Oxygen reduction reaction catalyzed by overhanging carboxylic acid strapped iron porphyrins adsorbed on carbon nanotubes

Article abstract of DOI:10.1142/S1088424619501232

A series of hybrid catalysts for the oxygen reduction reaction (ORR) has been investigated. They are composed of multi-wall carbon nanotubes (MWNTs) coated with iron strapped porphyrins. Two porphyrins have been probed; both are strapped with the same skeleton and differ only by the number of overhung carboxylic acid(s), either one or two. In this structure, the carboxylic acid group can act as a proton relay between the medium and the catalyst or as a polar group surrounding the dioxygen binding cavity. While the number of carboxylic acid group(s) does not exhibit a significant influence on the catalytic properties, the combination of both components-MWNTs and porphyrin-leads to a better catalytic activity than those of the nanotubes or the porphyrins taken separately.

Full text of DOI:10.1142/S1088424619501232

Journal of Porphyrins and Phthalocyanines
J. Porphyrins Phthalocyanines 2019; 23: 1–10
DOI: 10.1142/S1088424619501232
Published at http://www.worldscinet.com/jpp/
Oxygen reduction reaction catalyzed by overhanging
carboxylic acid strapped iron porphyrins adsorbed
on carbon nanotubes
Bernard Boitrel*a◊, Morgane Bougeta, Pradip K. Dasa, Stéphane Le Gaca,
Thierry Roisnela, Manel Hananab, Hélène Arcostanzob, Renaud Cornutb,
Bruno Jousselmeb and Stéphane Campidelli*b
aUniv Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes), UMR 6226, Rennes, F-35000, France
bLICSEN, NIMBE, CEA, CNRS, Université Paris-Saclay, CEA Saclay 91191 Gif-sur-Yvette Cedex, France
Dedicated to Professor Roberto Paolesse on the occasion of his 60th birthday.
Received 25 June 2019
Accepted 6 August 2019
ABSTRACT:A series of hybrid catalysts for the oxygen reduction reaction (ORR) has been investigated.
They are composed of multi-wall carbon nanotubes (MWNTs) coated with iron strapped porphyrins.
Two porphyrins have been probed; both are strapped with the same skeleton and differ only by the
number of overhung carboxylic acid(s), either one or two. In this structure, the carboxylic acid group can
act as a proton relay between the medium and the catalyst or as a polar group surrounding the dioxygen
binding cavity. While the number of carboxylic acid group(s) does not exhibit a significant influence on
the catalytic properties, the combination of both components–MWNTs and porphyrin–leads to a better
catalytic activity than those of the nanotubes or the porphyrins taken separately.
KEYWORDS: porphyrins, dioxygen reduction, synthesis, X-ray diffraction, catalysis.
it has been shown that “iron-only” tris(2-aminoethyl)
INTRODUCTION
amine(tren)-cappedporphyrins,thatiswithoutanycopper
Oxygen Reduction Reaction (ORR) is the paramount
in the cap, could be efficient catalysts of the 4-electron
reaction taking place at the cathode of a fuel cell. In
reduction of dioxygen as long as electron supply is not a
aerobic life on Earth, Cytochrome c Oxidase, the oxygen
limiting factor as at the surface of an electrode [6]. This
electrode of Nature’s fuel cell is a famous example of
surprising result was then shown true for strapped or
protein performing the reduction of oxygen through the
quinolinoyl picket porphyrins [7–9]. Finally, by grafting
four-electronfour-protonmechanism[1].Thus, aplethora
a carboxylic acid group around the binding site of the
of catalysts based on iron “porphyrinoid” derivatives,
macrocycle, the hangman catalyst family was also found
mimicking the structure of the active center of the protein
to be effective monometallic catalysts for the ORR
havebeendevelopedandextensivelyinvestigatedforORR
[10, 11]. This effectiveness was rationalized by the
[2–4]. However, where in nature a bimetallic center is
possible role of the carboxylic acid group as a proton-
required particularly under rate-limiting electron flux [5],
relay. The same concept has been extended to the study
of meso-tetra-arylporphyrins with four carboxylic acid
groups either in ortho or para position of the meso
aromatic cycles. In the case of the ortho substitution, the
selectivity was high for the four-electron process with a
production of hydrogen peroxide quasi inexistent [12].
One of us, in previous work [13], has reported that
the ORR activity of several nonfunctionalized Co and
◊ꢀSPP full member in good standing.
*Correspondence to: Univ Rennes, CNRS, ISCR (Institut des
Sciences Chimiques de Rennes), UMR 6226, Rennes, F-35000,
France, tel.: (33)2 2323 5856, fax: (33)2 2323 5637; email:
Bernard.Boitrel@univ-rennes1.fr, (Bernard Boitrel); stephane.
campidelli@cea.fr (Stéphane Campidelli).
Copyright © 2019 World Scientific Publishing Company

2
B. BOITREL ET AL.
Fe phthalocyanines and porphyrins immobilized by
π-stacking interactions on various types of carbon
nanotubes such as SWNTs, DWNTs, oxidized and
non-oxidized MWNTs, was improved by the presence
of the carbon nanotubes. However, as discussed by
Rigsby et al. the nature and structure of the adsorbed
catalyst are crucial parameters as well [14]. Thus, the
formulation itself of a catalyst ink containing the same
iron porphyrin, that is either deposited directly on glassy
carbon (GC) or mixed with Vulcan and then deposited
on GC electrode, has an effect at least as important as
changing the structure of the molecular catalyst itself.
In the same line, and more recently, we have reported
on the synergistic effect on ORR of strapped porphyrins
polymerized around carbon nanotubes [15]. It was shown
that the presence of a potential proton relay in the hybrids
materials in comparison with those lacking such a group
did not generate any significant improvement. However,
it is worth to mention that the spatial arrangement of
the molecular catalyst, namely the iron porphyrin,
properly functionalized with peripheral groups for the
polymerization was expected to be somehow controlled
by the covalent linkages between the porphyrin units. In
the same line, it has been evidenced that a concomitant
tuning of the ligand modification with the method of
immobilization of the catalyst as well as the supporting
material are three influential parameters [16–18].
to measure the ORR properties of strapped porphyrins
bearing either one or two polar overhanging carboxylic
acid(s) and second to evaluate the influence of the
communication between the nanotube and the catalytic
centers as well as the effects of the non-aggregation of
the porphyrins adsorbed on the nanotube as depicted in
Fig. 1.
