I???“ .33: ”4‘“!qu

I I "4 '-' -. J's}31" '1 .9 .4 I: he:

I ' *l'i‘IIIIIl34fi-Iu‘mi3rg, 3 2:;
‘ ‘ ‘ . : :J .'.' . .“hfi _‘_.‘.

, 2:30}: :I7_‘;,\.,' ‘2' .34.: 7 .

     
     
          

   
   
  

    

am ,A_
-N

.r__ .,':.~1§

‘:hII: : 3" 535; '2’: 41 .4
.. ’4‘: 4. “3a.;
Fig-II flzr'gflflge 83,133?
‘43“:4433433

r, . 43,, , .93:
' 1. . < :3 4I ~ "1
n “N L ' $3, firfig‘gn 31f x1331}
' I‘ I {I‘
~ 4:15 ~ $44454

. '1
Strl‘gznxn ;
‘.....r

‘

‘ '1’.‘

34,5 . 8 ,
533‘"! 3:25." I?
.23.

1 Y‘.‘ I q
r“ . ' q
‘1

    

3;.
v4,

 

   
 
   

       

__.‘
= h p-
p 5‘. '

N
‘—- 12‘...
‘- __>‘-—-.~
. . .

.4».

‘.....
«1 ’
u.

   
 
   
  

. "'..—‘..:

  
 
   
  
  
 

‘.." ‘4 __. -
d... _
- ~v-,_u

‘ _~.—~—
-1—‘v-

—- .
.~_
”'..:—

‘ ow
»
. _:I7
.
... 37—93},
L.
0.5....“

—I ’5...—

:M.‘
. -H-i

  
 

,-
“4.— _1 _
~.....- . v
.. -
c
a»

I}, ' ,Ie.
IIII 33:“ I'I III‘III IIIIIII
:33I:3:336I: 333‘” 3:: I

“PM:
.' “t I :9 M-‘II: :I'I.I 'I
~ 3 :::'Is=:=:::444IN II”
I II ':"I 334:.
2‘

 
 
 
 
    

  

     
 
    

~r ‘---..
v.4”
.-.._,
p 0‘

  

     
      
 
   
 

<o-‘o—
g... H
M..." ‘
‘—-—~‘ 5 .

'n‘
M.

  

Q
“:3:
‘~.

    
  

 

    

 
  

1'
§Ilv', H:-
”3. - . IIIII

In
I“-

    
    
   

.lym {11?}:
:"IIQ‘I?
':HIIf,

 
  
    

VIII] 0',” -----
f u I h a l l 'J
Ia’I' " III‘I'I’I’II I”? :II:IJII:»IEIIMIIRI
H . ..--7,:'I:,.o I 3:11,,3:,,,y'”_
, NJ: 1 .4'
, I401“: £33,.” if? SIIW‘E “in“ “433qu I‘II‘LI é‘N ‘15.}, Ifl!‘ ”3|
L 34.0,

: ' ~E.,' -.}',“‘, I I.
, ,:: 3,34,233‘J'2IQ‘ .333, 3 “:4 :"II*I"‘::’I‘1:III:‘III IIII3III IIIIIIIJ
- . I,:.Ih,:,,:~,.3,»,~’. ‘I III :II"I'I “3IIIYII3Ifn ' “II III

63th

  
       
  

   

, - . . ~ :4 : . :w I
raw . ’ .' r" (1’. ‘; '~:‘:'~:- HM: MI
u 3 : ,.- , :r- .v ,3:,-:,: : r$ :
: : E'I ,F,I'|1,:,:I 'l,,‘J "‘.II I ,,,,,-'i.l,' 33L;,I., IIIJHI: It I f:I :I‘ntfiuufi :

I' In!‘- I‘ II III: ""III 'I “1‘ .‘ -. ' '

. ~ I I : II : IIIII: 3- . KHZ!» IVE}? I If I ... L

I II: ,,,:Ib‘,II \I III ,III,‘ ,‘I'II I, I I,III II" (/1 VIII .III ‘.. I: IIIlI ”Illj‘jrf M I I '
IIII I“ I IIIII III III III, II I": I'I I : " III, II :IIIHIEJI’HILISJ lI IIIIIIIW) l M" l Iih MI l ‘ I" "I I I

tun-2818

1."

=5"
.

W¢~Ifiwmw Fan-aw -§ g

L g +:\ r ;R ‘1!” 33.1:
, 1.311 1 4“ vi:

5! ' .‘ °

v/l ( “‘ fl ~ c‘_ h

“Z~ afi';£?.m UKA“|}’§

' c.
‘ - . , k. . ‘
W ' Ii “5' {:1 12.; f r

t" id i“ s- ’1' hummr -am~r

l
n '%
I:

This is to certify that the

dissertation entitled

Surface Raman Spectra of tvl,2-bis(4-pyridyl)
ethylene and Amino Acids at a
Polycrystalline Silver Electrode

presented by

John J. McMahon

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Chemistry
\ _ ‘

Mffluzz

Major professor

Date (ll/N /y/_

 

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

 

 

 

RETURNING MATERIALS:
)V153I_J Place in book drop to ‘
LJBRAfiJES remove this checkout from
.AIIIIBIIIIL your record. FINES will

 

 

be charged if book is
returned after the date
stamped below.

 

 

 

‘l/yvv

SURFACE RAMAN SPECTRA OF T-l,2-BIS(4-PYRIDYL)ETHYLENE
AND AMINO ACIDS AT A POLYCRYSTALLINE SILVER ELECTRODE

BY

John J. McMahon

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1981

ABSTRACT

SURFACE RAMAN SPECTRA OF T-l,2-BIS(4-PYRIDYL)ETHYLENE
AND AMINO ACIDS AT A POLYCRYSTALLINE SILVER ELECTRODE

BY

John J. McMahon

The Raman spectrum of t-l,2-bis(4-pyridyl)ethylene
(t-BPE) adsorbed at a polycrystalline silver electrode is
reported. Assignments of vibrational modes were based
mainly on comparisons with the normal Raman spectra of
t-BPE, its dihydrochloride and its dideuterochloride
(also reported). The infrared spectrum of t-BPE and
the vibrational spectra of trans stilbene and substituted
pyridines were also utilized in the assigment of t-BPE
surface Raman lines. A short history of surface-enhanced
Raman scattering including examination of several of the
proposed enhancement mechanisms is presented. Spectro—
scopic evidence for surface complex formation between
t-BPE and the silver electrode is examined. The possibi-
lity of a resonance Raman enhancement mechanism resulting
from perturbation of the electronic energy state distribu-
tion upon complexation is discussed. The influence of
the metal conduction electron image field on the symmetry
of the scattering system is theoretically scrutinized.
Molecular reorientation of t-BPE molecules is observed

in the surface Raman spectrum under certain conditions.

John J. McMahon

The conditions for molecular reorientation at the
electrode surface are examined and a mechanism for the
rearrangement is proposed. The excitation profile of
t-BPE on silver is reported and demonstrates structure

in the red region and at other wavelengths. Perturbation
in the electronic state distribution of t-BPE upon
complexation to the silver surface is invoked to account
for the detailed structureixlthe excitation profile.
Spectroscopic features, characteristic of the monOproto-
nated bipyridine, are observed in the surface Raman
spectrum of t-BPE at pH 5.6. Photodegradation of
multilayers of t-BPE on silver is also observed. Surface
enhanced Raman scattering from L-glycine and L-leucine

is reported and suggestions of molecular orientation are

made.

TO CAITLIN

ii

tC

pe

th

re

9U

CO

ti

ch

in

Hu:

St;

um

ACKNOWLEDGEMENTS

I wish to extend particularly grateful appreciation
to Professor Gerald T. Babcock. Professor Babcock
permitted experiments on surface Raman scattering to
proceed utilizing his spectroscopic equipment, though
the studies had little direct connection to his own
research. Professor Babcock also provided considerable
guidance in both theory and experiment without which
completion of this thesis would have been impossible.

The assistance of Kendall L. Guyer in the construc-
tion of a suitable electrochemical cell,in the choice
of t-BPE (the ideal adsorbate), and in general electro-
chemical discussions is gratefully acknowledged.

The efforts of Beverly Adams and Margaret Lynch
in the preparation of this thesis are greatly appreciated.

Scientific discussions with Professors Katharine C.
Hunt, George E. Leroi and Michael J. Weaver often
stimulated my research and are acknowledged.

Special thanks to Rosemary and Caitlin for their

unwavering support.

iii

TABLE OF CONTENTS

Chapter Page

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . Vii

LIST OF FIGURES. . . . . . . . . . . . . . . . . . . . .ix

I.

II.

III.

IV.

Introduction. . . . . . . . . . . . . . . . . . . . 1

History of Surface Enhanced Raman Scattering
Scattering (SERS). . . . . . . . . . . . . . . . . .6

Theory of Resonance Raman Scattering from
Surface Complexes. . . . . . . . . . . . . . . . . 16

Experimental. . . . . . . . . . . . . . . . . . . .34
Raman Spectrometer. . . . . . . . . . . . . . . . .34
Samples. . . . . . . . . . . . . . . . . . . . . . 34
Lasers. . . . . . . . . . . . . . . . . . . . . . .37
Excitation Profiles. . . . . . . . . . . . . . . . 40
Laser Power Measurement. . . . . . . . . . . . . . 43
Visible Absorption Spectra. . . . . . . . . . . . .44

Sample Preparation: t-l,2-bis(4-pyridyl)ethylene
(t-BPE) O O O I O O O O O O I O O O O O O O O O O O 44

t-l,2-bis(4-pyridyl)ethylene dihydrochloride and
dideuterochloride. . . . . . . . . . . . . . . . . 44

Silver Complex of t-l,2-bis(4-pyridy1)ethy1ene
(Agn (t-BPE)mo o o o o o o o o o o o o o o o o o o o 45

Amino Acid Solutions. . . . . . . . . . . . . . . .46

AgOH, A920. . . . . . . . . . . . . . . . . . . . .46

iv

Chapter Page

V.

VI.

Polycrystalline and Solution Raman Spectra
of t-l,2-bis(4-pyridyl)ethy1ene, its

dihydrochloride and its dideuterochloride. . . . . 48
Symmetry Considerations. . . . . . . . . . . . . . 48
Frequency Region 0-600 cm_l. . . . . . . . . . . . 58

Vibrations 6a and 6b. . . . . . . . . . . . . . .. 58

Out-of—Plane Vibrations. . . . . . . . . . . . . . 61
Ring "Breathing" modes v1 and v12. . . . . . . . . 63
Frequency Region 1050-1300 cm-l. . . . . . . . . . 65
Frequency Region 1300-1400 cm-l. . . . . . . . . . 67

Vibration Pair 19a and 19b. . . . . . . . .. . . . 68
Vibration Pair 8a and 8b. . . . . . . . . . . . .. 69
Ethylenic C=C Stretch. . . . . . . . . . . . . . . 70
C-H, N-H(D), Stretching Motions. . . . . . . . . . 71

Raman Spectra of t-l,2-bis(4-pyridyl)ethylene
adsorbed at a polycrystalline silver electrode. . .73

Heavy Coverage Raman Spectra of t-BPE under
5145 A Excitation. . . . . . . . . . . . . . . . . 74

Molecular Reorientation on Complexation of t-BPE
at the Electrode Surface. . . . . . . . . . . . . .85

Excitation Profile of t-BPE on Silver. . . . . . . 98

Light Coverage Studies and the Raman Spectrum of
an Ag (t-BPE) Complex. . . . . . . . . . . . . . .110

Negative Potential Results and the Reduction
Product of t-BPE. . . . . . . . . . . . . . . . . 121

Monoprotonated t-BPE on Silver. . . . . . . . . . 133

Influence of C104- on the Surface Spectrum of
t-BPE. o o o o o o o o o o o o o o o o o o o o o o 141

Photodegradation of t-BPE Multilayers on Silver.. 148

Chapter Page

VII. Surface Enhanced Raman Spectra of Amino Acids
and Future Studies. .. . . . . . . . . . . . . . .155

Future Surface Raman Studies. . . . . . . . . . . 167
APPENDIX. C O O I C O O O O O O I O O O O O O O O O O O 170

LISTOFREFERENCES...................l72

vi

Table

1.

2a.

LIST OF TABLES

Page

Raman Spectra of t-l,2-bis(4-pyridyl)

ethylene, t-l,2-bis(4-pyridyl)ethy1ene
dihydrochloride and t-l,2-bis(4-pyridyl)

ethylene dideuterochloride. Excitation:

6471 A..... ........... ...... ................... 49
Correlation of symmetries of the pyridine

rings and ethylene to the molecular symmetry

of t-1,2-bis(4—pyridyl)ethy1ene and further
correlation to possible surface complex
symmetries ..... ....... ..... ....................56
Comparison of Raman frequencies of modes 6a

and 6b in representative molecules. Vibrational
frequencies are in wavenumbers (cm-l) ..... .....62
Raman Spectra of t-l,2-bis(4-pyridyl)ethylene
adsorbed onto a polycrystalline silver

electrode under 5145 A excitation. Comparison
made with polycrystalline t-l,2-bis(4-
pyridy1)ethylene ........ . ...................... 77
Excitation Profile of t-l,2-bis(4-pyridy1)
ethylene on silver at -600 mV versus sce
(corrected for w4)............................102
Potential Dependent Excitation Profile of the

655 cm-1 line (6b) in the Raman spectrum of
t-1,2-bis(4-pyridyl)ethylene on silver at

-600 and -400 mV versus sce ................... 108

vii

Table

10.

Page

Raman Spectrum of a polycrystalline silver
complex of t-1,2-bis(4-pyridyl)ethylene and
comparisons to the Raman Spectrum of
t-1,2-bis(4-pyridyl)ethy1ene on silver (light
coverage) at -400 mV sce and to the
polycrystalline t-BPE Raman Spectrum .......... 116
Raman Spectrum of the two-electron reduction
product of t-1,2-bis(4-pyridyl)ethylene on

silver at -l.2 v versus sce. Excitation:

5145 A........................................129
Raman Spectrum of Monoprotonated t-l,2-bis(4-
pyridyl)ethylene on silver at -50 mV and

-600 mV at pH 5.6 and at -600 mV after

base (OH_) is added. Excitation: 5145 A.
Region: 1000-1700 cm"1 .......... . ............ 136
Raman Spectra of L-Glycine and L-Leucine

adsorbed onto a polycrystalline silver

electrode at -600 mV sce, in the frequency
region 550-1700 cm-l. Excitation: 5145 A....l61
Raman Spectra of Polycrystalline AgOH and

A920 in the region 200-1600 cm.1 using a

back scattering geometry......................164

viii

Figure

LIST OF FIGURES

Page

Energy State Perturbation in t-l,2-bis(4-
pyridyl)ethylene upon Complexation to

RuII (NH3)5. The point groups for the

ligand complex and Ru (NH3)5 are given.

Ru(II) has a d6 electron configuration

and thus the lowest three levels in the

energy level diagram for the complex are

occupied ......... . ............................. 19
Surface Selection Rule Formalism. R is the
operation connecting the adsorbate dipole

with its conduction electron image. P is

an irreducible representation..................22
Experimental arrangement in the SERS experi-
ment. R = reference electrode (Ag/AgCl);

W = working electrode (polycrystalline

silver); C counter electrode (platinum
wire); FL = focusing lens; CL = collection

lens; = angle of incidence relative

¢inc
to the surface normal......... ..... ... ......... 38
Normal Raman spectrum of teflon and surface

Raman spectrum of t-1,2-bis(4—pyridyl)

ethylene on silver with teflon standard ..... ...41
Normal Raman Spectra of polycrystalline
t-l,2-bis(4-pyridy1)ethylene, its dihydro-
chloride and dideuterochloride. Excitation
Wavelength: 6471 A (Kr+) ............ . ......... 51

ix

Figure

6.

10.

Page

Normal Raman Spectra of t-l,2-bis(4-
pyridyl)ethylene in CH2C12, the aqueous

solution oftfluadihydrochloride, and the
dideuterochloride in D20. Excitation

Wavelength: 6471 A (Kr+)................ ...... 53
Heavy Coverage Potential Dependence of the
Surface Raman Spectrum of t-l,2-bis(4-
pyridyl)ethylene on Silver. Compared are:

the normal Raman spectrum of polycrystalline
t-BPE (bottom spectrum); the surface Raman
spectrum recorded at -600 mV immediately
following the oxidation-reduction cycle

(above polycrystalline spectrum); surface
spectrum at -50 mV; and the surface spectrum
recorded on returning to —600 mV (top

spectrum). Excitation Wavelength:

5145 A (Ar+) ........ . .......................... 75
Potential Dependence of Mode 10 in the

Surface Raman spectrum of t-BPE "flat" on the
silver surface. Excitation Wavelength:

5145 A (Ar+).. ...... ................... ........ 89
The Visible Absorption Spectra of I. t-1,2-
bis(4-pyridy1)ethylene and II. t-l,2-bis(4-
pyridyl)ethylene dihydrochloride....... ..... ...93
Excitation Profile at t-l,2-bis(4-pyridyl)
ethylene on silver at -600 mV sce in the

end-on configuration (Rhodamine 6G excita-

tion freqeuncy range). Data points are
normalized to the 935 cm.1 line of the
perchlorate internal standard and corrected

for the v-4 dependence of the normal Raman

cross section.... .............. ......... ....... 99

Figure

11.

12.

13.

14.

15.

16.

17.

18.

Page

Potential Dependence of the Excitation

Profile of the 655 cm"1 (613) Line in the
Surface Raman Spectrum of t-BPE on

Silver.. .......... . ..... . ..................... 106
Potential Dependence of the Surface Raman
Spectrum of a Light Coverage of t-BPE on

Silver. Excitation Wavelength: 6578 A

(Dye Laser Emission).

Normal Raman Spectrum of a Polycrystalline
Silver Complex at t-BPE (Agn(t-BPE)m).

Excitation Wavelength: 6471 A (Kr+) .......... 114
Potential Dependence of the Surface Raman

Spectra of t-BPE Potentials -400 mV, -600 mV,
-800 mV and -1.0 V versus sce. Excitation
Wavelength: 5145 A (Ar+) ..................... 122
Surface Raman Spectrum of the Two-Electron
Reduction Product of t-BPE on Silver at

—l.2 V versus sce. Excitation Wavelength:

5145 A (Ar+)................. ................. 127
Excitation Wavelength Dependence of the

Surface Raman Spectrum of Monoprotonated

t—BPE on Silver at -50 mV (sce). Electrolyte

pH: 5.6 ...................................... 134
Surface Raman Spectrum of Monoprotonated t-BPE

on Silver at -50 mV (see), -600 mV (sce),

and at -600 mV (sce) after Addition of Base
(OH-l. Excitation Wavelength: 5145 A

(Ar+)..... ............ . ....................... 139
Surface Raman Spectrum of Monoprotonated t-BPE

on Silver at -50 mV (sce) in the "flat"
Configuration. Excitation Wavelength:

5145 A (Ar+) ........... .. ..... ... ............. 142

xi

Figure

19.

20.

21.

22.

23.

Page

Influence of 0.5 M C104 on the Observed

Surface Raman Spectrum of t-BPE "flat" on the
Silver Surface at -600 mV (sce). Excitation
Wavelength: 5145 A (Ar+).....................145
Excitation Wavelength Dependence of Mode

11 (683 cm-1) Vibration t-BPE on Silver at

-400 mV versus sce...................... ...... 149
Photodegradation of Multilayers of
t-l,2-bis(4-pyridyl)ethy1ene on Silver: Loss

of Intensity at 684 cm-1 ...... ................151
Surface Raman Spectra of L-leucine and
L-glycinecx1Si1ver at -600 mV (sce). Excita-
tion Wavelength: 5145 A (Ar+)................159
Normal Raman Spectra of Polycrystalline A920

and AgOH Utilizing a Back-scattering

Geometry. Excitation Wavelengths: 5145 A

(Ar+) AgOH; 6471 A (Kr+) A920.. ............... 162

xii

LI ST OF ABBREVIATIONS

ABBREVIATIONS

FWHH: full width at half height
Ce: ethylenic carbon

IP: in plane

R: alkyl substituent

sym. str: symmetric stretch

HOMO: highest occupied molecular orbital
p: depolarization ratio

p: polarized mode

dp: depolarized mode

xiii

CHAPTER I

Introduction

Over the past three to four years there has been an
explosion of experimental and theoretical studies of Raman
scattering from metal surface adsorbates. Most of the
experimental efforts have concentrated on the observation
of enhanced Raman signals from pyridine at a silver
electrode. The emphasis of these studies has been on the
unmasking of the enhancement mechanisms which are apparently
unique to the surface Raman studies.

Many fundamental questions continue to elude adequate
explanation. For example, it is well understood that the
long mean free path metal conduction electrons arrange to
form an image dipole of a vibrating molecule which is
adsorbed at the metal surface. However, the influence of
this image field on the enhancement mechanism and on the
Raman scattering selection rules for surface molecules has
not been convincingly decided. Similarly, the question
of whether a surface complex between metal and adsorbate
is formed in the adsorption process remains unresolved.
Many efforts to observe metal-ligand bond stretches in the

surface Raman spectrum have failed.

sur
fie
the
knc
mer

thc

the
sig
dyx
met
km

de:

UR.
prl
Cr:
Vi<
L11:
ti,-

C0:

These points are fundamental to an understanding of
surface Raman intensities since the presence of an image
field or a surface complex or both may significantly alter
the symmetry of the light scattering species. Without
knowledge of the symmetry of the scattering system, assign-
ment of surface Raman spectra and interpretations based on
those assignments become meaningless.

However, if accurate conclusions could be drawn about
the symmetry of the adsorbates at a metal surface then
significant far-reaching advances into the orientation,
dynamics, and electrodynamics of molecules adsorbed at
metal electrode surfaces would be forthcoming. For example,
knowledge of adsorbate orientation can lead to more atomic
descriptions of the double layer region. The structure of
oxidation and reduction products observed in the surface
Raman spectrum would also be discernable. The effect of
symmetry distortions on the nature and wavelength of
electronic transitions in the adsorbate could be concluded
with the assistance of excitation profiles.

Raman studies of electrode-electrolyte interfaces are
unique to surface spectroscopies in that the surface can be
probed through solution owing to the small light scattering
cross section of water. Electronic methods such as ultra-
violet photoelectron spectroscopy, LEED, and Auger require
ultrahigh vacuum conditions. In solution, electrode poten-
tials may be maintained providing a new, researcher-

controlled parameter. The ability to vary electrode

potential continuously while monitoring structure-dependent
Raman spectra should lead to new insights into the effect
of electron transfer on molecular structure.

Continued investigations are certainly warranted and
in this thesis some of the above mentioned points are
addressed and experimentally examined. The influence of
the image field on the "molecular" symmetry probed in the
surface Raman experiment is scrutinized theoretically.
Electronic and symmetry perturbations experienced upon
complexation are discussed. Evidence of complexation
from Raman frequencies in regions well above the low
frequency metal-ligand stretching region are presented.

The system studied most extensively in this thesis is
that of trans-1,2-bis(4-pyridyllethylene (t-BPE) adsorbed
at a polycrystalline silver electrode. In order to obtain
an understanding of the surface spectra the normal Raman
spectra of polycrystalline and dissolved t-BPE were
recorded. The Raman spectrum of the dipyridine has not
been previously reported in the literature although the
infrared spectra of the cis and trans isomers were obtained
by Katsumoto.l The dihydrochloride and dideuterochloride
were prepared and the normal Raman spectra recorded to
assist in the assignment of vibrational modes.

Trans-1,2-bis(4-pyridyl)ethylene is a good choice of
adsorbate for several reasons. The extensive conjugation

of the bipyridyl ethylene produces a large normal Raman

scattering cross section facilitating experimental obser-
vation. The ligand offers an ethylene n system and two
nitrogens through which binding to the silver surface may
occur. The molecular symmetry is C2h which requires mutual
exclusion of Raman and infrared activities. Reductions in
symmetry upon complexation to the metal surface should be
recognizable with the appearance of infrared modes in the
surface Raman spectrum. The hydrophobicity of t-BPE is
demonstrated by its minimal solubility in aqueous solution;
saturation is reached at approximately 3 millimolar concen-
trations. It is likely that this hydrophobicity assists
in the adsorption of t-BPE to the silver electrode surface.
Intense surface Raman scattering can be observed from
molecules adsorbed from 250 nanomolar concentrations.
t-BPE also has two protonation sites and is electroactive,
undergoing a two electron reduction at -l.1 V versus SCE
(at neutral pH). Both of these capabilities can be
exploited in an accurate determination of surface coverage.
Finally, it is hoped that information received from
studies of the enhancement mechanism and surface orientation
can be extrapolated to the investigation of other molecules
at the electrode surface. Raman scattering from small
amino acids adsorbed on silver are recorded here. Amino
acid surface studies were undertaken so that results may
be applied to the study of large, biologically important
molecules at metal surfaces. The enhanced light scattering

from the amino acid residues of a protein molecule which

lie c1
of seq
Observ
by ads
fluore
metal

radiat
Raman

the pr

the ex

lie closest to the metal surface may lead to conclusions

of sequence and primary, secondary, and tertiary structure.
Observation of light scattering from proteins is facilitated
by adsorption to a metal surface owing to the lack of
fluorescent background. Through energy transfer to the
metal surface, the biological molecule gains a non-
radiative excited state decay pathway. Current resonance
Raman studies of proteins probe only the chromophore since
the protein substructures do not directly contribute to

the excited state in resonance.

a si
4

a1.

thee

subj

CHAPTER I I

History of Surface Enhanced Raman

Scattering (SERS)

Intense light scattering from pyridine adsorbed at
a silver electrode was first reported by Fleischmann et

al.2 in 1974, and in 1975 at copper e1ectrodes.3’4

In
these experiments a polycrystalline silver electrode was
subjected to a series of cyclic linear potential sweeps
between +200 and +300 mV relative to the saturated calomel
electrode for about 15 minutes. Oxidation of silver to
silver chloride and redeposition of silver metal resulted
from this process and produced an exceedingly rough surface.
Intensity enhancement in the light scattering experiment
which followed was not recognized since the increase in
the electrode surface area was large and considered
sufficient to support large numbers of adsorbed molecules.
In 1977 van Duyne and Jeanmaire5 repeated the silver
electrode experiments under stricter potentiostatic control
allowing a much smaller amount of charge to pass during
the oxidation-reduction cycle (typically 25 or 50 mC/cmz).
Therefore, van Duyne was the first to recognize incredible

intensity enhancement from pyridine adsorbed at a metal

electrode surface. An enhancement factor based.cm1estimates

of surface coverage was projected to be 4 to 6 orders of
magnitude above normal Raman scattering intensity from the
same number of molecules. Molecues which exhibit resonance
Raman scattering in solution were also examined (e.g.
methyl orange, and crystal violet) and found to yield even
higher enhancement factors up to 10 orders of magnitude.5’6
A wavelength dependence study of the pyridine on silver
experiment over the 4600 A to 6300 A region was reported
to follow the fourth power wavelength dependence of normal
Raman scattering presumably eliminating a resonance Raman
enhancement mechanism.

