a 3...

hawk»?

.v .
..

3:!

a}...

r
.p
£131

: x
Han;

_ 2. ‘
.L:§rk@%
I .. t W ‘

6
K

, F9045.»

.-
i

J

i V
‘I7

‘1.

(z :— .z
. .

 

 

20:7 \

This is to certify that the

thesis entitled

l

"Energy Migration and Morphology Control in Multilayersw

presented by

Shawn Marie Mehrens

has been accepted towards fulfillment

of the requirements for

Master of Science degree inflemisim

 

Date 31/ 23I/ 0/

0—7639

 

 

Major professor

MS U is an Affirmative Action/Equal Opportunity Institution

 

l
l
l

 

 

 

 

LIEI‘IARY
Michigan State ‘
Unlveraflty

a- ——_IA._‘ ....__

 

 

 

 

ve this checkout from your record.
n on or before date due.
IIer due date If requested.

PLACE IN RETURN BOX to remo
To AVOID FINES retur

MAY BE RECALLED with ear
DATE DUE DATE DUE

 

DATE DUE

h EH 2 6 'mnE
. ua

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

arm cycuac/omeotnpes-ms

 

ENERGY MIGRATION AND MORPHOLOGY CONTROL IN MULTILAYERS

By

Shawn Marie Mehrens

A THESIS

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

MASTER OF SCIENCE

Department of Chemistry

2001

“L‘_——-—

tilt

mt

pr:

had

[I

Chm

CON};

ABSTRACT
ENERGY MIGRATION AND MORPHOLOGY CONTROL IN MULTILAYERS
By
Shawn Marie Mehrens

This thesis focuses on the construction and characterization of multilayers
incorporating several different types of linking chemistry such as urea, urethane, amide,
ester and coordinate bonds. Thin films are important because of their potential use in
areas such as chemical sensing, optical information storage and nonlinear optics. By
monitoring the fluorescence lifetimes of optical donor and acceptor molecules in different
molecular layers, energy transfer between and within layers can be studied, which in turn
presents valuable structural information about the layers. Optically active chromophores
have been incorporated in covalent layered systems using amide linkages. These
multilayers have been characterized using fundamental techniques such as ellipsometry,
UV-visible spectrometry, FT IR spectroscopy and XPS. The population dynamics of the
chromophores, both in solution and in layered assemblies have been studied. The
fluorescence lifetimes of the chromophores on the surface are indicative of aggregation, a
finding consistent with that found by Home et at. We find that interlayer excitation
transport does not occur efficiently in these systems, and we believe that this is due to
more efficient donor-donor intralayer transfer within a single monolayer of chromophore.
This work has demonstrated the complexity associated with the design of multilayer
structures, and is a useful foundation for understanding interlayer excitation transport in

covalently bound layered structures.

In Memory of Grandma and Grandpa

iii

 

and
diff

Che

and
and
can'
cont

frorr

past
for 1}
Work
.VOU :
“'1 5h

3” Q‘

 

an),

“I h.

 

a
ti.

Acknowledgements

The person whom I must thank first is my thesis advisor, Gary Blanchard. From
the first moment I came to MSU he has always been encouraging, both in my research
and in my life. I want to thank him for always being there, especially through all of my
difficult decisions, and for giving me “doses of confidence” whenever the situation arose.
Cheers Gary, and thanks for all of the good times!

I would also like to thank my committee members: Greg Baker, David Weliky
and especially Merlin Bruening. Merlin: I appreciate all of your help, both as a professor
and as my second reader, and especially your jokes, such as “Head TA Mehrens”. I also
can’t forget Daniel Steffenson, my advisor at Albion College. Thank you for your
continuing support and for answering all of my “cheesy” questions about topics varying
from chemistry to life in general.

The next people who deserve recognition are my group members. To all of the
past members who have gone on since I came here: Jen, Wendy, Scott and Punit; thanks
for the fun times and all of the help! I can’t forget the current group members that I have
worked with as well: Michelle, Jaycoda, Stephen, John, Joe and Lee. I’ll never forget
you guys, we have a wealth of memories (I can’t mention the majority of them!), and I
wish you all the best of luck in your futures. Joe and Lee deserve special recognition for
all of their help with the “big laser machine,” without them I never would have gotten
any useful data! Also to Michelle: please let me know when you have the “epiphany,”
I’ll be waiting to hear about it!

I must also mention all of my friends at MSU. We went through this journey

called graduate school together. Through the good times and the bad, thanks everyone

iv

 

 

for

in c

[OI

it“:

(a
"1
(".1

CUP:

“as

 

 

for being there for me, as friends, teammates and scientific colleagues. To my other close
friends, namely Grace, Heather and Colleen: you guys are the best!

Finally, I can’t forget my husband Jared. His patience, understanding and ability
to cook have kept me sane through the past 2 1/2 years in graduate school. Without his
love and support, this thesis would not have been possible. My parents also deserve
credit. Thank you for always believing in me, and for raising a daughter who had
confidence in her abilities and wasn’t afraid to strive for her goals. I would not be the
person I am today without you guys!

Last but not least, good luck to the future Blanchard group members! May your
trip through the crazy world of graduate school be as rewarding and frustrating as mine

was!

 

L1~
L15

Ch

Ch.

Chat-I

Char

 

Table of Contents

List of Tables ..................................................................................................................... vii
List of Figures .................................................................................................................. viii
Chapter 1. Introduction ...................................................................................................... 1
1.1 Literature Cited .................................................................................................. 11
Chapter 2. Synthesis and Characterization of Multilayer Assemblies ............................. 13
2.1 Introduction ....................................................................................................... 13
2.2 Experimental ...................................................................................................... 14
2.3 Results and Discussion ...................................................................................... 22
2.4 Conclusions ....................................................................................................... 49
2.5 Literature Cited .................................................................................................. 51
Chapter 3. Spectroscopic Studies of Covalent Multilayer Assemblies ............................ 52
3.1 Introduction ....................................................................................................... 52
3.2 Experimental ...................................................................................................... 54
3.3 Results and Discussion ...................................................................................... 58
3.4 Conclusions ....................................................................................................... 80
3.5 Literature Cited .................................................................................................. 82
Chapter 4. Conclusions .................................................................................................... 84
4.1 Literature Cited .................................................................................................. 89

vi

:3 N i

 

 

Ta

Ta]

Tal

Tah

 

Table 2.1

Table 3.1

Table 3.2

Table 3.3

List of Tables

XPS peak areas, adjusted to the sensitivity factors of the instrument, for
both a primed monolayer and an acid chloride terminated bilayer
(contains both primer and acid chloride). The data show the presence of
the expected elements, and also the proper ratios between different
elements. While we still see the presence of C Is, two peaks are
observed, indicating two types of carbon with different binding energies..34

Calculated fluorescence lifetimes for one monolayer of 2,7-
diaminofluorene and one monolayer of MDA on quartz. The data were

fit to the equation y = y0 +A,e"”' + Aze'”r2 using Microcal Origin 6.0

software. The uncertainties are the 95% confidence intervals for the
measurements. ............................................................................................. 63

Measured fluorescence lifetimes of 2,7-diaminofluorene in a monolayer,
and capped with an additional monolayer of 4,4'-diaminoazobenzene.
The uncertainties are the 95% confidence intervals of the acquired data.
Note that within the experimental uncertainty, the values recovered are
identical. ...................................................................................................... 74

Lifetime measurements of a monolayer of MDA (donor, D); multilayers
of MDA (donor) and 4,4'-diaminoazobenzene (acceptor, A) with no
spacer layer; and one or two spacer layers incorporated between them.
The uncertainties are the 95% confidence interval of the measurements.
The spacer molecule used in these studies was 1,3-propanediol. ................ 75

vii

 

Fl;

Fi;

Fit

Fig

Fig:

Flgl

Figli

 

Figure 1.1

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

List of Figures

Structures and absorbance (—) and emission (—-—-) spectra (10‘5 M in
ethanol) of (a) 2,7-diaminofluorene, (b) 4,4'-diaminoazobenzene and (c)
3,3'-methylenedianiline. ................................................................................ 9

Scheme outlining the deposition of multilayers containing EDTA
molecules, which chelate with Zr4+ ions to form multilayer structures.
Note that the Y4' form of EDTA is shown for simplicity, and does not
necessarily imply that this is the form in the layers. ................................... 15

Synthetic scheme outlining the deposition of multilayers containing urea
linkages. Note that the dashed lines between N-H and 0-H groups
indicate the possibility of hydrogen bonding. This hydrogen bonding, if
it does occur, would give the multilayer structures an added degree of
lateral stability not seen in other assemblies. .............................................. 17

Synthetic scheme outlining multilayer deposition using amide linkages.
When forming ester bonds as opposed to amide bonds, 1,3-propanediol
was reacted instead of a chromophore, which is shown in this diagram. l9

Ellipsometric thickness vs. the number of layers deposited for
multilayers synthesized using EDTA as the layered material. The non-
zero y-intercept is due to the native oxide that is present on the surface
before the start of multilayer deposition. The error bars are the standard
deviations of six measurements. .................................................................. 24

Plot of ellipsometric thickness vs. the number of bilayers deposited for
multilayers constructed through the alternate deposition of 1,6-
diisocyanatohexane and 1,4-diaminobutane. The relevant linking
chemistry is a urea bond. The non-zero y-intercept is a result of a
combination of the native oxide and the primer layer. The error bars
represent the standard deviation of six different measurements .................. 26

FTIR spectra of films composed of bilayers constructed using urea
linkages. Note that the absorbance increases with an increasing number
of layers, up to three bilayers, as shown. The peaks at ~1634 cm'1 and
~1582 cm’1 are characteristic of the amide I and amide 11 bands of urea
compounds. The shoulder at ~1680 cm'l may be due to the carbonyl
stretch of another type of urea moiety in the multilayers. ........................... 28

Structures and solution phase absorbance (—) and emission (----) spectra
of the two dye molecule candidates for energy transfer studies. The top
structure is rhodamine 6G, and the bottom structure is cresyl violet.
Note that the emission band of the rhodamine 6G overlaps the absorption
band of the cresyl violet ............................................................................... 3O

viii

wes'

2n.

 

 

 

Fig

Fig

Figu

 

 

Figure 2.8

Figure 2.9

Figure 2.10

Figure 2.11

Figure 2.12

Figure 2.13

XPS spectra of surfaces that were primed with an amine-terminated
silane group (top) and one that has been primed and then further reacted
with an adipoyl chloride moiety (bottom). The insets depict the chemical
functionalities present on the surface during the measurement. .................. 33

Thickness vs. number of bilayers for multilayers incorporating 1,3-
propanediol reacted with adipoyl chloride to form ester linkages. Note
that the y-intercept is not zero, and this is due to the combination of an
initial priming layer and a native oxide. The error bars are the standard
deviation of six measurements. ................................................................... 37

Brewster angle FTIR spectra of four bilayers composed of 1,3-
propanediol and adipoyl chloride on silicon. Note the increase in
absorbance with an increasing number of bilayers. The peak at ~1725
cm'1 indicates the formation of ester bonds. ................................................ 39

Structures of molecules used in covalent multilayers that incorporate
amide and ester linkages: (a) MDA, (b) 2,7-diarninofluorene, (c) 4,4'-
diaminoazobenzene and (d) 1,3-propanediol ............................................... 41

Absorbance (a) and ellipsometric thickness (b) data for bilayers formed
by the reaction of 2,7-diaminofluorene with adipoyl chloride, which are
linked by an amide bond. Note that growth of up to 9 bilayers is shown. .43

Brewster angle FI'IR spectra of bilayers formed by the reaction of 2,7-
diaminofluorene and adipoyl chloride. Note the peaks at 1723 cm",
1536 cm'1 and 1426 cm", which are indicative of the formation of amide
bonds. As expected, the absorbance increases with the number of
bilayers, up to 9 bilayers, as shown. The spectra are offset for clarity ....... 44

Figure 2.14 Absorbance (a) and ellipsometric thickness (b) data for bilayers formed

Figure 2.15

Figure 3.1

Figure 3.2

by the reaction of 4,4'-diaminoazobenzene with adipoyl chloride. Note
the linear increases in both sets of data. In the ellipsometric data, the
average of six measurements is shown, and the error bars are the
standard deviations. ..................................................................................... 46

Absorbance data for bilayers formed by the reaction of MDA with
adipoyl chloride, forming an amide linkage. Note that the absorbance
increases linearly with the number of bilayers deposited. ........................... 48

Time correlated single photon counting (TCSPC) spectrometer used in
fluorescence lifetime experiments. .............................................................. 56

Wavelength shifts in the absorbance spectra of 10'5 M solutions (----)

and 5 bilayer (—) samples of the three chromophores used in this work:
(a) 2,7-diaminofluorene, (b) 4,4’-diaminoazobenzene, and (c) MDA. ....... 59

ix

 

Fig

Fig!

Figur

 

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Representative instrument response function (O) and fluorescence
lifetime of a 10’5 M solution of MDA in ethanol (0). The data are
plotted on a log scale for ease of presentation. Note that only a single
exponential decay is seen. ........................................................................... 62

Representative instrument response functions (O) and fluorescence
decays (O) of (a) 2,7-diaminofluorene and (b) MDA on quartz surfaces.
The data are plotted on a log plot for ease of presentation. ......................... 64

Vectors representing the different angles used to calculate K2, the

geometric factor in Forster energy transfer. The inset depicts the axial
view of the vector joining the dipole moments of the donor and acceptor
molecules. .................................................................................................... 68

Scheme depicting the kinetic model used for population decay dynamics
in the covalent monolayers used in these studies. In this scheme, A
represents an aggregate species, and M1 and M2 represent monomers in
different environments. ................................................................................ 71

Different multilayer structures constructed for the energy transfer
studies: (a) MDA + azobenzene, no spacer; (b) l spacer, (c) 2 spacers.
Note that the spacer layer used in these studies was 1,3-propanediol. ........ 76

THF!

2(

 

 

Chapter 1

Introduction

The popularity of thin films and hence the amount of research currently being
done in this area is due in part to the many potential applications of these materials. Just
a few of the possible uses of thin films include chemical sensing,l synthetic light
harvesting,2 electronic, electro-optic, and photonic devices,3’4 and nonlinear optics.5
Much research has gone into synthesizing new molecular assemblies whose growth can
be controlled in a layer-by-layer fashion. The ability to control the chemical composition
within the layers is also important. By expanding the chemical functionalities that can be
incorporated into multilayer structures, the utility of the resulting thin films increases.

The development of organized molecular assemblies on solid surfaces first began
with the introduction of Langmuir-Blodgett films. These films are composed of
amphiphilic molecules pre-assembled at an air-water interface. After their assembly,
they are transferred from this interface to a solid substrate. Langmuir-Blodgett films
were the first examples of monolayers constructed with a high degree of order, and have
been studied in areas ranging from the optical characteristics of chromophores to
chemical sensors.6 These films have proven useful for fundamental studies of molecular
layers, but their utility is limited because of thermal instability and the weak van der
Waals forces that bind the layers together.

The next development in the area of thin films was the self-assembling film,
which was more chemically stable and did not have to be pre-assembled like Langmuir-

Blodgett films. There are several types of self-assembled systems that result in organic

(\A,fi

\w
’\

 

I:

 

(Si
ani
rex
cr}'
In I
con
Cher
asse

difie

literar

monolayers. These systems include organosilicon compounds on hydroxlyated surfaces
(SiOx on Si or A1,,Oy on Al),7 alkanethiols on gold,8‘9 silver and copper,10 alcohols and

1 and carboxylic acids on aluminum oxide};13 Early studies

amines on platinum,l
revealed that certain molecules would adsorb to a suitable substrate in situ and form
crystalline-like monolayers, which could be further reacted to form a multilayer structure.
In multilayers, the film is connected to the surface, the molecules in the monolayers are
connected to each other, and the monolayers are bonded to one another by strong
chemical bonds. In other words, this type of network, which forms a multilayer
assembly, can be thought of as three dimensional, making these films significantly
different than Langmuir-Blodgett films.”

Allara and Nuzzo were the first to publish work on the most common type of self-
assembled monolayer, adsorbed alkanethiols on gold.15 In this type of system,
alkanethiols adsorb to the gold surface, forming an actual chemical bond between the
gold atoms and the sulfur head groups. Although alkanethiol monolayers have many
advantages over Langmuir—Blodgett films, and have been studied extensively in the

'6’” issues such as oxidative degradation of the thiol head-group18 and lability

literature,
of the gold-sulfur bond19 have limited their potential utility.

The most promising layered system to date has been metal phosphonate (MP)
multilayers.”21 Phosphonic acids form strong, sparingly soluble complexes with metal
ions, giving them significant advantages over other systems such as Langmuir-Blodgett
films and alkanethiols adsorbed on gold.22 The most widely implemented form of this

chemistry employs a tetravalent Zr4+ that reacts with a phosphonate group to form the

link between layers. Any type of molecule that can be functionalized to contain

\\.)

