9
5“,- I 1'5.
322;»
-_’ '3"?
:-2 '. - 54)"
1W ‘1?“ 2:5,":
' s. f." "‘
‘ I"
{m
f"
I
-' I
O
53:! }
m;
12.11. 1;
‘ :, cv1s'.“ft.‘.:)
“VI. 'K'asunl‘t’ '
(1‘
‘3’ V4.13" - "35’
syw'lï¬'s
32 2‘12ng ‘
‘l
I "-.. .; 5:â€: $7.7
I P .
' “1.11?
111,91; "3%., 3
"1’9"
’3 ‘ . I
, ’
’ ( â€Pu“ 1..."..:‘
y". . 1J.f.""’--f-~"F:ffagul) M.
gr...“ 7." ,;1/-M~; .551" .332:
‘4').
1.
flix'f';
: "‘
‘/"O’pl.
.’ :"..- (3,335
:1‘1"I’,:tl{ ‘d- v;
"a
'1' 2.5:»:
.1???
~ ' 4
‘14: .2. «3‘
‘3' 'r 11:!
I
11â€,“? HAWK." .-
fw ’- :1“ 31.513 ‘f- ‘3"
. 1.3.“:
:7 4 "a;
3.51.} n...
A)? a}.
I1.
79115.26'
""!u
ï¬w— ' I
:a:‘: - 1-. :2?
‘ 1119.5? “k ‘5'): f' *1 A
,1 39553;“ '
â€-..; a?
-.. 7f
I '.
'1 .
'1 9}:-
V .
' I 1
- t 1 A
f '1/2 .3232; 'fflt'v_- ’ V .
f‘ {1.113. I AI};- 1. - a. l. '1 hi; It!
431,1)1 «71:13); (")7'1: â€'1’?!" i n- 7â€; 'f'“
‘ I’. .A ‘
“l
$33317
{Apié
‘f
9‘ "lé I)
4913?; ‘0
â€r; â€v ‘
Jig/v;
[VJ-â€(.11-
[,1]: 13;./
7"1“.:'rd;’lff~$&£x\
9:211.-
gu'
“fly-M
.v::fulh ,
a "“+ 3‘ 33‘: â€1:31,."
â€â€™ :32; '1 .
I If???" 1;
.. ,‘
3;!" ' '.".
r, .
I}. . A 1
mi 1; ..
“"II‘
.â€.:1’
,
â€53'"?
3-,». .,
â€I, 1 .'
1.45
â€ruck-"3!
J- -.1 1 ~ _
flat“; ’r _
â€gem: '
1317'ol‘1‘6
MICHIGAN STATE UNIVER§ITY LIBHAR
Ill Hill} Mliffifllillililllilll,llIll
3 1293 00575 3920
ES
H
M
This is to certify that the
dissertation entitled
CHEMILUMINESCENCE: A MECHANISTIC PROBE
OF ELECTRON-TRANSFER REACTIONS
presented by
Robert Dorsey Mussell
has been accepted towards fulï¬llment
of the requirements for
Ph.D. degree in Chemistry
371ml 6 W
Major professor
Date September 15, 1988
MS U is an Afï¬rmative Action/Equal Opportunity Institution 0.12771
LIBRARY
Mkhigqn State
University
PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
DATE DUE DATE DUE DATE DUE
MSU Is An Afï¬rmative Action/Equal Opportunity Institution
CHEMILUMINBBCBNCB: A MECHANIBTIC PROBE OP
ELECTRON-TRANSFER REACTIONS
BY
Robert Dorsey Mussell
A DISSERTATION
Submitted to
Michigan state University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry
1988
ABSTRACT
CHRHILUKINBSCBNCR: A MECHANISTIC PROBE OF
ELECTRON-TRANSFER REACTIONS
by
Robert Dorsey Mussell
The chemiluminescent reactivity of M6X8Y62 (M = Mo, W;
X, Y = Cl, Br, I) clusters in nonaqueous solution has been
used to investigate the mechanism of electron transfer
reactions. The partitioning of the electrochemical
excitation energy upon annihilation of electrogenerated
M06C1143' with a series of W6X8Y6' ions has been determined
from overall electrogenerated chemiluminescence (ecl)
quantum. yields and chemiluminescence spectra. The
electrochemical excitation energy is partitioned to produce
M06C1142'* and W6X8Y62'* with essentially equal probability.
Analysis of the equal distribution with current electron-
transfer theories suggests that the electronic coupling and
reorganizational energy for the conversion of M5X8Y6‘--€>
2-*
Mssts
and M6X8Y63'——-§ M6X8Y62'* by simple electron
exchange ‘are equal» The free-energy dependence. of the
M6X8Y62’ ecl in acetonitrile and dichloromethane was
Robert Dorsey Hussell
investigated with four series of structurally and
electronically related electroactive organic compounds. The
yields for the formation of electronically excited M06C1142'
ion produced by the electron-transfer reaction of M06C1143â€
with electroactive organic acceptors and the reaction of
M06C114' with electroactive organic donors have been
measured over a wide potential range by simply varying the
reduction potential of the electroactive organic reagents.
-*
The dependence of the formation yield of M06C1142 , ¢est
on
the driving force of the annihilation reaction is similar
for the four series in both solvents. is immeasurable
¢es
(<1o'5) for reactions ‘with free energies positive of a
threshold value. Over a narrow free energy range just
negative of threshold, ¢es rapidly increases. And with
increasing exergonicity of the electron-transfer reaction,
fes asymptotically approaches a limiting value less than
unity. Analysis of these excited-state production yields
using Marcus theory reveals that unit efficiencies for
excited-state production are circumvented by long-distance
electron transfer. The distance this electron transfer
occurs can be mediated by solvent and solute interactions,
and calculations establish that the electron-transfer
distance is equal to the radii of the reactants plus the
diameter of two solvent molecules. Ecl efficiencies of the
hexanuclear cluster ions are not only perturbed by
intermolecular factors but also are dramatically
Robert Dorsey nussell
effected by ligand coordination sphere. Additionally, the
effects of temperature and potential step sequence on the
ecl efficiencies of the hexanuclear cluster ions have also
been investigated.
ACKNOWLEDGEMENTS
I thank all the Nocera group members, especially I-Jy
Chang, for making each day in the lab a unique and sometimes
scientifically stimulating experience. It especially thank
Dan Nocera for his friendship and unbridled (non-stop)
direction during my research career here at MSU. ’Dan’s
ability to always see the "big picture" definitely made my
graduate journey a much smoother trip. Mark Newsham, Dan
Kassel, Joe Skowyra and Randy "a fifth isn’t excessive" King
get a special thanks for their friendship and their efforts
to make sure that all chemistry students are not viewed upon
as nerds. I am indebted to Dr. Tom Pinnavaia for serving as
my second reader and Dr. Bob Cukier for his many helpful
discussions and also for serving on my committee. A big
thanks goes to Sharon Corner for being a good friend as well
as an excellent typist.
I would like to acknowledge the support of fellowships
from Dow Chemical Company and the College of Natural Science
during the 1987-1988 academic year.
Outside of chemistry I would like to express my thanks
to all the past and present team members of the "Froggers",
*which will go down as one of the great intramural teams in
MSU history, for the opportunity to participate in
basketball and softball with so many great individuals.
Also thanks go to all my friends, Tom Brege, Tommy Oosdyke,
V
John Childers, Richard Hahn, Tom Murphy and Bob Lane, who
have stood by me and helped me keep a prospective on the
true meaning of life. .
I want to especially thank my parents and family for
their support, understanding and love during all my school
years. Without them none of this would have been possible.
And finally, my deepest gratitude to my wife Joy. Her
patience, support, and love during the past five years have
made this experience much more enjoyable.
vi
TABLE OF CONTENTS
Page
LIST 0FTABLES..0...OOOOOOOOOOOOOOOOOOOO...OOOOOOOOOOOO ix
LIST OF FIGURES......... ..... .......................... xiii
I. Introduction........... ...... ..... ..... ........... 1
II. Experimental........................... ..... ...... 26
A. Synthetic Methods............... ..... . ....... 26
1. Preparation of Hexanuclear
Molybdenum Clusters........ ............. 26
2. Preparation of Hexanuclear
Tungsten Clusters. ....... .. ............. 29
3. Organic Donors and Acceptors ............ 32
4. Supporting Electrolytes. ....... . ........ 33
5. Solvents...... ............... ... ....... . 34
B. Experimental Methods.................. ....... 35
1. Characterization of Molybdenum and
Tungsten Clusters.. ..... ....... ......... 35
2. Electrochemical Measurements. ........... 36
3. Quenching Measurements............ ...... 36
4. Electrogenerated Chemiluminescence...... 37
1. Quantum Yields ..................... 37
ii. Spectra............ ..... ..... ..... . 42
III. Electrochemical Excitation Energy Partitioning
in Mixed Cluster Electron Transfer Reaction. ...... 45
A. BaCkgroundO0.00.00.00.00.0.00.00.00.00 ..... O. 45
B. Results and Discussion................. ...... 49
vii
IV. The Effects of Driving Force and Long-Distance
Electron Transfer on Chemiluminescence
Efficiencies......................................
A. BaCkgroundOO...OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
B. Results....... ..... ........ ....... .. ........ .
.C. Discussion ......... ............ ..... . ........
V. Environmental Effects on M6X8Y62'
Chemiluminescence Efficiencies....................
A. Background...................................
B. Solvent Effects................ ............ ..
C. Supporting Electrolyte Effects... ........... .
D. Ligand Coordination Sphere Effects. ...... ....
E. Temperature Effects..........................
F. Potential Step Program Effects...............
VI. Final Remarks.... ........ . ............... . ....... .
VII. REFERENCES......... ...............................
viii
Page
95
95
98
113
140
140
143
155
161
170
176
182
185
LIST OF TABLES
Excited State Energies and Electrochemical
Properties of Ru-Polypyridyl Complexes in
AcetonitrileO0.000000000000000000......OOOOOOOOOO.
Emission Maxima and Electrochemical
Properties for M06 Clusters in CHZCIZ o o o o e o o o e e e o 0
Emission Maxima and Electrochemical
Properties for W6 Clusters in CHZClz..............
Energy Transfer Quenching of M6X8Y62'
Clusters in CH2C120000000 oooooo so ooooooo ooeeo eeeee
Photophysical Properties and Overall Ecl
Quantum Yields of M6X8Y62' Clusters Used in
Mixed Cluster Ecl Reactions.................. .....
Driving Forces, Energy Partitioning Ratios
and M06C1142' Excited State Yields in Mixed
Cluster Ecl Reactions.............................
ix
Page
47
50
51
60
63
7O
10
11
12
Page
Rate Constants for Quenching of M6X8Y62'
Clusters by Nitroaromatics and Substituted
Benzoquinones in CH3CN at 23°C....OOOOOOOOOOOOOOOO 87
Rate Constants for Quenching of MGXBYGZ'
Clusters by Aromatic Amines in CH3CN at 23°C...... 88
Electron-Transfer Parameters Used in
Calculating Ai Values for Quenching Reactions ..... 93
Reduction Potentials, Quenching Rate Constants,
and Ecl Quantum Yield Data for Aromatic Amines
Used in Bel Studies in CH3CN....... .......... ..... 99
Reduction Potentials, Quenching Rate Constants,
and Ecl Quantum Yield Data for Aromatic Amines
Used in Bel Studies in CH2C12 ..................... 100
Reduction Potentials, Quenching Rate Constants,
and Ecl Quantum Yield Data for Nitroaromatics
and Aromatic Quinones Used in Bel Studies
in CHBCNeeeoe ssssss eeoeeoeeeeeeeeeooeoeoeeeoeeeoee 101
13
14
15
16
17
18
19
Page
Reduction Potentials, Quenching Rate Constants,
and Ecl Quantum Yield Data for Nitroaromatics
and Aromatic Quinones Used in Ecl Studies
in CHZCIZQOOQecoo.esoococooooeeoeeeeoeeooeoeeooooo 102
Reduction Potentials, Quenching Rate Constants,
and Ecl Quantum Yield Data for Pyridinium
Salts Used in Ecl Studies in CH3CN................ 103
Reduction Potentials, Quenching Rate Constants,
and Ecl Quantum Yield Data for Pyridinium
Salts Used in Ecl Studies in CHZClZ.......... ..... 104
Reduction Potentials and Ecl Quantum Yield
Data for Bipyridinium Salts Used in Ecl
Studies in CH3CN..... ..... ........ ...... . ......... 105
Electrochemical Properties of M06C1142' in
Various Nonaqueous Solvents.......... ............. 145
Physical Properties of Solvents Used in Ecl
Studies........ .......... . ..... ... ...... . ......... 146
Excited State Production Efficiencies for Ecl
Reactions in Several Nonaqueous Solvents.......... 14?
xi
20
21
22
23
24
Page
Electron Transfer Parameters and Reaction
Distances for the Ecl Reaction of M06C114'/
4'Cyano-N‘methYIPYI'idiniumo e e e e e o e e e e e e e o o e e e e e o e e 153
Solvent Diameters and Ar Values for
MOGC].14-/4-CyanO'N’methY1pyridinium. e e o e e o o o o e o e o e 154
Supporting Electrolyte Studies for the
M°6C1142- EC]. ReaCtiOn.. eeeee oeoeeeee ooooo eee eeeee 158
Dependence of Ecl Efficiencies on Supporting
EleCtrOIYte cation.........0.0000000000000000..... 160
Excited State Production Efficiencies for
M06C18ClnX6_n in CHZCIZOQQeoeeeooeeeeeoe eeeee so... 162
xii
LIST OF FIGURES
Page
Potential energy curve for an electron
transfer reaction accompanied by a net
chemical change. AG' is the driving force
for the electron transfer, AG* is the
activation barrier, and the splitting
at the intersection is equal to 2HAB'°°°°°°°°°°" 3
Potential energy curves for electron transfer
as a function of increasing driving force
(a) AGo > -A; (b) AG° = -A; (c) AG° < -A....... 9
Truncation of very fast electron transfer
rates by diffusion. The diffusion limiting
rate is represented by the dotted horizontal
line. Thus, the rates are leveled until AG°
becomes very large................................ 14
xiii
Potential energy curve description of
chemiluminescence. AG°es and AG' are
gs
reaction free energies for electron transfer
to produce excited- and ground-state products,
respectively, and AG* is the activation energy
for the excited-state reaction. k and k
es gs
are the electron transfer rates for production
of excited- and ground-state products,
respectively .....................................
Structure of M6X8Y62' ions: 0 = Mo(II), W(II):
X'Y=Cl' Br, 10............OOOOOOOOOOOOOOO. .....
Cyclic voltammograms of (0.1 NBu4PF6 at 23°C)
(a) M06C112 (3 mM in CH3CN); (b) M06c112122‘
(3 mM in CH2C12); (c) M06C112(SCN)22-
(3 mM in CHZClz).................. ...............
Electrogenerated chemiluminescence spectrum of
3— - - _
NBu4PF6 at 23°C) eeeee eeeeooeeeeoeooooooooeeeeoeeeo
xiv
Page
19
23
54
65
10
11
12
Page
Steady-state emission and ecl spectrum in
CH3CN for (a) w6I142', ————; M06C1142' °°°°';
M06C1142'/W6Il42' ecl, ----:
(b) Mosc1142‘/w61142',._____— ecl; fit from the
sum of a ratio of M06C1142’ and “61142- emission
spectra, °""’. Peak maxima are normalized
to an arbritraryvalueO0.0000000000000000000000000 68
Plot of log ¢es of M06Cll42' in the mixed
cluster ecl reaction (0) and in the reaction
of M06c1143‘/A+ (0) vs. AGes in cnzc12 at
23°C (p:001MNBU4PF6)eoeoeeeoooeeoeo eeeeeeeeeee 75
Molecular orbital diagram for M6x8Y62' ions ....... 79
Molecular orbital description for electron
- 3- .
transfer between W6X8Y6 and M06C114 Wlth
the excited state being produced from
(a) the W6X8Y6' ion and (b) the M06C1143' ion ..... 81
Depiction of the eg and 32g metal based
cluster orbitals ................. ..... ............ 84
XV
13
14
15
Page
Plot of kBTlnkq vs. AG° in CH3CN at 23°C
(numbering as in Tables 7 and 8) for
(a) M06C1142'* quenched by organic acceptors;
(b) M06C1142'* quenched by organic donors......... 91
Plot of log ¢es vs. Aces for the electron-
transfer annihilation reactions of the
M06C1143'/A+ (0), M06C114'/P (A), and
M06C114'/NA' (a) systems in acetonitrile.
The numbering scheme is defined in Tables 10,
12 and 14. The standard free energy change
for the excited-state reaction pathways was
evaluated as described in the text.. .............. 110
Plot of log ¢es vs. AGes for the electron-
transfer annihilation reaction of the
Mo6c1143'/A+ (a), M06c114'/p (A), and
M06C114'/NA' (0) systems in dichloromethane.
The numbering scheme is defined in Tables 11,
13, and 15. The standard free energy change
for the excited-state reaction pathway was
evaluated as described in the text.. .............. 112
xvi
16
17
18
Page
Molecular orbital description for competitive
electron transfer to give either ground- or
excited-state M06Cll42' by the reaction of
(a) M06C1143' with oxidized aromatic amines
(A+) and (b) M06Cl14 with reduced nitro-
aromatics (NA') or pyridinium ions (P).
Production of electronically-excited
acceptors and donors is an energetically
unfavorable proces ......... . ...................... 115
Distance dependence of the differential
bimolecular rate constant for the excited—
state (es) and ground state (gs) electron-
transfer channels for the reaction between
M06Cll4' and one—electron reduced 4-cyano-N-
methylpyridinium (CMP), calculated by solving
eqs 3, 9-14 between r and r + ar using
a = 1.2 r1 and HAB" = 200 cal .................... 124
Distance dependence of the differential
bimolecular rate constant for the excited-
state (es) and ground-state (gs) electron
transfer channels for the reaction of M06Cll4'
with: (a) a hypothetical one-electron reduced
pyridinium species with AG = -2.05 V,
gs
Aces. = -0.05 eV; (b) a hypothetical
xvii
19
20
21
22
Page
one-electron reduced pyridinium species with
AGgs° = -2.15 V; (3668' = -o.15 V;
4-cyano-N-benzylpyridinium; (d) 4-cyano-N-
methylpyridinium; (e) 4-carboethoxy-N-
benzylpyridinium; and (f) 4-carboethoxy-N-
methylpyridinium. The standard free energy
driving forces for (c)-(f) are given in
Table 14 .......................................... 124
Plot of log ¢es vs. (1/Dop- l/Ds) for: (a)
M06C114'/M06C1143 ; and (b) M06c114'/P
in various nonaqueous solvents............ ........ 150
Plot of log fes vs. ionic strength, p, for
- 3—
the electron transfer of M06C114 /M06Cll4
in CH2C12 (a) and CH3CN (I) at 23°C ............. .. 157
Plot of log ¢es vs. no. of bromides substituted
in the axial position for the M06C18ClnBr6_n2'
ecl reaction in CHZClZ at 23°C (p = 0.10 M
NBU4PF6)eeeeooeeoe eeeeeeeeeee oeeeoeooo eeeeeeeeeeee 164
Cyclic voltammogram (CHZClz solution at 23°C,
0.1 M NBu4PF6) for MOGClBBrGZ' (3 mM) --——;
MoGClaBrsz' (3 mM) and NBu4Br ( 1 mM) ———— ..... ... 167
xviii
23
24
25
26
27
Page
Mechanism for Br' interference of
M06C18ClnBr6_n'1/M06C18ClnBr6_n3' annihilation
reactionOOOOO0.00.00.00.00.........OOOOOOOOOOOOOOO 169
Decrease in M06C183r62' luminescence during
bulk electrolysis (
)3 Increase in
luminescence after adding Br- (----).............. 172
Plot of log ¢es vs temperature for the
. . . - 3- .
annihilation of M06Cll4 /M6Cll4 in
dichloromethane ( ); M06Cll4’/
4-carboethoxy-N-methylpyridinium in
dichloromethane (°"'); M06C114'/
Mo6Cll43' (----) in acetone at p =
0010 NBu4PF6 eeeeeeeeeeeeeeeeeee ooeoeoeeeeeeeeeeoee 174
Plot of log fes vs pulse frequency in CH3CN
for the Mo6c114‘/M06c1143 (a) and M06C114-/
4-amido-N-methylpyridinium (a) at 23“C
(â€=001MNBu4PF6)eee ooooooooo oooeoooeeo ooooooooo 178
Plot of the log ¢es vs no. of pulses in
CH3CN for Mo6c114’/M06c1143' (e) and
M06Cl14'/4-amido-N-methyl pyridinium (0)
at 23°C (p:001NBu4PF6)eoeeeeeeeeeoeeoeeeoo eeeee 180
xix
CHAPTER I
I. INTRODUCTION
Electron transfer reactions play a fundamental role in
chemical and biological processes. Many important chemical
reactions involve oxidation-reduction processes especially
including those in inorganic chemistry in which transition
metal complexes are versatile redox reagents. Small
molecule activationl, photocatalysis2'3, and homogeneous4
and heterogeneous catalysis5 are fundamental inorganic
processes involving the transfer of an electron or electrons
to or from a metal reaction center. In biological systems,
oxidation-reduction transformations at inorganic reaction
centers control several essential biological processes.5'7
Some of these include electron transfer between the heme
centers of cytochromes and reduction of 02 by cytochrome
oxidase in oxidative phosphorylation, the reduction of
dinitrogen at a molybdenum center of nitrogenase, and four-
electron oxidation of water to oxygen at the manganese
center of the oxygen-evolving complex in photosystem II.
Owing to the importance of oxidation-reduction reactions,
experimental and theoretical elucidation of the factors that
govern the rates of electron-transfer events, has been a
central theme of mechanistic chemistry during the past two
decades.8"11
Electron-transfer reactions can be described in
classical terms by activated complex theory. As first
proposed by Marcus,12 electron transfer can be represented
by potential energy curves for reactants and products such
as those depicted in Figure 1. In this diagram, the
1
Figure 1. Potential energy curve for an electron
transfer reaction accompanied by a net chemical change.
AG’ is the driving force for the electron transfer, AG*
is the activation barrier, and the splitting at the
intersection is equal to ZHAB.
ENERGY
REACTION COORDINATE
Figure 1
electron transfer reaction, which is defined in many-
coordinate space, (N-l, where N defines the positions of all
molecules, their orientations and their vibrational
coordinates) is simplified by choosing a one dimensional
generalized reaction coordinate involving a vibration which
is important to the reaction along a reaction coordinate, x,
which represents the positions of the reactant and product
molecules and their solvent coordination spheres. The
ordinate defines the relative potential energy of the
system. The intersection of the reactant and product
potential energy curves occurs at an intermediate
configuration, called the activated complex, where the
nuclei are in a position halfway between the reactants and
products. The barrier height from reactants to the
activated complex defines the activation energy, AG*, and
the rate of electron transfer can be described by the
ket = Z exp [AG*/kBT] (1)
classical expression shown in eq 1, where Z is the
collisional frequency of the uncharged reactions in
solution. Marcus has defined the contributions to AG* by eq
2 where AG* is the driving force for the reaction, A is the
AG* = wr + (AG° + A)2 / 4A (2)
reorganizational energy which contains inner-sphere, ii, and
outer-sphere, 1°, contributions and wr is the work required
to bring the two reactants together. Marcus has shown that
in a dielectric continuum, the outer-sphere reorganizational
energy is given by,
Ac = Aez (1/2a1 + 1/2a2 - 1/r) (1/Dop - 1/Ds) (3)
where a1 and a2 are the radii of the two reactants, r is the
distance between centers of the two reactants in the
activated complex (usually assumed equal to a:l + a2) , and
Dop and DS are the optical and static dielectric constants,
respectively. The inner-sphere reorganizational parameter,
which depends on differences in equilibrium bond lengths and
angles between reactants and products, is defined by eq 4.13
In this equation fi = 2f2f3/(f2 + f3) is a reduced force
Xi 1/2 E’fi (d2°-d3°)iz (4)
constant for the ith inner-sphere vibration and (d2°-d3°)i
is the corresponding difference in equilibrium bond
distances of the reactants and products. The summation is
over all the intramolecular vibrations. The work term, wr,
is approximated by a Debye-Huckel formalism where z
2
_ 2122 e 5
wr- ()
Dsr(l + flDH r 'V'Ti)
SOOOnNeZ 1/2
51')“ = (6)
1000 DskBT
and 22 are the usual charges of the two reactants and p is
the ionic strength of the solution. It is noteworthy that
this classical approach assumes that reaction to products
occurs from the activated complex with unity. For this case
the electron-transfer reaction is said to be adiabatic.
More generally, electron transfer can be mediated
significantly by the electronic coupling between the
reactant and product surfaces and nuclear tunneling through
the barrier. To this end, a more accurate expression of the
electron ‘transfer rate is given_ by eq; 7 'where nuclear
tunneling and nonadiabatic effects are accounted for by Pn
ket = ZwEPn exp [AG*/kBT] (7)
and 53, respectively.â€-17
Because nuclear tunneling will
increase the reaction rate, with respect to the activated
electron transfer, Pn takes on values 2 1. At room
temperature tunneling does not typically contribute to the
overall rate (1‘n 9‘- 1), but becomes significant when (1)
either the barrier height is large, (ii) the reaction is
very exergonic, or (iii) the temperature of the reaction is
low. These quantum mechanical aspects of nuclear tunneling
have been treated by several authors in recent years.18"20
More important to chemical and biological electron transfer
under typical reaction conditions is the mediation of the
overall rate by the electronic coupling strength.15'21'24
Quantitatively, the probability that the electron transfer
will occur in the activated complex is given by eq 8 where
ea in eq 7 is related to nE’ by 53 = 1-exp(-nE’) and RAB. is
the electronic coupling matrix element between reactant and
product surfaces,17 calculated usually by the Landau-Zener
‘3' = —— (3)
kBT AkBT
treatment of avoided crossings. This is represented in
Figure l by the splitting at the intersection which is equal
to 2 HABO' For adiabatic reactions, HAB° is large and nE =
13 reactions with EB < 1 are said to be nonadiabatic. The
semi-classical electron-transfer expression, eq 7 reduces to
the classical formalism when nE and Pn are unity.