RESULTS AND DISCUSSION
Strapped-porphyrin derivatives 1Fe and 2Fe bearing
one and two carboxylic acid functions respectively were
prepared according to a synthetic pathway previously
reported in the case of a single strap porphyrin bearing
3,5-dihydroxyphenyl in the 10,20 meso positions [19].
Chiefly, the synthesis of 2, summarized in Scheme 1,
began by condensation of 5-(4-methoxyaryl)dipyrryl-
methane 3 with 2-nitrobenzaldehyde in acidic conditions
to obtain porphyrin 4 whose nitro groups were reduced by
tin chloride in acidic medium leading to a mixture of the
two atropisomers of the resulting aminophenyl porphyrin.
Then, the atropisomer aa 5 was separated by silica gel
flash chromatography and acylated with 3-chloromethyl
benzoyl chloride in the presence of triethylamine to
obtain porphyrin 6. Condensation of diethyl malonate
in basic conditions allowed the formation of the strap in
the 5,15 meso positions of the resulting porphyrin 7 and
finally treatment by BBr3 of the latter supplied free base
porphyrin 2. Actually, both porphyrins 1Fe and 2Fe were
obtained by refluxing free-base porphyrin 2 in THF in the
presence of iron(II) bromide and 2,6-lutidine in a glove
box. During this process, the decarboxylation reaction
leading partially to 1Fe was observed and the two
complexes were obtained in roughly equal proportions as
indicated by TLC analysis and were separated by silica
gel chromatography out of the glovebox after oxidation
Thus, in the present work, we now report on the ORR
activity of MWNTs functionalized only by physisorption
with specific iron (III) strapped porphyrins in a pH
range from 13 to 6. These porphyrins contain a bridge
bearing one or two carboxylic functions between the
phenyl groups in 5 and 15 meso positions (Scheme 1).
The overhanging bridge should prevent the aggregation
of the porphyrins compared to the previous studies while
preserving one side available for the interaction with the
nanotubes by π-stacking. The goal of this study is first
Fig. 1. Schematic representation of the nanotube/porphyrins hybrids
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 2–10

OXYGEN REDUCTION REACTION CATALYZED BY OVERHANGING CARBOXYLIC ACID
3
O2N
MeO
N
H
O2N
HN
H
N
N
N
i
H
N
OMe
O
OMe
3
4
Cl
ii, iii
Cl
HN
N
H2N
N
MeO
MeO
iv
N
H
N
H2N
O
H
NH
N
N
H
N
H
N
OMe
OMe
6
5
v
EtO2C
HO2C
CO2Et
CO2H
O
O
HN
HN
N
MeO
HO
vi
N
N
O
O
H
H
H
NH
NH
N
N
N
H
N
N
OMe
OH
2
7
vii
CO2H
O
O
O
O
O
HN
HN
HO
HO
O
O
O
N
N
+
NH
Fe
NH
Fe
N
N
N
N
N
N
OH
1Fe
2Fe
OH
Scheme 1. Synthesis of iron porphyrins 1Fe and 2Fe bearing one and two overhanging carboxylic acid groups respectively, for MWNT
coating. (i) 2-nitrobenzaldehyde, CH2Cl2, BF3-Et2O, 2 h, then DDQ, 20%; (ii) SnCl2, HCl, 80%; (iii) silica gel chromatography,
CH2Cl2, 66%; (iv) 3-(chloromethyl)benzoyl chloride (3 eq), NEt3, THF, 85%; (v) CH2(CO2Et)2 (10 equiv.), THF, EtONa, room
temp., 12 h, 80%; (vi) BBr3, CH2Cl2, room temp., 12 h, 80%; (vii) FeBr2, 2,6-lutidine, THF, reflux overnight, silica gel column
chromatography after air oxidation and HCl (1M) washing, 1Fe (41%), 2Fe (37%)
and identified by MALDI-TOF mass spectrometry (see
experimental part).
right). Although a similar coordination polyhedron has
already been reported for the bis-carboxylato analogous
complex of 1Fe [15], in the present case, the strap is more
distorted with the carbon atom C2 diving toward the
macrocycle plane (3.935 Å from iron against 4.301 Å for
the non-decarboxylated complex) and with a distance of
the iron atom to the mean porphyrin plane of 0.449 Å,
against 0.512 Å for the non-decarboxylated complex.
The very same distorted conformation was also found
in an analogous zinc complex in which the distance of
Among the two final iron porphyrins, we were able
to obtain single crystals of 1Fe and to solve its X-ray
structure (Fig. 2 and Table 1). The latter indicates that
the iron cation is square-pyramidal five-coordinate with
the intramolecular carboxylate bound on it with a Fe-O1
bond lenght of 1.997 Å. This intramolecular binding
implies a significant distortion of the strap which retains
a “W-shape” as observed on the apical view (Fig. 2,
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 3–10

4
B. BOITREL ET AL.
C3
C2
C4
N4
O2
O1
C4
N3
C1
C1
C2
C3
O2
O1
N2
Fe
N4
N3
Fe
N2
N1
N1
Fig. 2. X-ray structure of 1Fe. Top: ORTEP lateral view (30% thermal ellipsoids) and bottom: apical rods view with the carbon
atoms of the strap colored in black. Selected distances (Å): N1-Fe 2.066, N2-Fe 2.049, N3-Fe 2.061, N4-Fe 2.056, O1-Fe 1.997,
C1-O1 1.232, C1-O2 1.230, (Fe, 24MP) 0.449
the metal to the mean porphyrin plane was only 0.386 Å
[20], clearly indicating that this type of strap remains
quite flexible. The macrocycle is strongly saddle-shaped
with almost no ruffling, as indicated by the angle with
the mean plane of the two pairs of diametrically opposed
pyrrole rings (N1 and N3: -13.02° and -12.68°; N2 and
N4: 11.14° and 14.06).
