Pure & AppL Chem., Vol. 58, No. 9, pp. 1 193—1206, 1986.
Printed in Great Britain.
© 1986 IUPAC
Photochemistry of metal coordination complexes:
metal to ligand charge transfer excited states
Tiunas J. Meyer
Department of nry, The University of North cro1ina, thapel Hill,
N.C. 27514, U.S.A.
Abstract - Systanatic variations appear in the othphysical and
photochnical properties of t4LC excited states whicth n be accounted for
qualitatively or even quantitatively based on the properties of the
nxlecules and of the surrounding rreditin.
In 1959,
Paris and Brandt first reported anission fran the cxinplex (RuUpy)
(bpy is 2,2'—
bipyridine) in solution (ref. 1). The excited state involved
sufficienty long-lived
(-8OO ns in water at roan tanperature) that the possibility existed for using it as a
thernical reagent and based on quanching studies using ccmplexes of Co(III), 1dainson and
excited
cazorkers suggested
a reducing agent (ref. 2). Their
suggestion was soon verified by additional quanching and flash çkiothlysis studies (ref.
3,4). It was also shawn that the excited state could act as an oxidizing agent (ref. 5),
is
that the
state could act as
that its redox potentials could be estimated experisontally by a kinetic qunnching techniqun
(ref. 6,7),
of oxidants
it
facile, bisolecular electron transfer with a varie
and that
could undergo
or reductants (ref. 8—10). For example, oxidative qcnching by paraquat (PQ ),
Scheme 1
(1)
Ru(Lpy)32 _Y) Ru(bpy)32
2+*
2+ kq
+
3+
+ PQ —p Ru(bpy). + PQ
Ru(bpy)
k
2
(2)
2
(3)
—*Ru(bpy)3+PQ
provides a basis for the conversion of visible light energy into a transiently stored redox
Similar electron transfer schanes had been established previously for organic excited
states
the high c±ianical
of the Ru ccznplex in sore than one
oxidation
and its high light absorptivity
the visible made it an attractive
candidate
exploiting the energy conversion possibilities associated
pair.
(ref. II), bit
state
for
stability
in
with E. 1-3.
The successful utilization of RU(bpy) 2+ and related cxiiplexes in energy conversion schanes
relies on establishing a basis for un&rstanding the light absorptivity and otophysical
(ref.
and *iotochanical properties of the cthrcrtophores involved
9a ,l2). The goal of this
account
describe those properties and to provide a basis for understanding than
the
is to
microscopic
level.
at
ELECTRONIC STRUCTURE. LIGHT ABSORPTIVITY
]2+ and (Os(phen) ]2+ (phan is l,l0-enantroline) tare with oanplexes like
(Re(phen)(o)3Cl], and (Ir(bpy) Cl)] the ground state electronic
configuratio (dii)6 (t in O syninetry) and vaant, 1CM lying ir*(bpy or men) orbitals on
the polypyridyl ligand.of apircpriate syrrmetry to mix with the dir orbitals. Visible light
absorption in these carlexes is usually cinated by intense absorption bands arising f ran
drr .- ir* metal to ligand charge transfer (MLCT) transitions, e.g., b(kpy)(CO) + hv r4o'(&y (CD) . For the lox oxidation state oarlexes of tb(O) and Re(I) the crbonyl groups
(Ru(1:py)
(Mo(bçy)CO) 1,
secessai4 to mix with and stabilize the dir levels thus decreasing sensitivity taGards
oxidation and bringing the dir_ir* (kçq or p&en) energy gap into the visible region of the
are
spectrun. In the higher oxidation state axilexes of Rh(III) or Ir(III) the dir levels are
so stabilized by the higher effective nuclear charge at the metal that the MLCT transitions
occur at high energies and can overlap with ir - lr* based transitions localized on the
polypyridyl ligands.
1193
1194
T.J.MEYER
TI• host straightforward experimental evidence supporting the assignant of the visible
absorption bands to M1CP transitions and of nission to the corresponding LbT transitions
nission barx energies and ground state
rathiction potentials for the appropriate natal and ligand—ased couples (ref. 13,14 ) e.g.
cctes fran linear crrrelations between absorption or
(LD, Q13O1 31 ...)
(y)Os1(L) + e -4
(y)OsU(L)
+ e ->
(bpy)OsL(L)+
(ty)OsU(L)
(E)
(4)
(Ered)
Experimontally, linear correlations exist between E, (=E —E
is the energy cuari€?ty f8
expected since, for example,
(y)Os'(L)
+
(by)Os(L)
of the intramolecular,
III
(t%y)Os
—
) and
E
or E as
outer-sre &avalent
-4 2ftçy)Os11(L)
(6)
INCT process that corresponds
to enission (ret. l4a).
2+*
II 2+
.4 (tpy)Os (L)4 + h
(L)4
(7)
A schenatic orbital and state energy level diagram for (dir)6-polypyridyl ccznplexes is shc&n
in Fig. 1.
H
l3(Hff*) ((H)l(H*)l)
2 * ______
1 2
13 (dd)
do ,do
((dlI)
13 (MLCT)
711
1'2'3
5(do*1
((dl!)
5*1
1(GS) ((71)6)
1L
(Orbitals)
(States)
Fig. 1. A schenatic orbital and state diagram for (dir)6 MEC excited states
excited states exists because of the presence of both filled dir and
(kpy or çien) levels and of unfilled ir* and &).* (eg O synnetry) levels. The relative
disposition of the states and the state or states which control photochanical and
photophysical properties follaiing excitation depend on the metal and its exidation state,
A ocxnplex melange of
ir
on the chrarcphoric ligand, on the additional, nonchrcxnophoric ligands, and can even depend
on the nature of the surrounding rrediun. Hcwever, in general:
1) Visible light absorption is usually &zxiinated by transitions to MLCr excited states
which are largely singlet in character -((dir)5(ir*)l and nission occurs fran M[CT states
which are largely triplet in character, 3((dir)S(ir*)h]. The energies of the MLCr states vary
systamatically as the rxrn-c±irancphoric ligands are varied. This is an expected result givçn
the considerable ifference in electronic configuration at the metal between the (dir)(ir*) 1excited, and (dir) ground states. Fbr example, good backbonding ligands like CO stabilize
(dir)6-Ru(II) k.y metal to ligand backbonding but not (dir)5-Ru(III) to any appreciable degree.
Consequnntly, incorporation of good backbonding ligands leads to an increased energy gap
betwgen the excited and gr9nd states. ¶Ite change in electronic configuration between the
(dir) (1r*)l excited and (dir) ground states involves dir orbitals which are largely nonbonding and only half population of anti-bonding ir (Lpy or phen) orbitals. Consequnntly,
relatively 3nall molecular distortions in the excited states are expected to exist.
Hcwever, given the significant change in radial distribution involved in the MECT
transition, absorption spectra and the energies of eguilibrated MLC excited states are
to be nediun dependent.
2) In the dd excited states (dir)5(do* )l,
expected
there is a significant increase in electronic
repulsion aI&ig the metal ligand sigra bonding axes because of the pranotion of a largely
non-bonding electron to a sigea anti-bonding th.* orbital (ref.. 15). cause of their
LaPorte forbiddenness, absorption bands arising fran dd transitions are weak and not
generally observed in the absorption spectra of (dir)6 polypyridyl complexes. In less
spectroscopically cluttered ccinplexes of rutheniun, dd-bauis appear at relatively high
1195
Photochemistry of metal coordination complexes
,
energies,
< 400 rin for [Ru(NH3) 6]2+ (ref . 16) .
Bever ,
because
of the significant
differences in structure along the netal-ligand bond axes between the (dir) 6
(dir)5 (doi)
confgur, when thermally uilibrated, excited states appear at much 1cer
energies. If accessible they cn play a siguificnt role in the thotochnica1 and
photophysical properties of a ocinpiex since they are relatively short lived in solution and
are often the source of *iotochnica1 instability arising fran ligand loss . The energies of
dcl excited states above the ground state in analogous cnplexes increase approximately 3O%
in sassing fran the first transition series to the second and —30% again fran the second to
It
is for this reason that MLCT excited states ksed on Fe(II) are
17 ,18) .
states lie lciest,
essentially unexploitable in energy conversion applications because
the çkotophysical properties of rutheniizn mplexes are scinetrnes quite ccznplex because
state are cunparable, and is coe of the endearing properties of
the energies of MLCP and
states are often sufficiently high in energy that they play no
unplexes of Os(II) where
the third (ref .
role.
3) _1*
states—Absorption
bands arising fran polyridine 1ocalizI 'rr-' transitions
appear to vary sczhat in energy with the netal and its ccidation state but generally
of the lack of dipole ctharacter
24O tin (r*2) •
appear at 3OO mi
for the 7T.41r* transition, Tr_.rr* eXCited states, are relatively insensitive to solvent
variations and at lc tanperatures vibrational structure arising fran araiatic ring based
stretching nodes can appear in their nission spectra (ref . 19 ) . Polypyridine based -i *
excited states are noarly always at higher energies than MLCr excited states except for
cases like Rh(III) where the relatively high oxidation state stalizes the drr levels belc
irU:py or *en) (ref . l2a,20 ,2l) or for mplexes like (Ru(biq) ] (biq is bis—isoquinoline)
where
(ref.
stabilization of the dri orbitals by netal to ligand backonding leads to ir-ir < MECT
19).
