New insight into photochemistry of ferrioxalate

Article abstract of DOI:10.1021/jp8040583

Optical spectroscopy and nanosecond flash photolysis (Nd:YAG laser, 355 nm, pulse duration 5 ns, mean energy 5 mJ/pulse) were used to study the photochemistry of Fe^m(C₂O₄)₃ ^3- complex in aqueous solutions. The main photochemical process was found to be intramolecular electron transfer from the ligand to Fe(III) ion with formation of a primary radical complex [(C₂O₄) ₂Fe^II(C₂O₄^.)] ^3-. The yield of radical species (i.e., CO₂^.- and C₂O₄^.- was found to be less than 6% of Fe^III(C₂O₄^.)^- disappeared after flash. [(C₂O₄)Fe^II(C₂O ₄^.)]^3- dissociates reversibly into oxalate ion and a secondary radical complex, [(C₂O₄)Fe ^II(C₂O₄^.)]^-. The latter reacts with the initial complex and dissociates to Fe^II(C ₂O₄) and oxalate radical. In this framework, the absorption spectra and rate constants of the reactions of all intermediates were determined.

Full text of DOI:10.1021/jp8040583

8316
J. Phys. Chem. A 2008, 112, 8316–8322
New Insight into Photochemistry of Ferrioxalate
Ivan P. Pozdnyakov,* Oksana V. Kel, Victor F. Plyusnin, Vjacheslav P. Grivin, and
Nikolai M. Bazhin
Institute of Chemical Kinetics and Combustion, Institutskaya 3, and NoVosibirsk State UniVersity, PirogoVa 2,
630090 NoVosibirsk, Russia
ReceiVed: May 8, 2008; ReVised Manuscript ReceiVed: June 20, 2008
Optical spectroscopy and nanosecond flash photolysis (Nd:YAG laser, 355 nm, pulse duration 5 ns, mean
energy 5 mJ/pulse) were used to study the photochemistry of FeIII(C2O4)3 complex in aqueous solutions.
3-
The main photochemical process was found to be intramolecular electron transfer from the ligand to Fe(III)
ion with formation of a primary radical complex [(C2O4)2FeII(C2O4 )]3-. The yield of radical species (i.e.,
•
CO2• - and C2O4• -) was found to be less than 6% of FeIII(C2O4)33- disappeared after flash. [(C2O4)2FeII(C2O4•)]3-
dissociates reversibly into oxalate ion and a secondary radical complex, [(C2O4)FeII(C2O4 )]-. The latter reacts
•
with the initial complex and dissociates to FeII(C2O4) and oxalate radical. In this framework, the absorption
spectra and rate constants of the reactions of all intermediates were determined.
1. Introduction
Fe(III)-O bond (between one oxalate ligand and Fe(III) ion)
and the C-C bond of the ligand, in which biradical complex
The photochemical reactions of Fe(III)-(poly)carboxylate
complexes influence many iron-dependent biogeochemical
processes and are responsible for photochemical production of
CO and CO2 and for oxygen consumption in natural waters.1-4
It is assumed that the photolysis of Fe(III)-(poly)carboxylate
complexes is one of the main sources of active oxygen species
[(C2O4)2FeIII(CO2 )2]3- 16 or CO2 radical are assumed as
primary intermediate(s).17,18
•
• -
In spite of the serious difference of the proposed mechanisms,
the successive formation of several (at least two) intermediates
is supposed in all aforesaid works. One of these species has an
absorption maximum near 400 nm and disappears in the
millisecond time range. It was also found that the decay of this
intermediate depends on both complex16-18 and ligand15,17
concentrations.TheFeII(C2O4)3 4-17 or the [(C2O4)2FeII(C2O4•)]3- 14,15
(•OH, HO2 , H2O2), the formation of which is catalyzed by Fe(II)
•
and Fe(III) ions in Fenton-like reactions.4-7 Active oxygen
species influence the chemical composition and the oxitation-
reduction capacity of natural water systems, in particular
providing oxidation of sulfur- and nitrogen-containing com-
pounds in the atmosphere.8,9 However, the primary mechanisms
of photoreaction for most Fe(III)-(poly)carboxylate complexes
are still not fully investigated.
and the [(C2O4)2FeIII(CO2 )]2-16 radical complexes were assumed
•
as the intermediate with the millisecond lifetime.
Unfortunately, information about the spectral and kinetic
properties of primary species formed immediately after excita-
tion of the FeIII(C2O4)33- complex is rather limited. This is due
to the application of methods with insufficient time resolution
Among Fe(III)-(poly)carboxylate complexes, the photo-
chemistry of trioxalate Fe(III) complex [FeIII(C2O4)33-] is of
intense interest, as this species not only exhibits high photo-
chemical activity in natural waters under sunlight8,9 but also is
widely used as a stable and sensitive chemical actinometer.10,11
The total process of FeIII(C2O4)33- photolysis could be presented
as12
in the majority of works devoted to the photochemistry of
3- 14,16,17
FeIII(C2O4)3
.
So suggestions about the mechanism of
FeIII(C2O4)33- photolysis presented in these works were mainly
hypothetical.
