 3~ —l~~-i—
on") °c'> °c'> % % % %
N e s: s e 92 8
Temperature (QC) Temperature (QC)

Figure 3.8 (a) Height of 393kD polystyrene macromolecules at room temperature after
heating to the corresponding temperatures found on the graph. At 180°C the
macromolecule’s height increases significantly. (b) Height of 393kD PS macromolecules
at room temperature after being heated to the corresponding temperatures on the graph.
The height of the macromolecules fluctuates as the temperature is cycled between 75°C
and 200°C. The raw data is shown as points to indicate the measurement errors caused by
polydispersity. All heights were obtained by AFM.

To gain some insight into the possible chain conformations on the substrate, and
into the cause of the “pop up” observed in the experiments, we performed Monte Carlo

(MC) simulations‘n“14

of an on-lattice, interacting, self-avoiding walk (ISAW)45 near an
attractive substrate“ 47. Previous simulation studies generated possible phase diagrams

for single chains near a surface48' 49, but did not examine the heights of the

macromolecules.l

 

i Dr. Erin McGarrity performed the Monte Carlo simulations described here.

48

Similar simulations have been done for isolated single chains in bulk to study the
globule to coil transition, which occurs as the temperature is raised from below to above
the 6 temperature, T347. The phase behavior of single chains becomes more complicated
in the presence of a surface since the chains will adsorb on the surface at temperatures
lower than the adsorption transition temperature, Ta. The phase behavior is fairly well
understood for T2, > T9, but not for T2, < T9 since polymer solutions phase separate at low
temperature”. However, this is not an issue here since we have single chains isolated on
the substrate in air. We expect that since air is a poor solvent for PS, the chains will
undergo their globule to coil transition at lower temperatures than those at which the
chains would desorb from the substrate, i.e. we expect that Ta > T9 in our case.

We chose to use a 12-connected lattice for our simulation because of its
algorithmic simplicity and speed and because the packing on such a lattice resembles a
hard sphere solid at low temperatures. The ISAWs were sampled using a variant of the
pruned- enriched Rosenbluth method (PERM)5 1’ 52 and ensemble averaged over a wide
range of normalized temperatures (kBT/e = 0.25 — 20). The energy of the chains is given
by

l N

N l
Urhain : _§Zk=l (ECk + 83 Sk ) + EZk=2 £8 gk (3-16)

where e is the monomer-monomer attractive energy and as is the monomer-surface
interaction energy. N is the chain length and ck, sk are the counts of contacts of the
monomer at site k with the other monomers and the surface, respectively. Finally, 8g is a

bending energy term and gk is a Boolean variable which takes on the value 0 for trans-

and l for gauche- angles between bonds (k + 1, k) and (k, k-l). This model is a hybrid of

l.47 1.“.

those used by Doye et a and Bachmann et a

49

The model parameters were chosen so that the ratio sales was similar to the
experimental system. The surface tension of PS is z 40 mN/m, while the surface energy
of the silanized substrate is :2 29 mN/m. We set e/kT = l and thus took es/kT = 0.25. PS is
fairly flexible so we modeled it as a floppy chain with $1ng = 0.05.

Once the chains were grown, statistics were computed at each temperature using

the usual procedure53'55,

(X) = :=,Wle (3.17)

where z is the number of chains including its Boltzmann factor and degeneracy, and Xk is
the variable of interest for chain i. It is important to note that only chains which touched
the substrate with at least one monomer unit were included in these averages. We thus did
not attempt to find the desorption transition for our model. The averages were calculated
using 2 ~ 108 samples at each temperature below 10 and 2 ~ 107 for the rest (kBT/e >10).
The additional sampling at low temperature is to account for the poor efficiency of the

chain growth algorithm in the dense regime.

50

 

 

 

   

 

 

 

 

 

 

 

 

so 4 12 4
70» 10
so a
of 50 A
e .c a
a: 40 v
V To=5.15
4
30* 2.5 I
20» 2 2
5 5.5 e
10 . . . . o . i . l . . . . .
02468101214161820 02468101214161820
kale kale

Figure 3.9. (a) Average Rg2 versus temperature for chains of length N = 124. (b) Average
macromolecule height from the surface as a function of temperature for N = 124. From
the inset, one can see the height begins to increase rapidly near T9. The scaled

experimental data is shown with error bars for comparison.

Our simulation shows a range of behaviors, as can be seen from the average
radius of gyration (Rg) and the average molecular height off the substrate (h) of the
macromolecules as functions of temperature, shown in Fig. 3.9 a and b. At low
temperatures the chains are in rod-like configurations (Rg2 ~ N2) adsorbed to the surface.
As the temperature is increased, the macromolecules collapse into 2D adsorbed globules,
with Rg2 ~ N2]3 and small values of h. Further increases in temperature cause the chain to
spread out on the surface, with Rg2 increasing but 11 staying nearly constant at the
adsorption blob size. Near a temperature of T z 5, the chains undergo their globule to coil
transition and adopt ideal random walk configurations, Rg2 ~ N, while still adsorbed on
the surface. Finally, they begin to desorb from the surface for T 2 6 and adopt ideal (Rg2

~ N) and swollen configurations (Rg2 ~ N65), with R,g and h increasing with both T and

chain length.

51

The globule to coil transition is found by following the technique outlined by

1.,56

Binder et a in which the 0 temperature is estimated from the crossover scaling of

chains of different lengths. In Fig. 3.10, we plot the normalized squared end-to—end
distance <R02>IN versus temperature. The inset shows that the longest chains cross at T 2

5.15, which we take as T9 for our model.

52

4.5

 

  
    
  

 

N364
4 ~ N=90

'6" N=1 24
l

 

 

 

 

 

 

 

 

 

 

0.5
1.73 -
4.9 4.95 5 505 at 5.15 5.2
O 1 1 l
0 5 10 15 20
ka/e

Figure 3.10. <R02>IN versus temperature for different chain lengths (N: 64, 90, 124).
The crossing point for the length 90 and 124 chains gives a rough estimate of the 0
temperature for the model, T9 2 5.15.

To compare with experiment, we need a relation between the model and
experimental temperatures. We assume that the experimental height increase begins at the
6 point in the model, and that this occurs at about 'I‘ = 413K. We thus obtain 8 = kB'I‘el'I‘m
= 80kg, so the highest experimental temperature of 473K corresponds in the simulation to
Tm = 5.9. Examining the inset of Fig. 3.9b we see that the simulated macromolecules
increase their height by about 0.8 monomer diameters between T9 = 5.15 and T = 5.9.
This agrees remarkably well with the experimental data, as a PS Kuhn segment is z 1 nm

in size.

53

3.5 Conclusion

Solvent and temperature strongly affect the conformations of polystyrene
macromolecules and nanoparticles adsorption on a substrate. Measurements using AFM
were used to study the conformations of polystyrene macromolecules and nanoparticles
on selected substrates. We have observed interplay between the solvents used during
deposition for the macromolecules and an irreversible height transition at low
temperatures for both macromolecules and nanoparticles. We have also presented
experimental evidence for a softening effect occurring near 40°C. As temperature
increases the measured AFM height of the particles decreases. This softening effect can
be attributed to the decrease in surface tension of polystyrene as temperature increases.
This can be explained by looking at the scale of a single macromolecule and nanoparticle
from that of the bulk. Using a simple model we found that ~50% of the monomer units
are on the outer area of the particle. The energy to remove one of these monomer units is
much less than from the bulk of the particle, suggesting that the temperature at which the
particle can deform is less than that for bulk polystyrene. We can also use the J KR theory
and the assumption that the particles adopt a spherical cap conformation to determine
how the change in surface tension of the polystyrene affects the modulus of the particles.
As temperature increases the modulus decreases for both macromolecules and
nanoparticles.

In final experiments, several heat cycles to 75°C were performed on the extremely
cross-linked polystyrene nanoparticles and it was determined that after each cycle the
nanoparticle gradually reduces in height. This gradual decrease can be contributed to the

adsorption of monomers onto the substrate with each heat cycle.

54

We have also presented experimental evidence for an expansion of single PS
chains on a substrate as the temperature is increased. The height increase is due in part to
the dynamics of the deposition process, in that the chains may not be fully equilibrated
when initially deposited on the substrate. However, the height data are very consistent
with an expansion in chain dimensions due to the globule to coil transition taking place in
the adsorbed chain. Monte Carlo simulations were used to generate and analyze the
configurations and heights of single chains on a surface, with parameters based on the
experimental system. From the simulations we find that the change in height due to the

globule to coil transition is of the same order as measured experimentally.