A considerable ccinplexity exists in the MLCP spectral region (ref . 22) arising fran a nunber
three dir orbitals are mixed and
of factors: 1) In the MrCT excited state, (dir)5(r*) 1,
further
At the polyrridyl ligands
by 1cM syrrinetry and spin orbit
level . 3) Even
electronic interactions between ligands can cause a splitting aisongst the
single polypyridyl—based &ranophores , the polypyridyl ligand has a series of ir *
which can provide the
split
coupling. 2)
in
levels
basis for multiple MLCr transitios. 4 Ther are both LCT singlet
and triplet states. Although the triplet transitions , [(dir) ] —> ((dir) 5(it*) ] , are
forbidden in the absence of spin-orbin coupling, they becaue allced when spin orbit
coupling is inclnded. Frau the theoretical point of viez the effect of spin orbit coupling
is to "mix" pire singlet and triplet states whiob imparts singlet äiaracter and allazedness
to transitions involving triplets.
¶[ypically, absorption bands in the tAT and visible are observed arising frau dir * and dir
9 transitions. ¶L1 electronic spectra can be quite xxnplex as illustrated by t?ie lcz
tenperature single crystal spectra of M(y) (M=F,Ru,Os) obtained by Ferguson, et. al.
(ref. 23). In the spectra nany absorption fatures can be observed in the dir - lT region.
-
Tbe various spectral features can be assigned to a series of MLC transitions to both 'MLCr
and 3?.Wr states and to their vibronic ccinponents based on a paraxneterized nolecular orbital
nodel which inclines the effects of spin orbit coupling (ref. 24,25). The appearance of
transitions to "MLCP" states at low energies s especially pronounced for Os where the
large spin orbit coupling constant (X3O00 au') ensures significant singlet-triplet mixing.
Analysis of the spectra using the nolecular orbital nodel shows that the lowest energy MLC
states arise fran a comson set of dir and ir* orbitals, are largely triplet in character, and
are mixed to different extents with low lying MLC singlet states.
Electronic structure. What happens following optical excitation?
excitation thronghout the dir -* ir* spectral region results
For
optical
in theRu(bpy)t
rapid, <1 ps (ref. 26),
quantitative appearance of an enitting 3t.WP state or states (ref. 27) whose auission is
sanewhat irediun dependent bit occurs around 600 un. Fran absorption spectral neasurenents
the energy gap between the lowest 1twr and 3rwr states is only 5000 au-. Since the
"singlet" and "triplet" states are mixed by spin—orbit coupling, it is not surprising that
excitation into the ljIp absorption nanifold leads rapidly to the emitting ½wr state or
states.
However, an interesting qstion retains in multiple chelates. Is the excited electron
localized on a single ligand as in ((bpy).RuIII(by)]2+*, possibly with rapid exchange
occurring between ligands, or dees the excited electron occupy a urolecular orbital which is
delocalized over all three ligarxls as in (Ru111 (bpy-l/3 )]2+ ? In fluid solution, the
excited electron in the enitting excited state(s) appear to be "localized" based on a
number of lines of evidence: 1) Transient Resonance Reman spectroscopy where discrete,
freguency shifted Raman lines are observed in the V (bpy) ring stretching region (1000-1500
air1) for both reduced and unreduced py ligands (ref. 28,29). 2) Site selection
Excited state
spectroscopy based on emission polarization rceasurenents (ref. 30).
correlations for related nono, bis, and tris-dielates like (Os(b) ] ,
(osUpy)
Ph2aI!PPh2) ]2+, and (Os(}çy) (cis- k2Hi1PPh2)2]2\which have c.osely correlaed excited
T.J.MEYER
1196
state properties without differences arising sihidi cxuld be attributed to 1oca1zation
effects (ref . 31) . 4 ) In related ligand reduced cxxnplexes like (( u(b,y) ] generated
e1ectrochnica11y, there is clear spectral evidence that the added electron is localized on
a single ligand and frau epr line broadening that the rate of electron Ixpping between
ligands is rapid (ref .
32,33).
experimental evidence for localization in fluid solution is convincing but the situation
is less clear in the solid state or in l tuperature glasses where the evidence ranains
sctrwhat uivocal (ref . 34 ,35 ) . The qistion of localization rsus cloca1ization is
Tbe
really a probln in mixed-valence dnistry where the issues involved have been discussed in
sane detail and applied nainly to ligand-bridged, mixed valence diiners (ref. lOa,36,37).
Tbe critical factors are the relative nagnitudes of the electronic coupling between the
electron &nor (b5y) and acceptor (bpy) sites , V1 and the vibrational trapping eflergy, X.
Vibrational traing arises fran the structural differences that exist between kir and Iy
and the solvent dipole reorientations that must occur for electron hopping to occur. ftaii
the theoretical point of view localizaion occurs when V > x/2 (ref. 36) . Fran
and related ccxnplexes it is clear that rxnspectroscopic neasurenents on (Ru(bpy)
neglible contri}xitions to_yibrational trapping exist fran the solvent in polar organic
600-800 an ) . Even in the solid state or in a frozen solution where solvent
solvents (X
dipole traping cn sot occur, contribetions fran intranlecular vibrational trapping exist
and there is so guarantee that sufficient electronic coupling between tçq and ty exists in
the excited state for localization to occur.
Experimental evidence is available shczing that in fluid solution the electron is initially
localized uring the electonic transition. The ground states of ociuplexes like
[RuUpy) ] + or [Os(tpy) ] + have D3 syrrinetry and rx permanent dipole nunent. If cptical
excitati6n produces a delocalized excited state, the D syniretry of the ground state would
be preserved and MECr absorption band energies should e independent of the solvent. (i the
other hand, if excitation leads to an electron localized on a single ligand, a dipole is
induced in the excited state and fran dielectric continutun theory (ref. 38) the energies of
MLCT transitions are predicted to vary with the cptical dielectric constant of the solvent,
Dop(n2), as (ref. 39),
2 1-D
(8)
In Eq. 8 ET is the solvent induced energy shift in the band maxitnun, t the dipole narent
induced in the excited stats and a the nolecular radius (6-7A for [Ru(bpy) ]2+) Fbr both
[Ru(bpy) j2+ and (Os(tpy) I absorption band energies vary as predicted by3Eq. (8) and f ran
the slops of the lines i can be estimated that the dipole iranents induced in the excited
states are t 14±6 D (ref. 39).
Low temperature measurements. Evidence for multiple MICT excited states
Fran
the results
tanperature dependent enission and lifetine studies, where the contributions of Crosby
40,41) and more recent enission polarization experiments
and co-workers are
is clear that following MLC excitation, contributions to excited state
(ref.
properties exist fran a series of states. In particular, below 77K there is clear evidence
for contributions to excited state decay and enission fran three low-lying Mi_cr states based
on changes in anission quantimi yields, lifetimes and enission polarization as the
tnperature is increased above 4K.
of
34a) it
rtable, (ref.
Experimental evidence also exists for a fourth Mi_Cr state at 600 an1 above the lowest
considerably shorter lived. Fbr [Ru(tpy) ]2+ the existence of a fourth state
state which
is obscured in tanperature dependent lifetime neasuranens in solution because of the
intervention of a strongly tsnperature dependent transition to a low-lying &1 state or
states that became important above 200K. However, for (Os(tpy)3]2 and related ccinplexes
the
evidence
Os(II), tnperature dependent lifetime studies provide
(ref. 41,42).
participation of a fourth state in
excited2tate decay at higher tEnperatures
Evidence for a fourth state for [Ru(kçy) ] has also been obtained fran tEnperature
dependent emission polarization experimetts (ref. 34a).
is
for
direct
of
A paraireterized molecular orbital model which includes spin orbit coupling and assumes
localization of the excited electron has been applied to the MI_cE excited states of
predicts the existence of four
model
tRuUçy) ]2+ and [Os(bpj) ]2+ (ref. 44).
correctly
low-lyin MICE excited states. It also predicts that for Ru the lowest three MLC states
is expected to have
have less than 11% singlet character, 26% for Os, while the fourth
a higher percent of singlet character.
state
Where the data
equivalent
are available,
canplexes of Ru
pattern of
excited states exists for
a closely related
and Os and state by state ccznparisons are revealing (ref. 45).
1197
Photochemistry of metal coordination complexes
the Einstein transition probability for spontan&us uission, the
radiative rate constant, k, j1 thflflS of the integrated &nittd intensity at energy ,
is given T Eq. 9. The term <v1(fI(v))/(fI()v3d\))E is the average va1i of v3
for the amission band corresponding to the radiative transiti&!I (ref. 46) and <d> =
the electric dio1e transition rrunent intera1.