There are only two works in which the photochemistry of
the FeIII(C2O4)33- complex was investigated with sufficient time
resolution,15,18 and it is worth noting that the conclusions of
these works are in a conflict with each other. In ref 15, two
intermediates with lifetimes 30 µs and 1 ms were detected by
means of a nanosecond flash photolysis technique. The spectrum
of long-lived intermediate corresponds well to transient spectra
observed in microsecond flash photolysis experiments.14,16,17 The
short-lived intermediate was assigned to the excited state of the
complex (*[FeIII(C2O4)33-]). This state decays due to electron
transfer from ligand to central ion with formation of the long-
hν
2FeIII(C2O2)33- 98 2FeII(C2O4
)
2- + 2CO2 + C2O4
2-
2-
2
(1.1)
The quantum yield of Fe(II) formation in reaction 1.1 is
independent of excitation wavelength in the range 270-365 nm
and equals 1.24.10,13 This means that the absorption of UV
quantum leads to the photoreduction of the excited complex
with a 60% probability.
lived [(C2O4)2FeII(C2O4 )]3- radical complex. It was assumed
•
There are two different points of view on the primary
photoprocess following by excitation of FeIII(C2O4)33-: (1)
intramolecular electron transfer from the ligand to Fe(III) ion
in which the long-lived excited state, the radical complex
that radical complex dissociates to FeII(C2O4)2 and C2O4
2-
• -
radical and oxidation of the latter species by the second
FeIII(C2O4)33- leads to formation of final photoreaction products.
In ref 18, the photochemistry of FeIII(C2O4)3 was investi-
3-
[(C2O4)2FeII(C2O4 )]3- or the C2O4 radical are proposed as
•
• -
primary intermediate(s);14,15 or (2) sequential cleavage of the
gated by combination of time-resolved extended X-ray absorp-
tion fine structure (EXAFS) spectroscopy, flash photolysis with
different time resolution (from femto- to milliseconds), a radical
* Corresponding author: e-mail pozdnyak@kinetics.nsc.ru.
10.1021/jp8040583 CCC: $40.75
2008 American Chemical Society
Published on Web 08/15/2008

New Insight into Photochemistry of Ferrioxalate
J. Phys. Chem. A, Vol. 112, No. 36, 2008 8317
scavenging technique, and theoretical calculations. Ultrafast
formation (<2 ps) of an absorption band with a maximum near
430 nm was observed, which was assigned to the excited state
of FeIII(C2O4)33-. At 4 and 9 ps after excitation, the length of
the Fe-O bond was found to decrease to 1.93 and 1.87 Å,
respectively. When it was taken into account that the experi-
mental values of Fe(II)-O bonds are always higher than those
of Fe(III)-O, the primary photoprocess was proposed to be
Fe-O bond cleavage in the excited state with formation of five-
coordinated [(C2O3)O-FeIII(C2O4)2]3- intermediate. This pro-
cess is accompanied by C-C bond cleavage of the ligand,
K3Fe(C2O4)3 (analytical grade) was recrystallized from aque-
ous solution before use. K2C2O4 (analytical grade) and methyl
viologen dichloride hydrate (Aldrich, 98%) were used without
further purification. Fe(II)-oxalate complexes were obtained
by mixing freshly prepared oxygen-free solutions of FeSO4 and
K2C2O4. FeSO4 solution was obtained by dissolution of carbonyl
iron powder (analytical grade) in 0.1 M sulfuric acid (chemical
grade) under constant argon bubbling. Fe(II) concentration in
the FeSO4 solution was determined by complexometry12 with
1,10-phenanthroline as a ligand. This method was also used to
determine the quantity of Fe(II) produced in the flash experiments.
The samples were prepared with deionized water. Unless
otherwise specified, all experiments were carried out with
oxygen-free samples at pH ) 6-7, ionic strength 0.1 (LiClO4,
chemical grade), and 298 K in a 1 cm optical cell. Oxygen was
removed by bubbling solutions with gaseous argon. Initial
concentrations of K3Fe(C2O4)3 and K2C2O4 were varied in the
ranges (1-12) × 10-4 and (3-300) × 10-4 M, respectively.
Under the aforesaid conditions, the main form (>90%) of Fe(III)
leading to formation of stable FeIII(C2O4)2- complex and CO2
• -
radical:
*[FeIII(C2O4)33-] f [(C2O3)OFeIII(C2O4)2]3-
f
• -
FeIII(C2O4)2- + 2CO2 (1.2)
• -
*[FeIII(C2O4)33-] f FeIII(C2O4)2- + 2CO2
(1.3)
The authors18 could not distinguish between reactions 1.2 and
1.3 due to insufficient time resolution of EXAFS method. For
both [(C2O3)O-FeIII(C2O4)2]3- and FeIII(C2O4)2-, theoretical
calculations predict a Fe-O bond length of about 1.9 Å. The
existence of reactions 1.2 and 1.3 was also supported by more
than 50% decrease in Fe(II) quantum yield upon the addition
of radical scavenger (thymine) in ferrioxalate solution.
At longer time scales (from tens of nanoseconds to 2 ms),
the transient absorption signal exhibits biexponential decay in
a similar manner as observed in ref 15. This behavior was
explained by the following set of reactions:18
in solution was FeIII(C2O4)3 complex.20 Samples for flash
3-
photolysis were used until a 10% decrease in UV absorption
occurred.