The energy dependence of the rate in both classical and
semi-classical treatments of electron transfer follows
directly from eq 2. Ignoring work terms for the moment, for
a weakly exergonic reaction (AG‘ > - A) the rate will
increase with increasing negative free energy, maximize when
the activation barrier is zero (A6“ = - A) and then decrease
for driving forces more exergonic than - A (i.e. AG° < - A)
(Figure 2). This latter region, called the inverted region,
is illustrated in Figure 2c, where the product curve, at
Figure 2. Potential energy curves for electron
transfer as a function of increasing driving force: (a)
AG' > -A; (b) AG° = -A; (c) AG° < -A.
m 939“.
AIVOQQ , Klflownw KIAOmuAV
E a: 2:
10
exergonicities greater than activationless transfer, climbs
up the back side of the reactant curve thereby introducing
again a positive activation energy.
Significant strides in the understanding of the
contributions of electronic, nuclear, and driving force
effects on the rate of electron transfer have been made in
recent years with the preparation of systems in which
electron donor and acceptor sites are molecularly linked
over fixed distances. One such approach to the design of
electron donor-acceptor systems is based on covalently
binding a transition metal complex (e.g. -Ru(NH3)52+) to
polypeptide residues of proteins sudh as cyctochrome c and
myoglobin.8'25a
In these semi-synthetic metalloproteins the
electron transfer rates between transition metal complex and
the heme center of the protein have been measured. A
modification of this approach has been to substitute Zn for
Fe in the heme center of hemoglobin, cytochromes, and
myoglobin.25'27
The Zn modified protein is structurally
similar to the native protein and hence can be complexed
with its biologically relevant electron-transfer counterpart
(i.e. cytc-chytc peroxidase). In these systems the
electron transfer is activated by absorption of a photon by
the long-lived Zn porphyrin. The photochemically activated
Zn site acts as an acceptor or donor with the heme center of
the complexed protein. In many instances the return
electron-transfer rate can also be measured. Results from
both. of these. approaches have led to» quantification of
11
biological inner-sphere reorganizational energies as well as
the effect of distance on the electronic coupling between
biological reaction centers.8'25'27 Alternatively a less
biological approach has relied on molecularly linking
organic acceptor and donor sites via rigid spacers.28'33
These systems have provided a direct comparison to
biological electron transfer and, for the case of aromatic
molecules bridged by steroid spacers, provided the first
verification of the inverted region.
Not surprisingly, the initial studies of electron
transfer, beginning with Rehm. and Weller’s studies on
fluorescence quenching of aromatic molecules,34 did not rely
on the design of synthetically complicated intramolecular
samples, but focussed on simple electron transfer reactions
between freely diffusing reactants. Since that landmark
study of Rehm and Weller, numerous experimental studies of
bimolecular systems have provided ample data for electron
transfer reorganizational energies, self-exchange rate
constants, and free energy dependencies in which the rate
increases and levels with increasing free energy (i.e., the
normal electron transfer region).35'40 However, unlike the
fixed distance electron donor-acceptor systems, observation
of a decrease in rate at high exergonicities (i.e., the
inverted region) has proven experimentally more elusive.
The inability to detect the inverted region prompted the
utilization of empirical free energy relationships, first
proposed by Rehm and Wellerfmt42 and modified by
12
others/‘3'44 to fit the observed data. In a more
quantitative approach, several theoretical studies have
involved quantum mechanical treatments to rationalize the
absence of the inverted region.17'19'45'49 Adthough these
studies attenuate the magnitude of the inverted effect
predicted by classical theories, a decrease in rate at high
exergonicity is maintained. The shortcomings of classical,
semi-classical and quantum mechanical rate expressions in
the highly exergonic region have been attributed to several
factors: (1) Truncation of the predicted rate curve by the
diffusion-controlled limiting rate as shown in Figure 3 will
obscure the inverted effect. Inverted behavior, which will
only be observed for rates below kd, occur at experimentally
inaccessible driving forces. (2) Electron transfer does not
proceed directly to ground state (in the inverted region)
but to electronically excited products (in the normal
region) which then decay efficiently to ground state
products.50 (3) And finally, the introduction of
competitive chemical pathways, such as H-atom transfer
followed by proton exchange with the solvent and exciplex
formation that can circumvent a simple electron transfer
pathway.51'52
More recently, Marcus and Siders have shown that the
inverted effect is diminished at large distances.53 This is
an important result because it was generally assumed for
bimolecular reactions that electron transfer takes place
only at closest contact. A typical bimolecular reaction
13
Figure 3. TTuncation of very fast electron transfer
rates by diffusion. The diffusion limiting rate is
represented by the dotted horizontal line. Thus, the
rates are leveled until AG° becomes very large.
log kobs
14
AG°(eV)
Figure 3
15
model for electron transfer is shown in Scheme 1. The first
step is the diffusion together of the two reactants to form
a precursor complex. This is followed by electron transfer
within the precursor complex to form the successor complex
kd k
M- + N3- ‘—_ M-eeeeN3-_—£t_9 M2-... N2"___9 M2- + N2-
k-d
Scheme 1
and the ultimate separation of the successor complex to form
products. The assumption of closest contact is valid for a
reaction in which equilibrium is established (i.e., ket <
kd). However, equilibrium is not achieved for the fastest
reaction (i.e., ket. z ikd) and longer' distance electron
transfer (r > a) is possible. Electron transfer at
distances larger than. closest contact. may occur if the
reactants are carrying a solvent shell or a counterion, both
of which can inhibit the closest approach of the reactants.
The long distance effect for bimolecular reactions can
be made quantitative by formulating the overall rate
constant as the harmonic mean of the diffusion-limited,
kdiff' and the activated, k rates,
act'
l/kobsd = l/kdiff + 1/kact (9)
Under steady-state conditions kdiff and k can be
154-56
act
approximated by eqs 10 and 1 where k(r) is the
16
41'1") a -2
kdiff = f 99(r) r dr (10)
1000 a
41rN co 2
kact = I gefr) k(r) r dr (11)
1000 a
unimolecular rate of the electron-transfer reaction between
reactants at a fixed center-to-center separation r, D is the
sum of the reactant’s diffusion coefficient, a is the
distance of closet approach, and ge(r) is the equilibrium
pair distribution function given in eq 12 where U(r)
represents the intermolecular potential between the
reactants.
qe(r) = exp [-U(r)/kBT] (12)
Typically, U(r) is described by the Debye-Huckel relation
given by eqs 5 and 6. Fbr nonadiabatic electron transfer,
k(r) is given by eqs 13 and 14.
ZHABZ «3 1/2 (A + AG°)2
k(r) = eXp - (13)
h AkBT 4AkBT
HAB2 = (HA3°)2 exp[-ï¬(r-a)] (14)
The distance dependence of the bimolecular electron transfer
rate can be assessed by substituting the previously
described nonadiabatic electron transfer rate expression
17
into eqs 10 and 11. Eq 14 accounts for the distance
dependence of HA3, where 5 is a measure of the conductivity
of the medium between the two redox centers, and follows
from the straightforward treatment of tunneling through a
classically impenetrable barrier.57
A major problem with testing the validity of
bimolecular theories, such as the ones described above, lies
in the difficulty with designing homogeneous systems which
incorporate both the effects of normal and inverted electron
transfer and the dependence of these two pathways on
distance. To this end, chemiluminescence (cl) is an
excellent probe of bimolecular electron transfer processes.
A chemiluminescence reaction is described by the potential
energy curve diagram shown in Figure 4 where driving force
to produce excited state and ground state products is
defined by AG°es and AG°gs respectively, and the activation
barrier for excited state production is AGes*°58 In this
figure the reaction is so exergonic that the reactant well
has become imbedded in the ground state product well.
Unlike a typical thermal reaction a luminescent excited
state product well is now energetically accessible and can
be populated by classical barrier crossing. The excited
state production efficiency, ¢est is related to the ratio of
the two competing pathways for excited state and ground
state production (defined by kes and k
gs! respectively).
fes = kes / (kes + kgs) (15)
18
Figure 4. Potential energy curve. description of
chemiluminescence. AG°es and AG°gs are reaction free
energies for electron transfer to produce excited- and
ground-state products, respectively, and AG* is the
activation energy for the excited-state reaction. Res
and kgs are the electron transfer rates for production
of excited- and ground-state products, respectively.
19
Figure 4
20
The‘ overall quantum yield of chemiluminescence, felt is
simply the product of fes and the steady-state emission
quantum yield fe'
¢c1 = ¢e¢es (16)
Since ¢e is an intrinsic property of the luminescent excited
state, it is ¢es that is fundamentally descriptive of the
efficiency of the chemiluminescent process. In a
chemiluminescent system, electron transfer in the normal
region (i.e., chemiluminescence pathway) will produce a
photon, while reaction in the inverted region (i.e., ground
state pathway) will be photometrically silent. Thus a
measure of the photons emitted per electrons transferred
allows ¢cl to be experimentally determined, and by eq 16
provides a direct probe of kinetics of electron transfer in
the normal and inverted region.
The issue of chemiluminescence efficiencies is not only
important for determining fundamental mechanistic features
of highly exergonic electron transfer but is also important
in a practical sense. Because cl represents a chemical
energy to light energy conversion process, several
applications of cl chemistry to the design of chemical based
laser systems,59"6o light emitting devices,62'63 and
electro-optical devices have been suggested. Of course the
21
practical development of such devices relies on developing
systems with high overall efficiencies.
Our interest in cl has centered on the electron-
transfer chemistry of the hexanuclear cluster system
M6X8Y62- (M = M0, w; x, Y = Cl, Br, I) whose structure
consists of an octahedral core of metal atoms coordinated by
eight face-bridging and six axial halides (Figure 5). These
cluster systems exhibit long-lived highly emissive excited
states [e.g. r0 = 180 psec, fe = 0.20 for M06C1142 in CH3CN
at 23°C] and also can. be. oxidized. and reduced. by one-
electron in nonaqueous solution.64a The magnitude of these
oxidation and reduction potentials, coupled with the low
energy' of the. emissive excited. state of ‘these clusters
permits the luminescent excited state to be populated
directly upon the exchange of an electron between M6X8Y6'
and M6X8Y63-. If the reactant precursors are generated
electrochemical 1y , the overal 1 process is cal led
electrogenerated chemiluminescence (ecl). The ecl chemistry
of the M5X8Y62' clusters is exemplified by the molybdenum
chloride cluster, whose spectroscopic and electrochemical
properties are summarized in the energy diagram in Scheme 2.
22
Figure 5. Structure of M6X8Y62" ions: 0 = Mo(II),
W(II): X,Y = Cl, Br, I.
Figure 5
24
2_*
M06C114
1.9 eV
3_ -l.56 V 2 +1.53 V _
M06C114 e M06Cll4 ) M06C114
Potentials versus SCE
Scheme 2
Electrogeneration of M06C114' and M06C1143' leads to red cl
attributable to the production of electronically excited
M06C1142'* according to the following annihilation
reaction,65
..«k 2...
+ Mo6Cll4 (l7)
M06C114" + M06c1143’———> M06C1142
The M6X8Y62- clusters offer an unique opportunity to
study the mechanism of bimolecular electron transfer in the
normal and inverted region, and provides an ideal system to
elucidate the factors governing chemiluminescence
efficiencies. Described herein is our electron-transfer
studies of the M6X8Y62' system. A fundamental issue in ecl
chemistry that heretofore has not been addressed is how the
electrochemical excitation energy between the
electrogenerated oxidized and reduced parent molecules is
distributed. As presented in Chapter III the cluster
systems have allowed the issue of energy distribution to be
addressed for the first time. Moreover, the cluster ions
possess unique properties which allow other important ecl
25
and electron transfer mechanistic issues to be investigated.
A problem of paramount importance that has eluded
identification is the factors governing partitioning between
the inverted and normal reaction pathways. The large
overpotential in the M6X8Y62' ecl reaction, has allowed the
driving force dependence of the eel efficiencies and hence
the partitioning, to be defined. These studies are
discussed in Chapter IV. with the information garnered from
the studies of Chapters III and IV, the effects of solvent,
supporting electrolyte, temperature and ligand substitution
on ecl efficiencies are described in Chapter V.
CHAPTER II
II. EXPERIMENTAL
A. Synthetic Methods
1. Preparation of flexangglea; Mglybgengm Clusters
Molybdenum dichloride, purchased from Cerac Inc.,
was dissolved in 6 M HCl and filtered to remove any
insoluble impurities. The volume of the yellow HCl solution
was reduced approximately to one quarter. As the solution
cooled, long narrow yellow crystals of (H3O)2M06Cll4 formed
in the beaker. The crystals were collected, heated to 150°C
for 2 h in vacuo to remove any excess HCl and H20, and
subsequent heating to 210 °C decomposed (H3O)2M06C114 to
MoGCllz. The 150 °C preheating step improved the quality of
the final M06C112 product. The tetrabutylammonium salt was
prepared by the addition of NBu4Cl (Southwestern Analytical)
to a 6 M HCl solution containing M06C112. The yellow
precipitate was collected and washed several times with
water and ethanol. The (NBu4)2M06Cll4 was multiply
recrystallized by slow evaporation of CHZClz previously
dried over MgSO4.
Disubstituted clusters, M06C112X22’ (X = Br, I, SCN),
were prepared by addition of X' to an ethanolic solution of
Mo6C112. For ‘the case of 1M06C112(SCN)22', the complex
slowly precipitated out of solution upon simple addition of
NBu4SCN (made by the metathesis of NaSCN and NBu4Cl in
ethanol) to M06C112 solution. For the halide clusters, a
small amount of the appropriate hydrohalic acid was
initially added to the ethanolic solution of Mo6C112.
Subsequent addition of excess NBu4Br or NBu4I, yielded the
26
27
tetrabutylammonium salt of the cluster. The purification of
thiocyanate and halide cluster complexes was accomplished by
multiple recrystallizations from methanol and
dichloromethane, respectively.
The axially substituted M06C18x62’ (X = Br, Cl, I)
clusters were prepared by dissolving M06C112 in ethanol and
then adding enough HBr or HI to yield a 1:1 volume ratio of
ethanol:hydrohalic acid.66 The exchange of the axial
chlorides with either bromide or iodide, was accomplished by
boiling the solution until the volume was reduced by 50
percent. A small volume of ethanol was added to redissolve
the solid that formed during the heating process. Addition
of NBu4I or NBu4Br to hot solutions promptly afforded a
yellow precipitate of (NBu4)2M06ClBBr6 or (NBu4)2M06C18I6,
respectively. The suspensions were gently heated and
stirred overnight to ensure complete exchange of the axial
halides. These cluster complexes were purified by using
procedures analogous to that of the tetrabutylammonium salt
of M06c1142‘.
The preparation of M06C18C16.nYn (n = 3,4,5: Y = Br, I,
SCN) necessitates stoichiometric control of the substitution
reaction at the axial ligand sites. This was accomplished
by removing the axial chloride ligands of M06C112 with the
appropriate number of equivalents of silver ion. For
example, (NBu4)2M06Cllo(SCN)4 was obtained with the addition
of two equivalents of silver p-toluenesulfonate (Aldrich) to
a methanol solution of M06C112. The AgCl precipitate was
28
removed from solution by filtration and the tetrathiocyanate
cluster was obtained by addition of an excess of NBu4SCN to
the filtrate. The precipitate was collected, washed with
ethanol, and recrystallized from methanol. The chloro-bromo
and chloro-iodo clusters were synthesized by similar
procedures: however, recrystallization of these compounds
was performed in CHZClz.
The preparation of NBu4M06C113 was accomplished by the
addition of slightly more than one equivalent of silver p-
toluenesulfonate to an acetonitrile solution of
(NBu4)2M06C114. The reaction was performed under dilute
conditions (< 5 mmolar) and stirred overnight to (i) ensure
the removal of only one chloride ligand and (ii) inhibit the
precipitation of Ag2M06C114. The AgCl precipitate was
removed by filtration and the filtrate was evaporated to
dryness in vacuo to yield NBu4M06C113. The crude product
was dissolved in CH2C12 and filtered to remove any insoluble
M06C112 that had formed during’ reaction, The filtered
solution was dried over MgSO4, and the CH2C12 was evaporated
to yield crystalline product.
Monosubstituted bromide , iodide , pyridine and
thiocyanate clusters were made by adding slightly more than
one equivalent of the corresponding ligand (tetra-
butylammonium salts of the anion donor ligand) to an
acetonitrile solution of M06C113'. The solvent was then
removed under vacuum and the resulting solid was washed with
29
methanol to remove any excess ligand. .All the
monosubstituted clusters were recrystallized from CH2C12.
The preparation of M06 bromide clusters was facilitated
by the fact that molybdenum dibromide could be purchased
from. Cerac Inc. The commercial iMoBrz was purified. by
dissolving it into ethanol followed by filtration to remove
any insoluble impurities. The ethanol was removed by
evaporation leaving a yellow-orange residue of M06Br12-
(HOCHZCH3)2. iMoGBrlz could be isolated. by’ heating the
ethanol complex under vacuum for several hours.
The tetrabutylammonium salt of MoéBruz' was prepared
by addition of NBu4Br to a ethanolic/HBr solution of
MoéBrlz. The dark yellow precipitate was washed with water
and ethanol, and recrystallized several times from CH2C12
which had been dried over MgSO4. The tetrabutylammonium
salts of the substituted molybdenum bromide clusters
MosBraBr6.nYn (n = Cl, I) were prepared and purified by the
analogous procedure described for that of the corresponding
substituted M06C18C16_nYn c1usters.57
2. e ra ' of Hex uc a t us
The method of Dorman and McCarley68 was used with
slight modification to prepare tungsten dichloride. In a
typical reaction 15 g of WClG, 1.35 g of Al metal, 6.75 g of
NaCl, and 10.00 g of AlCl3 were added in a dry box to a
quartz reaction tube. The tube was capped with a rubber
septum, removed from the drybox and connected to a high
30
vacuum manifold, evacuated for 1 h, and then flame sealed
under dynamic vacuum. The contents were thoroughly mixed
and the reaction vessel was placed into a high temperature
furnace. The furnace was heated to a temperature of 210 °C
to initiate the reaction which was allowed to equilibrate at
this temperature for 6 h. The temperature was then raised
to 450 °C over a 3 h period, held at 450 °C for 9 h, and
finally raised to 550 °C where it was held for 24 IL. The
tube was allowed to cool to room temperature. The contents
were collected by wrapping the tube in several sheets of
paper, and carefully cracking it open with a blunt object
(Caution: violent explosions sometimes resulted). The black
fused solid was dissolved in 6 M HCl/ethanol solution and
was filtered to remove any insoluble reaction products. The
light yellow filtrate was reduced in volume and, upon
cooling, greenish-yellow crystals of (H3O)2W6Cll4 formed.
The crystals were collected and heated in a furnace at 350
°C for 2 h under a dynamic vacuum to form W6C112. The
tetrabutylammonium salt of W6Cll42' was prepared by addition
of NBu4Cl to an ethanolic/HCl solution of W5C112. The
precipitate was collected, washed with water and ethanol,
and recrystallized several times from CH2C12.
Tungsten dibromide was prepared in a similar manner as
tungsten dichloride. To the quartz reaction tube 15 g of
WBrS, 0.72 g of Al metal, 7.50 g of NaBr, and 13.0 g of
AlBr3 were added in the drybox. The reaction conditions
were identical to those used for the preparation of W6C112.
31
The black solid product was dissolved into a solution of 6 M
HBr and ethanol, the solution was filtered, and evaporated
over' gentle heating to near’ dryness. The residue was
collected, dissolved in ethanol, and the insoluble alkali
salts were removed by filtration. The ethanol was allowed
to evaporate to afford greenish crystals of
W68r12(HOCH2CH3)2. The tetrabutylammonium salt of W6Br142'
was prepared by addition of NBu4Br to an ethanol:HBr
solution of W6Br12(HOCH2CH3)2. (NBu4)2W6Br14 was
recrystallized several times from acetonitrile solutions.
Tungsten diiodide was prepared by the method of Hogue
and McCarley69 with slight modifications. To a quartz
reaction tube 1.00 g of K2W6C114, 9.97 g of KI and 3.60 g of
LiI were added in the dry box. The reaction tube was
removed from the drybox, placed under vacuo, and after 1 h
it was flame sealed under dynamic vacuum. The reaction tube
was placed into a furnace and the temperature was raised
over a 1 1/2 h period to 550 °C. After 1 h at 550 °C the
tube was allowed to cool to room temperature and was opened
carefully (Caution: violent explosions sometimes resulted).
The black solid was washed with water to remove alkali salts
and iodine. The remaining yellow-brown solid was extracted
with ethanol to give a deep golden brown solution of W6I12
which was isolated by evaporation of ethanol and subsequent
heating of the solid under vacuum. The tetrabutylammonium
salt of W6I142' is obtained by addition of NBu4I to an
ethanol solution of W6I12. The dark yellow powder was
32
.collected, washed several times with ethanol, and
recrystallized several times from dry CH3CN to yield pure
(NBu4)2W6114.
Mixed tungsten halide clusters, W6X8Y62' (x, Y = Cl,
Br, I) were prepared by dissolving (W6X8)X4 in ethanol and
6 M hydrohalic acid HY where X x Y. The solubility of the
cluster in HY decreases along the series HCl > HBr > HI and
thus a larger amount of ethanol must be added. To ensure
complete exchange of axial halides the resultant solution
was evaporated to near dryness with moderate heating. When
preparing W6I8Y6 (Y = Cl, Br), gentle heating under a vacuum
aspirator was required. An insoluble precipitate formed
when the solution was heated too rigorously. The residue,
collected from solvent evaporation was redissolved in
ethanol: 6 M HY solutions. The tetrabutylammonium salt was
obtained by addition of excess NBu4Y to the ethanol/HY
solution. The mixed halide tungsten clusters (NBu4)2W6x8Y6
were recrystallized several times from either dry CH3CN or
CH2C12.
3. Orggnig Donors and Aggeprors
Nitroaromatics and aromatic amines, with the
exception of tris (p-tolyl)amine which was synthesized
following published procedures"), were obtained from
commercial sources (Aldrich Chemical Company, Alfa Products,
and Pfaltz and Bauer). Solids were purified by
recrystallization followed by vacuum sublimation and liquids
33
were purified by fractional distillation. The pyridinium
salts were synthesized by addition of either methyl iodide
or benzyl chloride to a 1:1 acetone/ethanol solution of the
appropriately substituted pyridine. Isonicotinamide
(Sigma), 4-cyanopyridine (Aldrich) and isonitotinic acid
ethyl ester (Sigma) were used without subsequent
purification. The bipyridinium salts were synthesized by
dissolving the appropriate bipyridine (4,4'-dimethyl-2,2'-
bipyridine, 4,4'-bipyridine, 2,4-bipyridine, 2,2'-bipyridine
and 1,2-bis(2-pyridyl)ethylene were purchased from Aldrich
and used as received) in a neat solution of methyl iodide,
1,2-dibromoethane (Aldrich), 1,3-dibromopropane (Matheson,
Coleman & Bell) or 1,4-dibromobutane (Aldrich), and by
gently heating these solutions overnight. The resulting
precipitate was collected and washed with ethanol and
acetonitrile to remove any starting compound or
monosubstituted products. The filtrate could be heated
further to generate additional disubstituted product.
Pyridinium and bipyridinium hexafluorophosphate salts were
obtained by the addition of ammonium hexafluorophosphate to
aqueous solutions of the bromo, chloro or iodo salt, and
were twice recrystallized from acetone/water solutions.
4. u or ' e t s
Tetrabutylammonium hexafluorophosphate and
perchlorate (Southwestern Analytical Chemicals) and
tetrabutylammonium tetrafluoroborate (Aldrich) were
34
dissolved in ethyl acetate, dried over M9804, and
recrystallized from pentane/ethyl acetate solutions. The
salts were dried in vacuo for 12 h at 60 °C to ensure that
the ethyl acetate was completely removed. Tetramethyl-
ammonium and tetraethylammonium hexafluorophosphate (Fluka)
were dissolved in acetonitrile and recrystallized from a
water/acetonitrile solution and dried in vacuo for 12 h at
100 °CL Potassium hexafluoroarsenate (Ozark-Mahoning) and
trifluoromethane sulfonic acid were converted into their
corresponding tetrabutylammonium salts by dissolving them in
water and adding excess NBu4Br. The resulting precipitate
was collected and dried in vacuo at 60°C for 12 h. Lithium
and sodium perchlorate (Fischer Scientific) were
recrystallized from acetonitrile solution and dried under
vacuo at 100°C for 6 h.
5' Solvents
Dichloromethane, acetonitrile, acetone, dimethyl-
formamide, and 1,2-dichloroethane purchased from Burdick &
Jackson Laboratories (distilled in glass grade), were
subjected to seven freeze-pump-thaw (fpt) cycles and vacuum
distilled onto 4-A molecular sieves (except acetonitrile
where 3-A molecular sieves were used) contained in a 1-liter
round-bottom flask equipped with a high-vacuum Teflon value.
Because acetone undergoes a condensation reaction in acidic
media to produce mesityl oxide, it was vacuum distilled from
the sieves after 24 h. Butyronitrile and benzonitrile were
35
purchased from Aldrich (Gold Label) and were used as
received. Propionitrile, purchased from Aldrich was treated
with dilute HCl to remove isOnitrile and after the
extraction, was sequentially dried over MgSO4 and CaHz, and
finally fractionally distilled from P205. Butyronitrile,
benzonitrile, and propionitrile were subjected to seven fpt
cycles and vacuum distilled onto 4-A molecular sieves.