L
«N?!

 

 

phOS]
more
connc
simil:
accon
solutn
grout!
been u

surfac ‘

alterna
include
Polyele
Charget
[\I'Q Ch
”Wing
This is i
"73! bai.
a Porem

the Sam.

 

Chem“:

 

phosphonate groups can be used to build layers with this chemistry. These layers are
more stable than the other systems mentioned mainly because of extended ZP sheets that
connect adjacent layers, forming a cage-like inorganic structure. These systems are
similar to self-assembling films in their ease of synthesis; the buildup of layers is
accomplished by alternate immersion of the substrate in metal ion and bisphosphonate
solutions. Due to the fact that there are two separate steps required for monolayer
growth, the possibility of spontaneous layer growth is eliminated. ZP multilayers have
been utilized in numerous studies, including nonlinear optics,5'23 electron transfer,24 and
surface modification.25

New methods and routes for layer formation are constantly being sought. Two

alternative types of linking chemistry that are currently being studied in the literature

26-28 29,30

include covalent linking chemistry and polyelectrolyte deposition.
Polyelectrolyte films are formed by the deposition of oppositely charged polymers onto
charged substrates, with the binding force being the electrostatic attraction between the
two charged species. This type of deposition chemistry has been used to study areas
varying from electron and energy transfer” to the permeabilities of membranes to ions.32
This is a robust method for layer deposition, but questions remain about the exact species
that balance the charges between layers and the lack of order in these assemblies may be
a potential problem in future applications. Covalent linking chemistry utilizes essentially
the same chemistry as MP films, except that the interlayer linkage is a covalent bond such

as an amide or ester linkage, as opposed to an MP functionality. This type of linking

chemistry has not been studied as extensively in the literature, but due to the nature of our

.‘l

2'\ N

 

F-

 

SIU(

hUr

opn

func

ornit
Chan
SCUB:
Chror
be us
hnpor

other“

Ihatthi
Msenm
[DUlUia}
IHFErs c

(K

F)
C")

Pit):

 

dependci

HUOIESCL

IncreuSc.

 

studies, this is the type of linking chemistry that we have solicited in our work (vide
infra).

One application of multilayers that the Blanchard group has been interested in is
optical information storage. In order to begin to understand this type of storage,
fundamental studies using ZP chemistry were performed in our group. When visualizing
optical information storage, it is important to understand both the interlayer and intralayer
characteristics of the assembly. Information storage assemblies will be built using a
series of layered assemblies that contain different chromophores in each layer; each
chromophore absorbs light at a different wavelength. An explicit wavelength of light will
be used to access a specific layer, and hence a specific piece of information. It is
important to understand if any excitation transfer will occur between different layers,
otherwise multiple layers may be accessed at once. The ideal situation is to hinder any
excitation transfer from occurring between layers, thereby keeping each layer discretely
accessible.

Horne and Blanchard studied energy transfer in ZP multilayer systems, and found
that this was not the dominant means by which excited chromophores relaxed in these
systems. Several reasons were cited for the inefficient energy transfer seen in the ZP
multilayer system. The first entails intralayer energy transfer. In the system studied, the
layers consisted of fixed concentrations of a donor molecule, and the concentration of
acceptor molecule was varied. According to the Forster model, there should be a
dependence of the donor lifetime on the amount of acceptor present. In other words, the
fluorescence lifetime of the donor should decrease as the concentration of the acceptor

increases. This phenomenon was not seen in these systems due to aggregation of the

IF!

76

,4

 

 

 

chroi
Excil

fount

differ.
alkant‘
b€III 6:
transit
for 1hr
COITIPCI
beta CC
diCifi‘Ctr
the 035
aCCeprg
momem
moiecm
belII‘een

Smelded

effecll \‘e
link the i

[he Bian

chromophores into islands that were spaced at distances larger than the critical radius.33
Excitation transport between the same chromophore (DC + D2 —» D1 + D;) was also
found to compete effectively with energy transfer between different chromophores (D* +
A —+ D + A’).

34 In these studies,

Interlayer excitation transport was also studied in ZP layers.
different layers were prepared which consisted of 100% donor, 100% acceptor or 100%
alkane spacer. Layers were designed to facilitate the highest efficiency energy transfer
between donor-acceptor pairs. These studies revealed that efficient Forster energy
transfer did not occur between layers of donor and acceptor molecules. One reason cited
for this result is that the intralayer energy transfer between aggregates and monomers
competes very effectively with interlayer transfer, lessening the likelihood of transfer
between layers. One assumption of the Forster model for energy transfer is that the
dielectric response of the space between the donor and acceptor is uniform. This is not
the case for the ZP linkages. The presence of the ZP layer between the donor and
acceptor acts as a “polarizable screen” which shields the oscillating transition dipole
moment of the donor molecule from the transition dipole moment of the acceptor
molecule. The basis of the Forster model for energy transfer is dipole-dipole coupling
between the donor and acceptor molecules (vide infra). If the dipoles are somehow
shielded from one another, then efficient energy transfer will not occur.

From this work it was concluded that in order to study energy transfer
effectively, the use of another type of system was necessary. The next logical step was to

link the layers covalently, thereby eliminating the need for the polarizable ZP moiety. In

the Blanchard lab, covalent linking chemistry studied by Kohli was the starting point for

 

HF!
r

26

 

 

 

the multilayer schemes used in this work. Kohli’s work was begun using
diphenylmethane derivatives35 functionalized with amines or isocyanates, in which the
relevant multilayer linking chemistry proceeds as follows. A diisocyanate-containing
monomer is reacted with a diamine-containing monomer, and growth is controlled layer-
by—layer by exposing the substrate to only one reactive monomer species at a time. The
only time that uncontrolled growth was seen was in the presence of water. Isocyanates
are water-sensitive, and water in the reaction may produce polyureas, due to the reaction
of hydrolyzed isocyanate (an amine) with excess isocyanate. It is interesting to note that
although layer growth is spontaneous when water is present, and thus thicker layers of
material result, FI‘IR demonstrates that the same urea functionality is present as in the
controlled growth layers.35

Another type of covalent linking chemistry has also been used in our lab, which
utilizes the formation of ester and amide bonds. In this case, diamine or dialcohol
functionalized molecules are reacted with a diacid chloride to form either an arrride or an
ester linkage. This chemistry proceeds more efficiently and in less time, allowing a
larger number of layers to be built more quickly. Using this chemistry, a multitude of
different chromophores have been included in multilayer assemblies“, and other work in
our lab has utilized this chemistry in polymer assemblies, demonstrating the versatility of
this linking chemistry. Another advantage to this type of linking chemistry, as opposed
to the urea chemistry, is the controlled growth of the assembly. Although the reactions
are performed under anhydrous conditions, if some water is present, uncontrolled growth
will not occur. The only consequence is the hydrolysis of the acid chloride to a

carboxylic acid, which will impede the reaction. The carboxylic acid can be converted

back to an acid chloride by reacting it with thionyl chloride, thereby allowing the reaction
to further proceed. The layer-by-layer control seen with this scheme and the shorter
reaction times are the main reasons that we have used this chemistry in our work.

There are two main goals in this project, the first of which is to build new types of
multilayer structures, which has been done using a variety of different linking
chemistries. The second is to better understand the local environment which molecules
sense in a multilayer, and this has been accomplished by studying the fluorescence
lifetimes of the molecules of interest. Utilizing fluorescence lifetime measurements, and
studying energy transfer within a multilayer assembly, two things can be accomplished.
First, the local environment of the molecules can be probed, demonstrating this
technique’s usefulness as a characterization method. Second, we believe that energy
transfer can be utilized in other applications as well. For example, this technique can be a
useful tool for irradiating a surface at one initial excitation wavelength. With layers
constructed appropriately, excitation could propagate unidirectionally through a thin film
by means of excitation transport. Such a structure could function as a broadband optical
antenna. Another useful application is enhancing photoelectric effects in solids by

37 In this case, a molecule in

optically sensitizing them with layers of fluorophores.
contact with a solid absorbs energy and then transfers it to the solid substrate. The
transfer of energy excites the solid, producing free carriers in the material. Sensitization
is useful when a solid is not absorptive in a region of interest; by coating it with a thin
film, the efficiency of the resulting photoelectric effect can be increased.6

In this work the chromophores 2,7-diaminofluorene, 3,3'-methylenedianiline

(MDA) and 4,4'-diaminoazobenzene were studied to analyze their potential use in energy

 

 

IfdflSi

impo:
appli.
these
Oilhc
chror.‘
intere~
diamii
measu
Same '.

CODIEX

[heiog

seen in

 

hnkage
“.6 \\ Lil

Efficien.

SI”died.
ImPOHuI

InUhfldV

the fund

transfer studies, based on their optical characteristics and chemical functionalities. An
important characteristic to consider when choosing chromophores for energy transfer
applications is the absorption and emission bands of the chromophores. It is important in
these studies that the emission band of the donor molecule overlaps the absorption band
of the acceptor molecule. The solution phase absorbance and emission spectra of these
chromophores are shown in Figure 1.1. Initially the donor-acceptor pair that we were
interested in consisted of 2,7-diaminofluorene as the donor molecule, and 4,4’-
diaminoazobenzene as the acceptor molecule. After preliminary fluorescence lifetime
measurements were performed, the donor molecule was changed to MDA, keeping the
same acceptor molecule. The fluorescence lifetime studies and energy transfer in the
context of the Forster model will be discussed in greater detail in Chapter 3.

We use fluorescence lifetime measurements as a characterization method to probe
the local environments of molecules, but we also wish to better understand the results
seen in the ZP multilayer systems. By using covalent linkages as opposed to ionic
linkages, we hope to support the reasons cited previously for inefficient energy transfer.
We want to conclusively determine if other factors may have been involved in impeding
efficient interlayer energy transfer in the ZP systems.

In this thesis, the characteristics of covalently linked multilayer assemblies are
studied. The multilayer structure, optical properties and local environment are all
important issues which must be better understood before the real-life applicability of
these covalent films can be realized. In Chapter 2, the synthesis of covalently linked
multilayers and alternate multilayer chemistries is described. Included in this chapter are

the fundamental characterization methods used, including ellipsometry, FTIR, UV-visible

 

2...: 57:3... _..<._=.:..::

F 7. mt,

rL 0 .U 0 0
«.24..» x»._;.__b:: TLNC-Z: 2.:

 

0
5:...

l
3 ‘_

 

-0 0 0 0

C 57:2... fi..x:=::::

Flgu re

0f(a)

—
o
O

Y I

F’
00

0.6

(a)

 

normalized intensity (a.u)
c
A

 

 

200 250A300‘350‘4oo‘450‘500

wavelength (nm)

 

 
    

(b)

‘
’

  

\
~-------—

.°
N

’

normalized intensity (a.u.)
o
A

.0
o

 

 

J. l L

300 400 500 600

wavelength (nm)

NHz

.° 2"
OO O

.°
o~

(C)

.9
N

 

 

normalized intensity (a.u.)
c
h

9
o

 

 

250 300 350 400 450

wavelength (nm)

Figure 1.1 Structures and absorbance (—) and emission (----) spectra (10'5 M in ethanol)
of (a) 2,7-diaminofluorene, (b) 4,4'-diaminoazobenzene and (c) 3,3'-methylenedianiline.

IF

 

 

 

spectrt
grow It
Chaptc’
homog
the th
diamir
exhibi:

SCCODC

spectrometry and X-ray photoelectron spectroscopy (XPS). We find that controlled
growth of multilayer assemblies occurs using all of the linking chemistries utilized.
Chapter 3 describes the investigations of the interlayer population dynamics of
homogeneous and heterogeneous multilayer structures and also provides a discussion of
the theory of Forster energy transfer. From these studies, we find that 2,7-
diarninofluorene and 4,4'-diaminoazobenzene, the first donor-acceptor pair, do not
exhibit efficient interlayer energy transfer. A decrease in the donor lifetime is seen in the
second donor-acceptor pair, MDA and 4,4'-diaminoazobenzene, but this is not due to
energy transfer. We believe that this decrease is a consequence of the structure becoming
more ordered with the addition of the azobenzene. The incorporation of spacer layers did
not affect the donor lifetime in this system and we conclude that inefficient energy
transfer is occurring. It was discovered that although the dilute solutions of these
chromophores exhibit single exponential decays, on the surface decays which are fit to a
double exponential are recovered. The double exponential decay may be due to
aggregation of the chromophores on the surface, and this will be studied in more detail in
Chapter 3. Chapter 4 summarizes the current work and details some future work of this
project, including the study of intralayer transfer within these covalent systems, rotational
dynamics of these chromophores in order to study the freedom that the molecules have on
the surface, and solution phase studies of model compounds which are analogous to the

surface species.

10

L
I2.) I\\ .71

 

tEwwya'u '- '- .' A‘ ..

 

1.5m
2.1%

3. But

4. Tar
5. Kat

6. Cir
Se.

7. Sag-

8. Bail

9. Pon

1.1 Literature Cited

1. Sun, L.; Kepley, L.J.; Crooks, R.M. Langmuir 1992, 8, 2101.
2. Kashcak, D.M.; Mallouk, T.E. J. Am. Chem. Soc. 1996, 118, 4222.

3. Batchelder, D.N.; Evans, S.D.; Freeman,T.L.; Haussling, L.; Ringsdorf, H.; Wolf, H. J.
Am. Chem. Soc. 1994, 116, 1050.

4. Tarlov, M.J.; Burgess, D.R.F.; Gillen,G. J. Am. Chem. Soc. 1993, 115, 5305.
5. Katz, H.E.; Wilson, W.L.; Scheller, G. J. Am. Chem. Soc. 1994, 116, 6636.

6. Ulman, A. An Introduction to Ultrathin Organic Films from Langmuir-Blodgett t0
Self-Assembly; Academic: San Diego, 1991.

7. Sagiv, J. J. Am. Chem. Soc. 1980, 102, 92.

8. Barn, C.D.; Troughton, E.B.; Yao, Y.T.; Evall, J.; Whitesides, G.M.; Nuzzo, R.G. J.
Am. Chem. Soc. 1989, 111, 321.

9. Porter, M.D.; Bright, T.B.; Allara, D.L.; Chidsey, C.E.D. J. Am. Chem. Soc. 1987, 109,
3559.

10. Stewart, K.R.; Whitesides, G.M.; Godfried, H.P.; Silvera, I.F. Surf. Sci. 1986, 57,
1381.

ll. Troughton, E.B.; Bain, C.D.; Whitesides, G.M.; Nuzzo, .G.; Allara, D.L.; Porter,
M.D. Langmuir 1988, 4, 365.

12. Allara, D.L.; Nuzzo, R.G. Langmuir 1985, 1, 45.

13. Allara, D.L.; Nuzzo, R.G. Langmuir 1985, I, 52.

14. Ulman, A. Adv. Mater. 1990, 2, 573.

15. Allara, D.L.; Nuzzo, R.G. Langmuir 1985, I, 45.

16. Nuzzo, R.G.; Dubois, L.H.; Allara, D.L. J. Am. Chem. Soc. 1990, 112, 558.

17.Dubois,L.H.; Zegarski, B.R.; Nuzzo, R.G. J. Am. Chem. Soc. 1990, 112, 570.

ll

l8. Kohli, P.; Taylor, K.K.; Harris, J.J.; Blanchard, G.J. J. Am. Chem. Soc. 1998, 120,
11962.

19. Karpovich, D.S.; Blanchard, G.J. J. Chem. Ed. 1995, 72, 466.

20. Byrd, H.; Snover, J .L.; Thompson, M.E. Langmuir 1995, 11, 4449.

21. Thompson, M.E. Chem. Mater. 1994, 6, 1168.

22. Cao, G.; Hong, H.G.; Mallouk, T.E. Acc. Chem. Res. 1992, 25, 420.

23. Flory, W.C.; Mehrens, S.M.; Blanchard, G.J. J. Am. Chem. Soc. 2000, 122, 7976.
24. Kumar, C.V.; Williams, Z.J.; Turner, R.S. J. Phys. Chem. A 1998, 102, 5562.
25. Dubois, L.H.; Zegarski, B.R.; Nuzzo, R.G. J. Am. Chem. Soc. 1990, 112, 570.

26. Ryswyk, H.V.; Turtle, E.D.; Watson-Clark, R.; Tanzer, T.A.; Herman, T.K.; Chong,
P.Y.; Waller, P.J.; Taurog, A.L.; Wagner, C.E. Langmuir 1996, 12, 6143.

