Iron porphyrin-catalyzed reduction of CO2. Photochemical and radiation chemical studies

Article abstract of DOI:10.1021/jp9628139

Several iron porphyrins have been reduced by photochemical and radiation chemical methods, in organic solvents and in aqueous solutions, from Fe^IIIP to Fe^IIP to Fe^IP and beyond. In aqueous solutions, the Fe^IP state is relatively stable for the tetrakis(N-methyl-2-pyridyl)porphyrin at high pH but is shorter lived in neutral and acidic solutions. The Fe^IP state of tetrakis(N-methyl-3-pyridyl)porphyrin and tetrakis(N-methyl-4-pyridyl)-porphyrin are short-lived at any pH. Decay of Fe^IP is accelerated by H⁺ and by CO₂, probably via reaction with the Fe⁰P state formed upon disproportionation of Fe^IP. These reactions may lead to formation of H₂ and CO, respectively, and to formation of the chlorin, Fe^IIPH₂, as a side product. The Fe^IP state is also observed as a stable product in several organic solvents. This is observed by photolysis of iron tetraphenylporphyrin and several of its derivatives (e.g., trimethyl-, dichloro- and pentafluorophenyl), mainly in dimethylformamide and acetonitrile solutions, using triethylamine as a reductive quencher. Further photoreduction in the presence of CO₂ results in catalyzed reduction of CO₂ to CO and formation of (CO)-Fe^IIP. The yield of free CO increases with time of photolysis and reaches turnover numbers of approximately 70 molecules of CO per porphyrin molecule.

Full text of DOI:10.1021/jp9628139

248
J. Phys. Chem. A 1997, 101, 248-254
Iron Porphyrin-Catalyzed Reduction of CO2. Photochemical and Radiation Chemical
Studies
J. Grodkowski,† D. Behar,‡ and P. Neta*
Physical and Chemical Properties DiVision, National Institute of Standards and Technology,
Gaithersburg, Maryland 20899
P. Hambright
Department of Chemistry, Howard UniVersity, Washington, DC 20059
ReceiVed: September 12, 1996; In Final Form: NoVember 5, 1996X
Several iron porphyrins have been reduced by photochemical and radiation chemical methods, in organic
solvents and in aqueous solutions, from FeIIIP to FeIIP to FeIP and beyond. In aqueous solutions, the FeIP
state is relatively stable for the tetrakis(N-methyl-2-pyridyl)porphyrin at high pH but is shorter lived in neutral
and acidic solutions. The FeIP state of tetrakis(N-methyl-3-pyridyl)porphyrin and tetrakis(N-methyl-4-pyridyl)-
porphyrin are short-lived at any pH. Decay of FeIP is accelerated by H+ and by CO2, probably via reaction
with the Fe0P state formed upon disproportionation of FeIP. These reactions may lead to formation of H2
and CO, respectively, and to formation of the chlorin, FeIIPH2, as a side product. The FeIP state is also
observed as a stable product in several organic solvents. This is observed by photolysis of iron
tetraphenylporphyrin and several of its derivatives (e.g., trimethyl-, dichloro- and pentafluorophenyl), mainly
in dimethylformamide and acetonitrile solutions, using triethylamine as a reductive quencher. Further
photoreduction in the presence of CO2 results in catalyzed reduction of CO2 to CO and formation of (CO)-
FeIIP. The yield of free CO increases with time of photolysis and reaches turnover numbers of ∼70 molecules
of CO per porphyrin molecule.
Introduction
the initial FeIIIP is reduced in three distinct steps to reach the
Fe0P oxidation state.
Catalyzed reduction of CO2 has been studied extensively as
a means of energy storage.1 A number of transition-metal
complexes have been utilized as electron transfer mediators to
achieve electrochemical2,3 or photochemical4-6 reduction of
CO2. Among these are nickel and cobalt cyclam (and related
tetraaza macrocycles),3-5 ruthenium and rhenium bipyridine
complexes,2,6 and several metalloporphyrins. Of the last group,
iron porphyrins have been demonstrated to be the most efficient
catalysts in the electrochemical reduction of CO2 to CO.7 The
mechanism of transition-metal-catalyzed reduction of CO2 has
been shown to involve an intermediate complex between the
metal and CO2. Certain such complexes have been observed
as stable products or short-lived intermediates, e.g., with Ni-
(cyclam)+ and with CoL+ (L ) various tetraaza macrocycles).5,8
The rate of formation of CO2 complexes and their stability
constants were strongly dependent on the ligand structure and
were particularly sensitive to steric effects. By comparison with
the flexible structures of cyclam and related macrocycles,
porphyrins have larger and more rigid planar structures which
are more likely to hinder the binding of CO2 to the metal center
or to destabilize the complex formed. Furthermore, because
of their conjugated π-system, porphyrin ligands readily accept
electrons in a process that may compete with the reduction of
the metal center and thus hinder reduction of CO2. Nevertheless,
several iron porphyrins have been shown to catalyze electro-
chemical reduction of CO2,7 and the mechanism was suggested
to involve formation of an intermediate Fe-CO2 complex after
In the present study we utilize photochemical and radiation
chemical techniques to examine several iron porphyrins in terms
of their one-electron reduction steps and their possible role as
catalysts for photoreduction of CO2. Iron(III) porphyrins are
known to undergo photoreduction to FeIIP upon excitation at
their ligand-to-metal charge transfer band (at wavelengths below
the Soret band), the axial ligand oxidized being either an
alkoxide9 or a halide.10
XFeIIIP
9
hν8 X• + FeIIP
(1)
The FeIIP produced in alcohol solutions was not affected by
further illumination,9 but we find that it can be photoreduced
in several solvents in the presence of strong reductive quenchers
such as triethylamine. Continued photolysis under certain
conditions is found to lead to reduction of CO2 to CO.
Reduction of FeIIP was studied also by radiolysis. The stability
of the FeIP produced depends on the porphyrin structure and
the medium.