However, subsequent studies by other researchers7-12
on excitation profiles of pyridine at metal surfaces
revealed an increased enhancement in the red after norma-
lizing for the w4 dependence. Pettinger et a1.8 were the
first to suggest resonance Raman enhancement of vibrational
mode intensities in a surface complex of silver and
pyridine. This conclusion was reached by showing that the
differential reflectance spectrum of a silver electrode,
electrochemically cycled in an aqueous pyridine solution,
fit the excitation profile exhibiting increased enhancement
in the red. In typical resonance Raman studies the
excitation profile of Raman intensities is compared with
the visible absorption spectrum. Absorption spectroscopy
is a transmittance experiment and is therefore inapplicable

to surface probes (except for very thin metal films, i.e.,

less than 500 A silver). Differential reflectance is

therefore the supportive method of choice. A drop in

the metallic reflectivity is assumed to be accompanied by
absorption at the surface. The application of this

single model to SERS was complicated by Mie scattering
from a rough surface which may interfere with reflectivity

10'11 however, maintained that

measurements.13 Moskovits
both the dip in the reflectance spectrum and the excita-
tion profile could be fit by modeling the rough surface as
a layer of metal spheres lying on a flat substrate with
adsorbate and medium filling the gaps. This model allowed
for resulting dispersion in the metal dielectric constant
which could be fit to the data.

It rapidly became evident that several theories could
account for surface enhanced light scattering. King,

14.15 originally prOposed an image

van Duyne and Schatz
field enhancement mechanism which they later showed
accounted for the excitation profile of pyridine on copper}6
Subsequently, a University of California, Santa Barbara,
group headed by Horia Metiu extensively studied the image
dipole theory in a semi-classical theoretical treatment.

In a series of papers Metiu et al.17-23 set up a Green's
function approach to describe the field experienced by an
adsorbate at a certain distance from the metal surfaces.
Included in the terms of the field equation were the
following: a) the field Ep’ produced by the incoming light

beam in the absence of the molecule but in the presence of

the surface (contains the Fresnel reflectivity coefficients),

and b)
a dista
of th

of the

where :
dieleC'

resona.

and b) the field, Es' emanating from the induced dipole at
a distance Zm from the surface and including reflection
of the radiation from the induced dipole off the surface
of the metal. The induced dipole was derived to be

01E

_ P
u — -l _3 (l)
l-a(2ne0) (22m)

 

where a is the molecular polarizability and 80 is the
dielectric constant of the medium. The distance dependent

resonance condition was therefore
a = (2m: )(2Z )3 (2)
0 m '

Thus, for typical molecular polarizabilities the resonance
condition could only be met at very short metal-adsorbate
distance, on the order of l A. Classical treatments, of
course, breakdown in modeling such atomic dimensions and
the image field theory remained an inadequate description
of the enhancement process.

Another theory of surface enhanced Raman scattering
gained prominence with a seriescxfelegant experiments.
These were coupling of electromagnetic radiation to surface
polaritons at the electrode-electrolyte interface. M.R.
Philpott24 first predicted the effect of coupling surface
plasmons to electronic transitions in molecules in 1975- In
general, surface plasmons are solutions to Maxwell's equa-

tions for the system, metal in contact with a dielectric:

which are non-radiative. That is, the phase velocity,

pha:
in

the
pla
hav

pla

rat

pl:

501
di:

as

pa
CC

of

a}

10

V . of the surface waves is less than the speed of light

phase
in the medium, C/nd, where nd is the refractive index of
the dielectric. In order to couple effectively to surface
plasmons they must become radiative. Currently, two methods
have been proven successful in the excitation of surface
plasma waves. They are l) excitation by evanescent waves
in the attenuated or frustrated total reflection configu-
ration and 2) coupling of light to non-radiative surface
plasmons utilizing surface roughness.25

Verification of roughness induced coupling has been
found experimentally by several researcherszs"3O and
discussed by Raether.29 Uniform roughness can be considered
as a regular array of step functions such as a diffraction
grating. The Fourier transform of this array is described
by a superposition of sine functions characterized by a

1, where d is the grating

particular wave vector F; = 2nd-
constant. A rough surface may be viewed as a superposition
of such gratings. Since the momentum of the wave is given

by h(§), where ?'= P; + f;, (where fig is the wavevector of
the incoming light),the phase velocity of the surface wave
may exceed C/nd if F; is large enough (conversely, the

phase velocity of the incoming photon W(޴ + kg) becomes
reduced). Thus, the surface plasmon becomes radiative

and may couple to molecular electronic transitions. Philpott

31'32 demonstrated enhanced Raman intensities from

et a1.
thin organic films adsorbed onto holographic metal gratings

(prepared by application of two interfering laser beams

i.e. 4

of 511

methoc
in 19!
call '

of re

11

i.e. 4579 A to photoresist followed by vacuum deposition
of silver).

Excitation of surface plasma waves in silver by the
method of attenuated total reflection (ATR) was described

33 The attenuated total reflection studies

in 1968 by Otto.
call for a system in which the dielectric medium (index

of refraction nd) containing the molecules of interest is
sandwiched between the metal and a high refractive index
(np) material (prism) such that np > nd. This arrangement
is known as the Kretschmann34 configuration. When electro-
magnetic radiation traverses the prism at the correct angle
¢ (according to Snell's laws) it undergoes total reflection
at the interface with the dielectric medium. An evanescent
wave is generated at the prism-dielectric interface by
virtue of conservation of momentum constraints. The
evanescent wave propagates through the dielectric spacer
layer with a phase velocity C/np. Since np is large the
low phase velocity of the evanescent wave converts it into
a radiative surface plasmon at the metal-dielectric inter-
face (for low k values). The radiative plasmon may then
couple to electronic transitions in the molecules of the
spacer layer. Several reviews of surface plasmon coupling
have appeared in the literature (e.g. see references (25)
and (35)). Experimental verifications of enhanced Raman
scattering from adsorbates through ATR coupling of surface

plasmons abound in recent publications.36-44

12

Though coupling to surface plasmons has been proven
to enhance the Raman scattering intensities, the enhancement

4--106 reported by

factors observed fall far short of the 10
van Duyne. Indeed, even exactly on resonance surface
plasmon coupling can only account for a factor of 10 to
100.38 Ultimately, the giant Raman intensities observed
may be accounted for by several cooperative mechanisms.
Other theories to explain the anomalous surface Raman
intensities have been proposed which have not enjoyed the
widespread acceptance of either the image field theory or
surface plasmon coupling. These include resonance Raman

45,46

from surface transients or radicals, normal Raman

scattering from molecules intercalated in a dense carbon
matrix at the electrode surface,47-49 resonance Raman
scattering from slightly perturbed surface and molecular

states,50 and radiative excitation and recombination of

51'52 Of these, Cooney's work on the

particle-hole pairs.
role of surface carbon helped researchers recognize the
important influencecfifsurface impurities and correctly
cautioned coverage interpretations. Regis and Corset46
have observed enhanced surface Raman scattering from
bipyridines on silver. Their assignments were particularly
useful in the assignment of the Raman spectrum of t—l,2-
bis(4-pyridyl)ethylene on silver in this report.

Studies of adsorbates deposited under ultrahigh vacuum
conditions have provided considerable information on

53,54

coverage dependencies of intensities, orientation of

13

molecules at the surface,55 energy transfer between adsor-

bates and meta1,56-58 influences of impurities,47 roughness

phenomena,59 and luminescent background identification.60'61

Demuth et a1.55 examined an interesting coverage dependent
phase transition on silver. When the coverage was increased
the Raman spectrum was interpreted to indicate a transition
from flat molecules to end-on configured pyridines. Con-
flicting reports from Bell Laboratories leave the question
of distance for maximum intensity enhancement from the
metal surface unresolved. Smardzewski et a1.53 initially
reported that molecules closest to the surface exhibited
the largest intensity enhancement. This conclusion was
reached by observation of Raman spectra from a monolayer
of deuterated pyridine followed by increasing coverages

of undeuterated pyridine. The deuterated pyridine closest
to the surface exhibited the most intense Raman scattering.
Rowe et a1.58 measured the enhancement as a function of
molecule-surface separation reaching the opposite conclu-
sion i.e. molecules beyond those immediately adjacent to
the metal surface also exhibited Raman intensity enhance-
ment. The controversy is representative of the overall
conflict currently in the SERS field. That is, is the
enhancement effect a chemical one, requiring the formation
of a chemical bond to the surface and enhancing only those
molecules attached to metal atoms? Or, is the enhancement

factor a result of electromagnetic phenomena, in which

the field at the molecular position is enlarged due to the

presen
in the

centre

direct
CN-, a
other
Pyraz
contr
surfa
pyrit

rule

14

presence of the surface, thus, allowing molecules simply
in the vicinity of the metal to become enhanced? This
controversy remains unresolved.

Although the greatest concentration of effort has been
directed at the study of pyridine, pyridine derivatives and
CN-, applications of surface enhanced Raman scattering to
other molecules have also appeared in the literature.

62'63 for example, has provoked considerable

Pyrazine,
controversy over the appearance of ungerade modes in the
surface Raman spectrum. Similar to t-1,2-bis(4-
pyridyl)ethylene, pyrazine has a center of symmetry and the
rule of mutual exclusion is operative. Forbidden bands

are also observed in the surface Raman spectrum of benzene
which have been discussed in terms of symmetry lowering

64'65 Biologically

important molecules have been observed on silver66-68 and

resulting from surface site symmetry.

demonstrate advantages of surface studies. Some benefits
of surface studies of biological molecules are the absence
of flourescent background owing to the non-radiative
excited state decay pathway provided by the metal substrate,
and general absence of decomposition problems from over-
heating in the beam with the metal surface acting as a heat
sink. Moskovits and DiLella69 observed surface Raman
scattering from ethylene and propylene on silver. These
spectra were helpfulithhe assignment of t-1,2-bis(4-
pyridyl)ethylene SERS. Several other molecules including

water,70 €0,71'72 and EDTA73 have also been observed on

15

silver substrates.

Several of these enhancement mechanisms and surface

74,75

techniques have been reviewed in the literature. In

this thesis only the mechanism of resonance Raman enhance-

ment by surface complexes is addressed.

CHAPTER III

THEORY OF RESONANCE RAMAN SCATTERING

FROM SURFACE COMPLEXES.

One mechanism which has recently advanced to the
forefront of surface Raman discussions is resonance Raman
scattering from surface complexes. The formation of a
medium strength chemical bond between cyanide and the

76’77 However,

silver surface has been demonstrated.
similar efforts to observe a silver-pyridine stretching
vibration have generally failed. In this chapter the
implications of surface complex formation on the intensi-
ties of Raman signals from adsorbates is discussed.

Complex formation at the metal surface requires locali-
zation of metal electron densities. Such localization was
examined by Goddard and McGill78 in a review of quantum
chemical methods appliedixasurfaces. The model found
suitable for surface electron localization was intermediate
between a lone metal atom and the delocalized metallic
matrix, i.e. a cluster of metal atoms. With the orbital
localization suggested by cluster formation a connection
is made between the roughness requirements of surface

79-81 ' 82-84

enhanced Raman scattering, the adatom hypotheses,

and a surface complex induced resonance Raman enhancement

l6

17

85'86 That is, clusters of metal atoms formed in

mechanism.
surface preparation procedures (e.g. oxidation-reduction
cycle) result in orbital localization which may be utilized
in overlaps of adsorbate electron densities.
Metal-adsorbate electron density overlaps may signi-
ficantly alter the distribution of electronic states. Thus
the possibility of coming into resonance with a low lying
excited state of the complex even with red excitation

becomes very real. Therefore, mechanisms of resonance

Raman scattering warrant further study. According to
87

 

 

Rousseau et al. the total scattering cross section is
8nw 4 <ngr Iev><evlr Igi> 2
_ s 2 l p K
0 — 4 fi- (5 _w ”if ) + ar (1)
9cl1 e,v g,e L I

where the bracket represents the pKth component of the
molecular polarizability, apK; wL is the laser excitation
frequency; ms is the scattering frequency. gi (electronic
state 9 and vibrational state i), ev and gj represent
respectively ground, intermediate and final vibronic states
in the scattering event. re and rK are electronic position
operators, with p and K representing incident and scattered
polarizations respectively. PI is the linewidth of the
intermediate state and ar represents antiresonance sums.
The sum runs over all vibrational-electronic states

except initial and final states, i.e., Igj> + |gi>. When

the frequency of the incident photon matches the frequency

18

of the electronic transition (resonance condition) the
scattering cross section becomes large and intensities
increase. Most of the adsorbates examined in SERS experi-
ments exhibit no resonance Raman scattering in solution
owing to the lack of low lying (visible) electronic states.
Therefore complex formation must be presumed if the enhance-
ment mechanism of resonance Raman scattering is invoked.

The metals which have demonstrated surface enhanced Raman

10 metals. This

scattering (Cu, Ag, Au, Hg) are all d
wealth of electrons may be used to occupy normally unoccu-
pied adsorbate molecular orbitals upon complex formation
as observed for the Ru(NH3)5(t-l,2-bis(4-pyridyl)ethylene)
complex88 (see Figure l) in which n* states in the bipyri-
dine ligand are occupied. In the pentaminebispyridylethy-
lene ruthenium II complex the complex symmetry is CS
removing degeneracies occurring in the C2v pentamine
ruthenium substrate. The attachment of t-l,2-bis(4-
pyridyl)ethylene to the silver surface should similarly
reduce the symmetry of the molecule probed in the Raman
experiment. Perturbation of the potential energy of the
normal modes in the adsorbate upon complexation may also
be invoked to account for minor shifts of vibrational
frequencies observed in the surface scattering experiments.
When considering the electronic distributions of
states in a surface complex in more detail some important
conclusions can be made. Excluding for the moment the

influence of the image dipole, formed by rearrangement

Figure l:

19

Energy State Perturbation in t-l,2-bis(4-

pyridyl)ethylene upon Complexation to RuII

(NH The point groups for the ligand,

3’5“
complex and Ru (NH3)5 are given. Ru(II)
has a d6 electron configuration and thus the

lowest three levels in the energy level

diagram for the complex are occupied.

20

 

Azzevnaet

NI

a.w$-u<za<_vo§<_oaov

 

 

 

83...:

3—32:me

 

 

21

of conduction electrons in the metal in response to the
adsorbate electron distribution, the complex would have
one—sided symmetry. Thus, all D group and higher symme—
tries would be eliminated owing to the absence of multiple
rotation axes. The low symmetry would then be expected

to lift most d electron degeneracies in the metal increasing
the electronic density of states as observed in some low
symmetry metal complexes.89 With a high density of states,
between which most transitions would be allowed, excitations
in the red may result in resonance Raman intensity enhance-
ment. Therefore, increased enhancement in the red, which is
experimentally observed in all SERS studies and must be
accounted for by any enhancement theory, is explained by

the concept of a low symmetry surface complex.

Returning now to the influence of the image field on
the symmetry of such a surface complex some previous misun-
derstandings are discovered. Pearce and Sheppard90 suggested
the influence of a "surface selection rule" operating on
the IR spectra of molecules adsorbed at a highly polari-
zable metal surface. In this model, the system which
interacts with an external electromagnetic field is not
the original charge distribution (a molecule) but the
system "molecule + image". As depicted in Figure 2, Pearce
and Sheppard visualized a vibrating dipole at a metal
surface producing dipole charge in the "perfect" metal.

For an oscillating dipole oriented parallel to the surface

the image would oscillate in the Opposite direction

22

Figure 2: Surface Selection Rule Formalism. R is the
operation connecting the adsorbate dipole with
its conduction electron image. F is an

irreducible representation.

in plane
7 :2

 

 

lx l-xl
R i = -1
ll ill

 

 

 

23

out of plane

 

 

{-1 oo\
ll= -10
001)

 

{1+1} is isomorphous with a c, axis.

f;= L1= L1
(i=8

For th :

2 III
,(end-on)
Ag(x’,y’,z’,xy)

E.”

z "X
(flat)

Ag(x29 yzizz)
3g (yz)

24

cancelling out that motion in the adsorbate. Alternately,
a dipole oscillating along the normal to the surface would
be reinforced by the oscillating image dipole.

Hexter and Albrecht91 considered in more detail the
symmetry effects of the "surface selection rule" on both
infrared and Raman activities. It is found that a new
symmetry operation, that of reflection in the plane (oh)
and charge conjugation C, is added to the total symmetry
of the molecule + image system. The new point group
generated by the new operation is M = G X {E_:_R } where G
is the molecular point group and R = ohC. The operation
{E—iiRdiis similar to a crystallographic screw axis, i.e.
it is isomorphous to a C2 axis. The effect of R on a
cartesian column vector is demonstrated in Figure 2 and the
R matrix developed with the normal to the surface desig-
nated 2. Therefore in the infrared spectrum only motions
polarized along 2 will be active i.e. P2 = A;

Px = FY = B; and only A representations are active. In

the Raman spectrum, in which components of the polariza-
bility tensor transform as the direct product of the
cartesian vectors, only symmetric direct products will be
active, i.e. P22 = sz = Fyz = ny = A; sz = Pyz = B. The
expected Raman activities for two different orientations of
a C2h molecule, of which t-l,2-bis(4-pyridyl)ethylene is

an example, on a metal surface are derived in Figure 2.

25

The effect of the image dipole would therefore be
to increase the symmetry of the system (molecule + image).
Increasing the symmetry can generally be expected to
decrease the number of Raman active modes. In many surface
studies the number of vibrations observed has increased
over that observed in solution suggesting "surface
selection rule" breakdown. The validity of the "surface
selection rule" lies in the degree to which the image
dipole charge distribution exactly mimics the adsorbate
oscillations on the metal surface. It must be understood
that the oscillations on the surface are those of nuclei.
The charge distributions in the metal are made up of
electrons alone and respond only to fluctuations in charge
distributions in the adsorbate. The response of the
conduction electron image to the adsorbate nuclear vibra-
tion therefore depends entirely upon the variation of the
electronic wavefunction with normal coordinate (Qk)' The
magnitude of that response decides the contribution to the
Raman scattered light intensity by the image dipole. A
sufficient but necessary requirement for "surface selection
rule" breakdown is that the image field show a different
nuclear coordinate dependence of the electronic wavefunc-
tions. Different electronic response functions on either
side of the silver surface will result in a lower symmetry
system as probed by the Raman experiment. The "surface
selection rule" requires an increase in symmetry, and thus

would breakdown. The specific electronic dependence on

26

nuclear coordinate in the Raman scattering tensor has been
reviewed by Clark and Stewart,94 and their treatment is
applied here to the adsorbate at a metal surface system.

The normal coordinate dependence of the electronic
wavefunction can be examined more closely by considering
the solutions to the Schrodinger equation for the system.
By introducing a zeroth-order Born-Oppenheimer approxima—
tion the wavefunction for the Raman scatterer in Eqn. (1)
becomes

= g
‘ng Og(€:Q)¢j (Q) (2)

where Og(€,Q) is the electronic wavefunction for state 9
and ¢j9(Q) is the jth vibrational state function of the
ground electronic state. 5 and Q represent electronic and
nuclear coordinates respectively. The integration over
electronic coordinates in Equation (1) leads to a
transition moment

.1 _ e .3 g
(Me’gi,ev"<¢v [(mp(Q))g,el¢i > (3)

and the corresponding moment of K polarization in (3).

(mp(Q))g e is the pure electronic transition moment
I

(mpmng,e = <9 r le > (4)

at nuclear configuration Q. In order to integrate Equa-
tion (3) over nuclear coordinates the explicit form

of the Q dependence in (mp(Q))g e is required. Following

Albrl

seril

Subs1

into

In a
the e
Equat
theor
Eqn.

Rayle

27

Albrecht92 and Rousseau87 (mp(Q))g e is expanded in Taylor
I

series about the nuclear coordinates
amp Q
(m (0)) = (m (Q l) + (———) K (5)
o g:e p 0 g,e 30k O,g,e,

Substituting the result into the transition moment and

into Equation (1), the molecular polarizability becomes

 

 

1 <1jv><vli>
a = E (m (0 )) (m (0 )) — 2 _ _
pK e p 0 g,e K 0 g,e h v (mg,e wL Fe)
+ 2 (ame) (amK) l E <j|v><v|QK|i>+<jIQK|v><v|i>
e BQK 0 BQK 0 h K (wg'e-wL-ire)

(6)

In a normal Raman experiment in which the v-dependence of
the energy denominators is negligible the integrals in
Equation (6) may be evaluated by invoking the closure
theorem (i.e. 3|v><v| = 1). Thus, the first term in

Eqn. (6), Albrecht's A term, will be responsible only for
Rayleigh scattering (i.e. j = l) and the second term which
contains the explicit nuclear coordinate dependence results
in Raman scattering. The absence of nuclear coordinate
dependence in the first term of Eqn. (6) in a normal Raman
experiment requires that the image field will not contribute
to the intensity of the Rayleigh scattering. In a resonance
Raman experiment the dependence of the energy denominators
on the excited-state vibrational quantum numbers cannot be

neglected and closure cannot be applied. The vibronic

28

expansion of the energy denominator, according to Tang and

Albrecht93 is given by

 

where weo is the vertical transition frequency evaluated at
the equilibrium nuclear configuration of the ground state.
The N = 0 term is the normal Raman case. Keeping only

the N = 1 term and recognizing that in the adiabatic

approximation we v|v> = % <elfl1e>lv> the first term in
I

Equation (6) becomes

1 <j|<e|fl>we|e>|v><v|i>

 

A' = 2 (mp(Q0))g'e(mK(Q0))

2:
e759 g’e H v

(8)

The Hamiltonian in Equation (8) can be expanded about the

equilibrium nuclear configuration following Clark and

Stewart.94
%-<e|:ic<o) - weo|e> =%- <e|ic<o> + (335900 + — weo e
= % <elflKO)|e> + <elggle>00 + ...-weo
and =11? <elg—‘gle>oQ (9)

 

*
Note: the damping factor has been omitted since the
expansion is strictly only valid far from resonance.

29

since % <e|fl(0)|e> = meo. This leads to a revised A term

which includes the resonance Raman case:

 

l <le|v><e|§5gle>O<vli>
A'=E(m(Q)) (m(Q)) 2-
p 0 g,e K 0 g,e h 0_ _ 2

(10)

Thus, if a force, given by the derivative of the potential
part of the Hamiltonian* with nuclear coordinate
(Fe0 = (3Ve/BQ)O), is experienced upon excitation into an
excited state, then the first term in Equation (6) may
contribute to observed Raman intensities. To experience
a force in the excited state the equilibrium position in
that state must be shifted relative to the ground state
equilibrium position. The shift in equilibrium position
will be only along totally symmetric normal coordinates
since the matrix element <eiggle> vanishes for non-totally
symmetric coordinates. Thus, resonance Raman scattering
by the A' term will result in enhancement of totally
symmetric vibrations.

In the second term of Equation (6), Albrecht's B term,
the dependence of the electronic wavefunction on nuclear
coordinate is explicitly contained in the coefficients

. 92
(amp/BQk)0 and (amK/BQk)o. Follow1ng Albrecht the

 

*

Note: In the adiabatic Born-Oppenheimer approximation the
potential energy of a nuclear vibration is equal to the
total electronic energy therefore

(aVe/aQ) 0 = <e|%%(|e>o.

30

coordinate dependence of the electronic moment may be
treated as a Herzberg-Teller perturbation i.e.

8m
(4343”2 = >3 mg'smoi ,1; <elaJC/aokls> . (11)
3

30k

Therefore, it is evident that contributions to the
Raman intensity from both the A term and the B term depend
on the quantity (aflyan). In the A term (axyao) produces
vibronic mixing within one excited state and in the B
term (aflyan) couples different electronic excited states.
Since the overlap of energy bands in the electronic manifold
of the valence and conduction electrons is extensive in the
metal and limited in the adsorbate the magnitudes of
(Bx/an) in the adsorbate and image dipole are likely to
be quite different. Since (BHYSQ) is responsible for the
intensity of the resonance Raman scattered light that
intensity emanating from either side of the silver surface
will not be the same. Therefore, contrary to the prediction
of the "surface selection rule", the image field can only
act to reduce the symmetry of the scattering system not
increase that symmetry.

