Reduction of cobalt and iron corroles and catalyzed reduction of CO2

Article abstract of DOI:10.1021/jp013668o

The role of cobalt and iron corroles in catalytic CO₂ reduction has been studied. Chemical, electrochemical, and photochemical reductions of the stable metal corroles Ph₃PCo^III(tpfc) (tpfc = 5,10,15-tris(pentafluorophenyl)

Full text of DOI:10.1021/jp013668o

4772
J. Phys. Chem. A 2002, 106, 4772-4778
Reduction of Cobalt and Iron Corroles and Catalyzed Reduction of CO2
Jan Grodkowski† and Pedatsur Neta*
Physical and Chemical Properties DiVision, National Institute of Standards and Technology,
Gaithersburg, Maryland 20899-8381
Etsuko Fujita*
Chemistry Department, BrookhaVen National Laboratory, Upton, New York 11973
Atif Mahammed, Liliya Simkhovich, and Zeev Gross*
Department of Chemistry and Institute of Catalysis Science and Technology, TechnionsIsrael Institute of
Technology, Haifa 32000, Israel
ReceiVed: October 1, 2001; In Final Form: January 30, 2002
The role of cobalt and iron corroles in catalytic CO2 reduction has been studied. Chemical, electrochemical,
and photochemical reductions of the stable metal corroles Ph3PCoIII(tpfc) (tpfc ) 5,10,15-tris(pentafluo-
rophenyl)corrole), ClFeIV(tpfc), and ClFeIV(tdcc) (tdcc ) 5,10,15-tris(2,6-dichlorophenyl)corrole) have been
carried out in acetonitrile solutions. Stepwise reduction to the [MII(tpfc)]- and [MI(tpfc)]2- states was observed
in all cases. Gradual reduction with sodium amalgam permitted recording of the optical absorption spectra of
the various oxidation states and showed that the MI state reacts with CO2. Cyclic voltammetry in Ar-saturated
acetonitrile solutions permitted determination of the following half-wave potentials: for Ph3PCoIII(tpfc), 1.11
V, 0.72 V, -0.42 V (Epc), -1.44 V, -2.3 V (Epc); for ClFeIV(tpfc), 0.44 V, -1.01 V (Epc), -1.60 V, -2.2
V (Epc); for ClFeIV(tdcc), 0.24 V, -1.18 V (Epc), -1.78 V vs SCE with a scan rate of 0.1 V s-1. Cyclic
voltammetry in CO2-saturated solutions indicated that the CoI and FeI complexes react with CO2 and that the
reduced Fe(tdcc) complex is the most efficient electrocatalyst for CO2 reduction, showing the largest catalytic
currents among these corroles. Photochemical reduction in CO2-saturated acetonitrile solutions containing
p-terphenyl (TP) as a sensitizer and triethylamine (TEA) as a reductant led to production of CO and H2.
These experiments also show that Fe(tdcc) is more effective than the other corroles as a CO2 reduction catalyst.
The present finding that the MI oxidation states of the cobalt and iron corroles can react with CO2 is in
contrast with the case of the respective porphyrins and phthalocyanines, which do not react with CO2 until
they are reduced beyond the MI state.
Introduction
porphyrins. For these reasons, corroles stabilize the metal center
at a higher oxidation state than porphyrins. On the basis of their
behavior in oxidation reactions,9,10 it is expected that their
reduction reactions will take place at more negative potentials
than the corresponding reactions of metalloporphyrins. If the
reduction to the MI state occurs at a very negative potential the
MI corrole may react with CO2. Therefore, chemical, electro-
chemical, and photochemical methods were used to study the
reduction reactions of several corroles and their interaction with
CO2.
Several transition metal complexes are known to act as
electron-transfer mediators for photochemical1,2 or electrochemi-
cal3 reduction of CO2.4 Recent studies have shown that iron
and cobalt porphyrins (MP) are effective homogeneous catalysts
for the electrochemical5 and photochemical6,7 reduction of CO2
to CO and formic acid. A more recent study has shown that Fe
and Co phthalocyanines also catalyze photoreduction of CO2
in homogeneous solutions.8 With both porphyrins and phtha-
locyanines it was found that the [MIP]- complexes do not react
with CO2 and that the products of the next reduction step are
the reactive species. These products were formulated as ze-
rovalent metalloporphyrins [M0P]2- and monovalent cobalt
phthalocyanine radical anion [CoIPc•]2-. The present study is
aimed at characterizing the species produced by reduction of
cobalt and iron corroles and examining their role in photochemi-
cal reduction of CO2. Corroles lack one meso bridge and thus
have a smaller macrocycle cavity than porphyrins. Their four
nitrogens bear three protons, as compared with two protons in
Experimental Section
The cobalt corrole was synthesized as the triphenylphosphine
complex, Ph3PCoIII(tpfc) (tpfc is the trianion of 5,10,15-tris-
(pentafluorophenyl)corrole) and the iron corroles as the chloride
complexes, ClFeIV(tpfc) and ClFeIV(tdcc), as described be-
fore.11,12
Triethylamine (TEA) was obtained from Aldrich and was
refluxed over sodium and distilled under nitrogen. Acetonitrile
(MeCN) used as solvent for sodium amalgam reduction was
purified by published methods13 and stored in a vacuum over
activated molecular sieves. Dried and distilled MeCN was used
* Corresponding authors.