Experimental Section11
Most photochemical experiments were carried out with iron
tetraphenylporphyrin, and most radiolysis experiments were
carried out with water-soluble porphyrins. The porphyrins used
and their abbreviations are as follows: tetraphenylporphyrin
(TPP), tetramesitylporphyrin (TMP, mesityl ) 2,4,6-trimeth-
ylphenyl), tetrakis(2,6-dichlorophenyl)porphyrin (TCPP), tet-
rakis(pentafluorophenyl)porphyrin (TFPP), 2,3,7,8,12,13,17,18-
octabromo-5,10,15,20-tetramesitylporphyrin (OBTMP), the picket
fence porphyrin R,R,R,R-tetrakis[2-(2,2-dimethylpropionamidyl)-
phenyl]porphyrin (PFP), octaethylporphyrin (OEP), tetrakis(4-
sulfonatophenyl)porphyrin (TSPP), tetrakis(N-methyl-4-pyridyl)-
† On leave from the Institute of Nuclear Chemistry and Technology,
Warsaw, Poland.
‡ On leave from the Soreq Nuclear Research Center, Yavne, Israel.
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, January 1, 1997.
S1089-5639(96)02813-7 CCC: $14.00 © 1997 American Chemical Society

Iron Porphyrin-Catalyzed Reduction of CO2
J. Phys. Chem. A, Vol. 101, No. 3, 1997 249
porphyrin (TM4PyP), tetrakis(N-methyl-3-pyridyl)porphyrin
(TM3PyP), tetrakis(N-methyl-2-pyridyl)porphyrin (TM2PyP),
and tetraphenylchlorin (TPC). They were obtained from Mid-
Century Chemicals (Posen, IL) in the form of chloroiron(III)
porphyrins, ClFeIIIP. Triethylamine (TEA) and triethanolamine
(TEOA) were from Aldrich; N,N-dimethylformamide (DMF),
2-propanol (2-PrOH), ethanol (EtOH), tetrahydrofuran (THF),
acetonitrile, sodium formate, and the inorganic compounds were
analytical grade reagents from Mallinckrodt. Water was purified
with a Millipore Super-Q system. Fresh solutions were prepared
before each experiment and were irradiated after purging with
Ar or He to remove oxygen or with CO2.
reducing species, which may reduce another FeIIIP.
Et3N•+ + Et3N f Et3NH+ + Et2NC˙ HCH3
(3)
Et2NC˙ HCH3 + ClFeIIIP f Et2N+dCHCH3 + FeIIP + Cl-
(4)
The efficiency of photoreduction was the same with 5 and 10%
TEA and was reduced by only 40% when the TEA volume
fraction was lowered to 1%. We chose 5% TEA (0.36 mol
L-1) as the standard concentration for most subsequent experi-
ments. Saturating the above solutions with CO2 (∼0.24 mol
L-1) had little effect on the rate of photoreduction.
Photolysis was performed with either a 300 or 900 W xenon
lamp, using water filters to absorb the IR and UV cutoff filters
as specified. Radiolysis was performed in a Gammacell 220
The quantum yield for photoreduction under the present
conditions was estimated by comparison with the reported value
for photoreduction of FeIIITPP in alkaline ethanol solutions,
which is 0.016 for photolysis at 355 nm.15 By running parallel
experiments with FeIIITPP in alkaline ethanol and in DMF
solutions containing 5% TEA, at 360 ( 20 nm (using a
combination of filters), we found the latter system to be about
3 times more efficient. This leads to a quantum yield for
photoreduction of FeIIITPP in DMF/TEA of approximately 0.05.
Photochemical Reduction of FeIIP in Organic Solvents.
Further illumination of the solutions of FeIITPP, obtained in
the above experiments, resulted in reduction of this species to
FeITPP, but the process was much less efficient. With the 325
or 385 nm cutoff filters, the efficiency of photoreduction of
FeIITPP was about 20% that of photoreduction of FeIIITPP under
the same conditions. With the 420 nm cutoff filter, however,
reduction of FeIITPP was about 10 times slower than reduction
of FeIIITPP. The mechanism must be similar to that outlined
above, i.e., excitation of the ligand-to-metal charge transfer band.
60Co source with a dose rate of either 0.8 or 2.7 Gy s-1
.
Changes in porphyrin structure were derived from spectropho-
tometric measurements, and the CO evolved was determined
by gas chromatography (Carboxen-1000 column, thermal con-
ductivity detector). To observe short-lived intermediates, pulse
radiolysis was carried out with the apparatus described before,12
which utilizes 50 ns pulses of 2 MeV electrons from a Febetron
Model 705 accelerator. All experiments were performed at
room temperature, 20 ( 2 °C.
Results and Discussion
Iron porphyrins were reduced by photochemical and by
radiation chemical methods in various solvents. Reduction of
FeIIIP to the stable FeIIP was readily achieved under anaerobic
conditions in various solvents, but reduction to FeIP was
discernible only at short times (by pulse radiolysis) or under
conditions that this species is stable for many minutes. The
photochemical experiments also provided evidence for catalyzed
reduction of CO2 to CO by some of the reduced porphyrins.
We describe first the stepwise photochemical reduction in
organic solvents, then the catalyzed CO2 reduction, and finally
the photochemical and radiolytic results in aqueous solutions.
Photochemical Reduction of FeIIIP in Organic Solvents.
Fe(III) porphyrins undergo photoreduction to the FeIIP state by
excitation at the ligand-to-metal charge transfer band.9,10 We
carried out photoreduction of ClFeIIITPP in DMF (and in
acetonitrile, dimethyl sulfoxide, tetrahydrofuran, benzonitrile,
and ethanol) and found that the process is greatly accelerated
by the addition of triethylamine. (The time of photolysis was
reduced by at least an order of magnitude.) In most experi-
ments, ClFeIIITPP was dissolved in DMF containing 5% TEA,
where it forms several possible complexes with the solvent and
the amine: Et3N(Cl)FeIIITPP, DMF(Cl)FeIIITPP, (Et3N)2FeIII-
TPP, (Et3N)(DMF)FeIIITPP. In dilute solutions the spectrum
had two main peaks at 409 and 569 nm. Upon photolysis, these
peaks decreased in intensity and were replaced with peaks at
432 and 560 nm ascribed to FeIITPP.13 Photoreduction was
very rapid with a 325 nm cutoff filter, but the rate decreased
by a factor of ∼30 with a 385 nm cutoff filter and by a factor
of ∼150 with a 420 nm cutoff filter. These results suggest that
excitation at the Soret band does not lead to reduction and that
the observed process is most likely an intramolecular electron
transfer from the axially bound TEA to the iron center.
