42 Photochemistry and Photophysics of Metal Complexes

2.4. PHOTOREACTIONS OF COMPLEXES OF COBALT(III)

The photochemistry of complexes of cobalt(ill) shows many similarities to the
photochemistry of complexes of rhodium(ill) and iridium(ill). The qualitative
energy level diagram shown in Fig. 2.5 can be used to interpret the spectra for all
of these d 6 metal ions in Oh symmetry. The lAl g ground state comes from the t~g
electronic configuration, while the low-lying ligand field excited states
3Tl g, 3T2g, ITl g, and IT2g result from the t~ge~ configuration. In addition, there is a
ST2g state that arises from the t~gei configuration. This quintet state lies above the
3Tl g state for cobalt(ill) complexes in a strong ligand field, as well as for complexes
of rhodium(ill) and iridium(ill) where the ligand fields are expected to be greater
than for cobalt(IIl). For complexes such as Co(NH3)f that have a smaller ligand
field strength, this ordering of states may, however, be reversed. Some support for
this premise that the ordering of states may be ligand-dependent comes from the
observation that whereas Co(CN)~- shows a phosphorescence due to the
3Tlg ~ lAl g transition, CO(NH3)~+ is nonemissive.(6S)

2.4.1. Photosubstitution and Photoredox Reactions

The major photochemical pathway for cobalt(ill) complexes in solution is
substitution. By contrast with the complexes of rhodium(III) and iridium(ill), the
quantum yields for photosubstitution for cobalt(II1) complexes are low. This feature
is shown in the homologous ammine complexes M(NH3)~, where the substitution
quantum yields for arnmine substitution in rhodium(ill) and iridium(ill) complexes

1 Tt,(t2,5 e ,l)

5 T , ( l2,4
2 e, 2)

3T2,

3 Tt, (l2,5 e ,t)

Figure 2.5. Qualitative energy level diagram for a d 6 transition metal ion in an octahedral (Oh)
ligand field.
Complexes of First-Row Transition Metals 43

log e

300 400 500 nm

Figure 2.6. Electronic absorption spectrum of Co(NH3)~+

are in the 10-2 range; whereas, those for cobalt(1Il) ammines fall in the range
10-3_10-4.(66.67) The absorption spectrum of the complex CO(NH3)t' shows absorp-
tion bands in the 200-500-nm range (Fig. 2.6). Upon irradiation into the ligand
field band at 365 nm, the complex undergoes photoaquation with the very low
quantum yield of -10-4 :
44 Photochemistry and Photophysics of Metal Complexes

(2.15)

The quantum yield at 472 run is approximately one order of magnitude lower,
verifying that the aquation quantum yield decreases with increasing wavelength in
the ligand field region.(68) By contrast, the photoaquation of the complex
Co(CN)~- with a stronger ligand field has a quantum yield that is much larger (0.31)
and is wavelength independent(69,70):

(2.16)

This difference can be explained on the basis of the excited states of the two sets
of complexes. Thus, while the lowest excited state of the hexacyano complexes
have triplet character, that of the hexaammine complex has quintet character. (71,72)
This difference results from the relative ligand field strengths of the ammine and
the cyanide ligands. The independence of the quantum yield with incident wave-
length for Co(CN)t in the ligand field region can be explained on the basis of an
intersystem crossing to the lowest wavelength triplet occurring with a quantum
yield of near unity. This premise has been tested by showing that the quantum yields
for photoaquation at 365 nm and 405 run are unchanged. In the former case the
pathway involves excitation to a singlet state, followed by intersystem crossing to
a triplet, whereas in the latter case the mechanism involves direct excitation
eTlg r IAlg) to the triplet.(73)
Irradiation at 254 run into the charge transfer band of the cobalt(Ill) complex
CO(NH3)r results in photoredox reactions to give cobalt(II) complexes. (68,74,75) The
quantum yield for irradiation into the charge transfer band at 254 nm is 0.16, which
is considerably larger than the corresponding photosubstitution quantum yield for
irradiation into the ligand field band. This quantum yield at 254 nm is independent
of the pH of the solution, and it does not depend on the concentration of either the
complex or a noncoordinating ion such as perchlorate. The quantum yield increases
slightly with temperature, and also increases when chloride ion is added to the
solution. The overall reaction is