For example , fran
<ei
IlPg>
(9)
kr (64ic4n3/3h)<'3>1<&2
In Eq. 9 ) and are the excited and ground state electronic wave functions arx n is the
refractiveeindex f the iieditin. ¶[ itagnitie of the electric dipole transition integral
depends cn the acunt of mixing of higher lyi1ng singlet excited states into the enitting
triplet excited states, 4 =
theory,
)
(l2)(31I)e) where fran first order perturbation
the mixing cceffi&ent oc is given by,
= <esoI3e>/(e3Ee)
(10)
In Eq. 10
is the spin orbit coupling cperator and 1Ee, e the iergies of the ire,
excit singlet and triplet states. Inclixling the mixing coefficient, the transition dipole
itanent integral is given by,
=
(11)
<1lPerj1iPg>
The square of the mixing coefficient is proportional th the spin orbit coupling constant
3000 an ) and it follcsws fran B. 9 that in cciuparing sqivalent
1100 an1;
staEs of Ru and Os,
[k/(Ean)3]Ru
u<eerg>Ru
(12)
(k/(Ean)3]os os<'eI'g>s
the ata are availble, a state by state coTparison shows that the ratio
[(k/(E) ]>J[kJ(E) 10s is l/3, the ratio of the spin-orbit coupling constants (ref.
Where
45).
This observ&ion clearly suggests that in terms of pire electronic effects the single
ncst important difference between Ru and Os is the magnitode of the spin orbit coupling
constant which determines the singlet content of the low—lying, enitting MLCP states.
The results of texLperature dependent experiuents clearly show that the "enitting MLCT
excited state" of Ru(tçy) and related ccmplexes is a ccxnposite of states. Although all
four lowlying states share a caiiion (d7r) 5(r*) electronic configuration, they are mixed to
different extents with low lying excited singlet states and their energies and decay
characteristic1are different. The energy spacings anongst the three lowest states are
enall (< 200an ) and at tenperatures near RP, to a good approximation, photophysical
properties can be treated as rising fran a single state having the averaged properties of
the three contributors. ¶[ energy gap to the fourth state is higher, it is shorter lived
because of its greater singlet character, and its population and decay can influnnce the
tenperature dependent properties of MLC cthraiophores near roan tenperature in solution.
MOLECULAR STRUCTURE
In (d7r)5(Tr*)1 MLC excited states, 1—electron occupation of the antibonding * orbital is
expected to lead to structural changes in the chranophoric ligand based on known crystal
structures of reduced or partly reduced polypyridyl ligands (ref. 47).
Structural differences between the ground and excited states can be resolved into a linear
cc*nbination of the nermal nodes for which there is a change in frequnncy between states or,
importantly, for which the change in electronic distribution leads to a difference
between states in the equilibritin displacenent. The pattern of contributing normal nodes
sore
can be discerned fran Resonance Ranan spectra by observing which nodes are resonantly
enhanced by scattering fran MLC transitions since resonance enhancenents can only be
observed for those nermal nodes which respond to the change in electronic configuration
between the states. Synuetry plays an important role. Fr example, if there is a
syninetrical change in structure between the ground and excited states only those normal
modes which belong to the totally synmetric representation of the point group can be
involved in the interconversion between structures. Pbr MtC excited states, molecular
orbital calculations have been used to predict the structural change upon excitation and to
evaluate qualitatively the contributing normal nodes (ref. 48).
Fran Resonance Ranan experiments on Ru(bpy)4 and related canplexes a series of 7-8 bpybased ring stretching nodes are enhanced in the range 1000-1600 an-', a ring torsional node
T.J MEYER
1198
at 6OO an' ar a series of 1cM freqncy nodes sane of which n be assigned to netal
ligand stretching vibrations (ref . 28 ,29 ,31 ,49 ) . Mditional information cn be thtain
frciu 1cz tnperaure nission spectra where well-defined vibrational progressions are often
observed for MLC enitters . The energy dependence of the anission intensity is given }ij
(ref. 50,51),
Iv, =(644c/3h)N( v,v
*)3<&2<xj,v
.
j,
>2
(13)
In Eq. j.3 the nission intensity as the nunber of quanta enitted per energy interval,
Iv,v*( v ) , is given as the prc4uct of the number of nolecules in the nitting state, N
the energy of the transition, v * , the speed of light, the transition dipole nunent, and
the product of the vibrational o&erlap integrals (Franck Condon factors) for the
contributing rcrmal nodes . The product is ocer all the contributing nodes, j, inclnding
solvent and other nditin ncdes . v and v are the vibrational quantum numbers in the ground
arid excit1 states. Assuming the validity of the harironic oscillator apprcximation, the
to nission fran a v*=0 level in the excited state to a level v in the ground
state is given by the square of the vibrational overlap integral (ref . 51),
contribition
<XvIXv*0> = v! exp(-S)
(14)
In Eq. 14, S is the electron-vibrational coupling constant which
difference in equilibrius displacanent between the ground
as shown in Eq. 15 where M is the reduced ness for
angularefrequency.
is a neasure of the
and excited states, Q
related to Q
is the
.
It is
the vibration nd w(=2iiv)
S=()(AQ)2
(15)
using appropriate expressions for the vibrational overlap integrals in Eq. 13, it is
possible to derive expressions for the anission spectral profile in terms of the properties
of the contributing vibrations. For [Ru(bpy) 3]2+ it is necessary to inclix3e contributions
By
from aditzn frequency node with hvM=hvMl300_l400 an-, a low frequency node with hwL3OO_
400 an and also contributions fran solvent nodes (dipole orientations) but treated
somiclassically (ref. 31,46,52,53). By canparing calculated and observed spectra, the
omission spectral fitting procedure can be used to obtain values for the parameters,
L' hwL, Eco and PWHM.FSHM is the full width at half naxinnxn (FWHM) for an
indiviual vibrational cosponent. 'Ihe solvent enters through FWHM as a broadening of the
individual vibrational crinponent and is related to X , the classical solvent vibrational
trapping energy for the electron transfer process co?responding to the excited to ground
state transition by (ref. 36)
= (FHM)2
l6kTln)
(FWflfl2
23l0 (in an—l at R)
(16)
Given the evidence from Resonance Ranan spectra for the participation of a series of nediin
and low frequency modes, experimentally derived values for S , hw are averages of
contributions from 7 or 8 ring stretching nodes and 5L' L f a eries of low frequency
modes incleding metal ligand stretching modes.
Application of the omission spectral fitting procedure to a series of polypyridyl canplexes
of Ru(II) and Os(IIj (ref. 31,46,52,53) have revealed: 1) Best fit values for hw in the
range 1250-1400 an consistent with an averaged value for the contributing ring des
observed by Resonance Reman. 2) Values of M that increase linearly with F.m, the v*=0, )
energy gap between the ground and excited sthtes. The linear increase can be reconciled, at
least qualitatively, as an increase in the extent of charge transfer between ground and
excited states as Eon increases.
The experimentally derived values of S lead to a direct approach for calculating excited
state structures (ref. 31). Anaver4e C-N,C-C bond distance change in the excited state
ccitpared to the ground state, Ar, can be calculated from by,
A=
(
(17)
where b is the number of C-C and C-N bonds in the ligand, b = 13 for py, and M is the
reduced mass. Applying this analysis to (Os(by)3]2+* gives Ar = 0.026+0.0005A.
Estimates for changes in average bond dispacoments between the ground and excited states can
also be obtained based on Badger's rule using the difference in vibrational frequencies
between the ground and excited states obtained by ground state Resonance Ranan and excited
state transient Resonance Reman neasurenents (ref. 28,29). The relationship between the
bond distance change, Ar = r*_ r , and the frequency differences is shown in Eq. 18.
Ar = r* — r =
vib//Vvib)hh'245
(18)
1199
Photochemistry ofmetal coordination complexes
Based on freqincy differences for five vUçy) nodes fran 1268-1610 cm, values of r in
the range O.0106-O.0173A have been c1cu1at in good agrenent with Ar fran nission
spectral fitting (ref. 31) . Fran the vibrational analyses, the bpy ring is rxt considerably
distort in the excited state. As expected, ccarisons within a series of relat
ccinplexes shz that r increases with the excited to ground state energy gap consistent
with a greater degree of c*iarge transfer. One ccxnplication that is yet to be clearly
etc. are significantly different for the three llying wr
resolved is whether S.d,
states • At or near than tnperature the results obtained t!j aission spectral fitting are
probably those of an averaged contribution fran all three states.
EXCITED STATE DECAY
In nany cases ctical excitation of MECP cthrancphores leads to the population of a series of
1CM-lying MLCT states whose radiative ar nonradiative properties &iiiinate excited state
ccay at least at 1CM tnperatures. At higher tanperatures additional, tEnperature
dependent processes often appear which play an important role in helping to dictate the
photochenical and photophysical properties of the ccmplexes. Before turning to a detailed
discussion of radiative and onnradiative decay fran MEICT states, it is of value first to
consider the cmpetitive processes which appear at higher tanperatures.
dd states
In 1976 Van Houten and Watts stedied luminescent lifetines and anission quantum
yields for (Ru(bpy) ]2+ in water fran 0 to 100°C (ref. 54). They found that they could fit
the tanperature depndent lifetine data to an expression of the form shCMn in equation 19
where
k was interpreted
constants
for
to be the sum of radiative, kr and Ix)nradiative, kurt decay
the lCMlying, "averaged" MEJCP state.
k + k°(E'/RT)
T(T)1=
rate
(19)
the interpretation given by Van Houten and Watts AE' was the energy gap f ran MECT to a
In
lCMlying dd state and k° the rate constant for the subsequent decay of the d+state. e
pathway for decay of the dd state(s) is ligand loss; t1otolysis of (Ruthpy) ] in
nonaqueous solvents in the presence of anions like Cl or S gives rise t6 photoinduced
anation (ref. 55,56). Ligand loss proceeds via unidentate intermediates like
[Upy) Ru(NCS)(-py)] , which subsequently loses y to give (Ru(bpy) (NCS)21 (ref. 57).