55

Chapter 4: Directional Self Assembly: Nanoparticles Drawn to Carbon Nanotubes
4.1 Introduction

Assembling arrays of nanostructures is an integral component to this possible data
storage technology. Forming arrays of nanoparticles has been approached using
numerous methods including chemical modification of nanostructures,21 self assembly,”'
58 and nanolithography.5%1 In one case electron beam lithography was used to create
patterned hole and trench templates in a silicon substrate. A drop of gold nanoparticle
suspension was then placed onto the patterned substrate and capillary forces served as the
driving force to move the nanoparticles into the trenches.‘51

In this work we look at aligning nanoparticles with the aid of single-wall carbon
nanotubes (SWNTS) which certainly has utility beyond our proposed application. A
posteriori we find the nanotubes pull the nanoparticles to them during deposition and so
by assembling or strategically locating the nanotubes one could then assemble
nanoparticles in a unique and robust manner. Extensive research has already been
performed on the alignment of nanotubes with modifications during SWNT synthesis
used to promote alignment.62454 SWNTS are grown from a silica substrate coated with
catalytic metal nanoparticles with subsequent chemical vapor deposition. Initially they
will be perpendicular to the substrate, but as growth continues they will eventually lay
down on the substrate forming an array of nanotubes that are parallel to the surface.

Currently, researchers can align nanoparticles on both single walled and multi-
walled nanotubes through chemical modification of the nanotubes.65458 In one case CdSe
nanoparticles were attached to acid-chloride modified SWNTS by amide bond

formation.‘56 Researchers have also attached nanoparticles to carbon nanotubes using

56

dielectrophoresis.” In this work SWNTS were linked between a Au/T i electrode bilayer
on a chip. A sinusoidal ac voltage was applied to the system after suspensions of
polystyrene and gold nanoparticles were deposited onto the chips. The dielectrophoresis
force applied to the system caused} the nanoparticles to agglomerate to the nanotubes.
Without the applied voltage the nanoparticles did not appear to attach to the nanotubes.

In this study, we are able to attach a variety of nanoparticles to SWNTS without
the aid of specific chemical modification or dielectrophoresis. The nanoparticles are
attracted to the nanotubes through van der Waals forces enhanced from the geometrical

effects of the cylindrical nanotubes.

57

4.2 Experimental

Functionalized SWNTS were synthesized using the HiPco process.7°' 7’ This
process is a gas phase method that uses an iron catalyst to catalytically disproportionate
high pressure CO gas to generate a mixture of metallic and semi-conductive SWNTS. The
direction the graphene sheet is rolled determines whether the nanotube will be conductive
or semi-conductive.72 The SWNTS used in our experiments was functionalized with a
butyl group via alkylation73 and this functional group produces a poor conducting
nanotube. The butyl group was added so that the SWNTS would be soluble in chloroform.
Even so, in all experiments, the solution was sonicated for 30 minutes prior to sample
preparation to ensure proper dissolution. The nanotube solutions were then spin coated at
5000 rpm for 40 seconds onto the selected substrates which included mica and silicon
wafers.

Several nanoparticles were used for alignment along the SWNTS; polystyrene
(PS), gold (Au), and cadmium selenide (CdSe). A detailed description of the
nanoparticles characteristics can be found in Chapter 2, section 2.2 and in previous
work.”' 74 We also used polystyrene macromolecules which were standards purchased
from Scientific Polymer Products, Inc. and had a weight average molecular weight of 393
kD and a polydispersity index of 1.2.

The Au nanoparticles were purchased from Meliorum Technologies and were
suspended in toluene by having an unknown steric layer that is ~l.5 nm thick. We
determined their average height via AFM to be 7.5 i 1.0 nm and so we believe the gold
core was ~6.5 nm in diameter consistent with transmission electron microscopy (TEM)

measurements. The CdSe nanoparticles75 contain a steric layer of pyridine76 that is

58

 

approximately 0.5 nm in length and are suspended in pyridine. Table 4.1 contains
characteristic details for the nanoparticles used in these experiments. Note that the
diameter listed for the polystyrene nanoparticles and molecules was predicted assuming

the macromolecule collapses to a sphere with the bulk density of polystyrene (p ~ 1.04
g/cm3) through D =[6Mn IJWApF”, where NA is Avogadro’s number and Mn is the

number average molecular weight. The diameter listed for the Au and CdSe nanoparticles
was determined via AFM height.
Table 4.1. Sample Code, Number Average Molecular Weight (kD), Solvent, Diameter if

the Macromolecules or Nanoparticles Collapse to the Bulk Density of Polystyrene and
the AFM Height (Diameter) of Au and CdSe NPS.

 

Sample Code Mn(kD) PDl Solvent D (nm)
T-P821 1 k 21 1.0 1.3 benzene, toluene 8.6
T-PS78k 78.0 1 .1 benzene 6.2
E-PS33k 33.0 1 .9 benzene,chloroform 4.7
P8393k 339.0 1 .2 benzene, toluene 10.1
Au toluene 7.5 :l: 1.0
CdSe --- --- pyridine 4.5 t 0.5

We prepared (0.1 — 10 ug/mL) solutions of all nanoparticles described above in
the solvents given in Table 4.1. All solutions were sonicated for approximately 30
minutes and then deposited onto a substrate that contained SWNTS which were
previously deposited by spin coating. The solutions were then imaged via AFM after the
solvent had evaporated.

Surface profile measurements were performed with a Pacific Nanotechnology
Nano-R atomic force microscope in close contact (oscillating) mode to generate height

images that were not altered other than a simple leveling procedure. Silicon tips with a

59

spring constant of 25-75 N/m, tip curvature <10 nm, and a resonance frequency of 200-
400 kHz were used for all experiments.

Transmission electron microscopy was also used to examine the alignment of
nanoparticles on nanotubes. The samples were prepared by depositing a drop of SWNT
solution onto a forrnvar supported TEM grid. Once the solvent had evaporated a drop of
solution containing the nanoparticles was deposited on the grid. The polystyrene
nanoparticles were then negatively stained with 1% phosphotungstic acid. The other
nanoparticles were not stained since they provide sufficient electron contrast to be seen

via TEM. The samples were then imaged with a JEOL 100CX TEM.

60

4.3 Results and Discussion

AFM and TEM images provide evidence of nanoparticle alignment on single-wall
carbon nanotubes. The nanoparticles appear to agglomerate to the nanotubes on both
freshly cleaved mica (surface energy, 7 z 300 mN/m) and silicon wafers (7 z 70 mN/m),
which indicate that surface energy does not play a role in the alignment for these rather
high energy substrates. AFM and TEM images of the nanoparticles’ alignment on
nanotubes are revealed in Figs 4.1 and 4.2 respectively. Note that a few of the SWNTS
agglomerate into small bundles (~10 nm). The SWNTS were deposited on a silicon wafer
and are shown in Fig. 4.1a to reveal a rather randomly oriented array. Following the
procedure outlined above, the E-33kD PS nanoparticles were deposited from chloroform
on mica, the T-PSlek nanoparticles were deposited from benzene on a silicon wafer,
and the Au nanoparticles were deposited from toluene on mica in Fig. 4.1b, c, and d.
Note that for all substrates (silicon wafer and mica), all solvents (benzene, toluene,
chloroform and pyridine), and all nanoparticle systems (Au NPS, PS NPS, and CdSe NPS)

nanoparticle agglomeration onto the nanotubes was found as seen in the figure.

61

 

0.00pm 0.97pm . 0.00pm 0.97pm

         

c d
33.92am 18.71a-
‘ 0.00m 0.0 1:
0.00pm 1.05pm pm 0.00pm 1. 6231111 3 .ZQnm

Figure 4.1. AFM height images of (a) SWNTS spin coated onto a silicon wafer, (b) E-
PS33kD NPs deposited from chloroform on mica (c) T-PSlekD NPS deposited from
benzene on a silicon wafer (d) Au NPS deposited from toluene on mica. Note that for the
AFM images in b,c, and d the SWNTS were deposited onto the selected substrate prior to
nanoparticle deposition.

TEM images of the nanoparticle alignment are shown in Figure 4.2. The images
in Figure 4.2a and b are of Au NPs and SWNTS. The Au NPs appear to agglomerate to
the nanotubes, but the large agglomerations of NPs do not as readily. As shown below the
vdW interaction between the SWNT and nanoparticles decreases as the size of the

nanoparticle increases. This suggests that the nanoparticles were agglomerated in the

solvent prior to deposition on the surface. Shown in Figure 4.2c is a TEM image of a line

62

- of CdSe nanoparticles on a nanotube and d is an image of the T-211kD PS NPs on the
SWNTS. The polystyrene nanoparticles must be stained in order to be visualized by
TEM, which causes a shadow effect that is not as visually appealing as micrographs for
the other systems; however, we believe there is an agglomerate of the polystyrene

nanoparticles present along the nanotubes.

 

Figure 4.2. (a) and (b) TEM images of Au NPS aligned along SWNTS. (c) CdSe
nanoparticles aligned along SWNTS and (d) T-211kD PS NPS agglomerated along
SWNTS.