3. Results and Discussion
3.1. Optical Spectroscopy of Fe(II) and Fe(III) Oxalate
Complexes. Depending on the pH and the concentrations of
Fe(III) and oxalate ions, complexes with one, two, and three
ligands in the coordination sphere of Fe(III) ion can exist in
aqueous solutions:20
Fe3+aq + C2O42- S Fe(C2O4)+
pK1 ) 9.4
pK2 ) 16.2
pK3 ) 20.2
FeIII(C2O4)2- + CO2• - f FeII(C2O4)34- + CO2 (1.4)
-
Fe3+aq + 2C2O42- S Fe(C2O4)2
2-
FeII(C2O4)34- S FeII(C2O4)22- + C2O4
(1.5)
3-
Fe3+aq + 3C2O42- S Fe(C2O4)3
The last reaction has to explain dependence of the intermediate
absorption decay on a ligand concentration observed in the
previous work.17 So in the mechanism of photoreactions
proposed,18 the main intermediates on the aforementioned time
scale are the oxalate complexes of Fe(III) and Fe(II). Both
species are supposed to have absorption bands in the near UV,
with maxima at 430 nm [FeIII(C2O4)2-] and 400 nm
[FeII(C2O4)34-].
All Fe(III)-oxalate complexes exhibit charge-transfer (CT)
bands with maxima about 260 nm overlapping with a strong
absorption band of the ligands at 210 nm.6,12,18 Intensity of the
CT band increases with increasing of number of ligands (Figure
1a).
The main product of FeIII(C2O4)3
photolysis is FeII-
3-
(C2O4)22-. This complex can reversibly dissociate to FeII(C2O4)
and free oxalate ion:21
Thus, the mechanism of FeIII(C2O4)33- photolysis is still open
to discussion. In this work, a nanosecond flash photolysis
technique was used to clarify the nature and the spectral and
kinetic properties of transient species formed after excitation
of ferrioxalate solution. The main attention was concentrated
on (1) reinvestigation of spectroscopy of stable Fe(II) and Fe(III)
oxalate complexes; (2) the search for the experimental evidence
Fe2+aq + C2O42- S FeII(C2O4)
Fe2+aq + 2C2O42- S FeII(C2O4)2
pK1 ) 3.05
pK2 ) 5.01
2-
The spectroscopy of Fe(II)-oxalate complexes is not well
described in the literature, so our first step was to obtain
information about the absorption spectra of these complexes.
At concentration of oxalate ions greater than 3 × 10-3 M, more
than 85% of Fe(II) ions are bound in oxalate complexes and
• -
for organic radical (i.e., CO2 and C2O4• -) formation in the
primary photochemical process; and (3) detailed investigation
of the kinetics of the intermediates’ decay in the wide range of
ligand and ferrioxalate concentration.
the contribution of Fe2+ to the observed optical spectrum is
aq
negligible (Fe2+aq complex absorbs weakly in UV region;22 λmax
) 239 nm, ε239 ) 20 M-1 cm-1). In this condition one can use
an effective absorption coefficient of Fe(II)-oxalate complexes
[ε(FeII)] that is related to absorption coefficients of the individual
complexes by
2. Experimental Details
Laser flash photolysis experiments were performed in a setup
with YAG laser excitation (355 nm, pulse duration 5 ns, mean
energy 5 mJ/pulse) analogous to that described elsewhere.19
Absorption spectra were recorded on an HP 8453 Agilent
spectrophotometer. For the numerical calculations of kinetic
curves, the differential equations were solved by means of the
fourth-order Runge-Kutta method. The composition of Fe(II)
and Fe(III) complexes was calculated with the Visual MINTEQ
2.32 program.
ε(FeII)
R1
R2
2R1
ε1 + ε
)
(3.1)
where R and ε are percentage and absorption coefficients of an
individual complex, and indices 1 and 2 correspond to FeII(C2O4)
and FeII(C2O4)22-, respectively.

8318 J. Phys. Chem. A, Vol. 112, No. 36, 2008
Pozdnyakov et al.
Figure 2. (a) Transient absorption spectra after the excitation at 355
nm of FeIII(C2O4)3
complex in aqueous solution at pH 7.
3-
Figure 1. Optical absorption spectra of (a) Fe(III)-oxalate and (b)
Fe(II)-oxalate complexes in aqueous solution: (a) (1) FeIII(C2O4)+, (2)
[K3Fe(C2O4)3] ) 5 × 10-4 M; [K2C2O4] ) 10-2 M. Spectra 1-3 were
taken at 0.05, 10, and 50 µs after the laser flash. (b) Typical kinetic
curve at 400 nm.
FeIII(C2O4)2-, and (3) FeIII(C2O4)33-. (b) (1) FeII(C2O4) and (2)
FeII(C2O4)2
.
2-
are 20 µs and 1 ms, which coincide very well with the published
data (30 µs and 1 ms).15
But does the intermediate with absorption maximum at 400
In Figure 1b, spectra of FeII(C2O4) and FeII(C2O4)2 are
presented. These spectra were calculated from a set of optical
spectra of Fe(II) solutions with different initial concentration
of oxalate ions by use of eq 3.1. Both complexes are character-
ized by low-intensity absorption bands in the near UV with
maxima at 340 nm (ε ) 70 M-1 cm-1) and 350 nm (ε ) 130
M-1 cm-1) for FeII(C2O4) and FeII(C2O4)22-, respectively.