† On leave from the Institute of Nuclear Chemistry and Technology,
Warsaw, Poland.
10.1021/jp013668o CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/06/2002

Reduction of Cobalt and Iron Corroles and CO2
J. Phys. Chem. A, Vol. 106, No. 18, 2002 4773
for electrochemical experiments and dry spectrograde MeCN
was used for photochemical experiments. All experiments were
performed at room temperature, (22 ( 2) °C.
Cyclic voltammetry measurements were carried out with
MeCN solutions containing 0.5 mmol L-1 of the metal corrole
and 0.05 mol L-1 Bu4NPF6 as the electrolyte, with scan rates
ranging from 10 mV s-1 to 10 V s-1. A conventional three-
electrode system consisting of a glassy carbon working elec-
trode, a Pt counter electrode, and a standard calomel reference
electrode was used with a BAS100B electrochemical analyzer.
Ferrocene (E1/2 ) 0.39 V vs SCE) was added as an internal
standard at the end of all experiments. All potentials are given
with reference to the standard calomel electrode. A glassy carbon
electrode was polished before each scan and several scans were
taken at different ranges of potential to determine the various
reversible reduction and oxidation steps.
The UV-vis spectra of reduced corroles were recorded in
MeCN solutions following stepwise reduction of Ph3PCoIII(tpfc),
ClFeIV(tpfc), and ClFeIV(tdcc) by sodium amalgam (0.5% Na
in Hg) in sealed glassware equipped with an optical cell. The
solution was exposed to the sodium amalgam for a few seconds
and then returned to the optical cell compartment and its
spectrum was recorded. The process was repeated many times
and the end point of each reduction was carefully monitored
by the loss of isosbestic points.
Photochemical experiments were performed with a 300-W
xenon lamp, using water filters to absorb the IR and Pyrex filters
to absorb the UV (λ < 310 nm). Solutions were placed in an
irradiation cell equipped with a water cooling system to maintain
the solution temperature, were bubbled with He or CO2, and
then were photolyzed. For recording of optical absorption spectra
before and after photolysis, the solution was irradiated in a
quartz optical cell, with 2 mm or 10 mm optical path length,
attached to a quartz tubing and sealed with a septum. For
analysis of the gases evolved, 35 mL of the solution was placed
in a 43 mL Pyrex bulb cooled by a water jacket, saturated with
the required gas, and then photolyzed. After various intervals,
the headspace was analyzed for CO and H2 by gas chromatog-
raphy (Carboxen-1000 column, thermal conductivity detector).
The total amounts of these gases in both solution and gas phase
were calculated from the known solubilities and are reported
as concentration as if all the gases remained in the photolyzed
solution.
Figure 1. Cyclic voltammograms of Ph3PCoIII(tpfc) in MeCN solutions
saturated with Ar (solid line) and with CO2 (dotted line), scanned
between +0.2 V and -2.0 V vs SCE. A wider scan of the Ar-saturated
solution between +0.2 V and -2.6 V is shown by the dashed line.
Scan rate 0.1 V s-1
.
TABLE 1: Half-Wave Potentials for Reduction of the
Corroles in MeCN
corrole
E1/2, V vs. SCE
Ph3PCoIII(tpfc)a
ClFeIV(tpfc)
-0.42b, -1.44, -2.3b
0.44, -1.01b, -1.60, -2.2b
0.24, -1.18b, -1.78
ClFeIV(tdcc)
b
a Oxidation took place at E1/2 0.72 and 1.11 V. Epc at a scan rate
of 0.1 V s-1
.
ClFeIV(tpfc), and ClFeIV(tdcc) are summarized in Table 1.
Ph3PCoIII(tpfc) undergoes two reversible one-electron oxidations
at E1/2 ) 0.72 and 1.11 V in MeCN under an Ar atmosphere.