27. Sabapathy, R.C.; Bhattacharyya, S.; Leavy, M.C.; Cleland, W.E.; Hussey, C.L.
Langmuir 1998, 14, 124.

28. Duevel, R.V.; Corn, R.M. Anal. Chem. 1992, 64, 337.

29. Decher, G.; Hong, J.D.; Schmitt, J. Thin Solid Films 1992, 210/211, 831.
30. Decher, G.; Hong, J .D. Ber. Bunsenges. Phys. Chem. 1991, 95, 1430.

31. Kaschak, D.M.; Mallouk, T.E. J. Am. Chem. Soc. 1996, 118, 4222.

32. Harris, J.J.; Stair, J.L.; Bruening, M.L. Chem. Mater. 2000, 121, 1941.
33. Home, J.C.; Blanchard, G.J. J. Am. Chem. Soc. 1999, 121 , 4427.

34. Home, J.C.; Huang, Y.; Liu, G.-Y.; Blanchard, G.J. J. Am. Chem. Soc. 1999, 121,
4419.

35. Kohli, P.; Blanchard, G.J. Langmuir 2000, 16, 4655.
36. Major, J.S.; Blanchard, G.J. Langmuir 2001, 1 7, 1163.

37. Meier, H. J. Phys. Chem. 1965, 69, 719.

12

F!

’6

I

 

 

 

2.1 II

the cu

into 2:

transfe

efficier

to inte

molecu

Signific.
on dipo
transitio
In Order
fUnCtion.
thEOTEIIC
OPPOSQd
Another ‘

fOCUS of -

Chapter 2

Synthesis and Characterization of Multilayer Assemblies

2.1 Introduction

As stated in Chapter 1, the results recovered by Home and Blanchard motivated
the current work. Their studies involved incorporating optically active chromophores
into zirconium phosphonate (ZP) multilayer assemblies in order to study the energy
transfer within these systems. It was found that interlayer energy transfer did not proceed
efficiently, in the context of the Forster model. It is believed that this result is due mainly
to interlayer shielding of the transition dipole moments of the donor and acceptor
molecules by the ZP functionality. It is also thought that intralayer transfer played a
significant role in the lack of interlayer energy transfer. Forster energy transfer depends
on dipolar coupling between the donor and acceptor molecules; if there is shielding of the
transition moments of the two molecules, then energy transfer will not occur efficiently.
In order to study energy transfer effectively in this type of multilayer system, the ionic ZP
functionality must be removed, and energy transfer in the now covalent system should
theoretically proceed efficiently. Systems that incorporate covalent linking chemistry, as
opposed to the ionic ZP linking chemistry have been explored in the current work.
Another goal of this project has been the construction of novel types of multilayers, the
focus of which has been covalent systems, but studies have not been limited to purely
covalent systems. In order to fulfill this goal of the project, several different types of
linking chemistry have been explored, and their potential utility evaluated. All of the
systems constructed have provided useful information, but they are not necessarily

appropriate for studying energy transfer. Some of the linking chemistries studied early

13

ex A‘

\\J

 

 

onin

and t

[Jpé'
layen
comp
secon
two d
anofli
synth.
funda

theex

L)
1“)
1‘71

(805“
Phanh
Subnk
laFCr

flcnxjn
dePOSI

Immer

the” HI

on in this work were not as facile as necessary, or there were experimental limitations,
and therefore these linking chemistries were modified. There are, in effect, two main
categories of multilayer synthesis that have been studied in this work. The first entails a
type of coordination chemistry, which is similar to ZP chemistry, but uses EDTA as the
layered material, rather than a bisphosphonated molecule. In this case, the EDTA
complexes with the Zr“, forming a coordinate bond as opposed to the ZP linkage. The
second type of multilayer synthesis that we have studied includes the incorporation of
two different types of covalent linkages, one that utilizes urea and urethane linkages, and
another that utilizes the reactions that form esters and amides. In this Chapter, the
synthesis of these different types of multilayers will be discussed. In addition, the
fundamental characterization of these assemblies will also be presented, demonstrating

the extent to which growth of these multilayer systems can be controlled.

2.2 Experimental

EDTA Multilayer Synthesis. Double-side polished, undoped Si(100) substrates
(Boston Piezo-Optics) were used for multilayer synthesis. Substrates were cleaned in
piranha solution (3:1 H2SO4/H202) for approximately fifteen minutes and were then
submerged in distilled water for approximately two minutes, producing a native oxide
layer of ~15 A, measurable by ellipsometry. The substrates were then rinsed with
flowing distilled water and dried under a N2 stream. The scheme that outlines multilayer
deposition is shown in Figure 2.1. The silicon substrates were phosphorylated by
immersion in a solution that was 0.2 M in both phosphorous oxychloride (Fluka) and
2,4,6-collidine (Aldrich) in anhydrous acetonitrile for 10 minutes. The substrates were

then rinsed with reagent grade acetonitrile and distilled water. The resulting surface was

14

OH ————> 0
POCl3,collidine 0__P <00
RT, 10 min. 0
0 2f
O—P—<—O
———>
8 Zr
ZrOClz solution 0—P—<—0
RT, 1 hour 0 Zr
Zr 0 0
0 y. ft
O—P O /
______> ——é0 O N 0.
EDTA solution, pH~7.0 o z' o <N 0_
RT, 1 hour O—P—fio _\>,l / \ <
/
Zr 0 O
Zr 0 O Zr
0 _\>__\ /__</_
O—P O / \
———> 0 o N o
ZrOCl r ‘ 0 z' < 2'
SO utron
RT, 1 hour 0— P+O R N 2 EDTA solution, pH~7.0
O /> <\ RT, 1 hour
Zr 0 O Zr

N

‘
'/O

/ T\O
N

H
'/O

rwx-f

\-

% rt:

0
I
i!
O

0 O O O
o z' <: z' <:
O O O
O—P—<—O \ / \
O /> <\ /> <:
Zr 0 O Zr 0

Figure 2.1 Scheme outlining the deposition of multilayers containing EDTA molecules,
which chelate with Zr4+ ions to form multilayer structures. Note that the Y4- form of
EDTA is shown for simplicity, and does not necessarily imply that this is the form in the
layers.

15

 

4T

 

expt
hour
pha
“as

subsr
duck
Subs

foHor

flhhc
Subsfi
dhnifl

“as

dhsoc}
ofmuh
NJ SOlu
SOIUIMJr
linkage

BQQV,

exposed to Zr4+ by immersion in 5 mM ZrOCl2 (Aldrich) in 85% EtOH/15% H2O for one
hour. After removal of the substrate from solution and rinsing with distilled water, it was
placed in an aqueous solution of 0.1 M EDTA (pH ~7.0). The pH of the EDTA solution
was adjusted with 50% KOH. After reacting in the EDTA solution for one hour, the
substrate was removed and rinsed with distilled water, dried under a N2 stream and the
thickness of the resulting layer was measured by ellipsometry (Rudolph Auto-El H).
Subsequent layers were added by first zirconating the substrate, as discussed above,
followed by immersion in the EDTA solution.

Urea Multilayer Synthesis. Single-side polished, undoped Si(111) substrates
(Silicon Quest International) were cleaned in piranha solution for fifteen minutes. The
substrates were then immersed in distilled water for two rrrinutes, rinsed with flowing
distilled water, and dried under a N2 stream. Priming of the oxidized silicon substrates
was carried out by reaction in a ~1.0% v/v solution of 3-
aminopropyldimethylethoxysilane (United Chemical Technologies) in reagent grade
toluene overnight at room temperature. After priming, the substrates were rinsed with
toluene, ethanol and distilled water, and dried in a N2 stream. Priming yielded amine-
terrninated surfaces that were ready for further reaction.

The following conditions were used for bilayer formation (a bilayer consists of a
diisocyanate layer capped with a diarnine layer) and the scheme outlining the deposition
of multilayers is shown in Figure 2.2. After priming, the substrate was placed in a ~0.2
M solution of 1,6-diisocyanatohexane (Aldrich) in anhydrous DMF under Ar. The
solution was allowed to react overnight at 60 °C. This initial reaction forms a urea
linkage between the amine-terminated substrate and the isocyanate, leaving a reactive

isocyanate group at the surface. The reacted substrate was rinsed with hot DMF, and

16

 

\ i1

‘VA
1"

 

 

nflml
hnkagfis.
possibilit)

Ullllayer

 

 

 

H3C CH3
CH3 “CH3 1 NH
OCH2CH3
y H3C __CH3
in toluene i " N H 2
20° c, 24 hours o W
OH
O=C=N MN =C=O
’
60 °C, overnight
under Ar
0
H C CH
3/\Si 3 /U\
N N
H

 

 

 

>
60°C,4hours
underAr
H3C cm i 0
/Si‘ N N /u\ /\/\/NH2
—0 W1 1 i i
no; it [H
H3C ._.CH3 )L \O
i JL NH
0/3 W171 I‘ll N NW 2
H H

Figure 2.2 Synthetic scheme outlining the deposition of multilayers containing urea
linkages. Note that the dashed lines between N-H and O-H groups indicate the
possibility of hydrogen bonding. This hydrogen bonding, if it does occur, would give the
multilayer structures an added degree of lateral stability not seen in other assemblies.

17

Y
.HF'

21

 

 

 

reacted for four hours at 60 °C in a ~0.4 M solution of 1,4-diaminobutane (Aldrich) in
anhydrous DMF under Ar. After reaction, the substrate was rinsed with hot DMF,
distilled water, acetone, distilled water again and dried in a N2 stream. At this point in
the multilayer deposition scheme, when an amine-terminated surface was present,
characterization of the surface was performed. This is due to the ease of hydrolysis of the
isocyanate, which yields an amine, hindering further reaction.

Acid Chloride Multilayer Synthesis. Quartz substrates or double-side polished, p-
doped Si(111) substrates (Silicon Quest International) were cleaned by immersion in
piranha solution for fifteen minutes, rinsed with distilled water and dried under a N2
stream. Clean substrates were primed immediately for further reaction. To prime the
surface, the substrate was placed in a ~1.0% v/v solution of 3-
aminopropyldimethylethoxysilane (United Chemical Technologies) in reagent-grade
toluene overnight at room temperature. The resulting primed substrate was removed,
rinsed with toluene, ethanol and distilled water and dried under a N2 stream. The
resulting amine-terminated surface was then ready for further reaction. The scheme that
illustrates the formation of multilayers constructed using the acid chloride linking
chemistry is shown in Figure 2.3.

The amine primer layer was reacted in a solution of ~0.2 M adipoyl chloride
(Aldrich) and excess N-methylmorpholine (Aldrich) in anhydrous acetonitrile, under Ar,
at room temperature for 25 nrinutes, to form an amide linkage. After this reaction, the
acid chloride-terminated substrate was rinsed with ethyl acetate and placed under Ar once
again. Either a ~10'3 M solution of diamine functionalized chromophore or a ~0.2 M
solution of 1,3-propanediol in anhydrous acetonitrile was used for reaction, with an

excess of pyridine added to neutralize any HCl that may be formed. This reaction

18

“Pd“

 

 

Figul‘e 2

. hen f(

 

 

H3C __CH3
NHz/V\\SI'\
OCH2CH3 H3C\ '__CH3
OH > o/S‘WNHZ
in toluene
OH overnight, RT
0
ii c':'—c1
cr—c/\/\/ O
b H3C‘ ’CH3 ii ii (:1
if _C/\/\/ —
in CH3CN 0/5 W111
N-methylmorpholine H

25 minutes, RT

>

 

in CH3CN
pyridine
1 hour, RT

0
H3C CH3 0 II 0.0
‘ - NH2

Si" N ”WC—NH
O/ \/\/l—C
H

Figure 2.3 Synthetic scheme outlining multilayer deposition using amide linkages.
When forming ester bonds as opposed to amide bonds, 1,3-propanediol was reacted
instead of a chromophore, which is shown in this diagram.

19

proceeded for one hour at room temperature. Following reaction, the substrate was
rinsed with reagent grade acetonitrile, ethanol and distilled water, then dried under a N2
stream. Subsequent layers were deposited by using the same sequence that is outlined in
Figure 2.3. It is important to note that characterization of the surface was performed after
the deposition of a bilayer. This bilayer consisted of an acid chloride-terminated group
capped with either a diarnine or dialcohol functionalized molecule. This is due to the
reactivity of the acid chloride group to hydrolysis. In order to prevent hydrolysis of the
acid chloride to a carboxylic acid group, which would be unreactive in this scheme,
bilayers were deposited before characterization.

Optical Null Ellipsometry. The thicknesses of primed surfaces and monolayers
were measured using an optical null ellipsometer (Rudolph Auto-El II) operating at 632.8
nm. The thickness measurements were determined using Rudolph DAFIBM software.
For all film thicknesses measured, a value of n = 1.462 + 0i was used as the film
refractive index. It has been suggested by Ulman that a value of n = 1.50 should be used
for monolayers of alkyl chains with chains of C10 and longer, while alkyl chains of Co
and shorter should use a refractive index of 1.45.1 Tillman et al. determined that an
increase of 0.05 in the refractive increase translates to a decrease in thickness of only ~1
AZ Given the small implications of a slightly inaccurate refractive index, the same value
of 1.462 was used for both the native oxide (SiOx) layer and the organic monolayers.

Brewster Angle FI’ IR. Infrared spectra of covalent films on silicon were obtained
using a Nicolet Magna 550 FT IR spectrometer. The Brewster angle rotation stage and
polarizer were obtained from Harrick Scientific Corporation. Spectral resolution was 4

cm]. The angle of the rotation stage was set at ~75°, which is the Brewster’s angle for

20

 

 

ourspe

using ti

kidnse
ofl.00.
nhcon L
oxuhzec
SubsUan
Wafers \\
In the c.
Consider:
mawnaL
AllhOUgh
3 P0111112:
IUOde. T
COnIDbUK
rough Shh
[5P6 of M
[136111] “h
enema] It
('1

a Car}. 3hr

onkunedzn

our specific p-doped silicon wafers. The Brewster’s angle of a material can be calculated

using the following equation:3

63. tan-(121] (2.1)
"i

In this equation, n; represents the refractive index of the ambient, usually air, with a value
of 1.00, and n2 represents the refractive index of the medium itself. In the case of the
silicon used, a value of n2 = 3.732 was measured by performing ellipsometry on a clean,
oxidized silicon surface, and this number was used to calculate the Brewster’s angle.
Substrates were prepared in the manner explained above, but double-side polished silicon
wafers were used rather than single-side polished, a necessity in Brewster’s angle studies.
In the case of external reflection, there are two modes of reflection which must be
considered, RTE and Rm. In a plot of reflectance versus incidence angle for a given
material, the angle at which Rm goes to zero is known as the Brewster’s angle.
Although RTE is not zero at this point, and therefore some percentage of light is reflected,
a polarizer was used in these studies, which should eliminate any contribution from this
mode. The majority of the light is transmitted, allowing both sides of the surface to
contribute to the signal. Since both sides of the surface are contributing to the signal, one
rough side would cause a large amount of scatter, disrupting the signal. Therefore, in this
type of spectroscopy, double-side polished wafers should be used. This technique is
useful when performing infrared characterization on silicon substrates, as opposed to
external reflection FTIR, a method used commonly to characterize gold substrates.

U V-visible Absorbance Spectrophotometry. Absorption spectra were taken using
a Cary 300 Bio UV-visible absorption spectrophotometer. Solution phase spectra were

obtained at a scan rate of 600 nm/min., while quartz surface measurements were obtained

21

ata:
Zr

surfat

constar
functior
mercapt

Was ther

 

first bilu
silicon (I
in studies

)1":
Were pert
radiation j

e\-",

 

2'3 Resui

As‘
hQI’e explo
Studies. 1h

mulmmeh

anaii'sis be

 

 

at a scan rate of 300 nm/min. Surface absorption spectra were obtained on a quartz
surface, using clean, blank quartz substrates as the references.

External Reflection FT IR. In order to utilize external reflectance FT IR (Nicolet
Magna 560), layers were also formed on gold-coated silicon substrates (2000 A of Au on
200 A of Ti, Silicon Quest International). These substrates were cleaned first in either a
UV cleaner (Boekel) for 10 minutes, or in piranha solution for three minutes. The optical
constants were immediately measured after cleaning by optical null ellipsometry. To
functionalize the gold surface, the substrates were immersed in a 10'4 M solution of 11-
mercaptoundecanol in ethanol at 20 °C for 24 hours. The hydroxyl-terminated surface
was then reacted with 1,6-diisocyanatohexane to produce a urethane linkage, forming the
first bilayer. Subsequent layers were grown using the same reaction scheme as that on
silicon (Figure 2.2). It should be noted that this characterization technique was only used
in studies where bilayers were grown on gold-coated substrates.

X-Ray Photoelectron Spectroscopy. XPS measurements on multilayer samples
were performed using a Perkin Elmer (I) spectrometer with monochromatized Al Ka
radiation (1486.6 eV). All values that are reported are referenced to the C Is line at 284.6

eV.