Thus, with the absence of any symmetry ordering
influence by the conduction electron image, the system
being probed in the Raman experiment returns to a one-
sided attachment of an adsorbate to the surface. Symmetry

reductions caused by the one-sided nature of the system,

31

the mismatch of the image field, and low site symmetry
experienced by an adsorbate at the surface of a metallic
matrix, result in expectations of extremely low symmetry
scattering systems. The most common point group expected,
even for highly symmetrical adsorbates, is C1. Of course,
in such low symmetry all modes will become Raman active and,
in general, an increase in the number of observed Raman
lines is expected. However, in a resonance Raman mechanism
only those modes effective in coupling electronic excited
states through electron density overlap are enhanced. Thus,
all possible vibrations may still not be observed. For
example, t-l,2-bis(4-pyridyl)ethylene adsorbed at a silver
surface may actually be attached through either of the
nitrogen atoms (but not both) in an "end-on" configuration
or may lie flat on the surface. In the end on configuration
only the nitrogen atom experiences electron density overlap
with the metal. Thus, it might be expected that the
surface Raman spectrum observed would not be much perturbed
from the normal Raman spectrum. In the flat configuration,
in contrast, the entire molecule experiences some overlap
of electron density with the metal and all vibrations
(both infrared and Raman active) are expected to be observed.
Conclusions drawn in this chapter are extrapolated to
the discussion of orientation and enhancement mechanism of
t-1,2-bis(4-pyridyl)ethylene at a silver electrode surface.
The experimental surface Raman spectra and discussions are

presented in Chapter 6.

32

The surface complex enhancement mechanism has been
slow to attain fruition. This has mainly been the result
of a lack of success in observation of a metal-adsorbate
stretch in the low frequency region. For all surface
Raman studies a very complicated low frequency region is
observed which has hindered assignments. The appearance
of a strong, Ag-Cl stretching mode in electrolyte studies
near 230 0111-1 also may cloud assignments in the region.
Pettinger and Wetzel85 recently applied reduced potential
difference spectra (RPDS) to the low frequency Raman
spectrum of pyridine on silver. In doing so the Bose-

w/kT - l).1 in the prefactor of the Raman

Einstein term (e
cross section could be eliminated revealing hidden struc-
ture in the low frequency region. A silver-pyridine sym-
metric stretch was assigned to a 210 cm.1 line.

Efforts to find a metal-adsorbate stretch may be
doomed at the outset. Such a mode can be expected to be
considerably damped owing to one atom of the vibrating
pair being associated in a large metal matrix (see, for
example the treatment of free damped motion by Rossgs).
The low amplitude resulting from damping will result in
low Raman intensity. Forces associated with the electric
field at a metal electrode also may act to damp the oscil-
lations of an adsorbate metal bond.

Experimental evidence of surface complex formation

in the system t-1,2-bis(4-pyridyl)ethylene at a silver

33

electrode is presented in this thesis. Assignments are

made without regard to the "surface selection rule".

CHAPTER IV

Experimental

Raman Spectrometer

 

All Raman spectra reported in this thesis were obtained
with a Spex Industries model 1401 Ramalog system. Photon
detection was accomplished with a cryostatted (-20°C) RCA
model C31034 GaAs photomultiplier tube in photon counting
mode. All spectra were recorded with a resolution better

than 2 cm-l.

Samples

Polycrystalline samples and aqueous solutions were
sealed in 1.7 mm capillary tubes. High vapor pressure
methylene chloride solutions were placed in a Raman cuvette.
A translating Raman cuvette holder and brass cooling jacket
were constructed for optimal sample management. Opaque
polycrystalline samples were observed in a back scattering
geometry. The incident laser beam was allowed to pass
through a hole in a 45° mirror just prior to striking
the capillaried sample. The back scattered radiation
was thus reflected into the entrance slit of the double

monochromator.

34

35

The first electrochemical cell utilized in surface
Raman studies was constructed from a hollowed-out lucite
cylinder. Raman spectra were observed through a quartz
optical flat. Because of difficulties in maintaining
the system free from impurities, the lucite cell was
replaced with a glass cell which could be cleaned by
immersion in concentrated nitric acid solutions. The
working electrode was a polycrystalline silver disk fitted
to a teflon sheath. The surface area of the silver disk
in contact with electrolyte solution was measured to be
0.11 cmz. The working electrode was generally placed very
close to the observation window only after the oxidation-
reduction cycle was complete. This was done to prevent
non-uniform coverage of the electrode surface caused by
concentration gradients which develop during the cycling
procedure. Subsequent movement of the electrode surface
close to the observation window permitted observation of
maximum intensity from surface species and a minimum from
the bulk solution.

The reference electrode was a silver-silver chloride
electrode. This was prepared by electrolyzing a coating
of silver chloride onto a silver coated platinum wire which
was immersed in a saturated silver chloride solution
(prepared by addition of a few drops of silver nitrate to a
saturated KCl solution). The body of the reference electrode
was k" glass tubing to which an ultrafine mesh porous plug

was attached. Generally, the reference electrode was

36

attached to another k" glass tube containing the supporting
electrolyte. This prevented potential drift accompanying
minor leakage. The reference potential of the Ag/AgCl
couple was 50 mV relative to the saturated calomel elec-
trode. All electrode potentials reported in this thesis
are referenced to the saturated calomel electrode (SCE).

In supporting electrolytes containing NaClO4 the reference
electrode was prepared with saturated NaCl solution instead
of KCl so as to prevent precipitation of KClO4 which is
minimally soluble in water. The counter electrode was a
heavy gauge platinum wire. Nitrogen gas (Airco Inc.,
99.99% pure) was bubbled into solutions prior to electroly-
sis and often during Raman observation to limit the forma-
tion of silver oxides. All surface spectra were recorded
at an angle of incidence of 60° for maximum scattering
intensity as observed by Pettinger et al.97 This system

is shown in Fig. 3.

The electrochemical cell was controlled with an Eco
Control Inc. model 550 potentiostat and model 731 digital
integrator. Generally, 50 mC/cm2 (or 5.5 mC for this
electrode) were allowed to pass during the oxidation-
reduction cycle (i.e. the potential was maintained at
+100 mV versus SCE until sufficient charge passed and then
the potential was returned to -600 mV versus SCE). This
cycling yielded a rough electrode surface. The reflecti-
vity of the metal surface drops during the oxidation-

reduction cycle producing a slightly yellow coating even

37

in the simple chloride blank. Varying amounts of residual
charge remained after completion of the ORC depending on
adsorbate and electrolyte. Potential drift was negligible
(~ct2 mV) over the course of a typical Raman experiment.
Potential cycling was carried out in the presence of
adsorbates. For solutions of t-l,2-bis(4-pyridyl)ethylene
at millimolar concentrations potential cycling produced a
dark brownish-black coating at the electrode surface.
However, similar Raman intensities were observed for

500 nM solutions of his pyridyl ethylene which upon cycling
yielded a yellow surface coating undistinguishable from
that observed in the chloride blank.

The silver electrode surface was subjected to mechani-
cal polishing before each experiment. This consisted of
applications of A1203 powders (Buehler Ltd.) of diminish-
ing size finishing with 3 micron powder. The electrode
was then rinsed with methanol and water and transferred to
the electrochemical cell with a drop of water attached to

inhibit oxide formation.

Lasers

A Spectra Physics model 165-08 argon ion laser (4 watts
all lines) with model 265 exciter was used for 5145 A
excitation and to pump a continuously tunable Spectra
Physics model 365-09 dye laser with three-plate bire-
fringent filter. A Classen filter was used to assume

monochromaticity of dye laser emissions. Rhodamine 6G

38

Figure 3: Experimental arrangement in the SERS experiment.
R = reference electrode (Ag/AgCl); W = working
electrode (polycrystalline silver); C = counter
electrode (platinum wire); FL = focusing lens;
CL = collection lens; ¢- = angle of incidence

lnC

relative to the surface normal.

 

4O

(Exciton Chem. Co.) was the dye used in obtaining excita-
tion profiles of surface molecules covering the range
5600 A to 6600 A. For red excitation a Spectra Physics

model 164-01 krypton ion laser was employed.

Excitation Profiles

 

Excitation profiles of species adsorbed to the silver
electrode surface were obtained in order to examine the
electronic manifold of states in the surface-molecule
system. The 935 cm.1 line of sodium perchlorate (0.5 M)
served as an internal standard. The intensity of the
935 cm.1 line could be varied by positioning the electrode
surface at various depths in the electrolyte. Because
of the potential dependence of the surface concentration
of C104- a better choice of internal standard is necessary
for comparison of potential-dependent excitation profiles.
The teflon sheath fitted to the silver electrode serves
this purpose well. Both the metal surface and the teflon
could be illuminated simultaneously with the application of
cylindrical focusing. The Raman spectrum of teflon is
reported here and is generally uninterfering with the
surface Raman spectrum of t-1,2-bis(4-pyridyl)ethylene
(see Figure 4).

It became necessary in the course of carrying out
excitation profile experiments on surface adsorbates to
assure that the laser beam remained focused to the same

point on the surface with each excitation wavelength.

41

Figure 4: Normal Raman spectrum of teflon and surface
Raman spectrum of t-l,2-bis(4-pyridyl)ethylene

on silver with teflon standard.

42

 

 

 

 

is... asses
Essa 5:: ...; a...” s we...

 

 

 

:3 8e
- q q _ £49 q Om.“ - £9— . 3N.— “ fi‘u _ 8W“
......m m ..
: m
z m n
m n
m
5:2 mm
A
...... m.
u
E.
J T.

 

 

 

 

 

 

 

 

 

43

Since tuning the dye laser requires passage through a
birefringent filter of refractive index greater than 1
each excitation frequency will traverse space with a
different vector (similar bending occurs during passage
through the Classen filter). Therefore an aperture was
placed just prior to the sample and the laser direction
adjusted to pass through the aperture and thus illuminate
the same point on the surface regardless of excitation
frequency. Spectra were recorded every 100 cm.1 during
the excitation profile.

The excitation profiles reported in this thesis are
corrected for the v-4 dependence of the Raman scattered
light intensity.5 Spectroscopic intensities were norma-
lized to the intensity of the 935 cm.1 line of the
perchlorate, but not normalized relative to intensities at
different excitation frequencies since doing so would
effectively double the relative error in each profile
data point. Therefore the most intense peaks in the
profile belong to the most intense peaks in the spectrum
itself. A program for v-4 correction and normalization
was written for a Hewlett-Packard 67/97 calculator and

is available upon request.

Laser Power Measurement
Laser powers were measured with either a Photodyne
model 44XL optical power meter or a Coherent Radiation

model 210 power meter.

44

Visible Absorption Spectra

 

Absorption spectra of t-l,2-bis(4-pyridyl)ethylene,
its dihydrochloride and a silver complex of t-BPE were
obtained on a GCA/McPherson model Eu-700 series UV/Vis

spectrometer.

Sample Preparation: t-l,2-bis(4-pyrigyl)ethylene(t-BPE)

 

Trans-1,2-bis(4—pyridyl)ethylene (Aldrich Chem. Co.)
was recrystallized from hot water; some orange color was
removed in the process. Normal Raman spectra of polycry-
stalline t-BPE were recorded. Since t-BPE is only sparingly
soluble in water, reaching saturation at about 3 mM
concentrations, spectra were recorded in methylene chloride
solutions. Solutions for surface studies ranged from
2.5 MM t-BPE (heavy coverage) to 250 nM t-BPE (light
coverage). Supporting electrolytes were either 0.1 M

KCl or 0.1 M NaCl and 0.5 M NaClO For "light coverage"

4.
surface studies fresh polishing pads were used owing to
the presence of residual t-BPE on used polishing pads,

which could influence bulk concentrations and therefore

coverage.

t-l,2-bis(4:pyridyl)ethylene dihydrochloride and dideutero-
chloride

Trans-1,2-bis(4-pyridyl)ethylene dihydrochloride was
prepared by dissolving t-BPE in methanol and adding concen-
trated hydrochloric acid until no further crystallization

was evident. The pinkish orange crystals were collected on

45

a 15 mesh buchner funnel and washed with methanol. The

Raman spectra of polycrystalline and aqueous solutions

of the dihydrochloride were obtained.
Trans-1,2-bis(4-pyridyl)ethylene dideuterochloride

was prepared by bubbling DCl into a methanol solution of

t-BPE. DCl was prepared98 by slow addition of 5 ml D20

to 30 ml benzoyl chloride and refluxing under light heating.

Unreacted D20 and benzoyl chloride were collected in an

ice trap and DCl in a liquid nitrogen trap. DCl was

bubbled through the t-BPE solution until no further

crystallization was evident. The pinkish-orange crystals

were collected and washed with methanol. A similar prepara-

tion was attempted by using chloroform as the solvent and

a pink compound was obtained after considerable DCl was

passed through solution. This compound was insoluble in

water and was air oxidized overnight to a yellow crystal-

line compound. It is believed that this pink compound

was the addition product of DCl across the ethylene double

bond. However physical evidence of its structure is

currently incomplete.

 

Silver Complex of t-l,2-bis(4-pyridyl)ethylene (Agn(t-BPE)m_
A silver complex of his pyridyl ethylene was prepared
by addition of an aqueous solution of silver nitrate to an
ethanol solution of t-BPE until no further crystallization
occurred. Crystallization appeared complete after addition

of silver nitrate to 1:2 (AgNO3/t-BPE) molar ratio with

46

t-BPE. Crystals were collected and washed with both
ethanol and water. The complex was found to be soluble
only in n-butyl amine and pyridine. The silver complex

was found to be insoluble in water, ethanol, CC14, CH3C1,
CHZCl, CHC13, dimethyl sulfoxide, tetrahydrofuran,

ammonium hydroxide, methyl amine, n-propyl amine, C82,
methanol, acetone, cyclohexane, n-hexane, diethyl amine,and
benzene. The Raman spectrum of the polycrystalline silver

complex was obtained and is reported here.

Amino Acid Solutions

 

Amino acid solutions were prepared to 0.05 M'in 0.1 M
KCl stock solution. Observation of Raman spectra of the
amino acids on silver adsorbed from neutral solution was
unsuccessful. However spectra were observed from solutions
in which the pH was adjusted above the isoelectric point
of the particular amino acids. Glycine and leucine surface

spectra were recorded at a bulk solution pH of 11.0.

Agonz Agzg

Raman spectra of polycrystalline silver hydroxide
(AgOH) and silver oxide (A920) were obtained for comparison
with surface spectra to examine the possible occurrence
of surface oxide formation. Commercially available silver
oxide (Fisher Chemical) was used without further purifica-
tion. Silver hydroxide was prepared by addition of excess
hydroxide (KOH) to an aqueous solution of silver nitrate

until the solubility product of silver hydroxide

47

(1.52 x 10-8; 20°C)99 was exceeded. Both silver oxide and

silver hydroxide are black polycrystalline substances and
therefore a back scattering geometry was required to obtain

the Raman spectrum. The specific geometry was described

previously in this chapter.

CHAPTER V

Polycrystalline and Solution Raman Spectra
of t-l,2-bis(4-pyridyl)ethylene, its
dihydrochloride and its dideuterochloride
In order to understand the surface Raman spectrum

of trans—1,2-bis(4-pyridyl)ethylene (t-BPE) more clearly
the normal Raman spectrum was recorded. The Raman spectra
of the dihydrochloride and dideuterochloride are also
reported to assist in assignment of vibrational modes. The
Raman spectra of these compounds have not previously been
reported in the literature although the infrared spectra

1’100 and cisl isomers have been reported.

of both the trans
Both solution and polycrystalline spectra are reported.
Solution depolarization ratios are also determined. The
Raman spectra are shown in Figures 5 and 6 and the vibra-
tional frequencies are listed in Table 1. The pyridine

ring vibrations are assigned according to Wilson notation.101

All normal Raman spectra reported here were obtained with

6471 A excitation.

Symmetry Considerations

 

The molecular point group symmetry of t-l,2-bis(4-
pyridyl)ethylene is C2h' The center of symmetry in the

molecule requires that infrared and Raman activities be

48

49

mom

moo
moo
mam

Hem
mum
Hmm
mvm
mow
mam
Ohm
«om

mwm
vmm
HVH
00H
mm
on
mm
vm
mv

H OHNH

H
H mom
Ohm

mom
Nmm
Hmm
ovm

Nov

mmm

 

UmHOm O

N N

HO

No

D mam!»

.mmfifiHMCOU H QHQMB

 

 

 

momH amm. omma
meoa
moom amm. mmm
mmm mmm
emm
mmm deem. mmm
mme name. mme
mee mee
ome

Hoe
omm
emm
oem omm
emm
mm
om
me
nmmom a 0mm
numaomm mmmubunu

 

mama mam. mmam .nbn e..o me me
emaa Ma
meom nmm 4 Ha
mam Ema. mmm m we Ha
emm omm nmm mm Hm
mmm emm m on ma
mmm nos mm ma
mmm meme. omm mom mm m
mme noes. mew be mm mm
flee amm. one no ma Ha
eme gem mm mm
med n ma
omm
mam mmm x N am
oem mmm nponnnm>
mom mnbmaexm
Ham
Hm
WOCOZ
wombbnq
me
neaom e mommo Am nowamse awn nmo >mo
iiiiiiii mmmiuiilil ucoEcmmmmm ilmuuoafiwmi

 

.< anew "COmumumoxm

.mpmuoHcooumusopmp ocoamnuoaawpmnwmivvmmb:~.aiu pcm mpmuoHsooupwnmp
ocoawnuwaawpmu%mivvmmbnm.aiu .wcmawnumfiawpmuwmlvvmmbum.alu mo muuoomm cmEmm "a mam¢a

50

 

 

maom moam
aeem
amom
emom
nneamm
neoeem
mmea amea
meea amea
mmea mmea
eoea mema
nnmoma mama
aama meea
emma
aama eema
mmma mmma
eema aama
mama
eaaom 0mm
aoma mam-»

 

 

uneoem
eemm nnommm
mmom aoam
amom
omem
amem
emea mmea
eeea nem. meea
mmea nem. meea
moea
anooma anaoma
mama nem. mama
mmma nee. mmma
eema mam. aama
mmma omma
emma mom. oema
m
eaaom e o n
:lmao m mamlbulu

emom oeom
amom

emom

maom

emea amea
mmea amm. aaea
aaea omea
mama nae. mama
aama nem. aama
aama

aaea nma. maea
aaea amm. mmea
aaea mam. aaaa
amma mmm. mmma
mama

aama naa. mama
emma nmm. mmma
mmma Ema. omma
eaaom a aommo
uuuuuuuu manipulnuu

.nbm muz a m<

 

 

A..Hu.m UHU x NV AN <
noumnum
mo
.aun mlmo a9 m4

aooma x my
.aun auz

.nbm e> m Ma mm mm

.uum ono < a

a< am

am me am

he < mm

e x m we am

naa < a<

baa we mm

aa + e 4 a<

aa x m w< am

ea < mm

be x m we a<

m mm mm

.mm@ mz ha mm mm

ea c am

1* :onaaze nma nmo >mo

ucmecmmmmfi ilmuumfifiwml

 

.poscmucou a canoe

Figure 5:

51

Normal Raman Spectra of polycrystalline
t-l,2-bis(4-pyridyl)ethy1ene, its dihydrochlor-
ide and dideuterochloride.

+
EXCITATION WAVELENGTH: 6471 A (Kr ).

lllI)

 

9»
fl
e\
c:
- -~ 3054
w ?3038
-3
g_ r 013
O
A 1676
\
__ -—1630
16"
g;- f 1591
c:
, 1549
1541
* ~1491
:;’ .1319
8' K
iii?
{ —1319
‘ - 123:
a (‘ 1225
:5. --—1198
Q t. ”80
§ \rm —995
‘958
300
’> 878
a
°.,,
0 \
‘1-611
641 _.
a: :'—-
°-_ 0
«=1 ’7’
2’.
‘31
‘F
005 :3
:L
=-
‘C
3 s
o- 0-0
370 {if
0
=
O
200
25 200
O
f‘:) 121
-, —;:—-— ~—-="9‘
4:: _____,3

 

52

 

“L 3005
3059
3050
jF’3031
1571
44
g ‘19 -————-1535
-1605
..1512
1500
1373
1346
1333
‘ - 1254
. -— 1203
1062
——aeiooa
A ~394
I- 319
,___55,
£643 7
:-
fi‘
5'
:5.
fr
'6
f4ao :5.
a.
1
9;
11”” ‘e
(m 5:
~— 240 ‘2;
fl
3'
E?
q -
EL
134 "
. 7“2§-
,.,, ‘~—-00
<17 g- ‘3
K

“672

‘1;
- 3092
“fl€f:;3060

3039
3020

a [643

gieos

.=-l510

—1206

1066
fe:~~ -e——-1005
I996

‘~=>-901
‘;~ use

ammmoomnw 30311111311“de-ms1q-Z'1-1

,_, ———-—-———— 1635

Figure 6:

53

Normal Raman Spectra of t-l,2-bis(4-
pyridyl)ethylene in CH2C12, the aqueous
solution of the dihydrochloride, and the
dideuterochloride in D20.

EXCITATION WAVELENGTH: 6471 A (Kr+).

INTENSIIY .

....Némsézaa.£3.25 9252222.? ... 8o

 

 

 

 

 

z
W 0
3 . n
u a 1
... I e _
r L L .-.IF .1 711.! .11». 1L! _ :1
wng mg: ~23 ~23 ~30
W 2 a“ “e u 22 I. 2 H
mm "u H» mm nununflr111¥ mu.u mg" mmw w a.
C, > 2 - E --.lllk A.
w.
m m
M. u“
.n
n o
u .._.~.zm3.3aa: £3.28 33:33.33 5 am:
a
m. _
I
_ m
J 7
n m w 11 u n .. m
6 a 1 3 .. a ,7
I. /K\/.lli.i\lu\ .(e. /. ..-Ili rll.(\\, (Qa/
.[IPI'IIL .— b. b . . . F -IFILI
mac: “:8 uwcc use .4.
5
A .._.~.zms.3aa._€2.12: ... 62%:
H
me
I _
1 7 n
‘ e s c
_ - ....
,. .... , ...
_ 8 2O ._ p

‘:Efilifao
>167!
{ggzo—"_
/ ..

,2 1549
>1492

‘ 1422
.7141:

‘ 133

1313

1

1

E172

‘- - -— 993
;570

954
3300
{cu2cb
f»ss7
Isso

l

I

225

woo

55

mutually exclusive. Since the molecule is made up of two
pyridine rings, vibrations in one ring may be in-phase or
out-of—phase with the same vibrations in the other ring.
In-phase vibrations will be of A symmetry (Ag, Au) i.e.
symmetric with respect to the C2 axis. Out-of—phase motions
will be antisymmetric with respect to the C2 axis or of

B symmetry (Bg, Bu). Analysis of the number of irreducible
representations for t-BPE predicts the observation of

23 Ag, 10 Bu' 11 Au' and 22 Bu modes in the vibration
spectra. Only gerade modes will remain Raman active.

Thus by correlating the symmetry representations of pyri-
dine (CZV) modes to the molecular symmetry (CZh) as in

Table 2 it becomes evident that each pyridine motion
correlates to one Raman active mode and one infrared

active mode. Combinations and overtones involving ungerade
modes, whose direct products transform as gerade representa-
tions, can also be observed in the Raman spectrum of t-BPE.
In Table 2 ethylene modes (DZh) are also correlated to the
molecular symmetry (CZh)' The vibrational modes associated
with the symmetry species in the correlated symmetries

102'103 included in Table 2. Further correla-

(CZV’DZh) are
tion to possible surface complex symmetries is also examined
in Table 2. The choice of axes is consistent with pyridine
assignments in reference (104). The crystallographic space
group symmetry was not determined in this study. Therefore,

correlation to crystal and site symmetries for polycrystal-

line Raman spectral assignment was impossible. Efforts

56

.mMDZHBZOU N mqmfifi

fl
baa .ma .m .ea
m mm ne.naa .nm .nm .nom

a A//// \\\\V am nea .aa .e .noa .m .nma

 

 

e”! ea
\v1

t
t;

 

 

 

\. 4 ova .MOH .oma
\\‘|‘nr.
:4 we .H .ma enma .Mm
.4 .1. a a< .naa .na .nem .ma .m
mu Emu >mo mcmpmuwm
11momuuweeww111 ImuumEE>m1 1>uumEE>m1 11111111 mmpoz 1111111111
xmamfipwzmommusm amasowaoz cocoamnwmw

 

.momuumeshm meQEOU mommusm
manmmmom Op EOHUMHEHHOU HQEUHSM paw ocmamnuomawpmumm vvmmnlmealu mo knuofiewm
unasomaofi mnu Op mamawsum can momma mampmuwm map m0 mmmupmaewm mo coaumamuuou "m mqmée

57

0 ~ "moon/m mo oomocu
X H
a

 

 

 

 

 

m
5am m9
0
< mm
II o
mam o9 .m>
.-< < \ 4 H) \N) fimp
..a
mo LNU zma «mcoawnuo
11mmmuuwaamm111 IwuumEE>m1 1>auoEE>m1 mopoz
onmEou 000musm unasooaoz venomounou

 

.QMDZHBZOU N mqmfie

58

to grow adequate single crystals for X-ray examination

are underway.