To explain this phenomenon we first considered the actions of the nanoparticles
as they settle to the substrate. A droplet containing the nanoparticles in solvent is placed
on the substrate already containing nanotubes. The benzene wets the substrates (both

mica and silicon wafer) and as the particle moves to the substrate we suggest, as

discussed below, they are attracted toward the SWNTS. Since the benzene wets the

63

substrate there is no pinning line to force the nanoparticles to deposit on the outer edge of
the droplet to produce the behavior referred to as the coffee ring effect.“ 78 During the
coffee ring effect a ring of particles is created due to evaporation at the droplet’s edge
which causes a pinned contact line. Solvent from the center of the droplet replaces
solvent lost by evaporation at the edge of the droplet; thus carrying particles to the
droplet edge through capillary flow. In our work the solvents wet the substrates, which
prevent a pinning line from occurring suggesting that the nanoparticles’ alignment on the
nanotubes is not due to capillary flow.

To understand why the nanoparticles are attracted to the carbon nanotubes,
consider the system depicted in Fig. 4.3. The nanoparticle diffuses freely in the solvent
above the substrate. As the solvent begins to evaporate, the particle is forced closer to the
substrate as well as the nanotube, and is drawn to them by van der Waals forces. By
analyzing the relative strengths of these forces one can show that the particle will tend to
be pulled preferentially toward the nanotube as the solvent evaporates and is mediated by

a combination of geometric and dielectric factors.

64

Solvent

  

 

(9’1

(d)
Figure 4.3. (a) Spherical nanoparticle (NP) diffusing in solvent) near a nanotube (NT) on

a substrate. (b-d) Coordinate systems for the geometries used in our systems: (b)
sphere/sphere (c) sphere/plate and (d) sphere/cylinder.

 

The interaction potential and force between a sphere and cylinder and sphere and
flat plane (substrate) can be determined using the Hamaker approach and pairwise
summation of all involved intermolecular forces. The van der Waals interaction energy

between two bodies U 2was first expressed by Hamaker79 as

U12 : _Cp1p2 L dVl L de '1}?

(4.1)
where ,0), ,0), V), and V2 are the atom number densities and the total volumes of the
objects, r is the distance between the volume elements, and C is the substance dependent

London-van der Waals constant.

We are interested in three particular geometric combinations: two spheres, a
sphere and plate and a sphere with a cylinder since these roughly describe the
nanoparticle/nanoparticle, nanoparticle/substrate and nanoparticle/nanotube interactions.

The potential energy between two spheres with radii R1 and R2 is given by79' 8°

65

 

 

A132[ RIRZ + RIRZ
3 cz—(R1+R2)2 cz—(Rl—R2)2
1 C2 -(R1 + R2)2 (4.2)

+—1n
2 cz-(R,-R,)2]

USS(c;R,,R2) =—

 

where c is the distance between their centers (see Fig. 4.3b) and A132 is the Hamaker

constant for substances 1 and 2 in the medium 3. In our calculations, we will use

hzc-Rl-Rz, the distance to contact as our measure. From the sphere/sphere calculation it is

possible to obtain the sphere/plate energy by taking one of the radii as infinity resulting in

A R R h
U h;R =——112- —+—+ln—
S’”( ) 6 [h 2R+h 2R+h] (43)

where R is the sphere radius, h is the distance to contact and A132 is the effective

Hamaker constant between the sphere 1 and plate 2 in the medium 3. The geometry for

this case is shown in Fig. 4.3c.

For the nanoparticle/nanotube potential, we treat the cylinder as infinitely long

since nanotube aspect ratios are typically 21000 and so the relation describing this system

reduces to a single integral which must be evaluated numerically“ 82

2 2 2 3
U SC (0, RS , RC ) = -A132 _R arcco 2 2
c 2rc r — RS

 

where Rs and R C are the sphere and cylinder radii, respectively, c is the distance between

their centers, c = h + R5 + RC , and A132 is the Hamaker constant. The coordinate system

for this system is depicted in Fig. 4.3d.

66

 

17. 80

To progress further the effective Hamaker constants for our system must be

determined. For the polystyrene/nanotube/Si02 system in benzene, we need the

parameter given in Table 4.2. These constants are for two bodies of the same composition
acting across vacuum and to account for the effect of the solvent, we use the following

approximation

A132 3 (J4: " X/A_33)X‘\/X2_2- " ‘\/-A—33-) (4.5)

where Aii is the self Hamaker constant in vacuum.

Table 4.2. Hamaker constants for typical substances used in our experiments.

 

 

 

 

 

Substance Hamaker Constant A (zJ)
Benzene” 5.0
Polystyrenel7 6.6
sroz” 25
N anotube83 50

 

 

 

 

For the polystyrene/nanotube/Si02 system in benzene we calculated the effective

Hamaker constants as well as the potential energies at a distance of h = 0.5131. The results
are tabulated in Table 4.3 along with the pertinent geometries where we assumed the
spheres had a radius of 4 nm and the cylinder had a radius of 1 nm, which correspond

approximately to the experimental measures of the nanoparticles and nanotubes.

67

Table 4.3. Hamaker constants and energies for typical substances and geometries used in

 

 

 

 

our experiments.

Substance Geometry A132(z.l) U12 (0.5/1) (21)
S i02/Benzene/Polystyrene Plate/Sphere 0.92 -0.86
Nanotube/Benzene/ Cylinder/Sphere 1.6 -0.43
Polystyrene
Polystyrene/Benzene/ Sphere/Sphere 0.1 1 -0.040
Polystyrene

 

 

 

 

 

By knowing the effective Hamaker constants, one may evaluate the potential

energy as a function of the closest distance between the two bodies, h as shown in Fig.

4.4. In this figure the combined effect of geometry and interaction strength can be seen.

The cylinder sphere potential is remarkably strong due in part to its geometry, but also

because the dispersion force is stronger between carbon and polystyrene than S102 and

polystyrene.

68

 

 

 

: —a— Sphere/Sphere
; +Sphere/Cylinder ;
-e-SpherelPIate .............................................

 

 

 

U12 (ZJ)

 

 

 

-10 A l n n n n n . n n A n A A
10'1 10° 10‘

h (nm)
Figure 4.4. Van der Waals potential energies for two body systems. In all cases, the

spheres are polystyrene, the cylinders are carbon nanotubes and the substrate (plate) is
$02.

From our calculations it is easy to see why the nanoparticles are attracted to the
nanotubes-there is a 150% gain in energy for doing so. Furthermore, the sphere/cylinder
interaction potential is extremely high validating the rather large structures seen in Fig.
4.1. So, as the nanoparticles are drawn toward the nanotube others are more likely to
follow. Similar effects have been studied using ultrafine fibers and aerosol particles. 84 As
the solvent is evaporating the fibers act like a filtration system that captures the particles

and pulls at them via their relatively strong van der Waals interaction.

69

4.4 Conclusion

As the nanoparticles settle closer to the substrate they are focused to the carbon
nanotubes through strong van der Waals forces. Using pairwise summation and the
Hamaker approach we were able to predict the strength of the van der Waals interaction
between the sphere and the nanotube compared to that of the sphere and the substrate.
From our results we concluded that the van der Waals strength between a sphere and
cylinder greatly increases as the particle gets closer to the substrate. This increase in van
der Waals interaction drives the nanoparticles to collect onto the nanotubes, which
creates an alignment of nanoparticles onto carbon nanotubes. By aligning carbon
nanotubes we will be able to achieve an array of nanoparticles that could benefit the

advances in future data storage devices and other possible nanodevices.

70

Chapter 5: Embedding Nanoparticles into a Cross-linked Network

5.1 Introduction
Robustly attaching the nanoparticle to the substrate is an integral component to
millipede technology. Since the nanoparticle will be deformed by an AFM tip, the
nanoparticles must be robustly attached to the surface to prevent tip forces from
disordering the nanoparticle array during deformation. Embedding the nanoparticles
within a polymer film thinner than the nanoparticle diameter is one way to achieve this.
However, this technology has other applications since others, are currently exploring the
optical properties produced by embedding silver nanoparticles into polymer films for
metal/semiconducting polymer systems.1°' 1‘ Silver is dispersed into vapor-deposited
nylon thin films (~100 nm) through heat treatment and are studied using optical and
vibrational spectroscopylo as well as transmission electron microscopy (TEM) to find that
nanoparticles placed on the surface will homogenously distribute throughout the film
after annealing. One of these studies10 was followed by another to find that chemical
composition of the nanoparticles is critical in controlling dispersion.85
Teichroeb and Forrest have investigated embedding Au nanoparticles in a polymer
film (~180 nm thick) by annealing the film above its glass transition temperature.86 The
Au nanoparticles were deposited on top of the film and as the system was heated the
nanoparticles slowly fell into the film. Atomic force microscopy (AFM) was used to
measure the height of the nanoparticles at different annealing times and as the annealing

time increased the height of the Au nanoparticles decreased suggesting they slowly

embedded into the polymer film. The interesting observation in this work was that the

71

 

nanoparticles were embedded into the film ~3-4 nm even when the film was at a
temperature below the bulk glass transition.