2.3 Results and Discussion

As discussed in the Experimental section, one method of layer formation that we
have explored is the use of EDTA in a method similar to ZP linking chemistry. In these
studies, the chelating characteristics of the EDTA molecule were utilized to build
multilayers by complexation with Zr4+ ions. EDTA is most commonly used for metal ion

analysis because it chelates 1:1 with many metal ions. In our experiments, we were

22

 

 

 

hoping
proton:

demon:

that tht
formatii
was pH
plot. the
values.
multilay
number .
l'4}ch is
thickness
the pH 1
making 1
Increasir
of the tu

finder-m,

HOVEL in
PH 2130
Emu-[h u
fact Ihut

hoping to utilize EDTA in its tetradentate form, Y4K In this structure, the four acidic
protons are dissociated, allowing chelation at all four sites. The a plot of EDTA
demonstrates that the Y4" form is prevalent only above pH 10.0.4

We believed that monolayer formation would be dependent on the conformation
that the EDTA molecule adopted in solution. In an attempt to optimize monolayer
formation, a range of pHs was employed. The only pH that yielded monolayer growth
was pH 7.0; layer growth was not seen when different pHs were used. According to the a
plot, the species in solution at pH 7.0 are ~65% HY3‘ and ~35% H2Y2'. Based on these
values, it seems unlikely that more than one monolayer of EDTA would form, but
multilayers were constructed, and a linear relationship between film thickness and the
number of layers was seen for up to four layers. The plot of thickness vs. the number of
layers is shown in Figure 2.4. Beyond four layers, linearity was not seen. Layer
thicknesses were not reproducible, and growth did not always occur. Attempts to raise
the pH to ~10.0, inorder to increase the concentration of Y4' in solution resulted in
making the silicon substrates brittle. The exact reasons for this behavior are unclear. By
increasing the pH of the Zr4+ solution, we tried to lessen the differences between the pHs
of the two solutions. This caused the Zr4+ ions to precipitate out of solution as Zr(OH)4,
rendering the solution useless.

Although the idea of using EDTA and metal ions together to form multilayers was
novel, in practice it was quite difficult. The chemistry itself is too pH sensitive, and high
pH also seemed to have adverse effects on the silicon substrates. Although some layer
growth was seen, the method was not as facile and reproducible as necessary. Due to the
fact that growth was seen when the pH should not be conducive to this raises questions

about the structure of the EDTA molecules in our solutions. Obviously the behavior of

23

'I'IE-

 

 

 

'\ C X \
. i ‘ ' II'Q I

I"igur'e 2,.
S.‘Tllhesize
natIVe mm
mm bill‘s .‘

 

ellipsometric thickness (A)
U.) A Ur O\ \l 00 \O
O O O O O O O
l l ' l l l I l

N
O
1 r I

p.—
O
I

 

O
—
p
-
h
—
p
—
n
—
-

number of layers

Figure 2.4 Ellipsometric thickness vs. the number of layers deposited for multilayers
synthesized using EDTA as the layered material. The non-zero y—intercept is due to the
native oxide that is present on the surface before the start of multilayer deposition. The
error bars are the standard deviations of six measurements.

24

THI

 

the EDTA in the multilayers was more complex than we had originally believed, and the
pH problems were too great to overcome. These facts combined to force us to seek a new
route for multilayer synthesis.

At this point, we turned to the urea and urethane chemistry under development in
our group. We were interested in modifying the original chemistry, which had involved
copolymerization using diphenylmethane derivatives and we instead reacted individual
monomer species. We first explored this method using alkane chains terminated with
diarnine and diisocyanate functionalities. The importance of demonstrating the alkane
layer growth was two-fold. The first important point was determining if the chemistry
was facile using diisocyanate and diarnine terminated alkane groups. Once this proved
fruitful, we could then incorporate these alkane functionalities as spacer layers in the
energy transfer studies.

Characterization of the interfaces was performed after bilayer formation so that a
stable, amine-terminated surface was present. It was imperative in all of the multilayer
syntheses that water not be present. The step most vulnerable to hydrolysis was the
addition of the diisocyanate. As previously stated, reaction of an isocyanate group with
water yields an amine. The amines and excess isocyanate can then polymerize, resulting
in uncontrollable layer growth. To protect against this, glassware was dried at 110 °C,
then cooled in a dessicator. Furthermore, the 1,6-diisocyanatohexane and 1,4-
diaminobutane were stored under Ar.

Linear growth of the multilayers was seen for up to 8 bilayers, and the average
bilayer thickness (from the slope of the line) was ~13 A. Figure 2.5 illustrates an

example of a plot of thickness vs. the number of bilayers. As stated previously, the

25

we

21

 

100
90
80
70

I'I'Tj

   

60
50
40
30
20

ellipsometric thickness (A)

10

O l 1 J 1 l L I 1 l I I l I J

0 1 2 3 4 5 6

U I I l I I U I I I U I 1 T I

 

number of bilayers

Figure 2.5 Plot of ellipsometric thickness vs. the number of bilayers deposited for
multilayers constructed through the alternate deposition of 1,6-diisocyanatohexane and
1,4-diaminobutane. The relevant linking chemistry is a urea bond. The non-zero y-
intercept is a result of a combination of the native oxide and the primer layer. The error
bars represent the standard deviation of six different measurements.

26

 

thickness of bilayers is reported because the stability of an amine-terminated surface is
much greater than that of an isocyanate-terminated surface. The expected thicknesses are
19 A/bilayer, as determined by HyperChem calculations, and this value is greater than the
average thicknesses that we report. There are several possible reasons for this
discrepancy. First, the HyperChem calculations are for a fully extended, all trans
molecule, and no assumption is made about the tilt of the molecules with respect to the
surface normal; it has been shown in the literature that molecules in a monolayer are not
aligned with the surface normal.5 Another possible reason for the discrepancy between
the expected and the experimental values may be due to incomplete surface coverage.
The spot size of the He-Ne laser on the sample is large compared to the size of the
molecules in the monolayer. Therefore when the beam is incident on the sample, a large
number of molecules are contributing to the change in polarization of the light. The
measured thickness is essentially an average of all of the contributing molecules, so a
partially covered surface could yield a lower measured thickness because of the
combinations of the bare and covered surface regions. The same result may occur if the
molecules on the surface have a tilt angle. This tilt may make the molecular thickness
appear lower than it actually is. These reasons must be considered any time that a lower
than expected thickness is returned when using ellipsometry.

Although the ellipsometric data exhibit bilayer thicknesses that seem low, the
FTIR spectra reveal that layers are being deposited, with the expected increase in
absorbance with an increasing number of layers (Figure 2.6). In addition to
demonstrating absorbance increases, FTIR data can also yield information about the
species that are present within the assembly. These spectra illustrate that urea bonds are

present. This fact is supported by peaks at ~1634 cm'1 and ~1582 cm", which represent

27

w:

21

 

0.010

0.008 r-

l

0.00.

0.004

absorbance (a.u.)

0.002

0.000

 

 

 

1800 1700 1600 1500

frequency (cm’l)

Figure 2.6 FTIR spectra of films composed of bilayers constructed using urea linkages.
Note that the absorbance increases with an increasing number of layers, up to three
bilayers, as shown. The peaks at ~1634 cm'1 and ~1582 cm'1 are characteristic of the
amide I and amide 11 bands of urea compounds. The shoulder at ~1680 cm'l may be due
to the carbonyl stretch of another type of urea moiety in the multilayers.

28

the amide I and amide H bands characteristic of urea compounds.6 As expected, the
absorbance of these peaks increases in proportion to the number of bilayers. A broad
peak centered at ~3334 cm'l (not shown) may represent numerous types of hydrogen
bonding occurring amongst the N-H groups of the urea moieties.

After the initial studies were performed using the alkane moieties, we chose
molecules to incorporate into the layers for energy transfer studies. The obvious choice
was organic laser dyes, due to their high fluorescence quantum yields and large extinction
coefficients. Both of these factors are important when studying excitation transport in
interfacial structures. Two molecules were chosen for these studies because of the amine
functional groups that they possessed, a necessity to form a urea linkage. The two dyes
which were chosen for these studies were rhodamine 6G and cresyl violet. Besides their
amine functionalities, these molecules were also chosen because of their optical
characteristics. The structures and solution phase absorption and emission spectra of
these two molecules are shown in Figure 2.7. When studying energy transfer between
molecules, their spectral overlap is an important characteristic to consider. For efficient
excitation transport to occur, the emission spectrum of the donor molecule must overlap
the absorption spectrum of the acceptor molecule. As seen in Figure 2.7, this
requirement is fulfilled for this donor-acceptor pair.

The same conditions were used as for the alkane moieties in an effort to
incorporate these dyes into the multilayers. These experiments were performed using
both oxidized silicon (to study ellipsometric thickness increases) and quartz (to study
absorption changes) as substrate materials. The formation of multilayers containing dye

molecules did not proceed as expected. There are several possible reasons that this

29

TH

 

NHCZH,

I CH3

cooczrr,

.6 .o .o
«k 0 00
I I T

normalized intensity (a.u.)
.9
N
I

 

.9
o

 

.1

 

450 500 550 600 650

wavelength (nm)

9 .0 SD :-
4: as 00 O
l l 7 l

normalized intensity (a.u.)
.3
N
I

 

.0
o

 

1 l u 1 A 1 1
450 500 550 600 650 700

wavelength (nm)

 

Figure 2.7 Structures and solution phase absorbance (—) and emission (----) spectra of
the two dye molecule candidates for energy transfer studies. The top structure is
rhodamine 6G, and the bottom structure is cresyl violet. Note that the emission band of
the rhodarrrine 6G overlaps the absorption band of the cresyl violet.

30

TH

 

chemical reaction scheme did not work. One reason deals with the reaction mechanism
itself. In this reaction, the isocyanate undergoes nucleophilic addition with the amine
functionality. The lone pairs located on the amine nitrogen attack at the partially positive
carbon in the isocyanate molecule. After the addition at this carbon, a proton from the
now partially positive amine undergoes a proton shift and moves to the nitrogen in the
isocyanate. As stated this is a nucleophilic addition, and therefore the attempted addition
of a cationic species would hinder the reaction. Unfortunately, we were unable to locate
any dye molecules that not only fulfilled the requirement of being either anionic or
neutral in nature, but also contained an amine functionality and had appropriate optical
characteristics. Another possibility that we have considered is that the dye molecules are
too sterically hindered and bulky to be incorporated into these assemblies. This situation
does not seem likely though, given the range and size of molecules that have been used in
multilayer assemblies reported on in the literature.

Although the alkane functionalized groups produced facile results using the urea
chemistry, the incorporation of the laser dyes into the systems was not successfully
accomplished. Along with exploring other chromophores that could be used in
conjunction with the urea/urethane chemistry, we also explored other possible types of
covalent linking strategies. The idea of using amide and ester bonds to link different
layers became a possibility for building new types of layered assemblies. The main
advantage to this type of chemistry was the facile nature of bond formation. In contrast
to the ~16 hours necessary to build a bilayer using the urea chemistry, the formation of
amides and esters was much quicker, requiring only ~1‘/2 hours for bilayer formation.
One obvious advantage to this linking strategy is that more layers can be built in less

time.

31

To summarize the experimental details outlined before, the following occurs
when building a bilayer with amide or ester bonds. Figure 2.3 outlines the synthetic
scheme in greater detail. First, the substrate is functionalized to yield an amine-
terrninated surface. The amine groups are then reacted with adipoyl chloride using
anhydrous conditions under Ar, which forms an amide bond. The next step is the
reaction of the acid chloride-temrinated surface with another group, either a diarnine or
dialcohol functionalized molecule, forming either an amide or an ester. As with the urea
linking scheme, characterization is performed after the formation of a bilayer, for the
reasons stated earlier.

Before moving on to more in-depth experiments with this type of multilayer
linking chemistry, we performed some preliminary XPS measurements. We wanted to
determine if the chemistry was occurring as we believed, starting first with the primer
layer. The main technique used to characterize the primer layer was ellipsometry, which
yields no chemical information about the species on a surface. FTIR is an alternative for
getting chemical information, but good spectra of primed surfaces could not be obtained.
Therefore we looked to XPS, which could give us useful information about the ratios and
types of elements on the surface. The first surface that we performed measurements on
was an amine-terminated surface, which contained only the primer layer. As shown in
Figure 2.8, numerous peaks are seen in the XPS spectrum. As expected, peaks for
oxygen, carbon and silicon were all seen. We were concerned about the presence of
nitrogen on the surface, and the C/N ratio, if nitrogen was present. A small amount of
nitrogen was seen, which we believe is due to the presence of the primer layer. The C/N

ratio was calculated to be ~2.5:1, which is lower than the expected value of ~5:1. This is

32

THI

 

intensity (a.u.)

intensity (a.u.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7500 C 5%
0 1s ’ l—EWNH"
\
6000 Auger C
4500 1 Si 23
/ Si 2p
C 1s /
3000 N Is 1
1500 " h " M
0
l 4 i n l 1 I n l n J
1200 1000 800 600 400 200 0
BE (eV)
9000 “1,?“ N_.C.,\/\/c_ou
0 18 I —O W};
\
7500
" Si 23
6000 Auger C
i C 1s
4500 1
~ N Is Si 29
3000 M /
1500 W
0
' r 4 r . r r r P r r r
1200 1000 800 600 400 200 0
BE (eV)

Figure 2.8 XPS spectra of surfaces that were primed with an amine-terminated silane
group (top) and one that has been primed and then further reacted with an adipoyl
chloride moiety (bottom). The insets depict the chemical functionalities present on the
surface during the measurement.

33

Table 2.1 XPS peak areas, adjusted to the sensitivity factors of the instrument, for
both a primed monolayer and an acid chloride terminated bilayer (contains both primer
and acid chloride). The data show the presence of the expected elements, and also the
proper ratios between different elements. While we still see the presence of C Is, two
peaks are observed, indicating two types of carbon with different binding energies.

Carbon 1 Carbon 2 Nitrogen Oxygen SiOx Silicon

 

Primed 546.75 0 217.10 4994.91 787.52 4386.91

 

Acid chloride 1491.50 310.91 126.28 2734.92 621.5 3734.72

 

 

 

 

 

 

an unexpected result, but due to the large amount of noise compared with the small
signal, the exact experimental ratio cannot be determined conclusively. On this surface
we were mainly concerned about confirming the presence of nitrogen, and although the
ratio was not exact, the presence of nitrogen was shown, and we moved on to the next
sample, which also contained a primer layer, but in addition had been reacted with
adipoyl chloride to form a bilayer. Figure 2.7 also depicts the molecules that should be
on the surface after the priming of the substrate.

In the second sample, much more information was recovered. As expected, the
relative peak areas of both Si and SiOx decreased, indicating that the layer was thicker,
and that the electrons could not penetrate through it. As shown in Figure 2.7, there
should be two types of carbon that are distinguishable in this sample, carbonyl carbon,
and methyl/methylene carbon. The overall carbon peak is larger in the second sample,
and a shoulder is seen which was not present in the first sample. From the tabulated
results shown in Table 2.1, we can see that there are in fact two different types of carbons
present in the second sample, which is a result of a difference in binding energies. From

this information, the ratio between the two types can be calculated, and it was found to be

34

Z

 

~4.5:l. This is the expected result, given that there are 9 methyl/methylene carbons, and
2 carbonyl carbons, which yields a ratio of 4.5:]. These results were extremely good, and
demonstrated that the chemistry on the surface was occurring as we believed.

Prior to incorporating chromophores into these layers, we first needed to
demonstrate that the linking chemistry was a reasonable route to multilayers. In order to
determine this, initial studies were performed using diarnine and dialcohol functionalized
alkane chains. The molecules used in these studies included 1,3-propanediol, 1,5-
penatanediol and 1,4-diarninobutane (all obtained from Aldrich). Initially, the use of the
alkane chains presented difficulties in the experiments. During the reaction, the solutions
were heated to 60 °C, and an excess of the appropriate reactive molecule was not used.
Although growth was seen under these conditions, it was consistent for only the first two
bilayers. Due to the freedom that is inherent when using these alkane chains, we believed
that both ends of the molecules were reacting at surface sites. This process could, in
effect, cap the layers, preventing further growth. To remedy this situation, the reaction
was performed at room temperature, and the concentration of the amine or alcohol was
increased. The concentration was increased in order to produce a large excess, which we
hoped would speed up the kinetics of the reaction, allowing all of the sites to react before
the molecules had a chance to cap the layer. One other change that we made was to use
only the 1,3-propanediol in these experiments. The 1,5-pentanediol is a longer chain
molecule with more flexibility and therefore has a greater chance of capping a layer.
Although we tried using 1,4-diaminobutane to form amides, the formation of more than
two bilayers was not seen. Even with the experimental changes, 1,4-diarninobutane and

1,5-pentanediol did not exhibit reproducible bilayer formation.