B. Experimental Methods
1. Characterization of Molybdenum and Tgngsren
Clusters
A thorough characterization of all the molybdenum
and tungsten cluster systems was performed by negative ion
Fast Atom Bombardment Mass Spectrometry (FABMS). The
technique utilizes a 10 keV Xenon beam which bombards a
sample placed in a high viscosity, low vapor pressure
matrix. The energy from the beam is transferred to the
matrix which causes desorption of matrix sample ions into
the gas phase. The negative ions generated during the
absorption process, were isolated and recorded to give
fingerprint spectra of each cluster. FABMS is a superior
analytical method for the synthesized clusters because
substituted halide clusters are easily detected. The
detailed experimental results and a general discussion of
the usefullness of this technique is presented elsewhere.71
36
2 - Ware
Formal reduction potentials of acceptors and donors
were determined by cyclic voltammetry using a Princeton
Applied Research (PAR) Model 173 potentiostat, Model 175
programmer, and a Model 179 digital coulometer. The output
of the digital coulometer was fed directly into a Houston
Instrument Model 2000 X-Y recorder. A three-electrode
system was used with a standard H-cell configuration. The
working electrode was a Pt button, the auxiliary electrode
was a Pt gauze and an Ag ‘wire served as an adequate
reference potential by using ferrocene as an internal
72 Potentials were related to the SCE reference
standard.
scale by using a ferrocenium-ferrocene couple of 0.31 V ye.
SCE. The molybdenum and tungsten cluster potentials were
measured under high vacuum conditions in a single
compartment cell (vida infra) to obtain accurate E1/2
potentials.
3. Qgegehigg Measurements
Electron—transfer quenching rate constants for the
reaction of M06C1142'* with nitroaromatics, aromatic amines,
and pyridinium compounds in CH3CN and CH2C12
([(NBU4)2MO5C114] = 3 mM, p = 0.1 M NBu4ClO4) were
determined from Stern-Volmer plots of the M06C1142'*
luminescence intensity. The quenching rate constants of
—*
M06C1142 by 4 , 4 ' -dimethoxydiphenylamine and 3 , 5—dichloro-
p-benzoquinone in CH3CN, acetone and propionitrile
37
[(NBu4)2M06C114 = 1 mM] were determined from Stern-Volmer
-*
plots of M06C1142 lifetimes. Additionally, the Stern-
Volmer lifetime method was used to study the energy-transfer
quenching of Mo6C1142’*
and w6x8Y62‘* by wsxsyéz' or
Mo5C1142' , respectively , in CH2C12 . Stern-Volmer
experiments were performed over a quencher concentration
range of 5 x 10"4 to 1 x 10'1 M and Stern-Volmer constants
*
were calculated by using ro(M06C1142') = 180 u sec in
CH3CN, = 160 p sec in CH2C12, = 170 p sec in acetone, = 148
*
u sec in propionitrile, 10(W6I142' ) = 19 )1 sec in CH2C12
and ro(w6188r62'*) = 16 p sec in cuzc12 at 23 °C.
Quenching experiments were performed in a specially
constructed high-vacuum cell, consisting of a l-cm cuvette
attached to a sidearm terminating with a 10-ml round-bottom
flask. Solvents were vacuum distilled into the quenching
cell and freeze-pump-thawed three times. All quencher
additions were performed under high vacuum conditions.
Luminescence intensities (*exc = 436 nm) were measured on a
high resolution emission spectrometer and emission lifetimes
(*exc = 355 nm) were acquired with a pulsed laser system
(Nd:YAG, FWHM = 8ns) . Both instruments were constructed at
Michigan State University and are described elsewhere.73
4. t e d ' s c
(i) Quantum_xields
A triple step square wave potential sequence
generated by the PAR 175 programmer was used to establish
38
ecl reactions. The potential limits of the program sequence
were chosen to ensure production of electrogenerated
intermediates in the mass-controlled region. The
electrochemical cell employed in ecl measurements was a
cylindrical, single-compartment high-vacuum cell. A sidearm
permitted solvents to be transferred into the cell by vacuum
distillation and two sample chargers allowed cluster and
electroactive acceptor or donor to be added independently to
the working electrode compartment while maintaining the
isolated environment of the electrochemical cell. Two
tungsten wires sealed in uranium glass served as electrical
leads to the Pt mesh auxiliary electrode and an Ag wire
quasi-reference electrode. The auxiliary and reference
electrodes were positioned parallel to a Pt disk working
electrode (A = 0.0314 cmz) which was positioned
centrosymmetrically along the cylindrical axis. of the
working compartment. The Pt disk was spectroscopically
viewed through a fused silica window which constituted the
bottom surface of the electrochemical cell. After each
experiment the Pt disk electrode was polished with 1 pm
diamond paste and 0.05 pm alumina purchased from
Bioanalytical Systems.
Ecl spectra and quantum yield experiments were
performed in solutions containing 0.1 M supporting
electrolyte and equimolar concentrations of M06C1142' and
electroactive acceptor or donor. Samples for all ecl
experiments were prepared by transferring the appropriate
39
amount of solvent under a high-vacuum manifold (1 x 106 to 5
x 10'6 torr) into the cell sidearm which contained
supporting electrolyte previously heated at 100 °C for 1 h.
After 3 fpt cycles, the solution was thoroughly mixed and
poured into the working chamber by slowly rotating the cell
by 90°. The current response of the solution containing
only supporting electrolyte was recorded before undertaking
ecl measurements. Background current densities of 48 pA/cm2
in CH3CN and 35 pA/cm2 in CHZClz were measured at potential
limits of -2.0 V and +2.0 V.
The quantum yield for ecl is defined by the fellowing
expression where I is the total ecl intensity
(einsteins/sec)
a
feel = I Idt/Q (18)
0
over a finite period of time t and Q is the total cathodic
or anodic charge. The ecl yield is equivalent to the number
of photons produced per electron transferred and
consequently fecl can be determined by measuring the number
of photons emanating from the electrode surface and the
number of equivalents of electrogenerated species. The
latter' quantity can. be :measured coulometrically' by
monitoring the anodic, QA, or cathodic, Qc' charge passed
into solution during an ecl experiment. QA and Qc were
determined by pulsing from the foot of one wave to the
40
diffusion controlled region of the other. In regard to the
former quantity, absolute ecl intensity measurements were
performed by using an EG & G Electro-Optics 550-19
integrating sphere and an EC & G Model 550-1
photometer/radiometer equipped with an EG & G Model 550-2
multiprobe detector. A flat response between 450 and 1100
nm was achieved by fitting the multiprobe with a radiometric
filter attachment provided by ES & G. Appropriate
corrections were made for the radiometric filter attachment
which allowed only 14 percent of the light to be
transmitted. Integration of the ecl intensity was
accomplished by using a Model 550-3 pulse integrator.
Calibration of the integrating sphere was performed by EG &
G Electro-Optics Division by using photometric sources
certified by the National Bureau of Standards. This
calibration led to a correction factor of 3774 at the
wavelength of M06C1142’ emission.
Ecl yields were calculated with appropriate corrections
for reflectivity of the electrode and. non-faradaic
contributions to the integrated current according to the
methods described by Bard.74 The reflectance of the
polished Pt disk electrode was taken to be 0.67 at the
wavelength of Mo6C1142' emission. The double layer charging
components qA and qc were measured by pulsing the electrode
between an anodic limiting potential set at the foot of the
oxidation wave and a cathodic limiting potential set at the
foot of the reduction wave. From these potential limits,
41
qA's and qc’s were measured in 200 mV increments by setting
the anodic and cathodic limiting potential 200 mV negative
and positive, respectively. Plots of qA and qc ye. AE3/2
were linear and the background current passed at potentials
used in ecl experiments were obtained by extrapolation of
this plot.
Measurements of the ecl efficiency of Ru(bpy)32+, which
has been determined in several previous studies, was
undertaken in an effort to allow us to check our
experimental apparatus and procedure. An acetonitrile
solution containing Ru(bpy)32+ (p = 0.1 M NBu4ClO4,
[Ru(bpy)32+] = 3 mM) was prepared in the high-vacuum
electrochemical cell and ecl measurements were performed
with the integrating sphere contained in a light-tight box.
An ecl yield for a single run was determined from twenty
measurements of the intensity generated from a single
triple-step potential sequence. The system was allowed to
equilibrate 30 seconds between each pulse sequence. The
overall yield calculated from five separate experiments was
¢ec1[Ru(bpy)32+] = 0.046 2 0.004. This value is in good
agreement with the previously reported efficiency of 0.05 in
cu3cu at 23°c.75
Quantum yield measurements of all ecl systems followed
procedures similar to those described above. For M06C1142'
/acceptor and donor systems the cluster ion and
electroactive reagent were contained in separate sample
chargers. Prior to the addition of a given acceptor or
calculated directly from the following expression, where
foecl is the ecl efficiency of
Q0
I
Q Io
¢ecl = foecl ' (19)
u06c1142“ (p = 0.1 M NBu4ClO4 in CH3CN or cnzc12 at 23°C),
Q“ and Q are the charges passed into solution, and 1° and I
are the measured integrated photon intensities of solution
containing cluster and solution containing cluster and donor
or acceptor, respectively. Ecl yields were calculated from
averaging three experimental runs of ten measurements: error
limits of ¢eclt measured by this method, were 2 15% in CH3CN
and 2 10% in CHZClZ.
Ecl quantum yields for the M06C1142'7W6X8Y62" systems
and. all systems. described in. Chapter’ V’ were calculated
relative to ¢ecl of M05C1142' for the former and to feel of
Ru(bpy)32+ for the latter.
(ii) Spectra
Ecl spectra of M06C1142'7acceptor and donor
systems were recorded between 350 and 1100 nm, and between
550 nm and 1050 nm for’ M°6C1142-/W6X8Y62- ecl spectra.
Spectra were obtained interfacing the specially designed
electrochemical cell directly to the detection side of the
emission spectrometer. The lock-in amplifier was referenced
to the ecl signal with the cycle synchronous output of the
43
donor to the working electrode compartment, the eel yield of
a solution containing only M06C1142' was determined from the
average of a minimum of ten pulses. Donor or acceptor was
then introduced to the solution and fecl was measured by
using a pulse sequence with potential limits appropriate to
the system under investigation. This procedure permitted us
to identify anomalous ecl measurements by monitoring the
M06C1142' ecl efficiency. Error limits for M06Cll42'/donor
and acceptor ¢ecl values, determined from three experimental
runs composed of ten ecl intensity measurements, were 2 12%
in CH3CN and r 10% in CHZClz.
Accurate determination °f¢ecl for acceptor and donor
systems exhibiting the weakest ecl intensities were hampered
by the low throughput of the integrating sphere. For these
systems, the electrochemical cell was positioned directly on
the face of the multiprobe detector. The ecl efficiencies
of M06C1142'/acceptor and donor systems were estimated using
M06C1142' as a relative standard. In an experimental run,
M06C1142', was initially added to the solution and the ecl
yield was determined from a minimum of twenty intensity
measurements. The electroactive organic reagent was then
added to the solution and the eel yield for the M06C1142'
/acceptor or donor system was recorded. In this manner,
errors due to geometric positioning of the cell on the
detector were minimized. Because the spectral distributions
of the two experiments are identical, the eel quantum yield
of M0601142'/acceptor and donor system, feclt can be
44
PAR 175 programmer. The single compartment cell is very
similar to the one previously described except the optical
window is on the side of the cell. The working electrode is
a platinum disk (area = 0.0707 cmz) sealed in glass but
whose face is now perpendicular to the bottom of the cell.
The Pt disk electrode was positioned at the focal point of
the collection lens on the detection side of the emission
spectrometer. The ecl was generated by using a cyclic
square wave (10 to 20 Hz) with potential limits appropriate
for the system under investigation.
Ecl spectra were recorded on a Zenith microcomputer.
The deconvolution of M06C1142'/W6x8Y62' ecl spectra was
performed by a band shape analysis in which varying ratios
of the steady state emission spectra of the discrete cluster
systems were added until their sum identically matched that
of the measured ecl spectrum. Energy partitioning values
were obtained by normalizing the calculated ratio with the
emission quantum yields of the two cluster systems. Errors
in the partitioning ratio were determined by three
measurements of the M06C1142'/W6XBY62' ecl spectrum and
multiple deconvolutions of each spectrum.
CHAPTER III
III. ELECTROCHEMICAL EXCITATION ENERGY PARTITIONING
IN MIXED CLUSTER ELECTRON TRANSFER REACTIONS
A. Background
A fundamental issue of chemiluminescence reactivity,
which heretofore has not been resolved, involves the
parentage of the luminescent excited molecule produced in
the electron transfer reaction between electrogenerated
oxidized and reduced reactants. For instance, in the
M6X8Y62' system, the energy released in the annihilation
reaction is sufficient to leave only one cluster anion in
its excited state. The important question here is whether
the electronically excited ion is generated from M6X8Y6' ion
or the M6X8Y63- ion. The energy partitioning between these
two parentages is ultimately related to differences in the
activation barrier heights and electronic variations of the
two discrete electron transfer pathways shown below.
2n 2-
-M6X8Y6 + M6X8Y6 (20a)
- 3..
M6X8Y6 + M6X8Y6
2- 2—*
M6X8Y6 + M6X8Y6 (20b)
Thus, partitioning of the electrochemical excitation energy
between reactants in ecl reactions has important electron
transfer implications and will shed light on factors that
limit the overall efficiency of ecl reactions.
For all ecl systems to date, the partitioning issue has
not been successfully addressed because there has been no
simple way of labelling the oxidized or reduced reactant and
therefore identifying the parent of the excited state. One
45
46
approach is to employ structurally similar reactants which
possess energetically distinct luminescent excited states,
thereby allowing the parentage of the excited state species
to be identified spectroscopically. An attempt to identify
the precursor of the excited state in an ecl system was
undertaken in a recent study of ruthenium polypyridine
complexes where the electrochemical properties of ruthenium
reactants and spectroscopic properties of the eel products
could be tuned with the polypyridyl ligation coordination
sphere.76
Relevant photophysical and electrochemical
properties of these systems are shown in Table 1. The
investigated reactions are shown in eq 21 and 22 where the
distinct emission energies give rise to different ecl
Ru(DTB-bpy)32+* + Ru(bpy)2biq2+ (21a)
Ru(DTB-bpy)33+ + Ru(bpy)2biq+
Ru(DTB-bpy)32+ + Ru(bpy)2b1q2+* (21b)
Ru(bpy)32+* + Ru(bpy)2DMC1-l2+ (22a)
Ru(bpy)33+ + Ru(bpy)2DMCH+
Ru(bpy)32+ + Ru(bpy)2DMCH2+* (22b)
spectra for the two pathways. However, several problems
were encountered with those ruthenium polypyridyl systems.
These problems included: (1) the free energy of reaction 21
is not large enough to produce the excited state of Ru(DTB-
47
wow .w> oHasoo +m\+msm or» now masï¬ucouoo coï¬uosnoe m<
mom .m> oaosoo +\+mcm one new mamï¬pcoaoc scauosoon we
madam oouï¬oxo wcï¬ueï¬so moonQEoo anaconuse one we mmuoco Ono
mssï¬oseeeesmeemoH.HLHn.nLoNemaneemseesw.hrosesEeens.m u more
mN.H+
®N.H+
mm.H+
HH.H+
HA+N\+MVN\HH
ocaaocï¬30ï¬oi.m.m u can one mcï¬nwnmcï¬o u woo
messesseeeeseseesmeee u Assumes
oo.H- ms.e o +mxmosavmxeeevsm
mm.eu me.m +mmxeesvsm
Hm.o- os.e e +mxseevmï¬seevsm
es.er we.m e +mmï¬seermeocsm
>\mom .u> s>oxenem
ol+\+se«\es
oauhaac0u00< sq mOKOHQEOU showuhnhuomlsm Ho
mowanonoum HdOï¬EGQOOHuooHM can neuwuonm madam coauoxm
H OHDSB
(6.0013094
48
b93032+
and thus reaction pathway 21a is energetically
unfavorable: (2) the oxidation potentials of Ru(bpy)32+ and
Ru(bpy)2DMCH2+ are similar and hence do not permit selective
production of Ru(bpy)33+ and therefore the annihilation
cannot cleanly be established: and finally ( 3) favorable
energy-transfer reactions between. the ruthenium. products
permitted the excited state energy to be redistributed
between both products thereby vitiating meaningful energy
partitioning ratios to be determined from measured ecl
intensities. For these reasons, the authors correctly
concluded that reactions 21 and 22 were not useful in
gathering quantitative energy partitioning data.
Many of the problems inherent to the ruthenium
polypyridyl complexes are circumvented by the M6X8Y62' ions.
Similar to the ruthenium complexes, the luminescent excited
state and electrochemical properties of M6X8Y62' ions can
also be varied with the ligating coordination sphere: and
therefore, the approach established by reactions 20a and 20b
can be pursued with mixed cluster systems (e.g. M6X8Y6' +
M6X8Y63'—-> M6X8Y62J or M6X8Y62'*). However, unlike the
complicating problem of the system described by reaction 21,
the relatively large oxidizing and reducing potentials of
M6X8Y62' ions compared with the relatively low excited state
energies allows for the possibility of either product to be
left in its electronic excited state in the annihilation
reaction. Furthermore, the problem associated with reaction
22 is avoided owing to significantly different redox couples
49
of substituted M6 cores. Of equal significance, the poor
overlap between the absorption and emitting states of
substituted M6 cores means that energy transfer is
inefficient. Thus, the measured ecl emission of.My§93' and
M6x8Y62' should accurately reflect the original partitioning
of the electrochemical excitation energy in the annihilation
reaction. The results presented in this chapter
successfully' address for' the first. time, the. effect of
energy partitioning in an ecl annihilation event as well as
shed. light on important electron 'transfer’ properties of
M6X8Y62- ions.
B. Results and Discussion
Electrochemical and photophysical properties of the
molybdenum and tungsten halide cluster systems in CH2C12 at
room temperature are shown in Tables 2 and 3. Most of the
cluster systems exhibit reversible one-electron oxidation
processes. The criteria used to establish reversibility
were ia/ic ratios varying between 0.95 to 1.05 and linear
plots of anodic and cathodic peak currents vs. (scan
1/2.
rate) Anodic to cathodic peak separations of the
reversible cluster systems were comparable to that measured
for ferrocene (125 mV), thereby establishing that deviations
of AEp from the theoretical value of 59 mV are due primarily
to uncompensated cell resistance. The remaining M6 clusters
either possess an irreversible oxidation couple or multiple
oxidation waves with the first being chemically
50
.mHaï¬ucoaoa xaoc one couhooou mosaa>
.moacsoo coï¬ucoï¬xo oHchasa p
.oaoemuo>oupï¬ ma oHonoo coï¬aocuom o
.mom .m> eacsoo im\immo= use rm\i®o: on» new maaaucouoq coï¬woscou m< o
.ochQmop houoaohpoocm scammï¬eo new concouuoo «Hommu ca oohsmsoe «Sexes coï¬mwï¬em m
omw.HI
omw.Hl
ONm.HI
OHw.HI
mb.Hr
mm.HI
0®.HI
h®.HI
m®.Hi
ON.HI
>\mcm
e1-e\-«cs\em
NHUNIU cu mucumSHO was new weauhocOhm unawaoa00haoo~m
N Oundh
oo.e+
HH.H+
so.e+
«H.H+
as.o+
eme.m .ooH.m
som.e .eeo.e .oee.e+
umv.a+
ems.e .smh.e .ome.e+
sm.e+
sm.e+
em.e+
.m>
DAIN\IVN\Hm
NNw
ooh
mph
mwh
omh
Hob
bmh
mam
mmh
0mm
mph
wmh
sss\xea
4
umeemsmoos
Immaowumwoz
INNHUNHHmmoz
-meahmmos
Namsmavmaï¬omo:
-mexzumceeeomoz
immxzomvmeeoeos
racemeoeoz
INNHNHHO®OE
-mosmweoeoz
INNSmNHHUGOS
INvflaomoz
whoawnï¬o 00:
one deals: scummnam
51
.NHonU Ho ocsoewxoso one cw we oaasoo Im\|Nm3 one o
.mom .w> mamnoo rm nos one see chHpcopoa coï¬posooe m< o
.omcoomoh sopo80hpoocm newmmweo new oouooesoo Hummu ca consumes wasaï¬xss conwHEm a
on wh.o+ oos immeomeos
o- mh.o+ cos usesmmees
on sm.o+ was -mvHHms
or vs.o+ mes :mmesmms
oh ss.o+ Hes -mteowsmms
or ow.o+ ems -meesmes
on em.o+ com -moeweoes
o- mm.o+ sew -mosmmeoes
or mm.o+ mmm -meeeoms
>\mom .m> essxxuse antennae @-
Aum\rucm\em exuu\uVN\Hm
NHUNMU cw whoamsflo oh how medaHOAOHA Houdamnoouuooam one sauna: scammeam
m OHQSH
52
irreversible. In general, clusters exhibiting multiple
oxidation waves contain ligands that are weakly bonded (e.g.
I' or SCN') in the axial positions. For example, the cyclic
voltammograms of M06C112L22' (L = SCN, I) and M06C112 are
shown in Figure 6. In panel a, scanning a Pt electrode
immersed in a CH3CN solution containing MoGCllz, anodically
produces two quasireversible waves at +1.73 V and +1.92 V
potentials. These two waves are preserved in the cyclic
voltammograms of M06C112(SCN)22 and M06C112I22 (Figure 6b
and 6c). The preceding irreversible wave at +1.43 and +1.61
in the cyclic voltammograms of M06C112122' and
M05C112(SCN)22' respectively, is due to oxidation of
dissociated I†and SCN'. That the potentials for this
oxidation do not correspond to those of free ligand suggests
dissociation subsequent to cluster oxidation. Of course the
absence of L in M06C112 precludes the appearance of this
preceding irreversible oxidation wave. The similarity of
the oxidation profiles of M06C112 and M06C112L22' suggest
facile dissociation of the heterodonor ligand from the
cluster core to give MoGCllz.
Whereas oxidation of the clusters is generally
reversible, the reduction of the clusters are for the most
part irreversible processes. This is not the case, however,
for [M06C181X62 and [Mo6C181X5L' (X = halide, L = donor
ligand) ions, which possess reversible or more typically
quasi-reversible one-electron reduction couples. All other
M06 clusters possess irreversible reduction waves. For the
53
Figure 6. Cyclic voltammograms of (0.1 NBu4PF6 at
23°C) (a) M06C112 (3 mM in CH3CN): (b) M06c112123' (3
mM in CH2C12): (c) M06c112(SCN)22' (3 mM in CH2012).
54
lpAi
l l l l 1 L L J I I 1 I
+0.9 +l.l +l.3 +l.5 +l.7 +|.9 +2.|
V/SCE
Figure 6
55
IpAI
l
J
+0.9 3i.)
L 1 L n l 1
+l.3 +|.5 +l.7 4-
V/SCE
Figure 6
|.9
+2.|
56
5pA1
L l 1 l 1 l I l 1
+09 jil +U3 +H5-H7 4+9
V/SCE
Figure 6
57
case of the tungsten clusters, the reduction waves are in
the background of CH2C12 thereby implying reduction at
potentials < -2.2 V.
Except for the thiocyanate and phosphine clusters, ecl
is observed for the M06 clusters shown in Table 2 at a
platinum electrode according to the simple annihilation
reaction shown in eq 17 (the properties of these ecl
reactions will be discussed in Chapter V). Conversely, the
tungsten halide clusters of Table 3 produce weak or no ecl
at a platinum electrode surface (feel 3 10's) in THF
presumably due to the fact that the reduced cluster anion
W6X8Y63' cannot cleanly be electrogenerated. However, the
issue of interest here is not ecl reactions of reactants
electrogenerated from the same parent molecule, but ecl
systems which permit the partitioning of the electrochemical
energy to be distinguished (i.e. reaction 20). The choice
of the appropriate cluster systems for such studies relies
on M6X8Y62’ and M6X8Y62" possessing sufficiently different
emission energies such that the electrochemically produced
excited state can be spectroscopically distinguished.
Experimentally a difference of 50 nm in the emission spectra
of M6X8Y62' and M6X8Y62' can easily be resolved. Moreover,
appropriate reduction potentials are critical to the
selective production of only one oxidized and one reduced
reactant in the annihilation reaction. The M6X8Y6‘/2‘
potential must be at least 100 mV negative of the M6X8Y6'/2—
potential and the M6X8Y62'/3' potential must be 100 mV
58
positive of the M6X8Y62’/3' potential for reaction 20 to be
established.
The choice of’ mixed cluster systems for ecl
annihilation studies can now be made with facility upon
inspection of the data in Table 2 and 3. Because the
tungsten halide clusters possess reversible oxidation
couples, but irreversible reduction couples, the W6 clusters
can only be used as monoanions in the annihilation reaction
of mixed cluster ecl systems. The choice for a trianion in
the mixed cluster ecl systems is limited to clusters with
the formula of M06C18Clnx6-n2' (n = 0—6) (X = Br, I) because
only these clusters possess reversible reduction potentials.
Owing to mechanistic problems (which will be discussed in
Chapter V) with the ecl reaction of clusters with Br or I
occupying axial coordination positions, the only reasonable
choice for the trianion in the mixed cluster ecl reaction is
M06C1143’.
Because the W6X8Y6'/2' redox couples are negative of
the M06C114"/2' redox couple and the W6X8Y62'/3' redox
couples are more negative than the reduction of Mo6Cll42',
the following electron-transfer reaction can cleanly be
-* -
+ w6x8262 (23a)
2
M06C114
2- 2-*
M06c114 + wéxgy6 (23b)
3 -
2- 2-
M06C114 + w6x8Y6 (23c)
59
established by standard electrochemical techniques.