(2.17)

The effect of added chloride is considered as due to the formation of an ion pair in
the solution. Such an ion pair is evidenced by the presence of a strong absorption
band at 250 nm. The increase in quantum yield can be explained if the ion pair
Co(NH3)r ... CI-, can more efficiently absorb light than can the free complex
CO(NH3)r.(68)
The quantum yield for the photoaquation reaction of Co(CN)~- can be de-
creased by the introduction of a supramolecular cation into the solution. When a
Complexes of First-Row Transition Metals 45

Figure 2.7. Structures of (a) 32-e.,-N~. (b) 32-NsHt". and (c) 24-N~.

small cationic linear alkyl ammonium ion such as the diethylammonium ion is
added to the solution, there is no significant decrease in quantum yield of the
photoaquation reaction. When, however, the macrocyclic alkylammonium cations
24-N6IIt, 32-C9'-N6Ht, and 32-NsHr (Fig. 2.7) are added to the solution, the
quantum yield drops from 0.31 to 0.15, 0.11, and 0.10 for these three adducts,
respectively.(76) In general, increasing the rigidity of the system results in an
increased quantum yield. If, however, it is assumed that the anions Co(CN)~- are
contained within the cavity of the cationic macrocycle, some of the cyanide ligands
will be bound to the polyammonium receptor, and will be thereby prevented from
escaping from the cage when the Co-CN bonds are cleaved in the excited state.
Photolysis of the mixed-ligand complex ions Co(CN)sxn- [where X is CN-.
N3-, cr, Br-, r, OH- (n =3) and H20 (n =2)] in the presence of thiocyanate ion
gives primarily the aquation product Co(CN)s(H20)2- [Eq. (2.18)], along with the
linkage isomers of the thiocyanate substitution product [Eq. (2.18)](77):

(2.18)

Co(CN)sXn- + SC~ ~ Co(CN)sSCN3- + Co(CN)sNCS 3- (2.19)

In Table 2.10 are shown the comparative data for photoaquation and photoanation,
which demonstrate that approximately 80% of the product is that resulting from the
photoaquation pathway. In Table 2.10 are also shown the ratio (SIN) of
Co(CN)sSCN3- to Co(CN)sNCS3- that is formed in both the thermal and the
46 Photochemistry and Photophysics of Metal Complexes

Table 2.10. Thermal and Photoanation Ratios ofthe S- and

N-Bonded Thiocyanate Isomer of Co(CN)sX3-
X SIN thermal SIN photochemical Aquation/anation
CI 4.3 ±O.2 8.5 ± 1.4 5.4±O.8
Br 4.5 8.8 ± 1.5 6.1±2.1
I 4.7 9.3± 1.3 7.5 ± 1.5
N3 4 8.7 ± 1.5 4.2± 1.1
OH 8.7
CN 8.1 ± 1.7

photochemical reaction for the different anions X-. From these data it is apparent
that the respective SIN thermal and photochemical ratios are 4.4 ± 0.4 and 8.7 ±
0.6 for the different anions studied. This constancy of the SIN ratio for the series
of anions X- used supports a pathway where the photoreaction is dissociative. Thus,
the photoaquation and photoanation reactions are in direct competition with each
other. The difference in the SIN ratio between the thermal and the photochemical
pathways shows that the photodissociated intermediate behaves differently from
the intermediate that is involved in the thermal reaction, with the photochemical
reaction resulting in a greater preference for the S-bonded isomer. No transients
have been observed in this reaction on the picosecond time scale, although low
temperature studies on Co(CN)~- (78,79) and cobalt(III) ammine complexes,(80,81)
have detected such species.
The favoring of S-coordination by the photochemical intermediate supports
the view that a species of greater polarizability is involved in the photochemical
reaction. Such a situation can arise with a 5-coordinate intermediate. If the photo-
chemical intermediate reacts with thiocyanate ion while still in the triplet excited
state, this t~get state can be expected to show some selectivity for sulfur coordina-
tion. Such a situation arises from the fact that the excited state will be more
polarizable than the ground state, and also that the frontier orbitals of the excited
state can extend further to react with the sulfur lone pair at longer range than the
equivalent lone pair on nitrogen.