35

After these changes were made, continual layer growth was seen for up to four
layers, which seemed to be the maximum number of layers that could be deposited. Due
to the lower than expected thicknesses seen, capping of the layers may still have been a
issue. To test this, we performed the following experiment. We used a solution that was
more dilute (~0.05 M compared to ~02 M), and the same reaction conditions as outlined
earlier. The first bilayer exhibited a slightly lower thickness than normal, and the second
bilayer showed an ellipsometric increase of only ~3 A, which is significantly lower than
expected. The addition of a third bilayer did not show any increase, demonstrating that
we had effectively capped the bilayer. It was not only important to demonstrate that the
chemistry was reactive, but these studies laid the groundwork for the construction of the
assemblies to be used in the energy transfer studies. In the energy transfer studies, the
distance dependence of energy transfer will be explored using non-optically active spacer
layers placed between the donor and acceptor molecules. It is important to understand
the chemistry that occurs in the spacer layers, so that accurate distances can be
approximated when separating the donor and acceptor molecules.

Based on the slope of the thickness vs. number of bilayers plots, an example of
which is shown in Figure 2.9, the average thickness of the 1,3-propanediolladipoyl
chloride bilayers was found to be ~7 A, which is significantly lower than the calculated
thickness from HyperChem. Assuming no tilt of the molecules on the surface,
HyperChem calculated a value of ~14 A. One adipoyl chloride layer represents ~8 A,
and the 1,3-propanediol has an approximate length of ~6 A. As discussed earlier, this
discrepancy could be due either to the tilt of the molecules on the surface, or to a partially

covered surface. Growth was seen consistently for up to four bilayers, yielding

36

ellipsometric thickness (A)

H H N N U) U) 4:- -h Ur
Ur O Ur O Ur O Ur O Ur O
I

O

I r I l

I 1 I I I

I

I

 

 

number of bilayers

Figure 2.9 Thickness vs. number of bilayers for multilayers incorporating 1,3-

propanediol reacted with adipoyl chloride to form ester linkages.

Note that the y-

intercept is not zero, and this is due to the combination of an initial priming layer and a
native oxide. The error bars are the standard deviation of six measurements.

37

reproducible film thicknesses of 7 Albilayer. FI‘ IR also confirms the growth of bilayers
using this synthetic scheme. Figure 2.10 demonstrates that the carbonyl peak which is
commonly seen for esters is located at ~1725 cm'lin these spectra. This peak increases
linearly with the number of bilayers, but another peak, located at 1649 cm‘1 does not
show a linear increase. We believe that this peak is another type of carbonyl, perhaps an
amide peak. This amide peak could be the result of the subsequent reaction of unreacted
amines at the surface. Another possibility is hydrolyzed acid chloride, although the
frequency seems low for this to be a possibility.

The next step was to incorporate optically active molecules into these assemblies.
It was important that these molecules contained an amine functionality, but the optical
characteristics of the molecules had to be considered as well. It must be reiterated that in
order to study energy transfer effectively there must be spectral overlap of the donor
emission and acceptor absorbance spectra. Another point to consider is how the optical
characteristics of the molecules change once adsorbed on the surface. We therefore had
to study the solution phase absorbance and emission spectra of different molecules, and
also their spectra once adsorbed on the surface.

We chose diarnine functionalized aromatic fluorophores for these studies, the
linking chemistry being the formation of amide bonds. Several molecules including 1,5-
diaminonapthalene, aminoanthracene, 3,8-diamino-6-phenylphenanthridine, 4,4'-
diamionazobenzene, 3,3'-methylenedianiline and 2,7-diaminofluorene were incorporated
into multilayer assemblies using the covalent linking chemistry outlined in Figure 2.3.
One problem we found with these molecules was the change in their absorption spectra

upon attachment to the surface. In solution several of these molecules

38

0.0020

l

0.0015

l
"
’

I
‘
a—
\
I

0.0010

absorbance (a.u)

0.0005

 

 

 

frequency (cm’l)

Figure 2.10 Brewster angle FT IR spectra of four bilayers composed of 1,3-propanediol
and adipoyl chloride on silicon. Note the increase in absorbance with an increasing
number of bilayers. The peak at ~1725 cm'1 indicates the formation of ester bonds.

39

had promising absorption and emission characteristics. Another problem that we
encountered was that many of these molecules, once adsorbed on the surface, exhibited
only a peak at ~250 nm, or a peak further to the blue region of the spectrum. Also, other
spectral features that were present in solution were not present on the surface, and
therefore the absorbances were no longer optimal for inclusion in these studies. Other
issues included irreproducible absorbance data, or solubility issues with the reaction
solvent, acetonitrile.

After these initial studies, we chose three different chromophores, which
exhibited promising optical properties, both on the surface and in solution. Although we
were interested in a donor-acceptor pair, and therefore only needed to choose two of the
three, all three chromophores were used. We then had several options available when we
began the energy transfer studies. The three molecules that we chose were 3,3'-
methylenedianiline (MDA), 2,7-diaminofluorene and 4,4'-diaminoazobenzene. The
structures of these molecules are shown in Figure 2.11. The first step, after choosing the
appropriate molecules, was to utilize fundamental characterization methods in order to
study these molecules in the covalent multilayer assemblies.

The first type of multilayer assembly that we studied incorporated 2,7-
diaminofluorene as the chromophore. In solution this molecule has an absorbance
maximum at ~275 nm, but once reacted with the surface this absorbance is red-shifted to
yield a broad peak centered around 300 nm. The solution emission yields a peak at ~4lO
nm, and due to the red-shifting of the absorbance spectrum, we expect the surface
emission spectrum to shift as well. Due to experimental difficulties, we were unable to
measure the steady state emission spectra of the surfaces (vide infra). The absorption

data for this molecule on the surface yield a linear absorbance increase for up to 9

4O

Figure 2.11 Structures of molecules used in covalent multilayers that incorporate amide
and ester linkages: (a) MDA, (b) 2,7-diaminofluorene, (c) 4,4'-diaminoazobenzene and
(d) 1,3-propanediol.

41

bilayers, demonstrating that equivalent amounts of material are being deposited during
each reaction cycle. The ellipsometric data for this molecule also yield promising results.
A linear increase in thickness was observed for up to 10 bilayers, demonstrating
controlled growth. The absorbance and ellipsometric data are shown in Figure 2.12. The
average bilayer thickness (from the slope of the line) was found to be ~17 A, very close
to the value of ~18.3 A which is predicted by HyperChem. The agreement between the
theoretical and actual thicknesses is very good. One reason for this may be due to the
rigidity of the molecule. 2,7-diaminofluorene consists of a five—membered ring
sandwiched between two benzene rings; all three rings are fused to one another. Due to
this rigidity, we would expect more ordered layers as compared to 1,3-propanediol,
which had more flexibility within the bilayers. The FTIR data also support the growth of
discrete, controlled bilayers. There are several peaks seen in the IR spectrum, all of
which support the presence of amide bonds. A representative spectrum of these bilayers
is shown in Figure 2.13. The largest peak is seen at ~1723 cm'l, and corresponds to the
carbonyl carbon of an aromatic amide. Another large peak occurs in the spectrum at
~1673 cm". Both of these peaks increase linearly with an increasing number of bilayers.
We are unsure of the nature of the peak at ~1673 cm". It may be indicative of the
formation of another type of amide such as a disubstituted amide, but the mechanism by
which this would occur is unknown. A peak in this region is present when forming both
ester and amide bonds, so it may be the result of the formation of amides from initially
unreacted primer amines, which would explain its presence in all of the spectra. Another
large peak occurs at ~1536 cm'1 and can be attributed to the C-N-H stretch-bend mode,

also referred to as the amide I] stretch. The last large peak occurs at ~1426 cm'1

42

0.20 -

0.15

(a)

absorbance (a.u.)

.9
O
Ur

 

 

0.00

 

l n 1 L l n I 1 l

200 250 300 350 400 450

wavelength (urn)

250 F

I
200 - I

I—5
LII
O
T

‘ (b)

“K

ellipsometric thickness (A)
8
\

 

0 2 4 6 8 10

number of bilayers

Figure 2.12 Absorbance (a) and ellipsometric thickness (b) data for bilayers formed by
the reaction of 2,7-diaminofluorene with adipoyl chloride, which are linked by an amide
bond. Note that growth of up to 9 bilayers is shown.

43

0.007

0.006 '
0.005 '
0.004 b
0.003 -

0.002

absorbance (a.u.)

0.001

 

0.000

 

 

 

 

1900 1800 1700 1600 1500 1400

frequency (cm’l)

Figure 2.13 Brewster angle FTIR spectra of bilayers formed by the reaction of 2,7-
diaminofluorene and adipoyl chloride. Note the peaks at 1723 cm}, 1536 cm'1 and 1426
cm'l, which are indicative of the formation of amide bonds. As expected, the absorbance
increases with the number of bilayers, up to 9 bilayers, as shown. The spectra are offset
for clarity.

and we assign this peak to ON stretching. A large peak in the 3300 cm'1 range (not
shown) is due to the N-H stretching mode.

The next molecule that we studied was 4,4’-diaminoazobenzene, which we were
also considering for the energy transfer studies. We are grateful to Dr. Punit Kohli for
synthesizing this molecule. The purity of the molecule was checked using TLC in three
different solvents—ethanol, acetone and ethyl acetate, all of which yielded a single spot.
The solution phase spectra seemed to be a good complement to the 2,7-diaminofluorene,
but this was not the case once the molecule was adsorbed on the surface. In solution, the
absorbance maximum is seen at ~400 nm, with an emission band at ~450 nm. On the
surface the maximum absorbance shifts to ~250 nm, and a small peak is seen at ~380 nm.
Once these molecules are covalently bound to the surface, their electronic structure and
hence their absorbance spectra change. We expect emission to occur from the 81 state,
which corresponds to the peak at ~380 nm, and therefore emission should occur at a
wavelength similar to the solution phase.

The same techniques used to characterize 2,7-diaminofluorene were used in
characterizing 4,4'-diarninoazobenzene. The absorbance increases linearly with the
number of bilayers deposited, indicating that controlled growth is occurring within these
assemblies. Up to six bilayers were grown and characterized using UV-visible
spectrometry. The ellipsometric data show a linear slope from a plot of thickness vs. the
number of bilayers. Ellipsometric data were obtained for up to six bilayers, yielding an
average bilayer thickness of ~13 A (Figure 2.14). This experimental thickness is much
lower compared to the predicted HyperChem value of ~21 A. The contrast between the

theoretical and actual bilayer thicknesses could possibly be due to the reasons cited

45

(107

0.05

0.03

absorbance (a.u.)

0.01

0.00

200

100

80
70

50
40
30
20
10

ellipsometric thickness (A)

Ir
\

p
E

0.02

 

 

(a)

L J l 1 I

350 400 450 500 550 600

wavelength (nm)

250

Irl'l

(b)

rffrrf‘ I

rT

 

_
-

A 1 l 1 l 1 1 1 1 1
2 3 4 5

number of bilayers

Figure 2.14 Absorbance (a) and ellipsometric thickness (b) data for bilayers formed by
the reaction of 4,4’-diaminoazobenzene with adipoyl chloride. Note the linear increases
in both sets of data. In the ellipsometric data, the average of six measurements is shown,
and the error bars are the standard deviations.

46

earlier; a partially covered surface and/or a tilt angle of the molecules on the surface.
One other possibility pertains to the flexibility of the azo bond within the molecule. The
molecule has the ability to twist or rotate about this bond, which could lead to changes in
thickness. Although the molecule should be “locked up” in the multilayer, there is still a
small possibility that it could move within a layer; rotational dynamics studies on this
chromophore in the future could help answer this question more definitively. Although
this is a lower thickness than expected, the bilayer thicknesses are reproducible between
samples. We also performed FTIR studies on these bilayers and discovered similar
results to the 2,7-diaminofluorene bilayers. The expected carbonyl peaks are seen,
indicating the presence of amides, and linear increases in absorbance were seen as well.
At this point we moved on to the characterization of the third and final chromophore for
these studies.

The last chromophore that we characterized was MDA. The absorbance spectrum
of MDA in solution yields an absorbance maximum at ~240 nm, and a small peak at
~290 nm. On the surface, as with the other candidates, the absorbance peaks shifted.
The peak at 290 nm was no longer present, except as a very small shoulder, and the
primary peaks were located at ~245 nm and ~230 nm. The solution phase emission
maximum was at ~350 nm. Due to the changes in the absorbance spectrum of MDA on
the surface, we can expect a shift in its emission spectrum as well. Absorption studies on
the surface showed that very reproducible bilayers were being deposited on the surface.
Unlike the other chromophores, we were unable to perform ellipsometric thickness
studies on this molecule, due to the unavailability of the priming solution, which occurred
late in this work. The absorbance data demonstrate that linear growth is occurring within

the assembly (Figure 2.15), and therefore the lack of ellipsometric data should not be an

47

0.09
0.08 -
0.07 .
0.06 ~
0.05 ‘
0.04
0.03 -
0.02 -

absorbance (a.u.)

m,
V
0.01 “M

;\‘l‘ A.
-'\“*§‘\~‘l\’u “A 'V/ ‘

 

0.00

_001 l I I I 1 I 1 I 1 4 I 1 I 1 J

200 220 240 260 280 300 320 340 360

wavelength (nm)

 

 

Figure 2.15 Absorbance data for bilayers formed by the reaction of MDA with adipoyl
chloride, forming an amide linkage. Note that the absorbance increases linearly with the

number of bilayers deposited.

48

issue. It is ideal to have both ellipsometric and UV-visible absorption data together to
support the growth of multilayers, but the absorbance data are sufficient evidence for the

controlled growth of this assembly.

2.4 Conclusions

One of the most stable multilayer assemblies to date are metal phosphonate
assemblies, particularly those incorporating zirconium phosphonates. In an effort to
explore different types of linking chemistry, multilayers were constructed by
coordinating EDTA with Zr4+, in a method similar to ZP chemistry. Although linear
growth was seen for up to four layers, the exact mechanism of layer formation was not
well understood. Based on the a plot of EDTA and comparison with our experimental
conditions, we believe that there are complicating issues within these multilayers. The
EDTA may have been adapting a different configuration than we had thought, but due to
the complex nature of what we saw, we moved on to different types of linking chemistry
that yielded more promising and understandable results.

The next type of multilayer scheme involved covalent linking strategies. In this
case, two types of covalent linkages were used, urea/urethanes, and amide/esters. Using
both schemes, reproducible and linear growth between samples was seen. Although we
attempted to incorporate laser dyes into the urea assemblies, due to difficulties with the
reaction mechanism this goal was not accomplished. The amide/ester linking chemistry
proved to be more facile, and therefore a larger number of layers could be built in less
time. Because of this, the amide/ester chemistry became the focus of our work. The
incorporation of three different chromophores, 2,7-diaminofluorene, 4,4'-

diaminoazobenzene and MDA into these assemblies yielded very promising results. UV-

49

visible spectrometry and ellipsometry prove that layer growth is controlled, and that
equal amounts of material are being deposited during each deposition cycle. Linear
increases in both absorbance and thickness, with an increasing number of layers, are
evidence of this fact. FI‘ IR also demonstrates the formation of amide bonds, as shown by
the carbonyl stretches associated with amides. Linear increases in IR absorbances also
indicate controlled growth. The synthesis of the multilayer assemblies incorporating the
desired chromophores has been optimized, and evidence of discrete, controlled growth
has been shown. In the next Chapter, the time-resolved and steady state optical

properties of the three chromophores both in solution and on the surface will be studied.

50

2.5 Literature Cited

1. Ulman, A. An Introduction to Ultrathin Organic Films from Langmuir-Blodgett to
Self-Assembly; Academic: San Diego, 1991.

2. Tillman, N.; Ulman, A.; Schildkraut, J.S.; Penner, T.L. J. Am. Chem. Soc. 1988, 110,
6136.

3. Pedrotti, F.L.; Pedrotti, L.S. Introduction to Optics; Prentice-Hall: Englewood Cliffs,
1987.

4. Skoog, D.A.; West, D.M.; Holler, F.J. Analytical Chemistry, An Introduction;
Saunders College: Philadelphia, 1994.

5. Home, J .C.; Blanchard, G.J. J. Am. Chem. Soc. 1998, 120, 6336.

6. Poacher, C]. The Aldrich Library for Infrared Spectra, The Aldrich Chemical Co.;
Milwaukee, WI, 1970.

51

Chapter 3

Spectroscopic Studies of Covalent Multilayer Assemblies

3.1 Introduction

In the previous Chapter, the synthesis and characterization of several different
types of multilayer structures was discussed. Characterization methods such as
ellipsometry, UV-visible spectrometry, XPS and FTIR spectroscopy were used to better
understand the synthesized layered assemblies. From these techniques, valuable
information such as ellipsometric thickness and types of bond formation was recovered.
The above techniques, while useful, are lacking in their ability to provide insight into the
local environment that individual molecules in a monolayer sense. These techniques can
reveal information about order within the layers, but they cannot tell us about aggregates
on the surface or about the efficiencies of the various relaxation processes that occur
when the molecules are excited. Currently, optical methods such as surface second
harmonic generation (SSHG)"3 and fluorescence lifetime‘t’5 measurements are being used
to probe the defect density and molecular environments within multilayer systems.

