LIBRARY

Michigan State
University

 
   

This is to certify that the

thesis entitled

Spectrophotometric, Electrochemical, Photoelectrochemical,
and Surface Analysis Studies of Copper
Phthalocyanine/Metal Oxide Electrodes

presented by

Vance Roger Shepard, Jr.

has been accepted towards fulfillment
V of the requirements for

M . S . degree in Chemi $12er

   

, k

“s
- “Ru“!!flfi
Major professor ‘

Date \\\\\‘\'\
\ \

0-7 639

SPECTROPHOTOMETRIC, ELECTROCHEMICAL, PHOTOE ECTROCHEMICAL,
AND SURFACE ANALYSIS STUDIES OF COPPER

PHTWALOCYANINE/METAL OXIDE ELECTRODES

8V

Vance Rogers Shepard, Jr.

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF SCIENCE

Department of Chemistry

1977

ABSTRACT

SPECTROPHOTCMETRIC, ELECTROCHEMICAL, PHOTOELECTROCHEMICAL,
AND SURFACE ANALYSIS STUDIES OF COPPER
PHTHALOCYANINE/METAL OXIDE ELECTRODES

By

Vance Rogers Shepard,Jr.

There has been a growing interest in the study of
chemically-modified electrode surfaces. Irreversible
adsorption or covalent-attachment of various molecules to
an electrode surface can impart specific catalytic prop-
erties which the electrode alone did not possess. Copper
phthalocyanine (CuPc) in its solution form, strongly ad-
sorbed, or covalently-attached to a metal oxide electrode

(SnO or Ti02) was studied. Spectrophotometric methods

2
aided in determination of the number of adsorbed or co-

valently-attached dye molecules present on the electrode
surface. Differential capacitance measurements for SnO2

and TiO and cyclic voltammetry of the tetrasodium salt of

2
tetrasulfonated copper phthalocyanine (CuPc(SOSNa+)h) in
DMSO and H20 resulted in energy mapping of the semiconductor
band structure in relation to the redox couples of the dye.
Cyclic voltammetry of adsorbed and covalently-attached dye
indicated retention of an electrochemically active species
on the surface. ESCA analysis of electrodes in various

states of electrochemical treatment showed two types of

phthalocyanine present and variation in the copper valence

state with change of solvent. Photocurrents were generated
by the adsorbed and covalently-attached CuPc electrodes.

Photocurrent studies indicated a potential dependent decom-
position/desorption process of the adsorbed dye electrodes.
A steady potential dependent response was observed for the

covalently-attached CuPc electrodes.

Dedication

To Mom and Dad

i1

"ACKNOWLEDGMENTS"

I would like to thank Neil and Monty for their advice
and encouragement. I would like to thank my parents, grand-
parents, Mrs. Elmer Clark, Mr. Alden Eddy and friends for
their encourgement, faith, and thoughts. And many thanks

to Victoria for her help in the preparation of this thesis.

iii

CHAPTER I.
A.
B.
C.
D.
E.
CHAPTER II
A.
B.

F.

G.

CHAPTER III.

A.

"TABLE OF CONTENTS"

INTRODUCTION. . . . . . . . . .
Semiconductors (n-Type). . . . . .
Chemical Modification of Electrode
ESCA . . . . . . . . . . . . . . .
Photocurrent Response Studies. . .
Phthalocyanines. . . . . . . . . .
. EXPERIMENTAL . . . . . . . . .
Dyes .-. . . . . . . . . . . . . .
Solvents and Electrolytes. . . . .
Spectrophotometric Studies . . . .
Electrochemical Studies. . . . . .
Covalent-Attachment. . . . . . . .

1. Sulfonamide Attachment . . .
2. Thiol. . . . . . . . . . . .
Photocurrent Studies . . . . . . .
ESCA Analysis. . . . . . . . . . .
RESULTS AND DISCUSSION. . . .
Solution Studies . . . . . . . . .

1. Spectrophotometric Studies of CuPc(SO

O

3

Na+)h. . . . . . . . . . . . . . . .

2. Electrochemistry of CuPc(SOSNa+)h. . .

iv

10
16
17
18
18
19
20
21
21
22
22
25
26

26

. 26

B. Adsorbed Dye Studies. . . . . . . . . .

1. Spectrophotometric Studies of
CuPC(SOSN8+)4 . . . . . . . . .

2. Electrochemistry of CuPc on Snoz.
3. ESCA of CuPc on Sn02. . . . . . .

4. Photocurrent Response of CuPc on 31102

C. Covalent-Attachment Studies . . . . . .

1. Spectrophotometric Studies of
Covalently-Attached CuPc. . . .

2. Electrochemistry of Covalently-
Attached CuPc-Sn02 Electrodes .

3. Photocurrent Response . . . . . .
CHAPTER IV. SUGGESTIONS FOR FUTURE WORK . . . .
LIST OF REFERENCES. 0 O C O O O O O O O O O O O C

41

41
43
45
52
58

58

60
64
67
69

"LIST OF TABLES"
Page

Table 1. ESCA Data CuPc/Sn02. . . . . . . . . . . . . .49

vi

Figure

Figure

Figure 3

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

13.

1h.

"LIST OF FIGURES"

Energy diagram of semiconductor-electrolyte

interface. . . . . . . . . . . . . . . . . 4
Structure of phthalocyanine. . . . . . . . . 11
CuPc covalent-attachment schemes . . . . . . 15

Absorption spectra of CuPc(SOSNa+)LL in DMSC

and H20. 0 e e o e o o o e e e o e o e o o 27

Cyclic voltammetry of CuPc(SO§Na+)u in 0.1fiL
TEAP/DMSO at (a) Sn02, (b) Ti02, and (c)
Pt 0 O O O O O O O O I O O O O O O O O O O 30

Cyclic voltammetry of CuPc(SOSNa+)u in R20

(pH u) at (a) Sno2 and (b) TiO . 32

2. O O O 0
Differential capacitance vs potential plots

of SnO2 in 0.1IL TEAP/DMSO and H20(pH h) . 35

Differential capacitance vs potential plots
of TiO2 in 0.1ll TEAP/DMSO and R20(pH h) . 37

Energy diagram of CuPc(SOSNa+)u redox couples

in 0.1 u TEAP/DMSO at SnOZ, Ti02and Pt . . 39

Energy diagram of CuPc(SOSNa+)LL oxidation in

R20(pH h) at SnO2 and TiO2 . . . . . . . . 40

Absorption spectra of clean SnO2 and CuPc sub-
limed on SnO2 electrodes . . . . . . . . . 42

Cyclic voltammetry of CuPc sublimed on SnO2
in0.1f.LTEAP/DMSO............ 44

Cu(2p1/2 3/2) ESCA spectra of. . . . . . . . 48
N(1s) ESCA spectra of. . . . . . . . . . . . 51

vii

Figure

Figure

Figure

Figure

Figure

15.

16.

17.

18.

19.

Photocurrent response of a CuPc/SnO2 elec-
trode in 0.05 Na2CZOu/pH 7 buffer. . . . . 53

Photocurrent response vs anodic bias poten-
tial plots of two CuPc/SnOP electrodes in
0.05M NaZCZOu/pH 7 buffer. . . . . . . . . 56

Absorption spectra of clean SnO2 and coval-

ently-attached-SnO2 electrodes . . . . . . 59

Cyclic voltammetry of a covalently-attached
CuPc-SnO2 electrode in 0.111 TEAP/DMSO . . 62

Photocurrent reaponse vs anodic bias poten-
tial of a covalently-attached CuPc-SnO2

electrode. . . . . . . . . . . . . . . . .66

Viii

CHAPTER I

INTRODUCTION

2

The photodecomposition of water into its elements (H2,
02) is one of the most attractive means of storing solar
energy. This decomposition process may be broken down
into three steps: light absorption, water oxidation, and
water reduction. Metal oxides (SnO2 and Ti02) doped as
n—type semiconductors have catalyzed the decomposition of
H20 (1 ). Certain dyes, by virtue of their high absorption
in the visible region and their facility in undergoing
oxidation-reduction reactions, should also be capable of
mediating some or all of these processes. Dye in solution
( 2), adsorbed (.3), and covalently-attached (li,ES) at
metal or semiconductor electrodes have been used to gener-
ate photocurrents. A brief discussion of semiconductor
electrochemistry, covalent-attachment (chemical modificat-
ion), and photocurrents is essential for a better under-
standing of light energy conversion. A useful tool for
surface analysis, electron spectroscopy for chemical

analysis (ESCA), is also discussed.
Semiconductors(n-Type)

Great progress in semiconductor electrochemistry has
been made since realization of their potential use in solar
cells. Various n-type semiconductors such as Sn02 and TiO2
have been used extensively for studies in both aqueous and
nonaqueous media (6:). Electrode processes on semicon-
ductors show certain characteristics that are different

from those of electrode reactions on metals. The main

3

difference in reactions on metals and on semiconcuctors is
that in the latter, the kinetics of the reaction may depend
on processes that occur within the solid electrode.