In contrast to the above studies we investigate embedding polystyrene
nanoparticles in a thin cross-linkable polymer film where the nanoparticle diameter is
greater than the thickness of the polymer film, which allows creation of a memory device
discussed above as well as a nano-rough surface. Researchers have recently begun to
explore the effect of nano-rough surfaces (features <100nm) on surface wettability.”91 In
one case swift ion irradiation was used to produce a random array of nanometer features
on a surface.87 It was discovered that as the defect concentration or roughness increased
the hydrophobicity of the surface also increased. Meli and Lennox88 have also seen an
increase in hydrophobicity with increased roughness, but with highly regular gold
nanoparticles and surface topology with roughness on the order of 10 nm in height. In
this work we study the wetting behavior of a nanorough surface having features that are

~3 nm high with a unique system where the nanoparticles are chemically identical to the

polymer film so roughness effects can be solely addressed.

72

5.2. Experimental

Polystyrene nanoparticles were used as a well characterized component and were
synthesized according to the procedure of Harth, et al.15 and kindly provided by Prof.
Craig Hawker. A linear, random, copolymer precursor consisting of styrene monomer
and benzocyclobutene (BCB) was first synthesized where the latter component dictates
the degree of cross-linking. In this work we use a BCB level of 20 mol%, denoted as
tightly (T) cross-linked. A dilute solution of the linear precursor polymer was then
dripped into a hot solvent activating the intermolecular cross-linking reaction to create
nanoparticles containing a single macromolecule whose size is dictated by the initial
precursor molecular weight as well as the degree of cross-linking. The nanoparticles were
prepared at concentrations of 50 ug/mL, 100 ug/mL, and 200 ug/mL in benzene, since
concentrations greater than 100 ug/mL revealed agglomeration of the nanoparticles in the
spin coated films discussed below. A PSZSkD linear precursor containing 2.5 mol% BCB
was used for the polymer film since it is chemically identical to the nanoparticles. Details
of the nanoparticle system and linear precursor can be found in Table 5.1. Note, to
ascertain the nanoparticle conformation on the substrate, we determined the diameter (D)
assuming the macromolecule collapses to a sphere with the bulk density of polystyrene (p

~ 1.04 g/cm3) through

 

Y
D=[6M"]3 (5.1)
WHO

where N A is Avogadro’s number and Mn is the number average molecular weight.

73

Table 5.1. Sample, Molecular Mass, Polydispersity Index (PDI), Degree of Cross-
linking, Nature (i.e. cross-linked or linear), and Diameter if the Linear Precursors or
Nanoparticles Collapse to the Bulk Density of Polystyrene (Eq.5.l) for the Samples.

 

mol%
Mn cross-linking diameter
Sample Code (kD) FBI agent (nm)
P825k 24.5 1.14 2.5 4.2
T-P8211k 211 1.32 20 8.6

Solutions of the linear precursor were prepared in benzene at a concentration of 5
mg/mL. The nanoparticles’ solution at each concentration and linear precursor solution
were mixed together in equal parts and then spin coated onto a SigmacoteTM silanized
wafer at 5000 rpm for 40 seconds. SigmacoteTM (Sigma-Aldrich) is a solution consisting
of 2.5% chlorosiloxane ((SiC12C4I-Ig)2O) and 97.5% heptane that functionalizes the
surface with short alkane chains. During the spin coating process the wafer was flooded
with the polystyrene solution and left to stand for ~2 minutes prior to starting the spin
coater. This delay was required due to the low surface energy of the substrate and the
requirements that a homogenous film must be manufactured. Note we used the silanized
wafer as a convenient way to create an ultrathin polymer film (~5nm) since the same
concentration spin coated onto a higher free energy substrate, such as a silicon wafer,
would have produced a 40 nm thick film. In addition, the subsequent cross-linking
procedure on a bare silicon wafer produced a highly adsorbed film that was difficult to
characterize by the technique discussed below to determine its thickness. Furthermore, a
higher energy substrate substantially deformed the nanoparticle used here.12

The samples were heated in an oven to 250°C for 30 minutes to produce a cross-

linked network of polymer film that will firmly embed the nanoparticle within the film.

74

At 250°C the pendent cross—linking group (benzocyclobutene, BCB) along the linear
precursor chain will bond with another cross-linking group to create a highly cross-linked
network.

Surface profile measurements were performed with a Pacific Nanotechnology
Nano—R atomic force microscope in close contact (oscillating) mode to generate height
images that were not altered other than a simple leveling procedure. Silicon tips with a
spring constant of 25-75 N/m, tip curvature of <10 nm, and a resonance frequency of

200-400 kHz were used for all experiments.

75

 

5.3. Results and Discussion
Initially only the linear precursor was spin coated onto the substrate and then
imaged before and after heating to ensure a smooth film was present (Fig. 5.1a). A razor
blade was used to create a thin scratch in the film which was subsequently imaged via
AFM to determine the film thickness. The measured thickness was 5 nm and a line
analysis of this is shown in Figure 5.lb. Below a thickness of 4 nm a continuous film was
unattainable, which made it clear that only the T-PSlekD nanoparticle system (highest
molecular weight nanoparticle system available to us)12 would still be elevated enough to
image after embedding.
The height of a bare T-PSZl lkD nanoparticle on a silanized wafer is ~8-9 nm (see
Fig.5 .2), suggesting an embedded particle will appear 34 nm in height. Note the systems
were imaged before and after heat and they were heated to 250°C to activate the cross-
linking process. The average height of the nanoparticles after the film was cured was ~
2.6 i: 0.4 nm, which suggests that the particles were embedded within the film and that

the nanoparticles were essentially in a spherical conformation within the film.

76

25kD L- PS

   

0 5 10 15 20 25
pm

Figure 5.1. (a) AFM height image of PSZSkD film on SigmacoteTM after being heated to
250°C for 30 min. (b) Line analysis of the scratch made in the film. The film appears to
be 4-5 nm thick.

An AFM height image of the embedded T-PSlek nanoparticles in the cross-
linked polystyrene film is shown in Figure 5.2 and 5.3. Nanoparticle concentrations of
SOllg/mL, 100 ug/mL, and 200 ug/mL were used to see if roughness of the surface would
increase with a greater number of nanoparticles embedded within the film. The number of
nanoparticles per umz (np) was calculated for the concentrations and was determined to
be ~9 for the 50 ug/mL sample, ~16 for the 100 [lg/mL sample, and ~3 for the ZOOllg/mL
sample. An AFM image of the nanoparticles prior to and after curing the film at the
50ug/mL concentration is shown in Fig. 5.3a and b. There was little to no change in the
height after being heated. A height profile for the nanoparticles before and after
embedment is shown in Fig. 5.2. The difference in height is ~5-6 nm, which suggests a

film thickness of ~5nm. The sample was also dipped in benzene for ~l minute to

77

determine if the film or nanoparticles would dissolve away. Since the film is cross-linked
the nanoparticles embedded within the film should remain intact and from the image in

Fig. 5.2c it is clear that the nanoparticles and film did not dissolve.

 

= - = g bi

4— T-P3211k NPs
10 :- - -- Embedded T-PSZ11k NPS *

 

 

 

.-. .. .-‘

 

 

 

 

 

 

 

 

 

 

 

gar ‘

* ~ nm

. . 3' ' ,fl- 1
L 2|_1_'. 1’ g a ’4..." .‘ "

V'l" : ..

0 1 2 3 4 5

pm

Figure 5.2. (a) Diagram of nanoparticle embedded in thin polymer film. (b) Graph of
the height of a T-PSlek nanoparticle on SigmacoteTM before embedment (black line)
and the height of a nanoparticle after embedment (dashed line). The average difference
in height between these particles is ~ 5-6 nm consistent with the film thickness
determined in Fig. 5.1.

78

0. 00111:: 2 .00pm . 2 .OOrlm

 

0.00pm 2.00pm . 0.00pm 2.001.111:

   

11.001211

0 . 00mm 2 . 00111:-

Figure 5.3. AFM height images of T-PSlek NPS embedded in a thin polymer film
(PS25kD linear precursor). (a) Height image prior to cure (50 ug/mL). (b) Height image
after heating to 250°C for 30 minutes (50 ug/mL). (c) Height image after the sample in
bwas dipped in benzene for 1 minute (SOMg/mL). (d) Height image after heating to 250°C
for 30 minutes (100 ug/mL). (e) Height image after heating to 250°C for 30 minutes
(200ug/mL).

The surface energy of the annealed polystyrene film and the higher concentration

(100 ug/mL) of embedded nanoparticles within the film were studied using contact angle

measurements. The advancing (0A) and receding (0R) contact angles for both systems

79

were 0A: 89.0° : 4.3, OR: 87.4 1: 2.7 and 0A: 81.10 :t 0.9, OR = 75.9 :1: 5.9 for the cross-
linked polystyrene film and the embedded nanoparticle film, respectively. Note the
advancing and receding contact angles for the 50 ug/mL concentration of embedded
nanoparticles were essentially equal to the contact angles measured on the polystyrene
film with no nanoparticles embedded, 0A = 89.7 i 3.0 and GR = 84.1 i 3.7. We note that a
sample not discussed here, 10 ug/mL, had similar contact angles, also, the samples spin
coated and cured with 150 and 200 ug/mL nanoparticle solutions showed contact angles
of ~89°, however, they showed phase separated regions (see Fig. 5.3e).