Et3NFeIIP
9
hν8 Et3N•+ + (FeIP)-
(5)
The concentration of TEA had a stronger effect here than in
the photoreduction of FeIIITPP; decreasing TEA from 5 to 1%
decreased the rate of photoreduction of FeIITPP by a factor of
3.5 (as compared with a factor of only 1.7 for the first reduction
step). This effect may be due to weaker binding of TEA to
FeIITPP, as compared with FeIIITPP, thus requiring higher TEA
concentrations to bind more of the FeIITPP in order to increase
the effectiveness of reaction 5.
It should be noted that photoreduction in alkaline ethanol
solutions not containing TEA was considerably less efficient.
The quantum efficiency for the FeIII to FeII reduction step was
about 3 times lower, as mentioned above, but for the FeII to FeI
reduction step it was extremely low. In fact, even after
photolysis for periods 100 times longer than those required for
reduction of FeIITPP in DMF/TEA, no reduction was observable
in alkaline ethanol.
Photoreduction of FeIITPP in most cases did not lead to
quantitative conversion into FeITPP (Figure 1a). As the peaks
of FeIITPP (431 and 562 nm) gradually decayed, new peaks
ascribed to FeITPP (∼520 nm peak and weak bands at 670 and
710 nm) were formed, but upon continued photolysis additional
peaks became apparent, mainly those of (CO)FeIITPP (sharper
peaks at 420 and 540 nm), the chlorin (where one â-pyrrole
double bond is saturated), FeIIPH2 (424 and 613 nm), and then
the isobacteriochlorin (where the â-pyrrole double bonds of two
adjacent pyrrole rings are saturated), FeII-PH4 (420 and 600
nm).
Et3N(Cl)FeIIIP
9
hν8 Et3N•+ + Cl- + FeIIP
(2)
Under CO2, however, a steady-state concentration of FeITPP
was achieved within a short time, and most of the porphyrin
was converted to (CO)FeIITPP (Figure 1b). Only much longer
irradiations resulted in reduction and eventual destruction of
The Et3N•+ radical cation is capable of oxidizing the FeIIP. This
back electron transfer, however, may be suppressed by a
competing reaction14 that converts the radical cation into a

250 J. Phys. Chem. A, Vol. 101, No. 3, 1997
Grodkowski et al.
Figure 2. Photochemical production of CO in DMF solutions
containing ClFeIIITPP, 5% TEA, and saturated with CO2. Corrected
for the small yield of CO observed under He (see text). The
concentrations of ClFeIIITPP were 1 × 10-5 (9), 2 × 10-5 (b), and
(2) 5 × 10-5 mol L-1. The solutions were irradiated in a Pyrex bulb
cooled by a water jacket, placed 10 cm away from an ILC Technology
LX-300 UV lamp.
FeIITPP and then partially to FeITPP was achieved within ∼5
min, and subsequently CO production could be detected within
15 min of photolysis (under the conditions of Figure 2). The
yield of CO increased with photolysis time and reached a
turnover number of 40 CO molecules per porphyrin molecule.
During photolysis, the porphyrin solution gradually turned green,
as a result of chlorin formation, and eventually turned yellow,
indicating destruction of the macrocycle, and production of CO
eventually stopped. Reduction of CO2 to CO also takes place
by the chlorin, as confirmed by an experiment in which authentic
ClFeIIITPC was used as the starting material. The results
indicate that the efficiency of the chlorin in reducing CO2 is
similar to that of the original porphyrin.
Figure 1. Photochemical reduction of FeIITPP (5 × 10-5 mol L-1
)
solutions in DMF containing 5% TEA: (a) under Ar and (b) under
CO2. The arrows show bleaching of the FeIITPP absorptions and
formation of the FeITPP absorptions. Under CO2 the absorptions of
both FeITPP and (CO)FeIITPP are seen (see text).
the macrocycle. The relatively rapid formation of (CO)FeII-
TPP under CO2 suggests that CO is formed from CO2. The
observation of some (CO)FeIITPP even in solutions bubbled with
Ar indicates that some CO is formed from another source as
well. This will be discussed below.
The various porphyrin products formed were distinguished
also by their sensitivity to O2. FeITPP is oxidized very rapidly,
FeIITPP is oxidized within minutes, (CO)FeIITPP is oxidized
much more slowly, within hours, and all yield FeIIITPP, but
the FeII-PH4 is oxidized to products with only weak absorp-
tions, indicating that the macrocycle is probably broken.16
(CO)FeIITPP also could be converted to FeIITPP by bubbling
the solution with Ar.
Catalyzed Photoreduction of CO2. Photolysis of the iron
porphyrins in DMF solutions containing 5% TEA and saturated
with CO2 led to stepwise reduction as described above and
formation of (CO)FeIIP as the most stable product. This already
indicates some reduction of CO2 to CO under these conditions.
Photolysis of (CO)FeIITPP is known to decompose this complex
to CO and FeIITPP with a high quantum yield.17 FeIITPP can
then recombine with CO or bind a different axial ligand such
as TEA. In the latter case, photolysis will lead to further
reduction of the iron and thus enable subsequent reduction of
CO2. Binding of CO to FeIITPP, therefore, inhibits the catalyzed
photoreduction of CO2, but only to a limited extent.
To estimate the amount of CO that may be produced from
other sources, similar photolysis experiments were carried out
in solutions saturated with He instead of CO2. (He was chosen
to deoxygenate the solution, rather than Ar or N2, because He
was the carrier gas in the GC and because Ar and N2 have GC
peaks just ahead of the CO peak, which prevents accurate
measurement of low CO concentrations.) The amount of CO
formed was about 4 times lower than that formed under CO2.