By monitoring fluorescence lifetimes, intermolecular energy transfer can be

studied. Energy transfer measurements have been especially useful in characterizing

6 10,11

roteins, aggregates of dye molecules,7'9 polyelectrolyte films, and Langmuir-
P

Blodgett films,”"14

and the theory may be applied to other systems. The experiment is
initiated when a donor molecule D is excited to its first excited singlet state, either
directly or indirectly. If an acceptor molecule A is in the proximity of D‘, then excitation
transfer from the excited donor molecule to the ground state acceptor molecule will

proceed. The efficiency of this process is determined by the distance and relative

52

orientations of D‘ and A, as well as the spectroscopic properties of the molecules, as
defined by the Forster critical radius, R0.” Fluorescence lifetime measurements are very
sensitive to the local environments of the molecules, whether in solution or in a
monolayer. This technique was used in this work in conjunction with the aforementioned
methods to better understand the way that molecules behave in a multilayer structure. In
addition, the possibility of energy transfer occurring between different donor-acceptor
pairs was also studied.

In this Chapter, the steady state and time-resolved spectroscopic behavior of three
chromophores, 2,7-diaminofluorene, 3,3'—methylenedianiline (MDA), and 4,4'-
diarninoazobenzene will be explored. These three chromophores were chosen because of
their useful optical properties both in solution and within a monolayer. We are interested
in studying the population dynamics of these molecules within multilayer assemblies in
order to determine if energy transfer can proceed efficiently between a chosen donor and
acceptor pair. To study these assemblies in the most effective manner, two different
donor-acceptor pairs were chosen; the first pair consisted of 2,7-diaminofluorene and
4,4'-diaminoazobenzene, and the second included MDA and 4,4’-diaminoazobenzene. In
solution, both of these pairs appeared to be good matches for energy transfer studies. The
results of these studies, and the theory of Forster energy transfer, the model for dipolar
energy transfer, will also be discussed in this Chapter. We find that interlayer energy
transfer is inefficient in these assemblies, which could be due to spectral shifting of the
molecules on the surface, making energy transfer less efficient, and/or aggregation of the
chromophores, causing intralayer donor-donor transfer to dominate the observed

response.

53

3.2 Experimental

Acid Chloride Multilayer Synthesis. The synthesis of covalently bound multilayer
assemblies was discussed in detail in Chapter 2 (Figure 2.3), and will be outlined briefly
here. Silica substrates, cut from sheets (Heraeus Optics) were cleaned by immersion in
piranha solution (3:1 HzSO4/HzOz) for fifteen minutes, rinsed with distilled water and
dried under a N2 stream. Clean substrates were primed immediately for further reaction.
To prime the surface, the substrate was placed in a ~1.0% v/v solution of 3-
aminopropyldimethylethoxysilane (United Chemical Technologies) in reagent-grade
toluene overnight at room temperature. The resulting primed substrate was removed,
rinsed with toluene, ethanol and distilled water, then dried under a N2 stream. The
resulting amine-terminated surface was then ready for further reaction.

The amine-terminated surface was reacted in a solution of ~0.2 M adipoyl
chloride (Aldrich) and excess N-methylmorpholine (Aldrich) in anhydrous acetonitrile at
room temperature for 25 minutes, under Ar in a roundbottom flask, to form an amide
linkage. After this reaction, the acid chloride-terminated substrate was rinsed with ethyl
acetate and placed under Ar once again. To functionalize the surface with an 'optically
active chromophore, 3 ~10'3 M solution of the diarnine functionalized chromophore in
anhydrous acetonitrile was used for reaction, with an excess of pyridine added to
neutralize any HCl that may be formed. This reaction proceeded for one hour at room
temperature. Following the reaction, the substrate was rinsed with reagent grade
acetonitrile, ethanol, distilled water and dried under a N2 stream. Optically inactive
spacer layers of 1,3-propanediol (Aldrich) were incorporated in the same manner as the

chromophore; the only difference was in the solution concentration.

54

Steady-State Optical Spectroscopy. The absorption spectra of solutions and
covalent multilayers were collected as detailed in Chapter 2. Surface absorption spectra
were obtained on a quartz surface, and blank quartz was used as the reference. In order
to collect the absorbance spectra, the surfaces were placed perpendicular to the incident
beam, and the beam was allowed to pass through the substrate, and into the detector.
Emission spectra of solutions were collected using a Jobin-Yvon Spex-3 fluorescence
spectrophotometer. The slits were adjusted based on emission intensity, but a five nm
bandpass usually provided sufficient signal for the experiments. The emission of the
solutions was collected at 90° with respect to excitation beam.

Time Correlated Single Photon Counting Spectroscopy. A schematic diagram of
the spectrometer used for the fluorescence lifetime studies of the three chromophores is
shown in Figure 3.1. A mode-locked CW Nd:YAG laser (Quantronix 416) produces ~11
W average power at 1064 nm, with 100 ps pulses at 80 MHz. An angle-tuned type I KDP
(potassium dihydrogen phosphate) crystal is used to frequency double the output to 532
nm. The 532 nm light excites a cavity-dumped, synchronously pumped dye laser
(Coherent 702-2) which is dumped at a repetition rate of 4 MHz, producing ~5 ps pulses.
The output from the dye laser is divided using a 90/ 10 beam splitter into two paths: 90%
of the light is used to excite the sample, and the other 10% is used for the reference. The
beam sent to the sample is frequency doubled (SHG) with a Type I LiIO3 (lithium iodate)
crystal. The fundamental is separated from the second harmonic using color filters, and
the excitation polarization is controlled using a polarization rotator (CV1). Emission
from the sample is Collected at 90° with respect to the incident light, and the polarization

of the emission was selected using a Glan-Taylor prism (solutions only). Monolayer

55

2(1) pulse
~ ._52:25::rearrange-.2113:r whack diagnostics

 

'- "'1." .‘I'tecflr'fih .‘ '13-;*.'. (JUL . .. 9 " .-.'o .‘. o
k '21-:{3‘359 -. ‘.'. . '3‘?' o-a -_~ ‘ I" ", my ”Iii”? 0.3.50
11" ’1‘? .‘o .45.}! n 1.13:... 1 '3‘. .291};sz .1: a It 1‘0..:.

. ‘

 

 

 

 

 

cavity dumped dye laser

‘ " ' ' fiberoptic

witchable

560 nm - 900 nm delay 7'“
SHG - .

TAC

 

. .35.?“ .

 

 

 

 

 

 

 

 

filter 1 .-

olarizatio ' ‘ j
p control "' " diode CPD delay -

scrambler ,
/ . .~I“‘i'?§§':': ratemetcr

v I 0 I I
_..-..'. 3w",-
1’ -‘,“."1'1'<".."Q,')~‘f
. - x “-1. \(1 t ~
, ,' K , 1.th . ,. L
‘, ;r1
' 1 1 e 1 1 .

 

sample

/ f I
. " G-T ’
reflective ii 8 m monochromator
objective

 

Figure 3.1 Time correlated single photon counting (TCSPC) spectrometer used in
fluorescence lifetime experiments.

56

samples were held approximately horizontally, with 5° tilts away from the horizontal in
two directions, toward the excitation beam and toward the detector. The polarization of
the light is then scrambled, and collection wavelengths are selected using a subtractive
double monochromator. The detection electronics have been described in detail
elsewhere,'6 and we briefly outline them here. The signal from the micro-channel plate
PMT (MCP-PMT, Hamamatsu) is sent to a constant fraction discriminator (CFD,
Tennelec TC454), where it is processed and sent to the time-to-arnplitude converter
(TAC, Tennelec TC864), operated in reverse mode. The TAC output counts are observed
using an oscilloscope (Tektronix 2230) and are sent to a multichannel analyzer (MCA,
Tennelec PCA-II) for collection. The reference consists of the remainder of the divided
laser beam (10%), which is sent through a fiber optic delay line, where we control the
time offset between signal and reference channels. A typical response function for the
instrument was ~40 ps. Excitation was performed at 315 nm (doubled Kiton Red,
Exciton) for both solution and surface measurements. The monochromator’s slits were
set at a 5 nm bandpass for solution measurements, and at 20 nm for monolayer samples.
For all monolayer samples, emission was collected at 450 nm (vide infra), while solution
emission was collected based on the maximum intensity from the emission curve.

Data Analysis. The lifetimes that we report here were fit to sums of exponentials
or single exponentials using Microcal Origin v. 6.0 software. For solution phase lifetime
measurements, the values reported are the averages of five measurements, with the
uncertainty reported as 95% confidence intervals. For monolayer samples, the values are

reported as the averages of between 15 and 40 measurements, and the uncertainty is

57

reported as the 95% confidence interval. The fits for all datasets were started at the point

where the instrument response function was at 5% of its maximum value.

3.3 Results and Discussion

As discussed in Chapter 2, we discovered that the absorbance spectra of these
chromophores changed significantly once bound to the surface. Figure 3.2 illustrates the
change in absorbance of the solution phase of each chromophore compared with a five
bilayer spectrum of the same molecule. This is an expected result, as the electronic
configuration of the molecule should be altered once it is covalently bound to a surface
and other molecules within a multilayer.

In these studies, we decided to use two different donor-acceptor pairs, the first
consisting of 2,7-diaminofluorene (donor) and 4,4'-diaminoazobenzene (acceptor), and
the second involving MDA as the donor, and the same acceptor molecule. In the initial
studies, we began by measuring the solution phase lifetimes of the chromophores. We
were able to measure the solution phase fluorescence lifetimes of two of the three
chromophores, due to the excitation wavelength accessible with the dye laser. The
wavelength of the incident beam was 315 nm, tunable from ~300-340 nm, and only 2,7-
diaminofluorene and MDA absorbed significantly at these wavelengths. Since these were
the two donor molecules that we were interested in using, it was most important to
determine their lifetimes. In these interface studies, we were monitoring the lifetime of
the donor, and its change in the presence of the acceptor. Knowledge of the acceptor’s
lifetime, while useful, was not central in these studies.

The steady state absorbance and emission spectra of the three chromophores in

solution were presented in Figure 1.1. From these spectra, it is apparent that the

58

F’
O\

(a)

absorbance (au.)
o
b

9’
N

 

Ԥ
§
5
.........

9’
o
, .

 

300 350 400 450

wavelength (nm)

A

9’
Ox

0))

absorbance (au.)
o
3:

gr
N

 

 

(C)

absorbance (a.u.)

 

 

 

wavelength (nm)

Figure 3.2 Wavelength shifts in the absorbance spectra of 10'5 M solutions (—--) and 5

bilayer (—) samples of the three chromophores used in this work: (a) 2,7-
diaminofluorene, (b) 4,4'-diaminoazobenzene, and (c) WA.

59

absorbance spectrum of the acceptor overlaps the emission spectrum of the donor, which
is essential for energy transfer to occur efficiently. If we assume though that the Stokes’
shift for each of the chromophores on the surface is similar to that in solution, then the
overlap between the donor and acceptor pairs may not be as great as in solution.

The ability to measure emission spectra from the quartz samples is an important
piece of information for these studies. The emission spectrum may help to identify the
presence of aggregates, and it is also revealing to measure the Stokes’ shift of surface
bound chromophores. Unfortunately, we were unable to measure emission spectra of the
surface bound chromophores on quartz substrates. Every effort was made in an attempt
to collect these emission spectra, but we were not successful. We believe that
background fluorescence from the quartz slides was a problem, and that scattered
excitation light was dominating the small signal from our monolayers. The quartz
samples were measured at varying incidence angles to the excitation source, yet there was
always a large background signal present. Up to eight bilayers of chromophore were
built on quartz slides and measured, but the signal could not be extracted from the large
background. To convince ourselves that this was not also an issue with the TCSPC
system, we attempted to get emission from a blank quartz slide. We were able to collect
only a very small signal from the blank quartz, and therefore this background signal was
not an issue in the experiments utilizing the TCSPC spectrometer.

Although we were unable to obtain steady state emission spectra, by adjusting the
monochromator of the TCSPC spectrometer, we were able to estimate the emission
maxima for the two surface bound donor molecules. Both 2,7-diaminofluorene and

MDA had emission maxima at ~450 nm on the surface. From the TCSPC spectrometer,

60

the data that we recover are not corrected for the sensitivity of the detector, nor are they
corrected for any fluctuations in source intensity. Because of this, these maxima may not
be exact, but these values should be very close to the actual ones. The large bandpass
used in these experiments may also have an effect on locating the actual maxima. We
were unable to elucidate information about the acceptor molecule, 4,4'-
diarninoazobenzene, due to the excitation wavelength used. This is not a surprising result
as we were unable to excite a 10‘5 M solution at this wavelength, and therefore a much
lower concentration of molecules would be less likely to become excited.

The solution phase lifetimes were collected using 10'5 M solutions made up in
ethanol. The solutions were excited at 315 nm, and the emission from the samples was
collected based upon the maxima from the emission spectrum. The monochromator
bandpass was 5 nm. In dilute solutions, a single exponential decay is the expected result,
and this was recovered for these solutions. Both chromophores exhibited a single
exponential decay, with relatively long lifetimes. The lifetime of 2,7-diaminofluorene
was measured as ~5.7 ns, and the lifetime for MDA was ~2.7 ns. It was very
encouraging that such long solution phase lifetimes were recovered. An example of the
instrument response function (~40 ps) and a representative plot of the solution phase
fluorescence lifetime of MDA is shown in Figure 3.3.

The longer lifetimes should allow a greater amount of time for energy transfer to
occur, on a timescale that we can easily monitor. The next step was to measure the
lifetimes of the molecules individually on the surface. To do this, samples were made up
which consisted of only one monolayer of the desired chromophore. It was important to

note that the acceptor did not absorb at this wavelength. This is an important piece

61

 

 

A
If}
E":
8 1'"
v 1 _
>5 ()0
3:
m
C
d)
H
.S
G
,9 10-
U)
E
0)
1- 000 mom 0 (D o 000

 

l 1 I 1 l 1 l 1 l

0.0 5.0 10.0 15.0 20.0

time (ns)

Figure 3.3 Representative instrument response function (O) and fluorescence lifetime of
a 10'5 M solution of MDA in ethanol (0), The data are plotted on a log scale for ease of
presentation. Note that only a single exponential decay is seen.

62

of information because emission from this molecule could interfere with the collected
emission from the donor molecule. We found that both 2,7-diaminofluorene and MDA
exhibited double exponential decays on the surface. The fluorescence lifetimes of both of
these molecules on the surface are shown in Figure 3.4. The recovery of a double
exponential decay was not surprising, and the primary reason for this change is probably
due to aggregation of the molecules on the surface (vide infra ).

The lifetimes recovered from the monolayer samples contained both a short and a
long lifetime component. The decay times recovered for both 2,7-diaminofluorene and
MDA in homogeneous monolayers are shown in Table 3.1. To summarize, the short time
component of the aminofluorene was measured as 336 i 13 ps, with a long time
component of 2189 .4.- 165 ps. The short time component of MDA was calculated as 434
i 13 ps, while the long time component was calculated as 2648 i 143 ps. Numerous
measurements were made on these monolayer samples, and once we were confident
about the lifetimes of the individual chromophores on the surface, we decided to
determine what change, if any, would occur upon the incorporation of another optically

active molecule into these assemblies. We believed that the incorporation of the acceptor

Table 3.1 Calculated fluorescence lifetimes for one monolayer of 2,7-
diaminofluorene and one monolayer of MDA on quartz. The data were fit to the equation

y = y() + A‘e’m' + Aze’”r2 using Microcal Origin 6.0 software. The uncertainties are the
95% confidence intervals for the measurements.

“'4

‘L .