Carbon (GC) electrode. The curves show that the catalyst
inks made by mixing the nanotubes with the porphyrins
exhibit higher current density and lower overpotential
(of about 0.1 V) than porphyrins alone. This result is
not surprising since similar observations were made by
Rigsby et al. in the case of iron porphyrins deposited on
different carbon supports [14]. In addition, it is interesting
to notice that the voltammetry curves of MWNT-1Fe
and MWNT-2Fe exhibit two waves at around -0.4 and
-0.7 V, explained by the own activity of the nanotubes
(see below). Figure 3c shows the activity of MWNT
and Fig. 3d compares the voltammetry curves (0 and
2000 rpm) of MWNT, 1Fe and MWNT-1Fe. First, we
observe that, at pH 13, MWNT reduced oxygen with
lower overpotential and higher current density than 1Fe.
Second, the reduction starts roughly at the same potential
(-0.2V vs. Ag/AgCl) for MWNT and MWNT-1Fe. Multi-
walled carbon nanotubes are active materials for oxygen
reduction in alkaline media [21, 22] and it is likely that
the nanotubes are more competitive than the nanotube/
porphyrin hybrids at pH 13. It can explain the two
waves observed on the reduction curves of the nanotube/
porphyrin hybrids. We believe that the reduction initiated
on the nanotubes is then performed by the porphyrin at
lower potential. Nevertheless, when the plateau is reached
at -0.8V vs. Ag/AgCl the current density is larger of ca.
2 mA/cm2 for MWNT-1Fe compared to MWNT alone.
Figure S48 shows the same comparison between the
ORR activity of MWNT, MWNT-1Fe, MWNT-2Fe and
those of 1Fe and 2Fe deposited on the GC electrode, but
at pH = 10. Once again, the curves clearly demonstrate
that the catalyst inks made by mixing the nanotubes
with the porphyrins exhibit higher current density than
porphyrins alone. The reduction of oxygen starts with an
overpotential of almost 0.4 V for 1Fe and 2Fe compared
to the same porphyrin deposited on the nanotubes.
In order to prepare the hybrid materials, MWNT
(Nanocyl NC3100TM) were purified by treatment with
nitric acid (35%) at 100°C for 5 h (see experimental
section). After treatment the nanotubes were diluted with
iced-water, filtered through 0.45 mm PTFE membrane and
extensively washed with water, methanol and then dried
under vaccum. The catalysts were prepared by mixing
MWNT with porphyrins 1Fe or 2Fe (in a 1:1 ratio in
weight) in THF under bath sonication. After mixing, THF
was evaporated and the catalyst inks were prepared by
dispersing, using bath sonication, the different MWNT/
porphyrin hybrids (3 mg) in 750 mL of ethanol and 75 mL
of Nafion solution (5% in alcohol). Similarly, the inks
of the reference compounds MWNT, 1Fe and 2Fe were
prepared by mixing the nanotubes or the porphyrins in
750 mL of ethanol and 75 mL of Nafion solution (5% in
alcohol). The catalysts were drop-casted on the Glassy
Carbon disk and tested in a series of Rotating Ring Disk
Electrode experiments at pH 13 (NaOH, 0.1 M), pH 10,
8 and 6 (phosphate buffers).
We first compared the electrocatalytic properties of
the different components: MWNT, iron porphyrins 1Fe
and 2Fe and the 1:1 mixture MWNT-1Fe and MWNT-
2Fe at pH 13 (Fig. 3) and pH 10 (Fig. S48). All the curves
correspond to the average (reduction and reoxidation) of
the cyclic voltammetry curves.
Figures 3a–3b present the comparison between the
ORR activity of MWNT-1Fe, MWNT-2Fe and those
of the iron porphyrins deposited directly on the Glassy
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 4–10

OXYGEN REDUCTION REACTION CATALYZED BY OVERHANGING CARBOXYLIC ACID
5
Table 1. X-ray structural data of 1Fe
a porphyrin that does not exhibit proton relay and
similar results to those obtained for MWNT-1Fe
were obtained [15]. The presence of the proton in
close proximity to the iron center does not seem
to be mandatory to improve the reaction in our
case. Indeed the presence of Nafion in the mixture
probably ensures the availability of protons close to
the reaction center.
It is worth mentioning that at pH 6 and lower, the
ORR activity of the porphyrins decreases rapidly
and almost completely vanishes after the 3 cycles
requiring the deposition of a new catalyst ink for
each rotation cycle. This loss of ORR activity can
be due to demetalation of the porphyrin during the
electrochemical processes [23, 24], this effect being
well documented in the case of iron phthalocyanine
[25–27] but still subject to debate in the case of
porphyrins.
Empirical formula
Formula weight
C62H41FeN6O6
1021.86 g/mol
150 (2) K
Temperature
Wavelength
0.71073 Å
Crystal system, space group
Unit cell dimensions
triclinic P-1
a = 14.499 (2) Å, a = 90.838 (5)°
b = 14.729 (2) Å, b = 109.567 (6)°
c = 17.848 (3) Å, g = 117.362 (5)
Volume
3124.0 (8) Å3
-3
.
Z, Calculated density
Absorption coefficient
F(000)
2, 1.086 g cm
0.291 mm-1
1058
Crystal size
0.320 × 0.260 × 0.110 mm
black
Crystal color
Figure 4 shows the RRDE curves registered at a
rotation rate of 400 rpm for MWNT, MWNT-1Fe
and 1Fe at pH 13 and 10. The numbers of electrons
n involved in the reduction, summarized in Table 2
were determined following the equation n = 4Id/
(Id + Ir/Nc) as recommended by Qiao et al.