These results were obtained with a porphyrin concentration of
1 × 10-5 mol L-1, but at higher porphyrin concentrations the
amount of CO formed under He was relatively smaller; with 1
× 10-4 mol L-1 porphyrin it was negligibly small compared to
that formed in the presence of CO2. The possibility that this
CO is formed following fragmentation of the porphyrin mac-
rocycle is ruled out, since CO began to be produced at the early
stages of the photolysis, before substantial amounts of the
porphyrin were decomposed. Furthermore, CO was also formed
to a similar extent in the photolysis of iron phthalocyanine under
the same conditions. The possibility that CO is a product of
TEA oxidation is also ruled out, since photochemical experi-
2-
ments with 1% TEA and 10 mM S2O8 in water, where UV
•-
photolysis leads to formation of SO4 which oxidizes TEA,
showed that the persulfate was decomposed but not CO was
formed. Most likely, CO is formed from the DMF solvent, not
by direct photolysis, but only in the presence of the porphyrin,
i.e., through partial complexation of FeP with DMF (instead of
TEA) and direct photolysis of the complex HCON(CH3)2-
Continued photolysis builds up concentrations of CO much
greater than the concentration of the porphyrin, and these could
be analyzed by gas chromatography. For example, with a
concentration of 2 × 10-5 mol L-1 FeIIITPP, photoreduction to
•
FeTPP to result in oxidation of the solvent-ligand to CON-

Iron Porphyrin-Catalyzed Reduction of CO2
J. Phys. Chem. A, Vol. 101, No. 3, 1997 251
•
(CH3)2. This radical may decompose to CO and N(CH3)2, as
indicated by the radiolysis of neat DMF, which yields consider-
able amounts of CO (G ) 0.27 µmol J-1) probably via the same
intermediate. This is further confirmed by replacing DMF with
acetonitrile as solvent for the photoreduction of FeP with TEA.
The yield of CO in solutions saturated with CO2 was similar in
the two solvents, but the yield of CO in the absence of CO2
was 30 times lower in acetonitrile than in DMF.
centered at 800 nm, and the chlorin, FeIIPH2, with a sharp peak
at 620 nm. Upon standing for several hours, the phlorin anion
changed to the chlorin,19 while the FeIP reverted to FeIIP.
Exposure to O2 converted these products to the chlorin, FeIII-
PH2 (655 nm), and the original FeIIIP.
Photolysis of FeIIITM2PyP in the presence of TEA and
NaHCO3 was also carried out in solutions that were bubbled
with CO2. Bubbling CO2 decreased the pH to 8.8, and since at
this pH the TEA is protonated to about 99%, the photolysis
was less efficient, by a factor of 20 for the initial stage and by
a factor of about 100 for the second reduction step. The
product at this pH was predominantly the chlorin, FeIIPH2 (peaks
at 418 and 616 nm). In an attempt to increase the efficiency of
photoreduction, we replaced TEA (pKa ) 10.7) with TEOA (pKa
) 7.8). The quantum yield for photoreduction of FeIIITM2PyP
was practically identical with these two amines and very similar
to that found with FeIIITPP in DMF. Photoreduction of FeII-
TM2PyP, however, was much less efficient with TEOA, by a
factor of ∼500 in 5% TEOA solutions at pH 11 and by a factor
of ∼4000 in neutral solutions. This may result from weaker
binding of TEOA to the FeII center and/or lower reduction
potential, as compared with TEA, owing to the lower electron
density on the amine nitrogen.
The yield of CO from CO2 was determined by subtracting
the amount of CO formed under He from that measured under
CO2. The corrected yield as a function of photolysis time
(Figure 2) is found to be similar for 1 × 10-5, 2 × 10-5, and
5 × 10-5 mol L-1 FeTPP. Under the conditions used in these
experiments, all the light is absorbed by the solution even at
the lowest porphyrin concentration used, and therefore, the yield
does not increase with concentration. The plateau level also
remains the same, indicating faster decomposition of the
porphyrin at higher concentrations. This may be due to
competition between reaction of Fe0TPP with CO2 and reaction
of Fe0TPP with other reduced porphyrin species leading to the
reduction or destruction of the macrocycle. The former reaction
is pseudo-first-order and thus independent of initial FeIIITPP
concentration, but the latter process is second-order and thus
will increase in importance at high absorbances (high local
concentrations of reduced porphyrin). As a result, decomposi-
tion of the macrocycle is more effective at higher concentrations.
With 1 × 10-5 mol L-1 FeIIITPP, the yield of CO at the plateau
level corresponds to a turnover number of 70.
Continued photolysis of FeIITM2PyP in aqueous solutions
containing 5% TEA, 0.1 mol L-1 NaHCO3, and CO2 led to
production of CO at concentrations similar to those of the
porphyrin. Without NaHCO3 the pH was lower, and the amount
of CO formed was at least 10 times lower.
In contrast with the above results for FeIIITM2PyP, the other
water-soluble porphyrins examined, FeIIITM3PyP and FeIIITSPP,
when photoreduced under similar conditions, formed mainly the
FeIIP products, with only small yields of the chlorin and
decomposition products. No stable FeIP state was observed,
and no CO was detected under the same conditions that FeIII-
TM2PyP gave considerable CO yield. It should be noted,
however, that when FeIIITSPP was photolyzed in DMF solutions,
substantial amounts of CO were detected.
Various iron porphyrins have been studied under the same
conditions for comparison with the FeTPP. Both the rate of
CO production and its total yield varied slightly. The rate was
similar with TPP and TMP, was slightly lower (∼20%) with
TCPP and TFPP, and slightly higher (10%) with OBTMP. OEP
was less effective (by 30%), PFP even less (50%), the
tetrapropylporphycene gave only one-third of the yield, and the
phthalocyanine was much less effective.
In electrochemical reduction of CO2 with iron porphyrins,7
it has been shown that the process is catalyzed by water and by
Bro¨nsted acids in general. Therefore, no effort was made in
the present study to dry the organic solvents from traces of
water. In fact, addition of 0.1-5% water was found to increase
the rate of production of CO from CO2. However, the rate of
CO production from the DMF solvent also was increased by
addition of water, and it was difficult to determine the effect of
water on the catalytic reduction of CO2. This effect could be
quantitated in acetonitrile solutions, where the yield of CO from
other sources was very low. Addition of 1% water increased
the yield of CO from CO2 by 30% while the yield in the absence
of CO2 remained very low. Experiments using water as solvent
are discussed below.
Radiolytic Reduction of Iron Porphyrins in Aqueous
Solutions. Radiolytic reduction of Fe(III) porphyrins in aque-
ous20 and in alcohol21 solutions has been shown to lead to
efficient production of Fe(II) porphyrins, i.e., with a radiolytic
yield corresponding to quantitative reaction of all reducing
radicals with the porphyrin. We obtained the same results with
the water-soluble porphyrins upon γ-irradiation in aqueous
•
solutions containing formate as a scavenger for H• and OH,
-
-
i.e., reduction of the porphyrin by all eaq and C˙ O2 radicals.