The wetting properties of rough surfaces92 have been vastly explored by many
researchers. They have found that as the roughness of the surface increases the surface
energy of the surface increases;52’ 93'” however, this increase in hydrophobicity occurs
when the contact angle on the smooth surface is greater than 90°. Others have found that
if the surface is initially hydrophilic (contact angle <90°) then by increasing surface
roughness, the contact angle will actually decrease creating a more
hydrophilic surface.96' 97 Bhushan and Jung97 have considered the hydrophobic and
hydrophilic nature of leaf surfaces to study the effect of rnicrobumps (3-7um in height)
and nanobumps (70-780 mm in height) on the surface to find nanobumps had a greater
effect on hydrophobicity than microbumps. Hydrophilic leaves were also studied to better
understand roughness changes on hydrophilic surfaces where it was determined that as
the roughness of hydrophilic surfaces increases the hydrophilicity increases.

Young’s equation is used to determine the contact angle (03) between a smooth

solid surface and a liquid droplet and is expressed as92

cos 65 = m (5.3)

7L

80

where 75, n, 751. are the surface tensions of the solid-gas, liquid-gas, solid-liquid
interfaces. Yet, Young’s equation does not account for surface roughness effects and
therefore would not apply to the polystyrene nano-rough surfaces described here. A
model to describe roughness effects on hydrophobicity was first presented by Wenzel in
193698 to relate contact angles for a chemically homogenous yet rough surface using a

roughness factor rf,

cos 0, = rf cos 65 (5.4)

where Or is the contact angle on the rough surface.

The Wenzel model can be used to predict that when rf increases the contact angle
will decrease if 08 is less than 90°, which is what we found in our study although the
amplitude of our roughness is ~3nm. The polystyrene film (with no nanoparticles) is a
slightly hydrophilic surface with a contact angle less than 90°. When nanoparticles are
embedded in the polymer film to increase roughness we see a decrease in contact angle
by ~10°. Using the advancing angles we can calculate Wenzel’s prediction of rf using eq.
5.4. For the high concentration of nanoparticles (lOOpg/mL) we find an rr value of 9.

The calculated rf value can also be considered by using a spherical cap model to
find the surface area of the nanoparticle embedded in the film. We first determine the
contact diameter (2a) of the nanoparticle embedded in the film (see Fig. 5.2a) using the

surface area of a spherical cap (Ase)
ASC = 7zDh = 7r(a2 + hz) (5.5)

where h is the height of the embedded nanoparticle (h~3nm), and D is the predicted
nanoparticle diameter (eq. 5.1). The ff value is the ratio of the actual area to the projected

area and can be calculated from

81

A +L2— 2
rf= 5" L2 m where L2=—1— (5-6)

up

 

The calculated If value for the 100 ug/mL sample is 1.0004, which is much less than the
rf value calculated from the Wenzel model (~9). This large variance in rf suggests that
nanoscale roughness has a large effect on wettability with the possibility of forming
extremely hydrophilic surfaces. Note the Cassie and Baxter model extends the Wenzel
model to heterogeneous surfaces.” 100 This model was also developed for both porous
substrates and rough hydrophobic substrates and therefore does not apply to our
particular case.

The contact angle hysteresis also plays a role in nano-rough surfaces. According
to the Wenzel model as the roughness of the sample increases, the contact angle
hysteresis should also increase. Contact angle hysteresis is the difference between the
advancing and receding contact angle (A0). For the cross-linked polystyrene film this
value is ~l.6 and for the embedded nanoparticles this value increased to ~52, suggesting

that this increase in hysteresis due to increased roughness holds true for our system.88

82

5.4. Conclusion

In this work we were able to robustly attach the nanoparticles to the substrate by
embedding them (locking them) into a cross-linkable film. The nanoparticles are now
capable of being deformed by an AFM tip without being displaced.

Nanoparticles embedded in a polymer film created nano—rough surfaces that
influenced the wettability of the polystyrene surface. Due to polystyrene’s hydrophilic
nature (contact angle <90°) the addition of polystyrene nanoparticles to create a nano-

rough surface increased the hydrophilicity of the surface.

83

Chapter 6: Conclusions

The conformation of intramolecularly cross-linked polystyrene nanoparticles on a
solid substrate was investigated. Optimal rigidity of the nanoparticles on the substrate
must be accomplished in order to achieve advancements in data storage systems. The
substrates tested include a high energy mica surface and a low energy silanized silicon
wafer. The nanoparticles collapse on the mica substrate, but remain robust and structured
on the silanized wafer, yet, an extreme amount of cross-linking is required for the
nanoparticles to retain their original spherical shape regardless of the substrate surface
energy. The nanoparticle behavior was also observed at elevated temperatures to reveal
that the height of the extremely cross-linked nanoparticles slowly decreases. The
temperature where a rapid size change occurs was well below the bulk glass transition
temperature.

Polystyrene macromolecules were also investigated at elevated temperatures and
it was discovered that these macromolecules fluctuate between two conformational states:
a collapsed state at low temperatures (25°C) and an expanded state at high temperatures
(200°C). We used a combination of experiment and Monte Carlo molecular simulation to
examine and explain this phenomenon.

Arrays of nanoparticles could then be produced by focusing them to SWNTS. As
the nanoparticles settle closer to the substrate they are focused to the carbon nanotubes
through strong van der Waals forces. The SWNT serves as a nucleation site that directs
the nanoparticles to it.

The nanoparticles were then robustly attached to the substrate by embedding them

within a cross-linked thin polystyrene film. Locking the nanoparticles in place allows one

84

to deform a single nanoparticle without displacing or disrupting the nanoparticle.
Deformation of single nanoparticles is not currently achievable due to the limits of the
AFM.

The items addressed in this'work build a path for the possibility of nanoparticles
being addressed as single bits of information along a surface for the advancement of

molecular memory.

85

Appendix A: Atomic Force Microscopy

Atomic force microscopy (AFM) was invented in 1986 by Binnig, et al.101 The
AFM reflects a laser from a solid state diode off a cantilever to a detector. A sharp probe
located on the end of the cantilever moves over the surface of the sample while
measurements of the signal changes in the bending of the cantilever are taken. A diagram

of the AFM set~up is shown in Figure A.1

 
   

X Piezoelectric
Y Piezoelectric
Z Piezoelectric

Feedback
Unit

Figure A.1. Atomic force microscopy diagram that illustrates the laser being emitted
from the diode to the cantilever and reflected back to the detector. The feedback unit uses
the input signal to produce the voltage that drives the piezoelectric ceramic.

A Pacific Nanotechnology Nano-R atomic force microscope in close contact
mode was used for all images presented in this work. Close contact (oscillating) mode
produces four outputs: Z(error), Z(height), Z(sensor), and Z(demodulated). A list is

provided below that contains a detailed description of each output.

Z(error): The cantilever deflection, as a function of the actuator position. This signal is
used to determine if the tip has made contact with the sample.

Z(height): As the tip moves over the sample, it periodically taps the surface, while
measurements of the signal changes in the bending of the cantilever are taken.

Z(sensor): This linearizes the Z(height) signal using the Z piezo. The Z(sensor) provides
a more accurate height measurement than the Z(height) because it is linearized.

86

Z(demodulated): (phase image) Measures changes in phase angle during scanning due to
energy dissipation during tip-sample interaction, which can be caused by changes in
topography, tip-sample molecular interactions, deformation of the tip-sample contact, and
experimental conditions. The phase image can also be used to determine if your tip is
broken or has something stuck to it such as a dust particle or nanoparticle. The phase
image can also determine if your sample is contaminated. Examples of good and bad
phase images are shown in Figure A.3.

   

211.3.st 2543.0'hlv

   

,_ ,~ _ ~ . rwrr"—' ' .
0.00pm 0.9 pm . - b.00pm 2.59'pm 5.19pm

 

Figure A.2. (a) Example of a good AFM phase image of 33kD E—PS NPS. (b) Example of
a bad phase image of 33kD E-PS NPs. Some of the particles in the image appear
triangular which indicates the tip is broken. The streaks that appear in the phase image
indicate that either the surface is contaminated (e.g. finger print) or something is stuck to
the tip such as a dust particle or nanoparticle.

For all AFM experimental data presented in this work, the particle heights were
determined by taking the average of 50 particles’ heights; however, the lateral size could
not be found due to convolution effects created by the tip.”’ 19

AFM tip damage before and after scanning was also investigated. Scanning
electron microscopy (SEM) was used to determine if the tip was damaged after tip
approach and after scanning a freshly cleaved mica substrate. SEM images of the AFM
tip at 3 different stages is shown if Figure A.3a, b, and c. An example of a damaged
AFM tip is shown in Figure A.3d. The damaged tip is rounded on the end suggesting that

the tip was damaged during tip approach or sample contact.

87

 

Figure A.3. (a) AFM tip prior to tip approach. (b) AFM tip after tip approach. (c) AFM
tip after scanning a mica substrate. ((1) Damaged AFM tip.