[28], from the disk and ring currents and with a
collection coefficient Nc = 0.2 determined using
the one-electron Fe(CN)36-/Fe(CN)46- redox couple.
From the curves, it is observed that the reduction of
O2 is accompanied by the production of hydrogen
peroxide both for MWNT and 1Fe. Conversely,
for MWNT-1Fe, almost no production of H2O2
is detected at the plateau. At pH 13 a bump in the
ring current between -0.20 and -0.70 V reflecting
the production of H2O2 is observed; however, this
phenomenon is attributed to the initial reduction
of oxygen by the nanotubes and appears before the
plateau at disk is reached. However, at both pH,
the evaluated value of n remains close to the ideal
value of 4.
Theta range for data collection
h_min, h_max
3.026 to 27.484°
-18, 18
k_min, k_max
-19, 19
l_min, l_max
-23, 23
Reflections collected/unique
Reflections [I > 2s]
Completeness to theta_max
Absorption correction type
Max. and min. transmission
Refinement method
Data/restraints/parameters
bS (Goodness-of-fit)
Final R indices [I > 2s]
R indices (all data)
Largest diff. peak and hole
aRint = ∑|Fo2 – 〈Fo2〉 |/∑ꢀ[Fo2].
57073/13966 [R(int)a = 0.1159]
9825
0.975
multi-scan
0.968, 0.776
Full-matrix least-squares on F2
13966/0/672
1.036
R1c = 0.0888, wR2d = 0.2340
R1c = 0.1213, wR2d = 0.2579
1.852 and -1.601 e-.Å-3
bS = {∑[w(Fo2 – Fc2)2]/(n – p)}1/2
cR1 = ∑ | |F | – |Fc| |/∑ |F |.
.
o
o
dwR2 = {∑[w(F 2 – Fc2)2]/∑[w(F 2)2]}1/2
.
o
o
CONCLUSION
w = 1/[s(Fo2) + aP2 + bP] where P = [2Fc2 + max(Fo2, 0)]/3.
Herein we formulated a series of catalyst inks
for oxygen reduction reaction containing strapped
porphyrins and MWNTs. The combination of the
nanotubes with the iron porphyrins systematically gives
better catalytic properties than those of two components
taken separately. While carbon nanotubes are known
to be slightly active in oxygen reduction, they produce
significant amounts of hydrogen peroxide. When
porphyrins are simply mixed in Nafion and deposited
on the glassy carbon electrode, a low efficiency is
generally observed. We attribute this behavior to the
lack of electrons available for the reduction. Indeed in
the catalyst ink, the porphyrins are embedded in Nafion
and only the porphyrins close to the glassy carbon
disk can benefit from efficient electron transfers from
MWNT-1Fe and MWNT-2Fe exhibit very similar
properties, it appears that the presence of one or two
carboxylate functions does not influence the properties
under alkaline conditions. At pH 10, MWNT become
less competitive for oxygen reduction than MWNT-1Fe
or MWNT-2Fe and only one reduction wave is observed.
The comparison of the ORR activity (RDE and RRDE)
of MWNT, 1Fe, 2Fe, MWNT-1Fe and MWNT-2Fe
for pH 8 and pH 6 are given in Fig. S46 and Fig. S47,
respectively. The catalytic activities of MWNT-1Fe and
MWNT-2Fe (and of the reference compounds) exhibit
the same trend with respect to those observed at pH 13
and 10. We also performed tests at pH 6 and pH 8 with
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 5–10

6
B. BOITREL ET AL.
Fig. 3. Polarization curves at different rotation rates (0, 400, 800, 1200, 1600 and 2000 rpm) recorded for ORR in O2-saturated 0.1 M
NaOH solution (pH 13) (scan rate = 5 mV/s, room temperature) on GC with predeposited (a) 1Fe (pink) and MWNT-1Fe (blue);
(b) 2Fe (green) and (c) MWNT-2Fe (orange); MWNT (black). (d) Comparison between MWNT, MWNT-1Fe and 1Fe at 0 and
2000 rpm
the electrodes. Conversely, carbon nanotubes create
a percolating pathway for the charges in the catalyst
film that ensure electron availabilities for the reduction
of oxygen. In this case almost no hydrogen peroxide
is produced and the reduction of oxygen is mostly
performed via a 4 electrons pathway. However, under
acidic conditions, at pH 6 and lower, the catalytic activity
of these inks decreased rapidly.
were recorded on an Uvikon XL spectrometer. Chemicals
were purchased from Aldrich and were used as received.
Solvents were purchased form Aldrich or VWR and
were used as received. THF (K/benzophenone, N2) was
distilled before use. MWNT commercial grade NC3100
(>95%) were purchased from Nanocyl.
Synthesis of porphyrin ligands
5-(4-Methoxyphenyl)-dipyrromethane 3. In a two
neck round bottom flask equipped with a stir bar and a
gas inlet, 4-methoxy benzaldehyde (92.8 mmol, 11.3 mL)
and pyrrole (161.2 mL, 25 equiv.) were mixed. The
reaction mixture was degassed for 15 min in argon under
dark at room temperature, and then TFA (707 mL, 0.1
equiv.) was added. The solution was stirred for further
half an hour. The reaction was monitored by TLC, after
that the reaction mixture was quenched by Et3N. The
excess pyrrole was recovered under reduced pressure.
The resulting solid was dissolved in CH2Cl2 and directly
loaded on a silica gel chromatography column. The
desired compound eluted with 70% CH2Cl2-cyclohexane
EXPERIMENTAL
Materials and methods
Mass spectra: ESI: Micromass MS/MS ZABSpec
TOFF spectrometer. MALDI-TOF: Microflex-LT Bruker
Daltonics were performed at the C.R.M.P.O. (University
of Rennes 1) using a-cyano-4-hydroxycinnamic acid
1
as the matrix. H- and 13C-NMR spectra were recorded
either on BrukerAvance500 or BrukerAvance400 spectro-
meters equipped with a BBFO probe. Spectra were
referenced with residual solvent protons. UV/vis spectra
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 6–10

OXYGEN REDUCTION REACTION CATALYZED BY OVERHANGING CARBOXYLIC ACID
7
Table 2. Number of electrons involved in the
a
reduction of O2 at -0.75V vs. Ag/AgCl
1Fe
MWNT-1Fe
pH 10
pH 13
3.22
2.81
3.54
2.82
3.98
3.97
aPotential chosen on the plateau for MWNT-1Fe.