The first reduction is irreversible and occurs at Epc ) -0.42 V
at a scan rate of 0.1 V s-1 and this reduction is coupled to
reoxidation peaks at Epa ) -0.3 and 0 V (Figure 1). This
reduction is assigned to CoIIIfII and the irreversibility is due to
a partial loss of the Ph3P ligand from [Ph3PCoII(tpfc)]-. The
second reduction (most likely CoIIfI) is reversible and takes
place at -1.44 V. The third reduction is irreversible and
occurs at -2.3 V at a scan rate of 0.1 V (Figure 1, dashed
line). The third reduction may be a multiple-electron reduction
of the fluorinated phenyl rings. Kadish et al.9a reported that
Ph3PCo(omc) (where omc ) 2,3,7,8,12,13,17,18-octamethyl-
corrole) has oxidation potentials of 0.19, 0.76, and 1.54 V, and
reduction potentials of -0.86 (Epc) and -1.92 V in PrCN. The
potentials of Ph3PCoIII(tpfc) are all shifted by about 0.35 V to
0.50 V toward the positive direction compared to Ph3PCo(omc)
due to electron withdrawing effects of the fluorinated phenyl
rings. A similar difference is found when comparing the
reduction potentials of a series of other metal complexes of tpfc
and omc,14 as well as the Co complexes of tetrakis(perfluo-
rophenyl)porphyrin7 and octaethylporphyrin.15
In the presence of CO2, the wave for the CoIIIfII reduction is
practically the same as those in the presence of Ar at any scan
rate, while the wave for the CoIIfI shows slightly increased
current for reduction but no current for reoxidation at scan rates
of 20 to 100 mV s-1 (Figure 1, dotted line). The changes in the
Results and Discussion
Electrochemical Reduction of the Corroles. The results
of cyclic voltammetry measurements with Ph3PCoIII(tpfc),

4774 J. Phys. Chem. A, Vol. 106, No. 18, 2002
Grodkowski et al.
Figure 2. Cyclic voltammograms of ClFeIV(tdcc) in MeCN solutions
saturated with Ar (solid lines) and with CO2 (dotted line). Scan rate
0.1 V s-1
.
CoIIfI peaks are more pronounced at lower scan rates and
disappear upon increasing the scan rate to 5 V s-1, indicating
a slow reaction of [CoI(tpfc)]2- with CO2 (<50 L mol-1 s-1).
At more negative potentials, high catalytic currents are observed
in the presence of CO2.
Figure 3. Absorption spectra recorded upon reduction of Ph3PCoIII-
(tpfc) by sodium amalgam in MeCN solutions. (a) Reduction of CoIII
to CoII. (b) Reduction of CoII to CoI. In each case, the initial spectrum
is the solid line, the intermediate spectra are the dotted lines, and the
final spectrum is the dashed line.
ClFeIV(tpfc) and ClFeIV(tdcc) (Figure 2) show similar elec-
trochemical behavior. ClFeIV(tpfc) is reversibly reduced at E1/2
) 0.44 V. Voltammograms at the narrow range of +1.0 V to
-0.2 V show reversibility at a scan rates of 0.1 V s-1, indicating
that the Cl- remains bound to the FeIII product on the time scale
of these experiments. The second reduction step is clearly
irreversible at a scan rate of 0.1 V s-1, indicating a loss of the
Cl- ligand during the reduction. The reduction of [ClFeIII(tpfc)]-
takes place at Epc ) -1.01 V and is coupled to a reoxidation
of Ph3PCoIII(tpfc) and ClFeIV(tpfc) by sodium amalgam in
MeCN solutions are shown in Figures 3 and 4, respectively.
The optical absorption spectrum of [(CO)FeII(tpfc)]- is given
in Figure 5. The UV-vis spectral data for all three corroles in
their various oxidation states are summarized in Table 2.
peak for FeII(tpfc) at Epa ) -0.43 V. The third reduction (FeIIfI
)
The UV-vis spectra obtained in this study for Ph3PCoIII-
(tpfc) and its first reduction product, [CoII(tpfc)]- (Figure 3a),
are somewhat similar to those reported for the Ph3PCoIII(omc)
and its reduced species.9a Figure 3b shows the second reduction.
Since the spectrum of the product is only slightly shifted from
that of [CoII(tpfc)]- and exhibits no absorption in the near-IR,
and by comparison with the case of similar cobalt porphyrins,
we ascribe this product to [CoI(tpfc)]2- rather than to the CoII
radical anion. Although this assignment is not solidly proven
by the present results, we adopt this notation for the present
article and will refer to both CoI and FeI corroles.
takes place reversibly at -1.60 V, and the fourth reduction
occurs irreversibly at -2.2 V (Epc) at a scan rate of 0.1 V. In
the presence of CO2, the wave for the FeIIf I reduction shows
increased current for reduction but no current for reoxidation
at scan rates of 20 mV s-1 to 1 V s-1. The wave becomes
practically the same as that under an Ar atmosphere at a scan
rate of 10 V s-1. The reaction rate of [FeI(tpfc)]2- with CO2
seems slightly faster than that of [CoI(tpfc)]2-. However, plots
of the catalytic current vs [CO2]1/2 with Fe(tpfc) at scan rates
of 10 mV s-1 and 20 mV s-1 do not show linear relationship
over a wide range but rather indicate saturation at high [CO2].