It is worth noting that all aforementioned results are in
disagreement with theoretical calculations18 predicting the
2-
nm appear to be the single primary species, or does this
• -
intermediate appear together with organic radicals CO2
/
C2O4• -, which absorb at λ < 300 nm23,24 and cannot be detected
directly in our setup? For detection of organic radicals in our
flash experiments, the well-known electron acceptor methyl
viologen dication (MV2+) was used.25,26 This species reacts with
both CO2 (k ) 1 × 1010 M-1 s-1) and C2O4 (k ) 4.3 ×
109 M-1 s-1) radicals with high rate constants.24 As a result of
the reactions, the cation radical MV• + was formed, which
• -
• -
maxima of absorption bands of FeIII(C2O4)2- and FeII(C2O4)3
4-
at 430 and 400 nm, respectively. The reason is most probably
due to neglect of solvation effects and spin orbital coupling in
the theoretical calculations.18 Both factors seem to be taken into
account for correct calculation of optical properties of such
highly charged complexes of transient metals like oxalate
exhibits strong absorption bands at 396 (ε ) 41 500 M-1 cm-1
)
and 605 nm.27
Addition of MV2+ to solutions containing FeIII(C2O4)33- leads
to formation and decay of an additional transient absorption
signal at 396 nm (Figure 3) after complex excitation. The
observed rate constant of signal formation depends linearly on
MV2+ concentration (Figure 3, inset), indicating the formation
of MV• + radical cation. The value of the bimolecular rate
complexes of Fe(II) and Fe(III) in a polar medium.
3.2. Laser Flash Photolysis of FeIII(C2O4
)
3
3- Complex.
2-
Excitation of an oxygen-free aqueous solution of ferrioxalate
at 355 nm gives rise to an intermediate absorption consisting
of two bands with maxima at 400 and 640 nm (Figure 2a).
Negative optical density at λ < 370 nm indicates the disap-
constant of MV• + formation calculated from the slope (kMV
≈
1.2 × 1010 M-1 s-1) is close to the value of the rate constant of
pearance of the ground-state absorption of FeIII(C2O4)3 after
3-
the reaction of CO2• - radical with MV2+ 24
. The nonzero value
excitation. The characteristic kinetic curve at 400 nm is shown
in Figure 2b. The change of optical density in the whole range
of wavelength (320-720 nm) obeys the biexponential law:
of the intercept of the plot in the inset of Figure 3 corresponds
to reaction of the organic radical with FeIII(C2O4)33-. The value
of the bimolecular rate constant calculated from the intercept
(kFe ≈ 5.5 × 109 M-1 s-1) is also close to the value of the rate
∆A(λ) ) A0(λ) + A1(λ)exp(-kfastt) + A2(λ)exp(-kslowt)
constant for the reaction of CO2 radical with the complex.28
• -
(3.2)
So one can assume that MV• + radical cation is formed in the
where A0 ) 2[ε(FeIII) - ε(FeII)]∆[FeIII] is the steady-state
change of absorbance of ferrioxalate solution (reaction 1.1),
ε(FeIII) is the absorption coefficient of FeIII(C2O4)33-, and ε(FeII)
is the effective absorption coefficient of Fe(II)-oxalate complexes.
This fact supports the supposition that excitation of
FeIII(C2O4)33- leads to successive formation of two intermediates
absorbing in the near UV.14-18 The observed rate constants kfast
and kslow are independent of a registration wavelength, which
indicates that only (pseudo)monomolecular reactions take place
in the system. The characteristic lifetimes of the intermediates
reaction of MV2+ with CO2 radical.
• -
In the absence of dissolved oxygen and other possible
oxidizers, the concentration of MV• + practically does not change
during several minutes.29 In our case, the decay of MV• + (Figure
3) seems to be determined by the reaction with FeIII(C2O4)3
3-
complex.
The initial concentration of CO2• - could be directly calculated
from the concentration of MV• + under conditions when all
organic radicals are quenched by MV2+. At [MV2+] ) 10-3 M,
the formation time of MV• + is less than 100 ns, and so the

New Insight into Photochemistry of Ferrioxalate
J. Phys. Chem. A, Vol. 112, No. 36, 2008 8319
Figure 3. Laser flash photolysis of FeIII(C2O4)3 (1.6 × 10-4 M) in
3-
the presence of MV2+. Curves 1-5 are kinetic curves (396 nm) at initial
concentrations of [MV2+] × 105 ) 0, 2.7, 8.4, 25, and 100 M,
of MV• + formation on MV2+ concentration.