1 equiv.) were dissolved in freshly prepared distilled
CH2Cl2 (600 mL) in a 1 L round-bottomed flask
containing molecular sieves, degassed with a stream
of Ar for 15 min. Then BF3-Et2O (110 mL, 0.1 equiv.)
was added slowly over 30 s. The reaction was stirred at
room temperature and monitored by TLC and MALDI-
TOF. After 2 h, DDQ (2.7 g, 1.5 equiv.) was added, and
the reaction mixture was stirred at room temperature
for a further 1 h. The complete reaction mixture was
quenched by Et3N and evaporated under reduced
pressure to give a black solid which was dissolved in
CH2Cl2. The mixture of two atropisomers i.e. ab and
aa-bis-2-nitrophenylporphyrin were purified by silica
gel column chromatography using CH2Cl2 as eluent. The
atropisomers could not be separated by the usual method
of column chromatography on silica gel due to the same
polarity. Overall yield: 20 % (600 mg, 0.78 mmol).
a-5,15-Bis-(2-aminophenyl)-10,20-bis-(4-methoxy-
phenyl)-porphyrin 5. ab and aa atropisomers of the
dinitroporphyrin 4 (5.2 mmol, 4 g) were dissolved in
the mixture of CH2Cl2-MeOH (100 mL), taken in a 2 L
.
conical flask along with a reducing agent SnCl2 2H2O
(11.8 g, 10 equiv.) and concentrated HCl (200 mL) was
added slowly to the mixture. The resulting green solution
was stirred for 2 days at RT. After completion of the
reaction (monitored by MALDI-TOF), it was quenched
by aqueous KOH solution at 0°C under ice. The resulting
violet solution was washed several times with water and
CHCl3. The organic layers were collected and dried over
MgSO4. Yield: 80% (2.95 g, 4.19 mmol).
Steric decompression of two atropisomers. In a 1 L
two necks round bottom flask equipped with a stir bar
and a condenser, 200 g of silica (60 mm) were added
to toluene (400 mL). The reaction mixture was heated
to 80°C and degassed with argon during 45 min.
Then 2.9 g of the ab and aa atropisomers of the bis-
2-aminophenylporphyrin obtained previously were
dissolved in toluene and added to the silica gel mixture in
toluene. After evaporation of the solvent, the compound
was dissolved in minimum amount of CH2Cl2 and purified
by column chromatography. The two atropisomers were
separated on a silica gel chromatography eluted with
CH2Cl2/MeOH (ab 0.2%, aa 0.5%). Yield: 5 aa (66%,
1.9 g, 2.70 mmol), ab (34%, 0.9 g, 1.28 mmol). It is worth
to note that the atropisomer aa 5 remains contaminated
with by-products resulting from scrambling reactions not
separable by silica gel flash chromatography and has not
been fully characterized.
Fig. 4. RRDE measurements of oxygen reduction (negative
current) and hydrogen peroxide oxidation (positive current) for
MWNT (black), 1Fe (pink) and MWNT-1Fe (blue) at (a) pH
13 and (b) pH 10 in O2-saturated solutions. The ring electrode
was polarized at 0.260 V vs. Ag/AgCl. Rotation rate: 400 rpm;
scan rate: 5 mV s-1
1
was obtained in 50% yield (11.65 g, 46.17 mmol). H
NMR (CDCl3, 298 K, 500.13 MHz): d 7.90 (2H, pyrNH),
7.18 (2H, d, 3J = 8.61 Hz, aro2), 6.91 (2H, d, 3J = 8.61 Hz,
aro3), 6.71 (2H, m, pyr4), 6.22 (2H, m, pyr3), 5.97 (2H,
m, pyr2), 5.44 (1H, CHa), 3.84 (3H, s, OMe). 13C NMR
(CDCl3, 298 K, 500.13 MHz): d 158.5, 134.3, 132.9,
129.4, 117.2, 114, 108.3, 107.2, 55.3, 43.2. ESI-HRMS:
calcd m/z = 275.1154 [M–H+Na]+ for C16H16N2NaO,
found 275.1159.
5,15-Bis-(2-nitrophenyl)-10,20-bis-(4-methoxyphenyl)-
porphyrin 4. Samples of 4-methoxyphenyldipyrrometh-
ane 3 (7.9 mmol, 2 g) and 2-nitrobenzaldehyde (1.19 g,
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 7–10