A bending of the bands occurs at the surface of a
semiconductor in contact with an electrolyte. Application
of an anodic potential results in formation of a depletion
or positive charged layer at the semiconductor surface (or
in the space charge region within it), thus leading to
upward bending of the bands (Figure 1). Application of a
cathodic potential gives rise to a downward bending of the
bands due to formation of a negative charge layer (7 ).

When a semiconductor is in contact with an electrolyte,
a significant part of the potential drop is across the
space charge region within the semiconductor; the rest is
across the double layer on the electrolyte side. In
experimental investigations of the double layer at semi-
conductor-electrolyte interfaces, an important parameter
is the capacitance of the interface, C, which varies as a
function of electrode potential V. The measured capacit-
ance C is composed of 03c, capacitance of space charges in
the semiconductor, and CH' the capacitance of the Helmholtz
double layer on the solution side of the interface. Since
the capacitances are in series and Csc is usually small as

compared to C the measured capacitance (C) is essentially

H'
equal to Csc ( 7).

A differential capacitance plot of 1/033 V8 potential
according to the Mott-Schottky equation

 

 

E.LL ./

 

Figure 1. Energy diagram of semiconductor-electrolyte
interface.

5
1/C:c= 2/660eonOCDS- kT/eo
yields information about the effective carrier density no
from the slope of the plot and flatband potential vfb
from the potential axis intercept. The flatband potential
is the potential at which excess electrical charge in the
semiconductor is zero and is indicative of the position of
the Fermi level in the semiconductor band gap region with
respect to the reference electrode. CSc is the capacitance
of the semiconductor electrode, E and 60 are the dielect-
ric constant of the semiconductor and permittivity of free
space, respectively, so is the absolute value of the elect-
ronic charge, no is the donor density, k is the Boltzmann
constant, and.qb8, is the potential difference between the
flatband potential and the potential at which the measure-

ment is made (7 ,8 ).
Chemical Modification of Electrode Surfaces

Chemical modification of surfaces is useful in research
areas such as chromatography and catalysts. Chemically-
modified solid supports have been use to improve chromato-
graphic column performance. This has depended on an
organosilane reaction with the surfaces of silica or
alumina particles. Homogeneous catalysts have been attached
to silica to provide heterogeneous hydroformylation catalysts
(<9). Acid-base dye indicators have been attached to
silica surfaces to create solid indicators (10).

Electrochemistry is a field where chemically-modified

6
surfaces are of great value. Chemical modification of an

electrode surface refers to strong binding of a selected
chemical reagent to the surface to endow it with some or
all of the chemical and electrochemical properties of the
selected reagent. Such chemically-modified electrodes rep-
resent new approaches to the study of electrochemical
reactions.

Modification based on covalent bond formation between
the reagent and electrode has been described for carbon (11),

Sno2 (12,13), and T10 (1h). Optically active amino acids

2
have been bound to carbon electrodes via amide bonds to form

a chiral electrode (15). The work on SnO2 and TiO2 elect-
rodes utilize organosilane reagents, which bear amine,
pyridyl, mercapto, and other functional groups. Schematically
the reaction of an organosilane with a surface hydroxyl

group can be represented as

a a
-M OH + SlYZR --§ -M€§—-OSiYZR + HA (13).
en face silane ‘ urface

The modified surface chemical stability is quite good except
in the presence of strong acid or base solution (13).
Rhodamine B and iron porphyrin have been attached to

SnO via an amide bond tit.16). Recently, Hawn and Armstrong

2
attached erythrosin to SnO2 using either an amide or a thiol

linkage (53). These chemically-modified electrodes have
displayed some very interesting electrochemical and photo-

electrochemical results in which the electrodes appear to

be chemically stable.

ESCA

The ESCA (electron spectroscOpy for chemical analysis)
technique was originated and developed by Siegbahn and co-
workers in Sweden during the 1960's (17,18). In ESCA,
nearly monochromatic x-radiation of 125hev (Mg K11'2) energy
impinge on a surface. Radiation in this range provides
sufficient energy for the ejection of core electrons. Al-
though the x-rays pentrate to thousands of angstroms, the
photoelectrons produced have an escape mean free path of
only S A0 to 100 A0, depending on the material. Core elect-
rons kinetic energies, Ek’ are measured accurately, making
it possible to determine the binding energies, Eb, of the
core electrons from the relation

Eb = hv I Ek ' qup’
where¢>8p is the work function or energy necessary to raise
a free electron from the Fermi level to vacuum level .
Binding energies are responsive to changes in the chemical
environment. Shifts in binding energies are produced by
the charge and valence state of an atom and the electronic
relaxation effect from neighboring atoms. This refers to
the effect of electronic charge flow from neighboring atoms
to the core hole in the positive ion produced by the photo-
emission (19,20).

ESCA is sensitive to all the elements but hydrogen and
helium. Since the measured electron binding energies of
most elements are unique, a broad range scan of electron

energy spectrum provides a good means of qualitative

8

analysis. The sensitivity of ESCA is on the order of 10"6
grams. The integrated intensities of the electron signals
are directly proportional to the number of similar atoms in
the sample, representing a pseudo-quantative tool. To insure
minimum surface contamination, mainly carbon, a 10-8 to 10'10
torr vacuum is required in the sample chamber. Samples may
be cleaned or depth profiled by an argon sputter gun (19).

ESCA has been used extentfully in adsorption (21),
surface oxide (22), and catalysts (23) studies. Surfaces of
clean (12.2h) and chemically-modified (13,11) semiconductor
electrodes have been characterized. This has lead to a

better understanding of the chemical and physical structure

of these surfaces.

Ehotocurrent Response Studies

n-Type semiconductors such as T102 and SnO2 will eat-
alyze the oxidation of water when irradiated with greater
than bandgap light energy (1 ,25). The electrooxidation of
water occurs as a result of vacancies produced in the valence
band of the highly-doped semiconductor (7 ). .A depletion or
space charge layer is maintained at the electrode surface by
means of an applied bias potential. Under these conditions,
the conduction band remains charge depleted, and holes
created in the valence band move to the surface and accept
charge in the oxidation of H20 or OH'.

It is well known that certain photoexcited dye molecules

in the vicinitv of the electrode interface are capable of

9

transferring charge to the conduction band of an n-type
semiconductor electrode (2, 26). Figure 1 explains this
process in terms of an energy diagram with the redox react-
ions of the dye in both the ground and excited states.
Photoexcitation of the dye occurs at light energies less
than the semiconductor band gap energy. The excited dye is
than capable of losing an electron to the conduction band
of the semiconductor.

Anodic bias results in charge depletion of the conduct-
ion band, and electron flow is from the dye towards the semi-
conductor conduction band. Photocurrent response as a
function of wavelength usually parallels the absorption
spectrum of the dye (22). If a reducing agent such as sod-
ium oxalate is present in solution with the dye, enhanced
photocurrents are observed (supersensitization). This is
due to a continual regeneration of the ground state of the
dye by the reducing agent. The kinetics of the supersensi-
tiaztion reaction depend upon the extent of overlap of the
energy distribution of the excited dye molecule population,
redox states of the reducing agent, and appropriate energy
level distribution of the semiconductor (26).

Photocurrent response studies have been carried out on
dyes in solution (2 ,26 ), adsorbed ( 3), or covalently-
attached (it) at both metal and semiconductor electrodes.
All of the dyes absorb strongly in the visible region.

Large extinction coefficients are necessary for maximum

capture of the sensitizing radiation. In most cases,

1O
multilayers of dye result in increased quenching and an
ohmic resistance for electron transfer is formed. This
prevents fast removal of electronic charge from the dye

layer (2).
Phthalocyanines

Phthalocyanine is a large organic heterocycle contain-
ing 7T electrons. Transition metal and post-transition
metal ions can be coordinated within this ring (Figure 2)
(27); these highly colored compounds are unique in many ways.
They are insoluble in most organic solvents and only slightly
soluble in solvents such as o-dichlorobenzene and pyridine.
Various functional groups such as halogens, amines, sulfonic
and carboxylic acids can be substituted for hydrogens on the
four benzene rings (Figure 2) (27). These functional groups
greatly affect the color and solubility of the phthalocyanines.
Colors vary from greens to blues; this is important in the
dye industry where the phthalocyanines have their greatest
use.

Absorption spectra of phthalocyanines in solution and
sublimed on glass have been reported (28,29). They absorb
strongly in the visible region, having large extinction co-

1 1

efficients (30,000M‘ cm‘ - 150,000M- cm'1). The strongest

absorption is in the 600nm-750nm (2.1ev-1.8ev) region and

corresponds toTT Trfi transitions of the ring system (30).

 

Cyclic voltammetry studies of CuPc(SO§Na+)u in .1 TEAP/

DMSO at mercury show two reversible one-electron reductions

n: H, SO3H' I, 80201, etc.

m: cm“? Co+2, Ge+u, Si+u, etc.