Frequency Region 0-600 cm-l

 

Assignments in this region of the Raman spectrum of
t-l,2-bis(4-pyridyl)ethylene are particularly difficult.
Other than the lattice modes, most of the bands in this
region are of very low intensity and quite broad. Lines
at 486, and 400 cm.1 have been assigned to pyridine motions
16b and 16a respectively following Spinner.104 The fre-
quency at 225 cm-1 has been assigned to a skeletal vibra-
tion in accordance with a similar motion in trans-

105 The X-ray crystal structure determination

stilbene.
is required in order for assignment of modes in the lattice
region of the polycrystalline Raman spectra to be made

possible.

Vibrations 6a and 6b

 

The in—phase (Ag) components of orthogonal pyridine
vibrations 6a and 6b appear at 641 cm.1 and 671 cm-1
respectively in the Raman spectrum of polycrystalline t-BPE.
Assignment of these modes is fortified by comparison with
the same motions in the dihydrochloride. In the Raman
spectrum of polycrystalline t-l,2-bis(4-pyridyl)ethylene
dihydrochloride the vibration 6b is observed shifted down
to 651 cm-1 while vibration 6a appears slightly shifted
up to 643 cm-1. The downward shift of 6b frequency is

exactly analogous to that observed in the Raman spectra

59

of 4-methyl pyridine and 4-methy1 pyridine hydrochloride as
reported by Spinner.102 There, the frequency of the 6b
motion in 4-methyl pyridine appears at 669 cm.1 and 651 cm-1
in the hydrochloride. In t-BPE, the behavior of vibration
6a upon formation of the hydrochloride follows that observed
for pyridine and pyridinium chloride. In pyridine the Ga

vibration appears at 605 cm-1; it falls at 607 cm.1 in

103 a similar shift to higher frequency

pyridinium chloride,
upon salt formation. A further consideration in the
assignment of vibrations 6a and 6b is the relatively weak
intensity of 6a relative to 6b in the Raman spectrum of
t-BPE and its hydrochloride. In some substituted pyridines
and pyridine hydrochlorides the Ga vibration is too weak

to be observed102 or appears with very weak intensity. The
intensity of 6b is also greater than that of 6a in the
Raman spectrum of trans-stilbene (the analogous molecule

105 Regis

to t-BPE with benzene rings instead of pyridine).
and Corset106 report a 6b/6a intensity ratio of 25/4 in
the Raman spectrum of methyl viologen (N,N'dimethyl-4,4'
bipyridine). In methylene chloride solution the Ga vibra-
tional mode, appearing at 650 cm-1, exhibits a depolariza-
tion ratio of 0.17 indicative of a polarized mode as
expected for a mode of Ag symmetry. Vibrational mode 6b:
however, appears depolarized with a depolarization ratio
of 0.64 though the symmetry of this mode is also Ag. The

difference in depolarization ratios arises because the

vibrational mode 6b transforms as the ayz component

60

of the polarizability tensor (in the coordinate frame

designated in Table 2) while 6a transforms asoL , a ,

YY
and azz. Therefore, in the polarizability invariants for

XX

vibration 6b only the derivative of the yz component of
the polarizability with respect to normal coordinate
(aayz/an) will be non-zero. The oyz component appears
only in the anistropic polarizability invariant and the
spherical invariant therefore vanishes, resulting in a
predicted depolarization ratio of 3/4. Situations such
as this where a totally symmetric vibration is expected
to demonstrate a high degree of depolarization, are not
common in group theory. In fact the only point groups in
which off-diagonal polarizability tensor components
transform as the totally symmetric representation are C2h
(the case examined here) and the very low symmetry point
groups C2, Ci' Cs' and Cl'

Vibration 6b cannot effectively couple to the N-H
deformation in the dihydrochloride since the motion of the
nitrogen atom and hydrogen atom are in the same direction
(the form of the normal coordinates for pyridine vibrations
are given in Appendix A).107 Thus, no shift is observed
upon deuteration of the dihydrochloride with vibration 6b

1 in both salts. Assignments of

appearing at 651 cm-
vibrations 6a and 6b are particularly important to the
understanding of the Raman spectroscopy of t-BPE adsorbed

at a silver electrode (see Chapter 6). Comparison of Raman

61

frequencies for modes 6a and 6b in representative molecules

is made in Table 2a.

Out-of—Plane Vibrations

 

All out-of—plane molecular vibrations appear in the
Raman spectrum with a frequency below 1000 cm-1. Only
the out-of—phase components of such vibrations are observed
in the Raman spectrum since the active symmetry is Bg for
out—of—plane modes. A pair of lines appearing in the
polycrystalline Raman spectrum of t-BPE at 878 cm.1 and
888 cm"1 have been assigned to pyridine ring vibrations 10a
and 10b respectively. The form of the normal coordinate
vibrations 10a and 10b as provided by Long and Thomas107
demonstrates that vibration 10a has no component of motion
involving the nitrogen atom while 10b involves nitrogen
motion (see Appendix A). Therefore, protonation or
deuteration of the ring is expected to have little effect
on vibrational frequency of vibration 10a. In the Raman
spectrum of both t-l,2-bis(4-pyridyl)ethylene dihydro-
chloride and dideuterochloride vibration 10a appears
unshifted at 879 cm-1. Vibration 10b, in contrast, appears
at 894 cm.1 in the polycrystalline Raman spectrum of the
dihydrochloride and at 901 cm.1 in the dideuterochloride
spectrum.

Analogously, pyridine ring vibration 5 has a component

of nitrogen motion and can therefore couple to the out—of-

plane N-H deformation. In the Raman spectrum

62

TABLE 2a: Comparison of Raman frequencies of modes 6a
and 6b in representative molecules.
frequencies are in wavenumbers (cm'1)*.

Vibrational

 

t-BPE

pyridine

4-methy1pyridine

t-BPE H2C12

pyridinium chloride
4-methy1pyridine hydrochloride
t-BPE D2C12

pyridine deuterochloride

4-methy1pyridine deuterochloride

a. Too weak to be observed.

b. No available data

641

605

643

607

645

602

671
652
669
651
637
651
651

634

 

Frequencies reported were those observed with polycrystalline
samples for all molecules except pyridine and pyridine
deuterochloride. Liquid pyridine Raman frequencies are
reported. Pyridine deuterochloride frequencies were

observed in D20 solution.

all

63

of polycrystalline t-BPE vibration 5 appears with a fre-
quency of 958 cm.1 but cannot be found in the Raman spectra
of the dihydrochloride and dideuterochloride. The form of
normal coordinate 17a (Appendix A) exhibits zero amplitude
at the nitrogen position as in vibration 10a. Therefore,
17a is not expected to couple effectively with the N-H
out-of—plane deformation and its frequency should be shifted
little in the Raman sepctrum of the dihydrochloride and
dideuterochloride relative to its position in the t-BPE
Raman spectrum. In the Raman spectrum of polycrystalline
t-BPE 17a appears at 976 cm.1 and at 970 cm.1 in methylene
chloride. In the dihydrochloride and dideuterochloride
salts 17a appears at 997 and 996, respectively, demonstra-
ting better than a 20 cm.1 shift. However,-in the Raman
spectra of the aqueous solutions 17a appears at 977 cm-1
and 970 cm"1 for the dihydrochloride and dideuterochloride:
respectively, which is shifted only slightly from the

motion in the solution t-BPE Raman spectrum. The out-of-
phase components of some of the out—of—plane modes do not
appear in the Raman spectrum of t-BPE though they transform
as Raman active representations (Bg). Those vibrations
absent are v4 and v . Interestingly, these can be observed

11

in the surface Raman spectrum of t-BPE.

Ring "Breathing: modes v1 and v12_

The totally symmetric ring "breathing" vibration v1

 

appears in the vicinity of 1000 cm-1 in all pyridines

64

substituted pyridines,enu1pyridine hydrochlorides. v1

is the most intense Raman line in the spectrum of pyridine

and occurs with high intensity in the spectrum of t-BPE.

In polycrystalline t-BPE v1 vibration appears at 995 cm-1.

In the Raman spectrum of the dihydrochloride v1 appears at

1008 cm-1. The form of the normal coordinate vibration

v 107
l

nitrogen motion which is directed among the N-H bond in the

(Appendix A) shows that the in-plane motion contains

hydrochloride. Therefore v1 may couple to the N-H stretch
and is expected to occur ataalower frequency upon deutera-

tion. In the dideuterochloride of t-BPE v appears at

1
1005 cm—1, which demonstrates the expected shift relative
to the dihydrochloride.

In the Raman spectrum of pyridine another very intense
line appears at 1030 cm-1 and has been assigned to the
trigonal ring "breathing" vibration v12. This vibration
is conspicuously absent from the Raman spectrum of t-BPE
and its chloride salts. The symmetry of this in-plane
deformation is Al symmetry in pyridine and thus the in-
plane analogue of v12 in t-BPE transforms as the totally
symmetric representation and is expected to be observed
in the Raman spectrum in the frequency region 1030-1040 cm-l.
The appearance of v12 in the surface Raman spectrum of
t-BPE and in the normal Raman spectrum of a polycrystalline
silver complex with t-BPE is important evidence for the

formation of a silver complex at the electrode surface (see

Chapter 6).

65

Frequency Region 1050-1300 cm-1

 

The in-phase component of in-plane ring deformation
v18a is very weakly observed in the polycrystalline Raman
spectrum of t-BPE at 1069 cm—1. The orthogonal counter-
part to vibration 18a, 18b is also expected in the frequency

1

region 1050-1100 cm- but is not observed in the Raman

spectrum of t-BPE.

1

At 1184 cm- in the polycrystalline t-BPE Raman spec-

trum a very weak line is observed on the tail of the very
intense Ceth - ¢ stretch at 1198 cm—1. This weak line is
probably a combination or overtone which is in Fermi
resonance with the very intense symmetric stretching vibra-
tion. If this is the case the symmetry of the combination
of overtone must be Ag. All 2nd overtones will be totally
symmetric since the result of the direct product of any
representation with itself is the totally symmetric repre-
sentation. However, the second overtone of a vibration

of frequency 592 cm.1 is required and no such vibration is
observed in either the infrared100 or Raman spectrum of
t-BPE. Interestingly, a very weak line near 600 cm.1 is
observed in the surface Raman spectrum of t-BPE which may
be the fundamental whose overtone comes into Fermi resonance

with the Ce - ¢ stretch. Assignment of this surface mode

th
is incomplete (see Chapter 6).
As mentioned above, the totally symmetric stretch

between the ethylenic carbons and the pyridine rings occurs

with high intensity in the polycrystalline Raman spectrum

66

of t-BPE at 1198 cm-1. Upon formation of the dihydrochlor-
ide and dideuterochloride this vibration shifts to higher
frequencies appearing at 1203 cm-1 and 1206 cm-1 in the
hydrogenated and deuterated analogues, respectively. The
frequency and intensity of the Ceth - ¢ stretch are
analogous to that observed in the Raman spectrum of trans-
stilbene.

Vibrations 9a and 9b appear next in the Raman spectrum
of t-BPE. 9a appears at 1225 cm-1 in the polycrystalline
spectrum of t-BPE, shifts under the intense neighboring

C - ¢ stretch in the spectrum of the dihydrochloride

eth
and is barely visible at 1215 cm.1 in the Raman spectrum

of the dideuterochloride. These observed shifts corres-
pond to those observed in the Raman spectrum of 4-
methylpyridine. As reported in reference (104), 9a appears
in the range 1220-1210 in 4-substituted pyridines and

102 The shift

shifts to 1205 cm-1 in 4-methy1pyridine.
to higher frequency upon deuteration of the dihydrochloride
is also observed in the Raman spectrum of pyridinium

chloride. In the hydrogenated salt 9a appears at 1198 cm-1

and at 1202 in the deuterated analogue.108

Vibration 9b is antisymmetric with respect to the C2
axis, transforming as the B2 representation in pyridine,
and thus the component of nitrogen motion (Appendix A) is
involved in N-H+ in-plane deformation in the dihydrochlor-

ide and dideuterochloride. In the Raman spectrum of poly-

crystalline t-BPE 9b is assigned to the intense line at

67

1234 cm”1 which shifts to 1254 cm_1 in the dihydrochloride
and drops sharply in intensity. Similar to vibration 6b,

9b exhibits a rather high depolarization ratio for a mode
of Ag symmetry since the vibration transforms as the yz
component of the polarizability tensor as discussed earlier.
In the D20 spectrum of the dideuterochloride 9b appears at
1294 cm-1. The shift to higher frequency by 30 cm"1 in

the solution (D20) spectrum is an indication of the

extensive hydrogen bonding occurring in solution.

Frequency Region 1300-1400 cm-1

 

In this region of the spectrum pyridine ring vibra-
tions v3 and v14 appear. The Raman line at 1319 cm-1 in
the spectrum of polycrystalline t-BPE is assigned to the
in-phase pyridine ring motion v3 in accordance with
assignments in other pyridine molecules104 and trans-

105

stilbene. Similar to vibration 9b, the frequency of

l

v increases upon salt formation, appearing at 1333 cm-

3
in the polycrystalline dihydrochloride.

In the Raman spectrum of polycrystalline t-BPE vibra-
tion 14 appears at 1351 cm-1 and at 1338 cm.1 in methylene
chloride. Vibration l4 correlates to the B2 representation
in the pyridine molecular point group as do vibrations 3,
6b and 9b. As for vibrations 3 and 9b, v14 demonstrates
a frequency increasing in going to the dihydrochloride
and dideuterochloride in the solution Raman spectra. Lines

at 1346 cm"1 in the solution spectrum of the dihydrochloride

and 1349 cm.1 in the solution spectrum of the

68

dideuterochloride are assigned to v14. On the low frequency

side of v in the Raman spectrum of polycrystalline t-BPE

14
appears a vibration at 1342 cm-1. This line is probably
the overtone 2x6b which is in Fermi resonance with the
totally symmetric vibration 14. The weak line appearing
in the Raman spectrum of t-BPE dihydrochloride and
dideuterochloride near 1375 cm.1 is probably an overtone
of vibration 11 which is of Au symmetry and not observed
in the Raman spectrum. In the surface spectrum of t-BPE,

however, appears near 685 cm.-1 as discussed in the

“11

next chapter.

Vibration Pair 19a and 19b

 

Pyridine ring vibrations involving non-hydrogen
stretches and bends occur in the frequency region 1400-
1650 cm-1. Vibration 19b is assigned to a weak line

1 in the Raman spectrum of polycrystal-

observed at 1419 cm-
line t-BPE. Similar to vibrations 6b and 9b, vibration

19b demonstrates the characteristic high depolarization
ratio, measured at 0.53. Vibration 3 (B2) demonstrates a
very low depolarization ratio (.19) which may be a result
of accidental degeneracy with a combination, perhaps 2x6a.
As with other vibrations which transform as the B2 symmetry
representation in the correlated pyridine symmetry (C2v)
19b couples with N-H+ deformation in the salts of t-BPE.

Thus, the vibrational frequency of motion 19b shifts to

1512 cm"1 and 1514 cm-1 in the dihydrochloride and

69

dideuterochloride, respectively. (Further evidence of
hydrogen bonding involving the nitrogen proton in solution
is evidenced by the shift of the frequency of 19b in the
dideuterochloride down to 1463 cm.1 upon dissolution in
D20.) On the low frequency shoulder of the 1422 cm-1
line a line appears at 1411 cm.1 in the Raman spectrum
of polycrystalline t-BPE. This is another example of
Fermi resonance in the spectrum. This mode is probably

the totally symmetric combination v4 + v As discussed

11°
earlier, neither v11 nor 04 appear in the normal Raman

spectrum of t-BPE. v4 is observed weakly at 740 cm.1 in

the infrared spectrum of t-BPE.109 v11 is observed at

685 cm'1

in the surface Raman spectrum of t-BPE as
mentioned above.

Pyridine vibration 19a is observed at 1491 cm-1 in
the Raman spectrum of polycrystalline t-BPE. 19a has a
component of nitrogen motion along the N—H bond and there-
fore is expected to be responsive to formation of the
salts. In the polycrystalline dihydrochloride and dideu-

terochloride 19a appears as shoulders at 1500 and 1505 cm-1,

respectively.

Vibration Pair 8a and 8b

 

Still another example of Fermi resonance in the Raman
spectrum of polycrystalline t-BPE appears at frequency
1541 cm.1 on the shoulder of fundamental 8b at 1549 cm-1.

The overtone is probably 2v11. The out-of-phase component

 

In

(M

We

70

of vibration 4 is not observed in the Raman spectrum as
discussed previously. However, the in-phase (Au) component
appears in the infrared spectrum100 at 740 cm-1 and in

the surface spectrum of t-BPE (Chapter 6) two lines are
observed at 720 cm.1 and 740 cm-1. Thus the Bg out-of-
phase component probably occurs at 720 cm.1 and the overtone
is expected at 1540 cm-1.

Typical of other b vibrations which may couple to N-H
deformation vibration 8b, appearing at 1549 cm“1 in the
Raman spectrum of polycrystalline t-BPE, shifts upon
dissolution of the dideuterochloride salt in D20 from
1606 cm.1 to 1567 cm-1. An analogous shift is observed
for vibration 19b. Vibration 8a appears as the most
intense line in the Raman spectrum at 1597 cm.1 in poly-

102 observed that 8a shifted

crystalline t-BPE. Spinner
by nearly 30 cm.1 to higher frequency upon formation of
the hydrochloride in the Raman spectrum of 4-methy1pryidine.

A similar shift of vibration 8a is observed in the Raman

spectrum of t-BPE with formation of the dihydrochloride.

Ethylenic C=C Stretch

 

The totally symmetric stretch of the central ethylene
carbon atoms appears at 1638 cm-1 in the Raman spectrum of
polycrystalline t-BPE. Upon formation of the dihydro-
chloride and dideuterochloride salts the carbon-carbon
stretch probably overlaps the intense 8a vibration. The

weak high frequency shoulder of vibration 8a in the

71

polycrystalline Raman spectra of the salts has been assigned
to the C=C symmetric stretch although the low intensity is

not completely understood.

C-H, N-H(D), Stretching Motions

 

The C-H stretching frequencies appear in the Raman
spectrum of t-BPE between 3000 cm-1 and 3100 cm-1. The
lowest frequency can be assigned to ethylenic carbon-
hydrogen spectrum following assignments in trans-stilbene.
Without isotopic substitution studies assignment of other
frequencies in the 3000-3100 cm"1 region are ambiguous.

In the dihydrochloride aqueous solution spectrum two broad
bands centered at 3270 cm-1 and 3400 cm-1 represent N-H
symmetric stretching motion. The broadness of these lines
is indicative of extensive hydrogen bonding with water. A
line appearing at 3266 cm.1 in the polycrystalline t-BPE
dihydrochloride Raman spectrum is probably the overtone
2X(C=C sym. str). In the deutrated salt the N-D symmetric
stretch appears as a broad doublet with centers at 2400 cm-1
and 2510 cm-1. The polycrystalline dideuterochloride Raman
spectrum exhibits no bands in this region.

The absence of N-H, and N-D stretches in the polycry-
stalline Raman spectra of the salts along with the large
shifts in frequency of modes coupling to N-H(D) deformations
upon dissolution appears to indicate the relatively weak

association of the proton (deuteron) for the nitrogen

atom in the crystal lattice. This association is expected

72

to be influenced by the presence of the highly electronega-
tive chloride ion nearby in the lattice.

The assignments in this chapter are central to an
understanding of the Raman activities of t-l,2-bis(4-
pyridyl)ethylene adsorbed at a polycrystalline silver

electrode (Chapter 6).

CHAPTER VI

Raman Spectra of t-1,2-bis(4-pyridy1)ethy1ene
adsorbed at a polycrystalline silver electrode.
Intense Raman scattering from trans-l,2-bis(4-

pyridyl)ethylene adsorbed at a polycrystalline silver
electrode has been observed in this study. The procedure
for electrochemical pretreatment of the silver electrode
is described in Chapter 4. Concentrations of t-BPE in
initial studies were 2.5 mM in 0.193KC1, but concentrations
as low as 250 MM t-BPE produced comparable Raman signals.
Normal Raman signals from aqueous solutions of 2.5 mM
t-BPE were too weak to be observed. Surface Raman spectra
of t-BPE, in contrast, generally demonstrated signal inten-

5 to 3 x 105 counts per second.

sity in the range 1 X 10
Under similar electrochemical roughening procedures t-BPE
molecules adsorbed to silver from bulk concentrations of

1 mM and above produced a brownish-black coating on the
electrode surface while adsorption from bulk concentrations
of 500 MM and less produced an electrode coating indis-
tinguishable from that observed following electrochemical
cycling in aqueous KCl solution alone. Thus studies of

surface Raman spectra from t-BPE molecules adsorbed from

bulk concentrations in excess of 1 mM are hereafter referred

73

74

to as "heavy coverage" spectra. Those spectra observed
from t-BPE molecules adsorbed from bulk concentrations less

than 500 MM are referred to as "light coverage" studies.

Heavy Coverage Raman Spectra of t-BPE under 5145 A

 

Excitation.

 

The Raman spectra of a heavy coverage of t-l,2-bis(4-
pyridyl)ethylene adsorbed at a polycrystalline silver
electrode under 5145 A excitation are given in Table 3.
Some important sections of the Raman spectra are shown
in Figure 6. Compared in Table 3 and Figure 6 are the
normal Raman spectrum of t-BPE (assigned in Chapter 5)
and the surface Raman spectra of t-BPE recorded at
-600 mV versus sce (i.e. relative to the saturated calomel
electrode) immediately following the oxidation-reduction
cycle, -50 mV versus sce and again at —600 mV versus sce
(following recording of the spectrum at -50 mV). Symmetry
assignments based on the molecular point group symmetry
and possibly surface complex symmetries (C2, C5) are also
included in Table 3.

In the Raman spectrum of t-BPE on silver recorded
immediately after the oxidation reduction cycle (which
appears just above the polycrystalline t-BPE Raman spectrum
in Figure 7) some interesting differences and similarities
to the normal Raman spectrum are observed. The most
pronounced difference is the appearance of a new peak at

1

655 cm- in the surface spectrum. Assignment of this

Figure 7:

75

Heavy Coverage Potential Dependence of the
Surface Raman Spectrum of t-l,2-bis(4-
pyridyl)ethylene on Silver. Compared are:

the normal Raman spectrum of polycrystalline
t-BPE (bottom spectrum); the surface Raman
spectrum recorded at -600 mV immediately fol-
lowing the oxidation-reduction cycle (above
polycrystalline spectrum); surface spectrum at
-50 mV; and the surface spectrum recorded on
returning to -600 mV (top spectrum).

EXCITATION WAVELENGTH: 5145 A (Ar+).

 

cmwll

O
O
N
we

I
I

76

-600 my vs. SCE

trans~1.2-bis-(4-pyridyl) ethylene
on Ag at

 

«Own.-
wmmue. 1

mmefil mme
V 1 1 l
moelilmluv meal. . mwe
MOwhhlllil/JM l. r
Smart/L1 - memalil
- -J --
coma . mama
23W
:2 as
m m .1 ll mama
m anmmar , aama
u m as:
mm: ... ... 5 mm:
(o 9:
gnaw 31-1 mac : 4
r/,-,. aama
3.3.». mm“...
Qmmu l. v . t lWLV

ms..a...:ll:.v - Eadlr .. -

.. Mace—...! .-Hll

:-.1 aeea
5 J

:55;—

-600 my

 

polycrystalline

 

l _l
700 650

L
I

1500 1400 1300 1200

1600

 

(:m'1

77

 

amm22aezoo m mam<ee

.cOabsaOn maommo ca also mee
¥

.>E oml um Esnuoomm mcmonooou mGHBOHHOM popuouou Esnuooam

+
.maoxo GOHUUDUTHIGOHumpme mnu uwumm mawumeoEEH Umpuoowu Esuuoomm
.1

 

 

k

 

ema nohnnha>
mumum
UmvHUXQ MOH
3> oem 2> oem 3> oem a ea e am
3> omm 3> omm 3> omm m =4 e we
eae mae 3> mme a ea aa aa
mee mee mee a .a e:ame he aa
mme mme mme a .a he a<
mee ne ma
ooe ooe oee maa am
3 omm m .a he an
3 amm 3 mm m .< me am
0mm mm mmm
omm omm omm
mom mom mom
omm omm omm oem
omm omm omm .nom acuaa
mom
nhoaa nhoaa nhoaa ama encahnnha>
am moahhna
me
+>s ooe: Il>s cm: >5 ooe: mo no mmmnh a cohaas manoEEmm
on am no ma * um m4 mamaaoummuo>aom ucoEGmmmm< Emu
co manna co manna no mamnh
.wamamhumxameanmaueenah:m.a1h mcaa
Iaoumwuowaom num3 mmma GOmmquEou .COHumumoxo < mvam Hopes mponuooam Hm>amm

mcmaamummuomaom m ouco UmQHOmpm ocmawcuomaxpmuwmlwvmeIN.H1u mo muuommm cmEmm "m mqmfla

78

Amm22aezoo m mameev

.>E omI no Esnuoomm ocapnoomu mCaBOaa0m pmpuoomu Esuuommm

+

.oao>o coauosponicOaumpaxo may noumm xaouoameEa popuoumn Econommm
an

 

cm

3>

3>

3>
3>

Nava
mmma

mNma
mama

aama
ooNa

Nooa
hmoa

mooa
Nno
vmo

mum
ovm

 

...->5 COO-I

on em

CO mmmlu

3>
3>

3>

mava
ovma

mmma
mama

onma

NVNa
ooNa
mmoa

oooa
ovoa

mooa
mum

omm
omw

>5 Oml
on am
so mnmuh

cm

3>

3>

3

Nava
mmma

mama
NoNa

mvNa
mNNa
mONa
maaa
vooa
mooa

oaoa
omo
mmo

owm
ovo

>E oool

¥

on am

SO mmmlu

<r< <I<F< a1<cnn2<r<

<1mcna1ms<

U

.4“
.<
.4
:¢
.<
:<

.4
.<

.<
.<

.<
=¢
:<
=<
=<
:¢

m
U

aava
amma
Nvma

mama

vaa
mNNa
mmaa

oooa

mom
who
mmo
hwm
mum

 

mmmlu

ocaaaMpmmnomaom

 

Qoa

aa + v

«a

he x N

ewe m..oo
m

ewe ha m..mo
GOaumuQa>
woman
wouaoxo e..mu
ha . on $2
om

.nhn e..o
ma

hma

naa
ceaumuoa>
oumum
UGUHUXG

a

m

mha

noa

moa

Ahhov e>

* nowaaz
ucmEGmmmm<

C‘UWT010

NoumEEN

L) m staimcnald

E
N

 

.QMDZHBZOU m mam<8

79

.>E oml um Esuuoomm ocaouooou ocaBOaaOM Uopnooou Esuuoomm

 

 

 

 

+
.oaowo ceauospoHIGOaumpaxo one Hmumm waouMacoEEa popnooou Esnuoomm

.4.
emea emea emea 4 .4 emea xhhee 4» x m m4
mmma oaoa avoa 4 .< mmma .uum ONO m4
4 .4 aama m o4
aooa mooa moea 4 .< mmma cm 4
emma amma a .4 ha mm
mmma mmma mama < .4 mama am 4
4 .4 aama e x m a4

moma oama m .4 oma
eaea maea eaea 4 .4 aaea naa a4
emea emea m =4 e x m a4

mvva EOaUMHoa>

mumum

omuaoxo
mmea mmea emea 4 .4 mmea haa a4
+>s ooe: >2 om: e>s ooe: mo no mmmuh # acnaas Nahossxn
no m< um o« no o< ocaaamummnomaom ucoficmammd END

GO mmmlu CO mmmlu :0 mmmlu

 

.QMDZHBZOU m mamflfi

80

Raman frequency is important to the discussion of molecular
orientation at the surface and is reserved until other
surface data are presented below.