Inspection of Tables 2 and 3 reveals that the W6X8Y62"
excited state energies are significantly different than that
of M05C1142'. These features permit the calculation of
partitioning ratios directly from ecl spectra of the
M06C1142'/W6X8Y62' systems if one assumes that subsequent
energy transfer between the products in reactions 23a and
23b is unimportant. In order to determine whether this
assumption is valid, energy transfer studies were undertaken
2.7::
where the quenching of W6X8Y6 and M06C1142'* by M06C1142
and W6X8Y62’, respectively, was measured. Quenching rate
constants in CH2C12 for the following reactions,
2-* 2- 2- 2-*
2- 2-* 2-* 2—
were deduced from classical Stern-Volmer analysis of the
cluster lifetimes and the results of these studies are
displayed in Table 4. Because both cluster ions emit, the
individual lifetimes were determined from a multiexponential
fit of the luminescence decay with the equation y = ae't/'1
+ be't/’2 by using the general nonlinear curve-fitting
program Kinfit77a where a and b represent the fractions of
total emission decay described by the excited-state
lifetimes 71 and 72, respectively. Convergence of the fit,
monitored by the sum of the squares of the residuals, yields
60
.mpsn oEHpoHHH Ho muflm Heepcosoaxoflpasa Eosw vocflEeopoU museumsoo open mcï¬nocosa o
.A‘HOIOWI GOIOWV N OOQ
â€sonocosv ow oeonQOESH map Bose pommcmsp sweoco new ooeoe mcH>HHQ m
sea v H.011
see A h H.o-A
toe x «.4 0.0
. 0.0
see A m he.o-
see x s mH.o-
neumeua\er s>o\.o<
NHONHU cw muoumsHO Ithwle no mswnocosa Romances
fl OHDSB
-mmsmweoes
umeeeoos
immeomsmes
INVHHUQOS
Imwï¬ï¬omos
INVHHUQOS
Honouona
INvHHUQOE
INVHHUQOE
IN¢HHD®OS
Imvï¬hmms
-meeems
ummemmHos
Openness:
61
values for a, b, r1 and 72. For an individual reaction
pair, obviously only one energy transfer reaction, 24a and
24b, will be energetically downhill. Because endergonic
energy transfer is a relatively inefficient process,77b only
the exothermic energy transfer reaction will result in an
attenuation of cluster’s lifetime. For instance for the
case of the M06C1142'/W6I142' system reaction 24a is
endergonic and therefore the Mo6C1142' lifetime does not
change. On the other hand, the exergonicity of reaction 24b
(-0.15 eV) results in a quenching of the W6I142 lifetime
with increasing concentration of M06C1142â€. The
experimental manifestation of these energetics is that the
M06C1142' is the quencher and its lifetime is constant while
W6I142' is the lumophore and its lifetime follows a
classical Stern-Volmer dependence. The largest quenching
rates are only 107 despite significant driving forces for
some reactions. In the context of a Forster energy transfer
mechanism, the poor spectral overlap of the absorbing and
emitting states precludes efficient energy-transfer
quenching.77b
In a Dexter energy transfer treatment, the
good orbital overlap of reactants required for efficient
energy transfer is precluded by the fact that the metal
localized emissive excited state of M6X8Y62' ions is
sterically shielded by the halide coordination sphere.77b
Thus, because the observed energy transfer rates are well
below the diffusion controlled limit of ~109, energy
transfer is unimportant in energy partitioning studies and
62
measured excited state production yields should accurately
reflect the partitioning of the electrochemical excitation
energy.
In Table 5 are shown the photophysical properties and
overall ecl quantum yields of the mixed cluster ecl systems
employed for partitioning studies. Chemiluminescence from
CH2C12 solutions containing M05C1142' and W6X8Y62' is
observed when the applied potential of a Pt electrode is
stepped into the oxidation wave of W6X8Y62' and the
reduction wave of M05C1142-. Overall ecl quantum yields
were determined by dividing the number of einsteins
emanating from the electrode surface by the number of
equivalents of M06C1143' or W6X8Y6' produced.
The partitioning ratios for reactions 23a and 23b for
the M06C1142'/W61142‘, W6I83r62', and W6C18Br62' systems can
be directly determined from ecl spectra. An exemplary mixed
cluster ecl spectrum of the M06C1142-/W61142' system is
illustrated in Figure 7. The broad featureless band is
characteristic of M6X8Y62' cluster emission. The larger
signal-to-noise ratio of the eel spectra, as compared to the
steady-state luminescence spectra, is due to mechanical
agitation of the solution over the duration of the
relatively' long' pulse sequence typically’ needed for’ ecl
experiments. The driving force to produce either M06C1142'*
or W6I142'*, calculated by summing the O-O energy of the
emitting excited state with the ground state reaction free
energy (AGes = AGgs + EO-O) as determined from standard
63
E H.0
000.0
000.0
000.0
0¢0.0
000.0
$00.0
snows
.o.mm be msoesmz
u : .NHONmO cs maopmmm Hoe sopmsao posse sow mnaoï¬a espouse How Hadno>0 o
.onm .mouv oasepchpoos ca muaoï¬m espouse scammï¬eo nousmsoe use «Hommo
ca maï¬powï¬a consumes one 809% oopseï¬emo weaves espouse scammï¬ao oedemrmcsoum o
.oomm us Naommo ca cousmdoe came» Enacsso coï¬wmï¬ao opdpmizosopm m
no.0 wH.0
0H0.0 wH.0
mH.0 , mH.0
00.0 wH.0
wH.0 wH.0
Om.0 wH.0
awhe SOOIe
-mosmweoms\umseeooo2
INVHHO@3\IN¢HHU®OS
INvHHm03\IN¢HHomoS
-moeomsmos\umeeeooos
IwmhmmH®3\INvHHO®OE
-meHHbs\umsHHooos
Ecumhm
hoausao some:
anewuocom Hum RoamSHU come: a“ comb mnouwSHU inmwmxwl
Ho moaowh savanna Hum Haehobo use newuhoaoum Acoum5o0090am
m OHDSE
64
Figure 7. Electrogenerated chemiluminescence spectrum
of Mo6011437w5114" in CH3CN (p = 0.1 M NBu4PF6 at
23°C).
65
EMISSION INTENSITY
550
650
760 ' 850
A/nm
Figure 7
I
séo
it Am -
I050
66
reduction potentials, is energetically downhill for both
reaction pathways. Figure 8a shows the ecl spectrum of
M06C1142'/W6I142' superimposed on the steady-state emission
spectra of the individual cluster ions. It is clearly
evident from Figure 8a, that the ecl spectrum is composed of
the emission spectra of Mo6C1142' and W61142'. The
partitioning between reactions 23a and 23b is obtained, as
shown in Figure 8b, by adding together varying amounts of
the steady-state emission spectra of M06C1142’ and W31142'
until the sum is identical to the measured ecl spectrum.
The partitioning value, PR for reactions 23a and 23b is
obtained by normalizing the ratio of the measured
contribution of the individual (6M06 and awe) to the overall
ecl spectrum with the cluster emission quantum yields (om)6
and 41w“), where M06 and W6 represent the M06C1142' and
W6X8Y62' cluster ions, respectively.
9 ¢
PR e M°o "6 <25)
For the M06C1142'7W6I142', W6183r62- and W6C18Br62' ecl
systems the above analysis provides partitioning ratios, and
these ratios are shown in Table 6. In each case, both
excited states are produced with essentially equal
probability (to within a factor of three) upon annihilation
of electrogenerated M06C1143' and W6X8Y6'.
67
Figure 8. Steady-state emission and ecl spectrum in
CH3CN for (a) “51142“: —; Mo6c1142“ Mo6c1142'
/W61142- eCl, -"-"'; (b) M06C1142-/W6I142- eel, —;
. . 2- 2..
fit from the sum of a ratio of M06C114 and W6114
emission spectra, Peak maxima are normalized
to an arbritrary value.
68
p a
I ‘W' ( )
‘a
>_ -' '\
t: i‘.
U) I ‘~ i:
Z r 't '
:2 I “a
Z " I
H I It
2 ' n\
9 i ‘
2,3 , ... .
2 I ‘I
LU , “\‘v
I, ‘I '-.
â€III a
’I v ‘‘I uni“.
i‘ r: V")! "'4. 1
I I", ' I I I U I V 3"? 44-
5 50 650 750 850 950 IOSO
A/nm
Figure 8
EMISSION INTENSITY
69
550
ii
660
760 ' 850
A/nm
950
(b)
. ..
I dial--.
IOSO
70
.0mm scaposoe a“ *imvHH0002 mo coï¬uoscosc one so“ moaoï¬m madam cmpï¬oxm
.omm 0:0 «mm coï¬posos coosuoo Capes wcï¬coHuï¬uusm
.omm coï¬uoaob so“ deuce msfl>wso madam pouï¬oxo one
.0mm mcoï¬uomoe uoH moses wce>ï¬eo madam nouï¬oxo one
.mao>suooamou .mocHHmo
Haï¬xa 0:0 mcwweï¬soloosm one ucommsaoe > 0:0 x memes Eopmsm How soumsao coxaz
000.0 00.0 n 00.0 00.0- 00.0- sm.H0
000.0 00.0 H 00n0 00.0- 00.0- H0.H0
0s0.0 00.0 H 0s.0 00.0- 00.0- sm.sm
0000.0 00.0 H 0s.0 00.0- 00.0- em.H
0H00.0 00.0 n 00.0 s0 0- 00.0- 0.0
o0oamuoe *0: n .00: o>oxmmooe. b>o\emooe «0.x
use -N0swx0-
\lflvdHDQOI
msoaaosos Hos bonuses some: he deems» oases seasons
IN¢HHU®OI one mauadm wcï¬coï¬uï¬uusm huhocm .moouom moï¬bahn
0 oases
90130)
d
71
Other mixed cluster systems support this observation
(such as those shown in Table 5) but their partitioning
ratios can only indirectly be determined from the following
equation,
‘I’ecl = ¢e51 ‘f’Moe + 4’e52 ¢w8 (26)
where °ec1 is the overall ecl quantum yield of the mixed
cluster ecl reaction and ¢es and ¢es are the excited state
1 2
a: 2..)-
production efficiencies for M06C1142- and W6X8Y6 ,
respectively, in the mixed ecl reaction. For example, the
small steady-state emission quantum yield of W6C1142' (rt-e
(wscll42‘) = 0.015 vs. ae (Mo601142') = 0.13) requires that
W5C1142'* would have to be produced at least two times more
often than M05C1142'* to be reflected in the mixed-cluster
ecl spectrum. The ecl spectrum of M06C1142'/W6C1142'
identically matches the steady-state emission spectrum of
M06C1142- and ‘I’ecl = 0.05. One explanation is that the
W6C1142' is not populated upon annihilation. In this case
W6C1142-is simply acting as an electron acceptor. However
this is unlikely because M06C1143’/A+ (A+ = organic electron
acceptor) annihilation (discussed in Chapter IV) reactions
of similar potentials to that of the M06C1142'7W6C1142’
system exhibit ‘I’ecl = 0.10 (CH2C12, p = 0.1 M NBu4PF6 at
23‘). Alternatively, the ¢ecl’s of the Mo6C1143-/A+ (¢ec1 =
0.10) and M06C1143'/W6C114' (aecl = 0.05) suggest that half
of the electrochemical energy is being distributed to the
72
W6C1142' excited state. For this case, eq 26 reduces to
°ecl = ¢esl fMo, because fes, 4’W. (= 4 x 10'3) is small
compared to fesl moo ( = 5 x 10-2), and hence the
calculated yield from eq 26 is consistent with the
experimentally measured ecl yield of 0.05. Our assumption
of a 0.50:0.50 partitioning ratio for the Mo5C1142'/W6C1142'
system is experimentally supported by the results of the
M06C1142'/W6C18Br62' system. As discussed above, the
partitioning ratio directly calculated from the ecl spectrum
of this system is 0.50:0.50. Owing to the similarity of
W6C1142' and W6C18Br62' our assumption of a 0.50:0.50
partitioning ratio for M06Cli42'/W6C1142' ecl system seems
reasonable. For the case of the M06C1142'/W68r142' system,
the emission bands are too close in energy [AEB (M06C1142',
W6Br142') = 15 nm] to be discerned in ecl spectra. Analysis
of this system’s ecl quantum yields with eq 26 is also
consistent with equal partitioning of the electrochemical
excitation energy.
There is one very satisfying aspect of the results of
partitioning experiments, the free energy dependence of the
M06C1142" excited state production is independent of the
type of electron transfer acceptor. Chapter IV describes
the free energy dependence of the ecl quantum yield of the
M06C1143’ with aromatic amines. In these systems the
W6X8Y6' has been replaced by an acceptor in which the
excited state is energetically inaccessible. A plot of the
experimentally determined ecl quantum yields (calculated by
73
dividing the integrated M06C1142’ ecl intensity by the
number of equivalents of electrons transferred) of
M0501142'* for the M06C1143'7W6X8Y6' and Mo6011437h+
(aromatic amine radical cations) systems vs. free energy is
shown in Figure 9. The free energy dependence for
M06C1142'* production in mixed ecl experiments is nearly
identical to that observed for the ecl reaction of M06C1143'
with aromatic amine acceptors. These results demonstrate
that the eel pathway is independent of whether the
electrochemical excitation energy is distributed to one or
between two excited states in the annihilation reaction.
Thus, the ecl studies on the W6X8Y52' systems clearly
establishes that the electrochemical excitation energy is
essentially equally distributed to both cluster reactants.
This equal partitioning can be rationalized by using current
electron-transfer theories. Annihilation of M05C1143' and
W5X8Y6' is described in Scheme 3 where kd is the diffusional
rate constant, kesi and kes are the rate constants to
2
produce the excited state of M06C1142- and W6X8Y62'
k0 1 .. -
’ M050|142 + W6X3Y52
2- 2--
M060.“ + staYs
kd
M050|143'+ staYé —— Mosel-f-‘m waxava
k‘d kg: 2- 2-
Nkbch4 +-“%X€Qs
Scheme 3
respectively, and kgs is the rate to produce both ground
state molecules. Calculation of electron transfer rate to
74
Figure 9. Plot of log fee of M06C1142' in the mixed
cluster ecl reaction (0) and in the reaction of
- + o O
Mo6C1143 /A (0) vs. Aces 1n 012012 at 23 c (n = 0.1 M
NBu4PF6).
75
"'
"I'I'Ifl'-
"'
0.5
‘ A cSes /ev
Figure 9
O.|
-O.75 -
-|.75 "
-275-
-375-
.000 00.
-4.75 -
76
ground state and excited state products, as discussed in
detail in Chapter IV, show that kgs is slow compared to ke81
and Res; Therefore, the ground state electron-transfer
pathway does not mediate the partitioning ratios.
Furthermore, a kinetic analysis of the rate of appearance of
2—*
Mo6C114
and W6X8Y62'* shows the production of
2-* . 2-* . .
M C l tiv t W t k. k and
[ O6 114 ] re a e o [ 6x8Y6 ] is jus e951/ eszt
is not controlled by the rate of diffusion. Thus an
understanding of partitioning ratios follOws directly from
electron-transfer nal si f k nd k .
a y s 0 e51 a es:
The excited state product rates kes and ke82 are given
- 1
by eq 27 where the variables have previously been described
2(HAB)2 «3 1/2 (A + AG°)2
k = -—-- exp - (27)
98' h AkBT 4AkBT
(Chapter 1). The almost equal production of M06C1142'* and
W5X8Y62'* in the mixed cluster ecl reaction implies that
k
es s kes . The values of keslt and hence partitioning of
1 2
the excited state energy, depends on the electronic coupling
element, HA3, the reorganizational energy, A, and the
driving force, AG°, of reaction 23a and 23b. From Table 6,
we see that if pathways 23a and 23b are sufficiently
energetic, then both excited states are produced with
probabilities independent of AG° (AG° < -0.05). Moreover,
the tungsten and molybdenum cluster compounds are almost
identical in size (11 to 12.5 A) and structure, and
77
therefore from eq 3, A0 is relatively constant for all mixed
cluster ecl reactions (changes in A0 are 5 0.05 eV in the
cluster series). More importantly, in a given mixed cluster
ecl reaction, A0 is independent of which reactant, Mo6C1143'
or W6X8Y6', is converted to the excited state. Therefore
the ratio of kesl and ke82 is independent of AG° and A0 and
energy partitioning depends solely on the HA3 and A1.
The contributions of A1 and HAS to partitioning in the
eel chemistry of the M6x8Y62- clusters can be understood in
terms of the hexanuclear cluster's frontier molecular
orbitals. Figure 10 summarizes the results of theoretical
studies in recent years aimed at describing the electronic
structure of the M6X8Y62' ions. Extended Huckel78 and SCF-
Xia-SW79 calculations predict the HOMO and LUMO to be
primarily metal in character and to possess molecular
symmetries eg and 32g! respectively. These results are
consistent with spectroscopic studies, which suggest that
the luminescence of the M6X8Y62' ions originates from an
excited state localized on the metal core.80 Additionally,
magnetic measurements establish a diamagnetic ground state
for M6X8Y62' ion and the oxidized M6X8Y62' cluster ions
display an axial EPR signal, which can be attributed to
tetragonally distorted metal core resulting from the single-
electron occupancy of the e level.64 On the basis of these
9
spectroscopic and theoretical results the eel chemistry of
the Mo6C1142'/W6X8Y62' is described by the molecular orbital
representation depicted in Figure 11. The two excited state
78
Figure 10. Molecular orbital diagram for M6X8Y62"
ions.
79
unoccupied 0mi-
bonding metal-
bosed orbitals
—029
80
Figure 11. Molecular orbital description for electron
transfer between W6X8Y6' and M°6C1143 with the excited
state being produced from (a) the W6X8Y6' ion and (b)
the M06C1143' ion.
81
829
LL
1L e9
WaxeYs' M060l142' .
Figure 11
82
pathways of M6X8Y62' ecl are electronically distinct: (1)
production of electronically excited Mo6Cll42'* from
M05C1143' involves the transfer of an electron from an eg
orbital of M06C1143" to the eg orbital on the W6X8Y6' ion:
and conversely (ii) transfer of an electron from the azg
orbital of M06C1143' to the 329 orbital on the W6X8Y6' ion
produces electronically excited W6X8Y62'*. Because kes
depends only on HAB and Ai and if, as observed, kesl = kesz
then Ai and HAB must be either equal or fortuitously counter
balance each other for the reaction pathways described by
23a and 23b.
Obviously, HAB will be different for the two excited
state reaction pathways (i.e. M6X8Y6-—-> M5X8Y62'* vs.
M6X8Y63'———) M6X8Y62'*) if the respective orbital overlap
of the a2g orbitals is different then that of the eg
orbitals. This does not appear to be the case for the mixed
cluster ecl system. The eg (HOMO) and azg (LUMO) molecular
orbitals are constructed from linear combinations of dxy
orbitals of adjacent metal atoms: these molecular orbitals
are shown in Figure 12. Owing to the similar radical
distributions of these metal-based orbitals, the electronic
factors of the conversion of M06Cll43' or W6X8Y6- to the
excited state should be closely related.81
The assumption of comparable HAB'S for the two ecl
reaction pathways implies that Ai should be similar for
reaction 23a and 23b. More specifically, this implies
similar nuclear reorganizational energies for electron
83
Figure 12. Depiction of the e and aZg metal based
9
cluster orbitals.
84
Figure 12
85
exchange involving either the azg or the e orbitals. The
9
A1 energies associated with electron exchange reactions
involving the azg and eg orbitals can independently be
measured by electron-transfer quenching studies of the
2.4
M6x8Y6
with a series of organic electron donors and
acceptors. The quenching studies are schematically
represented below. In this reaction scheme electronically
excited M6X8Y62'
2-*
M6X§Y5
“i
D A
M x Y 3' M x Y 2' M x Y '
6 3 6 6 3 6 6 8 6
Scheme 4
donates an electron from azg orbital to acceptor molecule A
to produce the reduced radical A" and the oxidized cluster
M6X8Y6'. Alternatively, in the presence of donor molecules
D, an electron is transferred to the cluster's eg orbital
to produce oxidized D+ and reduced cluster M6X8Y63'. The A1
value measured for the conversion of M6X8Y62'* to M5X8Y63'
in electron transfer studies is related directly to the
conversion of M6X8Y63 to M6X8Y62'* in the mixed cluster ecl
reaction. Conversely, the cluster's contribution to the
measured inner sphere reorganizational energy of M6X8Y62'*/A
86
electron transfer is equivalent to A1 for the conversion of
M6X8Y6‘ to M6x8Y62-* in the mixed cluster ecl reaction. The
quenching rate constants for the reaction of M6X8Y62'* ions
with benzoquinone (BQ) and nitroaromatic (NA) acceptors
(reaction 28a), and aromatic amine (AA) donors (reaction
28b) in CH3CN at 23°C
2-* \ — -
A = BQ, NA
2-* r 3- +
M6X8Y6 + o —> M6X8Y6 + 0 (28b)
D = AA
are shown in Tables 7 and 8, respectively. Rates were
determined from Stern-Volmer analysis of the lifetime of
M6X8Y62'* by the procedure described previously. From these
data, Ai can be evaluated with eq 29
ZHAB K3 1/2
kBT 1n kq = -1/2AG° + —A/4 + kBT 1n
h kBT
(29)
which is obtained by rearranging the electron transfer rate
expression of eq 27. The quadratic term of eq 27 has not
been included in this rearranged rate expression owing to
its small contribution to the overall observed quenching
rate at low driving forces. Consequently, eq 29 predicts a
87
.mpoosossmsoa oawpomï¬a opdumlzoeopm Bose cocHEeouoc mpcdpmsoo owns wcï¬cocoso o
.AI\0¢0 N\Hm I â€Grow I AI
sceposcon mascocoso A
000 x 0.0 se.0+
s00 x 0.0 Ha.0+
00H x 0.0 00.0-
000 x 0.0 00.0-
000 x 0.0 00.0-
nH-o-Ixos $3 .0 0
mm
ococwsooucocrc
ocoucoooupï¬cï¬cio
one:ï¬sooucoo-oioeoHcowcnm.m
ocoucoooepHcHo-Q
ocosï¬soousoc-a-osoHnoes-m.m
hoaocosa
vHH00020N\HmH u .00 809% copsanoHao
case-coewooao one new oOHOw wcH>He0 0
Imvaaomoz
-000000s
IN¢HHO®OS
Imoaumma
ImvHHmGOE
ouocqoasq
Comm as 20000 :0 meadowsoouoom cousawamasm use mowudEOHdOHaez
he mHoHMSHU Imwwwxol Ho wceaocosa new wansamaoo ovum
s 00009
.m
88
.mecoaonsmdos scammï¬ao ouwpmimcmopm Scum coneseouoc mpcepmsoo 0900 wcï¬cocosc o
.AO +QVN\HW + HOIOW + IMAINï¬HHOQOSVN\HmH fl oU<
0 0
Sony oopeasoawo coï¬uowoe woe: c050 powmcseplcospo
00H x 0.0 HN.0+ ocwewaaSHoquvaep
s00 0 0.0 s0.0+ oaï¬oeseoe-o-Hsaeosee-.z.z
s00 x 0.0 0H.0+ oasuaeseosoeoesseos-0H
00H x 0.H «0.0+ ocwusï¬cpocmca
00H x 0.0 00.0+ osssnesaoeasosxobeosso
oH-oaisxos a>o\.oa noooaoao
0.00 as zo0so :0 moosaa oeuaaone
he nemamSHU Immwwxwi Ho mcwcocosa Mom mandamdoo 0900
w OHQQB
one new money mc0>000 0
-000000os .0
-000000os .0
-000000o= .0
-000000os .0
-000000os .0
meanness:
89
linear plot of kBTlnkq vs. AG° with slope of -0.5 and an
intercept equal to the bracketed term.38'82
Plots for the rates of acceptor and donor quenching
pathways are shown in Figures 13a and 13b. The linear
dependence of the rate constant on AG° and slopes of -0.49
2—*
and -0.51. for' M6X8Y6
/A. and. 2M6X8Y62'*/D systems,
respectively, agree well with theory. By assuming adiabatic
electron transfer (HAB = .022 eV), overall reorganizational
energies A(=Ao + Ai) of 1.11 eV and 1.01 eV are calculated
from the intercepts of Figures 13a and 13b, respectively.
By accounting for A0 = 0.86 eV (eq 3), Ai values of 0.25 eV
and 0.15 eV are obtained for quenching reactions 28a and
28b, respectively. The relevant parameters for these
electron transfer calculations are summarized in Table 9.
The inner sphere reorganizational energy for the
M6X8Y62'/A system is composed of the nuclear reorganization
associated with M6X8Y62'* / M6X8Y6' and A / A' conversions.
Similarly, the M6X8Y62'* / M6X8Y63' and D / 0+ conversions
compose the A1 for reaction 28a. Calculations by using
self-exchange rate constants measured by EPR line broadening
techniques have shown that Ai's associated with the electron
transfer reactions of aromatic amines and nitroaromatics are
< 0.05 eV. Therefore the calculated values of A1 directly
reflect the inner-sphere reorganizational energy of
M6x8Y62‘*/M6x826‘ and M5X8Y62'*/ M6x8Y63' conversions,
respectivelyu The above calculations are predicated on
adiabatic electron transfer. The calculated values of 0.15
90
Figure 13. Plot of kBTlnk vs. AG° in CH3CN at 23°C
q
(numbering as in Tables 7 and 8) for (a) M°6C1142-*
quenched by organic acceptors: (b) M06C1142'* quenched
by organic donors.
kT In kg
0.6
0.5
0.4
0.3
91
(a) -
1' 0.2 0.0
A G°/eV
Figure 13
kT In kg
0.6
0.5
0.4
0.3
92
(b)
4' 0.4
4
1' 0.2
A G°/eV
Figure 13
0.0
93
Table 9
Electron-Transfer Parameters Used in Calculating
‘1 Values
Electron Transfer
Parameter
r/Aa
wr/eVb
wp/eVb
Ao/eVc
HAB/eVd
Int/eVe
li/eV
for Quenching Reactions
'6X8Y62’*/A '6X8Y62-*/D
9.5 9.5
0.00 0.00
+0.04 -0.11
0.86 0.86
0.022 0.022
-0.49 -0.51
0.25 0.15
a r is the separation between reactants during
electron transfer assumed to be the sum of the
reactants' radii.
b wr and w calculated from equations
5 and 6.
*0 calculated from equation 3.
d Typical value of HAB for an adiabatic reaction.
e Intercept obtained from Figures 13a and 13b.