Figure 2. Structure of phthalocyanine.

12
and an oxdation (30). Chemical stability of the monoanion,
dianion, and monocation is indicated by the reversible re-
dox couples. The anodic redox couple of copper, zinc, and
nickel phthalocyanine was inaccessible due to use of a mer-
cury electrode. Absorption spectra of negative ions formed
by sodium reduction of various phthalocyanines have indicated
addition of electrons to the ring system and not the central
atom (31).

The ability for electrochemical reduction of oxygen by
organic semiconductors such as phthalocyanines has attracted
the attention of many researchers in connection with their
possible use in fuel cells (32,33,3h ). Much of the phth-
alocyanine electrochemistry reported is centered around
this. Photoinduced reduction has been done on phthalocyanine
films on carbon and platinum in various pH buffers. It has
been shown that the catalytic properties of phthalocyanines
are largely determined by the nature of the central atom (35).
The order of decreasing electrocatalytic activity is Fe:>Co
;>Ni)>Cu;2H2(metal-free). This has indicated that phthalo-
cyanines with higher magnetic moments (or paramagnetic sus-
ceptibility) appear to exhibit electrocatalytic activity
toward the oxygen reduction reaction (36).

Recently, attachment of cobalt phthalocyanine to cross-
linked polyacrylamide produced a stable oxidation catalyst

with enhanced activity. It was coupled by means of cyanuric

Chloride t0 N32 groups of the polymer matrix (37).

ESCA analysis of phthalocyanines sublimed on COpper has

13
shown two types of nitrogen present in metal-free phthalo-
cyanine (four equivalent central nitrogens and four equi-
valent meso-bridging nitrogens). Presence of a metal atom
in the ring equalizes the energy of these nitrogens and re-
sults in a sharp nitrogen (18) peak with a weak satellite
(38. 39 ).

The choice of copper phthalocyanine for use in our studies
was for the following reasons: 1) structural simlarity to
naturally-occuring porphyrins, 2) large extinction coef-
ficient, 3) ease of copper detection in surface analysis by
ESCA. Our initial investigations of the CuPc in its solution
form, strongly adsorbed, or covalently-attached to SnO2 or
TiO2 electrodes are discussed. Spectrophotometric studies
were used to determine regions of maximum absorption and ex-
tinction coefficients for the CuPc solutions and electrodes.
Absorbance values were used to determine the number of dye
molecules present on the adsorbed or covalently-attached
dye/SnO2 electrodes.

Differential capactiance plots for SnO2 and TiO2 gave in-
formation concerning the band structure: the Fermi level
position and no, the donor density. Cyclic voltammetry of
CuPc(SOSNa+)h in DMSO and H20 was done to determine the pos-
itions of the redox couples at Sn02, T102, and Pt. From
these, energy diagrams showing the relation between the semi-
conductor band structure and redox couples of the dye could

be drawn.

CuPc sublimed on SnO2 was used in cyclic voltammetric

14
studies and the behavior compared to that of the solution

form. ESCA analysis of electrodes in various states of
electrochemical pretreatment in DMSO and H20 along with
standards was done to aid in understanding their electro-
chemical behavior.

Covalent-attachment of CuPc to SnO2 and TiO2 was
attempted via a sulfonamide or thiol formation using the
sulfonyl chloride or tetraiodated form of the dye, respect-
ively (Figure 3). The dye was coupled to the surface using
various organosilanes which had either a terminal amine or
mercapto functional group.

Cyclic voltammetry of the covalent-attached dye-semi-
conductor electrodes was carried out to determine the
chemical stability and position of redox couples. The
behavior of the covalently-attached dye was compared to
that of the solution and adsorbed form.

The spectrophotometric, electrochemical, and ESCA
studies characterized the behavior of the dye in its various
forms. This facilitated understanding of the photocurrent
response data. Photocurrents of adsorbed and covalently-
attached CuPc-Snoz electrodes in pH 7 buffer as a function
of anodic bias potential were explored. A reducing or
supersensitizing agent such as sodium oxalate was added to

enhance the photocurrent response.

15

mosenom pcosnomppmupCoHa>oo omSU .m opswfim

cm

\Cm _ x Q n 0
$53.25)“. Oil) L??? _ + 1m...» .6 I 8.2. + cm
13% 2 0

cm

2:
2
/

\/\

 

2
/

~/
\I
2
Di
I
\l
\
\

40.1%
Guam Owe
,cm
I
, i \
z z 2
.o~o z , .5 Guam azuriW/Q u m 01.6
QT. muolziaxofizmiox 4i \ fir\ m~o.u + :zifii 85. + 0\
z _ , HEW: /
ll . Z 1 cm \cm
p~o Gwen

MO_Z<ZOui_3m

CHAPTER II

EXPERIMENTAL

17
Dyes

Copper phthalocyanine was obtained from Eastman Kodak.

3

NaI)’4 was synthesized according to the procedure of Weber and

h,h',h",h'"tetrasulfonated copper phthalocyanine (CuPc(SO

Busch (U0). The monosodium salt of h-sulfophthalic acid
(0.162 mole), ammonium chloride (0.09 mole), uera (0.97 mole),
ammonium molybdate (0.0006 mole), and cOpper sulfate S-hy-
drate (0.0h8 mole) were added to no ml of nitrobenzene and
heated at 180°C for six hours. The crude product was then
washed with methanol, 1N hydrochloric acid, saturated with
sodium chloride and then dissolved in 0.1N sodium hydroxide.
After the solution was heated to 80°C and filtered, sodium
chloride was added to precipitate the solid product. The
pure blue product was filtered, washed with aqueous ethanol,
and dried in vacuo overnight.

h-iodophthalic anhydride needed for the synthesis of
tetra-h-iodo-copper phthalocyanine was prepared by the method
of Higgins and Hilton (h1). S-iodo-anthranilic acid (0.12
mole) was added to 500 ml of 3.3M sulfuric acid and cooled to
5°C. After the addition of 23 g of sodium nitrite and stir-
ring for two hours, the diazonium solution was poured onto
crushed ice and raised to a pH of 6.5 with sodium hydroxide.
A complex cyanide solution was added and the diazonium salt
solution was heated, cooled, and acidified. The resulting
solid product was added to a dry benzene and acetic anhydride
solution, refluxed, and cooled to give h-iodophthalic any-

dride. Tetra-h-iodo-cOpper phthalocyanine was prepared by

18
the method of Suzuki and Bansho (U2). h-iodophthalic anhy-
dride (13.2 g), urea (23 g), copper (II) chloride 2-hydrate
(h g), 1,2,h,trichlorobenzene (hOO g), and titanium (IV) chlo-
ride (0.2 g) were heated to 160°C for one heur, kept at 177-
18800 for fifty minutes, filtered, and washed with benzene,
methanol, 3% hydrochloric acid, water, and ethanol. The solid

blue-green product was dried in vacuo overnight.

Solvents and Electrolytes

Spectragrade dimethylsulfoxide (DMSO) was obtained from
Burdick and Jackson Labs. It was purified by passing through
an activated alumina column and then stored over activated
h A0 molecular sieves. The water used was distilled and pas-
sed through a Milli Pore Milli Q system containing an anion
exchanger, cation exchanger, and activated charcoal. Reagent
grade tetraethylammonium perchlorate (TEAP) was recrystal-
lized from ethanol and dried at 80°C overnight. Reagent
grade nitric acid, sodium hydroxide, sodium bicarbonate, po-
tassium dihydrogen phosphate were used as was for the pH h

and 7 buffer solutions.
§pectrophotometric Studies

Spectra of CuPc(SOSNa+)h and CuPcIu were obtained on a
Banach, Lomb/Shimadzu Spectronic 210 UV SpectrOphotometer
using quartz cells of one centimeter path-length. 8000 A0
to 3000 A0 were scanned at 500 A0 per minute using DMSO,

pyridine, and H20 as solvents. Sublimed films and covalently-

19
attached dye on glass and SnO2 were also studied at the same

scan rate and wavelength range (h3).

Electrochemical Studies

Flouride doped Sn02 on 0.38 inch thick glass was manu-
factured by Pittsburg Plate Glass Co. Sheet titanium metal
was obtained from Timet to produce TiO2 electrodes. After
being cut, polished, and cleaned the Ti electrodes were ox-
idized by heating red hot, then partially reduced by passing
nitrogen gas over them in a muffle furance for fifteen min-

utes (Sh). This gave doped TiO films having surface resis-

2
tances between SOOSI and 1.5kj1 per square. The thickness of
the Sn02 films varied between 5000 A0 and 7000 A0. The elec-
trodes were cleaned in an ultrasonic bath, followed by suc-
cessive washing with Alconox detergent, ethanol, and distil-
led Milli Q water. The electrodes were then vacuum dried and
stored in the glove box.