In a ]iter study in dichloranethane both lifetirre and photochenical qu&ntum yields for
ligand loss were observed for [Ru(bpy)3 2+ as a function of tanperature for a series of
anions (ref. 58). Fran those results the photochanial reaction schane in Fig. 2 was
proposed to count for the observed reactivity. In the absence of direct spectroscopic
evidence for the dd state, it is sot known whether it or 3Micr lies lower in energy for
[Ru(bpy) 312+.
(44)
/
/
/7
(bpy) Ru(py—py)
2
I
—L
L
0
(A)
(B)
2A. ergy vs. Ru-N displacnent
for ERuUçy) ]2+ illustrating the
relative dispositions of the three
Fig.
lowest and fourth 3MLCP states, the
lowest 1MLCr state and the energy
1
Gs
+L
barrier to the 3Mtcr--dd transition.
¶[te relative energy of the dd state
is sot known.
Fig. 2B. A schanatic diagram
illustrating ring opening and
subsequent reactions fran the dd
state.
Based on this schane different kinetic limits were proposed to explain the tanperature
dependence data. In one, the MIJCP and dd states are in dynamic eqilibritin as suggested by
Van Houten and Watts and /\E' is the energy difference between 3MLCr and dcl. In the other,
the 3MLCr-+dd transition is an irreversible surface crossing and E' the energy of
activation for the process. Fr cutipletely chelated cciuplexes like (Ruthpy) J, even
though photosubstitution and chelate opening can be relative1y2fficient in olvent like
water, subsequent chelate ring closure, ((tpy)2 Ru(H2O)(W-T)] - ((tçy)2Ru(bpy)] + +H20,
provides a useful pathway for "annealing" the systan and avoiding ligand loss
T.J.MEYER
1200
Although the ç&iotochnica1 iTechanin implied in Fig. 2 appears to be appropriate for
2+ it nay be far fran adequate to cover all possible ses. Fbr example, for
(Ru(kpy)3]
and related ecraplexes a striking difference between the
(Ru(bpyL,().,]2 (w pyridine)
tsnperatre dpendences for otoanation and lifetime (ref . 59 ,60) has led to the suggestion
that direct population of the c state(s) nay occur fran singlet MLCr states follcwing
initial M1CP excitation.
TI roles of solvent and ligand variations on the thermally activated 3iircr-&i transition
have been investigated by tenperature cpendent 1ifetin studies (ref. 52,61,62).
In the solvent dependence stixy (ref . 61) the 3t'wr-'da transition was treated as an
intranclecular electron transfer process,
5
5 *1
(dir) (it ) —>(dur) (do-a)
3(ME.CP)
1
(20)
(cu)
with k°
Ea being the pre-exponential term and energy of activation for the ligand to
metal electron transfer,
k = k'0exp-QE'/RT) = A exp(E/RT)
(21)
Significant variations in k ' and E' are also observed in the series of cauplexes (cisRu(bpy) .,(L) ,] ere the n8nchraicphoric ligand L s*as varied (ref . 52) . Although there are
insuffiient data to predict Ix variations in L inflince the relative energies of MECT and
&1 states, sane trends are discernible: 1) With appropriate variations in the surrounding
state
ligands , for example, by substituting 4 ,5-diazafltx,rene for Ipy the energy of the
can be selectively perturbed leaving the MLCP states relatively unaffected (ref. 63) . 2)
or CD, the L'Wr state is sthbi1ized relative
WIn L is a good backbonding ligand like
to the (11 state as evidenc kj lcz ergy barriers for the MWT —>dd transition, In
fact, the striking observation has been ziade that in the series (cis-Ru(bpy) (L) I L+ the
quantun yield for ligand loss increases linearly with the MLC enission eneriy (ef . 64).
T appearance of relatively lci lying &1 states in polyridyl xznplexes of ruthenitin,
although of saie fundamental interest, is a nuisance in the attanpt to design nz classes of
photosensitizers. The dd states2re short lived in solution and of little value in a
n '2 at roan tperature, 98% of excited state
cthenical sense. Ftr (Ru(en)
decay occurs through the ãl stte; 25% !or [Ru(bpy) .] (ref. 58). In addition, the dd
states represent a potentially najor source of oto1nstability. In the series (cis—
Ru(Iy) (L) ] L+ (Lpy, N-methylimidazole,...) quantum yields for ligand loss of >i have
been obervd (ref. 59,60). The presence of lcwlying dd states virtually rules out the
possibility of developing a useful family of MLCr-based photosensitizers based on iron. It
seams deubtful that any crznbination of ligands on iron will give both the necessary visible
light absorptivity and sufficient destabilization of the21 states to raise than above the
MIJCr states. Fbr example, MLCP excitation of EFe(bpy)3] leads to the rapid population and
decay (T =0. 6ns) of a lowlying &1 state (ref. 65).
There are approadies available for designing "de-free" MLCP sensitizers. The Irost
straightforward is to turn to Os where 10 Dq is higher by 30% and, in general, ccznplexes of
Os(II) are photoinert. Even with Os(II), &1 states can appear induced by: 1) strong
backbonding ligands like CD which destablize MtCT relative to dd, and 2) by decreasing the
syninetry of the oanplex in such a way as to lift the orthogonality of the d and th*
orbitals (ref. 66). The net effect is to "mix" the d and th* orbitals thus decreasing the
energy of the dd states. It has been suggested that the lowered syninetry at Ru in
(Ru(trgj) 12+ (trpy = 2,2' (6,6' ) ,2"—terpyridine) arising fran the inability of the
tridentat trpy ligand to span 180° leads to a lowlying dd state and a very short lifetime
at roan tenperature (ref. 67).
In
()y) ]2+ (n=l ,2; L-L is 2,2' -}yrazine or 2,2' -bis—
the mixed dielates (Ru(L—L)
isoquinole) (ref. 43b,68), pfIo€ochencal ligand loss is quenched for the ccxnplexes (Ru(Lin part because of the lower lying ir levels at (L-L). Studies of this kind
L)(bpy).,]
show pr&ise for the design of new classes of photochanically stable sensitizers based on Ru.
The fourth MLCT state
neted in a previous section, even in the absence of photocamical ligand loss, lifetime
studies as a function of tenperature reveal the existence of an additional thermally
activated process or processes for excited state decay of Os(II) (ref. 41,69) and Ru(II)
oanplexes (ref. 43,68,70).
A contribetion to the tenperature dependence cares fran thermal population and decay rf an a
fourth Mr_Cr (ref. 68) which on theoretical grounds is predicted to have a greater arcunt of
singlet character and be mere short lived than the three lowest MtCr0tates (ref. 44). With
this interpretation ' is the energy gap to the fourth state and k the rate constant for
decay of the fourth state. In contrast to the 3twr - dd transition where the parameters
k0' and AE' are characteristically in the range l&2 _1014 s1 and 2500-4000 air1, for the
As
Photochemistry ofmetal coordination complexes
1201
l are generally in the range i7-i s-1 a 400—bOO
nonhotochenica1 cases, k0'
au_ , respectively. As yet, there are insufficient data to estabbish the facthrs at the
nobecular leveb that dictate the nugnitude of the energy spbitting between the fourth and
1c'zest three MICT states and the situation is further cboind tij the fact that fits of
experilTtentab data to Eq. b9 rcessarily force the tnperature pendence into a singbe term.
In sane cases contributions to the experimental data nay cane fran population and cay of
both 1 and the fourth MLCr states.
T
availabibity of an ectended series of related ocnpbexes based on the sane Eranophore,
e.g.,
Radiative decay of MLCT states
((ky)Os(L)]2, offers the possibibity of exploring sane of the nost fundanentab
processes
stndies based on polypyridyl cxxnpbexes of 05(11) rather than Ru(II) . Although the
avaibable to excited states. cperinenthb1y, it is far simpler to rry xit such
properties of the MLCT excited states are entirely anabogous, it is necessary to .rry it
tnperature dapendent sttxIies for Ru in order to separate out contri}xitions to iuiradiative
dacay fran population and decay of
states.
In a radiative transition energy conservation is achieved 1q nission of a thothn and fran
Eq. 9 the nagnitude of k is proportionab to the square of the transition dipobe nanent and
to E3 . Ecperiinentab1y, radiative (k ) and ixrnradiative (k ) decay rate constants are
dateRined b7 a xinbination of bifeti& ft ) and enission quantun yiebd
b/T0=k+k;=kT0
(22)
expression for 4 in E. 22 assunes that the efficiency of formation of the enitting
state follaiing excitation is unity.
For the majority of MLC excited states whidi have been studied 4) is lcz (<10% at roan
T1-e
temperature) and difficult to neasure accurately hicth limits theeprecision Of kr.