94
eV and 0.25 eV represent upper limits of *i' which decreases
with increasing nonadiabaticity. Thus the observation of
small and equal Ai's (o < Ai < 0.20 eV) for reactions 28a
and 28b is preserved even if the original assumption of
adiabatic electron transfer is inaccurate. These small
inner-sphere reorganizational energies most probably result
from the fact the azg and eg orbitals are delocalized over
the metal core and any reorganization is dissociated over
the 6 metal atoms of the octahedron.
Energy partitioning studies unequivocally demonstrate
that electrochemical excitation energy in the MGXBYG’
/M6X8Y63' annihilation reaction is channelled to either
.reactant with equal probability. Electronic structural
similarities of the HOMO and LUMO orbitals are manifested in
similar electronic coupling and inner-sphere
reorganizational energies for the two excited state
production pathways; and hence similar electron transfer
rates. With the demonstrated ability to produce M6X8Y62'*
from either M6X8Y6' or M6X8Y63’ coupled with the evaluation
of the important electron transfer parameters such as
electronic coupling and reorganizational energies of the
M6x8Y6'/M6x8Y62‘* and M6x8Y63'/M6x8Y62‘* conversions, the
ecl reactions of the M6X8Y6' and M6x8Y63' ions can now be
independently investigated. An issue of particular
importance is the dependence of production of M6X8Y62'* from
M6X8Y6' or M6x8Y63" on the free energy of the electron-
transfer reaction.
cum-1m Iv
IV. THE EFFECTS OF DRIVING FORCE AND LONG-DISTANCE
ELECTRON TRANSFER 0N CHENILUNINEBCENCE EPFICIENCIEB
95
A. Background
One of the principal themes that has emerged from
mechanistic considerations of ecl and cl reactions is that
the efficiency of excited-state production is related
intimately' to 'the energetics of electron 'transfer.
Extensive investigations of ecl and cl reactivity have
established two pathways for excited-state production.83"89
The first pathway is shown in Scheme 5 where the driving
force for the electron transfer reaction between A" and D+
A' + n+—>A* + n
Schemeis
is larger than the energy required to populate the emitting
excited state of A or D. This process of directly forming
the emitting excited state upon electron transfer is called
an energy-sufficient route or S-route mechanism.
Alternatively, as shown in Scheme 6 the driving force of the
electron-transfer reaction is not sufficient to populate the
emitting excited state. The production of the emitting
excited state involves population of a nonluminescent
1" + D+-———)3A* + D
3M + 3A*——>1A*
Scheme 6
96
intermediate triplet excited state and consequent
annihilation of two of these lower energy excited states to
produce one emitting state. This two step energy deficient
mechanism is widely referred to as the T-route.
For typical organic ecl or cl systems, the high energy
of the luminescent excited state (usually a singlet)
precludes S-route reactivity and electron transfer produces
a non-emissive triplet intermediate which undergoes
annihilation to yield the emitting singlet state. Because
triplet-triplet annihilation processes are inherently
inefficient,7b the excited-state production yields of
organic systems are generally limited to a few
percent.88a'9°-92 In contrast, luminescence from transition
metal complexes usually originates from the lowest energy
electronic excited state and therefore S-route reactivity
for inorganic species is governed by modest energies. In
recent years, ecl and cl from a variety of inorganic
compounds including M(bpy)32+ (M = Ru, Os, Cr; bpy = 2,2’-
)93,95
bipyridine and related species,95"98 Re(I) diimine
complexes,99
100,101
binuclear complexes possessing metal-metal
bonds, 102
phthalocyanines, square planar complexes of
pd(11)1°3 and Pt(II)81, and Ir(III)(2-phenylpyridine)3,
Tb(III) thenoyltriflouroacetonate, Pt(II)(8-quinolinolate)2,
[Cu(I)pyridine(I)]4 complexes104 have been reported. For
all of these systems, the energy released from the electron-
transfer reaction between oxidized and reduced forms of the
parent molecule (i.e., commonly called the annihilation
97
reaction) is sufficiently energetic to directly populate the
luminescent excited state. Nevertheless, despite this
predicted and in some cases experimentally verified S-route
behavior,1°5'106
measured efficiencies for excited-state
production are well below unity.75 The reasons for the low
yields of some of these systems are known. For example, an
ecl yield of <10“5 for the Pt2(H2P205)44' ion1°7 can most
certainly be attributed to the relatively short lifetime of
Pt2(HZP205)55' in aqueous solution.108 And low excited-
state yields of RuL32+* (L = polypyridyl) produced in the
reaction of RuL33+ with CoL3+ have been shown to result from
an electron-transfer pathway competitive to cl in which a
non-luminescent excited state of CoL32+ is populated.109
For the most part, however, a general understanding of the
low ecl and cl yields of inorganic systems has not been
achieved.
The energetics of the M06C1142' ecl permit the energy
dependence of ecl chemistry to be defined. The magnitudes
of the M06C114‘/2' and M06C1142'/3' reduction couples
[El/2(M06C114-/2") = +1.53 V vs. sce, E1/2(M06C1142'/3’) = -
1.56 V vs. sce in CH3CN] and the relatively low energy of
the M06C1142" excited state [Eem(M06C1142'*) = 1.9 V] have
allowed us to observe ecl from the annihilation of Mo6C114’
and M06C1143' with a variety of electroactive donors (e.g.,
nitroaromatic radical anions) and acceptors (e.g., aromatic
amine radical cations), respectively. By varying the
reduction potential of the electroactive donor or acceptor,
98
the ecl dynamics of M06C1142' ion can systematically be
investigated over a wide potential energy range.
This chapter describes our efforts to elucidate the
factors which control the efficiency of MGXBYGZ' ecl system.
The results of the dependence of ecl quantum yields on the
exergonicity of the electron-transfer reactions of M06C114'
with a series of nitroaromatic radical anions (NA'),
pyridinium radicals (P), and bipyridinium cations (BP+) and
the reaction of M05C1143' with aromatic amine cations (A+)
in acetonitrile and dichloromethane are presented. Analysis
of these yields in the context of current electron transfer
theories is discussed. ‘This analysis suggests that
efficient ecl is circumvented by long-distance electron-
transfer which can explain the low excited state yields for
chemiluminescent reactions of other inorganic complexes.
B. Results
Electrochemical and quenching data in CH3CN and CH2C12
are displayed in Tables 10 and 11 for aromatic amines, in
Tables 12 and 13 for the nitroaromaticsllo, in Tables 14 and
15 for pyridinium ions, and in Table 16 for bipyridinium
ions employed as electroactive. reagents in ecl studies.
These acceptors and donors meet two important criteria for
ecl free energy dependence studies in that their reduction
potentials, determined by cyclic voltammetry, span a wide
potential energy range and all compounds exhibit reversible
one-electron processes in CH3CN and CHZCIZ. values of the
99
.$w+ ohm mpaaï¬a gonna
. .0I0H ma mamas Eupcmsq How one mo HHEHH coï¬pooumn c
Immaaomo: Ho mpnoam>ï¬=qo Ho Hones: you couscous mcopoca mo moHos mo Honesz w
.+< so
.mwcoaonsmmoa
scammaao wudumlzcdopm scam cocaEHopoc mucdumcoo mums wcï¬nocmsd m .HAImMIN¢HHO®OSVN\Hm
a no
- €305- u .mmm:
.mom .m> madsoo ox+¢ one new massacopOQ coï¬uosoos m< 6
upon 83H:OEEmH>psndnpou s H.c mcï¬cï¬cpcoo mHï¬HpH:0poow :H wows ohms mpcmsopsmcos Hag
unweaapm m
u+< seas ImvHHomo: Mo coï¬pocos one now ow
>wnoco ooh“
.Oom « mm as mpmnoHco
0
.A+<V Hwoacdm coHpao wcwam Oï¬amsoum one spas Imvaaomoa Ho scuaomos one new mcamï¬z
savanna Ham 9 A<v mocflsa oï¬pmaoud Acheson an IN Havoc: Ho msacososq oosoomosï¬asq m
«IOH x m.m moH x m.m em.m- Ho.H+ massaAHsamnaosognuevmsnp .m
mica x H.m moa x o.w NN.NI m®.o+ ocHEmAH>HOulnvapu .m
¢IOH x ©.© boa x O.N wH.NI m©.o+ osï¬cHSHoalnlamnmeï¬olz.z .v
euofl x o.m boa x m.m oï¬.m- mm.o+ manuaï¬npoamnaflsnpmsuoH .m
a once v moH x H.H wo.mu Hm.o+ massassamnasosxosposnu .m
n mica v woa x o.H mo.NI mm.o+ msHNmanocan .H
Him HI: > >
waooo max o.ome< c.N\HM Adv whounooo<
encapsum Hum nu coma mouua< vuwd80h< no“ Adana
cums» amended Hem cud .dmundamnou ovum maï¬aososa .nudwugoaom sewaosoom
OH OHndB
100
.&w« one mpHEHH nonsm
.mIOH ma mama» Bandung How one we HHEHH coï¬pooumo c .+< so Invaaomoz mo mucoam>ï¬sqo
Ho Hones: non couscond mucuoca Ho moHoa mo nonssz w .mpsoEossmwoa scammï¬ao madam
Imcmopm Sony ocHEsopoc mucmpmcoo mums wcï¬nocoso H .HAIm\I waaomosvm\am I Ao +<vm\HmHI
u on U< u+< spas ImvHHomoz mo acapomon one so“ omcano swamco comm msmccmum o
.mom .m> wagsoo o\+< one how mamwpsopoa coï¬poscos m< c .oom a mm as opmaoHnoaoa
asï¬coaamampsndupop s H.o mnwcflmpcoo ocmnaoEOHOHcoHc ca moms ohms mucoeousmmos HH< o
.A+<v ascends conmo ocwem oHmeOHc one spas I vaaomos mo scavenge one how mcaoï¬a
Sapnwsq Hem n .A<v mosHEm oï¬pmsosm Acheson an Imvmaomoz Ho mcï¬cocosq cocoomocï¬ssq a
mIoH x H.m ¢OH x m.© vb.mI vo.a+ ocï¬smAazcmcaoeounIvvase .m
«IQH x m.m moa x e.H sm.mu so.o+ measaAHsHopuavmane .m
mIOH x >.® moa x m.m mm.mI mm.o+ ocflmmflcaoconmamcpszoa .w
mIOH x b.m moa x m.H HN.NI Hm.o+ ocï¬cHsHowIQIamnmeHch.z .m
I boa x H.H 0H.NI m¢.o+ osflumwcuoconm .m
I boa x >.H mH.NI mv.o+ ocHEmamconHc>xonpoEHQ .H
HIm HI: > >
mauve H.6H O.omwc< U.N\Hm A<v mucunmoom
encapsum Hem aw comb monï¬ad oquEOH< HOH ocean
macaw EdundsG Ham and .dmasdamnoo ovum mnwaoaosc .mï¬dï¬asoaom scuacsbom
HH manna
. IoH ma cHoH> Bananas
Hoo one no pHSHH coHpoopoQ a .ï¬oa + oum mpHEHH sousm .I<z no IvHHOmos no opcoac>ï¬sqo mo
homage you coosoosa occuona mo moHos mo monasz m .mpcosousowos coï¬ooï¬so ops thcdopm Eon“
cocï¬eaopoc mpsapmcoo opus mawnocosd H .HAI\ «zvm\am I AIN Ioï¬aomosvm\HmHI u co O< mI<z spas
Iwafloooz we :oHpomos one now owccno mmuoco oonm csmccmpw o .ooo .m> oaasoo I o<z one
you macï¬pcopoa soï¬poscon m< o .oom H mm pm ououoHcosoq anacossmaapscaspop 2 H.o we campcoo
oHanHcoeoom ca once onos mucosonsmmoa HH< 0
one How mcaoï¬m aswccsq Hem n
101
.AImzv muocoo couscos spas IvHHomo: Ho :oHpomos
.A<zv osococ Hmuusoc >3 Invaaomoz mo msflnocosq oosoooocï¬asq a
vIoH x m.o moH x m.H mm.mu mm.HI mumssxuouogpsaum .m
«nod x m.m moH x m.H mm.m- om.HI maoflsxnauogpscum .w
mnoa x m.H mOH x s.m mo.m- mH.HI muonsmnogpsaueuonosnoua .s
mIoH x v.m «OH x o.n mm.mu mo.HI massmcaauamnonpï¬aus .o
mIoH x e.H woa x w.m we.m- no.0- mcssmcsauamnOHaï¬ana .m
muoa x H.H moa x m.m se.m- Hm.ou mamuqmnopuï¬caouo .e
euoï¬ x m.m mos x H.H Hm.mu ms.o- mamuumnoppsafleua .m
mIoH x H.e woe x m.e mm.mu ms.ou ozoaï¬soouamnuaussnpmssuuo.m .m
a muoflv mOH x m.m HH.mI mm.ou maoassuowamnuc .H
Hlm HI. > >
mace. was o..mmc< v.«\Hm A<zc mnoaon
mowcsum Hem aw comb monogwsa oHHdEOH< can wowudEOHdOHaï¬z HOH Adana
cdoqw Esaudsd mom can .douuduouoo ovum wuï¬aoaoad .onduanoaom noduoscom
NH oHndB
102
.mIOn mn onmnn saucmsa now one no unenn conuooumo m
.&0H« ona onnenn nonnm .I<z no Ivnnomo: no mucoac>nsqo no noses: non cousconq occaOs
no ooHOE no nonssz m .oasosonsomos :Onooneo ouauoIscwoum son“ cognEnouoc opcdaocoo
oamn mannocosa .HAImo<zvN\Hm I AINnIvnHomozvN\HmmI u omw0< "Imz nuns IvHHomos «o
sodaomon on» new omcwso z noso oonn cnwc mam o .ooo .m> onasoo I\o<z one new madnucoaoa
acnaoscon w< c .Oow « mm as ouanOnsonoc sancossannasbmnuou 2 H.o msncnansoo ocanuos
IonoHnonc an opus onos mucosonsoaos HH< o .AI<zv onococ couscon sans IVHHomoa no acnaoaon
onu new wcnnon Esususq new 9 .Amzv mnococ Hanusoc so Imvnnomo: no mcnnocosq oocoooocnesqd
mIOH x m.v eon x n.m no.m- mo.nu mcmnnxusuonnnnum .m
mIon x o.v wen x o.m nw.mI Hm.nu mamnsxononnnaum .m
mIoH x s.m non x s.n oo.m- om.n- muonsxusnonnnzun .s
mIoH x m.m men x n.m mm.m- mm.nu mcmucmnonunaIeIononnoIn .o
muon x m.v none mm.m- sn.n- ocnzoonauamnonunzus .m
n-0n x N wen x m.n mm.m- mo.n- moszmnnancmnonnnzua .v
VIOn x m men x o.m mm.m- sm.ou wcoucmnonnnnnouo .m
I men x n.n mm.m- mw.o- ozoncmnonnncnouq .m
I men x v.n Hm.m- mm.o- oceansconcmnuaunnnnmanoIo.a .n
nlm HI: > >
mauve HG: O.omm0< v.N\HM A<2v mnOEOQ
monusaw How an coo: monounso Ouaann< can mouuaaonmonanz new sauna
anon» asasdsa Hem can .dmusuuouoo one: unwaososa .ouanunoaom acaacscom
nH wand?
.$NH « ono ownsna nonnm
.A no Ioaaooo: no oncoad>nsdo no noses: non ooosoone occuond no monos no nonasz c .opcos
Ionsodoa oOHomneo owmpoIncwopo Bonn oosnanopoc masonocoo onon mannososa w .on +AVN\Hm I
AIN\I¢HHO©OSVN\HMHI u am Go “A nuns Iwaaomos no cenpooon ocu n0m owcwso nwnogo oonm mnocgonm H
.ooo .m> oaasoo c\+m one no“ maonucouOQ acnposoon m< o .ouaam opacemosnonosamwxom c
.oom H mm as opwnoHnonoe Esncoeeoanpsndnpop 2 H.o wcncnopcoo oHnnpncOpooo on ocma onos
mngoaonswoos HH< o .Amv sen soncncnnnm ooosoon connooHoIoco suns IvHHomos no oOnpomon one
now mononh aspowsq Hem n .A+mv osOH Eonsncnnzm an Imvaaomos no monsooosq ooooomosnasq a
muon x o.n eon x m.n mm.m- mo.nu sonononnnonnooosuzuoonaouo .o
anon x m.n eon x m.n mo.m- oo.ou sonononnnonnuooouzuoonsouo .o
m «Ion x N; «on x 9m 35.. wool sonononnnonnooosuzunxoï¬ooonoouo .1.
«Ion x m.n eon x o.e oo.m- om.o- sonononnnonnuooouznnxoooooonoouo .m
muon x n.o won x o.o om.m- os.ou sonononnnonnooosuzuooonouo .m
mnon x e.m son x m.m om.mu oo.o- sonononnnonnNoooIzIooonoIe .n
Hum HI: > >
a now. won n..omc< on)â€. 2+5 onoooo
o.oono=nm now on oomo monom sonononnnn non owed: ononn
aausdaa How can d.oandaonoo opus mnemonosa .ondwunoaom neuaoscom
¢H Odnda
104
.ï¬NHn onw muHEHH nonnm .m no
Ivaaomoz mo mpcoam>nsqo no nomssc nod cousconq occuona no on08 no nonssz .mnsosonsoooa
scammnao ououoIhcdo w Bonn coonsnouoc mucmpmcoo ouan wannocosa m .HA \+QVM\HM I AIN IvHHU
ImozvN\HmHI u co co Mm suns IvHHomos no cenpomon onu non owcono nmnoco oonm cnocmopm n
.ooo .m> oadsoo o +m on» now onanpsonoa coHnoscon m< o .onnmm oumsqooneonosnnaxom c
.oom H mm as ouwnononoQ Esnsossmï¬susnmnpoa 2 H.o msncnmncoo oHnnnH:0poom an onus onos
wpoosonsmmos HH¢ o .Amv sen Esncncnnne couscon conuooHoIoco nuns IvHHomo: no coHnooon one
now mcaonm Esacwsq Ham 9 .A+mv omen Eoncncnnma an Inwaaumos no wannocosq oocoomocnssq o
omo.o non x m oo.mI sw.oI sonononnnonnooosIZInxoooooonooIo .o
omo.o non x o om.mI mm.OI sonononnnonnuoooIzInxoooooooooIo .m
I son x m sm.mI os.oI eonononnnonnoooeIzIooosoIo .m
I son x m mm.mI mo.oI sonononnnonsnaooIzIooonoIo .n
nIo nI: > >
:Hooe max m omwu< oN\HM many onouon
ooonooam now on oooo manom sonononnnn non ooaao ononn
asusdsa Hem can d.oacduonoo ovum mnnaouosa .oHdHaaoaom nowaosuom
mH OHQGB
105
.ï¬ma a onm mansna nonnm .+mm no IvHHomoz no oncoad>n=qo no noses: non couscona oneness
no ooHos no nonssz m .on\+nvN\Hm I AIN\I¢HHO®OSVN\HmH u ommoo m+mm nuns Ioaaooos no
scapomon ocp non owcmno awnoso oonn snoocanm o .oom .m> onesoo + +Nmm one now mamnpconoe
cenaoscon m< c .opaoo ononaooseonosamoxom o .oom H mm as onus monQOnOSHmmxos Enncossw
Iampsnannop S H.o wsncnoucoo oHnnuncouooo an opus onos mucosonsmooa Ham 2 .A+mmv cOH
asncncnnnenn couscon conuooHoIoco nuns IvHHomoz no cenuooon onu new mononn sapcosq new a
mIon x m.o os.mI sn.nI aonononnsonoIo.mInnooosIz.z .o
«Ion x n.n mo.mI om.oI sonononnnono
I_m.«InnoooI.z.zInnooosI.o.o .m
mIon x o.o om.mI ss.oI eonononnnono
I.m.mInnoonoI.z.2InnooosI.o.o .m
«Ion x mm.m mmn.mI oo.oI eonononnsono
I.m.«IooonnoooI.z.zInnoooaI.o.o .n
> >
“Hooo ooowwc. UN\HM oA+Nmmv onoson
noonooum non on oooo
manna aonononnnonm non ounce ononn souoooo non ooo ononaoonon oonuoooom
on onoon
106
ratio of anodic and cathodic current maxima ic/ia varied
from 0.95 to 1.05 and plots of anodic and cathodic peak
currents vs. (scan rate)1/2 were linear with a zero
intercept. Anodic to cathodic peak separations (AEP) were
almost identical to ferrocene in both solvents, thereby
establishing that any deviations of AE from the theoretical
P
value of 59 mV are due primarily to uncompensated cell
resistance. Rate constants for the quenching of M06C1142'
luminescence in CH3CN and CHZClZ (p = 0.1 M NBu4C104 or
NBu4PF6 at 23 °C) were deduced from classical Stern-Volmer
analysis of the emission intensity. All Stern-Volmer plots
were linear over a quencher concentration range of 1-100 mM
and kq's were calculated from Stern-Volmer constants with r0
* *
(M06C1142' ) = 130 ps in cn3cu and 10(M06C1142- ) = 160 us
in cnzc12 (p = 0.1 M NBu4ClO4 or NBu4PF6 at 23 °C).
Because BP2+/+, P+/° and NA°/' reduction potentials are
positive of the M06C1142'/3' couple and A+/° reduction
potentials are positive of the M06C114'/2' couple, the
electron-transfer reactions in equations 30 and 31 can be
clearly established by standard electrochemical
techniques.111 Chemiluminescence from the Mo601142'/donor
107
NA
-* +
NA M06C1142+ P (30a)
Mo6c114‘ + BP2+
NA
M06C1142' + P+ (30b)
BP2+
-*
M06C1142 + A (31a)
3- +
M06C1142 + A (31b)
and acceptor systems is observed only when the potential
applied to the working Pt electrode is stepped into the
oxidation-reduction waves of the electroactive species.
Tables 10-16 list the free energy changes and the ecl
quantum yields, ¢ecl' for reactions 30 and 31 in CH3CN and
CHZClz. Owing to the formation of (BP)Mo6C114 salts,
bipyridinium systems were studied only in CH3CN. Even in
this relatively high dielectric solvent, ion pairing was
observed and hence ecl measurements are suspect. For this
reason the ensuing discussion does not include the M06C1142'
/BP2+ systems. Ecl quantum efficiencies were determined by
dividing the number of einsteins emanating from the
electrode surface by the number of equivalents of electrons
used to generate the oxidant or reductant (i.e. the
integrated anodic or cathodic charge passed into solution).
As evidenced by the relatively small quenching rate
108
constants listed in Tables 10-15, acceptors and donors are
inefficient quenchers of Mosell42' luminescence and
therefore the measured ecl intensities are not attenuated by
the presence of acceptor or donor. Of course quenching of
M06C1142' luminescence by the oxidized or reduced donors is
downhill and should be extremely efficient. However, the
concentration profiles of electrogenerated intermediates do
not significantly overlap in an ecl step experiment and
hence M06Cll42'* should not be quenched by the
electrogenerated cluster or electroactive organic reactants.
Even when the production of the electroactive organic
reactant was doubled, significant quenching of the ecl was
not observed. Fer systems exhibiting ecl, the spectrum is
identical with the emission spectrum of M06C1142' in CH3CN
or CH2C12. The absence of acceptor or donor luminescence is
consistent with spectroscopic data, which reveals that
population of the lowest energy electronic excited state of
these compounds collected in Tables 10-15 is an
energetically unfavorable process.112
As described in Chapter I, ¢es is the parameter which
best describes the efficiency of the ecl reaction. Plots of
the ‘es vs. the free energy driving force of the excited
state reactions (AG°eS = AGogs + 2.0 V) for acceptors and
donors listed in Tables 10-15 are shown in Figure 14 and 15.
The standard free energy, AG of the excited-state
ES’
electron-transfer pathway (reactions 30a and 31a) was
calculated from AG°gS = AG°eS - AGES where AGES is the free
109
Figure 14. Plot of log ¢es vs. AG for the
es
electron-transfer annihilation reactions of the
Mo6c1143'/A+ (0), Mo601147p (A), and Mo6c114‘/NA" (a)
systems in acetonitrile. The numbering scheme is
defined in Tables 10, 12 and 14. The standard free
energy change for the excited-state reaction pathways
was evaluated as described in the text.
110
l.0
‘AGeS
Figure 14
111
Figure 15. Plot of log ¢es vs. AG for the
es
electron-transfer annihilation reaction of the
Mo6c1143'/A+ (o), Mo6c114'/p (A), and Mo6c114'/NA‘ (0)
systems in dichloromethane. The numbering scheme is
defined in Tables 11, 13, and 15. The standard free
energy change for the excited-state reaction pathway
was evaluated as described in the text.
112
- A Ges
Figure 15
113
energy' content of’ the M06C1142‘ excited state. over the
ground state andAG’gs is the standard free-energy change of
the ground-state reaction pathway. AGES can be estimated
from the energy of 0-0 transition (ED-o = 1.9 eV) with
corrections for entropic contributions (TAS = 0.1 V) .113
The excited state yields were calculated with ¢e = 0.19 for
uo6c1142‘ in CH3CN and ¢e = 0.18 for No6c1142' in cnzc12 at
23 °C.
C. DISCUSSION
Electronically excited M06C1142' ion is produced by the
simple electron-transfer reactions of the electronically
generated Mo6C114' and M05C1143’ ions with electroactive
donors and acceptors, respectively. This observation is
consistent with energy partitioning studies of Chapter III
which clearly demonstrated that the excited state can be
produced by either oxidized or reduced cluster. As
discussed in Chapter III the chemiluminescent reactivity of
the oxidized and reduced forms of Mo6C1142' can be
accomodated in terms of the hexanuclear clusters electronic
structure. The ecl chemistry of the M06C114' and M06C1143'
ions with donors and acceptors, respectively, can be
described by the molecular oribtal representation depicted
in Figure 16. For the M06C1143'/A+ series, the azg orbital
is occupied prior to annihilation and therefore, transfer of
an electron from the e orbital to the appropriate acceptor
9
level will yield electronically excited cluster. Directly
114
Figure 16. Molecular orbital description for
competitive electron transfer to give either ground- or
excited-state M06Cll43' by the reaction of (a)
M06C1142' with oxidized aromatic amines (A+) and (b)
M06C114' with reduced nitroaromatics (NA') or
pyridinium ions (P). Production of electronically-
excited acceptors and donors is an energetically
unfavorable proces.