The solid state potentiostat was of conventional design.
The capacitance determination apparatus included a PAR 126
lock-in amplifier, Tektronix osciloscope, Krohn-Hite 5200
waveform generator, and Fluke digital voltmeter. Capacitance
measurements of the eletrodes were made by the method of
Gileadi and co-workers (uh). A twenty-five millivolt ampli-
tude triangular wave of between 100Hz and 1KHz was super-
imposed on a slowly changing, linear ramp potential. Cali-

bration of output current response was made by means of an

external circuit. Compensation for the iR drop between the

20
reference and the working electrode was made by positive
feed back to the control amplifier of the potentiostat.

The basic cell design used for the capacitance and elec-
trochemical studies allowed for sandwich—like positioning of
the semiconductor film electrode to the cell body (hS). The
cells were made of telfon or Lucite for DMSO and H20, re-
spectively, and had a volume of approximately 5 ml. The aux-
illary electrode was platinum wire and the reference elec-
trode was Ag wire for DMSO studies and an isolated Ag/AgCl
electrode for H20 studies.

Solutions were deoxgentated by either of two methods:
(1) Bubbling with dry, purified nitrogen for AS minutes.

(2) Pumping in vacuo with stirring for five minutes then
flushing with dry, purified nitrogen, these two steps being
repeated several times (h6).

Cyclic voltammetric studies of CuPc in the solution,
adsorbed, and covalently-attached state were done in DMSO and
H20. Solution studies of CuPc(SOSNa+)u were carried out

using Sn02, TiO and platinum as the working electrode. CuPc

2!
was sublimed onto SnO2 at S x 10'5 torr. The optimum time
was fifteen minutes at 250-27000, after which the electrodes
were cooled and stored in a nitrogen atmosphere (h7). This

procedure gave defined area, nonporous films of CuPc 20-100

monolayers thick.
Colvalent Attachment

Covalent attachment of CuPc to Sn02 and TiO2 was done

21
using two derivatives of CuPc. The Sn02 and Ti02 electrodes

were silanated using gamma-aminopropyl-triethoxysilane, N,
beta-aminoethyl-gamma-aminopropyl-trimethoxysilane or mer-
capto-propyl-trimethoxysilane. The electrodes were refluxed
in a 1% to 5% silane-dry toluene solution under nitrogen for
twelve hours after which the excess silane was removed by
refluxing the electrodes in dry toluene for one hour (13).

The silanated electrodes were stored in dry toluene in the
glove box. CuPc was covalently attached to the silanated
semiconductors via a sulfonamide formation using CuPc(SOSNa+)u

and a thiol formation using CuPcIu.

Sulfonamide Attachement

3

benzene, and 2 drOps dimethylforamide were refulxed with

0.1 g CuPc(SO Na+)u, 20 ml thionyl chloride, 20 ml dry
stirring for forty-eight hours under nitrogen. The dye was
then filtered and washed with dry benzene until the filtrate
gave a negative AgCl test. The silanated SnO2 and TiO2 elec-
trodes, 0.1 g CuPc(SOZCl)h, and 100 ml dry benzene were re-
fluxed under nitrogen for seventy-two hours (h8). The chem-
ically-modified electrodes were washed in a Sohxlet extractor
with benzene and water for six hours each, vacuum dried, and

stored in a nitrogen atmosphere.

Thiol

100 m1 of 1 x 10-3M CuPclu in pH 7 buffer (KH2P02,N30H)

was reacted with the silanated semiconductor electrodes at

22

25-3OOC under nitrogen for ten to twelve hours. The chem-
ically-modified electrodes were then washed in Sohxlet extrac-
tor with water for twenty-four hours, vacuum dried, and stored

under nitrogen.
Photocurrent Studies

Photocurrent responses of the adsorbed and covalently-
attached dye-semiconductor electrodes were studied in DMSO
and H20. The studies were done at wavelengths corresponding
to the absorptions of the semiconductors (below 350nm.and

below hOOnm for SnO and Ti02, respectively) and for the dye

2
(665nm, 630nm). Also, sensitizing agents such as hydroquinone
and sodium oxalate were added to the solvents, and photocur-
rent studies carried out (2). The experimental apparatus
included a Electra Powerpac Corp. power supply, hSOW or 1ooow
zenon arc lamp, PAR lock-in amplifier, chopper, Jobin Yvon

monochromator or colored filters, Tektronix oscilloscope, and

lab-built potentiostat.

ESCA Analysis

Adsorbed and covalently-attached dye-SnO2 electrodes,
from the electrochemical studies, were used for ESCA analysis.
Electrodes biased at different potentials past the various
redox peaks of the dye or combinations of peaks as well as
unused electrodes were analyzed. All manipulations of the

electrodes following the studies were done in a glove box

under purified, dry nitrogen. The electrodes were rinsed

23
throughly with ethanol, vacuum dried, and then mounted onto
appropriate holders for the surface analysis.
The ESCA data were abtained using a Physical Electronics,
Inc. (PHI) Model 5&8 ESCA/Auger Spectrometer which was equip-
ped with a Mg source (Mg Kcr1,2). The Mg x-ray beam was Oper-
ated at a power of hOOW. The pressure in the analyzer cham-

9torr during analysis.

here was maintained at less than 10'
Binding energies of the ESCA transitions were corrected for
charging effects by refercing to the Sn(3d5/2) line of
standard Sn02, h86.2ev (h9).

Data acquisition, storage, and processing, particularly
for obtaining signal averaged ESCA spectra, were accomplished
using a NOVA 800 minicomputer (Data General Corporation) which
was equipped with 32 K core of memory, 2 Diablo disk (1.2 M
bits) and x-y plotting facilities.

Part of the ESCA spectra were "deconvoluted" into sep-
arated spectral components. They were computer-simulated by
inputting the following parameters: (1) slope of the linear
spectral background; (2) the binding energy of each compo-
nent; (3) the full width at half maximum (FWHM) of each com-
ponent; (A) peak height of each component; and (S) the per-
centage of Gaussian contribution to the shape of the spectral
band. Parameters 3 and 5 were dependent on the pass energy
of the analyzer (25ev or SOev). Thus, parameters 1, 2, 3, and
5 were held constant for each simulation and parameter h

varied for each component until a "best fit" was attained for

the entire spectrum. Each simulated spectrum was also

24
corrected for the satellite peaks which were present in the

PHI instrument (without a monochromatic x-ray source) (50).

CHAPTER III
RESULTS AND DISCUSSION

26

Solution Studies
- +
Spectr0photometric Studies of CuPc(803Na )LL

Spectrophotometric studies of CuPc(SO§Na+)LL in DMSO and

H 0 were carried out to determine major absorption bands and

2
their extinction coefficients. The visible spectrum of CuPc
(SOSNB+)h in DMSO (Figure h) shows major transitions at 677nm

514-1 -1

and 3h9nm with extinction coefficients of 1.2 x 10 cm

and 0.h x 1OSM""'cm"1 respectively, corresponding to TF——1T*
transitions of the phthalocyanine ring. The spectrum in
water (Figureli) is different; two bands at 630nm and 665nm
(E = 6.1 x 10uM'1cm'1 and h.h x 1OuM'1cm-1, respectively) are
observed giving a broad absorbance in the SOO-700nm spectra
region. The spectral band at 336nm appears to correspond to
a shift of the 3h9nm band in DMSO, with little change in its
extinction coefficient. Association or self-solvation of

CuPc(SOSNa+)h in water is implied by the presence of the
broad 630, 655nm doublet (51).

Electrochemistry of CuPc(SOSNa+)I

The electrochemical reactions of CuPc(SO§Na+)u were ex-
plored by cyclic voltammetry in DMSO and 820 to determine the
position of each redox level of the dye with respect to the
conduction and valence bands of SnO2 and T102. The aim was
to probe the relation between the band structure and reac-
tivity of solution species and map the band gap region of the

semiconductors to determine the presence and energies of any

27

com

 

¢ ance—DNCO

 

 

new omzm c“ 4A+azmomvom50 mo «spoons Cowugpomp< .: shaman

2.5 5:22;:

 

 

 

 

azlll

 

32: III

28

intermediate bands or surface states.

The cyclic voltammetric behavior of CuPc(SOSNa+)u in
DMSO at highly-doped SnO2 and TiO2 electrodes as well as plat-
inum is shown in Figure 5. Redox couples a/a' and b/b' cor-
respond to the formation and reoxidation of the mono and di-
anion forms of CuPc (y1). The two one-electron processes
are nearly reversible on a platinum electrode ([preak <: 100mv
for v<< 100mv/sec) and have been attributed to addition of
electrons to the phthalocyanine ring and not the reduction of
copper (31). These processes appear to be slightly less re-
versible on SnO2 (APEpeak larger at the same scan rate). Re-
dox couples a/a' and b/b' are poorly resolved and appear even
less chemically (i.e. slightly disproportional peak current
ratios) reversible on T102; a/a' is also obscured by a de-
sorption process.