Nonetheless, cxirariscns anongst a series of related ccmplexes bike ((Icy)Os(L)A] 2+ sha the
existence of sane systematic trends: 1) Although a considerable scetter appear in the
data, kr increases roughly with <v 1>-i as predicted kxr Eq. 9 (ref. 46,71). The linear
increase shcMs that the transition dipole nanent must remain relatively constant through the
series. 2) There is on electronic distinction between }py and phen as cthranophoric ligands
(ref. 46) which is consistent with the results of nolecular orbital calculations which sha,z
that the % N character and energies of the lcwest lying * orbitals are similar (ref. 48).
3) Th_transition dipole narent integral can be evaluated f ran the slope of the plot of kr
<v 3>1 and fran it an estimate of the rrolar extinction coefficient?f th? absorption
band corresponding to emission can be obtained which for Os ( 4OO M an ) is in
reasonable agreement with valtes timated f ran the lcw temperire single crystal
(ref. 23).
absorption spectrum of [Os(bpy)3]
vs •
In order for a nonradiative decay process to occur with energy
conservation, the energy change associated with the change in electronic configuration
between the ground and excited states must appear in the surrounding vibrations. The
problem in excited state decay is energy disposal and only those vibrations for which Q
or w are different between the ground and excited states can accept the released energy?
The rate constant k is the product of two factors (ref. 72-75): 1) A vibrationally
induced electronic upling term. If the Born-Oppenheimer approximation is assumed the to
be valid, ground and excited states are orthogonal solutions of the same Hamiltonian and can
not imx. In fact, nolecular vibrations ck) lead to slight perturbations in electronic
structure. Fr selected vibrations (the "pranoting" nodes) one consequence of the
perturbation that they exert is a mixing of the excited and ground states which allcws the
transition between states to occur. 2) Vibrational overlap or Franck-Condon integrals. The
fractional distribution of energy released throughout the "acceptor" nodes for which LQ? or
pO depends on the magnitudes of the Franck-Condon overlap integrals (Eq. 14) subject to
the constraints imposed by energy conservation.
Nonradiative decay
In the limit of a single pranoting node or of an average contribution fran a series of
nodes, the vibrationally induced electronic coupling term is given by
=
Cuj(ls)(Tt/2)'2/(bOOO an1)
(23)
In Eq. 23
is the angular frequency of the pranoting node and the vibronic integral C is
defined in Eq. 24 (ref. 46).
=
h<'y.
In Eq.
wave
24
(hQk) Ih1f> = Vk( Whw)312
mcde,and
of the electronic wave functions and the vibrational
'Y = jx, Qk is the displacement for the pranoting
contains only non-Born-Oppenheimer
V f> = <1
and tYf are the products
for the prcaioting node,
function
the integral Vk =
operators.
<1
(24)
T.J.MEYER
1202
The vibrational
overlap integrals scribe Ii the energy is initially partitioned atongst
the various intrac1ecu1ar ncdes then the excited to ground state transition occurs. Fbr
the MLCr excited states radiative and nonradiative decay occur fran the sane states so that
fran the Resonance Ranan and nission spectral fitting studies trentioned earlier
oontrilxitions must exist frau 7-8 nediun freqi.ncy polyTridyl ring stretching vibrations,
lcw freqncy ring and netal ligand stretching vibrations, and the solvent. The
c,ontrilxitions fran niiu and li freqtncy ns .n be treated as averages through SM,hw,
and SL, 1ü)T Obtfld by nission spectra]. fitting. If for the energy gap between t1
excited ana ground states, E >> SJi, arx if h >> kT( 200 kcal/nole at RT) , it is
possible to evaluate the Franck—thidon integrals, Eq. t4, whicth gives for k (ref.
46,79,81).
k
Y
130.F =
8' (lftwE)"2Xp(-S) exp(_(YEO,41wM) + (y+l)2(Av/shwM)2/l6ln2]
(25)
(26)
= ln(E/Sh)—1
25 incltres contributions fran: 1) A thninant acceptor node (S , hw ) .
. In the classic1 liiit
freqincy intranclecular and solvent nodes through
Equation
given k:r,
=
(16kTln2(x0+X) ]
2) Lo,z
v is
=
[23lO(X0+X)](at RP in l)
In Eq. 27 x0 is the cxntrilxttion fran solvent and x =
(27)
x, is the
,
contribution fran the intranclecular vibrations; the surnation is over all of the
contributing 1c freqincy ncs with the prime noting that the higher freqtncy nodes are
exc]ed. 3) An energy gap term E whicth is related to the 0-0 iergy difference between
ground and excited states ,
E0E,-X-x
E, hj
(28)
E0 is the energy at which the vO - v =0 transition occurs in the euission spectrum for
pectra1 fitting.
mode M and cn be obtain1 by ission
the
It is always at higher energy than
apparent enission energy naxiinun in a structureless anission ianifold.
Equations like 25 predict that
ink
should vary approximately
linearly
with the energy gap,
have been referred to as "the ergy gap law" (ref. 72-75). In the limit E > SMhWM
energy dissipation through WM is kininant and the energy gap tezm, (E/iw) - v, dictates
in turn the vibrational quantzn number of the particular vibrational level which thainates
and
increases as decreases because vibrational overlap is
the energy acceptor role.
enhanced for na1ler valuas of v . Although vibrational levels with high1valts of vM ou1d
be required for a single acceptor node, e.g., v = 10 for E = 15,000 an and hwM =
1500 an , in fact, with 7—8 vUpy) nodes availble, the deSay process is thninated Iiy
transitions in which energy is dissipated into ocxnbinations of the v(bpy) nodes each of
which
has a relatively snail valuc of v. Contributions fran law frequancy nodes and the
in the /v term. Although they play a role in excited
solvent are included classically
state decay, the fraction of energy
dissipa€ed through these nodes is much snaller (< 10%).
The linear relation between ink and the energy gap predicted by the energy gap law has
been observed for families of aP&natic hydrocarbons (ref. 73,76,77). Hawever, perhaps the
nost systsnatically revealing cases have ocire fran studies of rx)nradiative decay in a series
of MLC excited states of Ru(II), Os(II), and Re(I) where linear correlations between ink
and F, have been observed: 1) fo families of related ociaplexes like [(phen)Os(L) j2+
(ref. 46,78), or ((tpy)Re(CO)3(L)] (ref. 79) (L = py,
, FCN, ...) where the basis for
the chrasophore is naintained and the energy gap varied b changes in L. 2) for
(Os(phen) 2+ and a related derivative, by varying th counter ion in dichiranethane
solution ref. 80). 3) for the family [Ru(bpy) (L) ] (ref. 43,52) or for (Ru(y) I 2+ in
a series of solvents (ref. 61) where tsnperatur de$endent studies ware used to diseitangie
contributions to excited state decay fran MECT and &i states. 4) for the families
(Upy)Os(L)4 2+ or [(phen)Os(L) 2+ where the energy gapwas changed by solvent variations
(ref. 81). 5) for (os(bpy) (a(py) ]2+ in the glass to fluid transition region in 4 :i(V/Si)
EtCU/B (90-140K) where tfie decrease in anission energy with tanperature is parailelled by
increases in lnkor (ref. 82).
In the energy gap law experiments based on MLCP excited state of Os(II) xx)t only are linear
relatioshi2 observed between lnk and the anission energy but the slopes (U) (lnkur)/)Ean=
-0.9x10 an ) and intercepts (27-30) are searly the same. This is a significant
observation since it suggests that the thranophoric basis for the excited states renains the
same throughout the series without significant variations in the electronic coupling term
B , that the same pattern of anceptor vibrations is involved throughout the series, and that
tRe energy gap law holds for MLCr decay regardless of Ii the energy gap is varied at the
microscopic level.
Nonetheless, the observation of a linear relationship between ink and Ean or E only
thnonstrates a qualitative agreanent with the underlying theory. "quation 25 is0not a
Photochemistry ofmetal coordination complexes
simple function of E as shc%7n kr the appearance of the term containing (E0)½
1203
the
dependence of S1 on E,. Consequently, there is rx quantitative information in the slopes
and intercepts df simple plots of lnk vs. E0.
A far sore insightful approach is to use the ocraplete results of enission spectral fitting
and the peranters S, tiw, Er,, ½ c]ath the vibrational overlap term F in
Eq. 25 . Based ai sudi calculations for a series of tpy and çhen complexes of (II) and the
relationship, ink = ThB + lnF fran Eq. 25 , a rnarkably ctailed insight is gain1 into
the xxnradiative cay process (ref . 46) : 1) TI najor oontributor (>90%) to the
vibrational overlap t.erm F ccxies fran the nediizn freqincy v(Içy) nodes and the term
-yE "M' 2) The slope of a plot of lnk vs . lnF is unity as predicted by Eq. 25. 3) Both
I:y°and phen as dircitcphoric ligands lie on the sane plot of lnk vs. lnF shczing that the
electronic cxupling term 13 is the sane for both consistent with the results of the
nclecular orbital calculations alhxied to earlier (ref . 44 ) . The increase by a factor of 3
in kl,T. for tpy onpared to then in analogous cxxnexes, arises because of systenatically
sraller vallEs of E arid larger valts of SM. 4) F E. 25 the intercept of a plot of
lnk vs . lnF gives°1n80 34 , 8 6y1lOl4l pi fran Eq. 24 , Vk 1300 an-. A vali. of
thiWrragnitixe for V is in line°with estimates mades fran absorption band intensities for
singlet9singlet traftsltions which are dipole forbidden het vibronically allawed (ref . 83).