Figure 4. Dependence of the observed rate constants (1) kfast and (2)
396
respectively. (Inset) Dependence of the observed rate constant (kobs
)
kslow on the concentration of FeIII(C2O4)3 at initial free oxalate ion
3-
concentrations of (a) 10-3 M and (b) 10-2 M. Curves are linear
approximations of the experimental data (note that the y-axis has a
logarithmic scale).
concentration of MV• + (and CO2• -) is easily calculated from
the difference in the initial amplitudes of the signals at 400 nm,
corresponding to kinetic curves 5 and 1 in Figure 3. One can
estimate that the yield of CO2 (ca. 1.2 × 10-6 M) is less
• -
than 6% of the number of FeIII(C2O4)33- complexes photolyzed
(ca. 2 × 10-5 M). This result is in contradiction with indirect
radical-scavenging experiments.18 The large (more than 50%)
decrease of Fe(II) quantum yield was observed after the addition
of thymine in ferrioxalate solution, which was interpreted by
• -
scavenging of CO2 radical formed in reactions 1.2 and 1.3
by thymine. Most probably the results18 should be explained
by oxidation of final Fe(II)-oxalate complexes or the transient
• -
species by thymine rather than by reaction of CO2 radical
with the scavenger.
Thus, the species with absorption maximum at 400 nm is
the main primary intermediate arising upon FeIII(C2O4)3
3-
photolysis. By carrying out a parallel determination of the
transient absorption signal at 400 nm and the concentration of
Fe(II) ions, one can calculate the absorption coefficient of this
species at 400 nm (ε ≈ 780 M-1 cm-1). It is worth noting that
the variation of MV2+ concentration has no effect on the lifetime
of the intermediate absorbing at 400 nm (Figure 3), indicating
the absence of reaction between these species.
Figure 5. Dependence of the observed rate constants (1) kfast and (2)
kslow on the concentration of free oxalate ions. The curve is a linear
approximation of the experimental data (note that both axes have
logarithmic scales).
3.3. Nature of Primary Intermediate in Photochemistry
of FeIII(C2O4)33-. As was mentioned in the Introduction, several
species are candidates for the role of primary intermediate
absorbing at 400 nm: (1) the FeIII(C2O4)2- complex, (2) the long-
lived excited state of FeIII(C2O4)33-, (3) the biradical complex
To distinguish between candidates 2-4, the dependence of
kfast on the pH and initial concentration of oxygen, ferrioxalate,
and oxalate ions was investigated in detail. It was observed that
kfast is independent of pH in the range 5.5-7.5 (similar to
published data15,16), concentration of oxygen (data not shown),
and ferrioxalate (Figure 4a,b) and increases linearly with
increasing concentration of free ligand (Figure 5). It is worth
[(C2O4)2FeIII(CO2 )2]3-, and (4) the [(C2O4)2FeII(C2O4 )]3- radi-
cal complex (hereafter in the text referred to as RC1).
•
•
As was shown in section 3.1, the FeIII(C2O4)2- complex does
not have an absorption band with a maximum at 400 nm, and
the absorption coefficient of this complex at this wavelength
(270 M-1 cm-1) is much less than the calculated one for the
primary intermediate (780 M-1 cm-1). Also, the formation of
2-
noting that the increase of C2O4 concentration has no effect
on initial amplitude of the transient signal at 400 nm but leads
to increased absorbance of the long-lived species (Figure 6a).
The absence of quenching of the primary intermediate by
FeIII(C2O4)33-, oxygen, and MV2+ means that the primary
species is not the long-lived excited state of FeIII(C2O4)33-. The
-
• -
FeIII(C2O4)2 has to be accompanied by formation of CO2
radical. As the yield of the latter species was found to be
negligible, the formation of FeIII(C2O4)2 in the photolysis of
-
FeIII(C2O4)3 must be ruled out.
3-

8320 J. Phys. Chem. A, Vol. 112, No. 36, 2008
Pozdnyakov et al.
As the quantum yield of FeIII(C2O4)3 photolysis depends
weakly on excitation wavelength,11 we can tentatively assume
that a THEXI state is a precursor of RC1.
3-
Below, interpretation of all experimental results in the
framework of RC1 formation at FeIII(C2O4)3 excitation is
3-
presented.
3.4. Kinetic Scheme of FeIII(C2O4)3
Photolysis. To
3-
explain the observed dependence of kfast on the concentration
of oxalate ions, the reversible dissociation of RC1 from the
secondary [(C2O4)FeII(C2O4 )]- radical complex (hereafter in
•
the text known as RC2) and oxalate ion was proposed:
k1
2-
RC1
98 RC2 + C2O4
(3.3)
(3.4)
k2
RC2 + C2O42- 98 RC1
This supposition is based on the smaller stability of Fe(II)-oxalate
complexes compared with Fe(III)-oxalate complexes.20,21 One
can estimate the values of the rate constants k1 (≈ 3.8 × 104
s-1) and k2 (≈ 3.4 × 106 M-1 s-1) from the linear fitting of
dependence of kfast on the concentration of oxalate ions (Figure
5).
kslow decreases with increasing ligand concentration (Figure
5, in agreement with published results17) and rises linearly with
increasing ferrioxalate concentration (Figure 4a,b). The slope
ratio of dependence of kslow on [FeIII(C2O4)33-] decreases with
increasing ligand concentration (Figure 4a,b), indicating that
only RC2 reacts with the initial complex:
Figure 6. (a) Kinetic curves at 400 nm. (Curves 1-5) [K2C2O4] ×
103 ) 1, 3, 10, 20, and 30 M, respectively; [K3Fe(C2O4)3] ) 5 × 10-4
M. (b) Kinetic curves at 400 nm. (Curves 1-5) [K3Fe(C2O4)3] × 104
) 1.3, 2.6, 5.1, 8.5, and 16.4 M, respectively; [K2C2O4] ) 10-3 M.