 

 

 

1'1 (PS) A1 1'2 (PS) A2
diaminofluorene 336 i 13 0.81 t 0.06 2189 i 165 0.19 i 0.03
MDA 434 z 15 0.76 1 0.07 2648 :r: 143 0.24 :r: 0.03

 

 

 

 

63

  
 
 

 

 

1000 r :
”a
a
a :3 (a)
3 '8
.>_.~ 100- °°
.5 ,8
I: °o
3 s
.E “s i
t: a:
,2 o 0‘94?)
w- are
E “:38
Q ‘3 ° ‘2: o

0.0 1 0 2.0 3.0 4.0 5.0
time (ns)
1000 l-
(b)

100-

10-

emission intensity (counts)

 

 

0.0 1.0 2.0 3.0 4.0 5.0

time (ns)

Figure 3.4 Representative instrument response functions (0) and fluorescence decays
(o) of (a) 2,7-diaminofluorene and (b) MDA on quartz surfaces. The data are plotted on
a log plot for ease of presentation.

64

molecule, 4,4'-diaminoazobenzene, into these assemblies would produce a decrease in
the lifetime of the donor molecule, due to facile interlayer energy transfer.

Before moving on to the results found in the heterogeneous multilayer studies, it
is instructive to introduce the theory of Forster energy transfer because regardless of
sample morphology, dipolar excitation transport will be operative in these systems and an
understanding of the essential physics is important. Excitation transfer in the context of
the Forster model is a phenomenon that does not explicitly involve emission and
reabsorption of a photon. This process can occur whenever the emission spectrum of a
donor molecule overlaps the absorption spectrum of an acceptor molecule. It should be
noted that the acceptor does not need to be fluorescent for this transfer to occur, because
the donor lifetime is monitored, not the acceptor’s. The donor and acceptor transfer
excitation by means of a resonance between their oscillating transition dipole moments,
the intensity of which depends on the distance separating them and their relative
orientations in space. The rate constant for Forster energy transfer is defined as: ‘7

3K2R06

3.1
22'DR6 ( )

 

DA:

In Equation 3.1, “CD is the decay time constant of the donor in absence of the acceptor; R0
represents the Forster critical radius and R is the distance between the donor and
acceptor; K2, the geometric factor, will be explained in more detail later in the discussion.
The critical radius, R0, is defined as the distance when 50% of the relaxation
processes in the molecule will occur by energy transfer, and the other 50% will occur by

other radiative or non-radiative means.6

To simplify this, we will assume that all
processes which occur when R>Ro are inefficient, and when R<Ro, we will assume that

the process is efficient. As expected, when R becomes much greater than R0, the transfer

65

efficiency between the donor and acceptor decreases, and the number of molecules that

relax by energy transfer lessens. The transfer efficiency can be equated by the following:6

R6
E=———6° 6
R0+R

(3.2)
As shown in Equations 3.1 and 3.2, both the transfer efficiency and the rate of transfer
have a R'6 dependence. This dependence stems from the dipole-dipole interaction energy
which scales as R3, and therefore the probability of the donor and acceptor interacting is
R6, yielding this distance dependence for dipolar energy transfer.

The term R0 is unique for each donor-acceptor pair, and must therefore be

determined for each individual case. The critical radius for a given donor-acceptor

system can be calculated using the following equation:6

 

90001n(10)K2¢ °°
”teem

In Equation 3.3 several terms must be defined. The first is (pp, the quantum yield of the
donor in absence of the acceptor; N, Avogadro’s number; n, the refractive index of the
medium; FD, the normalized fluorescence intensity of the donor in the wavelength region
of interest; 8A, the extinction coefficient of the acceptor at the wavelength of interest, and
the geometric factor K2, which describes the relative orientation in space of the transition
dipole moments of the donor and acceptor. For solution phase studies, the value of K2 is
typically assumed to be 2/3, which is appropriate for molecules that randomize rapidly by
rotational diffusion compared to the timescale of energy transfer. The value of K2 is

calculated using the following equation:6

K2 = (sin 60 sin 6,, cos¢ — 2cos 60 cos 6A) (3.4)

66

In Equation 3.4, 01) and 0A represent the angles between the transition dipole moments of
the donor and acceptor with the vector joining them. The angle (1) represents the angle
between the donor and acceptor transition moments along the axis of the connecting
vector. A schematic diagram illustrating these vector quantities is shown in Figure 3.5.
Equation 3.3 contains an overlap integral, which describes the degree of spectral overlap
between the donor emission spectrum and the acceptor absorption spectrum. This
equation allows the Forster radius to be calculated from the spectral properties of the
donor, acceptor and donor quantum yield.

There are several factors that have an effect on the critical radius R0 and the rate
of energy transfer. The two largest are the fluorescence quantum yield of the donor and
the degree of spectral overlap between the two molecules. The quantum yield does have
a large effect on the rate of energy transfer, which seems counterintuitive, due to the
nature of the process. It was stated previously that Forster energy transfer is a non-
radiative process, which does not involve the emission and reabsorption of a photon. The
greater the fluorescence quantum yield of a molecule, the larger the number of molecules
that will relax radiatively, and not by other processes. When energy transfer occurs, the
energy is transferred from a state that would otherwise decay radiatively. The more
excited molecules, the greater the probability that dipolar energy transfer will occur. If
the quantum yield is lower, then there are fewer molecules that will remain in the state
that is resonant with the acceptor, lowering the probability of dipolar excitation transfer.

The spectral overlap also plays a role in the determination of R0. In order for
energy transfer to occur, there must be coupling between the donor and acceptor, which

requires a resonance at a common energy for the two species. The spectral overlap is a

67

(D1

 

 

 

 

 

\
\
\
\
\
\ \
Donor 9/
¢ ¢
Acceptor
Axial view

 

 

 

Figure 3.5 Vectors representing the different angles used to calculate K2, the geometric
factor in Forster energy transfer. The inset depicts the axial view of the vector joining the
dipole moments of the donor and acceptor molecules.

68

measure of this resonance condition. Although there is no direct exchange of photons
between the donor and the acceptor, their spectral characteristics are nonetheless
reflective of this resonance condition.

The functionality of our decays is not what is expected from a simple Forster
treatment. This form of the data is reflective of the morphology of the interface and we
discuss this issue below. It is pedagogically useful to examine results found previously in
our group from probing energy transfer in zirconium phosphonate (ZP) multilayer
assemblies. The results obtained from those studies are also pertinent to our results in
covalent assemblies, and conclusions about the lack of energy transfer may also be
inferred from the previous work.

Horne and Blanchard studied both intralayer and interlayer energy transfer in ZP
multilayer systems. In their work, they reported double exponential decays for their
chromophores, although in solution the same chromophores yielded single exponential
decays. It is well known that molecules aggregate into island-like formations once they
are adsorbed on the surface, and the double exponential decay was attributed to
monomers in different environments (vide infra). Horne and Blanchard did not observe
interlayer energy transfer in systems where the multilayer structure should be conducive
to it. They performed a series of experiments incorporating both spectroscopic and AFM
data, which concluded that the chromophores were aggregated into islands ~50-100 A in
diameter, with nearest neighbor spacing varying between 40-100 A.18 The apparent
absence of intralayer energy transfer between the donor and acceptor molecules was
reflective of the violation of a central assumption of the Forster model; the homogeneity

of the monolayers and that the chromophores behave as individual entities. From these

69

studies, Home and Blanchard concluded that the double exponential decay functionality
of the donor emission was due to the contributions of monomers in two different
environments. The first type of environment was a monomer in close proximity to an
aggregate island, perhaps even within an island, but somehow isolated so that it is not
participating in the aggregate structure. The second case is when the monomer is not in
contact with the aggregate, and is residing at some point away from the island structure.
An exciton hopping model was used to calculate the size and shape of the aggregate
species.18

We believe, from the presence of a double exponential decay in our data, that
aggregation occurs on the surface. In an effort to test this theory, we have used an
exciton hopping model similar to the one proposed by Home et al.18 In this model,
depicted in Figure 3.6, an aggregated island of chromophores, A, is excited by a pulse of
light. The excitation can decay radiatively, which is not a likely event, due to the dark
state of the aggregate”, or it can “hop” about the aggregate island until it reaches a
radiative monomer to which it can transfer its energy. The radiative monomer can be
either present as a defect within or on the perimeter of the aggregate (M1), or may be
present outside of the aggregate (M2). A series of coupled differential equations that

describe the temporal evolution of monomer and aggregate populations are shown as

Equations 3.5-3.9.

d[A']

 

: _kxt[A‘]([Ml]+[M2])_knr[A*I (3'5)

fll—Ml—1 = —kM,[M,‘]+k,,[A‘][M,] (3.6)

70

A+M1+M2 l» A"'+M1+M2 —> A+M1+M2

Figure 3.6 Scheme depicting the kinetic model used for population decay dynamics in
the covalent monolayers used in these studies. In this scheme, A represents an aggregate
species, and M1 and M2 represent monomers in different environments.

71

a

 

 

dug, 1: —kM,[M,‘]+ kx,[A'][M2] (3.7)

M = kM,[M,‘]—k,,[A‘][M,] (3.8)
dt

“(1:21 = kM,[M,‘]-k,,[A‘][M,] (3.9)

These equations can be solved numerically to yield the time course of the populations of

M1 and M2, as done by Home et al. Two boundary conditions must be noted for these
calculations. The first is that the concentration of Ao“ equals unity. The second condition

is that at infinite times, M1 + M2 equals unity. The first quantity of interest in these
measurements is km, which is the transfer rate of excitation from an excited aggregate to a
radiative monomer. Experimentally, this is a difficult quantity to measure because it is
seen as a build-up in the single photon counting signal after time zero. Although our

i nstrument response function is short (~40 ps), we are unable to resolve this build-up in
the experimental data. Attempts to deconvolute the experimental data from the
instrument response function were unsuccessful. Horne et al. were able to deconvolute
the data, but the time constant was too short to be resolved clearly, and an upper bound
Value of 10 ps for the excitation migration time was reported. Based on this value, we
estimate that our value of km" must be less than 10 ps. Physically these small transfer
ti mes indicate the presence of small islands. We would expect a slow build-up in signal
With the presence of large islands, which would allow the excitation to “hop” around the
i 31 and for a longer period of time before transferring its energy to a radiative monomer, at
which point emission would be seen. Although we were unable to perform the
Si Illulations, due to the fact that we could not resolve a build-up, we can place an estimate

h the srze of the aggregate islands. Assuming k,“ = 1013 s 1, consrstent wrth excrtatron

72

hopping in LB films of stilbazolium dyes"), and a conservative estimate of (1)“ = 0.10 for

aggregated MDA, with In = 2700 ps for NHDA in solution, we estimate the nonradiative

rate constant, knr = 3.3 x 109 s'1 for the MDA aggregates. Although we did not perform
the simulations ourselves, we can use the results found by Home et al. and relate them to
our measurements. This is relevant because we are using the same types of surfaces and
analogous priming chemistry, so we expect similar environments on the surface.

In order to determine approximate island sizes, the time constant per excitation
event, or hop is assumed to be 0.1 ps, so in 10 ps, which was the upper bound used by
Home et al., an average of 100 hops can occur. Song et al. have published the results of
a random walk simulation used to model the motion of an exciton within an H-

19 Using their results, the mean distance an exciton can travel in 100 hops is

aggregate.
~40 A. Other simulations and estimates of ZP lattice sizes predict that the aggregate size
is on the order of 50 x 50 A. We estimate from these calculations that our aggregate
i slands have areas that are smaller, based on the faster value of km,

The first step in the heterogeneous multilayer studies was preparing layers that
consisted of donor and acceptor, separated only by the adipoyl chloride layer. The first
donor-acceptor pair that we studied was 2,7-diaminofluorene and 4,4'-
di arninoazobenzene. The fluorescence lifetimes recovered from this experiment

demonstrated that there was not significant excitation transfer occurring between these
two molecules. As discussed in Chapter 2, the overlap in the solution phase spectra of
tl‘lese two molecules seemed quite favorable, but due to the spectral shifting associated

with covalently bonding the molecules to an interface, the efficiency of energy transfer

was substantially reduced. The lifetimes recovered for samples containing both donor

73

and acceptor adsorbed on the surface suggest that the spectral overlap is no longer
favorable for energy transfer. The recovered times were the same, to within experimental
error, as those measured for a monolayer containing only donor. The results of these
lifetime studies are summarized in Table 3.2. From these data we conclude that energy

transfer between the two molecules is not efficient.

Table 3.2 Measured fluorescence lifetimes of 2,7-diaminofluorene in a monolayer,
and capped with an additional monolayer of 4,4'-diaminoazobenzene. The uncertainties
are the 95% confidence intervals of the acquired data. Note that within the experimental
uncertainty, the values recovered are identical.

1'1 (PS) Ar 1'2 (PS) A2

 

*

aminofluorene (D) 336 i 13 0.81 i 0.06 2189 1 165 0.19 i 0.03

 

¥

D + A 350 i 10 0.80 i 0.11 2154 i 200 0.20 i 0.02

 

 

 

 

At this point, we turned to the third chromophore that we had been studying,

MDA. In these studies, we were interested in using MDA as the donor, and 4,4’-
diaminoazobenzene as the acceptor molecule. The first set of lifetimes that we collected
involved only the donor and acceptor, which were separated by a C6 diarnide (adipoyl
Chloride) layer. We recovered lifetimes that were shorter than a monolayer of MDA
alone, a result that could indicate excitation transfer. Further studies demonstrate that
excitation transport does not account for these data. As discussed earlier in this Chapter,
tChe transfer of energy has an R'6 dependence. In an attempt to determine exactly what
Was occurring within these layers, we incorporated optically inactive spacer layers
betWeen the donor and acceptor layers. If energy transfer was proceeding according to

the Forster model, then the lifetime of the donor molecule MDA should increase with

74

 

increasing donor-acceptor separation. Figure 3.7 illustrates the chemical functionalities
that were present in each of the experiments incorporating the MDA and the azobenzene

in multilayer structures.

Table 3.3 Lifetime measurements of a monolayer of MDA (donor, D); multilayers of
MDA (donor) and 4,4'-diarninoazobenzene (acceptor, A) with no spacer layer; and one or
two spacer layers incorporated between them. The uncertainties are the 95% confidence
interval of the measurements. The spacer molecule used in these studies was 1,3-

 

 

 

 

 

propanediol.
11 (p8) A1 12 (p8) A2
MDA (D) 434 i 15 0.76 x 0.07 2648 x 143 0.24 i 0.03
D +A 369 1 9 0.79 i 0.14 2415 :t 152 0.21 i 0.04
T 1 spacer 371 z 16 0.82 1- 0.08 2332 i 138 0.18 s; 0.03
T 2 spacers 372 :1_- 18 0.85 i 0.51 2285 1- 216 0.15 :1; 0.06

 

 

 

 

The results of the lifetime measurements of the donor-acceptor layers with zero,
one and two spacer layers are shown in Table 3.3. These studies demonstrate that energy
transfer is not the main reason for the decrease in the lifetime of the donor molecule upon
i ncorporation of an acceptor layer. The incorporation of optically inactive spacer layers

into these multilayer assemblies did not have an effect on the lifetime of the MDA. This

finding led us to believe that there were other phenomena occurring in the layers that
1‘ equired further exploration.

Although there were no significant changes in the absorbance spectra with respect

to the solution spectra, which would indicate aggregate formation, the donor fluorescence

1i fetime senses changes in the local environment within the layers. To test if the

Stl‘llcture-dependent variation in donor lifetime is due simply to changes in local

75

N112

3Q~©~

all

)1
Z W
-—O/\/\O—ji

(a) (C)

i?

842'

i/ijmo_i/V\j—o/V\o—E/\/\/ité~< >72=2—< >—§

ii
I Amie %
E O 0

Figure 3.7 Different multilayer structures constructed for the energy transfer studies: (a)

MDA + azobenzene, no spacer; (b) 1 spacer, (c) 2 spacers. Note that the spacer layer
used in these studies was 1,3-propanediol.

76

 

dielectric response with the addition of an overlayer, we built a multilayer that consisted
of an MDA (donor) layer, then a hydrolyzed adipoyl chloride layer, but without the
azobenzene acceptor layer. We recovered a short lifetime component for the MDA of
442 i 22 ps, and a longer time component of 3047 i 226 ps. The short lifetime is the
same within the uncertainty as the uncapped MDA monolayer (434 i 15 ps). The longer
time components are similar (but not equal) for the two samples (MDA alone = 2648 i
143 ps).

The addition of an adipoyl chloride overlayer has little or no effect on the lifetime
of the MDA molecule. We propose that the addition of the azobenzene molecule results
in a more ordered multilayer structure, and hence the decrease in the lifetime of the
donor. The azobenzene molecule is planar, and will attempt to remain planar in a
multilayer structure, allowing for substantial intermolecular interactions between
acceptors. An MDA monolayer, and a monolayer with an overlayer of adipoyl chloride
yield the same lifetime. Once the azobenzene molecule is incorporated into the
assembly, the lifetime of the MDA becomes shorter. The addition of optically inactive
spacer layers (1,3—propanediol) results in the same shorter lifetime. Each assembly that
yields a shorter lifetime is capped with an azobenzene, which proves that the azobenzene
must be the reason for the change in the lifetime of the donor. We rule out excitation

transport based on the donor-acceptor distance independence of the lifetimes.