2.4.2. Photoisomerization Reactions

Irridation of cis-CoCh(en)i in methanol solution into the ftrst charge transfer
band results in isomerization to trans-CoCh(en)i [Eq. (2.20)].(82) The reaction
pathway has been proposed to involve the photoreduction to complexes of co-
balt(II), with these complexes acting as catalysts for the isomerization reaction.
Complexes of First-Row Transition Metals 47

+ b + (2.20)
cis-CoCh(enh ~ trans-CoCh(en)2

Subsequently, the isomerization of cobalt(III) complexes has been explained on the

basis of the angular overlap model. (54,56,83) This model is based on the assumption
that the photosubstitution in the complex MA~Y occurs by loss ofY in the lowest
triplet state. This loss of Y leads to the formation of an intermediate 5-coordinate
structure where the interconversions between apical and equatorial positions are
separated by a low-activation barrier via the trigonal bipyramidal structure. When
X is a poorer a donor than A, and the metal-to-ligand n-bonding is minimal, the
trans structure is favored, and the barrier to cis-to-trans isomerization is low as
compared to the triplet ~ singlet deactivation rate constant. Thus, trans ~ cis
isomerization is expected when X is a better a donor than A in trans-M~Y.
Confirmation of this approach is shown in the photochemistry of trans-
Co(en)2(CN)CI+ and trans-Co(en)2(CN)H20 2+. Loss of chloride ion (¢J =10-2) from
Co(enh(CN)CI+ or water from trans-Co(en)2(CN)H20 2+ leads to the same square
pyramidal intermediate Co(en)2(CNf+, where the cyanide ligand is trans to the
vacant site. In each case the photolysis product is cis-Co(en)2(CN)H202+, which
supports the angular overlap model of excited state reactivity.
Photoinduced linkage isomerization of cobalt(III) ammine complexes has also
been observed when one of the ligands is ambidentate. These reactions are shown
in generalized form in Eq. (2.21). Two such examples that have received attention
are the photoinduced isomerization of the thiocyanate and the nitrite ligands [Eqs.
(2.22) and (2.23)]:

(2.21)

(2.22)

(2.23)

Under continuous photolysis, the complexes Co(NH3)sSCN2+ and

Co(NH3)sNO~+ are converted into Co(NH3)sNCS 2+ and Co(NH3)sON02+, re-
spectively. The other product in the solution is the aquated cobalt(II) ion that
arises from a photoredox reaction.(84.85) These photoisomerization reactions are
believed to proceed via a pathway where irradiation leads to a thermally
equilibrated LMCT excited state. In this excited state, the complex dissociates
to give the radical pair