Another distinct difference between the normal Raman
spectrum and the surface Raman spectrum of t-BPE, observed
in 0.1 M Cl- electrolyte, is the appearance of an intense
line at 230 cm-1. Similar lines have appeared in the surface
Raman investigations of pyridine5 also and have been assigned
to the Ag-Cl symmetric stretch because of its appearance in
the surface Raman spectrum even in the absence of pyridine.8
We concur with this assignment by noting that moving the
electrode potential negative of the point of zero charge
(p.z.c.), which occurs near -0.97 V for silver,102a the
intensity of the 230 cm-1 line is attenuated to less than
half its intensity at positive potentials and does not
recover intensity upon return to positive potentials. The
intensities of the other lines in the surface Raman spectrum
of t-BPE do not show the same potential dependence. This
indicates that, as expected, Cl- ions leave the vicinity of
the electrode surface at negative potentials and the vacant
sites are filled by rearrangement of the molecules remaining
at the surface. Thus, the number of Ag-Cl bonds decreases
which accounts for the drop in intensity at 230 cm-1. This
is an important point: many researchers in efforts to locate
a metal-nitrogen stretch in the surface Raman spectra of
l

nitrogen-containing molecules have assigned the 230 cm-

line to such a motion.

81

Other differences between the normal and surface Raman
spectra of t-BPE include the appearance of new weak lines
at 685 cm-1, 600 cm-1, 550 cm-1, and a group of three
vibrations in the 300-400 cm"1 region. The 685 cm-1 and

600 cm‘1

lines can be assigned to the in-phase (Au) and
out-of-phase (Bg) components of vibration 11, respectively,
following similar assigments in the vibrational spectra of
trans-stilbene.105 The 550 cm-1 line is assigned to the
out-of-phase (Bu) component of vibration 6b by analogy to

105

trans-stilbene spectra and its appearance in the infra-

red spectrum of t-BPE.100 The low frequency motions have
not yet been assigned. A very weak line is also observed
at 1118 cm-1 which could be the in-phase (Ag) component
of vibration 15 which appears at 1148 cm-1 in pyridine104
but is not observed in the infrared or Raman spectra of
t-BPE. Another weak new line observed in the -600 mV
surface spectrum occurs at 840 cm-1 which may be assigned
to the in-phase (Au) component of model v4 of ethylene
which appears at 837 cm-1 in the infrared spectrum of
t-BPE.1

Other than the differences mentioned above, the surface
Raman spectrum of t-BPE observed at -600 mV under 5145 A
excitation after electrochemical cycling is very much like
the normal Raman spectrum of t-BPE. Absence of Fermi

resonance, intensity drops observed for vibrations v3 and

ng at 1317 and 1245 cm-1, respectively, and the shift in

82

the frequency of v6b from 671 cm.1 in polycrystalline t-BPE
to 667 cm.1 in the surface spectrum are all analogous to
solution effects observedixrthe normal Raman spectrum of
t-BPE (see Table 1). Most of the surface Raman lines
are only slightly shifted from their counterparts in the
polycrystalline spectrum.

Upon altering potential to -50 mV versus sce some
extraordinary changes in the surface Raman spectrum of t-BPE
are observed. The 655 cm-1 line becomes much less intense

and the in-phase component of vibration 11 (Au) at 685 cm-1

increases sharply in intensity. The very weak 550 cm-1

line disappears and a very weak line at 539 cm-1 appears

which can be assigned to the out-of—phase motion 6a since

it appears in the infrared spectrum.109 The intensity of

the Raman line at 880 cm.1 in the -600 mV spectrum drops,

1

accompanied by an increase in intensity at 832 cm- in the

-50 mV spectrum. The ethylenic v4 motion appears to

l O O O C I
and increases in intenSIty on mov1ng the

shift to 850 cm-
potential to -50 mV. The ethylenic carbon to ring carbon
symmetric stretch at 1200 cm.1 picks up some asymmetry

on its high frequency side accompanied by the appearance

of a new line at 1270 cm-1. The in-phase (Ag) ethylenic
hydrogen deformation appears at 1329 cm.1 and is assigned

by analogy with the corresponding mode in the trans-stilbene

vibrational spectra.105

A host of new vibrations appear in the 1400-1560 cm-1

range. Most of these lines can be assigned to the

83

out-of—phase (Bu) ring motions as shown in Table 3. The
out-of—phase component of vibrational mode 8b is assigned
to the new line at 1554 cm-1 by comparison to that observed
in the infrared spectrum of t-BPE where 8b occurs at

1560 cm-1. Similarly, the Raman line at 1510 cm-1 in the
-50 mV surface spectrum can be assigned to the Bu component
of vibration 19a which appears at 1504 cm.1 in the infrared
spectrum. The Raman band at 1480 cm.1 has no counterpart
in the infrared spectrum and is probably the totally
symmetric overtone of vibration 4 (740 cm-1) which is in
Fermi resonance with the totally symmetric component of

l in the

mode 19a at 1495 cm'l. The line at 1418 cm"
-50 mV spectrum can be assigned to the Bu component of
mode 19b which appears in the infrared spectrum1 also
at 1418 cm-1. A new line appears at 1449 cm.1 in the
surface spectrum at this relatively positive potential
which has not been assigned and will be discussed later
in this section. The totally symmetric ethylenic carbon
stretch at 1640 cm”1 is reduced to half the intensity of
that line in the -600 mV spectrum. The increased height
of the valley between the 1640 cm-1 line and the 1603 cm-1
line with decreased intensity at 1640 cm"1 indicates the
possible presence of an isosbestic point accompanying a
shift in the frequency of C=C to a frequency near 1600 cm-1.
Upon returning to -600 mV the initial spectrum observed

at this potential is not recovered. Instead, a spectrum is

observed which maintains many of the features observed in

84

the -50 mV spectrum. The in-phase and out—of—phase compon-
ents of the pyridine ring motions in the region 1400-

1560 cm”1 increase in intensity and the intensity of the
line at 684 cm.1 (V11, Au component) is maintained. The
intensity of the line at 655 cm-1 increases, but remains
much weaker than that observed in the original spectrum
recorded at -600 mV. The in-phase ethylenic hydrogen
deformation at 1325 cm.1 increases in relative intensity.

1, 1449 cm"1 and

Interestingly, the vibrations at 1270 cm-
832 cm-1 have disappeared from the spectrum upon moving

the potential to -600 mV. The intensity of the C=C stretch
at 1639 cm.1 has recovered at the more negative potential.

It must be noted that the spectroscopic changes observed
above occur irreversibly. That is, the original spectrum
observed at -600 mV is never recovered once the electrode
potential has been shifted relatively positive. Also, and
most importantly, these spectroscopic changes are observed
only under laser excitation with wavelengths in the green
or shorter. Excitation with low energy laser lines (e.g.
6471 A, Kr+) results in the observation of only the spectrum
initially observed at -600 mV versus sce and the potential
dependence of the surface Raman spectrum is completely
reversible. Another requirement for the observation of
irreversible potential dependent behavior is that the
electrode must have a heavy coverage of t-BPE at the

surface. In experiments with bulk t-BPE concentrations

at 500 MM and less only reversible potential dependence

85

of the surface Raman spectrum is observed. These experi-
mental data are addressed in the following sections.
Molecular reorientation and surface complex formation are
invoked to explain the anomalous potential-dependent

behavior of the surface Raman spectrum.

Molecular Reorientation on Complexation of t-BPE at the

 

Electrode Surface.

 

The experimental data presented in Figure 7 and
itemized above can be understood by considering the formation
of a surface complex between t-1,2-bis(4-pyridyl)ethylene
and surface silver atoms. In the surface Raman spectrum of
t-BPE observed at -600 mV immediately following the
oxidation-reduction cycle the molecule can be considered to
be bound to the silver surface through one of the nitrogen
lone pairs, i.e. in an "end-on" configuration. This conclu-
sion is reasonable in that the attachment of one point
limits the amount of electrode density overlap between
the metal and adsorbate. Thus, the surface Raman spectrum
of t-BPE at -600 mV appears little different from the
normal Raman spectrum. The major difference between the
normal Raman spectrum and the surface spectrum is the
appearance of an intense line at 655 cm-1. The 655 cm-1
line can be assigned to vibration 6b occurring in the
pyridine ring closest to the silver surface. When the

nitrogen atom is attached to a substituent, vibration 6b

always drops in frequency. In t-l,2-bis(4—pyridyl)ethylene

86

dihydrochloride and dideuterochloride 6b is observed at

651 cm.1 (see Table l), in N-methyl pyridium110 at 649 cm-1:
and in methyl viologen106 at 657.5 cm-l. The form of
107

vibration 6b according to Long and Thomas requires that
the substituent have a component of motion in the same
direction as the nitrogen atom and orthogonal to the
nitrogen-substituent bond. Thus, some aplitude of silver
motion is required in the vibration 6b when t-BPE is
attached "end-on" to the surface. This silver motion is
therefore available for vibronic coupling of metal
electronic states withtflmaincoming radiation and a resonance
Raman mechanism becomes a plausible explanation of the ob-
served large Raman intensities. In a resonance Raman
enhancement mechanism only those modes are enhanced which
most effectively couple to the electronic state in
resonance. In the "end-on" configuration only one pyridine
ring would show electron density overlap with the metal
thus enhancing motions in that ring which are effective in
vibronic coupling. Thus the 6b motion in the ring closest
to the surface (655 cm-1) appears with higher intensity
than the same motion in the ring more remote to the surface
(667 cm-l). Motions which do not include motion of the
nitrogen atom are expected to be absent from the surface
spectrum of the "end-on" t-BPE. Thus, vibration 10b

(888 cm-1) cannot be found in the surface spectrum of

t-BPE at -600 mV and mode 17a appears very weakly at

958 cm-1. A pyridine vibration which is very similar to

87

motion 6b is mode 15. Vibrational mode 15 is not observed

in the infrared or normal Raman spectrum of t-BPE but appears
very weakly at 1118 cm-1 in the surface Raman spectrum of
t-BPE in the "end-on" configuration. Most of the other

modes observed in the surface Raman spectrum at -600 mV

)

transform as the totally symmetric representation (C2h
and may therefore be an indication of a Franck-Condon
scattering mechanism (see Chapter 3).

On shifting the potential to -50 mV several ungerade
modes appear, along with other changes. Ungerade modes
appear exclusively in the infrared spectrum of t-BPE thus
implicating a reduction in the symmetry of the system
probed in the Raman experiment. The "surface selection
rule", studied extensively by Hexter and Albrecht91 does
not provide an explanation for the appearance of infrared
modes in the surface Raman spectrum of a centrosymmetric
molecule such as t-BPE (see Figure 2). A reduction in
symmetry upon complexation to surface silver atoms, in
contrast, adequately accounts for the appearance of infra-
red modes.

The new line at 1270 cm.1 is assigned to the symmetric
ehtylene-carbon-to-ring-carbon stretch in an excited n*
state of t-BPE. This conclusion is made on the basis of
comparison with the excited state vibrational spectrum of
trans-stilbene (observed experimentally by low-temperature,
high-resolution crystal absorption spectroscopy of trans-

109

stilbene in a dibenzyl matrix by Dyck and McClure ). In

88

trans-stilbene the Ceth - ¢ symmetric stretch is observed
at 1191 cm-1 in the ground state (fluorescence spectrum)
and at 1250 cm.1 in the excited state. A line at 1427 cm-1
in the excited state vibrations of trans-stilbene had no
counterpart in the fluorescence spectrum and remained
unassigned. A similar phenomenon occurs in the surface
spectrum of t-BPE at -50 mV. A line at 1449 cm-1 apparently
also has no counterpart in the ground state vibrations since
it disappears along with other proposed n* Vibrations upon
returning to -600 mV. The low-temperature, mixed crystal
absorption spectrum of trans-stilbene exhibits a sequency

of very strong lines separated by 1599 cm“1 which has been
assigned111 to the C=C symmetric stretch in the excited
state. A similar shift in the C=C stretch in the t-BPE
surface Raman spectrum at -50 mV is invoked here to explain
the reduction in intensity at 1640 cm-1. Such a shift

would bring the frequency of the ethylene motion under-
neath the peak of vibration 8a occurring at 1603 cm-1.
The shift in the frequency of mode 10 from 880 cm“1 to

832 cm-1 on moving to relative positive potentials is
analogous to a shift observed in going from the fluorescence

spectrum to the absorption spectrum of trans—stilbene111

l and 847 cm-1,

where lines are observed at 866 cm—
respectively. The potential dependence of the lines at
880 cm"1 and 832 cm.1 shown in Figure 8 is clear evidence

that these Raman lines are associated with the same

Vibration, not different ones.

89

Figure 8: Potential Dependence of Mode 10 in the Surface
Raman spectrum of t-BPE "flat" on the silver
surface.

4..
EXCITATION WAVELENGTH: 5145 A (Ar ).

 

Potential Dependence
of Mode 10b of t-BPE
‘flat' on Ag. .
Excitation: 5145A

mv

SCE
8
00

O
in
Q

'600

N
M
w
M
1

 

 

I I
900 800
CM'

91

Thus it appears that vibrations from the n* state in
t-BPE are observed in the surface Raman spectrum at -50 mV
versus sce. There are two mechanisms which might reasonably
account for the appearance of n* vibrations in the surface
Raman spectrum. The first is the influence of saturation
of the n* state by a high photon density excitation. If
saturation occurred, normal Raman scattering from the
excited state may be observed. However, these experiments
were performed with laser powers of only 50 mW or less.

A simple calculation of the number of photons per t-BPE
molecule (assuming a minimal surface coverage and a beam
waste of 0.5 nm radius) leads to conclusions that, for
saturation to be a viable explanation for the appearance
of excited state Vibrations, the excited state lifetimes
must be of the order of microseconds or larger, which is
not likely.

The alternative explanation for the appearance of n*
vibrations in the surface Raman spectrum of t-BPE is that
the n* state becomes significantly stabilized in energy
by electron density overlap with the metal. As discussed
in Chapter 3, the pentaminebispyridylethylene ruthenium II
complex88 demonstrates considerable stabilization of the
t-BPE n* state upon complexation such that the n* becomes
an occupied state in the complex. In the "end-on" configura-
tion at -600 mV there is little overlap of the molecular n
system with metallic electron density and the charge on

the silver atom is significantly diminished from the charge

92

on the ruthenium II atom. Hence no n* vibrations are
observed at -600 mV. Even at -50 mV the charge on the
surface silver atoms is still small and under red excita-
tion n* vibrations are not observed. Therefore, in order
to postulate that the appearance of n* in the surface Raman
spectrum is a result of stabilization brought on by overlap
of molecular and metal electron densities a reorientation
of the molecule at positive potentials must occur. This
reorientation must be intimately related to the requirement
of 5145 A (or shorter wavelength) radiation since reversible
potential-dependent behavior is observed with red excitation.
Consider first the t-BPE molecule in the "end-on"
configuration at -600 mV. In this configuration electronic
eigenstates in the ring closest to the surface can be
expected to be perturbed owing to electronic overlap with
the metal (particularly those electronic states with non-
zero electron density at the bonded nitrogen atom). Those
parts of the molecule not experiencing overlap will be little
perturbed. Thus of the two n states (i.e. two nitrogen
lone pairs) one is expected to be significantly perturbed
by interaction with the surface while the other remains
fairly constant in energy. For t-BPE in solution the
visible absorption spectrum shows a weak n-fl* transition
which peaks at 475 nm with a molar extinction coefficient

of 1 MI1 cm—1 (Figure 9). This shoulder on the near UV

110

n-n* transition can be assigned to n-n* since in the

dihydrochloride the intensity of the visible peak is

93

Figure 9: The Visible Absorption Spectra of
I. t-1,2-bis(4-pyridyl)ethylene and
II. t-l,2-bis(4-pyridyl)ethylene

dihydrochloride.

94

Asjfiueazmmz

 

cos one coo omm com amt ooe
_ 1 _ a _ .

‘

am: 5 «2.3.5225:
o:o_m._.oa_§:a.3n_h.m.a.a .= ..
mam-8 ... 2.23:.a_m_._:e-$n_h.m.a.a ._
Co 930on :o_3..emn< 0.5m;

 

1 m6

1o.a

m.—

 

 

1uaiomaoo uonoulixa 1910111

(I-wol-w)

95

attenuated (typical of the behavior observed for n-n*
transitions observed in other pyridine compounds upon
acidification). The unbonded nitrogen lone pair of t-BPE
in the "end-on" configuration at the surface is still
available to undergo such an n-n* transition when excited
with adequate energy. At 5145 A excitation, as evident

in Figure 9, the unbonded ring is slightly in resonance
with the n-n* transition. The low symmetry of the surface
complex may act to make the transition allowed and increase
the actual molar absoptivity.

On moving to a relatively positive potential (-50 mV)
an electron excited from the unbound nitrogen lone pair
into the ethylenic n* orbital would be expected to experience
a force of attraction to the positively polarized electrode
surface. The coulombic attraction of the n* electron for
the positively charged metal surface may cause the molecule
to shift its nuclear equilibrium configuration in the
excited state. The coordinate along which the shift occurs
is that coordinate connecting the n* orbital with the
positively charged electrode surface. In general, this
coordinate will include contributions from all normal
coordinates of the t-BPE molecule with components of
motion of the ethylene carbons out of the plane of the
molecule. Thus, an electron excited out of the unbound
n state to a n* state experiences a force which shifts
the equilibrium configuration of the excited state relative

to the ground state along the coordinate connecting the

96

n* orbital with the surface. The motion of the unbound
part of the molecule out of the molecular plane closer to
the electrode surface can be expectedtx>lead to further
electron density overlap with the metal. The increased
overlap afforded by the molecule "lying down" on the
surface can act to stabilize significantly the n* state to
within the occupied eigenstates of the silver-t-BPE complex.
Thus, the n* state appears as a ground state in the surface
Raman spectrum of t-BPE at -50 mV potential under 5145 A
excitation.

The absence of similar molecular orientation at -600 mV
(observed initially following anodization) is accounted for
by remembering that at -600 mV the positive charge at the
electrode surface is reduced (relative to that at -50 mV).
The reduced electrode polarization diminishes the force
experienced by an electron in the n* state. Reemission
occurs from an excited state only slightly shifted along
the coordinate connecting the n* orbital with the metal
surface. Thus reorientation is not affected.

With t-BPE molecules now "flat" on the surface, and
extensive electron density overlap of all sections of the
molecule with the metal, more vibrations are expected to
couple effectively to electronic transitions between
eigenstates of the surface complex electronic manifold.
Therefore, more modes appear, with greater intensity, in
the surface Raman spectrum of t-BPE by virtue of a

resonance Raman enhancement mechanism. Accordingly, the

97

ring motions in the region 1400-1560 cm.1 increase in
intensity since the rings are now flat on the surface.
Observation of the in-phase ethylenic hydrogen deformation
at 1329 cm-1 is further indication that other sections of
the molecule experience electron density overlap with the
metal at -50 mV.

The irreversibility in the potential dependence of
the surface Raman spectrum of t-BPE, evidenced by the
different spectrum observed on return to -600 mV, is a
manifestation of the extensive molecule-metal overlap in
the "flat" configuration. The stabilization afforded by
the overlap makes the "flat" configuration thermodynamically
favored over the "end-on". The disappearance of the n*
vibrations on returning to the -600 mV potential implies
that only at relatively positive potentials is the n*
state sufficiently stabilized to enter the manifold of
occupied states. Evidence in the excitation profile of
t-BPE at different potentials corroborates the stabiliza-
tion of energy states on going to positive potentials
(see the following section).

Not all t-BPE molecules have changed orientation at
the surface, as evidenced by the lack of complete attenua-

1, 1200 cm-1, and 880 cm-1 as well as the

1

tion at 1640 cm-
increase in intensity at 655 cm- after returning to
—600 mV. Regardless of the number of times the potential

is shifted positive (~50 mV), the intensity of the 655 cm"1

line is never completely attenuated. This indicates that

98

the surface can only accomodate a limited number of mole-
cules in the "flat" configuration with "end-on" molecules
filling the gaps.

The enhancement mechanism being proposed in this
discussion is one of a resonance Raman enhancement brought
about by the increased density of electronic states upon
complexation to the silver surface. Evidence that elec-
tronic energy distribution in the complex differs from
that of the molecule itself can be obtained from the
excitation profile of t-BPE on silver. This is discussed

in the next section.

Excitation Profile of t-BPE on Silver.

 

The variation in the intensity of lines in the surface
Raman spectrum of t-BPE with excitation frequency is
reported in Figure 10. The excitation frequency profile
was recorded in the rhodamine 6G range from 15350 cm-1 to
17350 cm-l. The internal standard to which all intensities
were normalized was the 935 cm-1 line of the perchlorate
ion (ClO4-) which was added to the supporting electrolyte
at a concentration of 0.5 M. The focusing used was point
focus and the intensity of the 935 cm-1 could be regulated
by positioning the electrode surface at different depths
in the electrolyte. The intensities of the t-BPE surface
Raman lines are corrected for the fourth power dependence
of the scattered frequency (ms) in equation (1) of Chapter

3. Normalization to intensities at one excitation frequency

was not carried out since doing so would effectively double

Figure 10:

99

Excitation Profile at t-l,2-bis(4-
pyridyl)ethylene on silver at -600 mV sce
in the end-on configuration (Rhodamine 6G
excitation frequency range). Data points
are normalized to the 935 cm.-1 line of the
perchlorate internal standard and corrected
for the 3’4 dependence of the normal Raman

cross section.

 

 

 

Excitation Profile of t-1,2-bis-(4-pyridyl)-ethylene

on Ag at -600 my vs. SCE (end-on).