H2O Df eaq-, H•, •OH, H+, H2, H2O2
(6)
(7)
(8)
(9)
-
H•/•OH + HCO2- f H2/H2O + C˙ O2
Photochemical Reduction of Iron Porphyrins in Aqueous
Solutions. These experiments were carried out mainly with
FeIIITM2PyP, which yields the most stable FeIP state. Pho-
tolysis of this porphyrin in aqueous solutions containing 5%
TEA and 0.05 mol L-1 NaHCO3 at pH 11.4, deoxygenated by
bubbling with Ar, resulted in rapid reduction of FeIIITM2PyP
(∼430 and ∼580 nm) to FeIITM2PyP (442 and 564 nm).
Subsequent photoreduction was ∼40 times less efficient and
produced FeITM2PyP. This product exhibits a weak maximum
at ∼430 nm and broad absorptions around 520 nm and around
600-800 nm, similar to those reported previously18 and to those
described above for FeITPP. Along with FeITM2PyP, two other
reduction products were formed in small yields, the phlorin
anion, FeIIPH- (the product of two-electron reduction and single
protonation at a meso position), exhibiting a very broad band
ClFeIIIP + eaq- f Cl- + FeIIP
ClFeIIIP + C˙ O2- f Cl- + FeIIP + CO2
Further reduction by γ-radiolysis led to formation of various
products from the various porphyrins studied under different
conditions. Radiolytic reduction of FeIITM2PyP in alkaline
solutions led to the spectral changes shown in Figure 3, i.e.,
decrease of the two main peaks of FeIITM2PyP, shift of the
Soret peak to 430 nm, and buildup of broad absorptions at 650-
800 nm, indicating formation of FeITM2PyP. As in the
photolysis experiments, this product was formed along with
small yields of the phlorin anion (broad 800 nm absorption)
and the chlorin (sharp ∼620 nm peak). The FeITM2PyP is

252 J. Phys. Chem. A, Vol. 101, No. 3, 1997
Grodkowski et al.
Figure 3. Radiolytic reduction of FeIITM2PyP (5 × 10-5 mol L-1
)
in deoxygenated aqueous solutions containing 0.1 mol L-1 HCOO-
and 1 mol L-1 KOH. The arrows show the stepwise increase or
decrease in absorbance after each irradiation with a dose of 29 Gy.
stable under these conditions for many hours while the phlorin
anion changes to the chlorin.19 At lower pH, FeITM2PyP
decayed more rapidly. For example, at pH 10 it decayed over
20 h and at pH 7 within 2 h. In neutral solutions, the radiolytic
yield of FeITM2PyP was about half that observed in alkaline
solutions and was produced along with considerable yields of
the chlorin (peaks at 424 and 620 nm).
Several experiments were also carried out by using 2-PrOH
as an OH scavenger instead of formate, and surprisingly, the
results were different. Instead of reduction of FeIITM2PyP to
FeITM2PyP in alkaline solutions, the predominant product was
the phlorin anion. The difference between these two systems
may be due to a different mechanism of reaction of the two
reducing species. Whereas C˙ O2- reduces FeIITM2PyP to FeI-
TM2PyP, the (CH3)2C˙ O- and (CH3)2C˙ OH radicals formed from
2-PrOH add to the metal center of FeIITM2PyP to form a
transient HO(CH3)2C-FeIIITM2PyP, which does not decompose
to FeITM2PyP rapidly. Instead, this species survives in solution
until it receives another electron. The electron is then added
preferentially to the ligand to form the radical anion HO-
(CH3)2C-FeIIITM2PyP•-. Release of (CH3)2CO and H+ from
this species and protonation at the meso position of the porphyrin
will form the phlorin anion observed.
Radiolytic reduction of FeIITM3PyP did not yield a stable
FeITM3PyP under any conditions. Reduction in neutral solu-
tions resulted in the formation of the chlorin (peaks at 428 and
620 nm) whereas reduction in alkaline solutions formed the
phlorin anion.22
Figure 4. (a) Differential absorption spectrum recorded by pulse
radiolysis of aqueous solutions containing 5 × 10-5 mol L-1 FeII-
TM2PyP and 0.01 mol L-1 HCOO- at pH 3.7, 20 µs after the pulse.
(b) Absorbance (×10) vs time over four time scales: 0.1 µs, 5 µs, 100
µs, and 4 ms per point, covering 2000 points, monitored at pH 8.6. (c)
Absorbance (×10) vs time over three time scales: 0.1, 3, and 700 µs
per point, covering 2000 points, monitored in solutions containing
bicarbonate and CO2 at pH 6.7.
Pulse Radiolysis Studies of Iron Porphyrins in Aqueous
Solutions. To observe the short-lived species that lead to the
above products, we utilized pulse radiolysis, where the solutions
were irradiated with a short electron pulse and the differential
absorption spectra were monitored within microseconds to
milliseconds after the pulse. For these experiments, the starting
materials, FeIIIP, were first reduced to FeIIP by H2 on Pd and
then pulse irradiated. The differential spectrum in Figure 4a
shows bleaching of the FeIITM2PyP absorption at 550 nm and
formation of strong absorption below 520 nm and broad
absorption at 600-840 nm. These changes are in agreement
with those found by γ-radiolysis for the reduction of FeII-
TM2PyP to FeITM2PyP. The results indicate that even in
slightly acidic solutions FeITM2PyP is the first product of
reduction of FeIITM2PyP. The transient spectra obtained with
FeIITM3PyP, FeIITM4PyP, and FeIITSPP are generally quite
similar to those obtained with FeIITM2PyP and indicate forma-
tion of the FeIP state by one-electron reduction. The alternative
possibility, i.e., formation of the π-radical anions FeIIP•-, is ruled
out, since no strong broad peaks around 700 nm, characteristic
for π-radical anions, have been observed in any of the above
experiments. Thus, all the FeII porphyrins studied in aqueous
solutions are reduced first at the metal to yield FeIP.23
The stability of the FeIP species was dependent on the
porphyrin and the medium. FeITM4PyP and FeITM3PyP
decayed over several milliseconds, both in neutral and in alkaline
solutions, via second-order processes. A rate constants of 2k
) (1.4 ( 0.4) × 108 L mol-1 s-1 was derived for FeITM4PyP
at pH 6.7, and a rate constant of 2k ) (1.0 ( 0.3) × 108 L
mol-1 s-1 was calculated for FeITM3PyP at pH 6.9 and 3.6.