(2.24)
48 Photochemistry and Photophysics of Metal Complexes

Geminate recombination of the Co(NH3)~+ and XY to give either Co(NH3)Sxy2+

or CO(NH3)SYX2+ competes with both aquation of the Co(NH3)~+ fragment, and
diffusion of the radicals into the bulk phase58S-87) For high selectivity in the
photoisomerization reaction, the reaction must occur with the two radicals held in
close proximity, or perhaps separated by a single solvent molecule.
For the thiocyanate complex the photoisomerization reaction of
Co(NH3)sSCN2+ to Co(NH3)sNCS2+ is not reversible, and Co(NH3)sNCS2+ does
not photoisomerize to Co(NH3)sSCN2+. If the CO(NH3)~+...SCN radical pair can
undergo secondary recombination to form eitherC0(NH3)sNCS2+orCo(NH3)sSCN2+,
but the Co(NH3)~+.. NCS radical pair can only form Co(NH3)sNCS2+, then it is neces-
sary to postulate that the secondary reactions of the radicals formed from excited
CO(NH3)sSCN2+ and Co(NH3)sNCS2+ differ in their properties. One explanation
for this difference is that the radical pairs have memories in the sense that they
undergo cage recombination to give the isomers from which they originated. The
experimental data also require that the radicals formed from Co(NH3)sSCN2+ and
Co(NH3)sNCS2+recombine in each case to give only CO(NH3)sNCS 2+. A plausible
explanation for this observation is that radical recombination proceeds in two steps.
In the first step a precursor complex is formed between CO(NH3)~+ and the
thiocyanate radical, and in the second step electron transfer occurs within the
precursor complex. The stability of this inner-sphere precursor complex is related
to the bonding in the thermally equilibrated LMCT excited state of the parent
complex. Since the LMCT excited state formed from Co(NH3)sNCS2+ is more
strongly bound than the LMCT excited state formed from Co(NH3)sSCN2+, the
electron transfer precursor complexes formed from identical radical pairs will favor
the more strongly bound N-bonded isomer rather than the more weakly bound
S-bonded precursor complex.(8S)

2.4.3. Photoreactions of Bimetallics

In order to effect the conversion of radiation energy into chemical energy,
multielectron processes will frequently be necessary. As a means to achieve such
processes, photoredox at a multiplicity of metal centers may be necessary. Such
complexes have been synthesized using a cyanide bridge from cobalt to other metal
centers. One of the earliest examples is the complex (NH3)SCOm(P-
NC)Rull(CN)s. This complex shows an intervalence transfer (IT) band due to a Rull
-+ Com transition at 375 nm, and irradiation into this band results in photoredox.(88)
Similar photoredox reactions are observed in complexes containing cyanide-
bridged COWoSll, C~Fell and FellOsm combinations.(89)
Intramolecular energy transfer also occurs between metal centers in bimetallic
complexes. Two examples are found in the complexes (CN)sCo(p-NC)C0(NH3)S
Complexes of First-Row Transition Metals 49

and (CN)sCo(,u-CN)Cr(NH3)s.(90·91) In each case ligand field excitation at one

metal center can lead to energy transfer into the other one.

2.4.4. Photohomolysis Reactions

Photohomolysis reactions of cobalt(llI) pseudohalide complexes can be used
to effect photoreduction to cobalt(II) complexes. Thus, intramolecular photoelec-
tron transfer in the complexes Co(CN)sN~- and Co(NH3)sN~+ leads to oxidation of
the azide ion to the azide radical, and reduction of the cobalt(llI) center to cobalt(II).
Evidence for the initial formation of the azide radical comes from the photolysis of
solutions containing Co(CN)sN~- and iodide ion, when the iodine anion radical I2
is observed in the solution. (92) This formation of I2 results from the photochemical
generation of the azide radical, which then oxidizes the iodide ion to an iodine atom
(Scheme 2.1). Subsequently, spin trapping experiments with phenyl-N-tert-butyl
nitrone has been used to verify the formation of azide radicals from the photolysis
of Co(CN)sN~- and Co(NH3)sN~+.(93)

2.5. PHOTOREACTIONS OF COMPLEXES OF COPPER(I)

2.5.1. Monomeric Complexes

Much of the recent interest in the photochemistry of copper(I) has centered on
complexes of the bpy type. Copper(I) with a dlO electron configuration generally
forms tetrahedral or pseudotetrahedral complexes that are kinetically labile in
solution. Copper(I) complexes with substituted phenanthrolines as ligands show
intense MLCT bands in the visible region of the electronic absorption spectrum.
For a series of complexes Cu(2,9-R2-phen)1 that have a range of different substi-
tuents R at the 2 and 9 positions of the phenanthroline, it is clear that the nature of
the group R in these positions has a strong influence on the properties of the excited
state. (94-96) When R is a phenyl group, the intensity of the charge transfer absorption
is decreased because delocalization of the n* -acceptor orbital over the phenyl

3 hv 3
CO(CN)SN3- ~ Co(CNh- + N3

Scheme 2.1.