 

16000 16500 17000

1
15500

cm"

101

the absolute random error. Therefore the most intense
points in the excitation profile arise from the most intense
line in the surface Raman spectrum. The excitation profile
was examined for a heavy coverage of t-BPE molecules on
silver in the "end-on" configuration observed at -600 mV
versus sce. Owing to intensity fluctuations with movement
of the laser beam to different points on the silver surface
precautions were taken to assure excitation of the surface
spectrum at the same point on the surface for each
excitation frequency. The experimental arrangement is
discussed in Chapter 4. The values plotted in Figure 10
and listed in Table 4 were obtained for each Raman line

at each excitation frequency by the equation:

Iv' -4 -1 18

F = (v ) (1 X 10 ) (12)
I935 S

 

where F is the value plotted, IV, is the intensity of the

vibrational Raman line at a particular exciation frequency,

I is the intensity of the internal standard (C104 )

935
line at 935 cm-1 and 3s is the scattered frequency in
‘wavenumbers. The absolute random error in the points (F)

was calculated from

 

 

 

A (AIV.1(US41'1(1 x 1018) (4193511v.(5541(1 x 1018)
F: +
I935 1335
(458141v.(1 x 1018)
+ __5 . (13)

I935V

102

.aIEo Ga ma mucosoonm coaumuaoxm
an

.wm comoxm no: 0U muonum Eoocmn muDaOmQ<

 

 

 

 

 

+
aa.eaa mm.am em.ma ae.ma am.ee mm.aa me.ma a.eemma .am
am.eoa me.mm am.ea am.ma me.em ea.ma me.ma o.ommma .am
em.am me.am aa.aa em.aa mm.aa mm.ma em.aa e.aeama .aa
mm.am ma.am ao.ma ma.aa mm.me me.ma ma.aa m.eeoma .ma
aa.ee am.mm ma.aa om.aa mm.me em.ma ea.oa m.omaea .ma
mm.me aa.am ma.aa me.ma aa.me mm.ma me.ma m.ammea .ea
aa.am mm.me aa.ea me.ea ma.am me.ea mm.ma m.ommea .ma
me.maa ma.aa mm.aa aa.aa em.mm am.mm em.em e.aeeea .ea
me.eea me.maa aa.am am.em me.ma em.am me.am m.aemea .ma
ee.mea em.eaa am.em mm.mm aa.aoa am.mm mm.mm m.omeea .ma
ma.aoa ea.oa eo.aa mm.am mm.am am.mm oe.em m.ommea .aa
ee.mm me.me om.ea me.ea ma.ae aa.ea mm.mm e.aemea .oa
mm.ma mm.me em.ma oa.ma aa.ae me.am me.em o.omaea .a
om.aa ee.me mm.ma am.ma am.me aa.am mm.mm m.meoea .a
mm.ma ae.am ee.ma em.am mm.mm em.mm em.mm m.omama .m
em.ma am.mm mm.ma am.am mm.mm aa.mm em.mm a.amama .e
mm.ma me.am oa.ma am.mm me.ma aa.am ee.mm m.ammma .m
aa.eaa ma.aa me.mm aa.am me.ma aa.am me.ma m.mmema .e
ae.mma mm.moa aa.mm me.am mm.aa me.mm aa.ee a.mmmma .m
aa.ema me.aoa mm.am ee.om aa.moa ee.mm ma.ae e.amema .m
am.mma aa.eaa ao.mm me.am ee.ama aa.ee mm.ae m.ammma .a
oeea aoea eema mama aama mooa mme .moeoswonm a

GOanmuaoxm
11111111111111111 A Boo mucoswonm amG0aumuoa>11111111111111111
+.A 3 How pmoowuuoov mum msmnw> >E oool
he ne>aan no ocmamhhehameanmalevnahum.a1h eo maaeona neahnhaoxm "e mam4m

103

The magnitude of the A quantities in eqn. (13) were

AIV, = 11 cm, = i1 cm and A58 = i2 cm—l. Thus, the

I935
absolute error in each point in Figure 10 is less than
5%.

Tfiuaexcitation profile shows the increased intensity
enhancement in going toward red excitation wavelengths as
observed in the surface Raman studies of pyridine and other
molecules at silver as discussed in Chapter 2. More impor-
tantly, however, is that the excitation profile exhibits a
peak at 16500 cm.1 which does not correspond to the frequency
of any absorption in the visible absorption spectrum of
t-BPE (Figure 9). The extinction coefficient of the t-BPE

1 (606 cm-1) is

visible absorption spectrum at 16500 cm-
near zero. The peak indicates a 0-0 transition from the
ground state to an excited state, i.e. from the state of
zero vibrational quantum number in the ground state to

that in the excited state. This is indicated by the
demonstration of a peak at 16500 cm- in the profile of
each Vibrational mode intensity. If the transition between
Vibrational states of the ground and excited state of
vibrational quantum number greater than zero the peak in
the profile would be shifted for each vibrational frequency.
There is some evidence in the profile of a 0-1 transition
also (i.e. a transition from the zeroth vibrational level
of the ground state to the first vibronic component of the

excited state). This evidence is the appearance of a

crossover of intensities of the 1200 cm-1 line and the

104

1608 cm_1 line at 15750 cm-l. This may be the 0-1 com-

ponent of the transition which peaks (0-0) in the red not

shown in this profile range. If this does indicate a 0-1

1 1

transition then the O-Oieiexpected at 14550 cm- (687 cm- )
which is near maxima observed in other surface Raman exci-
tation profiles (Chapter 2). If indeed the crossover at
15750 cm-1 implicates a 0-1 transition, then the 1608 cm-1
line is expected to show some relative intensity increase
approximately 400 cm-1 further to the blue. At 16150 cm-1
the 1608 cm.1 profile appears to increase slightly compared
to the 1200 cm-1 line which supports assignment of the 0-1
transition.

The appearance of a 0-0 transition at 16500 cm.-1
indicates the presence of an excited state not normally
accessible to excitation in the molecule itself. The
state however may be a complex state generated by electron
density overlap between t-BPE and the silver surface.

The appearance of an intense 0-0 component and a weak or
non-existent 0-1 component indicates that the excited

state position along all normal coordinates is little
shifted from the equilibrium configuration in the ground
state. Generally, to observe Franck-Condon scattering in

a resonance Raman spectrum the excited state is required

to be shifted along a nuclear coordinate such that upon
excitationtx>the excited state a force is experienced. For

Franck-Condom scattering a 0-1 transition or transition

to higher vibrational quanta is expected to be observed

105

with considerable intensity in the excitation profile.
Herzberg-Teller resonance Raman scattering requires no
such shift in the excited state to contribute to the
scattered intensity. Thus, a 0-0 transition is expected
to be the strongest peak in the excitation profile of a
Herzberg-Teller scatterer. It can therefore be suggested
that the intense 0-0 transition in the excitation profile
of t-BPE on the silver surface is a result of Herzberg-
Teller coupling of excited states of the surface complex.

The appearance of an intense absorption at 16500 cm-1
fortifies the hypothesis presentedixrthis thesis that the
distribution of states of the molecule at the surface is
significantly altered and increased in density upon forma-
tion of a surface complex. The complex formed in the
"end-on" configurationcxurbe expected to alter the electronic
state distribution of states which demonstrate electron
density at the bound nitrogen atom.

The peak in the excitation profile observed at -600 mV
is observed to be potential-dependent as shown in Figure 11.
In this figure the excitation profiles of the 655 cm.1 line
in the t-BPE surface Raman spectrum are compared at —400 mV
and at -600 mV. The actual intensities (normalized to
I935 and 534) are listed in Table 5. The 655 cm-1 mode is
the most characteristic line of the "end-on" configuration
of t-BPE on silver. It is observed in Figure 11 that the
intensity of the t-BPE spectrum at -400 mV relative to the

internal standard is reduced compared to that observed at

106

Figure 11: Potential Dependence of the Excitation Profile

of the 655 cm.1 (v Line in the Surface

61?

Raman Spectrum of t-BPE on Silver.

107

a.:=o.
oooha comma ooooa

comma

 

a__a.flra/_

\+

 

/+\7rl_l_

[7.2

I!
/_.\:/I,

_________a__

/.\+\1_/ ... .....-

1/

2.. coo./
_
\o

Z

.35 :3 s at: 1...: ...—5 m3
2: .e 2%... Songs Sages: 32.3....

/

 

(to) 101 paioauoo) AllSN 31m

 

108

.H E0 cfl wocmsvmum

I *

.wm mmmuxm no: ow muouum Eowcmu wusHOmn<_
- - -4.

'1 -l ‘1ft‘llt ..,.'l I .ll1-\ll Ilrl v1 tin-1|"..“III II'I‘I'I‘I'IIIII’y.II, '..llll.,-|".knl
I.I‘:.‘I\l'il . I ilil.l.ellll Ii. I'll-III!!! '.II'I .' I. I‘I.i‘llu.l"lll..' ' [I'll't

wN.m m.mvmna >¢.ma H.¢¢Mha .Nm

 

.IIIIIIJ"|yI"||‘;I|v‘III. .Il'l.4.- I ll Ill‘ |||l|||l|ulll

 

mm.OH 0.0mmba mw.mH 0.0mmha .HN
ma.aa m.mmHhH mm.aa w.mthH .om
mm.ma O.Hmona mm.aa m.wvo>a .ma
mm.ma m.omme om.oa N.ommmH .ma
em.ma m.ommoH me.ma m.Hmm©H .ma
¢h.mH m.mvhma mm.ma m.omhma .ma
mm.ma N.om©©H mm.vm «.mvowa .ma
NH.BH N.vmmmH mm.mm m.mvmoa .VH
mm.ma m.omvma mm.mm m.omvma .ma
wh.mH N.ommmH ow.©N m.ommoa .NH
0H.NN h.hVN©H mm.mm «.mvmwa .HH
vm.vm m.HmH®H Nv.vm 0.0meH .oa
mm.mm o.mvomH mm.mm m.mvooa .m
©N.ON N.HmmmH «m.mm b.0mmma .m
me.ma m.¢mmmd em.mm w.ammma .5
Hv.NN o.mmhma v0.5m m.HmhmH .w
mm.mm h.¢mme mo.hv m.mmwma .m
mh.mm o.mmmmH mm.vv 0.0mmma .v
mm.vm o.vamH mo.wv w.amvma .m
mm.mm o.vmmmH mm.oo h.HmmmH .N
hm.hN m.mmmma .H
:80 mmwv wocmsvmum A :60 mmmv «xocmsvmum le
uflmcmucH :oflumufloxm uflmcmucH aoflumufloxm

mum m5muo> >8 oovl Amomv >E cowl

' l«|;.|‘"|‘ .I .lll-l-1| ‘-.. "l‘ 1"... II" lll’lllll'gdlllll-A'II'I‘lnllll‘.ll.lt‘lu|.\|‘l£1l‘“ i, . ‘-l"lll 1! l O.

 

+.wom mswum> >8 oovl mam com:
um um>HHm co mamahnuwAaxwfiuwmlvvmHQIN.HIu mo Esuuowmm :mEmm mnu
CH Anw>v mafia HIEo mmo mnu mo mHHmoum coflumufluxm ucwwcwmma Hafiucmuom um mqmde

109

-600 mV. This is a manifestation of the combined increase
in concentration of C104- ion near the more positive
surface at -400 mV with the decrease in the number of
adsorbed t-BPE molecules on moving to positive potentials.
The important point to be taken from the potential
dependence of the excitation profile is that the peak
shifts to lower energy with the corresponding shift to
more positive potentials. The peak in the profile at
-400 mV occurs at 16150 cm-1 relative to the 16500 cm-1
peak in the -600 mV profile. The shift to lower energy
at positive potentials is consistent with the appearance
of the n* state in the -50 mV spectrum of t-BPE on silver
(Figure 7) and its disappearance on returning to -600 mV.
That is, stabilization is sufficient at positive poten-
tials for n* to become an occupied state but not at more
negative potentials.
Frequently in SERS studies researchers have reported
potential dependence of surface Raman intensities and
drawn conclusions about surface coverage and other infor-
mation essential to analysis of the enhancement mechanism.
The potential dependence of the excitation profile of t-BPE
at a silver electrode demonstrates that such conclusions
can be premature and misleading without knowledge of the
shifts in intensity expected by excitation profile arguments.
The excitation profiles reported here were determined by
continuous variation of excitation frequency with spectra

1

recorded every 100 cm- (~H4Jlnm). In other excitation

110

profiles reported in the literature (see Chapter 2 for
references) spectra have been reported in intervals of
500-1500 cm’l. The full width at half height (FWHH) of
the peak observed in the excitation profile of t-BPE on
silver is approximately 300 cm-1. Therefore, early
profiles reported may actually have missed important
structure in the excitation profiles and consequently
the density of states for adsorbates at a metal surface
has not previously been recognized as large. Most
researchers have concentrated on explanation of the

increased enhancement to the red which can be easily

understood if the density of states is high.

Light Coverage Studies and the Raman Spectrum of an

Ag (t-BPE) Complex.

 

Another requirement for the observation of the irrever-
sible potential-dependent behavior observed for the t-BPE
surface Raman spectra in Figure 7 is that the adsorbate
coverage must be heavy. As mentioned earlier, when the
bulk concentration of t-BPE prior to the oxidation-reduction
cycle does not exceed 500 ng the adsorbate coverage of the
electrode is light and only reversible potential dependence
of the SERS spectra is observed. Some insight into the
effect of light coverage can be gained by analysis of the
potential dependence of the light coverage spectrum of
t-BPE as shown in Figure 12. The spectra shown were excited

with 6578 A radiation (Rhodamine GG emission) although any

111

Figure 12: Potential Dependence of the Surface Raman
Spectrum of a Light Coverage of t-BPE on
Silver.
EXCITATION WAVELENGTH: 6578 A (Dye Laser

Emission).

112

 

35...... ...—.2538 ... .. E... 833“...
a 7r»....m......E.E....=i§. .... R.
Q 323......" «mun»

_

1200

 

2D

   

“‘3
‘ :‘ f" 1013
3655

~_-%
- .TD 1839
‘__>
D ..3
—“‘ 1&4

.2)
.)

(cw

  

 

 

93S

 

 

 

«3.552 3...:

113

wavelength excitation of light coverages produces only
reversible potential dependence of the observed surface
spectrum. An interesting feature of the potential depen-

dence shown in Fig. 12 is the dynamics of the 655 cm-1

. . . . . -1
line. On mov1ng to p051t1ve potentials the 655 cm
line decreases in intensity and shifts to higher frequency,

coalescing with the 664 cm.1 line (v ). This strengthens

6b
the previous assignment of the 655 cm-1 line to motion 6b
in the ring attached to the silver surface as discussed in
an earlier section of this chapter. A similar b-type
mode, 9b, observed at 1244 cm.1 in the surface spectrum
exhibits similar potential-dependent behavior, decreasing
sharply in intensity on going to positive potentials.

This and other changes in the surface Raman spectrum
of a low coverage of t-BPE can best be understood by com-
parison withtflmenormal Raman spectrum of a polycrystalline
silver complex of t-BPE shown in Fig. 13. The stoichio-
metry of the complex has not yet been determined but some
conclusions can still be drawn in the comparison. In the
polycrystalline silver complex spectrum the 1201 cm.1 line

(C h - ¢ symmetric stretch) appears at the same intensity

et
as the v1 mode at 1025 cm-1. Similarly in the surface
spectrum of t-BPE at -100 mV versus sce the two vibrations
appear with similar intensity. Two lines appearing at
1037 cm.1 and 1040 cm-1, probably split by site symmetry

in the crystalline matrix, in the polycrystalline complex

spectrum can be assigned to the pyridine ring vibration v12

114

Figure 13: Normal Raman Spectrum of a Polycrystalline
Silver Complex at t-BPE (Agn (t—BPE)m) .

EXCITATION WAVELENGTH: 6471 A (Kr+).

 

115

 

 

Polycrystalline Silver Complex of

t-1.2-bis(4-pyridyl)etllylene

 

 

 

 

 

[02! ~'—4~—_._.._l

 

 

{[91

 

 

1
200

1
400

J -
600

1
800

cm“

4
1200 1000

1400

1600

3100 3050

 

116

..mezHBzou v mamme. .coHusHOm NHUNSU cH ommH um muoooo coHumubH>e
musmHm .um>HHm co

mcmecumAHmpHummlvvanIm.HIu wo uon HMHucmuoqucmuuoo mzu CH 858me8 HmooH mo ucHom

+

.NH. mnomHm cH popuoowu mum muuommm womum>oo ubmHH one
.4.

 

 

 

 

 

HmMH mmmH ommH ¢H
mHMH mHMH mmmH m
«mmH mmmH mva bm
*mNNH mmmH vmmH Mm
mmHH anH HomH .uum e I U
mmOH moOH vBOH omH
ovOH NH
VMOH NMOH
mNOH H
mom MHoH mHOH
Ham eem no.
mum mmm mmm mOH
Hue moo vow
mmm no
mmm .hum ZIm<
ovm omm vmm .HQH> Hmummem
omm .uum Huumm
HNH MQOHH mvH
mm mopoz
Hm moHuumq
me mm
mmqu .wom. >8 ooea um E.mmmuu.cm¢ #:OmHHz
mcHHHmummuowHom um>HHm co mmmlu mcHHHmummuo>Hom uswEQmHmm¢

Illf .I..|Il \...I. ‘1’ I'll. I, ll Cl II

.ESHuoQO cm8mm

mmmn Iu mcHHHmummuoxHom on» on cam .mom >8 oovn um .ewmmuw>oo uanHv Hw>HHm co

mcmHhcuwfiprHummI avaQI N HI u mo Eouwommm cm8mm may ou mCOmHHMQEOo pom wcmH>£um
IAH>©HH>QI vvaQI m HI u mo waQEOo um>HHm mcHHHmuwwnowHom m mo Eouuommm CMEmm no mqmée

117

 

 

 

 

«mom

mmom
«mom OMHm coumupm
MHom owom mIU
mmoH vmmH oqu .upm UHU
hmmH mooH MHoH om
mmmH v + on
mva mmmH mmmH am
HmvH quH mmvH m.mH
NNvH HmvH vmvH an
mmmlu .mom. >5 ooen um E.mmmuu.cm< e comHHz
wCHHHmHmmuowHom Hm>HHm so mmmlu mcHHHMHmwuomHom #:08cmHmmfi

 

. DMDZHBZOU o mqmflfi

118

which occurs in the Raman spectrum of pyridine at 1037 cm-1
and is the most intense line in the spectrum. Vibration 12
is conspicuously absent from the heavy coverage spectrum
but appears weakly at 1035 cm.1 in the light coverage
spectrum, reaching maximum intensity at the -400 mV
potential. The appearance of v12 in the light coverage
surface Raman spectrum of t-BPE and its appearance in the
Raman spectrum of the polycrystalline silver complex
strengthens the hypothesis of the existence of a surface
complex. The C=C symmetric stretch at 1639 cm—1 attenuates
in intensity going to positive potentials and us currently
not completely understood. It seems that most of the
scattering observed in the light coverage originates in

the part of the molecule closest to the surface. Thus,

the Ce - ¢ symmetric stretch (1200 cm-1) and the C=C

th
symmetric stretch are attenuated at positive potentials

and the v1 vibrations (1013 cm-1) appears to narrow. The
reason for more intensity arising from the closer part of
the molecule can only be decided when the state in which

the excitation wavelength is in resonance (assuming a
resonance Raman enhancement mechanism) becomes known. There
are several electronic excited states which, by using
typical building up principles, exhibit electron density,
only at the rings and not in the ethylene bridge. Resonance
Raman excitation to such a state would be expected to

result in increased intensity for the bound ring only. The

attenuations of the above mentioned line intensities may

119

not be observed in the heavy coverage because of intense
scattering from multilayers.

Although some of the features of the light coverage
surface Raman spectra are not completely understood other
features can be utilized to understand the requirement of
heavy coverage for the observation of the "flat" configura-
tion. At positive potentials the silver surface atoms have
developed nearly a unit positive charge per silver atom.

The silver atom in the polycrystalline silver complex is,
of course, formally in the +1 oxidation state, experien-
cing a full positive charge. Withealight coverage of t-BPE
molecules at the surface, chloride and perchlorate ions

are expected to occupy more surface sites. The increased
concentration of anions at the positively polarized surface
acts to screen the positive charge seen by the neutral t-BPE
molecule at the surface. The screening effect is invoked
to account for differences between the light coverage
spectrum and that observed for heavy coverages and for the
polycrystalline silver complex.

Modes with components of motion orthogonal to the
direction of the attraction force at the surface (i.e.
those modes in the end-on configuration which are orthogonal
to the surface normal) and are not expected to be influenced
by charge fluctuations at the surface. However, modes with
components of motion along the direction of the attractive

force are expected to be influenced by charge fluctuation

120

at the surface. Modes which involve asymmetric stretch
along the surface normal (end-on configuration) demonstrate
the largest sensitivity to surface coverage. Examples of
this are now discussed. Vibration 9a, which appears at
1234 cm.1 in the polycrystalline silver complex Raman
spectrum demonstrates asymmetric stretching along the
surface normal. This vibration appears at 1233 cm.1 in

the heavy coverage surface Raman spectrum of t-BPE at

-100 mV but at 1225 cm-1 in the light coverage spectrum.
Similarly, vibration 19a appears at 1498 cm.1 in the
polycrystalline silver complex spectrum and at 1497 cm-1

in the heavy coverage spectrum of t—BPE on silver at

-100 mV. The motion 19a appears at 1493 cm-1 in the

light coverage spectrum shown in Figure 12. These observed
shifts and the shifts of other a-type modes in the light
coverage spectrum relative to the heavy coverage and
complex spectra are therefore understood by considering
the reduced positive charge seen by the light coverage

of t-BPE molecules owing to anion screening.

This screening is undoubtedly responsible for the lack
of reorientation of the t-BPE molecule in a light coverage
under 5145 A irradiation. As discussed in the previous
section, reorientation of the t-BPE molecule occurs through
excitation into a n* orbital which, owing to the proximity
of the positively charged surface, experiences a force along
the coordinate connecting the n* orbital with the silver

surface. With the positive charge on the electrode screened

121

by the presence of large numbers of anions in the light
coverage situation the force exerted on the w* electron
is reduced. The reduction in the force is apparently
sufficiently large enough such that reorientation cannot
be affected. Thus, under light coverage conditions, the
"flat" configuration never materializes (at any excitation
wavelength) and completely reversible potential-dependent
behavior of the surface spectrum is observed.

The effect of exciton splitting of an excited state
of one t-BPE molecule relative to the excited state of
its neighbor on the surface may also account for the
observed coverage-dependent phenomena. However, physical
evidence for both the exciton model and the screening model

are currently lacking and further study is needed.

Negative Potential Results and the Reduction Product of

 

t-BPE .

 

Interesting potential-dependent spectroscopic phenomena

are also observed on moving to potentials negative of

-600 mV. As shown in Figure 14, the surface Raman spectra
of t—BPE develops features characteristic of the "flat"
configuration. This is observed by comparing the "end—

on" spectrum at -400 mV with spectra observed on moving to
-1.0 V versus sce. In the spectrum observed at -800 mV
intensity has developed at the 681 cm-1 line (vll' Au)

1

with decreasing intensity at 655 cm— . This indicates

that the "end-on" configuration is diminishing in

Figure 14:

122

Potential Dependence of the Surface Raman
Spectra of t-BPE Potentials -400 mV, -600 mV,
-800 mV and -1.0 V versus sce.

EXCITATION WAVELENGTH: 5145 A (Ar+).

123

 

- 1639

“C=C on I; 0.5! cm;-

. >16“

at -1.0 V v5.50!

' > 1342

1320

 

IIIEISIIV

1 “ 600 .v
J VMW‘ wry),

- 400 IIV
‘i

J Wu”

 

cm“

£00 1

|

‘ 5- 1010

...“,H

__._ l____J - ,.____1_____
1600 1500 1400 1300 1200 1100_ "1000

124

prominence and the "flat" configuration takes its place.
The increase in intensity of the vibrational lines in the
region 1400-1560 cm.1 is further indication that the rings
are beginning to lie down on the surface and experience
electron density overlap with the metal. The increased

1

intensity at 1320 cm- (C - H in-place, in-phase,

eth
deformation) is consistent with that observed in going

to the flat configuration. By comparing the relative
intensities of the C=C symmetric stretching vibration at
1639 cm.1 and the intensity at 1200 cm-1, it is apparent
that the 1639 cm.1 band becomes attenuated relative to the

1

C - ¢ stretch at 1200 cm- . This may be indicated of

eth
the electronic state from which the Raman signal is origi-
nating from.

At potentials near the p.z.c. (point of zero charge)
of silver (i.e. near -900 mV sce in 0.1 g Cl-) it might be
expected that electronic excited states would become
occupied by transfer of electron density from the metal
to the adsorbate. The lowest such excited state for t-BPE
can be determined by using known electron density distribu-

111 and

tions for the two rings and the ethylene bridge
applying building-up principles. The most probable lowest
excited state is one of B9 symmetry which therefore can

contain a contribution from the ethylenic n* orbital which
transforms as Bg symmetry in the molecular C2h point group.

In this state there is no electron density at the ring

carbon involved in bridging to the ethylene carbon. Thus,

125

no overlap will exist between the W* orbital and the ring
carbon at the 4-position. The force constant of the

Ceth - ¢ bond is therefore not expected to be different
from that in the ground state. Accordingly, the Ceth - ¢
symmetric stretch at 1200 cm.1 is not attenuated by a shift
to another frequency even though the system is in a n*
state. The C=C symmetric stretch, however, is expected to
be attenuated since the force constant of the bond between
the ethylene carbons will change and move under the sym-
metric ring stretch near 1600 cm_1.

A new peak appears in the negative potential surface
Raman spectra of t-BPE at 1291 cm-1. This mode can be
assigned to the ethylenic hydrogen out-of-phase, in-plane
deformation (Bu) because of its appearnace at 1289 cm-1
in the infrared spectrum of t-BPE as observed by Katsumoto.l
Interestingly this mode does not appear in the spectrum of
the "flat" configuration at -600 mV as shown in Figure 7.
This may indicate that the state bound in the "flat"
configuration is a state which contains only in-phase
electron structure (i.e. is of A type symmetry) while
the state observed on sweeping to negative potentials is
probably an out-of-phase state as denoted by the B9
symmetry assigned above.

The reasons for the molecular reorientation at negative
potentials are similar to those for reorientation under
5145 A irradiation at -50 mV (Fig. 7). With the occupation

of a n* orbital at negative potentials reorientation may

126

occur in order that better adsorbate-metal overlap be
affected. The force experienced by the w* electron in the
vicinity of the still relatively weakly positive electrode
surface is undoubtedly small and accounts for the minimal
amount of reorientation observed spectroscopically. Reori-
entation of the end-on molecules is further facilitated by
decrease in surface concentration of anions from the
supporting electrolyte. The surface sites vacated by
anions at negative potentials can be filled by molecular
reorientation of the neutral t-BPE molecule at the surface.