The near cxnstancy in B implied by the linear correlation between ink and F suggests that
the pattern cit k*ilying°MLCr states must ratain rarly the sane througRut tue series of
diratcphores.
The quantitative success of Eq. 25 is quite striking and its implications profound. As a
result of the analysis once 8 is evaluated for a particular class of thranophore,
can
be Ca1CU1ate to within
Tflce
like
a facor of 3 nission spectral fitting. The situation is much
that for EEe appearance of netalto netal charge transfer bands in mixed-valence diners
[(y)Ru1I(4,4'_birridine)RuCl()çrj)]3+ where absorption band spectral profiles
been used to estimate rate constants for itranclecular electron transfer (ref . lOa,84In either case the spectral profiles for the appropriate spectroscopic process,
(ty)Ru" .- bpyRu11 or Ru IL III Ru11-pji II, contain all the needed information
required to calculate the vibrational contri}xitions to the electron transfer or nonradiative
have
85 ) .
cay rate constants.
Medium effects
Fran Eq. 25 the solvent
plays a minor role as energy acceptor in the
initial non-radiative decay process. Fllawing the transition between states and subseqnt
vibrational relaxation, the energy &es ultimately end up in the surrounding solvent thermal
pool.
The solvent also plays a role in determining the magnitude of k through the energy gap
term yE /lw which dictates the pattern of v(tpy or en) levels which ninate as energy
accepto?s. MFr [Ruthpy) ]2+ or (Os(bpy) ]2+ where a dipole is created in the excited state,
the excited state is preferentially stabilized relative to the ground state by the
surrounding solvent dipoles as shn by the fact that E tends to decrease as the static
dielectric constant of the solvent increases (ref. 86) . As noted above, the changes in F0
are tracked by lnk. as predicted by the energy gap law. Hcever, the microscopic origins
of the solvent shifts observed in F,3 are not understood (ref. 81,86) and dielectric
continuum theory, which àes provide a basis for accounting for solvent shifts for MLCT
absorption bands (ref. 39), is not successful for E (ref. 86).
Experimental evidence is available to suggest that in hydrcxylic solvents, especially H 0 or
CE CE, high frequnncy v (0-H) nodes at 3500 atr1 can also play a role as energy acceptos in
noradiative decay (ref. 81,86). The evidence includes the fact that for the family of
c,cinplexes (Os(phen)(L) A ]2+ in both polar organic and hydroxylic solvents, plots of lnk vs.
shari that excited tate decay is always faster in the hydrcxylic solvents and that a
significant k120/kD o kinetic isotope effect exists (2). Even though S is anall for such
vibations, they nee a contribution because of their relatively high frequnncies (-P3500
an ) (ref. 86).
The anission and lifetime characteristics of MLCP excited states, must notably for
(Ru(bpy) 2+, have been investigated in a variety of media ranging fran crystals to glasses
and polyzeric films (ref. 34,41,70,87—90). In such media where dipole reorientations in the
surrounding medium are restricted and different chanical envirorzrents nay exist for the
excited state, non-exponential decay kinetics are often observed and intranolecular
processes such as the MLCP-+dd transition can be inhibited. An important contributor to
such effects is, no deubt, the fact that in a hard glass, polymeric film, or the solid
state, dipole reorientation times are long on the tine scale for excited state decay.
Dipole orientations surrounding the nolecule are nocessarily constrained to be those
appropriate to the electronic configuration of the ground state. By contrast, in solution
solvent dipole reorientation times are shorter than excited state lifetimes for all except
the shortest lived excited states. ¶[1e solvent dipoles have the opportunity to reorient to
configurations appropriate to the electronic configuration of the excited state before it
decays. Consequnntly the excited state is stabilized in fluid solution ccinpared to a rigid
medium.
T.J.MEYER
1204
In glasses, in the inteimliate tuperature range between the hard glass and fluid states,
interesting tine dependent enanena cn be observed as the correlation times for dipole
reorientations approach the thne scale for excited state decay (ref . 91-92) . In this dcxnain
a xznplex time dependence for eccited state decay is predicted since as solvent dipoles
reorient a tiri pendence is introduced into the energy gap, F=F(t) leading to a tiii
dependence in k, lnk(t) - yE0(t)/h, Eq. 25 . In this limit ireasurnents taken at a
single wave length give rise to rxriexponential decay kinetics it dipole reorientation and
excited state decay n be convoluted using the full nission spectrum as a function of
time. Based cii the deconvolution, excited state decay neasurnts c.n be used to probe the
dynamics of dipole reorientation as the glass softens.
SYNTHETIC DESIGN OF EXCITED STATE PROPERTIES
Fran
the synthetic point of view, the availability of MLC thrancphores based on variations
in the c*irarcphoric and xxn—dirctrcphoric ligands (ref. 43,60,93—96) is quite ranarkable
since it offers the possibility of "designer" excited states where otophysical and
photocheitical properties n be controlled in a systanatic way by synthetic variations. Ps
a result of the otochuical and çtiotcphysical stndies described here and the ranarkably
consistent patterns that nerge, it is possible to formulate a set of "rules" that cn both
account for excited state properties and guide the sign of nz cxinplexes . 1) Radiative
cxxnplexes of Os(II) and Ru(II) having the
decay rates vary approximately with 3 .
sane enission energies , kr for Os wilr be greater by 3 because of tk higher spin-orbit
coupling constant for Os. 2) Nonradiative decay is &ininated by the energy accepting zole
of a series of polypyridjne-based ring stretching vibrations. Fbr equivalent by and en
complexes the py ootnplexes undergo nore rapid son-radiative cay by a factor of -3 because
of naller valuss of F. (-F) and larger valuss of S. Non-radiative decay rates are more
rapid in H.p than in polar organic solvents because of contributions fran v (0-H) nodes as
energy accptors. 3) 1 anission energy is the deciding factor in determining excited
state properties (ref. 14a). In a family of related cxxnplexes where <d> and 13 ranain
nearly constant, k varies with 3, k with exp(-E) and +e with kr and k°. 4) Excitedstate redox potentials can also be estited fran the Enission energy using ground state
redox potentials as shcin in Eq. 29.
r
o'
E
2+*/+
[(phen)Os(L)4
] = Eo' [(phen)Os(L)42+/+I
+
E
E0((en)0s(L)43+#/2+] = E0((pen)0s(L)4)3+'2
5) Since
the
-
(29a)
E
(29b)
is
relatively insensitive to variations
ligand-based [(phen)9s(L)4]21'+ couple
with °
the M(III/II)-based couple. As a consequence, in a
closelyelated series of ccirlexes, excited
emission energies, lifetines, and
emission
estimated simply
rreasuraient of the potential of the ground
state 05(111/Il) couple (ref. 14a). 6) In cctnplexes of Ru(II) ccznplications exist fran
lailying dd states and unravelling their contributions to excited state properties demands
linearly
effencies can all be
in L, E varies
for
state
temperature dependent lifetine studies. As yet, there are insufficient data and an absence
of the theoretical insight needed to sake clear predictions about I relative energies of
the MIL'T and di states vary with variations in ligands. Ha'zever, in certain mixed cthelates
containing at least one good polyridyl-based ir-accepting ligand, the energy gap between
the MLCP and IF states can be sufficiently large that the population of dd states is
relatively insignificant at roan temperature. Photochamical ligand loss is inhibited for
dielates where the dielate can reclose follcwing dissociation and is favored in polar
organic solvents in the presence of coordinating anions like Cl which can capture open
coordination sites.
Acknowledgement
has
I weuld like to acknazledge support of n n work in this area by the
Energy and of the contributions of J.
provided much of the basis for this article.
U.S. Departnent
of
V. Caspar and E.
M. Kober whose work
REFERENCES
1. T. P. Paris and W. W. Brandt, J. Am. then Soc., 81, 5001 (1959).
2. a) H. D. Gafney and A. W. Adainson, J.Am. then Soc., 94, 8238 (1972); b) T. N.
Danas and A. W. Adenson, J. Am. then. Soc., 95, 5159 1L973).
3. a) C. R. Bock, T. J. Meyer and D. G. Whitten, J. Am. Chen. Soc., 96, 4710 (1974);
b) R. C. Young, T. J. Meyer and D. G. Whitten, J. Am. Chen. Soc. ,7, 781 (1975).
4.
5.
G. Navon and N. Sutin, Inorg. then., 13, 2159 (1974).
a) C. P. Anderson, D. J. Salnon, T. J. Meyer and R. C. Young, J. Am. then. Soc.,
99, 1980 (1977); b) M. Mstri and M. GrAtzel, Ber. Bunsenges. Phys. then., 81, 504
—
(1977).
R. Bock, T. J. Meyer and D. G. Whitten, J. Am. Chen. Soc., 97, 2909 (1975).
a) C. R. Bock, .J. A. Conner, A. R. Gutierrez, T. J. Meyer, D. G. Whitten, B. P.