FeITM2PyP exhibited a different decay behavior. Despite the
above findings from the γ-radiolysis experiments that this
species is stable under certain conditions, at least partial decay

Iron Porphyrin-Catalyzed Reduction of CO2
J. Phys. Chem. A, Vol. 101, No. 3, 1997 253
was observed in the millisecond range. This observation is
rationalized by the existence of four atropisomers of the
TM2PyP ligand (due to inhibition of free rotation of the four
N-methylpyridinium groups at the meso positions).24 For
porphyrins with more highly inhibited rotation, such as the
tetrakis(2-(N-methylisonicotinamido)phenyl)porphyrin, where
the four isomers could be separated, it has been shown that the
pKa for protonation of the free bases varied between 2.2 and
2.9, and the rates of metal incorporation varied by an order of
magnitude for the different isomers.25 Since decay of the FeI-
TM2PyP is probably affected by protonation, it is likely that
the different isomers will exhibit different decay behavior.
The decay of FeITM2PyP was monitored at 700 nm by pulse
radiolysis of solutions containing 0.01 mol L-1 formate ions
under various conditions (two representative traces are shown
in Figure 4b,c). At pH 8.6 (buffered with borate), 30% of the
initial absorbance decayed within ∼3 ms, 15% decayed within
∼30 ms, and the remaining absorbance was stable beyond the
limit of the pulse radiolysis detection, i.e., >10 s. This
remaining absorbance was also observed several minutes after
irradiation in the γ source and decayed within hours. At pH
6.7 (in solutions containing formate but no added buffer), the
decay was similar, and the remaining absorbance was slightly
lower, 50% of the initial value. At pH 6.7 with added phosphate
buffer (2 × 10-3 mol L-1), larger fractions of the absorbance
decayed in each step, within ∼3 and ∼100 ms, leaving only
30% for the long-lived fraction. When the pH was decreased
to 2.9 and 2.3, or when the solution was at pH 6.9 but was
bubbled with CO2, the fraction of remaining absorbance at 700
nm was only 10%; most of the absorbance decayed in two
distinct steps, within ∼2 and ∼300 ms. The decay of FeI-
TM2PyP was also monitored throughout the wavelength range
studied (from 450 to 840 nm). At most wavelengths, the
differential absorbance returned to zero, indicating that most
of the original FeIITM2PyP is recovered; some absorbance
remained around 610 nm, which indicates formation of the
chlorin.
and small amounts of the chlorin are formed, suggests that the
decay takes place via disproportionation.
2(FeIP)- a FeIIP + (Fe0P)2-
(10)
Possibly, very little Fe0P is present in this equilibrium, but the
reaction is shifted to the right due to reaction of Fe0P with water,
H+, or CO2. These reactions lead to formation of short-lived
hydride or carboxylate complexes.
(Fe0P)2- + H2O f (HFeIIP)- + OH-
(Fe0P)2- + CO2 f (CO2-FeIIP)2-
(11)
(12)
Subsequent protonation of these species will lead to formation
of H2 or CO.
(HFeIIP)- + H2O f H2 + FeIIP + OH-
(13)
(CO2-FeIIP)2- + H2O f CO + FeIIP + 2OH- (14)
Protonation of the macrocycle to form the chlorin also takes
place concurrently.
Earlier studies on iron protoporphyrin27 suggested that the
FeIP state, even at pH 10, undergoes protonation to form the
hydride, which reacts with water to yield H2 and FeIIIP. The
present results indicate, however, that FeITM2PyP does not
undergo protonation in neutral solutions. The difference is
clearly due to the difference in reduction potentials of these
two porphyrins. Electrochemical results18 indicated that the
reduction waves for FeII/FeITM2PyP and for H+/H2 overlapped
at pH < 4.2 but became more separated as the pH increased.
Thus, protonation of FeITM2PyP should be unimportant in
neutral solutions.28 Even at pH 2.9, our results indicate second-
order decay behavior and no formation of HFeIIIP or FeIIIP,
suggesting that direct protonation of FeITM2PyP does not
compete effectively with the proton-enhanced second-order
decay. Electrochemical results with FeTPP29 also indicated that
reduction of H+ to H2 is catalyzed by the Fe0 and not by the
FeI state.
The first and second decay steps were sufficiently separated
in time to permit kinetic fitting of each step independently of
the other. The first step obeys second-order rate law quite
closely, at different dose per pulse, and the calculated rate
constants were (1.5 ( 0.5) × 108 L mol-1 s-1 in neutral solution
and about 50% higher in the presence of CO2 or in acidic
solutions. The slower step was also fitted to second-order
kinetics, although in this case, because of the smaller signals
and increased noise (analyzing light instability), the second-
Summary and Conclusions
Fe(III) porphyrins are readily reduced by photochemical and
radiation chemical methods to the stable FeIIP products and then
to FeIP. Photochemical reduction in DMF solutions takes place
by excitation of the ligand-to-metal charge transfer band, where
the ligand is an axially bound triethylamine. The quantum yield
for the first reduction step is ∼0.05 but for the second reduction
only ∼0.01. Continued photolysis in the presence of CO2 results
in the formation of CO. However, the porphyrin ligand is also
partially reduced, first to the chlorin, which also catalyzes
photochemical reduction of CO2, but eventually the porphyrin
macrocycle is destroyed and CO production stops before a
turnover number of 100 is achieved. This value is considerably
lower than the turnover number achieved in the electrochemical
reduction of CO2 by iron porphyrins,7 probably because pho-
toinduced side reactions in the homogeneous system lead to
attack on the porphyrin macrocycle. Several substituted por-
phyrins are examined, but no great enhancement of catalytic
activity or long-term porphyrin stability is found. Encapsulation
or binding of the porphyrin may offer some advantage in this
respect. It is noted, however, that despite the susceptibility of
order fit and the derived rate constants (106-107 L mol-1 s-1
)
were considerably less certain. The second-order decay may
involve reaction of an unstable isomer of FeITM2PyP with any
of the isomers, until the unstable ones are depleted. If the four
isomers of FeITM2PyP have different reactivities, up to four
decay steps may be observed. In neutral solutions, two decay
steps are observed, presumably for two or three isomers, and
the remaining one or two isomers are stable. In acidic solutions,
or in the presence of CO2, the first decay step is larger and
appears to encompass both isomers that were seen to decay in
two steps in neutral solutions; i.e., the decay for the second
step is accelerated by the acid or CO2 and merges with the first
step. The first step is also expected to be accelerated, but its
rate does not appear to increase since it already approaches the
diffusion-controlled limit.26 The isomer that was stable in
neutral solution decays in acidic solutions (but more slowly than
the second step in neutral).