At potentials negative of the p.z.c. (i.e. -l.0 V
versus sce) anomalous spectroscopic changes are observed
which further indicate electron transfer from the metal to
the adsorbate. The totally symmetric ring motion (8a) at
1604 cm.1 becomes attenuated in intensity accompanied
by appearance of a new line at 1560 cm-1. The high fre-
quency side of the C=C symmetric stretch at 1639 cm.-1
develops a shoulder in the surface Raman spectrum of t-BPE
at -l.0 V. Some distortion in the 1240 cm-1 range is
also observed. These changes can be best understood by
continuing to alter potential in the negative direction.
At -1.2 V versus sce an obviously distinct surface Raman
spectrum of t-BPB is observed (Figure 15).

In Figure 15, the 1561 cm.1 peak, which was developing
in the spectrum at -l.0 V, has become the most intense line
in the spectrum. This line is assigned to the totally

symmetric pyridine ring motion 8a in the "new" molecule by

Figure 15:

127

Surface Raman Spectrum of the Two-Electron
Reduction Product of t-BPE on Silver at
-l.2 V versus sce.

EXCITATION WAVELENGTH: 5145 A (Ar+).

128

.....e

GOG GOG." O0: OONH OOMH 09v.— OOmH cow.— OONH,
_ _ _ _ _ _ _ H _

 

 

 

299!

 

 

....3 3.2.3.... €2.88
... .... .2. .. w. .. . . .... . l e

 

 

 

 

 

 

...Olwl. _.

I99!

.. 3H

L! r n —

 

 

 

 

 

l;

AIISNJINI

 

129

TABLE 7; Raman Spectrum of the two-electron reduction
product of t-l,2-bis(4-pyridy1)ethylene on
silver at -l.2 v versus sce. Excitation: 5145A.

 

 

 

Vibrational
Frequency Suggested Assignment (Wilson #)
(cm’ )
641 (6b) in—plane ring deformation
871 10a out-of-plane CH deformation
1005 (1) symmetric ring stretch
1071 (18a) in-plane C-H deformation
1085 (18b) in-plane C-H deformation
1150 (15) in-plane C-H deformation
1196 (9a) in-plane C-H deformation
1214 ?
1232 ? central carbon symmetric stretch
1245 (9b) in—plane N-H deformation
1287 out-of-phase Ce-H deformation
1330 14 ring stretch
1403 19b ring motion
1440 ?
1507 (19a) ring stretch
1540 (8b) ring stretch
1561 (8a) symmetric ring stretch
1593 ? unreduced t-BPE vibration
1630 ? unreduced t-BPE vibration

1652 Ce-C¢ stretch

“_

130

reason of the attentuationcxfthe 1604 cm-1 line at -1.0 V

accompanying the growth at 1560 cm-1. It is clear that the

very intense Ceth - ¢ symmetric stretch occurring at 1200<mn—l
in the -l.0 V spectrum is absent from the 1200 region in the
surface spectrum at -l.2 volts. The shoulder observed on
the high frequency side of the ethylenic C=C stretch in

the -l.0 V spectrum has developed into a sharp line at

1652 cm"1 in the -1.2 v spectrum and the 1639 cm‘1 line

is essentially completely attenuated. The high frequency
motion at 1652 cm.1 is clearly a C=C stretch. These changes
and others observed in the -1.2 V surface Raman spectrum

of t-BPE on silver can be understood by first examining

the current-potential profile of t-BPE which is inset in
Figure 15. At approximately 1.1 V versus sce the his
pyridyl ethylene molecule undergoes a two-electron reduc-
tion. This reduction was examined by Volke and Holubek110
in 1962. They found that in solutions of neutral pH the
yellow reduction product of t-l,2—bis(4-pyridyl)ethy1ene

was unstable and quite different from the ethane reduction
products of less symmetrical isomers than the 4-substituted
bipyridyl ethylene. The difference was manifested by
reoxidation of the yellow reduction product and the
reappearance of the original reduction after electrolysis
which indicated a reversible electrode process. The ethane
reduction product of t-1,2-bis(4-pyridy1)ethylene was only

obtained by reduction in acidic solutions. Volke and

Holubek suggested the structure of the reduction product

131

to be that shown in Figure 15. The suggestion was that the
two electron reduction involved protonation at the nitrogens
in the pyridine rings rather than at the central ethylene
carbons. The 1962 reference,110 however, offered no
experimental evidence of the proposed structure of the
reduction product. The surface Raman spectrum of the
reduction product shown in Figure 15 is the first physical
evidence to support the suggested structure. The 1652 cm-1
line can be assigned with confidence to the C=C symmetric
stretch betweentflmaethylene carbon and the 4-position ring
carbon. The increased force constant relative to the
ethylenic C=C symmetric stretch (1639 cm-1) is consistent
with the large C=C force constant in substituted ethylenes
(R2C=CH2) in which one end of the double bond has a carbon
which is bonded only to carbon atoms. The symmetric C=C

stretch occurs in the range 1658-1644 cm in such com-

pounds.103 Therefore, the absence of the single bond

C - ¢ stretch at 1200 cm”1 is also accounted for.

eth

Other assignments in the surface Raman spectrum of
the t-BPE reduction product can be made. The sharp line
at 1196 cm.1 is probably vibration 9a in the rings of the
reduction product. Vibration 9a appears to move consis-
tently to lower frequency in going to negative potentials
as observed in Figure 14. There 9a appears at 1224 cm-1
in the end-on spectrum at -400 mV but shifts under the

C - ¢ stretch at 1200 cm.1 in the -1.0 V spectrum.

eth
The sharp line at 1232 cm”1 is assigned to the symmetric

132

stretch of the central carbon single bond in the reduction
product for lack of a better assignment for the relatively
intense line. Other modes in the reduction product do not
appear significantly shifted from their positions in the
unreduced t-BPE surface Raman spectrum. For example, the
intense line at 1005 cm.1 can certainly be assigned to the

totally symmetric ring motion v in the reduction product.

1
The appearance of a weak line at 1440 cm”1 is interesting
in comparison to the spectrum observed at -50 mV which
contained W* vibrations. In that spectrum, a line was

1 which eluded assignment. The

observed at 1449 cm-
1449 cm.1 was characteristic of the n* state in that it
disappeared upon returning to -600 mV. The reduction
product of t-BPE is related to the occupation of n*

orbitals at positive potentials. The reduction product

is, of course, the result of addition of electrons to the
t-BPE molecule which necessarily occupy fl* orbitals in the
unreduced molecule. Thus the 1440 cm-1 mode in the Spectrum
of the reduction product and the 1449 cm."1 mode in the

-50 mV spectrum of t-BPE are probably the same normal
coordinate and assignment as such fortifies the hypothesis
of presence of n* states in the surface Raman spectrum of
t-BPE at -50 mV under 5145 A irradiation. Suggested

assignments of many of the observed lines in the t-BPE

reduction product spectrum are listed in Table 6.

133

Monoprotonated t-BPE on Silver.

 

In Figure 16 the surface Raman spectrum of t-BPE was
recorded in an electrolyte of pH 5.6. The top spectrum in
the figure was recorded under 5145 A excitation and demon-
strates some significant differences from the usual spectrum
of t-BPE on silver at -50 mV. These results were observed
for low coverages of t-BPE on silver and therefore at
-50 mV only the end-on configuration is evident.

The differences between the pH 5.6 spectrum (top,
Figure 16) and the neutral pH spectrum are consistent with
spectroscopic changes observed in comparison of the normal
Raman spectra of t-BPE and its dihydrochloride. For
example, in the solution spectra of t-BPE (Figure 6) and
its dihydrochloride the symmetric Ceth - ¢ stretch is
observed at 1195 cm-1 and 1210 cm-1, respectively. In the
surface Raman spectrum of t-BPE at pH 5.6 (5145 A excita-
tion) the 1200 cm.1 line (Ceth - ¢ stretch) appears markedly
attenuated and the neighboring 1222 cm-1 line appears to
have gained intensity. The shift in the Ceth - ¢ symmetric
stretch to higher frequency on formation of the hydro-
chloride can account for these intensity variations. A
new line in the SERS spectrum of t-BPE at pH 5.6 occurs at
1582 cm-1 at -50 mV. This line can be assigned to motion
8b in the protonated ring by virtue of the shift to higher
frequency of mode 8b in the normal Raman spectrum of t-BPE
upon formation of the hydrochloride. In t-BPE mode 8b

appears at 1549 cm-1 in the normal Raman spectrum and at

134

Figure 16: Excitation Wavelength Dependence of the
Surface Raman Spectrum of Monoprotonated
t-BPE on Silver at -50 mV (sce) .

ELECTROLYTE pH: 5.6.

135

 

 

hue—

flood

00‘—

-50 mv vs. SCE

\

{H

3.2.8.. W.

 

 

QMOH

 

 

 

1100 1000

1200

4400

1500

 

 

>b_mzuh.2_

cm"

 

136

 

 

 

 

TABLE 8: Raman Spectrum of Monoprotonated t-l,2—bis(4—
pyridyl)ethylene on silver at -50 mV and
-600 mV at pH 5.6 and at -600 mV after
base (OH’) is added. Excitation: 5145 A.
Region: 1000-1700 cm-l.
Monoprotonated _
Assignment t-BPE on Ag Base (OH ) Added
Wilson # —50 -600 mV -600 mV
1 ‘ 1017 1006 1012
18a 1063 1056
18b 1100 1095
Ce-¢ sym. str. 1200 1199 1201
9a 1222
9b (NH def) 1242 1243 1246
Ce-—H def.(out-of 1287 1286
3 phase) 1316 1322
14 1340 1338 1338
4 + 11 1414 1409
19b 1427 1424 1422
19a 1495 1490 1491
8b 1545 1538 1543
8b (outsof- phase) 1555
8b (protonated ring) 1582 1574
8a 1607 1603 1606
Ce=Ce sym.str. 1639 1637 1639

 

_.* -__ .. -——— _—.._v_ H.

137

1605 cm-1 in the dihydrochloride. The increased intensity
of mode 18b at 1100 cm"1 in the surface Raman spectrum of
t-BPE at pH 5.6 can be understood in that 18b is expected
to couple to N-H deformation. Thus it appears that the
t-BPE molecule on the electrode surface is monoprotonated
at pH 5.6.

This conclusion is not unreasonable. In the end-on
configuration the t-BPE molecule has one nitrogen lone
pair extending into the electrolyte solution. Lavalee and
Fleischer88 report values of 5.92 and 4.49 for the pK2

and pK values (negative logarithm of the acid dissociation

l
constants), respectively, of t-BPE. Though the acid
dissociation of the complex (t-BPE at the silver surface)
is expected to be somewhat different from the dissociation
constants in the ligand, the difference is probably small
owing to several factors. In the pentaminebispyridylethy-
lene ruthenium II complex the pKa is observed to be

5.0 i 0.1.88 For t-BPE on silver the charge built up on
the silver atom is significantly lower than that on the
ruthenium atom in the +2 oxidation state. Also, as
discussed earlier the increased concentrations of
electrolyte anions at the surface are expected to screen
the full positive charge on the electrode surface. Thus it
is expected that the pKa of the t-BPE molecule—silver
surface complex will not be significantly lower than the

pK2 of the t-BPE ligand alone. Future surface titration

experiments should clarify this point.

138

Some further evidence that the surface spectrum
observed at pH 5.6 is that of the monoprotonated t-BPE
molecule is presented in Figure 17. There the potential
dependence of the pH 5.6 surface spectrum and the spectrum
observed after addition of base (KOH) are shown. The
appearance of the characteristic lines of the protonated
ligand (i.e. 1574 cm-1, 1095 cm.1 and other changes) at
-600 mV as well as at -50 mV shows the insensitivity of
the pKa value on the true polarization at the electrode
surface. The disappearance of the lines, characteristic
of the monoprotonated t-BPE-silver complex, upon addition
of base is very convincing evidence for the protonation
phenomenon at the silver surface.

Returning to Figure 16 briefly, some interesting exci-
tation frequency dependence for the monoprotonated spectrum
is discovered. The lines characteristic of the monopro-
tonated species are absent under red excitation but grow
in intensity with increasing frequency of the exciting line.
This is consistent with the notion that for light surface
coverage at positive potentials most of the light scattering
intensity originates in the parts of the t-BPE molecule which
are closest to the silver surface. The intensity from the
ring more remote to the silver surface increases only as a
consequence of coming into resonance with a weak n+w*
transition in the visible absorption spectrum of both t-BPE
and its dihydrochloride (see Figure 9). This evidence also

corroborates the hypothesis that the n+n* transition from

139

Figure 17: Surface Raman Spectrum of Monoprotonated t-BPE
on Silver at -50 mV (sce), -600 mV (sce), and
at -600 mV (sce) after Addition of Base
(of) .

EXCITATION WAVELENGTH: 5145 A (Ar+).

 

 

140

 

lNTEllSITY

  

 

 

Monoprotonated t-BPE on Silver.
Excitation: 51451

1201

 

 

 

i
E
I
s
i

 

 

1700

1500 1500 1400 1300 1200 1100 1000

141

the outside ring can still occur though the distribution of
electronic states in the ring closer to the silver surface
is significantly perturbed from the distribution of states
in the molecule in solution. That the electronic states
of the section of the t-BPE molecule not experiencing
electron density overlap with the silver surface in the
end-on configuration, are n93 significantly perturbed is

a requirement for the reorientation mechanism proposed in
earlier sections of this chapter. The lines which have
been interpreted as characteristic of the monoprotonated
t-BPE molecule on the surface also appear in the heavy
coverage spectrum observed at -50 mV under 5145 A excita-
tion i.e. the spectrum of the flat molecule containing n*
vibrations. This spectrum is shown in Figure 18. An
interesting experimental fact observed in the studies of
the monoprotonated surface spectrum of t-BPE on silver

is that upon addition of sufficient acid all surface Raman
signals, identifiable as originating from t-BPE molecules,
disappear. It is not difficult to understand this fact.

A doubly—protonated t-BPE molecule is quite positively
charged and is not likely to approach the positively
polarized electrode. Also protonating both nitrogens
eliminates the possibility of end-on complexation to the

metal surface.

Influence of C10,- on the Surface Spectrum of t-BPE.

 

The addition of 0.5 N perchlorate for purposes of use

as an internal standard in excitation profile experiments

142

 

Figure 18: Surface Raman Spectrum of Monoprotonated t-BPE
on Silver at -50 mV (see) in the "flat" 1
Configuration.

4..
EXCITATION WAVELENGTH: 5145 A (Ar )

 

 

 

1143

 

 

t-BPE on Silver at - 50 mv SCE, pH5.6, 1” 51451 .

N
o
o
l

600

 

l
700

l
800

058
I 88

 

 

l
1000

810! z

0901
'90!

1.60! —

1100

 

l l l
1400 1300 1200

1
1500

 

(.09!

 

1600

 

 

 

‘ Iusuaml

 

1700

cm

144

has some important influences in the Raman spectrum observed
for t-BPE on silver. In general, the screening of the
positive charge developed at the silver electrode surface

is invoked as responsible for the effects produced by the
presence of ClO4-. For the potential dependence of the

t-BPE surface spectrum in the negative potential range

effects observed in simple 0.1 g chloride electrolytes are

4

to negative potentials t-BPE molecules in the end-on posi-

exaggerated with 0.5 5 C10 present. That is, on going
tion "lie down" at more positive potentials with half molar
perchlorate ions present than in their absence. In both

0.5 I_I 0104’/0.l I_I Cl" and 0.1 I}; 01" alone electrolyte
solutions when altering potential in the positive direction
the surface Raman signal is attenuated owing to the dis-
placement of adsorbed t-BPE molecules by anions at the
metal-solution interface. This reduction in surface Raman
intensity at positive potentials is understandably augmented
by the presence of 0.5 g C104-.

In the spectrum of t-BPE "flat" on the silver surface
some very interesting influences of the 0.5 g C104- presence
are observed. After initiating the molecule reorientation
of t-BPE molecules from the end-on configuration to the
"flat" configuration at -50 mV under 5145 A excitation the
potential was moved to -600 mV and the spectra shown in
Figure 19 were observed. In 0.1 g Cl- only the spectrum
at ~600 mV typifies what has been assigned to the "flat"

spectrum in earlier sections of this chapter. Relative

145

Figure 19: Influence of 0.5 g ClO4 on the Observed
Surface Raman Spectrum of t-BPE "flat"
on the Silver Surface at -600 mV (sce).

EXCITATION WAVELENGTH: 5145 A (Ar+).

 

146

 

INTENSITY

"“F:>1039

 

1001

 

1554
1530

CH-

 

1480

__;>1494

1425

Influence of cro; on the spectrum of
t-l,2 -bls-(4-pyrldyll_-ethylene 'flat' on A:
at -600 Mi! vs. SCE.

Excitation: 5:45 A

1325

1330

230
1242

.SM era;
.1» cr §

.1 M cr only

 

 

l l l“. “a
‘ ' A
.N J I
l V 2 .I' f
2 I / v a
i v w
l ' l l
'J l 0\ fl
l'
! I! [I \‘ l ‘l V
V g ‘ I" I
//,// V/ \\.H// 1
wfl\
\J l
l l I l A l l 1 n l \J l
1500 1500 1400 1300 1200 1100 1000

 

 

 

147

to that spectrum, in the spectrum observed at the same

4

differences are evident. The intensity of the ethylenic

potential with 0.5 M C10 also present some distinct

C=C symmetric stretch at 1639 cm"1 remains attenuated in
the -600 mV spectrum in 0.5 M ClO4-/0.1 M C1- while regain-
ing intensity in 0.1 M Cl- alone. Also, noticed in the
perchlorate spectrum is the appearance of a line at

1288 cm.1 which also appears when the t-BPE molecules lie
down at negative potentials (earlier section) and which
has been assigned to the out-of-phase component of an
in-phase ethylenic hydrogen deformation. The weak line
at 1120 cm-1 is probably the pyridine ring motion 015.
The reason for these differences could rest in the idea
that the state that is observed is the highest occupied
molecular orbital (HOMO). The HOMO in the case with

0.5 M C10 - present may be different from that when only

4

0.1 M Cl- present as a result of screening. The high
concentration of anions diminishes the effective charge at
the surface and thus stabilization of different states

may result. The state observed in the 0.5 M ClO4—/0.1 M Cl-
solution is probably the same state that begins to appear

at negative potentials with the t-BPE molecules in the
end-on configuration. That is, a H* state which involves

no overlap of electron density between the n* orbital and
the closest pyridine carbon atom which consequently causes

no shift in the Ce - ¢ stretch (1200 cm-1) but shifts

th

the C=C stretch to lower frequency, attenuating the line at

148

1639 cm-1. The spectrum observed at -600 mV with only
0.1 M C1- present may involve electron density mixing with
an unexcited n state since neither of the lines at

l

1639 cm- and 1200 cm.1 are shifted or attenuated in

intensity.

Photodegradation of t-BPE Multilayers on Silver.

 

In the observation of the excitation profile of t-BPE
on silver some anomalous frequency dependence is observed.
Figure 20 demonstrates this anomalous behavior. The
intensity of pyridine ring vibration v11 appears to increase
in intensity with increasing excitation wavelength. Ini-
tially this phenomenon was attributed to the distinction
of the red state appearing in the excitation profile. That
is, it was thought that v11 was effective in vibronic
coupling to the red excited state but not to other higher
energy excited states. However, further investigation
proved these assumptions erroneous. The 683 cm"1 line
(v11) is absent from the surface Raman spectrum of a
light coverage of t-BPE molecules at the electrode-
electrolyte interface. The spectra recorded in Figure 20
were obtained with a heavy coverage of t-BPE molecues on
the surface. Also, as shown in Figure 21 the intensity of

the v vibration is photosensitive to red excitation. In

11
Figure 21 a series of consecutive spectra were recorded in
the 625-700 cm_1 region. The only variable changing in

this experiment was the time the sample was irradiated. It

149

Figure 20: Excitation Wavelength Dependence of the

v11 (683 cm-1) Vibration t-BPE on Silver at

-400 mV versus sce.

INTENSITY

 

Excitation Wavelength Dependence
of trans-l,2-bis-(4-pyridyl)-etlrylene
on Ag at -400 m vs. SCE

936 (cm)

>>%%%%

I

6431 jk
sssojk

 

 

 

cm"

151

Figure 21: Photodegradation of Multilayers of

t-l,2-bis(4-pyridyl)ethy1ene on Silver:

Loss of Intensity at 684 cm-1.

152

 

4 2.... 5.5.5... «523......

 

mum mflw mflm mum mum
. _ . H .
<9? i é 9
9
9’
S
S 9
8
.7
..Ill...
.....o em a «Re 2.22.4.3.“

....” ...... .... o3. ...
we .... o..o~....o...........-$1...- «....
... 2o 23...... ... 5:233:32...

 

AIISNJINI

 

 

153

1

is clear from the figure that the 684 cm- line is attenu-

ated in intensity with increasing irradiation tinmh However,
the neighboring vibrations 6b at 667 cm.1 and 655 cm”1
retain full intensity throughout the time of the experiment.
The phenomenon pictured in Figure 21 occurred over a time
range of 30 minutes at an incident laser power of 50 mW.
Obviously, there is evidence of photodegradation but the
signal from the molecules at the surface is not degraded.
The degradation must occur in the multilayers of t-BPE
molecules at the surface. This can be deduced from the
absence of the 684 cm-1 line in the light coverage spectrum
and the absence of degradation of the characteristic surface
mode at 655 cm-1. The photodegradation must necessarily

be accompanied by a visible absorption. Since the 684 cm.-1
line increases in intensity in the red so too must the
absorption be in the red. But the t-BPE molecule has no
visible absorption spectrum above approximately 600 nm.
However, in the multilayers the rapid adsorption of t-BPE
molecules to the electrode surface probably produced a
very low site symmetry seen by each t-BPE molecule. The
site symmetry in the multilayers is expected to be C1.
Therefore, just as the symmetry of the one-sided attachment
of t-BPE molecules to the silver surface is liable to

make forbidden transitions allowed so also will the low
symmetry in the multilayers. The actual transition could

also be an intermolecular one which for stacked flat

molecules is likely to occur along the stacking direction.

154

An out-of-plane mode, such as pyridine vibration v11:
has nuclear coordinate components only in the stacking
direction and may therefore be expected to couple
effectively to intermolecular transitions.

The visible signcflfphotodegradation in the multilayers
is the appearance of a burn hole in the surface coating
where the laser beam has been focused. The absence of
attenuation in the Raman signals from the molecules
directly associated with the silver surface (denoted by
the characteristic 655 cm-1 line) is probably a consequence
of the ability of the metal substrate to serve as a heat
sink, offering an alternative excited state decay pathway
to decomposition.

The appearance of surface Raman signals from the
multilayers should be treated as a caution signal in
interpretations. Rowe et a158 and other researchers have
reported surface enhanced Raman signals from multilayers.
The possible influence of site symmetry lowering of the
molecular symmetry on the absorption of incident radiation
was not addressed. The observation of surface enhanced
Raman signals from absorbed molecules not in direct contact
with the metal surface is inconsistent with the proposed
enhancement mechanism of surface complex resonance Raman

scattering offered in this thesis.

CHAPTER VII

Surface Enhanced Raman Spectra of
Amino Acids and Future Studies.

Surface enhanced Raman scattering has been shown to be
a versatile probe of dynamics at a metal-dielectric inter-
face. For t-l,2-bis(4-pyridyl)ethy1ene adsorbed at a
polycrystalline silver electrode nearly every vibrational
mode of the molecule, both infrared and Raman active modes,
has been observed in the SERS spectrum. Molecular orienta-
tion at the surface could be interpreted from irreversible
potential-dependent behavior of the surface Raman spectrum.
The structure of the reduction product was determined with
confidence. Similar studies can be applied to other
molecules.

The new researcher-controlled experimental parameter
of continuous electrode potential variations has direct
application to systems incorporating electron transfer. The
importance of electron transfer in biological systems makes
those systems logical candidates for study by surface Raman
techniques. Generally, Raman studies of large biologically
important molecules have been concentrated in resonance
Raman investigations of the chromophore. These studies are
very effective in determining contributions to UV/vis

absorption spectra by different chromophores and distinction

155

156

of similar chromophores in regard to substituents which

may couple to the electronic transitions. The influence

of the protein substructure on energy transfer is often
invoked to describe dynamics observed in the chromophoric
resonance Raman spectrum. The actual structure of many
proteins has eluded determination owing to the inability

to crystallize many of the large molecular weight
molecules. The dynamics of protein structure change during
oxidation and reduction at the redox center is essentially
unknown.

Because of the intense light scattering observed from
molecules adjacent to the metal electrode surface in the
SERS experiment it is reasonable to assume that the outer
amino acid residues of an adsorbed protein should exhibit
intense SERS signals. Thus, the SERS experiment may be
applicable to analysis of structure, amino acid sequence,
and energy transfer at the outer sections of a large
protein molecule. The combination of potential dependent
Spectroscopic phenomena and structure and orientation
interpretations with excitation wavelength dependence of
an adsorbed protein promises to make deep inroads into the
Imechanisms of energy transfer in biologically important
molecules. Of course, distinction of amino acid residues
and their orientation at the electrode surface are pre-
requisites to any interpretation of outer protein structure.
In this chapter the surface enhanced Raman spectrum of two

small amino acids, L-glycine (NHZCHZCOO-) and L-leucine

157

[(CH3)2CHCHNH COO-l are presented and interpreted in terms

2
of orientation at the electrode surface. It is hoped that
distinctions in the surface enhanced Raman spectrum of
these and other amino acids may be extrapolated to assign—
ments in surface spectra of adsorbed protein molecules in
the future.