8
B. BOITREL ET AL.
3
a-5,15-Bis-(2-[{3-chloromethyl}benzoylamido]-phenyl)-
10,20-bis-(4-methoxyphenyl)-porphyrin 6. A 500 mL
two neck round bottom flask equipped with a stirrer and
cooled in an ice bath was charged with compound 5 (0.99
mmol, 700 mg), dry CH2Cl2 (300 mL) and NEt3 (350 mL,
2.5 equiv.). 3-(chloromethyl)benzoyl chloride (420 mL,
3 equiv.) was then added dropwise under argon
atmosphere. The reaction mixture was allowed to stir
for three hours. Then the reaction was quenched by
water and the organic layer was separated. The solvent
was removed under vacuum. The resulting solid was
dissolved in CH2Cl2 and directly loaded on a silica gel
chromatography column. The expected compound eluted
with0.2%CH2Cl2/MeOH, wasobtainedin90%yield(902
mg, 0.89 mmol). 1H NMR (CDCl3, 298 K, 500.13 MHz):
d 8.93 (4H, d, 3J = 4.81 Hz, bpyr), 8.87 (4H, d, 3J = 4.81
Hz, bpyr), 8.93 (2H, d, 3J = 8.06 Hz, aro2′), 8.13 (2H, d,
3J =7.49 Hz, aro5′), 8.11 (2H, d, 3J = 7.60 Hz, aro9), 8.02
1.23 (4H, bs, CH2 ester), -0.55 (6H, t, J = 6.58 Hz,
CH3 ester), -2.40 (2H, s, NHint). 13C NMR (CDCl3, 298 K,
500 MHz): d 167.5, 164.4, 159.6, 138.8, 135.9, 135.2,
134.9, 133.8, 133.7, 132.4, 131.2, 129.9, 128.2, 127.4,
125.9, 122.8, 120.7, 119.7, 113.7, 112.5, 59.9, 58.8, 55.5,
42.5, 11.9. ESI-HRMS: calcd m/z = 1097.4232 [M + H]+
for C69H57N6O8, found 1097.4234, calcd m/z = 1119.4051
[M + Na]+ for C69H56N6O8Na, found 1119.4047. UV-vis
(DMF): l/nm (10-3 e, dm3 mol-1 cm-1): 425 (380), 520
(15.8), 558 (8.2), 596 (4.4), 653 (3.4).
a-5,15-Bis-({{2,2-(3,3-[2,2-(dicarboxylicacid)propane-
1,3-diyl]-dibenzoyl-amido]-diphenyl)-10,20-bis-(4-
hydroxyphenyl)-porphyrin 2. Boron tribromide (2.5 mL,
50 equiv.) was added to compound 7 (0.54 mmol,
600 mg) was dissolved in DCM (100 mL). After 12 h of
stirring at RT, the reaction was completed. The mixture
was quenched by water. The precipitated compound was
filtered and the green solid was washed with water at
pH = 7. The product was purified by silica gel chromato-
graphy column and eluted with CHCl3/MeOH/AcOH
3
3
(2H, d, J = 7.60 Hz, aro9′), 7.91 (2H, t, J = 8.06 Hz,
3
aro3′), 7.63 (2H, s, NHCO), 7.59 (2H, t, J = 7.49 Hz,
3
1
aro4′), 7.28 (4H, bs, aro8, aro8′), 6.73 (2H, d, J = 7.68
(90/9/1). Yield: 80% (440 mg, 0.43 mmol). H NMR
(DMSO-d6, 298 K, 500.13 MHz): d 9.92 (2H, s, OH),
3
Hz, aro4), 6.55 (2H, d, J = 7.90 Hz, aro6), 6.49 (2H, t,
3J = 7.68, aro5), 6.33 (2H, s, aro2), 4.09 (6H, s, OMe), 3.30
(4H, s, CH2 bz), -2.61 (2H, s, NHint). 13C NMR (CDCl3,
298 K, 500.13 MHz): d 164.7, 159.7, 138.7, 137.1,
135.7, 135.6, 134.9, 134.7, 133.6, 131.7, 130.8, 129.9,
128.4, 126.2, 126.1, 123.2, 120.9, 120.7, 113.7, 112.5,
112.4, 55.6, 44.3. ESI-HRMS: calcd m/z = 1009.3030
[M + H]+ for C62H47N6O435Cl2, found 1009.3031, calcd
m/z = 973.3263 [M–HCl+H]+ for C62H46N6O435Cl, found
973.3268. UV-vis (DMF): l/nm (10-3 e, dm3 mol-1 cm-1):
426 (364), 521 (14), 558 (8), 599 (4), 652 (3.4).
8.84 (4H, d, J = 4.55 Hz, bpyr), 8.77 (4H, d, J =
4.55 Hz, bpyr), 8.59 (2H, s, NHCO), 8.37 (4H, bs, aro2′,
aro5′), 8.12 (2H, bs, aro9), 7.92 (2H, t, 3J = 7.92 Hz, aro3′),
7.84 (2H, d, 3J = 6.25 Hz, aro9′), 7.75 (2H, t, 3J = 7.60 Hz,
aro4), 7.24 (2H, d, 3J = 8.04 Hz, aro6), 7.16 (4H, bs, aro8,
aro8′), 6.90 (2H, t, 3J = 7.49 Hz, aro5 ), 6.72 (2H, d, 3J =
7.77 Hz, aro4), 4.57 (2H, s, aro2), 1.19 (4H, s, CH2 bz),
-2.70 (2H, bs, NHint). 13C NMR (DMSO-d6, 298 K,
500 MHz): d 171.5, 165.7, 157.8, 139.2, 136.2, 135.9,
135.7, 135.3, 134.9, 134.8, 132.1, 132.0, 129.7, 128.1,
126.8, 126.4, 124.6, 124.3, 120.7, 115.3, 114.3, 59.3,
40.5. ESI-HRMS: calcd m/z = 1013.3293 [M + H]+ for
C63H45N6O8, found 1013.3288, calcd m/z = 1035.3112
[M-H+Na]+ for C63H43N6NaO8, found 1035.3102. UV-vis
(DMF): l/nm (10-3 e, dm3 mol-1 cm-1): 428 (344), 522
(19), 560 (12), 599 (7.6), 655 (6).
Iron insertion. A free-base solution of porphyrin 2 in
THF in the presence of an excess of iron bromide and
2,6-lutidine was heated at reflux overnight inside a glove
box. During this process, the decarboxylation reaction
leading to both 1Fe and 2Fe was observed and the two
complexes were obtained in roughly equal proportions
as indicated by TLC analysis. The resulting mixture
was taken out of the glove box, washed with HCl (1M),
and dried. There were easily separated by silica gel
chromatography using a gradient of MeOH in CHCl3
(from 0.2% to 2.2%) and identified by MALDI-TOF
mass spectrometry (1Fe: 41%, calcd m/z = 1022.25 [M +
H]+ for C62H42FeN6O6 found 1022.30 ; 2Fe: 37%, calcd
m/z = 1066.24 [M + H]+ for C63H42FeN6O8 found 1066.41).