115
m: 959“.
=| OmLFF =
o IE.
4« .24“. = o s:
+ +
ll 80' ll 80+
4 I «Boos. < Ixuoooz
.q I “Boos 2:
I:
=
+
00$
ouOllII
.o .. ...“...ooos.
=
J«
116
opposing this excited-state pathway is removal of the
electron from the 32g orbital to afford ground-state cluster
ion. In the case of the ‘H06C114'/NA' and P systems,
transfer of an electron from the donor level to the
cluster’s 329 orbital directly yields electronically excited
ion whereas exchange to the eg orbital brings the cluster
ion to its ground state.
It is evident from Figure 16 that cl is directly
competitive with the ground-state reaction. Specifically,
the yield for excited-state production, given by eq 15 in
Chapter I, is more conveniently expressed as,
k /k
¢es = es gs (32)
(Res/kgs) + 1
where Res and kgs are the rate constants for electron
transfer to produce excited-state and ground-state products,
respectively. The functional dependence of ¢es on the
driving force of the electron-transfer reaction is similar
for the different donors and acceptors in CH2C12 and CH3CN
(See Figures 14 and 15). Namely, the M05C1142’/acceptor and
donor systems show no ecl at low driving forces. Ecl is
observed at a threshold energy and at free energies negative
of this threshold, ¢esv rapidly increases. Finally, with
increasing exergonicity, the ¢es approaches a limiting value
well below unity.
The energy dependence of 4’es for ecl measurements in
CH2C12 and CH3CN is almost equivalent. However, ensuing
117
electron transfer analysis will focus on CH3CN because (i)
more organic acceptors and donors are soluble in CH3CN, (ii)
the effects of work terms are minimized in CH3CN, and (iii)
electron transfer models are most accurate for reactions in
high dielectric mediums.
Substitution of the asymptotically limiting values of
the excited-state yields for the Mb6C1143’/A+, Mo6Cll4T/P,
and M06C114'/NA' systems into eq 32 gives Res/kgs ratios of
0.15, 0.083, and 0.013, respectively. These values clearly
establish that the excited-state reaction pathway is
kinetically competitive with the ground-state reaction, even
though the latter is favored thermodynamically by 2.0 V.
This kinetic enhancement of the excited—state pathway may be
understood within the context of an electron-transfer model
for cl, first proposed by Marcus,58 in which electron
transfer to produce ground-state products is so exergonic
that it lies in the inverted region and therefore is
inhibited. In contrast, the modest exergonicity of the
exchange reaction to produce excited-state products occurs
ï¬x: the normal region. and. consequently electron ‘transfer
proceeds at relatively rapid rates. More quantitatively,
the ratio of the excited-state and ground-state rates is
given by12a
118
ks 1 1
2‘3RT 10g i = — (Ags-Aes) + - (AGgSO-AGGSO) +
k 4 2
gs
e 2 o 2
1 (AG ) (AG )
" ——gs - es (33)
4 *gs Aes
where Aes and "gs are the reorganizational energies for
excited-state and ground-state reactions. This rate
expression assumes that electron transfer is adiabatic and
occurs at a reaction distance of closest contact (i.e., rij
= ai + aj where ai and aj are the radii of the two reactants
and rij is the distance between their centers).
The reorganizational energy for electron transfer
comprises inner- and outer-sphere contributions (1 = "i +
10). The outer sphere reorganizational parameter is given
by eq 3 in Chapter I. The structural similarities of the
acceptors and donors listed in Tables 10-15 are manifested
in a nearly constant value of *0 = 0.86 z 0.05 eV for
reactions 30 and 31. The values for ai in eq 3 were
calculated with radii equivalent to the sphere of equal
volume, using the relation. a. = 1/2(d1d2d3)1/3 where di
represents the van der Waals diameter along the three
molecular axes. The inner-sphere reorganizational parameter
depends on differences in equilibrium bond lengths and
angles between reactants and products and thus is composed
of the inner-sphere contributions of the acceptor or donor
and cluster reactants. The contribution to the inner-sphere
119
reorganizational energy by acceptors and donors can be
determined with measured self-exchange rate constants by
using the Marcus self-exchange relation, k = Z exp(-A/4RT)
where Z = 1011 s'1 and k is the measured self exchange rate
constant. Ai's associated with the electron-transfer
reactions of these compounds are <0.05 eV.114 For cluster
reactants, the mixed cluster ecl and electron-transfer
quenching studies described in Chapter III, establish the
value of Xi for the conversion of either Mo6Cll4’ or
‘* is 0.2 eV.
M06C1143' to the excited state ion, M06C1142
These results suggest that the addition and removal of an
electron from the azg orbital 'requires almost. the same
reorganization of the nuclei as the addition and removal of
an electron from an eg orbital. Therefore ii for reactions
30 and 31 is ~0.2 eV and values of A (=11 + 10) for the
excited-state pathway or ground-state pathway of the
M06C1143'/A+, M06C114'/NA' or M06Cll4'/P electron transfer
reactions should be nearly equal and on the order of
magnitude of 1.10 s 0.10 eV at a reaction distance of
closest contact.
With the appropriate values of Aes and *gsr kes/kgs can
now be evaluated for reactions 30 and 31. For purposes of
comparison between the three series, we focus on the
electron-transfer reactions of Mo6C1143'/tris(4-
bromophenyl)amine+ (BPA; Aces. = -0.54 V, AG = -2.54 V),
gs
MoGCl14'/p-nitrobenzaldehyde' (NBA; AGes° = -0.48 V, AG
gs
-2.48 V), and MoGCl14â€/4-cyano-N-methylpyridinium (CMP:
120
AGes" = -0.30 V, AGgs° = -2.30 V) because these systems
exhibit asymptotically limiting values of 4’es for their
respective series. Using eq 33, we calculate values of
Raï¬/kgs = 5.6 x 105, 2.5 x103, and 8.3 x 105 for the
Mo6c1143‘/BPA+, Mo6c114'/CMP and Mo6c114’/NBA' systems,
respectively. These values are 105 - 108 greater than those
determined from the measured excited-state yields listed in
Tables 10 , 12 , 14 [keg/kgs (Mo6c1143'/BPA+) = o . 15 :
0.0072]. This striking discrepancy between the
theoretically predicted and experimentally measured rates of
the ground- and excited-state electron transfer is not
specific to M05C1142' ecl but, as mentioned above, is
typical of many inorganic transition metal complexes
displaying chemiluminescent reactivity75. Deviations from
inverted-region behavior have been attributed to a variety
of reasons including decomposition of the reactants before
annihilation and to a failure of the Marcus model in the
inverted region owing to the presence of competitive
reaction pathways such as H-atom transfer or the formation
of non-emissive excited-state products.85a'88a None of
these reasons, however, satisfactorily explain the results
of MoGCll42' ecl. For example, invoking a competitive
electron-transfer pathway to rationalize the low yields of
systems in this study is not reasonable because acceptors
and donors were judiciously chosen such that population of
their excited states is an energetically unfavorable
121
process. In addition, we can explicitly rule out deviations
from theoretical predictions resulting from the chemical
instability of the reactants on the basis of the
electrochemical reversibility of the cluster and
electroactive organic reactants. Thus differences in
calculated and observed rates of the M06C1142'/acceptor and
donor systems bear directly, by design, on the mechanistic
features of electron transfer at high exergonicities.
A crucial mechanistic feature of reactions 30 and 31
not explicitly accounted for by the simple Marcus
expressions used to derive eq 33 is that cl results from
bimolecular electron transfer which can occur over a range
115
of distances. A more general expression is given by eqs
9-14. From these expressions, the distance dependence of
the observed bimolecular rate constants of the excited-state
and ground-state electron transfer annihilation reactions of
the M06C1142'/acceptor and donor systems, is explicity
defined with a knowledge of *i HABo' and the driving force.
The distance dependence of the excited-state and ground-
state pathways is most easily illustrated by differentially
solving the integrals in eqs 10 and 11 between r and r + dr
for reaction separation distances from an arbitrarily large
value of 22 A to a closest contact distance of 9.5 A. For
this calculation the integrand in eq 11 can best be
approximated by kact = (4x N/1000)ge(r) k(r) rzsr (-1 8-1)
where 6r = 0.8 A.116 We initially focus on the results of
the Mo6C1142'/pyridinium series because calculations for the
122
electron-transfer annihilation reactions of this system are
simplified by the fact that ge(r) = 1. Figure 17 shows a
plot of the excited-state and ground-state differential rate
constants (kes,difn and kgs,difn' respectively) as a
function of r for the electron-transfer reactions of
M06Cll4' with CMP,
2- +
-M06Cll4 * + Nc-@N -CH3 (34)
Mo6c114" + NC-@N-CH3
2- +
M06Cll4 + Nc-@I -CH3 (35)
Equations 10 and 11 were evaluated by using an encounter
distance, a, of 9.5 A, a diffusion coefficient of 5x10"6 cm2
s' , RAB. = 200 cal and fl = 1.2 A’1 which are typical values
for electronic coupling terms of transition metals in
homogeneous solution. 18
The large value of kes,difn (=
kgs,difn x 105) at r = a clearly establishes that formation
of electronically excited M05Cll42' is preferred for
electron transfer occurring at a separation distance of
closest approach. In this regard, the results of eq 10 and
11 at r = a are consistent with those obtained using eq 33.
With increasing distance, however, kes,difn and kgs,difn
exhibit striking differences in their functional dependences
on r. This contrasting behavior of kes,difn and kgs,difn is
derived from opposing contributions of *o to the edectron-
transfer rate in the normal and inverted region._ As
described in Chapter I, the electron-transfer rate is
123
Figure 17. Distance dependence of the differential
bimolecular rate constant for the excited-state (es)
and ground state (gs) electron-transfer channels for
the reaction between M06C114' and one-electron reduced
4-cyano-N-methylpyridinium (CMP), calculated by solving
eqs 3, 9-14 between r and r + or using fl = 1.2 A"1 and
HA3. = 200 C31.
10
124
log ki
1
CS
1 n l
gs
9.5
12.0
14.5 17.0
r/A
Figure 17
19.5
22.0
125
related to the separation distance via the electronic
coupling element and outer-sphere reorganizational energy
(11 is independent of r). From eqs 3 and 14, an increase in
r causes 10 to increase and HA3 to decrease in magnitude.
For reactions in the normal region (i.e., -AG° < A), as is
reaction 34, an increase in 10 raises the activation barrier
to electron transfer and the rate becomes attenuated.
Couple this effect with an abatement in rate due to
decreasing HAB and, as observed in Figure 17, an increase in
r is accompanied by a steady diminution in kes,difn°
Conversely, although an exponential decrease of HAB with r
contributes to a decrease of the electron-transfer rate in
the inverted region (i.e., -AG° > A), it follows directly
from eq 13 that the increase of *0 causes an enhancement of
the electron transfer rate in the inverted region. These
opposing effects of HAB and 10 on the electron- transfer
rate is reflected in a maximum of k at r = 13 A. The
gs,difn
disparate behavior of differential excited-state and ground-
state rates with separation distance has interesting
implications for the chemiluminescent reactivity of the
M06C1142'/CMP+ system. As Figure 17 clearly illustrates,
the contribution of the ground-state pathway to the overall
rate comes from r > a, while most of the contribution for
excited-state production comes from r ~ 0. Thus the
appreciable 'values of kgs,difn. at r :> 0' suggests that
electron transfer to yield ground state products is
competitive with excited-state production.
126
The integral (or overall) excited-state (k and
es)
ground-state (kgs) rates are explicitly related to the
experimentally measured chemiluminescence yields by eq 32.
Accordingly, the reaction distance for electron transfer can
be determined by integrating eqs. 10 and 11 from r = m to a
value of r that yields a keg/k ratio commensurate with
gs
that calculated from the observed chemiluminescence yields.
For the M06C114'/CMP annihilation reaction, an observed
kes/kgs ratio of 0.033 yields an electron-transfer reaction
distance of 18 A. This result clearly implies that approach
of the electrogenerated reactants to a distance of closest
approach (a = 9.5 A) is impeded. Recent studies of outer
sphere electron transfer reactions of inorganic metal
complexes in nonaqueous solution have demonstrated that ion
pairing decreases electron transfer rates by increasing the
electron transfer distance (discussed in Chapter V).
Considering the relatively high ionic strengths used in our
ecl experiments, ion association between the supporting
electrolyte and charged reactants is likely, and in this
case, reaction at short distances will be inhibited.
Indeed, we have observed a marked dependence of the sol
intensity on the concentration of supporting electrolyte
(More detailed investigations aimed at assessing the
influence of solvent and ion association on ecl will be
reported in Chapter V). Thus our calculations indicate that
electron transfer between M06C114' and CMP occurs at
127
reasonably rapid rates over large separation distances to
produce M06C1142' ion.
The above analysis not only accounts for ¢es values of
less than unity, but it also qualitatively explains the
general dependence of ¢es on AG° for the acceptor and donor
systems depicted in Figure 14. Differential ground- and
excited-state rates obtained by numerically solving eqs. 10
and 11 for the remaining pyridinium systems are summarized
in Figure 18. We have also included in Figure 18
calculations performed for hypothetical pyridinium systems
with exergonicities below and near the ecl threshold free
energy; these results are indicated by dashed lines.
(Electron-transfer annihilation reactions between M06C114'
and pyridinium radicals with driving forces less than the
ecl threshold free energy, inferred from extrapolation of
the data shown in Figure 14, were not investigated owing to
our inability to find pyridinium reagents meeting the
necessary criteria required of electroactive reagents for
ecl studies.) Annihilation reactions possessing driving
forces below the ecl threshold energy exhibit comparable
excited— and ground-state differential rate constants at
distances near close contact. Consequently, the ground-
state electron transfer pathway is dominant over all r and,
therefore, kgs >> kes and ¢es << 1. As the driving force of
the annihilation reaction increases, electron transfer to
yield excited-state products becomes competitive with the
ground-state reaction pathway as evidenced by the
128
Figure 18. Distance dependence of the differential
bimolecular rate constant for the excited-state (es)
and ground-state (gs) electron transfer channels for
the reaction of Mo6C114' with: (a) a hypothetical one-
electron reduced pyridinium species with AG = -2.05
gs
V, AGes° = -0.05 eV; (b) a hypothetical one-electron
reduced pyridinium species with AG = -2.15 V; Aces.
gs
= -0.15 V: 4-cyano-N-benzylpyridinium: (d) 4—cyano-N-
methylpyridinium: (e) 4-carboethoxy-N-benzylpyridinium3
and (f) 4-carboethoxy-N-methylpyridinium. The standard
fee energy driving forces for (c)-(f) are given in
Table 14.
129
10
I09 ki
(a)
(...
GS
l 1 l n I L l
9.5
12.0 14.5 17.0 19.5 22.0
r / A
Figure 18
130
10
log ki
l e l
(b)
9.5
12.0
14.5 17.0
r/A
Figure 18
19.5
22.0
131
10
log k,
12.0 14.5 17.0 19.5
r /A
Figure 18
22.0
132
10
(d)
log k;
o 1 l 1 l 1 l A l
9.5 12.0 14.5 17.0 19.5 22.0
r/A
Figure 18
133
10
(9)
'09 ki
9.5 12.0 14.5 17.0 19.5 22.0
r / A
Figure 18
134
10
(1)
log ki
9.5 12.0 14.5 17.0 19.5 22.0
r/A
Figure 18
135
attenuation of kgs,difn and concomitant increase in kes,difn
over all r. At large exergonicities, electron exchange to
produce electronically excited cluster ion will predominate
and 4’es should be unity. That ¢es appears to approach an
asymptotically limiting value of less than unity (Figure 14)
suggests that the ground-state reaction rate is not
attenuated to the extent predicted by eq 13. We believe
part of this anomalous behavior is nested in the fact that
eq 13 is a classical expression and does not include nuclear
tunneling effects which can significantly enhance the rate
of electron transfer for reactions with large
exergonicitiesll7.
The effects of nuclear tunneling can be evaluated with
a semi-classical treatment of electron transfer. The
nuclear tunneling factor is defined as118
K
r = A (36)
(KB)Q
nA is the nuclear transmission coefficient given by
AG° 1 hu (AG°)2 hu 1/2
IcA = exp -— -— coth -— - —— + csch2 —
2
2kBT hv 2kBT 1 231'
AG° AG° by
+ — sinh- — sinh (37)
136
In the classical limit (i.e. hu << kBT), eq 37 reduces to,
eq 38,
(A+AG°)2
(NA)°° = exp - —— (33)
41kBT
where hu is the intramolecular tunneling frequency. By
assuming that the symmetrical metal-metal vibration, "a2g
(M06) = 120 cm'l, is the important vibrational frequency, a
nuclear tunneling factor of 1.4 x 102 is calculated for a
ground state reaction with a driving force of -3.0 V. This
is manifested in a direct enhancement of the ground state
rate by two orders of magnitude while the excited state rate
remains unaffected. However, this nuclear factor drops off
to 3.0 at a driving force of -2.2 V. These results show
that although nuclear tunneling does contribute to an
increased contribution of the ground state pathway to
annihilation at large driving forces, it is not large enough
to fully account for a leveling of the excited state quantum
yields at high driving forces. Parallel to this semi-
classical approach, a complete quantum mechanic treatment of
the ground state electron transfer reaction also increases
electron transfer rates in the inverted region, however,
this increase is relatively small owing to the low energy
vibrations of cluster ions. Therefore, not even a quantum
mechanical treatment can fully explain the leveling of ¢es
at large driving forces.
137
Calculations of the integral rates kes and kgs for the
M06C1142'/A and NA systems are similar to those of the
Mo6Cll42'/P+ system, however, the equilibrium pair
distribution function must be evaluated for the former
series. Parallel to the results described above, although
formation of excited-state M06C1142' is favored for electron
exchange between proximate reactants, the long-distance
electron transfer channel yielding ground-state products
contributes significantly to M06C1143'/A+ and Mo6C114'/NA'
annihilation. Solving eqs 10 and 11 with the experimentally
measured yields of the M06C1143'/A+ and M06C114'/NA' systems
listed in Tables 10 and 12 gives reaction separation
distances ranging from 18 A to 20 A.
Evaluation of eqs 10 and 11 for the M06C1142'/acceptor
and donor systems necessarily relies on estimates of RAB.
and. 5. It is satisfying that the general conclusions
derived from Figures 17 and 18 do not significantly depend
on these estimates. Specifically, the relative dependence
of the ground- and excited-state rates vary only marginally
over the rather large interval 0.8 A"1 < 5 < 1.8 A"1 which
includes any reasonable value of p for the reactions of the
type described in our ecl studies. Furthermore, RAB. is a
constant and therefore the excited- and ground-state
electron-transfer pathways exhibit a parallel dependence on
the electronic coupling element. This result is predicated
on our tacit assumption that HA3“ is similar for the ground-
and excited-state pathways. As discussed in Chapter III the
138
annihilation involves electrons residing in the metal-based
eg and 32g orbitals of the cluster core. Owing to the
similar radial distributions of these metal based orbitals,
the electronic factors of the excited- and ground-state
electron transfer pathways are more closely related than
those of any cl or ecl system studied to date.
Nevertheless, our assumption of similar values of HA3. for
the two reaction pathways, at best, is tenuous.
The electron-transfer chemistry of M06C114- and
M06C1143' ions can be described in terms of two competing
reaction channels: a highly exergonic electron-transfer
pathway yields ground-state products and less exergonic
exchange leads to the formation of electronically excited
MoGCll42- ion. The ratio of the electron-transfer rates for
these two channels, deduced from measurements of ecl yields,
is a powerful experimental quantity which has provided us
with the opportunity to address fundamental aspects of
electron trans fer in highly exergonic regions .
Specifically, the observation of ecl from M06C1142'/acceptor
and donor systems is evidence of the Marcus inverted region.
Moreover, the cl electron transfer chemistry, interpreted
within the context of the theoretical prediction of Marcus
and Siderslg'53 and of Brunschwig, Ehrenson, and Sutin115
that the electron transfer rate in the inverted region will
accelerate with increasing distance owing to an increase in
the solvent reorganizational parameter, suggests that
139
excited state production yields of less than unity result
from facile electron transfer over long distances.
CHAPTER V
v. ENVIRONMENTAL srrscrs 0N nsxsvsz'
CEEMILUMINESCENCE EEPICIENCIES
A. Background
The electron transfer formalism developed in Chapter IV
provides a framework in which to elucidate intrinsic
mechanistic details of cl or ecl reactions. The role of
distance of the electron-transfer annihilation reaction
bears directly on partitioning between the normal and
inverted regions and hence the ecl efficiency. Therefore
the influence of solvent and supporting electrolyte in
mediating distance will be important to the development of
efficient cl or ecl systems. For the case of M6X8Y62' ecl,
these influences will be augmented by the fact that the
electrogenerated reactants carry second coordination spheres
composed of supporting electrolyte and a tight solvent shell
owing to the high charges of the reactants.
The role of solvent in electron-transfer reactions has
recently come into question because of the inability of
current electron-transfer theories to rationalize the
absence of the inverted region and other anomalous behavior
of bimolecular electron-transfer events.118'122 Breakdown
of a dielectric continuum treatment of the solvent during
electron ‘transfer’ has come ‘under’ intense scrutiny.
Experimentally, this issue has been addressed by considering
the role of solvent and ionic strength for the following
reaction123
FeICpiz + COIdm9)3(BF)2+——> Fe(Cp)2+ + COIdmg)3(BF)2 (39)
(dmg = dimethylglyoxine, Cp = cyclopentadiene)
140
141
An observed decrease in rate of the electron transfer with
increasing ionic strength was attributed to increased ion-
pairing of the charged reactant. Moreover, for a given
ionic strength, plots of the log of the rate vs. (1/Dop"
1/Ds) , which is directly proportional to the outer-sphere
reorganizational energy, showed no obvious correlation.
These results suggested that the solvent and ionic strength
might affect the transition state structure (localized
structure about reactants different than the bulk solution)
or increase the distance of electron transfer.
The idea of a discrete solvent structure (i.e.,
nondielectric continuum model) about the transition state
during electron transfer has been treated by Ulstrup and co-
workers by using a nonlocal electrostatic theory.124 In
this approach, a nonlocal dielectric constant, which arises
from solvent not being subjected to full dielectric
polarization by an ionic field at molecular distances, is
used instead of the bulk dielectric constant of the solvent.
The nonlocal dielectric constant emphasizes the contribution
of electrostatic potentials in determining reaction
distances, and is calculated from eq 40 where
Deff(r) = 05/ [1 + «Os/Di)" 1) exp (-r/A)] (40>
Di is a "short-range" dielectric constant corresponding to
electronic and molecular' polarization and. A is a
142
"correlation length" . The dependence of the electron-
transfer rate of cobalt polypyridyl complexes in different
solvent mediums of varying ionic strength has been shown to
correlate, not with a continuum dielectric constant, but
with the effective dielectric constant described by eq 40.
A more radical deviation from conventional dielectric
solvent models has been proposed by Truong who considers
every reactant to possess a specific interaction with its
solvent shell . 125
At equilibrium conditions, the solvent
shell is treated as a charge transfer ligand which acquires
the partial charge of the reactant. Electron exchange to or
from the reactant must occur through the solvent shell,
which is now formulated as a hard sphere imbedded in a
dielectric continuum. Effectively, inclusion of the
immediate solvent shell as part of the overall charge-
transfer complex increases the reaction distance by at least
the diameter of two solvent molecules, and its
redistribution upon electron transfer contributes
significantly to the inner-sphere reorganizational energy.
A striking prediction of the theory is that *i is a linear
function of AG° (=El/2°x - El/Zred). The variation of
electron-transfer rate is predominated by *i and, therefore
log kobs will vary linearly over all driving forces at a
given pH and ionic strength. Although these predictions are
substantiated by data for several electron-transfer studies,
the theory is still under examination, and has yet gained
wide acceptance. Nevertheless, this work clearly
143
demonstrates that specific solvent and supporting
electrolyte structures about the transition state will
mediate simple electron exchange predictions.
Experimental parameters other than solvent or solute
interactions may also bear directly on ecl efficiencies.
Variations in the nature of the face-bridging and axial
ligands may alter electronic coupling and hence from eq 13
the partitioning between normal and ground-state electron
transfer pathways.24'126
Temperature can play an important
role in electron-transfer reactions127 and hence, ecl
chemistry by changing the population distribution between
the transition state of the normal and inverted regions.
Alternatively, the effects of ligand substitution and
temperature may not be so subtle. The stability of the
electrogenerated reactants in M6X8Y62' ecl may be enhanced
at low temperatures or preferred for clusters with specific
ligand coordination spheres. In these cases, ecl efficiency
will simply' be related. to temperature and ligand
substitution inasmuch as they affect the stability of the
electrogenerated reactant. This chapter describes efforts
to define the effect of solvent, solute, ligation sphere,
temperature and other experimental variables such as the eel
potential pulse sequence on the efficiency of M6X8Y62- ecl.
B. Solvent
The excellent solubility of (NBu4)2M6X8Y6 clusters in
many different solvents permits the investigation of
144
M6X8Y62' ecl under a variety of solvent conditions. The
electrochemical properties of the M06C1142' ion in several
nonaqueous solvents is shown in Table 17. The reversibility
of the M06C114'/2' redox couple is maintained throughout the
solvent series. However, as the coordinating ability of the
solvent decreases the M06C1142'/3" couple becomes quasi-
reversible as evidenced by the anodic and cathodic waves
becoming broader and more separated. The wide range of
dielectric constants of the solvents, shown in Table 18
allows the dependence of the sol efficiency on solvent to be
ascertained by analysis of the solvent reorganizational
energy. These solvents, which possess longitudinal
relaxation-times too fast for dynamical electron transfer
effects, have been shown to behave ideally in electron-
transfer studies.128'135
The excited state production quantum yields of the
annihilation reaction between M06C114- with Mo6C1143' or
neutral pyridinium radicals (reactions 17 and 30) in
different nonaqueous solvents are summarized in Table 19.