Redox couple c/c' corresponds to the formation of the
monocation form of CuPc and its reduction. The cyclic voltam-
metric studies indicate that the oxidation is followed by a
slow chemical reaction. The peak current ratio 1pc/ipa for
this process approaches 1.0 on SnO2 electrodes as the scan
rate is increased. This redox couple appears to be more chem-

ically reversible on SnO2 than on platinum or TiO An ex-

2.

planation for this will be given in a later section.
Redox couples d/d' and e/e' are unexplained reductions

which do not appear very well resolved unless an anodic scan

has been performed previously. The trianion of CuPc has been

formed via sodium reduction (52); it is possible redox couple

29

Figure 5. Cyclic voltammetry of CuPc(SO-Na+) in 0.1

TEAP/DMSO at (a) Sn0 (b) TiO and (0) Pt.

2’ 2

 

'50

l-——-1500 mv

 

38mvlsec
(a)
I l5-5uAlcm2
C
c,
(b)
I 31-3uAlcm2
C
cl
(C)
131.3uAlcm2

C

-
a:

 

31
d/d' is the formation and reoxidation of the trianion. It is

also conceivable that one of these couples correspond to the
reduction of the product of the chemical reaction consuming
the oxidative intermediate of CuPc(SOSNa+)h. These redox
couples are not resolved as well from the background currents
on SnO2 as on Ti02 or Pt.
The cyclic voltammetric behavior or CuPc(SOSNa+)LL in H20
(pH h) was found to be more complicated than in DMSO (Figure 6).
Two poorly resolved reduction peaks were observed on SnO2 at
-0.8 volts and -0.93 volts vs Ag/AgCl. The reductions appear-
ed chemically irreversible, and anodic wave was observed on
the return scan at -0.1 volts corresponding to the oxidation
of products formed on the cathodic sweep. These reductions
were less resolved on Ti02. A very poorly defined oxidative
wave was observed at 1.0 volts and 1.2 volts vs Ag/AgCl for
SnO2 and Ti02, respectively, which was chemically irreversible.
The difference in electrochemical behavior of CuPc(SC§
Na+)u in DMSO and H20 probably results from the difference
in the chemical state of CuPc(SOSNa+)u in both solvents, the
stability of the redox intermediates in both solvents, and
the state of the semiconductor surface in both solvents. An
associated form of CuPc(SOSNa+)u in H2O may inhibit charge
transfer; the irreversible electrochemical behavior in H20

was not surprising (S1).

A differential capacitance plot of 1/02SC vs potential ac-

cording to the Mott-Schottky equation

2 -
1/csc - 2/€ oeono ¢>S- kT/eO

yields information about the effective carrier density nO from

32

38mvlsec

  

(a) [EMA/cm2

 

[31.3

[ti-3uA/cm2
(b)

 
 

"'—"'|500 mv

Figure 6. Cyclic voltammetry of CuPc(SO3

at (a) Sn0 and (b) Ti0

Na+)u in H201pH u)

2 2°

33
the slope of the plot and the flatband potential Vfb from the

potential axis intercept. The results of these measurements
for SnO2 and TiO2 in DMSO and 820 are shown in Figures 7 and
8. The effective carrier densities, assuming dielectric con-
stants of 12.5 and 127 for SnO2 and Ti02, respectively, were
as follows: -1.65 1 0.1 volts vs AgRE (SnOZ/DMSO), -1.h3 t
0.1 volts vs Ag/AgCl (Sn02/H20), -2.0 1 0.1 volts vs AgRE
(TiOZ/DMSO), and -1.h volts vs Ag/AgCl (Tiog/H2O)' From the
donor density, flatband potentials and band gaps of SnO2 and
TiO2 (3.5ev and 3.0ev, respectively) the energy positions of
the valence band and conduction band edges (Ev and EC) are
determined from the relation
Ec(ev) = eOVfb + kT 1n (no/Nc

where NO the density of states at the bottom of the conduc-
tion band is about 10'19 carriers cm.3 in most semiconduc-
tors (53).

Figures 9 and 10 show the relation between the band struc-
tures of the semiconductors and the redox couples of CuPc(SOS

Na+)u in DMSO and H O. The reductions and oxidation of CuPc

2
(303Na+)u occur at potentials coincident with the band gap
region of SnO2 and Ti02. The fact that the processes appear
chemically reversible or pseudo-reversible in DMSO may indi-
cate the contribution of surface states to the charge transfer
process or electron tunneling through the narrow depletion
region since the number of charge carriers is high. Surface
states can be thought of as intermediate energy levels of

narrouw width in the band gap region. Adsorbed species or sur-

face defects of the semiconductor generate these surface states

Figure 7. Differential capacitance vs potential plots of

SnO2 in 0.1IL TEAP/DNSO and H2O(pR h).

 

55

xas_a>v .a._m

=.~- 9—-

+-
C1!

+
1

Guinea

:34 3:...

I

l.)

1..-

N

b
4|

:5

=5 .—

b
1‘1

l
fin
J
i.
___l

IINO.

L-e=.

iTaa.
~e\_
lrss.

 

2.

LVN—.

 

ro-
:1

i.
o.
r.
9.

. u

:.

-- v-‘(nn
.a / '. a
a J. nJ'v

0".-
p
n

,l

la
/ a;
1+-
sf:
‘1’

37

3...: :5
9...

‘1?"
.1.
«s
.i

q
fir

.
q

is.

atmo.
lessee

NE.
.4...
sea 3:..-

ts.
ass? a: e

2 Na:
.8.

 

38

Figure 9. Energy diagram of CuPc(SOSNa+)LL redox couples in
O'1FL TRAP/DMSO at Sn02. TiO
(Fc) at Pt is given for reference.

2 and Pt. Ferrocene

 

39

 

 

 

u

a L
umduflumauicamava m .
w _
amdmadlsfl W m
a l

H
1.143: 5m: s m .
a . w I

1, m m W >ma

ashmfiu.fin- _ _ .
_ _ i

teammangen- n >oe - - _

asamswsew- . L
madmasnaw-
lungzu has? a a
m sch?
>00. u._ on
_02co.om _OZC0.0& 7 _OZCO.O&
xopea woven .xppea

& «0: N05

40

SM] Til)

N
N

 

‘_ -1” "'45" Ec\ \ \ \ .1.4V

0
J. ,
r;/
l

l
l
i

l
i
1]
ii

 

;///

//

-_ _-__a.///
/

—-— _...._L_
.—

 

 

 

 

3.0V
3.51! l
l
i 1
a 2”,, “.2.-.—
l Lzuzuiaza'
l
i Ly “‘7‘; 7‘1‘6V
! / /
z 4 /*
[VT/’fizz-OSV / ///A
/,/:// 3/
/////

Figure 10. Energy diagram of CuPc(SO'3'Na+)h oxidation in H20
(pH h) at SnO2 and Ti02.

41
through which electron transfer can occur. As the carrier
concentration increases, the charge depletion region at the
surface decreases in thickness. When the carrier density is
sufficiently high enough, this depletion thickness will be
10 A0 or less, which enables electron tunneling to occur (5h).
The redox couples of the dye and ferrocene in DMSO at plati-
num are shown in Figure 9 for reference.

The redox couples at SnO2 and Ti02 are shifted cathod-
ically from the potential at which they occur at pltinum.
When the number of charge carriers at the electrodes sur-
face approaches Nc, degeneracy begins and electrode behavior
becomes more metal like. Thus, the higher the carrier den-
sity, the more metal-like a semiconductor electrode's be-
havior is, i.e. reductions of CuPc(SOSNa+)u occur at cathodic
potentials closer to potentials corresponding to the reduc-
tions at platinum (8). The couples at SnO2 are shifted less
and the carrier density higher indicating the SnO2 is more

metal-like than the Ti02.

Adsorbed Dye Studies

SpectrOphotometric Studies of CuPc on SnO9

An absorption spectra of copper phthalocyanine (CuPc)
deposited on SnO2 is shown in Figure 11. Absorption maxima
occurs at 69hnm and 620nm in comparison to 691nm and 62hnm
for CuPc deposited on glass (29). The spectra indicated the
films to be in the at form. The CuPc thin films could be

vacuum deposited on the SnO2 surface in one of three ways:

42

 

 

CuPc on
SM]
2
I\
I:
3
O
r
t:
a
[1
c
e
Sn02
l l
l T
400 550 750
Wavelength (n m)
Figure 11. Absorption spectra of clean SnO2 and CuPc sub-

limed on SnO2 electrodes.