MWNT. MWNTs (60 mg) were sonicated in nitric acid
(35vol%)(150mL)withasonicbath(Fisherbrand,37kHz,
power 100% for 10 min and then 40% for 30 min) and
then heated at 100°C for 5 h. The suspension was then
cooled and vacuum filtered through a 0.2 mm PTFE
3
3
a-5,15-Bis-({2,2-(3,3-[2,2-(diethoxycarbonyl)propane-
1,3-diyl]-dibenzoyl-amido]-diphenyl)-10,20-bis-(4-
methoxy-phenyl)-porphyrin 7. Sodium metal (182 mg,
10 equiv.) was added to the absolute alcohol (20 mL)
in a small round bottom flask and stirred for few
minutes until the complete consumption of Na. Diethyl
malonate (1.2 mL, 10 equiv.) was added to this solution
at room temperature and stirred for half an hour. The
resulting mixture was added to a solution of porphyrin 6
(0.79 mol, 800 mg, 1 equiv.) in CH2Cl2 (600 mL) and
the solution was turned immediately from violet to green.
After 2 h of stirring the reaction was quenched by H2O,
the organic layer was separated and removed under
vacuum. The desired product was purified on a silica
gel chromatography column eluted with 0.3% CH2Cl2/
MeOH. The expected compound was obtained in 80%
1
yield (702 mg, 0.64 mmol). H NMR (CDCl3, 298 K,
500 MHz): d 9.15 (2H, d, 3J = 8.39 Hz, aro2′), 8.93 (4H,
d, 3J = 4.58 Hz, bpyr), 8.86 (4H, d, 3J = 4.58 Hz, bpyr),
8.22 (2H, s, NHCO), 8.19 (2H, bs, aro9), 8.01 (2H, d, 3J =
7.54 Hz, aro5′), 7.98 (2H, bs, aro9′), 7.91 (2H, t, 3J = 7.70
Hz, aro3’), 7.70 (2H, d, 2Hh, 3J = 7.98 Hz, aro6), 7.55 (2H,
t, 3J = 7.70 Hz, aro4′), 7.31 (4H, bs, aro8, aro8′), 6.98 (2H,
t, 3J = 7.90 Hz, aro5), 6.74 (2H, d, 3J = 7.90 Hz, aro4), 4.99
(2H, s, aro2), 4.08 (6H, s, OCH3), 1.70 (4H, s, CH2 bz),
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 8–10

OXYGEN REDUCTION REACTION CATALYZED BY OVERHANGING CARBOXYLIC ACID
9
membrane and washed with water. The nanotubes were
with a (CMOS) PHOTON 100 detector founded by
FEDER, Mo-Ka radiation (l = 0.71073 Å, multilayer
monochromator), T = 150 (2) K; triclinic P -1 (I.T.#2),
a = 14.499 (2), b = 14.729 (2), c = 17.848 (3) Å, a =
90.838 (5), b = 109.567 (6), g = 117.362 (5)°, V = 3124.0
(8) Å3. Z = 2, d = 1.086 g.cm-3, m = 0.291 mm-1. The
structure was solved by dual-space algorithm using
the SHELXT program [29], and then refined with full-
matrix least-squares methods based on F2 (SHELXL)
[30]. The contribution of the disordered solvents to the
calculated structure factors was estimated following the
BYPASS algorithm [31], implemented as the SQUEEZE
option in PLATON [32]. A new data set, free of solvent
contribution, was then used in the final refinement. All
non-hydrogen atoms were refined with anisotropic atomic
displacement parameters. H atoms were finally included
in their calculated positions and treated as riding on their
parent atom with constrained thermal parameters. A
final refinement on F2 with 13966 unique intensities and
672 parameters converged at wR(F2) = 0.2340 (R(F) =
0.0888) for 9825 observed reflections with I > 2s(I).
redispersed in NaOH 2 M (100 mL) using the sonic bath
(100% for 10 min) and then filtered through a PTFE
membrane and washed with deionized water, and then
HCl 1 M followed by deionized water until the filtrate
was neutral.
Electrochemical experiments
Sample preparation. For the preparation of the sample
containing nanotubes (MWNT-1Fe, MWNT-2Fe), a
mixture of purified MWNT (9 mg) and porphyrin 1Fe
(9 mg) or porphyrin 2Fe (9 mg) in dry THF (10 ml) was
homogenized using a sonic bath (Fisherbrand, 37 kHz,
power 100%) for 15 min. The THF was gently evaporated
with a steam of N2 and the mixtures were dried under
vacuum. 3 mg of mixture were dispersed in 750 ml of
ethanol and 75 ml of Nafion solution (5% in alcohol). The
mixtures were homogenized using a sonic bath until they
formed homogenous inks. For the MWNT ink, the same
procedure was followed but without adding porphyrins.
For the reference 1Fe and 2Fe inks, 3 mg of porphyrins
were directly dispersed in 750 ml of ethanol and 75 ml of
Nafion solution (5% in alcohol).
Acknowledgments
Electrode preparation. Before each measurement,
the glassy carbon (GC) disk (5 mm, 0.196 cm2) used as
rotating electrode was polished with aqueous dispersions
of synthetic diamonds (1 mm), then rinsed and sonicated
with water. 5 ml of the catalyst inks were deposited by drop-
casting onto the GC disk, then dried in air. For pH 6, a new
ink was deposited onto the GC disk for each rotation step.
Electrochemical measurements. The instrument
used was a VSP bipotentiostat (Bio-Logic SAS). The
electrochemical tests were carried out in 0.5 M H2SO4
solution in a three electrode glass cell, thermostated at
25°C. A “CE to Ground” connection with a saturated
KCl Ag/AgCl electrode as reference and a graphite plate
as counter electrode was used. As working electrode, a
Pine rotating ring disk electrode (RRDE) with catalyst-
loaded GC disk (0.196 cm2) and Pt ring (0.110 cm2)
was controlled by a speed control unit from Princeton
Région Bretagne as well as CNRS are deeply
acknowledged for their significant support to our
research team. This work was also supported by the
JST-ANR program TMOL “Molecular Technology”
project MECANO (ANR-14-JTIC-0002-01) and project
MAGMA (ANR-16-CE29-0027-01).