The second-order behavior of the decay processes, along with
the findings that most of the original FeIITM2PyP is recovered

254 J. Phys. Chem. A, Vol. 101, No. 3, 1997
Grodkowski et al.
(15) Hoshino, M.; Ueda, K.; Takahashi, M.; Yamaji, M.; Hama, Y. J.
Chem. Soc., Faraday Trans. 1992, 88, 405.
the porphyrin macrocycle to stepwise reduction and eventual
cleavage, the turnover numbers for CO production found in the
present study compare favorably with those reported in photo-
chemical studies with different complexes of iron or nickel.4a,5a
Radiolytic reduction was studied only in aqueous solutions
to avoid the possibility of attachment of radiolytically produced
organic radicals to the FeIIP species (to form R-FeIIIP). The
FeIP state is less stable in water than in organic solvents. Only
some of the atropisomers of FeITM2PyP exhibit long-term
stability in water, mainly in alkaline solutions. Time-resolved
studies suggest that the FeIP species does not reduce H+ or CO2
directly, but rather disproportionates to FeIIP and Fe0P, and the
latter serves as the reductant (in line with the conclusions of
the electrochemical studies)7. Catalyzed reduction of CO2 to
CO was less efficient in water than in organic solvents, although
the reaction in organic solvents is enhanced by small amounts
of water.
(16) The identity of the reduced products was confirmed by experiments
with an authentic sample of ClFeIIITPC (tetraphenylchlorin). Solution of
this compound in DMF with 5% TEA exhibited main peaks at 405 and
624 nm. Upon photolysis this was reduced to the FeIITPC (424 and 614
nm), and further photoreduction produced FeIITPiBC (420 and 600 nm).
Exposure to O2 resulted in the oxidation of FeIITPC to FeIIITPC, but FeII-
TPiBC gave products with only weak broad absorptions, probably due to
cleavage of the macrocycle.
(17) Hoshino, M.; Ueda, K.; Takahashi, M.; Yamaji, M.; Hama, Y.;
Miyazaki, Y. J. Phys. Chem. 1992, 96, 8863.
(18) Chen, S.-M.; Sun, P.-J.; Su, Y. S. J. Electroanal. Chem. 1990, 294,
151.
(19) Transformation of phlorin anion to chlorin has been observed
previously with TM2PyP (Mosseri, S.; Nahor, G. S.; Neta, P.; Hambright,
P. J. Chem. Soc., Faraday Trans. 1991, 87, 2567) and other porphyrin
derivatives (Fuhrhop, J.-H. Struct. Bonding (Berlin) 1974, 18, 1).
(20) Brault, D.; Bizet, C.; Morliere, P.; Rougee, M.; Land, E. J.; Santus,
R.; Swallow, A. J. J. Am. Chem. Soc. 1980, 102, 1015. Brault, D.; Neta,
P. J. Am. Chem. Soc. 1981, 103, 2705.
Acknowledgment. This research was supported by the
Division of Chemical Sciences, Office of Basic Energy Sciences,
U.S. Department of Energy. We thank Prof. J.-M. Save´ant for
helpful discussions.
(21) Baral, S.; Neta, P.; Hambright, P. Radiat. Phys. Chem. 1984, 24,
245.
(22) Radiolytic reduction of FeIITM3PyP at pH 12 in aqueous solutions
containing 0.013 mol L-1 2-PrOH resulted in the formation of the phlorin
anion with a strong peak at 816 nm. This product was readily oxidized by
O2 to yield FeIIITM3PyP. In the absence of O2 it was stable and did not
change to the chlorin (in contrast with the findings for the ortho isomer,
FeTM2PyP). Radiolytic reduction of FeIITM3PyP at pH 12 in the presence
of 0.01 mol L-1 HCOO- resulted in the formation of a product with a
different absorption, a strong broad peak at 666 nm. This product also
was stable in the absence of O2 and also was oxidized by O2 to the porphyrin
FeIIITM3PyP. Thus, it behaves like the phlorin anion, and it is not clear
why the spectrum is different than that found in the presence of 2-PrOH.
References and Notes
(1) Catalytic ActiVation of Carbon Dioxide; Ayers, W. M., Ed. ACS
Symp. Ser. 1988, No. 363. Behr, A. Carbon Dioxide ActiVation by Metal
Complexes; VCH: Weinheim, 1988. Electrochemical and Electrocatalytic
Reactions of Carbon Dioxide; Sullivan, B. P., Ed.; Elsevier: Amsterdam,
1993.
(2) Sullivan, B. P.; Bolinger, C. M.; Conrad, D.; Vining, W. J.; Meyer,
T. J. J. Chem. Soc., Chem. Commun. 1985, 1414.
(23) One-electron reduction of FeIIP, in principle, may lead to production
of either FeIP or FeIIP-. Both species are unstable in neutral aqueous
solutions and may decay by disproportionation and protonation. Decay of
metalloporphyrin π-radical anions (MP•-) is known to lead to formation
of either a phlorin anion, MPH-, in alkaline solutions or a chlorin, MPH2,
in neutral solutions. Both of these products have been observed with the
iron porphyrins under different conditions. This result, however, is not
sufficient evidence for formation of the radical anion since FeIP also may
decay to give similar products. The decay may be via disproportionation,
involving electron transfer from the metal to the ligand, in a process that is
enhanced by protonation (similar to that suggested before for nickel
porphyrins: Nahor, G. S.; Neta, P.; Hambright, P.; Robinson, L. R.;
Harriman, A. J. Phys. Chem. 1990, 94, 6659).