The intensity of the surface enhanced Raman signals
from the adsrobed amino acids are significantly reduced
from the signals observed from t-l,2-bis(4-pyridyl)ethylene
on silver. The signals observed from the amino acids on
silver generally were in the range 3 x 103 counts per
second, or less, whereas the signals obtained from t-BPE on

5 to 3 X 105 counts per second. The

silver ranged from 1 X 10
reasons for the sharp drop in intensity are consistent with
the ligating ability of t-BPE, over that of the amino acids,
in conjunction with a resonance Raman enhancement mechanism
involving a silver complex. The normal Raman scattering
cross section of the highly conjugated t-BPE molecule
over that of the saturated amino acids examined here may
also account for the remarkable intensity of t-BPE surface
Raman signals. As mentioned previously, the hydrophobic
character of t-BPE undoubtedly assists in the facile
adsorptions of the bipyridine to the silver surface. The
amino acids are polar molecules and are therefore hydrophi-
lic.

Notwithstanding the low intensity of the surface Raman

signals from glycine and leucine some interpretations of

158

molecular orientation at the electrode surface can still
be made. The Raman spectra of L-leucine and L-glycine on
silver at -600 mV under 5145 A excitation are shown in
Figure 22 and the frequencies and assignments are listed
in Table 9. These spectra were recorded in a bulk pH of
11.0. Since at this pH both amino acids are above their
respective isoelectric points the molecules carry a nega-
tive charge associated withtflmacarboxylate group. No
surface Raman spectrum was observable from amino acids at
the silver electrode when the pH of the bulk solution was
below the isoelectric point of the amino acids. Therefore,
it can be assumed that the negatively charged amino acids
adsorb more strongly to the positively polarized electrode
than do the neutral compounds in agreement with the t-BPE
results.

The pH adjustment from neutral solution wasaccomplished
by addition of KOH. The formation of silver oxide (A920)
and silver hydroxide (AgOH) was possible in the hydroxide
media. In order to determine the presence of the oxides at
the surface and their influence on the surface Raman spec-
trum observed for the amino acid system the normal Raman
spectra of the two polycrystalline oxides was obtained.
Both A920 and AgOH are black opaque solids which required
a back scattering geometry for observation of the normal
Raman spectra. These spectra are shown in Figure 23 and
the frequencies listed in Table 10. The normal Raman

spectra of these two silver oxides have not been previously

159

Figure 22: Surface Raman Spectra of L-leucine and
L-glycine on Silver at -600 mV (sce).

+
EXCITATION WAVELENGTH: 5145 A (Ar ).

p—_ _ .

5145 I

Amino Acids on Silver at -600 mv VS. SCE.

Excitation

LLEI

 

l- leucine

1051

£09

119

£81
900

0Z6

0001
1201

9101

0911
£111

891!
0181

OZ?!
9771

£15!
6851

1291
£59!

160

 

l- Glycine

uogssgwa lugseI-uou

 

600

1400

1600

 

 

[IgsuaIuI

 

161

TABLE 9: Raman Spectra of L—Glycine and L-Leucine
adsorbed onto a polycrystalline silver
electrode at -600 mV sce, in the frequency
region 550-1700 cm’l. Excitation: 5145 A.

_——_—*_____—
_...-. - _.___.___..

 

 

 

Assignment L-glycine L-leucine
NH2 torsion 555 557
C00” wag 605 607
coo" bend 671
teflon? 736
790 783
806
CH2 rock 920
CH3 rock 1000
CCN asymmetry stretch 1035 1027 w
1078 1076
NH2 rock 1142 1140
stretch of CH-(CH3)2 1177
CH2 twist 1270 1268
CH2 wag 1310
C00' symmetry stretch 1380 1377
methyl group motion? 1424
CH2 bend 1450 1446
NH2 symmetry deformation 1498 1501
1577
C00' asymmetric stretch 1588 1589
NHZ degenerate deformation 1621

1653

 

 

162

Figure 23: Normal Raman Spectra of Polycrystalline A920
and AgOH Utilizing a Back-scattering Geometry.
EXCITATION WAVELENGTHS: 5145 A (Ar+) AgOH

6471 A (Kr+) A920

I.ma

0001

0091

0071

0021

009

009

002

163

INTENSITY

 

 

 

 

 

 

 

L- -— 1560
1510 I H-
Q? h
(:9 £r:
1300 = c
l
‘l
non-lasing emission I
l
1050
979
957
104 1 115
678
see
538
“E 492
he
3
403
332
348
non-lasing emission
290 280‘,69
232

1489

butane: Suglaueos noes

 

 

164

TABLE 10: Raman Spectra of Polycrystalline AgOH and AgZO
in the region 200-1600 cm- using a back
scattering geometry.

 

—_. __—__._._.—_--
w —.———-—. .. _—...-..- . .—

 

 

Polycrystalline Polycrystalline
AgOH AgZO
232 269
290 286

348

382

403

492

538

588

678 715

704 729

748

857

957

970

979 980

1050 1489
1300 1560

1510

 

165

:reported in the literature though the infrared spectrum

(of A920 has been reported.112"113

No attempt at assignment
of observed frequencies is made. Further study is required
for such assignment. Qualitatively, the effect of hydrogen
loonding in the back scattering spectrum of AgOH is mani-
zfested in the broadening of all Raman lines in comparison
‘to those observed in the spectrum of AgZO. The spectra of
'the two silver compounds were obtained mainly to examine
‘their interference in the surface Raman spectrum of the
amino acids. No interference was observed, however, and

it can be concluded that either the silver oxides are not
present or that the intensity of their surface signal is
too weak to be observed.

Returning now to the surface Raman spectra of the
amino acids shown in Figure 22 some interesting anomalies
are observed. Assignments of the observed surface Raman
spectra for L-leucine and L-glycine are made by comparison
to other vibrational spectroscopic studies of amino
acids.ll4-120 The suggested assignments are listed in
Table 9. The most important and distinct difference between
the surface spectra of L-leucine and that of L-glycine is
the appearance of a line at 671 cm.1 in the leucine spectrum
which is absent from the glycine. This mode has been
assigned to the ~COO- bend in leucine and is expected to
occur in the same region of the glycine surface spectrum.

The -COO_ bending vibration occurs at 690 cm"1 in the

infrared spectrum of glycine and at 694 in the Raman spectrum

 

166

of polycrystalline y-glycine.117 The most intense line in

both surface spectra appears at 1377 cm-1 in L-leucine and
1380 cm-1 in L-glycine and is probably the symmetric
stretching motion of the -C00- group. The line at

1424 cm-1 in the surface spectrum of L-leucine has no
counterpart in the spectrum of L-glycine and is accordingly
assigned to motion in the methyl substituents in leucine.
The COO- asymmetric stretch appears as a doublet at

1589 cm-1 and 1577 cm-1 in the surface spectrum of L-
leucine, but as a singlet at 1588 cm.1 in the spectrum of
L-glycine. The appearance of splitting in the asymmetric
stretching motion of the C00. group in the leucine spectrum
may be a consequence of bonding of one oxygen atom at the
surface while the other remains free. This explanation is
also invoked to account for the appearance of a C00- bend
in the leucine spectrum and not in glycine. In glycine

if both oxygen atoms are bound to the surface the bending
vibration would be significantly dampled and probably
shifted in frequency. The -NH2 deformations above 1600 cm-1
in the surface Raman spectrum of L-leucine appear more
intense relative to the same vibrations in the glycine
spectrum. This may also be a manifestation of the way in
which glycine is attached to the surface. If the glycine
molecule is attached through both carboxylate oxygens it
may be forced into a configuration which keeps the -NH2
group from coming in contact with the surface. With only

one oxygen attached in the leucine situation the molecules

167

maintain flexibility on the surface and allow the amine
group to contact the surface. The clear distinction
between the surface Raman spectra of L-leucine and L-
glycine on silver may be extrapolated to distinguish the
same amino acid residues in a protein adsorbed at an

electrode surface.

Future Surface Raman Studies.

 

The analysis of experimentally observed surface Raman
spectra presented in this thesis can be applied to nearly
any molecule. Molecules which have demonstrated ability
as ligands in metal complexes undoubtedly will exhibit
intense surface Raman scattering. Those molecules which
are of particular interest to the verification of the
surface dynamics of t-l,2-bis(4-pyridyl)ethy1ene on
silver are the isomers to the highly symmetrical 4,4'-
bipyridylethylene (i.e. the isomers with nitrogens in
positions other than the 4 position). The very low frequency
region (0-100 cm-l) of the surface Raman spectrum of t-BPE
should demonstrate significant spectroscopic changes
accompanying the molecular reorientation from the end-on
to the flat configuration. Gerack et a1.121 attributed
potential dependent dynamics of the low frequency surface
spectrum of pyridine to reorientation of the molecule to
the "standing up" position at less negative potentials.
Determination of surface coverage is particularly important

to the designation ofenlenhancement factor. The

168

monoprotonated t-BPE molecule generates a surface Raman
spectrum distinct from the unprotonated molecule and may
therefore be exploited in accurate determination of
surface coverage through titration. The titration tech-
nique has some obvious advantages over electrochemical
techniques which require sweeping of electrode potential.
Also the surface Raman spectrum of monoprotonated t-BPE
is observed even at very light coverages (250 nM bulk
concentrations). Electrochemical measurements are gener-
ally insensitive or irreproducible at such low coverages.

Excitation profiles in regions of the visible spectrum
other than the rhodamine 6G region are necessary to analyze
the energy distribution of eigenstates in t-BPE and other
adsorbates. Profiles should also be obtained at other
potentials, as shown to be important in the discussion of
potential-dependent surface Raman intensities. The use
of teflon as an internal standard should be exploited in
the preparation of the excitation profiles in that C104-
demonstrates its own potential dependent behavior at the
electrode surface.

The application of SERS to biological molecules
should be accelerated. The advantages of observation of
these molecules at the electrode surface include flour-
escence quenching by the metal surface, absence of
photodecomposition problems owing to the behavior of the

metal as a heat sink, and the ability to vary continuously

 

169

potential while observing the effect of electron transfer
on the surface Raman spectrum.

Further experimentation and theoretical investigation
on the influence of the multiple enhancement mechanisms,
proposed to account for the anomalous surface Raman
intensities, are necessary, but at this time applications
of the SERS technique are more generally worthwhile.

The most important conclusion which can be drawn from
this thesis is that any discussion of origination of
Raman intensities from surface adsorbates must include
mention of surface complex formation and the resonance
Raman intensity enhancement expected from the perturbation

of molecular electronic state distributions.

 

170

APPENDIX A

The form of the normal coordinates of some pyridine
vibrational modes, as determined by Long and Thomas,107
are depicted in this appendix. Discussion of the form of
the normal coordinates was central to the assignment of
lines in the normal Raman spectrum of t-l,2-bis(4-

pyridyl)ethylene (see Chapter 5).

171

 

List of References

172

10.
11.

12.

13.

14.

15.

16.

173

REFERENCES

T. Katsumoto; Bull. Chem. Soc. Japan 33, 1376 (1960).

. M. Fleischmann, P.J. Hendra, A.J. McQuillan; Chem.

Phys. Lett. 23(2), 163 (1974).

R.L. Paul, A.J. McQuillan, P.J. Hendra,14.Fleischmann;
J. Electroanal. Chem. 33, 245 (1975).

A.J. McQuillan, P.J. Hendra, M. Fleischmann; J.
Electroanal. Chem. 33, 933 (1975).

D.L. Jeanmaire, R.P. van Duyne; J. Electroanal. Chem.
33, l (1977).

D.L. Jeanmaire, R.P. van Duyne; J. Electroanal. Chem.
33, 235 (1975).

J.A. Creighton, M.G. Albrecht, R.E. Hester, J.A.D.
Matthew; Chem. Phys. Lett. 33, 55 (1978).

B. Pettinger, U. Wenning, D.M. Kolb; Ber. Bunsenges.
Phys. Chem. 32, 1326 (1978).

J.A. Creighton, C.G. Blatchford, M.G. Albrecht; J.
Chem. Soc. Faraday II 13, 790 (1979).

M. Moskovits; J. Chem. Phys. 33, 4159 (1978).
M. Moskovits; Solid State Comm. 32, 59 (1979).
J.A. Creighton, C.G. Blotchford, J.R. Campbell; Proc.

of the VIIth Int. Conf. on Raman Spectroscopy, 398-399
(1980) (W.G. Murphy, ed.).

 

J.I. Gersten; J. Chem. Phys. 22(10), 5779 (1980).

F. King, G.C. Schatz, and R.P. van Duyne; J. Chem.
Phys. 33, 4472 (1979).

G.C. Schatz, R.P. van Duyne; Surface Sci. 101, 425
(1980).

C.S. Allen, G.C. Schatz, R.P. van Duyne; Chem. Phys.
Lett. 13(2), 2101 (1980).

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.
28.
29.
30.

31.

32.

33.
34.

35.

36.

174

S. Efrima, H. Metiu; Chem. Phys. Lett. 99(1), 59
(1978).

S. Efrima, H. Metiu; J. Chem. Phys. 29(4), 1602
(1979).

S. Efrima, H. Metiu; J. Chem. Phys. 29(5), 2297
(1979).

S. Efrima, H. Metiu; J. Chem. Phys. 29(4), 1939
(1979).

S. Efrima, H. Metiu; Israel Journal of Chem. 29, 17
(1979).

G. Karzeniewski, T. Maniv, H. Metiu; Chem. Phys. Lett.
29(2), 301 (1980).

P.K. Avavind, H. Metiu; Chem. Phys. Lett. 29(2), 301
(1980).

M.R. Philpott, J. Chem. Phys. 92(5), 1812 (1975).

E.N. Economon, K.L. Ngai; Adv. Chem. Phys. 22, 265
(1974).

Y.Y. Teng, E.A. Stein; Phys. Rev. Lett. 29, 511
(1967).

D. Beaglehole; Phys. Rev. Lett. 22, 708 (1969).
I. Pockrand; J. Phys. 99, 2423 (1976).
. Raether, Nuovo Cimento 299, 817 (1977).
Raether, Phys. Thin Films 9, 145 (1977).

H
H
A. Girlando, M.R. Philpott, D. Heitmann, J.D. Swalm,
R. Santo, J. Chem. Phys. 22(9), 5187 (1980).

A

H

A

. Girlando, J.G. Gordon II, D. Heitmann, M.R. Philpott,
. Seki, J.D. Swolen; Surf. Sci. 101, 417 (1980).

. Otto; Z. Phyzik 241, 398 (1968).
E. Kretschmann; Z. Phyzik 241, 313 (1971).

R.M. Hexter, M.G. Albrecht; Spectrochim. Acta 35A,
233 (1979).

A. Brillante, I. Pockrand, M.R. Philpott, J.D. Swalen;
Chem. Phys. Lett. 92(3) 395 (1978).

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.
51.

52.

53.

54.

175
W. Carius, O. Schroter; Experimentelle Technik der
Physik 29(1), 35 (1980).

B. Pettinger, A. Tachjeddime, D.M. Kolb; Chem. Phys.
Lett. 99(3), 544 (1979).

W.P. Chen, J.M. Chen; Surf. Sci. 92, 601 (1980).

B. Pettinger, U. Wenning, H. Wetzel; Surf. Sci. 101,
490 (1980).

R. Donnhaus, R.E. Benner, R.K. Chang, I. Chabay; Surf.
Sci. 101, 367 (1980).

A. Otto; Surf. Sci. 101, 99 (1980).

I. Pockrand, A. Brillante, D. Mobins; Chem. Phys. Lett.
99(3), 499 (1980).

J.G. Gordon II, S. Ernest; Surf. Sci. 101, 499 (1980).

A. Regis, J. Corset; Chem. Phys. Lett. 29(2) 305
(1980).

A. Regis, J. Corset; Extended Abstracts, 81—1
Spring Meeting of the Electrochemical Society, 921
(1981).

 

M.W. Howard, R.P. Cooney, A.J. McQuillan; J. Raman
Spec. 9(4), 273 (1980).

R.P. Cooney, A.J. McQuillan, M.W. Howard; Proc. of
the VIIth Int. Conf. on Raman Spectroscopy, 400
(1980) (W.G. Murphy, ed.)T

 

 

J.C. Tsang, J.E. Demuth, P.N. Sanda, J.R. Kirtzey;
Chem. Phys. Lett. 29(1) 54 (1980).

G.W. Robinson; Chem. Phys. Lett. 29(2), 191 (1980).

E. Burstein, Y.J. Chen, C.Y. Chen, S. Lundquist,
E. Tosatti; Solid State Comm. 29, 567 (1979).

C.Y. Chen, E. Burstein, S. Lundquist; Solid State
Comm. 92, 63 (1979).

R.R. Smardzewski, R.J. Colton, J.S. Murday; Chem.
Phys. Lett. 99(1), 53 (1979).

S.R. Kelemen, A. Kaldor; Chem. Phys. Lett. 29(2),
205 (1980).

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.
68.

69.

70.

71.

72.

176

J.E. Demuth, K. Christmann, P.N. Sanda; Chem. Phys.
Lett. 29(2), 201 (1980).

R.E. Chance, A. Prock, R. Silbey; Adv. Chem. Phys.
92, 1 (1978).

D.A. Zwemer, C.V. Shank, J.E. Rowe; Chem. Phys. Lett.
29,(2), 201 (1980).

J.E. Rowe, C.V. Shank, D.A. Zwemer, C.A. Murray; Phys.
Rev. Lett. 99(26), 1770 (1980).

H. Seki, M.R. Philpott; J. Chem. Phys. 29(10), 5376
(1980).

C.Y. Chen, I. Davoli, G. Ritchie, E. Burstein; Surf.
Sci. 101, 363 (1980).

J.P. Heritage, J.G. Bergman, A. Pinczuk, J.M. Worlock;
Chem. Phys. Lett. 92(2,3), 229 (1979).

R. Dornhaus, M.B. Long, R.E. Benner, R.K. Chang,
Surf. Sci. 99, 240 (1980).

G.R. Erdheim, R.L. Birke, J.R. Lombardi; Chem. Phys.
Lett. 99(3), 495 (1980).

W. Krasser, H. Ervens, A. Fadini, A.J. Renouprez;
J. Raman Spec. 9(2) 80 (1980).

M. Moskovits, D.P. Dilello; J. Chem. Phys. 29(12),
6068 (1980).

T.M. Cotton, S.G. Schultz, R.P. van Duyne; J. Am.
Chem. Soc. 102, 7960 (1980).

M.B. Lippitsch; Chem. Phys. Lett. 29(2), 224 (1981).

E. Koglin, J.M. Sequaris, P. Valenta; J. Mol. Str.
99, 421 (1980).

M. Moskovits, D.P. DiLella; Chem. Phys. Lett. 29(3),
500 (1980).

M.R. Philpott, J.G. Gordon II, B. Pettinger, Extended
Abstracts, 81-1 Spring Meeting of the Electrochemical
Society, 896 71981).

 

T.H. Wood, M.V. Klein; J. Vac. Sci. & Tech. 29, 459
(1979).

M. Moskovits, D. DiLe11a, P. McBreen, R. Lipson,
A. Gohin; Proc. of the VIIth Int. Conf. on Raman
Spectroscopy, 394 (1980) (W.G. Murphy, ed.).

 

 

73.

74.

75.

76.

77.

78.

79.

80.
81.

82.

83.

84.

85.

86.

87.

88.

89.

177

H. Wetzel, B. Pettinger, U. Wenning; Chem. Phys.
Lett. 29(1), 173 (1980).

E. Burstein, C.Y. Chen, S. Lundquist; Proc. of the
Joint US-USSR Symposium on Theory of Light Scattering
in Condensed Matter, (Plenum Press, New YOrk, 1979).

 

 

R.P. van Duyne; Chemical and Biochemical Applications
of Lasers, Vol. 4, (C.B. Moore, ed), (Academic Press,
New York, 1978).

 

T.E. Furtak, Solid State Comm. 29, 903 (1978).

T.E. Furtak, G. Trott, B.H. Loo; Surf. Sci. 101,
374 (1980).

W.A. Goddard III, T.C. McGill; J. Vac. Sci. Technol.
29(5) 1308 (1979).

A. Otto, J. Timper, J. Billman, G. Kovacs, I. Pockrand;
Surf. Sci. 92, L55-L57 (1980).

A. Otto; Surf. Sci. 92, 145 (1980).
J.I. Gersten; J. Chem.Phys. 22(10), 5779 (1980).

H. Wetzel, H. Gerischer, B. Pettinger; Chem. Phys.
Lett. 29(2), 392 (1981).

J.R. Lombardi, E.A. Shields Knight, R.L. Birke; Chem.
Phys. Lett. 29(2), 214 (1981).

J. Billmann, A. Otto, I. Pockrand, J. Timper; Proc. of
the VIIth Int. Conf. on Raman Spectroscopy, 424
(1980) (W.G. Murphy, ed.).

 

B. Pettinger, H. Wetzel; Chem. Phys. Lett. 29(2),
398 (1981).

J.R. Lombardi, R.L. Birke; Extended Abstracts 81-1
Spring Meeting of the Electrochemical Society, 894
(1981).

 

D.L. Rosseau, J.M. Friedmann, P.F. Williams, in:
Topics in Current Physics, Vol. 2 Raman Spectroscopy
of Gases and Liquids, (A. Weber, ed.), chapter 6.

 

 

D.K. Lavalee, E.B. Fleischer; J. Am. Chem. Soc. 99(8),
2583 (1972).

H.B. Gray, C.J. Ballhausen; J. Am. Chem. Soc. 99,
260 (1963).

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.
102.

102a.

103.

104.

105.

178

H.A. Pearce and N. Sheppard; Surf. Sci. 99, 205
(1976).

R.M. Hexter, M.G. Albrecht, Spectrochim. Acta 35A,
233 (1979).

A.C. Albrecht, J. Chem. Phys. 99(5), 1476 (1961).
J. Tang, A.C. Albrecht, in Raman Spectroscopy, Theopy

and Practice, Vol. 2, pp. 33-68 (1970) (H.A. Szymanski,
ed.).

 

R.J.H. Clark, B. Stewart; Structure and Bonding 99,
1 (1979).

S.L. Ross; Introduction to Ordinary Differential
Equations (Xerox College Publishing, MA) (1966)
Chapter 5, pp. 152-160.

 

 

B. Pettinger, U. Wenning, H. Wetzel; Chem. Phys.
Lett. 92(1), 192 (1979).

U. Wenning, B. Pettinger, H. Wetzel; Chem. Phys.
Lett. 29(1), 49 (1980).

D.P. Shoemaker, C.W. Garland, J.I. Steinfeld; Experi-
ments in Physical Chemistry, 3rd ed. (1974) pp. 455-
457.

 

Chemical Rubber Co; Handbook of Chemistry and Physics,
52nd ed. p. B232.

H.H. Perkampus, E. Baumgarten; Spectrochim. Acta 29,
359 (1964).

E.B. Wilson; Phys. Rev. 99, 706 (1934).
E. Spinner; J. Chem. Soc., 3870 (1963).

G. Valette, A. Hamelin; J. Electroanal. Chem. 99,
301 (1973).

G. Herzberg; Molecular Spectra and Molecular Structure II
Infrared and Raman Spectra of Polyatomic Molecules,
(Van Nostrand Reinhold Co.) (1945), 247-249.

F.R. Dallish, W.G. Fateley, P.F. Bentley; Characteris-
tic Raman Frequencies of Organic Compounds (Wiley-
Interscience) (1974) pp. 263-272.

 

Z. Meic, H. Gusten; Spectrochim. Acta 34A, 101 (1978).

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

179

A. Regis, J. Corset; Extended Abstracts 81-1, Spring
Meeting of the Electrochemical Society, 921 (1981).

 

D.A. Long, E.L. Thomas; Trans. Faraday Soc. 99, 783
(1963).

E. Spinner; Aust. J. Chem. 29, 1805 (1967).

R.H. Dyck, D.S. McClure; J. Chem. Phys. 99(9), 2326
(1962).

J. Volke, J. Holubek; Coll. Czech. Chem. Commun. 22,
1777 (1962).

W.L. Jorgensen, L. Salem; The Organic Chemists Book
of Orbitals (Academic Press, NY) (1973).

 

 

E.F. Gross, F.I. Kreingold; Optics and Spectroscopy
29, 211 (1961).

V.M. Kirillova, E.N. Lotkova; Optics and Spectroscopy
29, 181 (1966).

M. Tsuboi, T. Onishi, I. Nakagawa, T. Shimanouchi,
S. Migushima; Spectrochim. Acta 22, 253 (1958).

R.K. Khanna, P.J. Miller; Spectrochim. Acta 26A,
1667 (1970).

R.K. Khanna, M. Horak, E.R. Lippincott; Spectrochim.
Acta 22, 1759 (1966).

M.A. Salimov, V.A. Pchelin, A.V. Kerimbekov; Russ. J.
Phys. Chem. 92(10), 1231 (1963).

M. Tsuboi, T. Takenishi, A. Nakamura; Spectrochim.
Acta 29, 271 (1963).

K. Machida, A. Kagayama, Y. Saito, Y. Kuroda, Y. Uno;
Spectrochim. Acta 33A, 569 (1977).

R.S. Krishnan, V.N. Sankaranarayanan, K. Krishnan;
J. Ind. Inst. Sci. 99(2), 66 (1973).

A.Z. Genack, D.A. Wertz, T.J. Gramila; Surf. Sci.
101, 381 (1980).

   

MICHIGAN STATE UNIVERSITY LIBRARIES
IIIllUllllHI("WNWIIHIIIIUIHIIHIIIHIUIIPIIIHIM
3 1293 03145 5805