The smooth curves are fits by numeric solutions of the system of the
differential equations corresponding to eqs 3.3-3.8, with the average
values of the parameters listed in Table 1.
formation of this state was based on flash experiments15 in which
the decreased lifetime of the primary species with increased
ferrioxalate concentration was observed. This fact was inter-
k3
RC2 + FeIII(C2O4)33- 98 2FeII(C2O4)22- + 2CO2 (3.5)
preted as self-quenching of the excited state of FeIII(C2O4)3
3-
Most probably, that reaction between RC1 and FeIII(C2O4)3
3-
by the initial complex. However, the lifetime of primary
intermediate was measured at a constant [Fe(III)]/[oxalate] ratio,
that is, with a synchronous increase in both complex and ligand
concentrations.15 So one can conclude that the results15 are well
explained by the increase of kfast (which is a reverse lifetime of
primary species) with increasing ligand concentration (Figure
5). Moreover, the lifetime of *[Fe(C2O4)33-] (ca. 30 µs)15 is
1-2 orders of magnitude higher than the typical lifetimes of
the excited states of the metal complexes.30
is very slow and is not observed under our experimental
conditions due to strong Coulomb repulsion (both complexes
have a high negative charge, q ) -3).
The value of the rate constant k3 (g9 × 105 M-1 s-1) could
be estimated from the slope ratio of dependence of kslow on
[FeIII(C2O4)33-] at low ligand concentration (Figure 4a). The
low value of k3 is explained by negative charges of both
reagents. The nonzero value of the intercepts of the plots in
Figure 4a,b corresponds to the monomolecular dissociation of
both radical complexes with approximately the same rate
constants (k4 ≈ k5 ≈ 150 M-1 s-1):
The independence of kfast from the concentration of oxygen
and MV2+ also conflicts with the hypothesis that the
[(C2O4)2FeIII(CO2 )2]3- biradical complex is formed in a primary
•
photochemical process, because CO2• - radical reacts with both
species.24 The fast intramolecular electron transfer from coor-
dinated CO2• - on the central Fe(III) ion could also be expected.
Moreover, dependence of kfast on [C2O42-] cannot be explain
k4
9
• -
RC2
8 FeII(C2O4) + C2O4
(3.6)
k5
in the framework of hypothesis of [(C2O4)2FeIII(CO2 )2]3-
•
• -
RC1
9
8 FeII(C2O4)22- + C2O4
(3.7)
formation.
Thus, radical complex RC1 is the most probable candidate
for the role of the primary intermediate. It worth noting that
formation of radical complexes is a typical feature of photo-
chemistry of transient metals and is shown for the number of
coordination compounds of copper, chromium, manganese,
cobalt, nickel, and platinum.31,32 Unfortunately, our time resolu-
tion is not sufficient to determine the precursor of RC1. It could
be the primary excited state populated upon a ligand-to-metal
charge transfer- (LMCT-) type transition or a long-lived
secondary excited state (so-called THEXI state) populated after
relaxation of the LMCT state.30 One piece of evidence for
THEXI state formation is independence (or weak dependence)
of the quantum yield of photolysis on excitation wavelength.
Oxalate radical formed in reactions 3.6 and 3.7 reacts most
probably with the initial complex:
k6
2-
C2O4• - + FeIII(C2O4)33- 98 FeII(C2O4)22- + C2O4
+
2CO2 (3.8)
The rate constant of reaction 3.8 cannot be determined directly,
as oxalate radical exhibits an absorption band in a wavelength
region unregistered by our setup (λ < 300 nm24). We propose
that the value of k6 is close to value of the rate constant of the
reaction of CO2 radical with FeIII(C2O4)3 and exceeds
• -
3-

New Insight into Photochemistry of Ferrioxalate
J. Phys. Chem. A, Vol. 112, No. 36, 2008 8321
TABLE 1: Parameters Used in Numerical Simulation of Kinetic Curves at 400 nm
ε(FeIII), M-1 cm-1 ε(RC1), M-1 cm-1 ε(RC2), M-1 cm-1 k1 × 10-4, s-1 k2 × 10-6, M-1 s-1 k3 × 10-6, M-1 s-1 k4 × 10-2, s-1 k5 × 10-2, s-1
140a
780a
370 ( 30
3.8a
3.4a
1.1 ( 0.2
1.7 ( 0.7
1.6 ( 0.7
a Fixed parameter.
109 M-1 s-1. This assumption is based on proximity of the
values of the rate constants for reactions of both radicals with
MV2+. As the decay rate of the oxalate radical in reaction 3.8
(k6[FeIII(C2O4)3]3- > 5 × 105 s-1) exceeds greatly its generation
rate in reactions 3.6 and 3.7 (k4 + k5 < 500 s-1), one can use
the method of stationary concentrations with reference to this
intermediate.