ClFeIV(tdcc) exhibits reduction potentials more positive (≈0.2
V) than ClFeIV(tpfc). ClFeIV(tdcc) is reversibly reduced at E1/2
) 0.24 V and irreversibly reduced at Epc ) -1.18 V at a scan
rate of 0.1 V s-1. This reduction is coupled to a reoxidation
peak at Epa ) -0.55 V. A small reduction step near -0.6 V is
probably due to a small amount of FeIII(tdcc) produced by partial
loss of chloride following the first reduction. The third reduction
(FeIIfI) takes place reversibly at -1.78 V (Figure 2, solid line).
In the presence of CO2, the wave for the FeIIfI reduction shows
increased current for reduction but no current for reoxidation
at scan rates of 20 mV s-1 to 10 V s-1 (Figure 2, dotted line).
At all scan rates, the ratio between the current under CO2 and
the current under Ar, I(CO2)/I(Ar), was the highest for this
complex. For example, at a scan rate of 100 mV s-1 the ratio
was 2.67 for Fe(tdcc), 1.67 for Fe(tpfc), and 1.05 for Co(tpfc).
Thus, Fe(tdcc) is the best CO2-reduction electrocatalyst among
these corroles, probably due to its more negative reduction
potential.
On comparing the spectra of the starting materials, it is
noticed that while Ph3PCoIII(tpfc) exhibits two peaks in the Soret
region with equal intensities (Figure 3a), Ph3PCoIII(omc) exhibits
mainly one Soret peak and a shoulder.9a Since the second peak
in both cases is at the same wavelength as the intense peak of
the CoII complex, we initially suspected that the second peak
of Ph3PCoIII(tpfc) may be due to the presence of some CoII
complex in the starting materials.16 However, when we at-
tempted to oxidize an acetonitrile solution of this material at
0.6 V vs. SCE, we did not observe any significant current nor
spectral change. Furthermore, Ph3PCoIII(tpfc) has been shown
to be pure by TLC and was found not to contain any
paramagnetic impurities.17 The possibility that the two peaks
are due to two forms of the complex, resulting from partial
dissociation of the Ph3P ligand in solution, is ruled out since
addition of a 1000-fold excess of Ph3P did not change the
spectrum.
The results for reduction of ClFeIV(tpfc) in three distinct steps
to [FeI(tpfc)]2- are shown in Figure 4 and indicate clean and
Reduction of the Corroles by Sodium Amalgam. The
optical absorption spectra obtained during stepwise reduction

Reduction of Cobalt and Iron Corroles and CO2
J. Phys. Chem. A, Vol. 106, No. 18, 2002 4775
TABLE 2: Spectroscopic Properties of the Corroles in Acetonitrile Solutionsa
corrole
λ )
max, nm (10-3 ꢀ, L mol-1 cm-1
Ph3PCoIII(tpfc)
[CoII(tpfc)]-
[CoI(tpfc)]2-
ClFeIV(tpfc)
FeIII(tpfc)
258 (24.8), 301 (17.5), 374 (50.0), 406 (49.9), 520 (7.4), 546 (8.1), 583 (8.0), 806 (2.4)
264 (37.1), 322 (20.5), 412(>95), 498 (7.8), 522 (12.2), 630 (5.5)
260 (40.0), 332 (34.2), 408 (52.8), 574 (22.6)
370 (44.3), 394 (46.8), 606 (3.5)
326 (32.1), 406 (45.7), 532 (13.4), 706 (4.8)
[FeII(tpfc)]-
[COFeII(tpfc)]-
[FeI(tpfc)]2-
ClFeIV(tdcc)
FeIII(tdcc)
304 (20.5), 410 (56.3), 426 (55.0), 476 (23.5), 526 (8.5), 594 (18.8)
403 (95.5), ∼530 (11.9), 553 (12.4)
423 (37.8), 472 (27.2), 622 (13.3), 782 (3.6)
360 (38.2), 394 (41.3), 500sh (8.1), 550sh (5.3), 600sh (3.7), 744 (1.3)
326 (29.7), 410 (44.8), 540 (13.6), 730 (5.3)
305 (21.5), 415 (54.1), 432 (51.1), 482 (22.7), 530 (10.5), 564sh (12.3), 604 (22.5)
406 (>100), 526sh (12.2), 560 (15.3)
434 (36.9), 478 (24.6), 562 (11.5), 606 (15.0), 638sh (12.1), 776 (5.6)
[FeII(tdcc)]-
[COFeII(tdcc)]-
[FeI(tdcc)]2-
a Molar absorption coefficients ꢀ were calculated by assuming 100% conversion from starting materials. Their estimated standard uncertainties
are (10%.