From straightforward electron-transfer analysis, a decrease
in A will be accompanied by a decreased activation barrier
in the normal region and increased barrier in the inverted
region. Consequently from eq 15, the ecl yield should
increase with decreasing A and hence (1/Dop- 1/Ds) .
Although the production yields depend significantly on
solvent, a plot of ¢es vs. (l/Dop- 1/Ds), which is directly
proportional to the outer sphere reorganizational energy,
H
KIQUIIbOJN
a As
vs
b As
vs
145
Table 17
Electrochemical Properties of I0601142‘ in
Various Nonaqueous Solvents
Solvent
Acetonitrile
Propionitrile
Butrylonitrile
Acetone
Benzonitrile
Dichloromethane
1,2-dichloroethane
E1/2(-/2-)a
+1.53
+1.49
+1.45
+1.46
+1.48
+1.38
+1.36
vs SCE V
reduction potential for the M05C114’/2‘ couple
SCE at 23°C.
reduction potential for the Mo6C1142‘/3‘ couple
SCE at 23°C.
c Reduction couples are quasi-reversible.
21,2(2-I3-)b
.56
.60
.64c
.73
.61
.70c
.73c
146
on .Qn no moone>
0®.H
ON.H
wH.v
ww.m
bo.v
mm.m
mm.m
Eo.=mo\wHOH N
noose: onoonc
.Qn
monocoemonnoo Eonm cosnEnouoc oEnp denumxmaon uco>HOm Hdsncsunwcoq a
.chH can wwmï¬ .mon EOnm moEnn cenpmxmaon onnom a
NHIOH x ®.H NHIOH x m.®
NHIOH x v.0 NHIOH x msH
NHIOH x w.m NHIOH x w.m
NHIOH x m.o NHIOH K m.m
NHIOH x m.o NHIOH x m.®
NHIOH x N.O NHIOH x m.m
nooo\qn sooo\:n
w.oH
wo.m
N.mm
b.0m
m.om
N.hm
n.5m
we
wo.N
mo.m
wm.m
vw.H
N®.H
nw.H
ow.H
non
oeeeeoononsonoIm.n .e
ocmsuosonoH20nQ .0
onnnuncoucom .m
osouoo< .v
onnnenoonnnesm .m
onnnoneonoooe .m
onnoeneooooe .H
moucsum Ham on com: wu=o>nom HO monunononm adouoham
wH oHndB
147
.m 0o Bonn mmnoso ascenumuncmwnoon onosqo nopso on» an Enop ennuooï¬ona 0
.oneesmz n.o u s .comm no ssnenonnsonneeoeIzIooonoIo\Ivnncoos
coosnon cenuowon nomossnu connooao one an INVHHUQOE no caonz Bandeau ououm copnoxm a
.oneesmz n.o u s .comm no Imennooos\Iennooos
coospon cennooon nonmcanu connoono one an INvHHUmoz no cnonn ï¬neness opmpm copnoxm d
om.o oo.o oo.o oeesooononeonoIm.n .e
wm.o om.o om.o ooeeeosoooneono .o
mm.o mn.o , on.o onnnoneooeom .o
oe.o noo.o eno.o oeooooe .e
se.o moo.o oeo.o onnnonoonnnesm .m
oe.o on.o ooo.o onnneneonoone .m
mm.o oso.o ooo.o onnnoneooooe .n
OAmG\HIQOG\Hv neooe ammoe
mucobaom msoosudsoz Hdno>om an
anewHOdom Hum new monosonowmum senaosconm oasum eoaï¬oxm
mH OHDdB
148
shows no obvious relation between dielectric and ecl
efficiency (Figure 19). The similarity of the plots for
reactions 17 and 30 demonstrates that a correlation between
¢es and (l/Dop- l/Ds) is not obscured by work term
contributions. These observations are consistent with
similar plots of several other inorganic bimolecular
electron transfer reactions in nonaqueous solutions.123
The data in Figure 19 clearly establish that electron
transfer is not occurring between reactants at a distance of
closest contact in a dielectric continuum. Alternatively,
the data suggests that the solvent is directly mediating the
electron-transfer distance. Inspection of Table 19 reveals
that efficiency of the eel reaction is loosely correlated
with the solvents’ dipole moment (Table 18). A tight second
solvent shell coordination sphere will inhibit the reactants
from approaching closest contact distance. As discussed in
Chapter IV, distances greater than closest contact cause the
contribution of the ground state electron-transfer pathway
to become more competitive with, and in some cases surpass
that of, the. excited state jpathway. Electron transfer
distances for the electron-transfer reaction in these
solvent systems can be ascertained from ¢es by the analysis
described in Chapter IV. As previously discussed,
calculations of the reaction distance are facilitated by the
absence of work terms. The parameters of the integrand in
eqs 10 and 11 and the reaction distances for the M06Cll4'/4-
cyano—N-methylpyridinium in various solvents are shown in
149
Figure 19. Plot of log ¢es vs. (1/Dop" l/Ds) for:
(a) M06C114"/M06Cl]_43 ; and (b) M06C114"/P in various
nonaqueous solvents, numbering as in Table 19.
'09 (P 953
150
( 1/Dop- I/Ds )
Figure 19
0.0
(a)
6 u I: 7
-0.5 '-
-1.0 - u 5
D 3
-1.5 "
D 4
-2.0 I I l
0.35 0.40 0.45 0.50
0.55
151
-0.2
(b)
6 a o 7
~06 -
v
3
9‘ P o 2
U)
.9
D 5
1.0 '-
D 3
D 1
D 4
1.4 I A I a I
0.35 0.40 0.45 0.50
( 1/Dop- 1/Ds)
Figure 19
0.55
152
Table 20. In each solvent the calculated r value is much
larger than the closest contact of 9.5 A. That the
difference between the observed reaction distance and
distance of closest contact, (Ar) listed in Table 21, in
each solvent varies despite the same supporting electrolyte,
implies ‘that the observed. reaction. distance is directly
related to the solvent. We see, with inspection of the
kinetic diameters of each solvent (Table 21), that ‘Ar is
approximately equal to the diameter of two solvent
molecules. Thus, these data suggest that electron transfer
occurs between reactants separated by two solvent molecules.
Our prediction that two solvent molecules mediate the
electron transfer distance of M6X8Y62‘ ecl is supported by a
qualitative comparison of the eel quantum yields of
reactions 17 and 30 in various solvents. Substitution of
the neutral pyridinium for M06C1143' is accompanied by an
increase in the ecl efficiency in all coordinating solvents.
The :more highly charged trianion should possess a ‘much
larger solvation shell than the neutral pyridinium donor and
hence electron transfer of the former system should occur
over longer distances.
The solvation studies show that solvent plays a
critical role in M6X8Y62' ecl. The origins of solvent
effects are not directly related to the solvents dielectric
but from the more subtle contributions of the solvent in
mediating the electron transfer distance. For the case of
M6X8Y62-/4-cyano-N-methylpyridinium ecl chemistry, the
153
nonossnp nonpooao Houoa acnnnmcsnn Hwononu Bonn consanpmo psosoao wsnaosoo
< am
< NN
¢ m.wH
< 5H
o<\n
.>H nopoono on com: cospos one nuns nonmasoamo oesoponc nommcmnp connooam o
.Imonaomoz no opepm connoxo one ooscono 0p oonon wsn>nno c
.>o «.0 Op Hesoo acnpsnnnpcoo
ononoo noccn a can: m oo Eon“ coannonoc nwnoco HosOHuancownoon Hopoe o
.oomm no oneesmz s n.o u s .o oee o moo soon ooeensoneo
>o on.o oo.o nn.o
>o om.o oo.o mn.o
>o em.o es.o on.o
>o sm.o oe.o on.o
so om.o oo.n oeo.o
o>oxoooe o>o\« o>o\es
aauducnnhmï¬hdaoaIZIossh0I¢\I¢H~O®Ol no :OHHOdom How one now
moondaown deï¬nedom one onoaoasndm nonossnh nonuoodm
ON OHDdB
0.0
0.0
0.0
0.0
0.0
n>o\nB
Swoo.o
bwoo.o
bwoo.o
bwoo.o
hwoo.o
d OQQN
manop xnos n
.ocOnnowon
enconuooam s
oannpncousom
oconoo<
oHnnpnconmusm
oHnnnncoHQonm
onnnenoonooe
154
Table 21
Solvent Diameters and IAr values for I060114-l
4-cyano-N-methylpyridinium Ecl Reactions
A r/Aa kd/Ab
Acetonitrile 7.5 4.0
Propionitrile 9 4.7
Butyronitrile 12.5 5.7
Acetone 11.5 5.5
Benzonitrile - 8.0
a Is the difference between the observed reaction distance
and closest contact.
b Kinetic diameter of the solvent calculated from a sphere
of equal volume using the relation kd = (d1d2d3)1/3
where di is the diameter along the three molecular axes.
155
proposed model consisting of two reactants separated by a
distance of approximately two solvent molecules, is
concordant with Truong’s approach in which the reactants
solvent shell is preserved in the transition state. A
localized solvent microstructure, which is typically not
addressed in electron transfer studies, most likely is
important not only in ecl reactions, but bimolecular
electron transfer reactions in general.
C. Ionic Strength Effects
Supporting electrolyte can play a role in determining
ecl efficiencies. For the M06Cll4'/Mo6Cll43' ecl reaction,
in CHZClz the concentration of supporting electrolyte
dramatically alters ¢es' (Figure 20). The effect of
supporting electrolyte concentration on ecl efficiency is
attenuated in CH3CN. This trend is consistent with a
decrease in work terms associated with M06C114'/M06C1143'
electron transfer as the medium's dielectric is increased.
Similar arguments have been used to previously interpret
ionic strength effects observed for other ecl systems in
which both reactants are charged.136'137
From simple charge considerations the anion of the
supporting electrolyte should not perturb the M06C1142' ecl
efficiency. This expectation is confirmed by the data shown
in Table 22, changing the anion from BF4', C104â€, AsFG',
PFG’, and CF3S03', while keeping the cation constant
(NBu4+), has little effect on the overall excited state
156
Figure 20. Plot of log ¢es vs ionic strength, p,
for the electron transfer of M06C114'/Mo6C1143' in
CH2C12 (a) and CHBCN (I) at 23°C.
0.5
0.0
-0.5
log (p 93
-1.0
-1.5
-2.0
157
D
11 A 21 I
1.1 x101/M
Figure 20
Supporting
Electrolyte
NBU4PF6
NBu4C104
NBU4BF4
NBu4AsF6
NBU4CF3SO3
NBU4PF6
NBU4PF6
NBu4PF6
Supporting Electrolyte Studies
for the l06C1142' Ecl Reaction
conc./N 433(CH2012)a
0.10 0.50
0.10 0.33
0.10 0.50
0.10 0.33
0.10 0.47
0.30 1.00
0.02 0.55
0.04 0.28
a Excited state quantum
of M060114'/M0601143‘
b Excited state quantum
of M06C114'/M06C1143'
158
Table 22
0 e8(CH3CN)b
0.065
0.025
0.074
0.060
yield for the annihilation reaction
in
CH2012 at 23°C.
yield for the annihilation reaction
in
CH3CN at 23°C.
159
yield. The decreased durability of M06Cll42' ecl and
varying day to day ecl yields were observed when C104- was
the supporting electrolyte anion. These observations are
explained by the more nucleophilic behavior of C104" and its
tendency to promote interfering, but ill-defined side
reactions.137b'138
Much larger perturbations of the ecl yield by the
supporting electrolyte should be observed with changes of
the cation. If ion-pairing is important in determining
reaction distance, then variation of the cation’ s size
should significantly affect the rate of excited state
123,126
production. A dramatic decrease in
oes in CH3CN is
observed along the series NBu4+ > NEt4+ B NMe4+ (Table 23).
These data conflict with ion-pairing considerations which
predict an increase in ‘i’es with decreasing size of the
cation. Indeed, the behavior of the eel yield with the
supporting electrolyte appears to be much simpler in origin.
The decrease in ¢es is related to the decreased solubility
of M06C1142' in CH3CN containing NEt4+ and NMe4+. We expect
the trianion to be even less soluble than the dianion.
Indeed, after scanning the reduction wave of M06Cll42' in
CH3CN solution containing NMe4+ or NEt4+ large anodic peaks
due to cluster absorbed on the electrode are observed.
Moreover, decreased ecl yields are not observed for
annihilation reactions between M06C114' and reduced
pyridinium (Table 23) . These results clearly suggest that
the decrease in ecl efficiencies is not due to larger
160
Table 23
Dependence of Ecl Efficiencies on
Supporting Electrolyte Cation
Supporting
Electrolyte conc/I Ies5a 4936b
NBu4PF6 0.10 0.065 0.10
NEt4PF6 0.10 0.006 0.16
NMe4PF6 0.10 0.008 0.22
a Excited state quantum yield for the annihilation
reaction of M060114‘/M0601143‘ in CH3CN at 23°C.
b Excited state quantum yield for the annihilation
reaction of M06C114'/4-amido-N-methylpyridinium
in CH3CN at 23°C.
161
reaction distances owing to ion-pairing, but to depleted
concentration of the trianion resulting from the formation
of insoluble NMe4+ and NEt4+ salts in the diffusion layer.
B. Ligand Coordination Sphere Effects
Study of the effect of various ligands on the
production efficiency of the excited state in the Mo6X8Y6'
/Mo6X8Y63' annihilation reaction is facilitated by the
ability to synthesize virtually any axially or face-bridging
substituted cluster complex. The electrochemical properties
of various Mo6X8Y62' clusters in CH2C12 was presented
previously in Chapter III. The excited state production
efficiency of the ecl reactions of these clusters is shown
in Table 24. Inspection of Table 24 shows the ecl
efficiency of these M°6C18C1nx6-n (X = Br,I; n = 0-6)
efficiency depends dramatically on the axial substitutent:
¢es is greatly diminished when bromide or iodide is
substituted in the axial positions (Figure 21) . For the
Mo6C112122- and Mo6C18162' ions, the extremely low values of
¢es are most likely attributed to degradation of cluster
oxidation as evidenced by the multiple oxidation waves in
the cyclic voltammogram. Hewever, this explanation is not
valid for the bromide substituted Mo6C1142'
(Mo6C18ClnBr6_n2') cluster owing to its chemically and
electrochemical reversibility. If the bromide-substituted
trianion is removed from the annihilation reaction and
162
Table 24
Excited State Production Efficiencies for
I06013014x6-n in CR2C12
Cluster 48813 Ieszb
M0601142’ 0.50 0.50
M060113Br2' 0.005 -
M06C112Br22‘ 0.003 0.18
M06C1llBr32‘ 0.002 -
M06C13Br62' 0.001 0.027
Mo60112122- - -
M06C18I62- - -
a Excited state quantum yield for the annihilation reaction
of Mo6C1801nX5-n'/M06C18C1nX6_n3' in CH2C12 at 23°C.
b Excited state quantum yield for the annihilation reaction
of M06C18ClnX6-n/4-carboethoxy-N-methylpyridinium in
CH2C12 at 23°C.
163
Figure 21. Plot of log ¢es vs. no. of bromides
substituted in the axial position for the
M06C18C1n8r6_n2' ecl reaction in CH2C12 at 23'C (p =
0.10 M NBu4PF6).
164
~0.5 -
U)
03 -1.5 I-
9-
. U)
2
I
-2.5 r I
I
-3.5 ‘ 1 . 1 . l 4
0 1 2 3 4
2-
[MOGCIBICInBr6,n
Figure 21
165
replaced by a neutral pyridinium radical (reaction 41) the
eel efficiency increases
2-* +
substantially (Table 24). These data clearly show that the
diminished ecl yield is associated with the M06C18C1n8r6_n3'
ion. Addition of bromide to the solutions of
M06C18ClnBr6_n2' significantly improved the reversibility of
the Mo6C18ClnBr6_n2'/3' couple (Figure 22). These results
strongly suggest that the low ¢es yields of Mo6C18ClnBr6_n2'
ecl chemistry is due to bromide dissociation from the
cluster core upon its one-electron reduction. The decrease
cannot be ascribed to low ecl yields of a coordinatively
unsaturated intermediate because the M06Cll3' was prepared
and the efficiency for production of the excited state in
the eel reaction (eq 42) is almost equal to that of
Mo601142' (4es(Mo6Cll3') = 0.40).
2- —* -
This result demonstrates that the Br" directly interferes
with the cluster's ecl chemistry. A mechanism consistent
with these observations is shown in Figure 23. Reduction of
166
Figure 22. Cyclic voltammogram (CH2C12 solution at
23°C, 0.1 M NBu4PF6) for Mo6c188r62‘ (3 mM) ----: MoGCl
8Br62- (3 mM) and NBu4Br ( 1 mM) ————.
167
I I I I I l I l I l
-l.90 -l.7O -l.50 -l.30 -|.|
V vs. SCE
Figure 22
168
Figure 23. Mechanism for Br' interference of
Mo6C18ClnBr6_n'/Mo6Cl8ClnBr6_n3' annihilation reaction.
169
-...xsoeos. + - x
N 5x868:
r
..m
. 8 exeoeos
I o
- Nexeosos.
’
1‘
+
no 9:9“.
-mesosos.
- «£5682
.. 8x568:
IQ
- Nexeosos.
5
II
x
-mexscoos +
h.
- Nexsoeos.
170
Mo6C18ClnBr6-n2- causes prompt dissociation of bromide from
the cluster core to produce unsaturated reduced cluster,
Mo6C18ClmBr5_m2' (m = 0-5) and free bromide. Ensuing
oxidation of the bromide by M06C18ClnBr6_n' yields the
dianion and radical halide, which in turn can react with the
unsaturated reduced cluster to produce Br' and
M06C18ClmBr5_m'. Subsequent addition of Br" to the
coordinative unsaturated cluster yields starting dianion.
Thus, the eel mechanism (i.e. ' M06C18C1nBr6_n'
/Mo6C18C1nBr6_n3' annihilation) is efficiently circumvented.
The crucial step of this mechanism, namely oxidation of
bromide by the one-electron 'oxidized cluster has been
independently verified. Figure 24 shows the decrease of the
Mo6C18ClnBr6_n2' luminescence during bulk electrolysis to
produce Mo6C18ClnBr6_n'. Addition of Br' to freshly
oxidized solutions leads to virtually complete recovery of
the luminescence intensity. This mechanism does not appear
to be important for all chloride clusters because the ¢es
values of Mo6Cll4'/Mo6C1143 and. M06C114'/P annihilation
reactions of equal driving forces are nearly identical.
E. Temperature Effects
The effect of temperature on Mo6C1142 ecl in
dichloromethane and acetone is shown in Figure 25. The ecl
efficiency increases substantially' as the ‘temperature is
lowered. Similar observations of other ecl systems have
been. attributed. to increased. stability' of 'the
171
Figure 24. Decrease in Mo6C188r62- luminescence
during bulk electrolysis (————): Increase in
luminescence after adding Br" (----).
172
em 959“.
mozoumm
can own ON.
. . vaIi .
~----------
AilSNaiNI EALLV'EIH
173
Figure 25. Plot of log ¢es vs. temperature for the
annihilation of Mo6C114'/Mo6C1143' in dichloromethane
(————): Mo6Cl14'/4-carboethoxy-N-methylpyridinium 1J1
dichloromethane ("°'); Mo6Cll4'/Mo6Cll43’ (----) in
acetone at p = 0.10 NBu4PF6.
174
O I.
“.5 "
-l.0 r-
3
S
U’
9.
-l.5 1'
-2.0 -
I l I t l
-90 -60 -30 0 30
T/C°
Figure 25
175
electrogenerated reactants at low temperatures.75'139 This
may be the case here, but our experience is that with the
appropriately chosen pulse sequence (vida infra) the
electrogenerated reactants are stable on the time scale of
the ecl experiment.
A more intriguing explanation for the temperature
dependence lies in the analysis of the ecl energetics. The
driving force for the excited state pathway (AGes = 1.1 eV)
is equal in magnitude to the total reorganizational energy
of the electron transfer reaction. From eq 2, the excited
state pathway is activationless (ignoring work terms) and
therefore the population of the activated complex should be
temperature independent. However, this is not the case for
the inverted pathway. Reaction to ground state must
surmount a non-zero activation barrier and hence the rate
will diminish with decreasing temperature. Thus, the
chemiluminescence electron-transfer pathway will
increasingly dominate as the temperature is lowered.
Further support of this hypothesis comes from the attenuated
temperature dependencies of M06C1142-/A. and M06C1142'/P+
ecl. For these reactions, the activation barrier to excited
states is non-zero. The temperature dependence of normal
and inverted pathways in these systems will be more similar
and therefore the ecl efficiency will, as observed, exhibit
a smaller temperature effect. This work provides support of
Faulkner and Kim’s earlier contention that temperature
effects in ecl could be explained by differences in
176
activation barriers between normal and inverted region
pathways.137a
F. Potential Step Program Effects
The ecl efficiency of the M06C114'/M06C1143' system
depends on the potential step sequence. .An. anodic to
cathodic potential pulse sequence (initial production of
M06C114-) produces ecl 6 times the intensity of a cathodic
to anodic pulse program (initial production of M06C1143').
This observation supports the fact that Mo6Cll43' is
chemically unstable over long periods of time. For the
latter pulse sequence Mo6Cll43' is diffusing away from the
electrode during the cathodic pulse and then diffuses back
to the electrode during the anodic pulse. The longest
residency time of the Mo6C1143' generated for a 100 msec
pulse sequence in the diffusion layer is approximately 200
msec. On the other hand, the residency time of Mo6C1143'
during' the former' pulse sequence is less than. 10 msec
(diffusion away from the electrode). That the eel yields of
experiments possessing pulse sequences, which result in long
Mo6C1143' residency times, are low clearly demonstrates that
the trianion undergoes decomposition reactions. We
approximate the decay reaction to occur with a half life on
the order of 100 ms.
Two empirical observations relating ecl efficiencies
with potential step sequences are shown in Figures 26 and
27. A plot of pulse frequency vs. log ¢es (Figure 26) shows
177
Figure 26. Plot of log 4’es vs. pulse frequency in
CH3CN for the Mo60114'/Mo601143’ (o) and Mo6011474-
amido-N-methylpyridinium (a) at 23°C (p = 0.1 M
NBu4PF6).
178
cm
on
N: \ .ooi moan.
ov
om osom
on
ON 0—
d
d 4 -
EN-
ON-
NA-
sad) 00)
179
Figure 27. Plot of the log ¢es vs. no. of pulses in
CH3CN for Mo6c114'/Mo6c1143 (e) and Mo6C114'/4-amido-
N-methyl pyridinium (O) at 23°C (p = 0.1 NBu4PF6).
180
0..»
CM
R 9.5:
woman. co 52
ON
d
181
that ecl efficiency is fairly constant at long pulse
frequencies, but as the pulse frequency decreases the ecl
efficiency diminishes. This effect is attenuated for the
M06C1142'/pyridinium ecl reaction. Figure 27 shows the
dependence of the ecl efficiency on the number of
consecutive pulses. For Mo6Cll4"/Mo6Cll43' ecl, the
efficiency steadily diminishes as the number of pulses
increases to a value ten times less than the maximum. Once
again, the effect is attenuated for the Mo6C1142'/pyridinium
ecl system. Both studies suggest side decomposition
reactions of M06C1143' over long times. Substitution of the
reduced pyridinium for Mo6C1143' eliminates this problem,
and the eel efficiency is restored.
CHAPTER VI
VI. FINAL REMARKS
Most electron-transfer reactions only allow the study
of rates in either the normal or inverted region.
Chemiluminescence is unique because normal and inverted
region electron-transfer pathways are competitive and
changes in reaction conditions are directly reflected in
differences in ground and excited-state rates. The M6X8Y62-
systems have provided insight into the parameters which
control ground (inverted region) and excited state (normal
region) reaction pathways and hence, the efficiency of the
ecl process. An important result of the work described
herein is that the efficiency of ecl is circumvented by
long-distance electron transfer. This observation implies
that the most efficient cl or ecl systems will be those
possessing annihilation reactions between redox centers
chemically linked over short separation distances. An
obvious avenue of future exploration is to construct
electrode microstructures in which the cluster and its redox
partner are covalently or electrostatically bound at fixed
distances. The versatile substitution chemistry of M6X8Y62'
clusters permits straightforward covalent attachment of a
variety of redox active groups (e.g. pyridinium) in the
cluster’s axial positions. In these systems, ecl can be
established between the reduced substituent in the axial
position and the oxidized metal core. Of course, the
success of this electron-transfer chemistry assumes that the
hole and electron are localized on the respective cluster
core and axial substituent (i.e. weakly coupled system).
182
183
Spectroscopic properties of Mo6X8Y5L clusters indicate that
the excited state and oxidation-reduction properties are not
strongly coupled with ligands in axial positions. Thus the
assumption of localized state appears to be valid and ecl
experiments between an axial substituent and cluster core
appear to be feasible. Alternatively, bridging ligands such
as pyrazine can be used to form dicluster units immobilized
to metalloxide or activated graphite electrode surfaces by
using well established methods.“’0 Because the cluster is
charged, the ions can be incorporated into polyelectrolyte
electrode films by simple ion-exchange methods.1‘"°'1‘1'2'143
For’ instance, .M6X8Y62' clusters bound. in Ibipyridium. and
pyridinium films have recently been prepared. The
observation of an oxidation wave of the cluster and
reduction wave of the polymer has resulted in the
observation of weak chemiluminescence from annihilation
reactions analogous to those found in homogeneous solution.