43
with the crystallites parallel with the surface, with the

crystallites standing obiquely to it, or some combination of
these two arrangements (63). Assuming an electrode area of
0.8 cm2, a dye diameter of 1h.3 A0, and an extinction coef-
ficient of 3.32 x iohm'1cm'1 at 618nm (27), one could de-
termine the approximate thickness of the films from Beer's

Law (A = be). The SnO contribution to the adsorption at

2
618nm was subtracted before calculating the thickness of the
films. This method resulted in films varying between 20-100

monolayers with 3.91 x 1013 molecules per monolayer.
Electrochemistry of CuPc on SnO2

The electrochemistry of CuPc sublimed on Sn02 in DMSO
is shown in Figure 12. Cathodic scans showed the first re-
duction wave (1) at -0.75 volts vs AgRE, shifited 0.13 volts
cathodically from the peak potential of the solution compo-
nent (Figure 5). Scanning back anodically resulted in an ox-
idation wave (2) at 1.h volts vs AgRE. Both waves appeared
chemically irreversible. Upon scanning to the second reduc-
tion peak (3) potential, a large symmetrical desorption cur-
rent peak was observed, coincident with the visible loss of
CuPc from the electrode surface. Returning anodic scans in-
dicated the reversible oxidation of the two reduction inter-
mediates, (3',1') but a resorption process wasn't observed.
If the potential was cycled repeatly between the cathodic and
anodic limits, the desorption occurring at the second catho-

dic peak potential slowly decayed away, the oxidation wave

44

 

 

 

 

 

 

 

38mv/sec i_i?200mv 2
3JuA/cm
(a) 2'
I— r
2
AI t—dSOOmv
IRMA/cm2
3
38mv/sec
(b)
3 l-——'l 50011”
I I 8.3uA/cm2
75mv/sec
(C)
T
2

Figure 12. Cyclic voltammetry of CuPc sublimed on SnO2 in
0.1,; TEAP/DMSO.

45

(2) and two reduction waves (1,3) retained a constant magn-
itude indicating the fact that some of the dye remained after
the desorption process and was electrochemically active. All
three peaks appeared to be chemically irreversible, i.e. dis-
proportional peak current ratios.

These electrochemical experiments indicate that the sub-
limation process produces a tightly-bound form of dye and a
loosely-held form which can be desorbed at a potential very
near the second reduction. The formation of the dianion Spe-
cies may force this desorption. If the dye remained purely
surface bound at all potentials, electron transfer to and
from it should occur with symmetrical reduction and oxidation
peaks with a‘QSEpeak near zero. Oxidation of the reduction
intermediates should be diffusion controlled since they were
still near the electrode surface immediately after being de-
sorbed. Diffusion control is indicated by the cyclic be-
havior observed, i.e. [iEpeak23100mv.

The cyclic voltammetric studies of CuPc sublimed on SnO2
in H20 resulted in poorly defined and irreversible reduction

and oxidation processes. No desorption of the dye was ob-

served during the scans in any aqueous media.
ESCA of CuPc on SnO9

To facilitate the understanding of the adsorbed CuPc/
SnO2 electrochemistry, ESCA analysis of electrodes in various
states of pretreatment along with standards was carried out.

Changes in surface concentration and valence states of the

46
adsorbed dye occurring as a result of these electrochemical

pretreatments are easily observed with ESCA. The Cu(2p1/2,
3/2) ESCA spectra are shown in Figure 13. Since Sn02 was a
common component for all these materials, the binding ener-
gies of all components were corrected to the Sn(3d5/2) line
of standard Sn02, h86.2ev (h9). The Cu(2p1/2,3/2) peaks ex—
cept for spectrum (g)(H20) are accompanied by multiplets at
binding energies approximately 9-10 ev higher than the 2p
transitions. This type of multiplet Splitting has been
previously reported for metal (paramagnetic) oxides. This
splitting has been contributed to the presence of oxygen in
the oxide lattice (Sn02), adsorbed oxygen, or another oxygen
containing molecule (55). Binding energies of the Cu(2p3/2)
and Cu(Auger) peaks for the various samples are given in Table
1. Because there is little binding energy shigt in the Cu
(2p1/2’3/2) peaks for the various copper species, the most
useful information can be obtained from the Cu(2p3/2) - Cu
(Auger) binding energies. This binding energy difference
does shift with a change in the copper valence state. CuPc/
SnO2 pressed pellet, CuPc/SnO2 unused electrode, and CuPc/
SnO2 electrode (DMSO) [NBE'S (597.1ev, 596.8ev, and 597.0ev,
respectively) correlate well with that of CuO (596.9ev) (57),
thus indicating the presence of Cu(II) species. The KNEE
for CuSOu (anhydrous)/Sn02 standard pellet is slightly lower,
595.8ev. The absence of multiplet splitting and the BE

(59h.9ev) for the CuPc/SnO electrode in H20 seems to sig-

2
nify the presence of a copper species more reduced than

Figure 13.

47

Cu(2p1/2 3/2) ESCA spectra of (a) CuSOu/Snoz
powder pressed pellet, (b) CuPc/SnO2 powder
pressed pellet, (c) CuPc/SnO2 unused electrode,
(d) CuPc/SnO2 electrode in DMSO (scanned past
1st reduction peak and back anodically to the
oxidation peak), (e) CuPc/SnO2 electrode in
DMSO (scanned past 2nd reduction peak and back
anodically to the oxidation peak), (f) CuPc/
SnO2 electrode in DMSO (scanned until desorp-
tion process complete), (g) CuPc/SnO2 electrode
in H20 (scanned until desorption process
complete).

 

Cu (2P1/2,3/2)

 

N(E)

 

1300 cts

o——-—4

500 cts 1

 

f)

iWcm
9)
WW 1300 C15

300 Cts

fl/

(1)
W 00 cts

if

1300 cls

 

 

l l
970 945

920

BINDING ENERGY (eV)

 

49

on.

as.

oz.

ADV

O
z

Anmv oogopomom

Aomv monogamomA

m.mqm

o.mom

0.50m

o.®om

 

leessamazv.m.m

on
on
on

Ir

on

Av

ea

 

mofiwmocm mcwpcfim Asows<vzo u Am\momvaoaov

mofipw>fiufiuCoo mmx opfipuaop mom popoomnoo

 

 

a Loo
Aodv oogopohom .>om.om: u Am\mpmvgm 0p popooaaooAav
m.mom so
m.oom omso
o.oom one
Acvm.mmo one
omm\ooosp
o.:om xe.oeovo.emm w.mmo -ooao mocm\omso
omxo\oooae
o.eom Ae.oeove.emm e.smo -ooao mosm\omso
pandas coon»
m.oom Am.oeovm.emm 0.3mm -ooHo mocm\omso
popzoa Nocm\uoHHon
e.eom Ae.oeovs.emm m.:mo cascades omso
e.mem he.o.ees.emm e.mme soaso
.m.mnw xovxxsoms<vsoe.m.m Aswddm\mmmvsov.m.m mmmmmm
mosm\omse ease «owe .e canoe

Figure 1h.

50

N(1s) ESCA spectra of (a) CuPc/SnO
trode, (b) CuPc/SnO2
past first reduction peak and back anodically to
the oxidation peak), (c) CuPc/SnO2 electrode in
DMSO (scanned past the second reduction peak and
back anodically to the oxidation peak), (d) CuPc
/Sn02 electrode in DMSO (scanned until desorp-
tion process is complete) and (e) CuPc/SnO2
electrode in H20 (scanned until desorption pro-
cess is complete).

2 unused elec-

electrode in DMSO (scanned

 

51

 

 

 

 

 

 

N ( Is)
a)

1300 cts A\
b)

1300 cts
1300 cts I C)
N( E) .
1300 cts . I d)
.‘k .
1300 C13 8)
' 46o sic
4K3

BINDING ENERGY (eV)

52
Cu(II) (57).

N(1s) spectra of the CuPc films consisted of two comp-
onents, a major peak at 398.0ev and a satellite at 399.7ev
(Figure 1h) as previously reported (38). After cycling the
electrode potential past the second reduction peak, which
resulted in a large desorption, the ESCA spectra showed a
diminished CuPc surface concentration (Spectrum d). The
N(1s) spectrum was then equally divided between the 398.0ev
and 399.7ev component, indicating the probably existence of
de-metalled phthalocyanine in the films (38). It seems rea-
sonable that the sublimation process could produce a small
amount of the de-metalled form of phthalocyanine. Since the
Cu(2p1/2’3/2) spectrum of this electrode indicated almost no
copper present, the de-metalled form was probably deposited
first. This new type of phthalocyanine system made up at

most approximately 5-10% of the total phthalocyanine film.
Photocurrent Response of CuPc on Sn09

The dark and light cyclic voltammgrams of CuPc depos-
ited on SnO2 in 0.05M Na2C20u(supersensitizer or reducing
agent as described on page )/pH 7 buffer are shown in
Figure 15. Since the photocurrent response corresponded to
the absorption spectrum of the CuPc, IR, UV, long pass h70nm,
and long pass 5h0nm filters were used. This combination of
filters resulted in a wavelength window in which maximum
absorption of the dye occurred (550nm-750nm); it also fil-
tered out any semiconductor response (<: hOOnm). The photo-

current response appeared to begin at approximately 0.5 volts

53

.soeeso a mo\:omomsz zmo.o as

opoapoofic mocm\omso a mo ongoomoh unopeneouocm .mr oaswfim

  

 

xmqo 3:: cm d1.