Figure 5. Absorption spectrum of [COFeII(tpfc)]- in MeCN solutions
containing 5% TEA (solid line) compared to the absorption spectrum
of [FeII(tpfc)]- (dotted line).
fraction of [(CO)FeII(tpfc)]-. We interpret these findings by the
following reactions:
[MI(tpfc)]2- + CO2 f [(CO2)M(tpfc)]2-
(1)
(2)
2-
[(CO2)M(tpfc)]2- + CO2 f MIII(tpfc) + CO + CO3
MIII(tpfc) + [MI(tpfc)]2- f 2 [MII(tpfc)]-
[FeII(tpfc)]- + CO / [(CO)FeII(tpfc)]-
(3)
(4)
The [MI(tpfc)]2- complexes react with CO2 to form an adduct,
which decomposes to CO and MIII(tpfc). This decomposition
may be enhanced by CO2, since no protons are available to
accept the oxide. The subsequent reaction 3 must be fast to effect
practically complete conversion of the corrole into [MII(tpfc)]-.
This product does not react with CO2 but, in the case of the
iron complex, it can bind with CO to form an equilibrium
mixture. A higher fraction of [(CO)FeII(tpfc)]- was obtained
in the photochemical experiments and is discussed below, but
the spectra for the cobalt complexes show no evidence for
binding of CO to [CoII(tpfc)]-.
Photochemical Reduction of the Corroles. As in the earlier
experiments with cobalt and iron porphyrins7,18 and phthalo-
cyanines,8 the corroles were photoreduced in deoxygenated
MeCN solutions containing triethylamine (TEA) (5 vol %, 0.36
mol L-1) and p-terphenyl (TP) (0.03 to 3 mmol L-1). The reason
for the use of TP is that direct photochemical reduction of the
MII complexes by TEA is not efficient since there is no
significant binding of TEA to the metal center. The use of
Figure 4. Absorption spectra recorded upon reduction of ClFeIV(tpfc)
by sodium amalgam in MeCN solutions. (a) FeIV f FeIII, (b) FeIII
f
FeII, (c) FeII f FeI. In each case the initial spectrum is the solid line,
the intermediate spectrum is the dotted line, and the final spectrum is
the dashed line.
quantitative reduction in each step. Only the FeII corrole forms
a strong complex with CO and its spectrum is shown in Figure
5. The FeII corrole for this experiment was produced by
reduction with sodium amalgam or by photochemical reduction
as described below.
Addition of CO2 to the MeCN solutions of [MI(tpfc)]2-
resulted in a color change indicating formation of MII complexes.
The spectrum of the product obtained from [FeI(tpfc)]2- was a
combination of the spectra of [FeII(tpfc)]- and a very small

4776 J. Phys. Chem. A, Vol. 106, No. 18, 2002
Grodkowski et al.
p-terphenyl as a sensitizer is based on the highly efficient
formation of the p-terphenyl radical anion by reaction of the
singlet excited state of TP with TEA,19
with 408 and 574 nm for the CoI-corrole). The peak positions
shifted gradually upon photolysis and it is possible that CoI-
corrole is one of the intermediates in this process. On the other
hand, the main peaks for the Fe(tpfc) product were at 407, 517,
556, 591, and 634 nm, i.e., quite different from those of the
[FeI(tpfc)]2- listed in Table 2, indicating that the FeI complex
is unstable under these conditions. At this stage, addition of O2
led to substantial recovery of the FeIII(tpfc). However, upon
further photolysis (without O2) the absorbance at all peaks
decreased and then the solution became almost colorless; at this
stage, no recovery of FeIII(tpfc) was observed following addition
of O2, indicating degradation of the macrocycle.
TP
9
hν8 TP*
(5)
(6)
TP* + Et3N f TP•- + Et3N•+
and the rapid reduction of the metal complex by TP•-, which
has a highly negative reduction potential (-2.45 V vs SCE in
dimethylamine).19
TP•- + [MIII(tpfc)] f TP + [MII(tpfc)]-
TP•- + [MII(tpfc)]- f TP + [MI(tpfc)]2-
(7)
(8)
Photochemical reduction in CO-saturated solutions led to
different spectral changes. The first reduction product, [FeII(tpfc)]-,
strongly binds CO and the spectrum of the complex
[(CO)FeII(tpfc)]-, exhibits a very intense Soret peak at 403 nm
and two small peaks at ≈530 and 553 nm (Figure 5). Extensive
photolysis also led to bleaching. Similar results were obtained
with Fe(tdcc) (but no spectral evidence was found for binding
of CO to [CoII(tpfc)]-).
The TEA radical cation reacts with a TEA molecule to form a
carbon-centered radical (E1/2 ) -1.12 V vs SCE in MeCN)20
that can reduce [MIII(tpfc)] but not [MII(tpfc)]-.