(3) Fisher, B.; Eisenberg, R. J. Am. Chem. Soc. 1980, 102, 7361. Beley,
M.; Collin, J.-P.; Ruppert, R.; Sauvage, J.-P. J. Am. Chem. Soc. 1986, 108,
7461. Fujita, E.; Haff, J.; Sanzenbacher, R.; Elias, H. Inorg. Chem. 1994,
33, 4627.
(4) (a) Grant, J. L.; Goswami, K.; Spreer, L. O.; Otvos, J. W.; Calvin,
M. J. Chem. Soc., Dalton Trans. 1987, 2105. (b) Craig, C. A.; Spreer, L.
O.; Otvos, J. W.; Calvin, M. J. Phys. Chem. 1990, 94, 7957. (c) Kelly, C.
A.; Mulazzani, Q. G.; Venturi, M.; Blinn, E. L.; Rodgers, M. A. J. J. Am.
Chem. Soc. 1995, 117, 4911.
(5) (a) Fujita, E.; Brunschwig, B. S.; Ogata, T.; Yanagida, S. Coord.
Chem. ReV. 1994, 132, 195. (b) Ogata, T.; Yanagida, S.; Brunschwig, B.
S.; Fujita, E. J. Am. Chem. Soc. 1995, 117, 6708. (c) Matsuoka, S.;
Yamamoto, K.; Ogata, T.; Kusaba, M.; Nakashima, N.; Fujita, E.; Yanagida,
S. J. Am. Chem. Soc. 1993, 115, 601. (d) Ogata, T.; Yamamoto, Y.; Wada,
Y.; Murakoshi, K.; Kusaba, M.; Nakashima, N.; Ishida, A.; Takamuku, S.;
Yanagida, S. J. Phys. Chem. 1995, 99, 11916.
(24) Rodgers, K. R.; Reed, R. A.; Su, O.; Spiro, T. G. Inorg. Chem.
1992, 31, 2688.
(6) Hawecker, J.; Lehn, J.-M.; Ziessel, R. J. Chem. Soc., Chem.
Commun. 1985, 56.
(25) Hambright, P.; Turner, A.; Cohen, J. S.; Lyon, R. C.; Katz, A.;
Neta, P.; Adeyemo, A. Inorg. Chim. Acta 1987, 128, L11. Valiotti, A.;
Adeyemo, A.; Williams, R. F. X.; Ricks, L.; North, J.; Hambright, P. J.
Inorg. Nucl. Chem. 1981, 43, 2653.
(7) Hammouche, M.; Lexa, D.; Save´ant, J.-M.; Momenteau, M. J.
Electroanal. Chem. Interfacial Electrochem. 1988, 249, 347. Hammouche,
M.; Lexa, D.; Momenteau, M.; Save´ant, J.-M. J. Am. Chem. Soc. 1991,
113, 8455. Bhugun, I.; Lexa, D.; Save´ant, J.-M. J. Am. Chem. Soc. 1994,
116, 5015. Bhugun, I.; Lexa, D.; Save´ant, J.-M. J. Am. Chem. Soc. 1996,
118, 1769.
(8) Fujita, E.; Szalda, D. J.; Creutz, C.; Sutin, N. J. Am. Chem. Soc.
1988, 110, 4870. Creutz, C.; Schwarz, H. A.; Wishart, J. F.; Fujita, E.;
Sutin, N. J. Am. Chem. Soc. 1989, 111, 1153; 1991, 113, 3361. Fujita, E.;
Creutz, C.; Sutin, N.; Szalda, D. J. J. Am. Chem. Soc. 1991, 113, 343. Fujita,
E.; Creutz, C.; Sutin, N.; Brunschwig, B. S. Inorg. Chem. 1993, 32, 2657.
(9) Bizet, C.; Morliere, P.; Brault, D.; Delgado, O.; Bazin, M.; Santus,
R. Photochem. Photobiol. 1981, 34, 315. Hoshino, M.; Ueda, K.; Takahashi,
M.; Yamaji, M.; Hama, Y. J. Chem. Soc., Faraday Trans. 1992, 88, 405.
(10) Hendrickson, D. N.; Kinnaird, M. G.; Suslick, K. S. J. Am. Chem.
Soc. 1987, 109, 1243. Suslick, K. S.; Watson, R. A. New J. Chem. 1992,
16, 633.
(11) The mention of commercial equipment or material does not imply
recognition or endorsement by the National Institute of Standards and
Technology, nor does it imply that the material or equipment identified are
necessarily the best available for the purpose.
(12) Neta, P.; Huie, R. E. J. Phys. Chem. 1985, 89, 1783.
(13) Brault, D.; Rougee, M. Biochemistry 1974, 13, 4598.
(14) DeLaive, P. J.; Foreman, T. K.; Giannotti, C.; Whitten, D. G. J.
Am. Chem. Soc. 1980, 102, 5627.
(26) The diffusion-controlled limit for average neutral species in water
is 6 × 109 L mol-1 s-1. This value will be decreased by the Coulombic
interaction between charged species. For doubly charged ions the decrease
is by a factor of 50 and for triply charged ions by a factor of 104 (Debye,
P. Trans. Electrochem. Soc. 1942, 82, 265). The triply charged FeITM2PyP
does not exhibit the full effect of the three charges since these are not
localized at the same point but rather at distinct points at the edges of the
molecule (Nwaeme, J.; Hambright, P. Inorg. Chem. 1984, 23, 1990).
(27) Sorek, Y.; Cohen, H.; Meyerstein, D. J. Chem. Soc., Faraday Trans.
1 1985, 81, 233.
(28) Protonation may be important for the protoporphyrin because its
reduction potential is much more negative. Iron protoporphyrin is expected
to have an FeII/FeI reduction potential about 0.2 V more negative than that
for FeIITPP (by comparison with the value for FeIIOEP: Lexa, D.; Save´ant,
J.-M.; Wang, D. L. Organometallics 1986, 5, 1428) whereas pyridyl
porphyrins have reduction potentials more positive than that of TPP.
Therefore, the likelihood of protonation and water reduction by the FeIP
state is diminished and becomes unimportant in our experiments.
(29) Bhugun, I.; Lexa, D.; Save´ant, J.-M. J. Am. Chem. Soc. 1996, 118,
3982.