Sullivan and J. K. Nagle, J. Am. Chen. Soc., 101, 4815 (1979); then. Phys. Lett., —
61, 522
6. C.
7.
(1979); b) R. Ballardini, G. Varani, M. T. Irr1i, F. Scandola and V. Baizani,
J. Am. then. Soc., 100, 7219 (1978).
8. a) N. Sutin and C. Creutz, Pure Appi. Chen., 52, 2717 (1980); b) N. Sutin,
J.
Photothen.,
10, 19 (1979).
Photochemistry of metal coordination complexes
9.
a) V. Baizani, F. Boileta, M.
T. Gandolfi arid M. Mstri ,
1p.
1205
Curr. then, • 75
,
1
(1978) ; b) K. Kalyanasundaram, Coord. then. Rev. , 46, 159 (1982) .
10. a) T. J. Meyer, Acc. then. Res. , 11, 94 (1978) ; T. J. Meyer, Prog. Inorg. then.,
(1983); C) D. G. Whitten, Acct. then. Res., 13, 83 (1980).
11. D. Reliit and A. Weller, Ber. Bunsenges. Phys. then. , 73, 834 (1969) ; Isr. J. then.,
i,
8, 259 (1970).
12. a) G. A. crosby, Acc. then. Res., 8, 231 (1975); b) M. K. DeAnond and C. M.
Carlin, Coord. then. Rev. , 36, 325 (1981) ; c) T. J. KeLp, Prog. Read. Kinetics, 10, 301
(1980) ; d) J. Frguson, F. Barren, E. R. Krauss, M. Maeder and J. Vrbancich,
Coord. then. Rev., 64, 21 (1985).
13. a) T. Saji and S. Aoyaqui, J. EleCtrOanal. Chen., 60,
14
.
1 (1975); 58, 401 (1975); b)
V. Sharda, M. J. Cook, A. P. Lads, G. S. G. McAuliffe, A. J. Thanson, and D. J. Rothins,
J. C. S. Perkin Trans. II, 1309 (1984).
a) E. M. Kober , J. L. Marshall , W. J. Dressick , B. P. Suflivan, J. V. thspar and
T. J. Mayer, Inorg. then., 24, 2755 (1985); b) E. S. Dodsworth and A. B. P. Lever,
then. Phys. Lett. , 119, 61 (1985).
Solaxn, J. Mt. Chen. Soc. , 102, 4085 (1980).
A. Gamier and J. tnon, Inorg. thim. Acta. , 21, 119 (1977).
to Ligand Fields, McGraw-Hill, New York (1962).
A. B. P. Lever, Inorganic Electronic Spectrospy ( 2nd ed. ) , Elsevier , New York
15. R. B. Wilson arxl E. I.
16. D. Guenzbirger,
17. C. J. Ballhausen, Introduction
18 .
19 .
(1984).
a) P. Belser , A. Von
20.
J. Ptiys.
a) T. thno and S. Kato, Bull. Chen. Soc. JprL, 57, 3391 (1984); b) V. T. Cocinbe,
mcci
F.
and V. Balzani,
Ze1esky, A. Juris , F.
101 (1982) ; b) F. Barigelletti, P. Belser, A. von Zelesky and V.
then. Phys. Lett. ,
Balzani,
then., 89, 3680 (1985).
Barigeletti ,
Yellc1ees,
G.
A. Heath, A. J. MacKenzie and L. J.
Inorg. then. , 23, 3423 (1984).
M. Nishizawa, T. M. Suuk, S. Sprouse, R. J. %tth and P. C. Pbr Inorg. Chen.
23, 1037 (1984).
22. T G. M. Bryant, J. E. rgusson and J. K. J. Pciie1l, Aust.
then. , 24, 257
(1971) ; G. M. Bryant and J. E. Fergusson, Aust. J. then. , 24, 275 (197U.
23. a) F. Felix, J. Ferguson, H. U. Gixiel and A. Lndi,
Ferguson, H. U. Gixiel and A.
Am. then. Soc., 102,
(1979);
21 .
J.
then. Phys. Lett., 62, 153
Lndi, J.
4096
(1980); 5. Decurtins, F. Felix, J. Ferguson, H. U. Gndel and A. Lisli, J. Am. then. Soc.,
b) F. Felix, J.
102, 4102 (1980).
24. 15i. Kober
and T. J. Meyer, Inorg. then., 21, 3967 (1982).
then.
25. J. Ferguson and F. Barren,
Phys., 76,T5 (1983).
26. A. D. Kirk, P. E. Haggard, G. B. Porter, M. G. Rockley and M. W. Windsor,
199
(1976).
then. Phys. Lett., 37,
18, 3177 (1979).
27. J. N. Denas and D. G. Taylor, Inorg.
28. P. G. Bradley, N. Kress, B. A. Hornberger, R. F. Dallinger and W. H. Woodruff,
then.,
J.
29.
Am. then. Soc., 103, 7441 (1981).
a) M. Pbrster and R. Hester, Chen. Phys. Lett., 81, 42 (1981); b) W. K.
Snothers and M. S. Wrighton, J. Am. then. Soc., 105, 1067 (1983).
30. C. M. Carlin and M. K. DeArnond, then. Phys. Lett., 89, 297 (1982);
J. Am. Chen. Soc., 107, 53 (1985).
31. J. V. Caspar, T. D. Westmoreland, G. H. Allen, P. G. Bradley, T. J. Meyer and
H. Woodruff, J. Am. then. Soc., 106, 3492 (1984).
32. G. A. Heath, L. J. Yellciilees an. S. Braterman, J. C. S. then. Ccxan., 287
W.
(1981); then. Phys. Lett., 92, 66 (1982).
33. D. E. Morris, K. W. Hanck a M. K. DeArmond, J. Am. then. Soc., 103, 3032 (1983).
34. a) H. Yersin and E. Gallhuber, J. Am. Chen. Soc., 106, 6582 (1984); b) J. Ferguson
and E. R. Krausz, then. Phys. Lett., 93, 21 (1983).
35. a) R. F. Dallinger, Inorg. then., in press; b) S. J. Milder, J. S. Gold and D. S.
Kliger, J. Phys. then., 90, 548 (1986).
36. a) C. Creutz, Prog. Inorg. then., 30, 1 (1983); b) K. Y. Wang and P. N. Sdatz,
Prog. Inorg. Chen., 28 369 (1981); c) N. S. Hush, Prog. Inorg. Chen., 8, 391 (1967);
Electrochim. Acta, 13, 1005 (1968).
37. P. 5. Braterman, G. A. Heath, L. J. Yellc'zless, J. C. S. Dalton, 1081 (1985).
38. R. A. Marcus, J. then. Phys., 43, 1261 (1965); J. Chen. Phys., 39, 1734 (1963).
39. E. M. Kober, B. P. Sullivan and T. J. Meyer, Inorg. Chen., 23, 2098 (1984).
40. a) D. E. Lacky, B. J. Pankudi and G. A. Crosby, J. Phys. then., 84, 2068 (1980);
G. D. Hager and G. A. Crosby, J. Am. then. Soc., 97, 7031 (1975); G. D. Hager, R. J.
Watts and G. A. Crosby, J. Am. then. Soc., 97, 707 (1975).
41. 5. R. Allscpp, A. Cac, T. J. Ke, W. J. Reed, V. Carassiti and 0. Traverso,
then. Soc. Faraday Trans. 1, 75, 353 (1979).
42. a) R. S. Lu*in and T. J. Meyer, J. Phys. then., in press; b) E. G. Megehee and
T. J. Meyer, in preparation.
43. a) G. H. Allen, R. P. White, D. P. Rillena and T. J. Meyer, J. Am. then. Soc.,
106, 2613 (1984); b) F. Barigeletti, F. Juris, V. Balzani, P. Belser and A. von Ze].eisky,
Inorg. then., 22, 3335 (1983).
44. E. M. Kober anT. J. Meyer, Inorg. then., 23, 3877 (1984).
45. E. M. Kober and T. J. Meyer, in preparation.
46. E. M. Kober, J. V. Casper, R. S. Liznpkin and T. J. Meyer, J. Phys. then., in press.
47. a) M. H. thisho]in, J. C. Huffman, I. P. Rothwell, P. G. Bradley, N. Kress and W.
Am. then. Soc., 103, 4945 (1981); b) L. J. Padonovich, M. W. Eyring, T.
H. Woodruff,
Groshens and K. J. Klabende, J. Am. Chen. Soc., 104, 2816 (1982).
48. E. M. Kober and T. J. Meyer, Inorg. then., 24, 106 (1985).
49. 0. Poizet and C. Sourisseau, J. Phys. then., 88, 3007 (1984).
50. H. Yersin, H. Otto, J. I. Zink and G. Glienann, J. Am. Chem. Soc., 102, 951
(1980); G. Herzberg, Molecular Spectra and Molecular Structures.
J.
J.
J.
I. Spectra of Diatonic Molecules (2nd Ed.), . 20 and 200, Van Nostrand Reinhold, N
51. C. J. Ballhausen, Molecular Electronic Structures of Transition Metal Ccznplexes,
York (1950).
Chap. 4,
McGraw-Hill,
Nerz York (1979).