.52: to as

.mzze344

. .t _ divoo. um::.m.
%t€<1m~ > 00.

54

vs Ag/AgCl reference. A maximum value of 1.5 }LA/cm2 was ob-
tained at approximately 0.9 volts vs reference after which
the photocurrent response decayed back to essentially zero.

In another experiment, the light source was chopped at
13 Hz and the photocurrent measured using the lockin tech-
nique. This technique employed the light chopper frequency
as a reference for the lockin amplifier, thus allowing easier
detection of the electrode photocurrent response. The photo-
current as a function of anodic bias potential was studied.
The results for two electrodes having approximately the same
dye coverage are shown in Figure 16. Electrode A was biased
at 0.2 volts, 0.5 volts, 0.9 volts, then 0.5 volts and 0.2
volts vs Ag/AgCl reference and the photocurrent response re-
corded. The response appeared to reach a maximum between 0.5
volts and 0.9 volts. The second readouts at 0.5 volts and
0.2 volts were slightly less than the final steady response
at 0.9 volts.

The other electrode (B) was biased at 0.9 volts and an
initial reading taken. The response decayed over a period
of several minutes to a steady photocurrent approximately one-
fourth of the initial value. As with electrode A, the photo-
current response at 0.5 volts and 0.2 volts was slightly less
than the final steady 0.9 volts response.

After the photocurrent studies, visual examination of
the electrodes indicated partial desorption and/or decomposi-
tion of the CuPc film. This phenomenon appeared to be poten-

tial dependent; it occured at potentials anodic of 0.5 volts.

55

Figure 16. Photocurrent response vs anodic bias potential
plots of two CuPc/SnO electrodes in 0.05M Na
C2Ou/DH 7 buffer.

2 2

56

IR filter
UV filter
LP47 filter
(.1354 filter

93 monolayers

Lockin sensitivity SOO hv

 

 

 

 

 

 

 

 

 

 

 

 

.14.
( l
J J
2.-
i 2
12 PC.
response
. 0 ll (V)
i
i UWQ 9
I0 .8 .6 .4 .2 .0
Bios (v)
90 monolayers
(9 4
43
‘2 Dc.
response
1 -«I (V)
u— +0 0
(0 .8 .6 .4 .2 0

Bios (v)

57
As the bias potential was moved anodically, the photocurrent

response increased. This would be indicative of a higher
number of excited dye molecules and faster rate of electron
flow from the dye to the semiconductor conduction band. As

a result, the supersensitizing agent (oxalate ion) concentra-
tion at the dye-solution interface could be rapidly depleted
since diffusion of oxalate ion to the interface is not fast
enough to maintain a high concentration. Thus, there is a
build-up in the CuPc+ species concentration which with the
bias potential and light probably enhance the desorption and/
or decomposition process. Cyclic voltammetry studies indi-
cated a tightly-bound form of CuPc and a more loosely-bound
form on the electrode surface. Since a steady photocurrent
response was noted after the desorption—decomposition process,
it seems the tightly-bound form of dye is retained on the
surface. The photocurrent response then appeared to be less
potential dependent.

Increasing the oxalate ion concentration appeared to
slow down the desorption-decomposition process. Stirring the
solution could also retard this process by increasing trans-
port of oxalate ion to the interface. However, the magnetic
stirrer frequency coupled with the chopper frequency. This
resulted in noise which obscurred the photocurrent response.

The actual composition of the retained or tightly-bound
layers on the electrodes is not known. Assuming the layers
are essentially the same in the electrochemical and photo-

chemical studies, ESCA studies appear to indicate the presence

58

of an electrochemically active species. It is possible the
retained layers are not de-metalled phthalocyanine, but a
decomposed form of CuPc; this decomposed form is electroch-

emically active, i.e. contains double bonds.
ggyalent Attachment Studies

CuPc was attached to SnO2 via sulfonamide formation
using gamma-aminopropyl-triethoxysilane or N,beta-aminoethyl-
gamma-aminopropyl-trimethoxysilane and the sulfonyl chloride
form of the dye (Figure 3). The sulfonamide procedure in-
volves loss of HCl and the formation of a nitrogen-sulfur
bond. Attachment via a thiol formation using mercapto-
prOpyl-trimethoxysilane and tetraiodated CuPc (Figure 3) was
unsuccessful because the dye was insoluble in aqueous media.
The reaction should involve loss of HI and the formation of
a carbon-sulfur bond between the dye and silane. Since
CuPth was soluble in pyridine, either it or water/pyridine
mixtures were also tried without success. The thiol form-
ation has previously worked only in pH 7 buffer with water
soluble dyes (58).

Spectrophotometric Studies of Covalently-Attached CuPc

An absorption spectrum of CuPc covalently-attached to
SnO2 is shown in Figure 17. Absorption maxima were at
620nm and 69hnm corresponding to those of the dye sublimed
on Sn02. Absorption due to the dye (0.02h) indicated approx-

imately 2.h5 x 1013 dye molecules present on the surface

59

CuPc--Sn0

 

 

 

2
A
b
s
o
r
b
a
n
c
e
Sn02
_i l
T l
400 550 750

Wavelength tnm)

Figure 17. Absorption spectra of clean Sn0 and covalently-

attached CuPc-SnO2 electrodes.

60
assuming a dye diameter of 1h.3 A0, extinction coefficient

-1 (27), and a 0.5 cm2 sample area. If a

2

of 3.h x 10uM-1cm
SnO2 molecule area was 30 A0 and the dye was arranged
perpendicular to the surface, there should be, at most,
eight dye molecules for every 550 A02 of surface, or approx-
imately 7.35 x 1013 molecules/0.5 cm2. Approximately 1.03

x 1013 molecules/0.5 cm2 would be present if the dye mole-
cules were arranged parallel with the surface. The actual
dye arrangement is probably some combination of the above
mentioned. The extent of silane coverage, number of active

dye attachment sites, and steric hinderance limit the dye

coverage.

Electrochemistry of Covalently-Attached CuPc-SnO2

Electrodes

The cyclic voltammetry of a chemically-modified elec-
trode in DMSO is shown in Figure 18. Scanning cathodically
(cyclic (b)), three reduction waves were abserved at -0.55
volts, -1.25 volts, and -1.80 volts vs AgRE. The peaks
appeared chemically irreversible with three anodic waves
at -0.58 volts, -0.98 volts, and -1.6 volts corresponding
to oxidations of the reduction products. An irreversible
oxidation of the dye occurred at 1.h volts vs AgRE. Stir-
ring the electrolyte solution then scanning cathodically
again, resulted in cyclic (c); two poorly resolved reduc-
tions and an oxidation of the dye were observed. These
waves maintained constant magnitude indicating the presence

of an electrochemically active species.

61

Figure 18. Cyclic voltammetry of covalently-attached CuPc-

SnO2 electrode in O'1flL TEAP/DMSO.

 

62

l-—-—-l 500mv

I :Hul/cm2

   

38mV/sec
(a)

7.4mV/sec

(b)

38mv/sec

(c)

63
Cyclic (c) of the covalently-attached dye is quite

similar to cyclic (c) of the adsorbed dye (Figure 12). It
is conceivable the chemically irreversible waves noted are
really the catalytic oxidation and reduction of molecular
oxygen since it is well known the phthalocyanines are good
catalysts (59). The presence of 10'uM - 10'6M oxygen would
be enough for the adsorbed or covalently-attached dye to
catalyze its electrochemical reduction and oxidation. It
is also possible the cyclic corresponds to the oxidation
and reduction of a decomposed form of CuPc.

The above spectrophotometric and electrochemical
results do not totally confirm the success of the covalent
attachment. Therefore, unsilanized SnOZ and TiO2 were used
as controls to insure the results were not those of adsorbed
dye. Unsilanized SnO2 was allowed to react with the sulf-
onyl chloride form of the dye. It was then extracted in
benzene and water in the manner the derivatized electrodes
were. Spectrophotometric studies indicated a clean SnO2
surface; cyclic voltammetry of the electrodes also confirmed
the absence of any adsorbed dye (Figure 18 (a)).

It is conceivable that the silanized surface of the
SnO2 could enhance adsorption. However, if the dye was
adsorbed, extraction with water would decompose the sulfonyl
chloride sites on the CuPc to form tetrasulfonated CuPc

which is soluble. Therefore, the dye is most probably

covalently-attached to the semiconductor electrode.

64

Photocurrent Response

 

The photocurrent response vs anodic bias potential

plot of a sulfonamide-linked CuPc-SnO electrode in 0.05M

2
sodium oxalate/pH 7 buffer is shown in Figure 19. The re-
sponse appears to be potential dependent with its initia-
tion beginning at approximately 0.5 volts vs Ag/AgCl ref-
erence electrode. The photocurrent at any given anodic
potential was steady over a period of several minutes, and

is comparable to the photocurrent response of the adsorbed
dye electrode after the desorption/decomposition process

is complete. The covalently-attached dye electrode response
was approximately one-fourth the final adsorbed dye electrode
photocurrent response. Further conclusions concerning the

stability and photocurrent response of the covalently-at-

tached dye can not be made without more experimentation.