We propose that the ordering of the assembly occurs in the following manner.

The first monolayer of MDA is organized to some extent on the surface. The addition of
the azobenzene forces the entire assembly into a more ordered arrangement. The new

Order within the multilayer assembly alters the structure of the aggregates in this

77

 

environment. The radiative monomers are now in closer proximity to a medium with a

20’“ that the

larger dielectric constant than before. It has been shown in the literature
result of decreasing the distance between a dipole and a dielectric is to shorten the
lifetime of the dipole, in this case, a radiative monomer.

From these studies, it does not appear that interlayer energy transfer plays a major
role in the relaxation dynamics of the chromophores that we have studied in this work.
We saw no change in the donor lifetime of the 2,7-diaminofluorene/azobenzene pair,
even though energy transfer should be favorable in this system. Although we see a
decrease in MDA lifetime with the second donor-acceptor pair, this is most likely due to
ordering within the layers initiated by the planar azobenzene molecule.

Horne and Blanchard also probed the ZP multilayers for the possibility of
interlayer energy transfer.22 In this work, they calculated that, based on the orientation of
the chromophores within a layer, that interlayer energy transfer should be twice as
efficient as intralayer energy transfer. They found no dependence of the donor lifetime
on the presence of acceptor, indicating that efficient interlayer energy transfer was not
occurring. They thought that the donor lifetime in those studies was dominated by same
molecule (D1‘ + D2 —2 D1 + D2*) energy transfer. Another underlying assumption of the
Forster model is that the dielectric medium is spatially uniform between the donor and
acceptor. Due to the ZP functionality present between the two molecules, this
requirement is not necessarily fulfilled. They concluded that interlayer energy transfer

was mediated by the presence of the polarizable ZP moieties, which shielded the donor

and acceptor transition dipole moments.

78

 

 

 

There are no polarizable linkages between layers in our covalent assemblies, so

the exact reason for the lack of energy transfer is unsure. Work in the literature involving
assemblies such as LB7‘9 and polyelectrolytem’” films demonstrate that energy transfer
can proceed efficiently in these multilayer assemblies. Evans et al. studied energy
transfer from excited aggregates to monomers in LB films, and found it to be very
efficient.23 In these experiments they showed that in a system with primarily aggregates,
(<0.5% monomer) the monomer dominated the emission spectrum.23 This fact was true
even when the system was excited at the absorbance maximum of the aggregate, and not
the monomer. They concluded that the aggregate is a dark state, and only relaxes non-
radiatively, and any emission must be from a transfer of energy to the monomer. Under
this assumption, the lifetime that we see must be due only to the monomer, and not to
both the monomer and the aggregate. This also supports Horne and Blanchard’s
conclusion that the double exponential decay seen on the surface in their studies was due
to the monomer in two different environments.

One other point of interest is the effect of the primer layer on the resulting
assembly. Horne and Blanchard also looked at the effect of priming on the dynamics of
their system. The normal priming method used in our studies was 3-
aminopropyldimethylethoxysilane (APDMES), which is known to be less stable than
another type of primer used by Home et al., 3-aminopropyltriethoxysilane (APTES).
ARIES is known to polymerize“, producing a cross-linked structure due to three reactive
ethoxy groups, compared to the one ethoxy group present in APDMES. In AFM studies,
they demonstrated the differences in the surface topology with the two different

primers.22 In contrast to the results found by Home et al., when we primed a substrate

79

with APTES, and then added a monolayer of MDA to this surface, we recovered a
lifetime which was longer than MDA on a surface primed with APDMES. The short and
long time components of MDA on this APTES primed surface were 544 i 18 ps and
27 87 1 410 ps, respectively. This difference in lifetime can be attributed to the different
types of bonding present in each system. In ZP systems, the bonding is primarily ionic,
While in our systems the bonding is covalent. The polymerized primer is also a covalent
network, and will therefore contribute to the covalent bonding that occurs in subsequent
layers. This change in the lifetime of MDA demonstrates very effectively the sensitivity

of this technique, and why the information gained is so valuable.

3 .4 Conclusions

In this work we constructed multilayer assemblies using optically active
chromophores. The steady state and time-resolved dynamics of these chromophores both
in solution and within multilayers were studied. The first donor-acceptor pair measured
consisted of 2,7-diaminofluorene and 4,4'-diaminoazobenzene. The fluorescence
lifetime of the donor in this pair, 2,7-diaminofluorene, was studied both in a
homogeneous monolayer, and in multilayers that also contained azobenzene. No change
in the donor lifetime was recovered in the presence of the acceptor molecule. Another
donor-acceptor pair, MDA and 4,4'-diaminoazobenzene was then considered. Once
again the fluorescence lifetime of the donor, in this case MDA, was studied both in a
homogeneous monolayer and in the presence of the acceptor. Unlike the first donor-
acceptor pair, in this instance, a decrease in the donor lifetime was seen, which could be
indicative of energy transfer within the assembly. Studies were performed which

incorporated both one and two spacer layers into this assembly, and no change was seen

80

between these results and the donor and acceptor directly next to one another. We
therefore concluded that energy transfer was not occurring as previously thought, and that
the presence of the azobenzene was ordering the assembly, thereby decreasing the
lifetime of the donor molecule. Based on the double exponential decay that was
recovered for the lifetime of both donor molecules we believe that the chromophores are
aggregating on the surface and perhaps impeding interlayer energy transfer. An upper
bound for the size of an aggregate island was estimated based on an exciton hopping
model. From this model, we found that while we could not determine the exact island
sizes, we could place an upper bound of 50 x 50 A on their area. Although interlayer
energy transfer should compete effectively with intralayer transfer between aggregates
and monomers, this was not seen in these assemblies. Further studies which probe the
intralayer energy transfer within these assemblies will give a clearer picture of the exact

behavior in these systems.

81

3.5 Literature Cited

1. Flory, W.C.; Mehrens, S.M.; Blanchard, G.J. J. Am. Chem. Soc. 2000, 122, 7976.
2. Lvov, Y.; Yamada, S.; Kunitake, T. Thin Solid Films 1997, 300, 107.

3. Heflin, J.R.; Figura, C.; Marciu, D.; Liu, Y.; Claus, R.O. Appl. Phys. Lett. 1999, 74,
495. r

4. Matsui, J .; Mitsuishi, M.; Miyashita, T. Macromolecules 1999, 32, 381. i
5. Mabuchi, M.; Ito, S.; Yamamoto, M.; Miyamoto, T. Macromolecules 1998, 31, 8802. i

6. Lakowicz, J .R. Principles of Fluorescence Spectroscopy; Kluwer/Plenum: New York,
1999; Ch. 13.

 

7. del Monte, F.; Mackenzie, J.D.; Levy, D. Langmuir 2000, I6, 7377.

8. Kano, K; Kazuya, F.; Wakami, H.; Nishiyabu, R.; Pastemack, R.F. J. Am. Chem. Soc.
2000, 122, 7494.

9. Wang, M.; Silva, G.L.; Armitage, B.A. J. Am. Chem. Soc. 2000, 122, 9977.

10. Kaschak, D.M.; Mallouk, TE. J. Am. Chem. Soc. 1996, 118, 4222.

11. Baur, le.; Rubner, M.F.; Reynolds, J.R.; Kim, s. Langmuir 1999, 15, 6460.

12. Hisada, K.; Ito, S.; Yamamoto, M. Langmuir 1196, 12, 3682.

13. Mabuchi, M.; Ito, S.; Yamamoto, M.; Miyamoto, T. Macromolecules 1998, 31, 8802.
14. Matsui, J .; Mitsuishi, M.; Miyashita, T. Macromolecules 1999, 32, 381.

15. Forster, T. Disc. Faraday Soc. 1959, 27, 7.

16. DeWitt, L.; Blanchard, G.J.; LeGoff, E.; Benz, M.E.; Liao, J.H.; Kanatzidis, M.G. J.
Am. Chem. Soc. 1993, 115, 12158.

17. Fleming, G.R.; Chemical Applications of Ultrafast Spectroscopy; Oxford University:
New York, 1986; Ch. 6.

18. Home, J.C.; Blanchard, G.J. J. Am. Chem. Soc. 1999, 121, 4419.

82

19. Song, Q.; Bohn, P.W; Blanchard, G.J. J. Phys. Chem. B 1997, 101, 8865.

20. Drexhage, K.H. J. Lumin. 1970, 1,2 693.

21. Girard, C.; Martin, O.J.F.; Dereux, A. Phys. Rev. Lett. 1995, 75, 3098.
22. Home, J.C.; Blanchard, G.J. J. Am. Chem. Soc. 1999, 121, 4427.

23. Evans, C.E.; Bohn, P.W. J. Am. Chem. Soc. 1993, 115, 3306.

24. Haller, I. J. Am. Chem. Soc. 1978, 100, 8050.

83

 

Chapter 4

Conclusions

This thesis has focused on the synthesis and characterization of novel multilayer
systems. Multilayer assemblies have been built using both coordinate chemistry through
the chelation of EDTA with Zr4+ ions, and covalent linking chemistry, which utilized
urea, urethane, ester and amide bonds. The covalent linkage studies were of importance
because of the possibility of probing energy transfer within these layered systems. Home
and Blanchard studied energy transfer in ZP multilayer systems and found that interlayer
energy transfer was not the dominant means of relaxation in these systems. This result
motivated us to replace the ZP functionality with something less polarizable, and focus
on the resulting covalent assemblies, in the hope that energy transfer could be mediated
structurally in these systems.

We have shown that the synthesis of multilayers containing urea, urethane, amide
and ester linkages proceeds in a controlled manner. Fundamental characterization of
these systems using techniques such as ellipsometry, FTIR, UV-visible spectrometry and
XPS was performed. All of these techniques show the expected trend of a linear increase
in response with an increasing number of layers. The incorporation of optically active
chromophores into the covalent assemblies utilizing amide bonds was also accomplished.
The growth of bilayers containing functionalized chromophores was also shown to
proceed in a controlled manner.

Using these covalent multilayer systems we hoped to better understand the role of

energy transfer in the relaxation dynamics of these molecules. It was discovered that a

84

~_ .-. .pl‘U-KT41

 

til!

 

donor-acceptor pair containing the molecules 2,7-diaminofluorene and 4,4'-
diaminoazobenzene exhibited no decrease in the lifetime of the donor molecule,
indicating a lack of energy transfer. Another donor-acceptor pair was also studied,
incorporating the same acceptor molecule, but using a different donor, MDA. In this
system a decrease was seen in the lifetime of the donor in the presence of the acceptor.
Subsequent experiments that separated the donor and acceptor with optically inactive
layers showed no change in the lifetime of the donor with increasing donor-acceptor
separation, and we concluded that the decrease in lifetime was not due to Forster energy
transfer.

One possibility for this decrease in lifetime is that the azobenzene molecule is
forcing order upon the system upon its addition to the multilayer. This explanation seems
reasonable since a more ordered assembly should exhibit a shorter lifetime due to
structural mediation of the local dielectric response. There are several reasons that
interlayer energy transfer is not occurring efficiently, even in situations where it should
be favorable. The first reason stems from aggregation of the chromophores on the
surface. Evans et al. studied the aggregation of chromophores in LB films, and found
that Forster energy transfer from the aggregate to the monomer was the prevalent means

l Home and Blanchard also reported on the

of relaxation for these chromophores.
significance of aggregates in their work.2 In their studies of both intralayer and interlayer
energy transfer, they discovered that the chromophores were aggregated into islands that
were separated by distances greater than the Forster critical radius. As a result, even

though it should have been efficient, no transfer of energy was seen between different

chromophores in these assemblies. In this instance the result was believed to be due to

85

 

 

Tmrrzslu- YM“

the ZP moiety separating the donor and acceptor molecules. In our work, as shown in
Chapter 3, we studied the dynamics of a multilayer assembly that did not contain this
polarizable ZP functionality. We discovered that interlayer energy transfer was not an
efficient means of relaxation in these covalent systems. This result is surely because of
chromophore aggregation on the surface.

Chromophore aggregation is seen in numerous multilayer assemblies}5 in the
literature, and there does not appear to be any simple way to eliminate this problem. The
priming step is an area of great concern. AFM studies performed by Home et al.

demonstrated that while the addition of a primer to the surface fills grooves or

 

depressions in the surface, the silanes react in areas with dense active silanol sites,
forming island-like structures.6 These island-like structures inhibit the formation of a
uniformly smooth surface. As shown in Chapter 3, different primers had an influence on
the resulting donor lifetime. The triethoxy silane primer is known to polymerize on the
surface, and the initial layer is therefore harder to control.7 The dimethylethoxy silane
primer may be easier to control, but the resulting structure is less stable due to the lack of
cross-linking, which is seen when using the triethoxy silane primer. The difference in
donor lifetimes for monolayers made with the two primers demonstrates that the priming
step can have a profound effect on the structure of the resulting layers.
It will be very important in future work to devise a priming method that will result
in a more uniform initial layer. There is current work in our laboratory that will make use
Of a polymer templating layer to remove the functional group density and distribution of

the surface. A dramatic change should be seen in the surface morphology of the resulting

86

layers if the initial layer is more uniform. Excitation transport measurements will be
critical in the evaluation of this “remapping” chemistry.

The issue of aggregation must be studied in greater detail in these assemblies. It
would be helpful to utilize AFM surface images and combine them with results from
intralayer energy transfer studies. In our lab we are in the process of devising a second
harmonic generation (SHG) imaging system, which could provide a wealth of
information about the groups on the surface of a substrate. This combination could
provide more useful information about the exact aggregate island sizes and spacing
between aggregates. This SHG imaging system may also prove to be another technique
that could be used to gather topological information. The only drawback to this
technique may be the resolution, which will be on the micron scale. This may not be a
small enough scale for us to look at island formation. However the studies are
performed, the intralayer population dynamics of these assemblies must be studied in
greater detail.

In Chapter 2, the rotational freedom of the acceptor molecule, 4,4'-
diaminoazobenzene was questioned. Although the energy barrier for the isomerization of
azobenzenes is estimated as ~50 kcal/mol, it would be interesting to perform
isomerization and anisotropy studies on these covalent systems. These rotational
dynamics will be pertinent to the optical information storage work currently underway in
our laboratory. The read/write function of chromophores in multilayers will be mediated
by a cis-trans isomerization. It will also be very important to determine if the molecules

have motional freedom while in a multilayer structure. These results may also help

87

 

l
17’

explain the lower thicknesses that were seen in the ellipsometric studies of the
azobenzene molecule.

It would be interesting to perform solution phase studies on analogs of the surface
bound species. This would be a useful set of studies for several reasons. First, it would
be helpful to note if the shifts in the absorbance spectra of the chromophores are due
primarily to changes in their electronic structures, aggregation of the molecules, binding
them to a surface or combinations of these factors. By forming amide bonds in solution,
the chromophore’s characteristics in this environment can be measured and shifts in the
absorbance and emission spectra can be noted. The second reason that solution phase
studies could be important is for energy transfer studies in solution. These solution phase
energy transfer studies would provide a controlled reference experiment with which to
compare the surface studies, and may or may not demonstrate efficient energy transfer.
Whatever the results of these studies, they would provide a wealth of information that is
not currently accessible.

There is still much work to be done in the area of covalent multilayers. We have
shown that the deposition of layers occurs in a seemingly controlled manner. The use of
a more sensitive technique, fluorescence lifetimes, has demonstrated that this may not be
the case. There is obviously aggregation of the chromophores on the surface and we need
to understand this mechanism better. As long as aggregation occurs, the potential of
covalent systems is very limited. The incorporation of different chromophores may
alleviate the problem, but this is doubtful considering the number of studies that have

been performed, which demonstrate aggregate formation.

88

 

 

4.1 Literature Cited

1. Evans, C.E.; Bohn, P.W. J. Am. Chem. Soc. 1993, 115, 3306.
2. Home, J .C.; Blanchard, G.J. J. Am. Chem. Soc. 1999, 121, 4427.

3. Kano, K; Kazuya, F.; Wakami, H.; Nishiyabu, R.; Pastemack, R.F. J. Am. Chem. Soc.
2000, 122, 7494.

4. Wang, M.; Silva, G.L.; Armitage, B.A. J. Am. Chem. Soc. 2000, 122, 9977.

 

5. Kaschak, D.M.; Mallouk, T.E. J. Am. Chem. Soc. 1996, 118, 4222. ,j
6. Home, J.C.; Blanchard, G.J. J. Am. Chem. Soc. 1999, 121, 4419.

7. Haller, I. J. Am. Chem. Soc. 1978, 100, 8050. I;

89