'
9OL
'
'A 15d553
'J '1'
PPP '(861)
'blOul "1Eq3
'ç (5'lsdsg)
'2 'A 'isdss 'ci'qa As1a4p1_'UOflS1SBSTU 0 41ON 'uttoiv 861 '2 'A
'U 'S 'uptduzvi 'H 'N
'J '2 a15 I8
'2
'p
UA uanoa
'r
'st
'5043 "309
5(8L61) ___________
'bloul "5043
_______________
T8E '(8L61)
pUS
'2
ISciON PUS
'WV
pus 'U
'2
'LI
'55 'N
'15412W
'ouw
'vr1sID 'a
'D osTbbsa PUS
'(8L61)
'H pisb6oa pus
(S
(Cl
UT
aLT
'a '5flb0xOS _________________________
'2 '5043 '305 'uWqO "uflUIlO3 59
'a
') '5 '1a10a '2 'WV '5043 '3°S '001 LSPI 5(8L61) (ci 'M 'N
TI5M PUS 'a 'H 'pis6fc 'bloul "5043 '61 IPW '(0861)
'LS 'S 'wSqlflU
"I 'qSIt '3 '[ ISVSD Pus 'J '2 'ieAs 'bloul "043 '61 098 '(0861)
'85 'S 'UOqlflU '2 'A '15d853 '2 'U 5T6N pUS 'L '2 '15412W _______________
'1 'WV '5043 "305 'POT £081'
'95
'1'
'(861)
'J
1P8 '(P861)
'65 'U 'A PTU50a PU 'S 'tuSqlflU 'bioul "5043
'09 'M 'N azT()43 'U 'V '43S415flV 'U 'H 'TqPS 'N OUTITO PU 'M 'U 'Axiep
'biouI 'U0t(3 'P 8SLI '(5861)
'19 '2 'A 111dS53 Pus 'L '2
'1' 'WV '5043 "OOS '501 £855 '(E861)
L2 '(9861)
'M 'a 'zIoW)SM 'U 'V qosqit US 'U 'H
"5043
'£9 "I '1
'12
'M 'U
'2 "uuiipooqa '8? £P1 '(ST)
OPPI '(P861)
'P9 'd 'A PTUUTa P5 'S 'WStjlflU 'bioui
a
'N 'uprtS _______________
'2 'WV 'Uq3 "305 '0I
'59 '3 'ZflaT3 'N 'flCX) 'L "I '(SZSN 'N S1W1tO0
'
a1p
aj
'
5
'5
a5
6OEI '(0861)
'99 'D 'USITV 'a
USâTUflS puS
'2 'la&aw '2 '3 'S 'uq3 "51103 £6L '(1861)
'WV 'uWq3 '309
'L9 '2 'U 'Oq31U 'U 'U aus.fl.)( 'a 'V OU1EW *15 '2 'a 'a6WuW5 '2
_______________
'a
'LOT
8U
'(S861)
'89 't) 'H
a 'U 'd 'eqM 'U 'd 1ITU pus ' '2
£T9 '(P861)
'69 (5 'U 'S uprii&i'i PU 'J '2 'ieAe UT SUOflVlVdSld
UT
'UOtS15dG1d
'OL 'S 'U 'dOsflN 'V 'WO
'(8L61)
'v 'a
'IL 'N '2
'sutqqOU '2 '3 'S
''
'?L
Aiuen PUs M
'JI
5
Saq8f( Pus
'H
(Cl
'2
'1S/2W
'I
'1' '3 'S Aspsisa
'pWa ________________
'PL SLIT
'2 duuiU pus 'M '2
a(
'
'2 'WV 'UE3 "305 '9rfl
'ieAuW _______________
'U '2
'S(flV3N 'A aspxSq5 'V '2
upiea 'suWXJ
£Ofl '(P861)
apu5c15S 3TUO ISIfl3SICWs35Aqc1OOqd 'bA) (I a 'ds'
'saqa 'IOA)
a uopuori5(L61) 'a 'U 5uqa a 55SIUOTSTPSUM '..sesseooza
(SI a 'bSIlSA-ISBUTLdS I'WN POA 5(9L61) 'a '5T10AV 'M 'N 1sq(s() PUS 'N 'V
uuiq3
'LL £6L 5(LL6T) 'U a
a.59 all PL 5(8L61) 'H
(
a.U
peto 'se'4s;S ovpy
3TUPV 'sd
'L
't'L
'
a
5(6L61)
6SL
(
'2
a. tqy 53 ) 'i
.q3
'a
P
'H u' pus
996fl
'S
eTsdTTuD
'H '1 ISSSSTID
5
'A '1dS5Z) 'H 'N
'18 'r 'A
a
'2
'i
a
'v
'fl
'U3
'd
'PTTM
TTSIIOA pus 11
a
iaqoa
'd
'r
a5
ap94S
'i'
r
a,
a
a
pu .j
Aqspo
a1ej 'uqD
P1b11
'sAqd
r
aUtq
p
p1fl
1T1M pu r •)I 'SUEW r
.
a
a93 s
.Q
LI'W (0861)
.2
'
'opqy
'(8961)
.
'scqd a a7 S6
a
'5'4
a68 S601 '(S861)
a ______________a
(
•(986t)
£6 A
'
'1
.1•1
iessi:sipc t*X •D
r TT1S
aj
IAeI. 'pe:TuITns
S6
96
'iip
.1'
scq a
____________
.
LIP £(6L6T)
'r
series 'cxi aT.
Staj'aürr4ua3
sz urTsous
(q
)j
.sicqd
H UWI
a
.1,1
•ai,
.
OI'S'_a99
Zfl13 xie
r r
r
'a •i
i
'ai
qO
'
a 'W
'11
•uoideid
8I'u(86T)
a
'
.
'LB 066 (E861)
(86T)
r_____________
sAqa •uq 'ii
a
'r
u
qN
' '06
88fl
pmz
p;m .y uo 'ANsI.TZ
' v sirtr' çee6pa30' d ies-(e'sexeTdurO
• s1add .:sui ;o sicqd. x1
ETU UI •ubcs 3OO
si ' 'Z 8L (P861) pm seouaie;a Uta18t4
6T 86T)
•r
d stij pm _'UOSuIXIL_•r_•v
upiej
301d
6• w.1
af
'sseid
S8T (S86T) .
ssnezi pu w 'iepw
'uosnb1e%
sAqd •Uq3 '
Z6
s upduivi we a r 'ieIW r SILjd '.tiq
Ssseid
'UOSTeTUQ
PU S enzvj ____________
1fltupI H
' •T
sAqd .UqD
'UI)I •A
16
(t'861)
'Ui
a
_______________
'(E861)
'Siqd
'r
9)
'
i's:
TL6 (8961)
I
q
2 'uq3
ajj EL
ieA
3TU10I :Azstuq3
'3'U '(86T)
affO(ST43
'a •
'r 'A
a •i s up5uwj pim •j
.j •y
.,
r iwi uo °os ' 'öY
i Y TTQ
PU r
•9 a.3J3
i$WfT1
68 2
06
'f
'd
'sUta
£61 '(6L61)
LIT £(086T) (q 'y at(9J.IaCet3
al,8
(t'L6T)
'D 'V ':.ThwaN ')i
aJ9 8SST
'S8 (P 'ai
'88
'
SSSTUOflSTPS '5UOTT5USlJ
a ____________
'sAqa a
a;o5
'U13a.sA
USATflT1S Pus 'J
'ai
w
.uq3
a '1'
'L8
'3
'_
a
"-
'UOS155 '2
'sAqa
'3 'U1 'UqD
H
•r
'r
16 '(861)
pU •L r 'ieiw '1' 'sAqd
'8
8 'i 'Sr upiuri'i
'q3e1q(y 'j' 'U3 a.sAqd
'98
ar
'H
'r 'sAqa 'u S6a.p
pu 'i' a19ç
ids
•uq
'd atUns•
iaqoj pu '
O9 '(86T)
'6L '1' 'A xsdsg) pus L
'08 'M 'r a5(flflA '1' 'A
'
'S
(1OA (0861) S 'H '2 19TTSH pus 'j
,
'9L 'D 'U
8L
•
a
'do 'I'V
3 aW
(E861)
96I (96I) 'D 'M
'D 'M UO5UOU PU 'U 'a
'Uui3 'sAqa a
'q3 "SAtjd '5
'r 'uq3 'sAqaa '8E L811 (E961) 'D 'M IOITCTOT pus •3 'V
__________
—
(PL61)
pus
•9fly
'UflH
a59())J
'SI 16S (96T) 'r 'a
'a 'a
pu5
'(5961)
i59 'sny
a
'st pusxqes pus u 'a 'suistluM 'uq3 a.s)qa aft £01' 5(L961) 'M p'qTS
pus 'U
a55J5(-fl.4
a.SA4,j a6?, 0981 (8961) (3 'S 'H arJ
a.sAqj app
'S s
'LL
a
'5
UO
"(000 'V
spttj' A
(186T)
y 'ias w
TUZT' .d
'IeU3M
•3 •s
'
•sueij;
'ii
'°Y ''9 SLTZ
A1911 •urrqD
iesTe ptm 'V uoi A(sMøTeZ _____________
j 3 sezg puv
H cj
'eu bioui 'iD
'TZ S801 (Z861)