AFM calibration for lateral resolution was performed using a calibration grid.
Since the particles imaged for this work were less than 10 nm, gold nanoparticles (Au
NPS) were used to calibrate the AFM in the vertical direction. Au NPs were purchased
from Meliorum Technologies and are suspended in toluene. The particles average height
via AFM was ~7 nm. We use these particles to ensure that the measured height values for
other particles are accurate.

A hot stage was designed for the Nano-R AFM so that individual molecules and
nanoparticles could be observed at elevated temperatures. The hot stage consisted of a
solid metal cylindrical sample puck with an alumina insulator, heater element, and
thermocouple connected to a PID controller to monitor the temperature (see Figure A.4).

AFM scans at temperatures above room temperature require that the hot stage be turned

88

off during scanning due to electrical interference effects; therefore, the high temperature
experiments have a temperature range of approximately : 5°C as the heater cools slightly
during AFM scans. In all AFM elevated temperature experiments the sample and AFM
components were kept at the desired temperature for ~3 hours to ensure that all
components, specifically tip and sample, had reached the desired temperature, which
aided in the prevention of thermal drift.

Thermocouple
Sample

Heater Element

Alumina (0.094” thick)

 

AFM Nana-R Puck

Figure A.4. AFM heater design. Samples can be heated to a maximum temperature of
100°C without damaging the AFM.

Calibration of the hot stage was performed by imaging Au NPs spin coated on a
mica substrate at temperatures ranging from 25°C to 75°C. At each temperature the
average AFM height of the Au NPs was ~7 nm.

The final task of this research is to indent a single nanoparticle via an AFM tip.
Unfortunately due to limitations of the AFM this was not achievable. In an attempt to
indent a nanoparticle on a single-walled nanotube it was discovered that the AFM x-y
resolution was not accurate. AFM images of nanoparticles and SWNTS before and after
tip indentation is shown in Figure A5. The circled area designates where the tip should
have struck the sample and the dotted circle is the debris from where the tip actually

struck the sample.

89

 

. 0.
0.00pm 1 .75pm . 0.00pm 1.0 pm

Figure A.5. (a)AFM height image of nanoparticles aligned on single-walled nanotubes
before tip approaches sample (b) AFM height image of nanoparticles aligned on single-
walled nanotubes after tip approaches sample. In both (a) and (b) the black circle
represents the targeted tip approach area and in (b) the dotted circle represents where the
tip actually struck the sample.

90

Appendix B: Height vs. Temperature Data for Polystyrene Macromolecules and
Nanoparticles

The graphs displayed below contain height vs. temperature data for polystyrene
macromolecules and nanoparticles. Figures BI and B2 contain graphs of polystyrene
macromolecules that were heated from 25°C to 75°C. In both cases a softening
temperature occurs at ~ 40°C. Both samples were sprayed onto a SigmacoteTM substrate
from a solvent/non-solvent solution.

Nanoparticle height vs. temperature data can be found in Figures B3 and B.4.
Both of these systems were drop deposited onto a SigmacoteTM substrate in benzene. In
both cases a softening effect at ~ 40°C is observed.

A graph of the 393kD macromolecules is shown in Figure B5. The
macromolecules were heated in an oven to 200°C to 75°C for 30 cycles. The values
plotted in the graph are for heights after cycles 20 and 30. This proves that the
macromolecule fluctuation is still observed after several cycles. Heat cycles between
200°C and 75°C were also performed on the polystyrene nanoparticles. The AFM height
of the 211kD T-PS NPs was observed for 3 heat cycles and is plotted in Figure B.6. The
nanoparticles were deposited in pure solvent (benzene) and in solvent/non-solvent
(benzene/Z-MEA).

Fourier transform infrared spectroscopy (FT IR) and X-ray photoelectron
spectroscopy (XPS) were performed to prove that 2-MEA was not present in the sample
after initial preparation. Figures B7 and B8 contain FTIR and XPS spectrums of the
sample after initial preparation of the 393kD polystyrene macromolecules in benzene/2-
MEA. Both spectrums reveal that there is no residual solvent or non-solvent still present

after initial preparation.

91

 

 

: E
'- E 5 d
a 1 i
l E
l l =
l l I
A 5 l A : ——-- .
2 1 i
E I , . .
E . ‘ : x
C ; ~ = 2 i
4—0 ”_ l t l .
i 2 2
c i 3 I 3
U) z . x :
o i l 2 Z 1
1 i i i 2
l ' ' i
i

i

 

o. 44kD PS ‘
O 44kD PS (after heat)
----- Predicted Height
3 T l - - - - . - - -
30 40 50 60 70

i

 

 

 

 

 

 

Temperature (9C)

Figure B.l. Height vs. temperature for 44kD polystyrene macromolecules heated from
25°C to 75°C. The sample was spray deposited from a solvent/non-solvent solution onto
a silanized wafer.

 

 

6— g -—
A .
E _ s
5 f
s it 0 '
4_ i .:_
‘ o 1.5MD PS ‘

 

 

 

 

3 Lrl A; 1.8 MD PS (afterheat):
30 40 50 60 70

 

 

Temperature (QC)

Figure B.2. Height vs. temperature for 1.5MD polystyrene macromolecules heated from

25°C to 75°C. The sample was spray deposited from a solvent/non-solvent solution onto a
silanized wafer.

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 ; 78kD rips Ne ' r
r o 78kD T-PS NP (after heat) 4
7 ---- Predicted Height
E 6 ---""'== ===== ——--- = ----- cc.
1; - _ _
" 5
:‘E B t} . _
4 . Q .
3 , J l I I
30 40 50 60 70 80

Temperature (9C)

Figure 8.3. Height vs. temperature for 78kD T-PS nanoparticles heated from 25°C to
75°C. The sample was drop deposited onto a silanized wafer from benzene.

 

-

it! it
’0

Height (nm)
0)

r
1 l l l I
. . ~ I .
l r I l

8 —.......3... ................I..................-r_....-......-......r....-......-......
i . .
. . 0
i r
r l .
.

n
l
: _
i
”T“ 1 I ..—
1
-
i
: i : :
1.........e.......-!......-..........H.T...-......-........l.....—
. g . r .
3 l . E i -
l i I I l
2 g .
l i l
......4....................................... ...-.............. “.-...—
’ ' r i .1
l q

 

4 _--..i o 211kD T-PS NP

 

 

.{fi

1 --.-- Predicted Height

I 211k T-PS NP (after heat) :

 

 

 

-.... L

L

J

 

30 40 50 60
Temperature (“0)

Figure B.4. Height vs. temperature for 211kD T-PS nanoparticles heated from 25°C to
75°C. The sample was drop deposited onto a silanized wafer from benzene.

93

70

 

 

 

6'0 | o 393kDa PS (cycle 20 & 30)]‘

5.5 -----w l --.._. :

 

: .- i E E
. . . . =
I i I I '.
l l : I :

l i 1 i r t
3 1 i l l r
. l . . 1 .

1

Height (nm)

1 J i l I
Z l l l 3 1
l l l l l 1
- . r 1 l : :
' : I l l : :
i i 1 1 I 1
r i ' i i

1 l i : l E

3 i 3 i

I .' i 1 I 1

4 0 —.s- .......;.-......-..........1-......-........~;.............-.._'_

I . i i K 1 i

i 3 I I l

3 i l l i

1 i l i ‘

 

 

 

259C

759C (C-20)
20090 (C-30)

200°C (C-20)

Temperature (”C)

Figure B.5. Height of 393kD PS macromolecules at room temperature after being heated
to the corresponding temperatures on the graph. The height of the macromolecules
fluctuates as the temperature is cycled between 75°C and 200°C. The graph shows the
heights obtained after cycles 20 and 30.

 

 

 

68 .11..-- l .. lfi. .. ._ l... I I. 6.8 1.1. ---.-. . ..---ML . ---! -.--1....... ..-.I......-.....I....._--!_

6'6 7” b.— 2211KpaT'PS Npsl- 6.6 --I;o— 211koa r-Ps :NPSl-

o ‘ _
' l l l l l r l
l 1 r r i l I
l i 1 l i
l l l l l
_ 5 8 ....--“ ......... . . -.--.. ”mph.--“—
. . .‘ .
g i
' l
I ' l
l

. . | 5.6 r? \ Sigmacote i—

, , 5.4 ....... (Bmzene’Z-hEA).l_
5.0 ~l~ --l lml ~

. . 5.0rl “i
25" 75° 200°

 

 

 

     

 

Height (nm)
Height (nrn)

 

 

 
 
   

 

 

 

r i
l i i
l l :

       

    

   

 

 

 

 

 

 

 

l

 

200° 759 20091759 259 200° 759 '2009 759 2009 759
Temperature (°C) Temperature (“0)

Figure B.6. Height vs temperature graphs of the 211kD T-PS NP system. (a) NPs were
deposited in pure benzene and (b) NPs were deposited in benzene/Z-MEA. The
nanoparticles do not undergo the same two phase behavior that is observed in the
macromolecules.