3.5. Simulation of Experimental Kinetic Curves at 400
nm. If d[C2O4• -]/dt ) 0 and fast equilibrium between
Fe(II)-oxalate complexes takes place, one can express the
change in concentrations of the reagents and intermediates by
the following set of differential equations:
concentration because in our conditions k3[FeIII(C2O4)33-] <
1500 s-1 , k1 ) 3.8 × 104 s-1. kslow exhibits linear dependence
on ferrioxalate concentration and is inversely proportional to
ligand concentration.
To obtain more accurate values of rate constants (eqs 3.3-3.8)
and to determine the absorption coefficient of RC2, all
experimental kinetic curves at 400 nm were fitted by numerical
solutions of the system of differential equations (eq 3.9) by the
fourth-order Runge-Kutta method. As fixed parameters, the
absorption coefficients of FeIII(C2O4)3 and RC1 [ε(FeIII) and
3-
ε
RC1] and rate constants of reversible dissociation of RC1 (k1
and k2) were used. As variable parameters, the absorption
coefficient of RC2 (εRC2) and the rate constants of disappearance
of RC1 and RC2 (k3, k4, and k5) were used. Dependence of
ε(FeII) on oxalate ion concentration was also taken into account,
although this parameter has a weak influence on results of fitting.
The results of fitting of about 90 experimental kinetic curves
[with different initial concentrations of oxalate ions,
FeIII(C2O4)33-, and RC1] are summarized in Table 1. The
smooth lines in Figure 6 show calculations with the average
values of parameters presented in Table 1. Full agreement
between calculated and experimental curves was observed over
a wide range of concentrations of reagents and at the same time
with values of variable parameters close to the experimental
one (Table 1). These facts allow us to maintain the adequacy
of the proposed mechanism of the photolysis of [FeIII(C2O4)3]3-
complex (Figure 7).
d[RC1]
) -(k1 + k5)[RC1] + kꢀ2[RC2]
dt
d[RC2]
) k1[RC1] - (kꢀ2 + k3ꢀ)[RC2]
dt
d[Fe(III)]
) -k5[RC1] - kꢀ3[RC2]
dt
d[Fe(II)]
) 2k5[RC1] + 2kꢀ3[RC2]
(3.9)
dt
where [Fe(III)] and [Fe(II)] describe the change in concentrations
of FeIII(C2O4)3 and Fe(II)-oxalate complexes, respectively;
3-
k′2) k2[C2O42-], and k3′ ) k3[FeIII(C2O4)33-] + k4.
The system of differential equations (eq 3.9) can be solved
analytically on the assumption that both [C2O42-] and
[FeIII(C2O4)33-] . [RC1]. This assumption works well at
[C2O42-] > 5 × 10-4 M and low intensity of excitation, when
less than 20% of initial complexes are decomposed in irradiated
volume. The analytical solution of the system (eq 3.9) for change
of optical density at 400 nm gives an expression absolutely
identical with eq 3.2, where
Information about rate constants (eqs 3.3-3.8) allows one
to calculate the absorption spectra of the radical complexes.
Figure 8a presents the results of global fitting of experimental
kinetic curves in the range 300-520 nm to eq 3.2. The best
agreement of the calculated and experimental kinetic curves was
kfast ) k1 + k2[C2O42-] + k3[FeIII(C2O4)33-] + k4 + k5
kslow ) (k3[FeIII(C2O4)33-] + k4){k1/(k1 + k2[C2O42-])}
Components A1 and A2 (eq 3.2) are complex functions of the
absorption coefficients of initial, radical, and final complexes
and the rate constants of reactions (3.3-3.8). It is very important
that the analytical expressions of kfast and kslow by the rate
constants of elementary stages fully coincide with our experi-
mental data (Figures 4 and 5). kfast exhibits linear dependence
on ligand concentration and is independent of ferrioxalate
Figure 8. (a) Spectra of (1) A0(λ), (2) A1(λ), and (3) A2(λ) components
(eq 3.2), corresponding to kfast ) 4.5 × 104 s-1 and kslow ) 670 s-1. (b)
Calculated absorption spectra of (1) RC1 and (2) RC2. [K3Fe(C2O4)3]
) 5 × 10-4 M; [K2C2O4] ) 10-3 M.
Figure 7. General scheme of FeIII(C2O4)3 photolysis.
3-

8322 J. Phys. Chem. A, Vol. 112, No. 36, 2008
Pozdnyakov et al.
obtained at kfast(fit) ) 4.5 × 104 s-1 and kslow(fit) ) 670 s-1
.
(3) Voelker, B. F.; Morel, M. M.; Sulzberger, B. EnViron. Sci. Technol.
1997, 31, 1004.
These values are in fair agreement with the calculated ones from
data in Table 1 [kfast(calc) ) 4.2 × 104 s-1 and kslow(calc) )
650 s-1]. At the ligand concentration used (10-3 M), the
sum of amplitudes A1 and A2 corresponds to the absorption
spectrum of RC1 and amplitude A2 corresponds to the absorption
spectrum of RC2. The calculated absorption spectra of radical
complexes are presented in Figure 8b. Both radical complexes
exhibit absorption bands with maxima at 330 and 380 nm. It is
worth noting that stable Fe(II)-oxalate complexes have absorp-
tion bands in the same region (Figure 1b). The intensity of the
absorption bands of the radical complexes decreases with
decreasing coordination number, which is typical for spectros-
copy of the stable Fe(II)- and Fe(III)-oxalate complexes
(Figure 1a,b).