Et3N•+ + Et3N f Et3NH+ + Et2NC˙ HCH3
Et2NC˙ HCH3 + [MIII(tpfc)] f
(9)
Photochemical reduction in CO2-saturated solutions resulted
in similar spectral changes to those found under Ar in the early
stages of photolysis. When the photolysis was continued until
nearly half the [FeII(tpfc)]- was reduced, a weak absorbance at
403 nm was observed, indicating partial formation of
[(CO)FeII(tpfc)]-. Continued photolysis produced large amounts
of CO (see below).
Et2N+)CHCH3 + [MII(tpfc)]- (10)
A pronounced difference between the photochemical reduction
and the reduction by sodium and by electrochemistry lies in
the fact that the photochemical process leads to production of
acid (e.g., via reaction 9). As a result, reduced species such as
[MI(tpfc)]2- may undergo protonation. They may form a hydride
and subsequently H2 and may undergo hydrogenation on the
macrocycle (as observed with porphyrins).6-8,18
Photolysis of He-saturated acetonitrile solutions containing
TEA (5 vol %), TP (3 mmol L-1), and either the cobalt or the
iron corrole led to production of H2, after a slight delay for the
reduction of the starting material to the active species (Figures
6a and 7a). The maximal rate of H2 production was similar with
both Fe(tpfc) and Co(tpfc), ≈4 × 10-7 mol L-1 s-1. After
several hours the production rate decreased and at longer times
production of H2 stopped. Photolysis of similar solutions
saturated with CO2 led to production of both H2 and CO (Figures
6a and 7a). Both of these gases were produced at similar rates
in the early stages of photolysis, after a slight delay, and the
sum of [CO] + [H2] was slightly less than the [H2] produced
under He. After several hours the rates of production of the
two gases diverged; the ratio [CO]/[H2] was >1 for Co(tpfc)
and <1 for Fe(tpfc), and at longer times production of the gases
stopped. The results obtained with Fe(tdcc) (Figure 7b) are
significantly different from those of Fe(tpfc) and Co(tpfc) in
that the yield of CO in the early stages of photolysis was much
higher than that of H2 and higher than that found with the other
corroles. This result is in line with the electrochemical experi-
ments showing that Fe(tdcc) was a more efficient electrocatalyst
for reduction of CO2 than the other two corroles.
With 4.2 × 10-5 mol L-1 Co(tpfc) under CO2, the total
concentrations of both CO and H2 reached a plateau after ≈10
h of photolysis. The plateau level of CO (4.2 mmol L-1) was
higher than that of H2 (2.5 mmol L-1). The results with similar
concentrations of Fe(tpfc) and Fe(tdcc) under CO2 were
somewhat different; the yield of CO leveled off (at ≈2.2 mmol
L-1) but the yield of H2 continued to rise (to >8 mmol L-1
with Fe(tpfc)). In contrast with the above results, production of
H2 in similar solutions under He reached a higher plateau level
with Co(tpfc) (≈14 mmol L-1) than with Fe(tpfc) (≈6 mmol
L-1).
[MI(tpfc)]2- + Et3NH+ f [HM(tpfc)]- + Et3N (11)
[HM(tpfc)]- + Et3NH+ f MIII(tpfc) + H2 + Et3N (12)
Before photolysis, the absorption spectra of the corroles were
first recorded in MeCN solutions. Addition of TEA resulted in
a major change in the spectra. From comparison of the spectra
recorded under these conditions to those recorded following
reduction with sodium amalgam we conclude that TEA not only
replaces the axial ligand but also reduces the metal center even
under air. In the case of Ph3PCoIII(tpfc), the spectrum of the
starting material in MeCN exhibits two peaks at 374 and 406
nm of nearly equal intensities and after addition of TEA and
equilibration for several minutes the spectrum has only one Soret
peak at 412 nm corresponding to that of [CoII(tpfc)]-. (This
solution was unstable under air and was slowly bleached, in
contrast with the hexa-coordinated CoIII corroles with aromatic
amines, which are air-stable and display Soret peaks of about
440 nm17). In the case of ClFeIV(tpfc) and ClFeIV(tdcc), addition
of TEA led to immediate reduction of the compounds to the
FeIII corroles. These solutions were then photolyzed (after
addition of TP and bubbling with Ar) and their spectra were
recorded before photolysis and after different intervals of
photolysis. Photoreduction of FeIII(tpfc) and FeIII(tdcc) led to
formation of the FeII corroles. The spectra of the FeII and CoII
corroles were identical to those determined by reduction with
sodium amalgam (Table 2).