94

' Iflfl“: t. a. InapI‘III 1"... 1

4538.5:
do 2528 3:23.. Me com: 0: mm 0.85 .58 05 E “some“. 0.3 $32. ocohbmbom i=0
gob; 35:36 a 95 3.3:? we? <92-N esonEoucon “Neon e8 noun—cm AER: fie
< doses—amen coca .3026 3302088an mm 95% me 5533 EE KA— 0.53..—

céov Eon—5533

 

 

 

 

 

 

 

 

 

.N08.?
508.0
3890

1ecood

 

.886 W
......m
.286
.886
“T286

“.286

 

352.

1800.0

95

$3 >995 9:25

4528-5:
5 8038 3:282 as: mo 8:: o: a Bush. 6::8wxoan 05 :8 :8: 83 has: @3628
2:. 3%? tour—«=8 : 3:0 @388 8:3 <§-~ $33583 Que we E528 4.5m:
fio < donut—38m St: 389:: 338—08888 mm 9.3m b8 8:588 men .wfi ”...-firm

 

 

 

 

 

 

 

0 GOP CON 00m Gov com 80 con com com 000—. CO _. _.
_ _ _ _ _ _ 1 _ _ _ o
.0
u a n N
l v
x
w
I .l C
9
B m»
I a . I m
S 0
w w
1
7 l o—
.l l Nw
I. 1 v_.
I o L m:
m.»
— - _ L _ — n h L I]. _ CF
0.. x

m8$$§wo

9/0

96

References

l. Vettiger, P.; Cross, G.; Despont, M.; Drechsler, U.; Durig, U.; Gotsmann, B.;
Haberle, W.; Lantz, M. A.; Rothuizen, H. E.; Stutz, R.; Binnig, G. K. Nanotechnology,
IEEE Transactions on 2002, 1, (1), 39-55.

2. Eleftheriou, E.; Antonakopoulos, T.; Binnig, G. K.; Cherubini, G.; Despont, M.;
Dholakia, A.; Durig, U.; Lantz, M. A.; Pozidis, H.; Rothuizen, H. E.; Vettiger, P. Ieee
Transactions on Magnetics 2003, 39, (2), 938-945.

3. Durig, U.; Cross, G.; Despont, M.; Drechsler, U.; Haberle, W.; Lutwyche, M. 1.;
Rothuizen, H.; Stutz, R.; Widmer, R.; Vettiger, P.; Binnig, G. K.; King, W. P.; Goodson,
K. E. Tribology Letters 2000, 9, (1-2), 25-32.

4. Vettiger, P.; Despont, M.; Drechsler, U.; Durig, U.; Haberle, W.; Lutwyche, M. I.;
Rothuizen, H. E.; Stutz, R.; Widmer, R.; Binnig, G. K. Ibm Journal of Research and
Development 2000, 44, (3), 323-340.

5. Mackay, M. E. Ieee Transactions on Nanotechnology 2005, 4, (5), 641-644.

6. Xue, G.; Lu, Y.; Shi, G.; Dai, Q. Polymer 1994, 35, (4), 892-894.

7. Rana, N.; Yau, S.-T. Nanotechnology 2004, 15, (3), 275-278.

8. Teranishi, T.; Sugawara, A.; Shimizu, T.; Miyake, M. J. Am. Chem. Soc. 2002,
124, (16), 4210-4211.

9. Mathur, A.; Brown, A. D.; Erlebacher, J. langmuir 2006, 22, (2), 582-589.

10. Akamatsu, K.; Tsuboi, N.; Hatakenaka, Y.; Deki, S. J. Phys. Chem. B 2000, 104,
(44), 10168-10173.

11. Pore], 8.; Singh, S.; Harsha, S. S.; Rao, 1). N.; Radhakrishnan, T. P. Chem. Mater.
2005, 17,(l), 9-12.

12. Dukette, T. E.; Mackay, M. E.; VanHorn, B.; Wooley, K. L.; Drockenmuller, E.;
Malkoch, M.; Hawker, C. J. Nano Lett. 2005, 5, (9), 1704-1709.

13. Zhang, Q.; Clark, C. G.; Wang, M.; Remsen, E. E.; Wooley, K. L. Nano Lett.
2002, 2, (10), 1051-1054.

14. Richardson, M. J. Journal of Polymer Science: Part C 1965, (3), 21-29.

15. Harth, E.; Horn, B. V.; Lee, V. Y.; Germack, D. S.; Gonzales, C. P.; Miller, R. D.;
Hawker, C. J. J. Am. Chem. Soc. 2002, 124, (29), 8653-8660.

16. Israelachvili, J. N.; Alcantar, N. A.; Maeda, N.; Mates, T. E.; Ruths, M. Langmuir
2004, 20, (9), 3616—3622.

97

17. Israelachvili, J. N., Intermolecular and Surface Forces. Academic Press: 1992.
18. Ebenstein, Y.; Nahum, E.; Banin, U. Nano Lett. 2002, 2, (9), 945-950.
19. Markiewicz, P.; Goh, M. C. Langmuir 1994, 10, (1), 5-7.

20. Mecke, A.; Lee, I.; Baker, J. R.; Holl, M. M. B.; Orr, B. G. The European
Physical Journal E 2004, 14, 7-16.

21. Chubynsky, M. V.; Thorpe, M. F. Current Opinion in Solid State and Materials
Science 2001, 5, (6), 525-532.

22. Johnson, K. L.; Kendall, K.; Roberts, A. D. Proceedings of the Royal Society of
London. Series A, Mathematical and Physical Sciences 1971, 324, (1558), 301-313.

23. Kendall, K.; Padget, J. C. International Journal of Adhesion and Adhesives 1982,
2, (3), 149-154.

24. Fowkes, F. M. Ind. Eng. Chem. 1964, 56, (12), 40-52.
25. Dee, G. T.; Sauer, B. B. Colloidal Interface Science 1992, 152, (1), 85-103.

26. Ferry, J. D., Viscoelastic Properties of Polymers. 3rd ed.; John Wiley & Sons:
1980; p 672.

27. Rubenstein, M.; Colby, R. H., Polymer Physics. Oxford University Press: 2003; p
440.

28. Dobrynin, A. V.; Deshkovski, A.; Rubinstein, M. Physical Review Letters 2000,
84, (14), 3101.

29. Dobrynin, A. V.; Rubinstein, M. Macromolecules 2002, 35, (7), 2754-2768.
30. G. Xue, Y. L., G. Shi, and Q. Dai. Polymer 1994, 35, (4), 892-894.

31. Mundra, M. K.; Ellison, C. J .; Behling, R. E.; T orkelson, J. M. Polymer 2006, 47,
(22), 7747-7759.

32. Priestley, R. D.; Ellison, C. J .; Broadbelt, L. J .; Torkelson, J. M. Science 2005,
309, (5733), 456-459.

33. See, Y. K.; Cha, J .; Chang, T.; Ree, M. Langmuir 2000, 16, (5), 2351-2355.

34. Shin, H. S.; Jung, Y. M.; Oh, T. Y.; Chang, T.; Kim, S. B.; Lee, D. H.; Noda, I.
Langmuir 2002, 18, (15), 5953-5958.

35. Serghei, A.; Tress, M.; Kremer, F. Macromolecules 2006, 39, (26), 9385-9387.

98

36. Erichsen, J .; Dolgner, K.; Zaporojtchenko, V.; Faupel, F. Macromolecules 2004,
37, (24), 8813-8815.

37. Fryer, D. S.; Peters, R. D.; Kim, B. J .; Tomaszewski, J. E.; de Pablo, J. J .; Nealey,
P. F.; White, C. C.; Wu, W. l. Macromolecules 2001, 34, (16), 5627-5634.

38. Giessibl, F. J.; Quate, C. F. Physics Today 2006, 59, (12), 44-50.
39. Sugden, S. Journal of the Chemical Society 1924, 125, 32-41.

40. de Gennes, P., Scaling Concepts in Polymer Physics. Cornell University: Ithaca,
1979.

41. Netz, R. R.; Andelman, D. ArXiv: cond-mat/0002266 2000, 1, pp. 1-44.

42. Frenkel, D. a. B. S., Understanding Molecular Simulation. 2002; Vol. Academic
Press.

43. Hammersley, J. M.; Handscomb, D. C., Monte Carlo Methods. John Wiley &
Sons, Inc.: New York, 1964.

44. A. Michel, S. K., and B. Wunderlich In Molecular dynamics simulations of the
adsorption of single chains on surfaces, 12th annual workshop, Recent developments in
computer simulation Studies in condensed matter physics, 2000; 2000.

45. Flory, P. J ., Statistical Mechanics of Chain Molecules. Interscience Publishers:
1969.

46. Chandrasekhar, S. Reviews of Modern Physics 1943, 15, (1), 1.

47. Doye, J. P. K.; Sear, R. P.; Frenkel, D. Journal of Chemical Physics 1998, 108,
(5), 2134-2142.

48. Bachmann, M.; J anke, W. Physical Review Letters 2005, 95, (5), 058102-4.

49. Rajesh, R.; Dhar, D.; Giri, D.; Kumar, S.; Singh, Y. Physical Review E 2002, 65,
(5), 056124.

50. K. Binder, J. B. M. M. W. P. F. R. Macromolecular Symposia 2006, 237, (1),
128-138.

51. Grassberger, P. Physical Review E 1997, 56, (3), 3682.
52. Liu, Y.; Chen, X.; Kim, J. H. Nanotechnology 2006, 17, (13), 3259-3263.