(4) Gao, H.; Zepp, R. G. EnViron. Sci. Technol. 1998, 32, 2940.
(5) Behra, P.; Sigg, L. Nature 1990, 344, 419.
(6) Zuo, Y.; Holdne, J. EnViron. Sci. Technol. 1992, 26, 1014.
(7) Faust, B. S.; Zepp, R. G. EnViron. Sci. Technol. 1993, 27, 2517.
(8) Zuo, Y.; Zhan, J. Atmos. EnViron. 2005, 39, 27.
(9) Zuo, Y.; Hoigne, J. Science. 1993, 260, 71.
(10) Calvert, J.; Pitts J. N. Photochemistry; John Wiley and Sons: New
York, London, and Sydney, 1967.
(11) Kurien, K. C. J. Chem. Soc. B 1971, 2081.
(12) Balzani, V.; Carassiti, V. Photochemistry of Coordination Com-
pounds; Academic Press: London and New York, 1970.
(13) Goldstein, S.; Rabani, J. J. Photochem. Photobiol., A 2008, 193,
50.
(14) Parker, C. A.; Hatchard, C. G. Nature 1955, 176, 122.
(15) Nadtochenko, V.; Kiwi, J. J. Photochem. Photobiol., A 1996, 99,
145.
(16) Patterson, J. I. H.; Perone, S. P. J. Phys. Chem. 1973, 77, 2437.
(17) Cooper, G. D.; DeGraff, B. A. J. Phys. Chem. 1971, 75, 2897.
(18) Chen, J.; Zhang, H.; Tomov, I. V.; Wolfsberg, M.; Ding, X.;
4. Conclusions
Rentzepis, P. M. J. Phys. Chem. A 2007, 111, 9326.
(19) Pozdnyakov, I. P.; Plyusnin, V. F.; Grivin, V. P.; Vorobyev, D.
Yu.; Bazhin, N. M.; Vauthey, E. J. Photochem. Photobiol., A 2006, 182,
75.
Laser flash photolysis experiments carried out with a wide
range of initial parameters provide new evidence that intramo-
lecular electron transfer from ligand to Fe(III) ion is a main
photochemical process in FeIII(C2O4)3 photochemistry. As
3-
(20) Sillen, L. G.; Martell, A. E. Stability constants of metal-ion
complexes; The Chemical Society: London, 1964.
primary intermediate, the radical complex (C2O4)2FeII(C2O4 )
• 3-
(21) Micskei, K. J. Chem. Soc., Dalton. Trans. 1987, 255.
(22) Jortnor, J.; Stein, G. J. Phys. Chem. 1962, 66, 1264.
(23) Neta, P.; Simic, M.; Hayon, E. J. Phys. Chem. 1969, 73, 4207.
(24) Mulazzani, Q. G.; D’Angelantonio, M.; Venturi, M.; Hoffman,
M. Z.; Rodgers, M. A. J. J. Phys. Chem. 1986, 90, 5347.
(25) Feng, W.; Nansheng, D.; Glebov, E. M.; Pozdnyakov, I. P.; Grivin,
V. P.; Plyusnin, V. F.; Bazhin, N. M. Russ. Chem. Bull., Int. Ed. 2007, 56,
900.
is formed; the yield of organic radicals in the primary photo-
process is negligible.
A kinetic scheme for FeIII(C2O4)33- photolysis was proposed,
including formation of (C2O4)2FeII(C2O4•)3-, its reversible
dissociation to the oxalate ion and secondary radical complex
• -
(C2O4)FeII(C2O4 ) , and the decay of both radical complexes.
In the framework of this scheme, the absorption spectra and
rate constants of formation and decay of all intermediates were
determined.
(26) Togashi, D. M.; Costa, S. M. B. Phys. Chem. Chem. Phys. 2002,
4, 1150.
(27) Mulazzani, Q. G.; Venturi, M.; Hoffman, M. Z. J. Phys. Chem.
1985, 89, 722.
(28) Hislop, K. A.; Bolton, J. R. EnViron. Sci. Technol. 1999, 33, 3119.
(29) Patterson, L.; Small, R.; Scaiano, J. Radiat. Res. 1977, 72, 218.
(30) Vlcek, A., Jr. Coord. Chem. ReV. 2000, 200-202, 933.
(31) Muller, J.; Kikuchi, A.; Bill, E.; Weyhermuller, T.; Hildebrandt,
P.; Ould-Moussa, L.; Wieghardt, K. Inorg. Chim. Acta 2000, 297, 265.
Acknowledgment. This work was supported by RFBR
(Grants 06-03-110, 06-03-32110, 05-03-39007GFEN, 08-03-
90102-Mol, and 08-03-00313) and by Program of Integration
projects of SB RAS-2006 (Grant 4.16).
References and Notes
(32) Ivanov, Y. V.; Plyusnin, V. F.; Grivin, V. P.; Larionov, S. V. Chem.
Phys. Lett. 1999, 310, 252.
(1) Miles, C. J.; Brezonik, P. L. EnViron. Sci. Technol. 1981, 15, 1089.
(2) Zuo, Y.; Jones, R. D. Water Res. 1997, 31, 850.
JP8040583