Photoreduction of [FeII(tpfc)]-, [FeII(tdcc)]-, and [CoII(tpfc)]-
produced species with absorption spectra that are different from
the spectra of the MI corroles listed in Table 2. The main peaks
for the Co(tpfc) product were at 408 and 587 nm (as compared
As discussed above, H2 is formed via reactions 11 and 12,
and CO is formed via reactions 1 and 2. The finding with
Co(tpfc) and Fe(tpfc) that the rates of production of the two
gases are similar suggests that reaction with protons (reaction

Reduction of Cobalt and Iron Corroles and CO2
J. Phys. Chem. A, Vol. 106, No. 18, 2002 4777
Figure 6. Photochemical production of CO and H2 in MeCN solutions
of cobalt corrole (a) and cobalt porphyrin (b). Production of H2 (4) in
He-saturated solution and production of CO (b) and H2 (o) in CO2-
saturated solution. The solutions contained 5% TEA, 3 mmol L-1 TP,
and (a) 4.2 × 10-5 mol L-1 CoIII(tpfc), (b) 5.2 × 10-5 mol L-1 CoII-
(TTP). The standard uncertainties are (10%.
11) efficiently competes with reaction with CO2 (reaction 1)
despite the large excess of CO2 over protons in the solution.
With Fe(tdcc), however, the reaction with CO2 predominates
in the early stages of photolysis. In parallel with these reactions,
the macrocycle undergoes hydrogenation. The partially hydro-
genated complexes also catalyze the photochemical production
of CO and H2 but exhibit different reactivities. After very long
irradiations, production of all gases stops. This may be due to
decomposition of the metal complex into an unreactive product.
An additional factor is the loss of TP as a sensitizer. The TP•-
radical anion transfers an electron to the metal complex to
recover TP (e.g., reactions 8, 9). However, TP•- also undergoes
protonation to form TPH• and this leads to loss of the sensitizer.
We found that a considerable fraction of the TP was lost during
these experiments and this loss was related to the termination
of the photochemical production of the two gases. The loss of
TP was estimated from the UV absorbance of the solution before
and after photolysis. In those cases where the final yields were
lower, the remaining TP concentration was quite low, indicating
that the lack of TP limited the yield. In those cases where the
final yields were higher, the remaining TP concentration was
also higher, indicating that other factors limited the yield. Thus
the end of the photochemical process is determined by a
combination of several effects, as discussed before for the
metalloporphyrins.18
Figure 7. Photochemical production of CO and H2 in MeCN solutions
of iron corroles (a,b) and iron porphyrin (c). Production of H2 (4) in
He-saturated solution and production of CO (b) and H2 (o) in CO2-
saturated solution. The solutions contained 5% TEA, 3 mmol L-1 TP,
and (a) 4.8 × 10-5 mol L-1 ClFeIV(tpfc), (b) 4.5 × 10-5 mol L-1 ClFeIV-
(tdcc), (c) 5.0 × 10-5 mol L-1 FeII(TTP). The standard uncertainties
are (10%.
very different from those of the Co complexes but the results
with Fe(tdcc) (Figure 7b) are close to those obtained with
Fe(TTP) (Figure 7c); both of the latter two complexes are more
effective in CO production in the early stages of photolysis.
The total concentrations of CO and H2 produced in these
experiments, with both sets of metal complexes, were ≈300
times larger than the concentrations of the metal complexes.
The total concentrations of CO produced were 50 to 100 times
larger than the concentrations of the complexes. The findings
that the results for the corroles are generally similar to those
obtained with the corresponding porphyrin indicate that despite
the fact that the corroles react with CO2 in their MI oxidation
state while the porphyrins have to be reduced beyond the MI
state to react with CO2, in both cases the reduced species are
hydrogenated at the macrocycle and are deactivated to similar
extents.
For comparison of corroles to porphyrins, Figures 6b and 7c
show the results obtained by photolysis of Fe(TTP) and Co(TTP)
(TTP ) tetrakis(3-methylphenyl)porphyrin)21 under identical
conditions (MeCN/TEA/TP, solute concentrations, cell size,
photolysis source) to those used with the Fe and Co corroles
(Figure 6a and 7a). For the Co corrole (Figure 6a) and porphyrin
(Figure 6b), the yield of H2 under He and the yields of H2 and
CO under CO2 are each similar for the two complexes, within
a factor of <2. The results with Fe(tpfc) (Figure 7a) are not
Acknowledgment. This research was supported in part by
the Division of Chemical Sciences, Office of Basic Energy
Sciences, U.S. Department of Energy; work at NIST under
Contracts DE-AI02-95ER14565; and work at BNL under
Contract DE-AC02-98CH10886. Work at the Technion was
supported by the Israel Science foundation under Grant 368/00
and by the Petroleum Research Fund.

4778 J. Phys. Chem. A, Vol. 106, No. 18, 2002
Grodkowski et al.