53. Rosenbluth, M. N.; Rosenbluth, A. W. The Journal of Chemical Physics 1955, 23,
(2), 356-359.

99

54. Wall, F. T.; Erpenbeck, J. J. The Journal of Chemical Physics 1959, 30, (3), 634-
637.

55. Wall, F. T.; Erpenbeck, J. J. The Journal of Chemical Physics 1959, 30, (3), 637-
640.

56. Binder, K.; Baschnagel, J.; Muller, M.; Paul, W.; Rampf, F. Macromolecular
Symposia 2006, 237, (1), 128-138.

57. Sohn, B. H.; Choi, J. M.; Yoo, S. I.; Yun, S. H.; Zin, W. C.; Jung, J. C.; Kanehara,
M.; Hirata, T.; Teranishi, T. J. Am. Chem. Soc. 2003, 125, (21), 6368-6369.

58. Jang, J .; Oh, J. H. Langmuir 2004, 20, (20), 8419-8422.

59. Haynes, C. L.; McFarland, A. D.; Smith, M. T.; Hulteen, J. C.; Van Duyne, R. P.
J. Phys. Chem. B 2002, 106, (8), 1898-1902.

60. Cui, Y.; Bjork, M. T.; Liddle, J. A.; Sonnichsen, C.; Boussert, B.; Alivisatos, A.
P. Nano Lett. 2004, 4, (6), 1093-1098.

61. Kuo, C. W.; Shiu, J. Y.; Chen, P. Chem. Mater. 2003, 15, (15), 2917-2920.
62. Huang, S.; Woodson, M.; Smalley, R.; Liu, J. Nano Lett. 2004, 4, (6), 1025-1028.

63. Kusunoki, M.; Rokkaku, M.; Suzuki, T. Applied Physics Letters 1997, 71, (18),
2620-2622.

64. Li, W. Z.; Xie, S. S.; Qian, L. X.; Chang, B. H.; Zou, B. S.; Zhou, W. Y.; Zhao,
R. A.; Wang, G. Science 1996, 274, (5293), 1701-1703.

65. Banerjee, S.; Wong, S. S. J. Am. Chem. Soc. 2003, 125, (34), 10342-10350.

66. Haremza, J. M.; Hahn, M. A.; Krauss, T. D.; Chen, S.; Calcines, J. Nano Lett.
2002, 2, (11), 1253-1258.

67. On, Y. Y.; Huang, M. H. J. Phys. Chem. B 2006, 110, (5), 2031-2036.

68. Zanella, R.; Basiuk, E. V.; Santiago, P.; Basiuk, V. A.; Mireles, E.; Puente-Lee,
1.; Saniger, J. M. J. Phys. Chem. B 2005, 109, (34), 16290-16295.

69. Zheng, L.; Li, S.; Brody, J. P.; Burke, P. J. Langmuir 2004, 20, (20), 8612-8619.

70. Chiang, I. W.; Brinson, B. E.; Huang, A. Y.; Willis, P. A.; Bronikowski, M. J.;
Margrave, J. L.; Smalley, R. E.; Hauge, R. H. J. Phys. Chem. B 2001, 105, (35), 8297-
8301.

71. Nikolaev, P.; Bronikowski, M. J.; Bradley, R. K.; Rohmund, F.; Colbert, D. T.;
Smith, K. A.; Smalley, R. B. Chemical Physics Letters 1999, 313, (1-2), 91-97.

100

72. Mchen, P. L.; Fuhrer, M. S.; Hongkun, P. Nanotechnology, IEEE Transactions
on 2002, 1, (1), 78-85.

73. Saini, R. K.; Chiang, I. W.; Peng, H.; Smalley, R. E.; Billups, W. E.; Hauge, R.
H.; Margrave, J. L. J. Am. Chem. Soc. 2003, 125, (12), 3617-3621.

74. Tuteja, A.; Mackay, M. B.; Hawker, C. J.; Van Horn, B.; Ho, D. L. Journal of
Polymer Science Part B: Polymer Physics 2006, 44, (14), 1930-1947.

75. Asokan, S.; Krueger, K. M.; Alkhawaldeh, A.; Carreon, A. R.; Mu, Z. Z.; Colvin,
V. L.; Mantzaris, N. V.; Wong, M. S. Nanotechnology 2005, 16, (10), 2000-2011.

76. Skaff, H.; Emrick, T. Chemical Communications 2003, (1), 52—53.

77. Deegan, R. D.; Bakajin, 0.; Dupont, T. R; Huber, G.; Nagel, S. R.; Witten, T. A.
Nature 1997, 389, (6653), 827-829.

78. Deegan, R. D.; Bakajin, O.; Dupont, T. R; Huber, G.; Nagel, S. R.; Witten, T. A.
Physical Review E 2000, 62, (l), 756.

79. Hamaker, H. C. Physica 1937, 4, (10), 1058-1072.

80. Parsegian, V. A., van der Waals Forces. Cambridge University Press: New York,
2006.

81. Kirsch, V. A. Advances in Colloid and Interface Science 2003, 104, (1-3), 311-
324.

82. Rosenfeld, J. I.; Wasan, D. T. Journal of Colloid and Interface Science 1974, 47,
(1), 27-31.

83. Snow, E. S.; Campbell, P. M.; Novak, J. P. Applied Physics Letters 2002, 80,
(11), 2002-2004.

84. Kirsh, V. A. Colloid Journal 2004, 66, (4), 444-450.

85. Akamatsu, K.; Kawamura, T.; Nabika, H.; Deki, S.; Strunskus, T.; Faupel, F.
Journal of Materials Chemistry 2002, 12, ( 12), 3610-3614.

86. Teichroeb, J. H.; Forrest, J. A. Physical Review Letters 2003, 91, (1), 016104.

87. Ramos-Canut, S. Nuclear Instruments and Methods in Physics Research Section
B: Beam Interactions with Materials and Atoms 2006, 245, (1), 322-326.

88. Meli, M. V.; Lennox, R. B. Langmuir 2007, 23, (4), 1619-1622.

89. G. R. J. Artus, S. J. J. Z. H. P. G. K. M. S. S. Advanced Materials 2006, 18, (20),
2758-2762.

101

90. Cao, B.-Y.; Chen, M.; Guo, Z.-Y. Physical Review E (Statistical, Nonlinear, and
Soft Matter Physics) 2006, 74, (6), 066311-7.

91. Rios, P. F.; Dodiuk, H.; Kenig, S.; McCarthy, 8.; Dotan, A. Journal of Adhesion
Science and Technology 2006, 20, (6), 563-587.

92. Adamson, A. W.; Gast, A. P., Physical Chemistry of Surfaces. 6 ed.; John Wiley
& Sons, Inc: New York, 1997; p 358.

93. McHale, G.; Shirtcliffe, N. J.; Newton, M. I. Langmuir 2004, 20, (23), 10146-
10149.

94. Lundgren, M.; Allan, N. L.; Cosgrove, T.; George, N. Langmuir 2003, 19, (17),
7127-7129.

95. Shastry, A.; Case, M. J .; Bohringer, K. F. Langmuir 2006, 22, (14), 6161-6167.

96. Cheng, Y.-T.; Rodak, D. E.; Angelopoulos, A.; Gacek, T. Applied Physics Letters
2005, 87, (19), 194112-3.

97. Bhushan, B.; Jung, Y. C. Nanotechnology 2006, 17, (11), 2758-2772.
98. Wenzel, R. N. Ind. Eng. Chem. 1936, 28, (8), 988-994.

99. Baxter, S.; Gregg, S. J.; Razouk, R. 1.; Hirst, W.; Adam, N. K.; Crisp, D. J.;
Cassie, A. B. D.; Shuttleworth, R.; Kemball, C.; Barkas, W. W.; Nieuwenhuis, K. J.;
Schofield, R. K.; Foster, A. G.; Gregg, S. J.; Bond, R. L.; Ashpole, D. K.; Maggs, F. A.
P.; Wark, I. W.; Bangham, D. H.; Joly, M.; Hirst, W.; Griffith, M.; Dresel, E. M.;
Duncan, J. F.; Neale, S. M.; Barrer, R. M.; Razouk, R. 1.; Atkinson, D. I. W.;
Klinkenberg, A.; Pearse, J. F.; Brinkman, H. C.; Ernst, E.; Hartley, G. S.; Denbigh, K. G.;
Childs, E. C. Discussions of the Faraday Society 1948, 3, 94-&.

100. Cassie, A. B. D.; Baxter, S. Transactions of the Faraday Society 1944, 40, 0546-
0550.

101. Binni g, G.; Quate, C. F.; Gerber, C. Physical Review Letters 1986, 56, (9), 930.

102

MlCHlGAN STATE U IVERSITV LIB‘

\llllllllll \lllllllll\llllllllllll

3 1293 02956