65

Figure 19. Photocurrent response vs anodic bias potential
of a covalently-attached CuPc-SnO electrode in

0.05M Na2020u/pH 7 buffer.

2

 

 

 

NT

 

«I.

1.0

I
0“”

66

LOCKIN SENSITIVITY lfllluV

IR FILTER
UV FILTER
LP 47 FILTER
LP 54 FILTER

J_
j

.6 ,4
Bias (volt: s)

1)-

0
ID"

P.c.
(volt 3)

b6
.

 

CHAPTER IV

SUGGESTIONS FOR FUTURE WORK

68

Future work would consist of several closely related
investigations. Better methods of covalent-attachment
to increase the phthalocyanine surface coverage would be
studied. Efficiency of the photosensitizing process could
be increased by using electrodes with several monolayers
of covalently-attached dye. This could be accomplished
by two methods: (1) coupling of phthalocyanine molecules
with cyanuric chloride (60), or (2) coupling of a silicon
phthalocyanine derivative with itself (61,62).

Extensive photocurrent studies of the adsorbed and
covalently-attached dye electrodes would be carried out.
Use of a pulse dye laser would increase the light inten-
sity and photocurrents. Photocurrent response studies
using different supersensitizing agents, light aging
studies with the chemically-modified electrodes biased
at different potentials, and ESCA studies of electrodes
in various states of electrochemical and photocurrent
treatment would aid in a better understanding of the
chemical stability of the dye and kinetics of the photo-

sensitizing process.

"LI ST OF REFERENCES"

10.
11.

12.

130

1h.

70

"LIST OF REFERENCES"

M.S. Wrighton, D.S. Gimley, P.T. Wolczonski, A.B. Ellis,
D.L. Morse, and A. Linz, Proc. Nat'l Acd. Sci., U.S.A.,

1a, 1518 (1975).

H. Gerischer, Photochemistry and Photobiology, 16, 2h3
(1977). __

S. Mishitsuka and K. Tamsru, J. Chem. Soc., London, 13,
705 (1977). -

T. Osa and M. Fujihira, Nature, 26h, 3R9 (1976).

D. Hawn and N.R. Armstrong, unpublished results (man-
uscript in preparation).

T. Osa and T. Kuwana, J. Electroanal. Chem., 22, 389
(1969). -
V.A. Myamlin and Y.V. Pleskov, "Electrochemistry of

Semiconductors," Plenum Press, New York, 1967.

S. N. Frank and A.J. Band, J. Amer. Chem. Soc., _1, 7h27
(1975)-

K.G. Allum, R. Hancock, and P. Robinson, J. Opganomet.

G.B. Harper, Anal. Chem., g1, 3h8 (1975).
C.M. Elliott and R.W. Murray, Anal. Chem.,gg, 12h? (1976).

N.R. Armstrong, A.W.C. Lin, M. Fujihara, and T. Kuwana,
Anal. Chem., pg, 7h1 (1976).

P. R. Moser, L. Wier, and R. W. Murray, Anal. Chem., Q_,
1882 (1975)

P.R. Moser and R.W. Murray, J. Amer. Chem. Soc., in press.

71

15. B.F. Watkins, J.R. Behling, E. Koriv, and L.L. Miller,
J. Amer. Chem. Soc., 31, 3su9 (1975).

16. D.G. Davis and R.W. Murray, Anal. Chem., 82, 19h (1977).

17. K. Siegbahn, C. Vordling and A. Fahlmann, et al., "Elec-
tron Spectroscopy for Chemical Analysis," Almquist
and Wilkells, Uppsala, Sweden, 1967.

18. K. Siegbahn, C. Vordling, K. Hamru, and J. Redman, et al.,
"ESCA-Atomic, Molecular and Solid State Structures
Studied by Means of Electron Spectroscopy," Almquist
and Wilksells, Uppsada, Sweden, 1967.

19. Kane and Larrabee, "Characterization of Solid Surfaces,"
Plenum Press, New York, 197R.

20. J.T. Yates, Chem. and Eng. News, Aug. 26, (197k).
21. J.T. Yates and T.E. Madey, Surface Sci., 8;, 257 (197A).

22. K.S. Kim, T.J. O'Leary, and N. Winograd, Anal. Chem.,

23. M.J. Dressers and T.E. Madey, Surface Sci., 88. 533 (197h).

2h. A.W.C. Lin, N.R. Armstrong, and T. Kuwana, Anal. Chem.,
in press.

25. J.M. Bolts and M.S. Wrighton, J. Phys. Chem., 89, 26h1
(1976 . -

26. R. Memming, Photochem. and Photobiology, 18, 325 (1972).

27. F.H. Moser and A.L. Thomas, "Phthalocyanine Compounds,"
ACS Monograph Series, No. 157 (1963).

28. D.H. Busch, J.R. Weber, D.H. Williams and N.J. Ross,
J. Amer. Chem. Soc., 88, 5161 (196k).

29. E.?. ggpia and F.D. Verderame, J. Chem. Phys., 8, 267R
19 . “

30. L.D. Rollman and R.T. Iwamot, J. Amer. Chem. Soc., 28,

31. A.N. Sidorov, Russian J. Structural Chem., 18, 255 (1973).

33.

3h-

35.

36.

37.

38.

39.

h0.
M1.

1.2.
43.

AS.

h6.

h7.

R8.

72

E.A. Rodyushkuna, B.D. Bereyin, and N.R. Tarasevich,
Elektrokhimiya, 2, R10 (1973).

Y.M. Stolovitskii, A.Y. Shkuropatov, and V.V. Shichhov,
Biofizika, 12, 820 (197A).

V.S. Shumov and M. Heyrovsky, J. Electranal. Chem., 88,
h69 (1975). -—

G.A. Alferov and V.I. Sevast'yanov, Elektrokhimiya,
827 (1975).

J.P. Randin, Electrochimica Acta., 19, 83 (197h).

J

D

11

T.A. Maas, M. Kuifier, and J. Zwart,
Comm., 88 (1976).

. Chem. Soc., Chem.

Y. Niwa, H. Kobayaashi, and T. Tsuchiya, J. Chem. Phys.,

M. Zeller and R.G. Hayes, J. Amer. Chem. Soc., 25, 3855
(1973)- ‘-

J.H. Weber and D.H. Busch, Inorg. Chem., 8, h69 (1965).

R.W. Higgins and C.L. Hilton, J. Org. Chem., 18, 1275
(1951). -

Japan Patent No. 19,225.

W.F. Kosoncky, S.E. Harrision and R. Stander, J. Chem.

M. Babai, N. Tshermikooski, and E. Gileadi, J. Electro-
chem. Soc., 119, 1018 (1972).

R.K. Quinn and N.R. Armstrong, J. Vac. Sci. Technol., 18,

D. Sawyer and J. Roberts, "Experimental Electrochemistry
For Chemists," John Wiliz and Sons, New York, 197k.

Y.S. Shumov, V.I. Sevast'yanov, and 0.0. Komissarov,
Biofizika, 18, RB (1973).

A. Vogel, "Practical Organic Chemistry," 2nd ed., Wiley,
New York, 1966.

R9.

50.
51.

52.

53.
5h.
55.

56.

S7.
58.
S9.

60.

61.
62.

63.

73
P.A. Grutsch, M.V. Zeller, and T.P. Fehlner, Inorg.
Chem., 2’ 11431 (1973).

N.R. Armstrong and T.V. Atkison, private communication.

K. Bernauer and S. Fallab, Helvetica Chemica Acta, 88,
1287 (1962).

D.W. Clac, J.B. Raynor, L.P. Stodulski, and N.R. Symons,
J. Chem. Soc. A, 997, (1969).

W.P. Gomes and F. Cardon, Z. ngs. Chem., 88, 333 (1973).

N.R. Armstrong, private communication.

T. Novakov and R. Prins, in "Electron Spectroscopy,"
Proc. Intern., Conf. Held at Asilomar, Calif. 7-10
Sept. 1971, Ed. D.A. Shirley, Worth-Holland, Amster-
dam, 1972.

P.A. Larson, J. Elec. Spect. and Related Phenom., 8,
213 (197h)o

G. Schon, Surface Science, 35, 96 (1973).
8.8. Hasenoff, Cand. J. Biochem., 82, 7h2 (1971).

J. .R. Go§denstein and A.C. C. Tseung, Nature, 222, 869
1971 -——

K. Venkatasaman, " The Chemistry of Synthetic Dyes,"
Academic Press, New York, 1972.

Japan Patent 5,711.

J.N. Esposito, J. E. Lloyd, and M.E. Kenney, Inorg. Chem.,
5. 1979 (1966).

R.S. Kirk, Molecular Crystals, 5, 211 (1968).

l
Ill-II
|
Iii
Ii!
i
III
I