(9) (a) Kadish, K. M.; Koh, W.; Tagliatesta, P.; Sazou, D.; Paolesse,
R.; Licoccia, S.; Boschi, T. Inorg. Chem. 1992, 31, 2305. (b) Van
Caemelbecke, E.; Will, S.; Autret, M.; Adamian, V. A.; Lex, J.; Gisselbrecht,
J.-P.; Gross, M.; Vogel, E.; Kadish, K. M. Inorg. Chem. 1996, 35, 184. (c)
Kadish, K. M.; Adamian, V. A.; Van Caemelbecke, E.; Gueletii, E.; Will,
S.; Erben, C.; Vogel, E. J. Am. Chem. Soc. 1998, 120, 11986.
(10) Gross, Z.; Simkhovich, L.; Galili, N. J. Chem. Soc., Chem. Commun.
1999, 599. Meier-Callahan, A. E.; Gray, H. B.; Gross, Z. Inorg. Chem.
2000, 39, 3605.
(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.
References and Notes
(1) Hawecker, J.; Lehn, J.-M.; Ziessel, R. J. Chem. Soc., Chem.
Commun. 1985, 56. Grant, J. L.; Goswami, K.; Spreer, L. O.; Otvos, J. W.;
Calvin, M. J. Chem. Soc., Dalton Trans. 1987, 2105. Craig, C. A.; Spreer,
L. O.; Otvos, J. W.; Calvin, M. J. Phys. Chem. 1990, 94, 7957. Kelly, C.
A.; Mulazzani, Q. G.; Venturi, M.; Blinn, E. L.; Rodgers, M. A. J. J. Am.
Chem. Soc. 1995, 117, 4911.
(2) (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.
(3) Fisher, B.; Eisenberg, R. J. Am. Chem. Soc. 1980, 102, 7361.
Sullivan, B. P.; Bolinger, C. M.; Conrad, D.; Vining, W. J.; Meyer, T. J. J.
Chem. Soc., Chem. Commun. 1985, 1414. 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) For reviews see: Catalytic ActiVation of Carbon Dioxide; Ayers,
W. M., Ed.; ACS Symposium Series, Vol. 363, 1988. Behr, A. Carbon
Dioxide ActiVation by Metal Complexes; VCH: Weinheim, 1988. Electro-
chemical and Electrocatalytic Reactions of Carbon Dioxide; Sullivan, B.
P., Ed.; Elsevier: Amsterdam, 1993. Sutin, N.; Creutz, C.; Fujita, E.
Comments Inorg. Chem. 1997, 19, 67.
(5) 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.
(6) Grodkowski, J.; Behar, D.; Neta, P.; Hambright, P. J. Phys. Chem.
A 1997, 101, 248.
(7) Behar, D.; Dhansekaran, T.; Neta, P.; Hosten, C. M.; Ejeh, D.;
Hambright, P.; Fujita, E. J. Phys. Chem. 1998, 102, 2870.
(8) Grodkowski, J.; Dhanasekaran, T.; Neta, P.; Hambright, P.;
Brunschwig, B. S.; Shinozaki, K.; Fujita, E. J. Phys. Chem. A 2000, 104,
11332.
(12) Simkhovich, L.; Galili, N.; Saltsman, I.; Goldberg, I.; Gross, Z.
Inorg. Chem. 2000, 39, 2704.
(13) Riddick, J. A.; Bunger, W. B.; Sakano, T. K. Organic SolVents,
Physical Properties and Methods of Purification, 4th ed.; Wiley: New York,
1986.
(14) Simkhovich, L.; Mahammed, A.; Goldberg, I.; Gross, Z. Chem.
Eur. J. 2001, 7, 1041.
(15) Fuhrhop, J.-H.; Kadish, K. M.; Davis, D. G. J. Am. Chem. Soc.
1973, 95, 5140.
(16) It is a known phenomenon that many CoIII porphyrins in the solid
state contain varying amounts of CoII porphyrins, depending on the method
of preparation and drying.
(17) Mahammed, A.; Giladi, I.; Goldberg, I.; Gross, Z. Chem. Eur. J.
2001, 7, 4259.
(18) Dhanasekaran, T.; Grodkowski, J.; Neta, P.; Hambright, P.; Fujita,
E. J. Phys. Chem. A 1999, 103, 7742.
(19) Matsuoka, S.; Kohzuki, T.; Pac, C.; Ishida, A.; Takamuku, S.;
Kusaba, M.; Nakashima, N.; Yanagida, S. J. Phys. Chem. 1992, 96, 4437.
(20) Wayner, D. D. M.; McPhee, D. J.; Griller, D. J. Am. Chem. Soc.
1988, 110, 132. Armstrong, D. A.; Rauk, A.; Yu, D. J. Am. Chem. Soc.
1993, 115, 666.
(21) Further details on metalloporphyrins were presented in refs 7 and
18.