This inherently low ecl efficiency is caused by negligible
exergonic excited state driving forces, and thus different
polymer environments providing driving forces well above the
ecl threshold are currently being sought. The results from
these linked ecl systems will allow the electron-transfer
distances in ecl reactions to be precisely defined and
mediating factors such as solvent and solute interactions to
be quantitatively investigated.
The criteria of high electrical efficiencies over long
lifetimes has deterred successful development of ecl
184
59-63
systems. A renewed interest in developing practical
applications of ecl systems, has been rekindled with many of
the discoveries presented in this thesis. Not only will the
construction of electrode ‘microstructures improve
efficiencies by maximizing the fundamental factors crucial
to efficient excited state production, but attachment of the
cluster ions to the electrode surface should improve
luminosity and durability of cluster based ecl devices while
structures will also minimize the adverse solvent and solute
effects. In this regard, these systems potentially
constitute the fundamental building blocks of solid state
electroluminescent devices.
The research described herein establishes the utility,
at a quantitative level, of chemiluminescence in determining
electron transfer mechanism. The importance of electron
transfer distance and factors mediating this distance (e.g.
solvent, solute) in governing the electron—transfer pathway
are results which are not only prerequisite for
chemiluminescence reactions but pertain to all bimolecular
electron-transfer reactions as well. The existence of
numerous redox active, luminescent molecules coupled with
the dynamic interplay between electron-transfer theory and
experiment ensures that chemiluminescence will continue to
provide ‘valuable insight into the :mechanism. of electron
transfer reactions.
VII . REFERENCES
10.
REFERENCES
Kalyanasundaram, K.: Gratzel, M.: Pelizzetti, E.
CQQLQ. Chem, Rev. 1986, 62, 57-125.
Gratzel, M. (Ed.), "Energy Resources Through
Photochemistry and Catalysis", Academic Press: New
York, 1983.
Fendler, J.H. J, Ehyg, thm. 1985, 82, 2730-2740.
Masters, C., "Homogeneous Transition-Metal Catalysis",
Chapman and Hall: New York, 1981.
Sutterfield, G.N. "Heterogeneous Catalysis in
Practice", McGraw-Hill: New York, 1980.
Addision, A.W.: Cullen, W.R.; Dalphin, D.: James, B.R.
(Eds.), "Biological Aspects of Inorganic Chemistry",
Wiley-Interscience: New York, 1977.
Govidjee (Eds.), "Photosynthesis: Energy Conversion by
Plants and Bacteria", Academic Press: New York, 1982,
Vol. 1.
Isied, s.s. Frog. Inorg. one . 1984, 32, 443-517.
(a) Marcus, R.A.: Sutin, N. c ' ' t h s
Act; 1985, 811, 265-322. (b) Newton, M.D.; Sutin, N.
Ann, Rev. Phys, Chem. 1984, 35, 437- 480.
McLendon, G., Guarr, T.: McGuire, M.; Simola, K.:
Strauch, S.; Taylor, K. Coo d. Cem. v. 1985, 64,
113-124.
185
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
186
Meyer. ToJ- Eregi_lngrgi_gheml 1983. 12. 389-440-
(a) Marcus, R.A- Ann1_8e!1_£hxsl_£h§m. 1964. 15. 155-
196. (b) Marcus, R.A. M. 1956, 23,, 966-
978.
Hush, N.S. Trans. Fagaaay Soc. 1961, 51, 557-580.
Hush, N.S. EIQQI_IDQIQI_Qh§E- 1967, a, 391-444.
Hopfield. J- I1.2r221_natili_Asadl_§sil 1974. 11. 3640-
3544.
Chance, 3.: DeVault, D.C.; Frauenfelder, H.:
Schrieffer, J.R.: Sutin, N., Eds., "Tunneling in
Biological Systems", Academic Press: New York, 1979.
Marcus, R.A. in "Third International Symposium on
Oxidases", King, T.E.: Mason, H.S.; Morrison, M., Eds.,
July 1-4, 1979.
Brunschwig, B. 8.; Logan, J.: Newton, M. D.; Sutin, N.
Q, Am. Chem. SOC. 1980, 19;, 5798-5809.
Sutin, N. Age, Cham, Res. 1982, 15, 275-282.
Siders, P.: Marcus, R. A. a, Am. Cheh. Soc. 1981, 103,
741-747.
DeVault, D. "Quantum-Mechanical Tunneling in Biological
Systems", Cambridge University:New York, 1984.
Brunschwig, B.S.; Sutin, N. Qammehhs Inorg. Cham. 1987,
Q, 209-235.
Endicott, J.F.: Ramassami, T. g, ghys. Chan. 1986, 99,
3740-3747.
Balzani, V.; Scandola, F.; Orlandi, G.; Sabbatini, M.:
Indelli, M.T. J, Am, Chgm. 899. 1981, 102, 3370-3378.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
187
Nielson, R.M.: Golovin, M.N.: McManis, G.E.: Weaver,
M.J. lm_Am1_§hgmm_§Qg1, 1988, 11a, 1745-1749.
(a) Mayo, S.L.: Ellis, W.R.: Coutchley, R.J.: Gray,
H.B. Science 1986, 21;, 948-952. (b) Axup, A.w.: Albin,
M.: Mayo, S.L.: Crutchley, R.J.; Gray, H.B. J. .
h . cc. 1988, 119, 435-439.
McLendeon, G. 599. Qhem, Reg, 1988, 21, 160-167.
Elias, H.; Chou, M.H.: Winkler, J.R. J, Am. Chem. Soc.
1988, 119, 429-434.
(a) Miller, J. R.: Calcaterra, L. T.; Closs, G. L. ,1,
Am, Chem. Soc. 1984, 19a, 3047-3049. (b) Miller, J. R.;
Beitz, J. V.; Huddleston, R. K. J, Am, Chem. Soc. 1984,
199. 5057-5068. (c) Closs, G. L.; Calceterra, L. T.;
Green, N. J.: Penfield, K. W.; Miller, J. R. ,1, Phys,
M3. 1986, fl, 3673-3683. (d) Closs, G.L.; Miller,
J.R. Sgience 1988, gag, 440-447.
Wasielewski, M. P.; Niemczyk, M. P.; Svec, W. A.;
Pewitt, E. B. J. Am. Chem. Soa, 1985, 1g1, 1080-1082.
Heitele, H.; Michel-Beyerle, M.E. J, Am, Cham, Soc.
1985, 191, 8286-8288.
Harrison, R.J.: Pearce, B.: Beddard, G.S.: Cowan, J.A.:
Sanders, J.K.M. Chem, Phys. 1987, 11 , 429-448.
Guerr, T.; McGuire, M.E.: McLendon, G. J. . .
gag. 1985, 191, 5104-5111.
Oevering, H.; Padden-Row, N .M.; Heppener, M.: Oliver,
A.M.; Cotsaris, E.: Verhoeven, J.W.; Hush, N.S. J, Am.
gnamy_§aa., 1987, 199, 3258-3269.
34.
35.
36.
37.
38.
39.
40.
41.
188
Rehm, 0.; Weller, A. I§r1_11_ghem. 1970, 8, 259-271.
(a) Creutz, C.: Sutin, N. J, Am. Chem, Sog. 1977, 99,
241-243. (b) Lin, C.-T.: Bottcher, W.; Chou, M.;
Creutz, C.: Sutin, N. a, Am, Chem, 59g, 1976, 29, 6536-
6544.
(a) Ballardini, R.: Varani, G.: Indelli, M. T.:
Scandola, F.; Balzani, V. J, Am, Qham, Sag. 1978, 199,
7219-7223. (b) Scandola, F.: Balzani, V.; Schuster, G.
5. J. Am. Chem. Soc. 1981, 19;, 2519-2523. (c)
Indelli, M. T.; Ballardini, R.: Scandola, F. J. Phys.
ghgmm 1984, 99, 2547-2551.
Nagle, J. K.; Dressick,‘W. J.: Meyer, T. J. J. .
Qham, $09. 1979, 191, 3993-3995.
(a) Nocera, D. G.: Gray, H. B. a, Am, Chem, §oa, 1981,
191, 7349-7350. (b) Marshall, J. L. : Stobart, S. R.:
Gray, H. B. J. Am, Cham. $99, 1984, 199, 3027-3029.
Heuer, W. B.; Totten, M. D.: Rodman, G. 8.; Hebert, E.
J.; Tracy, H. J.; Nagle, J. K. J. Am, Chem. Soc. 1984,
lag, 1163-1164.
Kavarnos, G. J.: Turro, N. a, Chem. Bay, 1986, 99, 401-
449.
Evidence for the inverted region has been observed in
geminate radical pairs: (a) Ohno, T. M
1985, ‘99, 5709-5713. (b) Ohno, T. Yoshimura, A.:
Mataga, N. J. Phys, Chem, 1986, 9Q, 3295-3297. (c)
Ohno, T.; Yoshimura, A.: Shioyama, H.: Mataga, N. _J_._
Ehys, Chem, 1987, 91, 4365-4370.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
189
Rehm. 0.; and Weller. A. Ber1_Bunsgengeai_£hrsi_§hem-
1959, 1;, 834-839.
(a) Agmon, N.: Levine, R.D. ghama_£hyay_Lahh. 1977, 5;,
197-201. (b) Agmon, N. o . .
1978, 19, 388-404.
Scandola, F.: Balzani, V. J: Am, Cham. Sac. 1979, 191,
6140-6142.
(a) Van Duyne, R. P.; Fischer, S. F. Cham. Rhys, 1974,
5, 183-197. (b) Fischer, S. F.: Van Duyne, R. P. CREE;
ghyaa 1977, aa, 9-16.
Kestner, N. R.: Logan, J.; Jortner,IJ. J. Ehys. Cham.
1974, 19, 2148-2165.
(a) Ulstrup, J.; Jortner, J. J, Qham, Ehya, 1975, 91,
4358-4368. (b) Jortner, J.: Ulstrup, J. J, Am, Chem.
§QC, 1979, 101, 3744-3754.
Newton, M. D. 1mg. 9. Quantam Chem. 1980, 14, 364-391.
Webman, 1.; Kestner, N. R. J, Chem. Ehys. 1982, 7_’7_,
2387-2398.
Beitz, J.V.; Miller, J.R. J. Cham, Phys. 1979,
Id
(...I
4579-4595.
Marcus, R.A. Int, a, Chem, Kihah. 1981, 1;, 865.
(a) Weller, A.; Zachariasse, K. Qhem. Ehys. Lat . 1971,
19, 590-594. (b) Joussot-Dubian, J.; Albrecht, A.C.;
Garischer, H.; Knox, R.S.; Marcus, R.A.: Schott, M.;
Weller. A-: Willie. F. W399. 1979. 12.
129.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
190
Marcus, R.A.; Siders, P. J. Ehgg, thm. 1982, 99, 622-
630.
Marcus, R. A. Qigcus, Eazagay 909. 1960, 29, 129-131.
Noyes, R. M. gngm_gaaapy_giha§. 1961, 1, 129-160.
Debye, P. Iran§1_Elsetreshemi_§22- 1942, 9;, 255-272.
McLendon, G.; Guarr, T. Coopd, gham. Bav. 1985, 99, 1-
52.
Marcus, R.A. J, Qham. Ehya. 1955, 4;, 2654-2657.
(a) Schaper, H.; Schnedler, E. 9. Ehys. Chem. 1982, 99,
4380-4385. (b) Schaper, H.: Kostlin, H.; Schnedler, E.
Ehilips Teah. Rev. 1982, 99, 69-80.
Laser, D.; Bard, A.J. 9.431ectroghem, Soc. 1975, 22,
632-640.
Brilmeyer, G.H.; Bard, A.J. J. Elactrochem. Sop. 1980,
121, 104-110.
Heller, C.A.; Jernigan, J.L. Appl. Optics 1977, 19, 61-
66.
(a) Measures, R.M. Appl. Optics 1975, 14, 909-916. (b)
Measures, R.M. Appl. Optics, 1974, 19, 1121-1133.
Maverick, A. W.; Najdzionek, J. 8.; MacKenzie, D.:
Nocera, D. G.; Gray, H. B. J Am. Chem. Soc. 1983, 195,
1878-1882. (b) Zietlow, T.C.: Gray, H.B. Inorg. Chem.
1986, 99, 631-634.
Nocera, D.G.; Gray, H.B. J. Am, Chem, Soc. 1984, _99,
824-825.
Sheldon, J.C. 9. Qhem. $09. 1960, 1007-1014.
Sheldon, J.C. J. Qhem, S99. 1962, 410-415.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
191
Dorman, W.C.; McCarley, R.E. M. 1974, 11,
491-493.
Hogue, R.D.: McCarley, R.E. 1h9:91_9ham. 1970, 9, 1354-
1359.
Walter, R.I. J, Q, Chem. 99 . 1955, 11, 5999-6002.
Mussell, R.D.; Kassel, D.: Allison, J.: Nocera, D.G.,
manuscript in preparation.
Gagne, R.R.: Koval, C.A.: Lisensky, G.C. 199:91_9ham.
1980, 19, 2854-2855.
Newsham, M.D., Ph.D. Dissertation, Michigan State
University, 1988.
Keszthelyi, C.P.; Tokel-Takvoryan, M.E.; Bard, A.J.
Ahal1_9h9m. 1975, 41, 249-255.
(a) Wallace, W.L.: Bard, A.J. 91_2hy§1_9ham. 1979, 99,
1350-1357. (b) Itoh, K.; Honda, K. Chem. Le 3;. 1979,
99-102.
Bolleta, F.; Bonafede, S. Pure & Appl, Chem. 1986, 99,
1229-1232.
(a) Dye, J.L.; Nicely, V.A. J. Chem, Ed. 1971, 49, 443-
448. (b) Turro, N.J. "Modern Molecular Photochemistry",
Benjamin/Cummings:Menlo Park, CA, 1978; Chapter 9.
Hughbanks, T.: Hoffman, R. J, Am, Chem. Spa. 1983, 199,
1150-1162.
(a) Seifert, G.; Grossman, G.: Muller, H. ,11_JmaLg
9919991 1980, 94, 93-102. (b) Tyler, D. R., University
of Oregon, personal communication, 1986.
80.
81.
82.
83.
84.
85.
86.
87.
192
Zietlow, T. C.; Nocera, D. G.; Gray, H. B. lhppg1_9ham1
1986, 19, 1351-1353.
(a) The effect of disparate values of the electronic
coupling element on chemiluminescent reactivity has
been addressed for annihilation reactions involving
transition metal polypyridyl complexes.81b (b)
Bonafede, 8.; Ciano, M.; Bolletta, F.; Balzani, V.:
Chassot, L.; Zelewsky, A. J. Ehya, Qham, 1986, _9,
3836-3841.
Scandola, F.; Balzani, V.: Schuster, G.B. J. Am. Chem.
Soc, 1981, 103, 2519-2523.
Faulkner, L. R.: Bard, A. J. J. Am. Chem. 9oc. 1969,
.91, 209-210.
Bezman, R.; Faulkner, L. R. 9. Am, Qhem, 99c. 1972, 94,
6317-6323.
(a) Hercules, D. M. Acg. gham, has, 1969, 9, 301-306.
(b) Hercules, D. M. "Physical Methods in Organic
Chemistry" Weissberger, A.; Rossiter, B. Eds.;
Academic: New York, 1971; Part II, Chapter 13.
Chandross, E. A. TIQDS: 9,1, 9999, 991,, 9e:. 2 1969,
91, 571-583.
Weller, A.: Zachariasse, K. "Chemiluminescence and
Bioluminescence" Cormier, M.: Hercules, D. M.: Lee, J.,
Eds.; Plenum: New York, 1973.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
193
(a) Faulkner. L- R- MW
1975, 9, 213-263. (b) Faulkner, L. R. W
1978, 91, 494-526. (c) Faulkner, L. R.; Glass, R. 8.
"Chemical and Biological Generation of Excited States"
Adam, W.; Gilento, G., Eds.; Academic: New York, 1982.
Park. S--M-: Tryk. D- A- W 1981.
4, 43-79
Pighin, A. 9am, Q, Chen. 1973, 91, 3467-3472.
Bard, A. J.; Keszthelyi, C. P.: Tachikawa, H.: Tokel,
N. E. "Chemiluminescence and Bioluminescence" Cormier,
M., Hercules, D. M.: Lee, J., Eds.: Plenum: New York,
1973.
(a) Itoh, K.: Honda, K.: Sukigara, M. 91999199h1m1_99§a
1979, 94, 1195-1198. (b) Itoh, K.: Honda, K.;
Sukigara. M. £1_Elsgtreanali_shem1 1980. 110. 277-284.
Tokel, N. E.; Bard, A. J. ,1, Am, 9hem. Spc, 1972, fl,
2862-2863.
Abruna, H. D. J, Elecppoanal. Cham. 1984, 119, 321-326.
Kane-Maguire, N.A.P.; Guckart, J.A.; O'Neil, P.J.
199:9, Chem. 1987, 99, 2340-2342.
Tokel-Takvoryan, N. E.: Bard, A. J. J, Am. Cham. Soc.
1973, 99, 6582-6589.
Abruna, H. D. J. Elactrocham, 99c. 1985, 199, 842-849.
(a)Gonzales-Velasco, J.: Rubinstein, I.; Crutchley, R.
J.: Lever, A. B. P.: Bard, A. J, Lhopg, Qhem, 1983, 29,
822-825. (b) Lee, C.-W.: Ouyang, J.; Bard, A.J. ,1,
31999999981, Chem, 1988, 244, 319-324.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
194
Luong, J. C.; Nadjo, L.; Wrighton, M. S. M
£99, 1978, 120, 5790-5795.
Vogler, Ad KunkelY. H- W
1984, 21, 316-317.
Ouyang, J.; Zietlow, T. C.: Hopkins, M. D.; Fan, F.-R.
F.; Gray, H. 8.; Bard, A. J. ,1, Ehys, 911m, 1986, 29,
3841-3949.
Wheeler, 8. L.; Nagasubramanian, G.; Bard, A. J.:
Schechtman, L. A.; Dininny, D. R.; Kenney, M. E. J, Am.
Qhem, §QC. 1984, 106, 7404-7410.
Tokel-Takvoryan, N. E.: Bard, A. J. thm, Ehys, LQLE.
1974, 25, 235-238.
Volger, A.: Kunkely, H. WEE 1987, 381,
155-168.
Luttmer, J. D.; Bard, A. J. J. Rhys. Chem, 1981, Q,
1155-1159.
Glass R. 8.; Faulkner, L. R. J, Rhys. Chem, 1981,
I.»
1160-1165.
Kim, J.; Fan, F. F.; Bard, A. J.; Che, C.-M.; Gray, H.
B. gngm, Pays, ngg. 1985, lg_, 543-546.
Che, C.-M.; Atherton, S. J.; Butler, L. G.: Gray, H. B.
W 1984. 1.0..6. 5143-5145.
Liu, D. K.; Brunschwig, B. 8.; Creutz, C.; Sutin, N.
Q, Am, Qhem, $02, 1986, 198, 1749-1755.
110.
111.
112.
113.
114.
195
Owing to our inability to find nitroaromatics that are
reduced at potentials less negative then p-
dinitrobenzene, some aromatic guinones were employed
for the reaction of M06C114' with negatively charged
donors.
Faulkner, L. R.: Bard, A.J. MW. 1977,
19, 1-95.
The lowest excited-state energies for donors and
acceptors used in ecl studies are: 2,6-dimethyl-p-
benzoquinone, ET = 2.32 V (Trommsdorff, H. P. J. C
ghye, 1972, eg, 5358-5372): p-dinitrobenzene, ET = 2.54
V (Brown, R. G.: Harriman, A.: Harris, L. J. Chem,
So Fara ns. 2, 197a, lg, 1193-1199); 0-
dinitrobenzene, ET = 2.55 V, and phenothiazine, ET =
2.62 V (Alkaitis, S. A.: Gratzel, 14.: Henglein, A.
Berg, flunsenges, Phys, ghee, 1975, 12, 541-546); 10-
methylphenothiazine, ET = 2.66 V (Rothenberger, G.;
Infelta, P. P.: Gratzel, M. J, Phys, Chem. 1979, _8_3_,
1871-1876) : tris (p-tolyl) amine, ET = 3 . O V
(Zachariasse, K. A. Ph.D. Thesis, Free University
Press: Amsterdam, 1972): tris(4-bromophenyl)amine, ET
> 3.1 V (Terenen, A.: Ermolaev, V. Sevie; ED¥§iQS
Qeï¬legy 1962, g, 600-602).
Faulkner, L. R.; Tachikawa, H.: Bard, A. J. J. .
Qnemy_§ggy 1972, eg, 691-699).
Kowert, B. A.: Marcoux, L.: Bard, A. J. J. . .
$09, 1972, 24, 5538-5550.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
196
Brunschwig, B. S.: Ehrenson, S.: Sutin, N. ;y_Amy_§hemy
§Qc, 1984, 196, 6858-6859.
Sutin, N.: Brunschwig, B. S. W 1982, 128,
105-125.
Sutin, N. W. 1983, ;_Q, 441-498.
Kakitani, T.: Mataga, N. ,L_2hys_,_ghem_, 1987, 2],, 6277-
6285.
Hupp, J.T.; Weydert, J. M; 1987, 2_6_, 2657-
2660.
Ennix, K.S.: McMahon, P.T.: Rosa, R.; Curtis, J.C.
lhogg, Chem, 1987, ee, 2660-2666.
Kahlow, M.A.: Kana, T.J.: Barbara, P.F. J, Phye, Qhem,
1987, 21 6452-6455.
Boyd, D.C.: Rodman, G.S.: Mann, K.R. J. . S c.
1986, 10;, 1779-1784.
(a) Borchardt, D.: Pool, K.; Wherland, S. Ihorg, Chem.
1982, .;;, 93-97. (b) Nielson, R.M.; Wherland, S.
lhozg. Chem. 1984, 2;, 1338-1344. (c) Borchardt, D.:
Wherland, S. Ino . m. 1984, ,g;, 2537-2542.
(d) Nielson, R.M.: Wherland, S. IDQIQ- Chem. 1986, 2;,
2437-2440. (e) Stebler, M.; Nielson, R.M.: Siems,
W.F.: Hunt, J.P.: Dodgen, H.W.; Wherland, S. M;
Qhem; 1988, e1, 2893-2897.
(a) Kjaer, A.M.: Ulstrup, J. M... 1986, 2_5,
644-654/ (b) Kjaer, A.M.: Ulstrup, J. I or . m.
1987, 26, 2052-2058.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
197
(a) Truong, T.B. J, Ehye. Chem. 1994, fl, 3906-3913.
(b) Truong, T.B. gene & A921, ghem. 1986, ea, 1279-
1284.
Rustman, P.: Brown, T.L. J, Am. Qhem, §oc, 1987, 199,
3632-3639.
Roller, H.B.; Hawkridge, F.M. mm 1985,
1g1, 7412-7417.
(a) Nielson, R.M.: McManis, G.E.: Golovin, M.N.;
Weaver, M.J. J. Phys. Chem. 1988, 9;, 3441-3450. (b)
Nielson, R.M.: Golovin, M.N.: McManis, G.B.; Weaver,
M.J. J. Am. Chem, 809. 1988, 11;, 1745-1749. (c)
McManis, G.B.: Mishra, 'A.K.: Weaver, M.J. J, Phys.
Qhemy 1987, §_6_ 5550-5556. (d) McManis, G.B.: Golovin,
M.N.: Weaver, M.J. J. Phys, Chem. 1986, eg, 6563-6570.
(a) Nadler, W.: Marcus, R.A. J. Chem. Phys, 1987, ;§,
3906-3924. (b) Sumi, H.: Marcus, R.A. . E e a a
Chem. 1986, 29;, 59-67.
Calef, D.F.: Wolynes, P.G. J. Phys. Chem. 1983, _1,
3387-3400.
Hynes, J.T. J. Phys, Chem. 1986, eg, 3701-3705.
Simon, J.D. ACC. Chem, Res, 1988, 2;, 128-134.
Cukier, R.I. Q, ghem. Phys. 1988, 21, 3938.
(a) Rips, 1.; Jortner, J. W113; 1987, 1;3_,
411-414. (b) Rips, 1.: Jortner, J. J, Chem, Phys.
1988 88, 818-822.
McGuire, M.: McLendon, G. _J_,_Ehye,__e‘hem; 1986, 29,
2549-2551.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
198
Keszthelyi, C.P.: Tokel-Takveryan, N.E.; Tachikawa, H.:
Bard, A.J. W 1973, 2;, 219.
(a) Kim, J.; Faulkner, L.R. J, Am, Qhem. ï¬ec, 1988,
11.0.. 112-119. (b) Kim, J.: Faulkner, L.R. J_._
W 1988, 252,, 107-121.
Zoski, C.G.; Sweigart, D.A.: Stone, N.J.: Rieger, P.H.:
Mocellin, E.: Mann, T.F.: Mann, D.R.: Gosser, D.K.:
Doeff, M.M.: Bond, A.M. ,1, Am. Qhem, §oe. 1988, _1_Q,
2109-2116.
Gonzales-Velasco, J. J, Phys. Qhem, 1988, _9_2, 2202-
2207.
Murray, R.W. "Electroanalytical Chemistry": Bard, A.J.,
Ed.: Dekker:New York, 1983: Vol. 13, p. 191-368.
(a) Potts, K.T.: Usifer, D.A.: Guadalupe, A.: Abruna,
H.D. ,1, Am. Qhem, gog. 1987, .192, 3961-3967. (13)
Guadalupe, A.: Usifer, D.A.: Potts, K.T.: Hurrell,
H.C.; Mogstad, A.-E.; Abruna, H.D. J. Am. Chem. Soc.
1988, 10, 3462-3466.
Sharp, M. J, Electroanal. Chem. 1987, 2;0, 109-124.
Hupp, J.T.: Meyer, T.J. J, Elee§reene1, Chem, 1987,
32:1, 59-65.
Bettelheim, A.: White, B.A.: Raybuck, S.A.: Murray,
R.W. W 1987, 2_6, 1009-1017.
Redepenning, J.: Anson, F.C. J, Phys. Chem, 1987,
1.
4549-4553 .