Supporting information
Detailed spectroscopic data and detailed electro-
chemical measurements, Figures S1–S48 are given in
the supplementary material. This material is available
free of charge via the Internet at http://www.worldscinet.
com/jpp/jpp.shtml. Crystallographic data (excluding
structure factors) for the structures reported in this paper
have been deposited at the Cambridge Crystallographic
Data Center under deposition number 1916858. Copies
can be obtained on request, free-of-charge, at www.
ccdc.cam.ac.uk/conts/retrieving.html or from the
Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK (fax: +44 1223-336-033
or email: deposit@ccdc.cam.ac.uk).
Applied Research Model 636 Electrode Rotator. The
-1
.
voltammograms were recorded at 5 mV s in stationary
conditions (with various rotating rates: 0, 400, 800,
1200, 1600, and 2000 rpm) in O2-saturated solutions.
An average current was calculated from the forward and
backward scans. All potentials reported in this paper
refer to that of the Ag/AgCl electrode. H2O2 production
REFERENCES
was monitored in the RRDE configuration at 400 rpm
1. Momenteau M and Reed CA. Chem. Rev. 1994; 94:
659–698.
2. Lemon CM, Dogutan DK and Nocera DG. In Hand-
book of Porphyrin Science, Vol. 21. Kadish KM,
Smith KM, Guilard R. (Eds.) Word Scientific: Sin-
gapore, 2012; pp. 1–143.
3. Bezerra CWB, Zhang L, Lee K, Liu H, Marques
ALB, Marques EP, Wang H and Zhang J. Electro-
chem. Acta 2008; 53: 4937–4951.
-1
.
with a CV at the GC disk (5 mV s ). The collection
coefficient of the RRDE (0.20) was measured using the
one-electron Fe(CN)63-/Fe(CN)64- redox couple, according
to the manufacturer’s instructions.
X-ray crystallographic studies
CCDC 1916858, C62H41FeN6O6, M = 1021.86. D8
VENTURE Bruker AXS diffractometer equipped
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 9–10

10
B. BOITREL ET AL.
4. Kim E, Chufan EE, Kamaraj K and Karlin KD.
Chem. Rev. 2004; 104: 1077–1133.
5. Collman JP, Devaraj NK, Decréau RA, Yang Y, Yan
Y-L, Ebina W, Eberspacher TA and Chidsey CED.
Science 2007; 315: 1565–1568.
19. Balieu S, Halime Z, Lachkar M and Boitrel B. J.
Porphyrins Phthalocyanines 2008; 12: 1223–1231.
20 Le Gac S, Najjari B, Dorcet V, Roisnel T, Fusaro
L, Luhmer M, Furet E, Halet J-F and Boitrel B.
Chem. — Eur. J. 2013; 19: 11021–11038.
21. Kruusenberg I, Alexeyeva N and Tammeveski K.
Carbon 2009; 47: 651–658.
6. Ricard D, Andrioletti B, L’Her M and Boitrel B.
Chem. Commun. 1999: 1523–1524.
7. Ricard D, L’Her M, Richard P and Boitrel B. Chem.
— Eur. J. 2001; 7: 3291–3297.
8. Ricard D, DidierA, L’Her M and Boitrel B. Comptes
Rendus Chimie 2002; 5: 33–36.
9. Ricard D, Didier A, L’Her M and Boitrel B. Chem-
BioChem 2001; 2: 144–148.
10. Rosenthal J and Nocera DG. Acc. Chem. Res. 2007;
40: 543–553.
11. Rosenthal J and Nocera DG. Prog. Inorg. Chem.
2007; 55: 483–544.
12. Carver CT, Matson BD and Mayer JM. J. Am.
Chem. Soc. 2012; 134: 5444−5447.
13. MorozanA, Campidelli S, Filoramo A, Jousselme B
and Palacin S. Carbon 2011; 49: 4839–4847.
14. Rigsby ML, Wasylenko DJ, Pegis ML and Mayer
JM. J. Am. Chem. Soc. 2015; 137: 4296–4299.
15. Hanana M, Arcostanzo H, Das PK, Bouget M, Le
Gac S, Okuno H, Cornut R, Jousselme B, Dorcet
V, Boitrel B and Campidelli S. New. J. Chem. 2018;
42: 19749–19754.
22. Morozan A, Jégou P, Pinault M, Campidelli S, Jous-
selme B and Palacin S. ChemSusChem 2012; 5:
647–651.
23. van Veen JAR and Colijn HA. Ber. Bunsenges.
Phys. Chem. 1981; 85: 700–704.
24. Choi CH, Baldizzone C, Polymeros G, Pizzutilo E,
Kasian O, SchuppertAK, Sahraie NR, Sougrati MT,
Mayrhofer KJJ and Jaouen F. ACS Catal. 2016; 6:
3136–3146.
25. Meier H, Tschirwitz U, Zimmerhackl E, Albrecht
W and Zeitler G. J. Phys. Chem. 1977; 81: 712–718.
26. Baraton S, Coutanceau C, Roux C, Hahn F and
Léger JM. O J. Electroanal. Chem. 2005; 577:
223–234.
27. Li W, Yu A, Higging DC, Llanos BG and Chen ZB.
J. Am. Chem. Soc. 2010; 132: 17056–17058.
28. Zhou R, Zheng Y, Jaroniec M and Qiao S-Z. ACS
Catal. 2016; 6: 4720−4728.
29. Sheldrick GM. Acta Cryst. 2015; A71: 3–8.
30. Sheldrick GM. Acta Cryst. 2015; C71: 3–8.
31. Sluis Pvd and Spek AL. Acta Cryst. 1990; A46:
194–201.
16. Zhang W, Lai W and Cao R. Chem. Rev. 2017; 117:
3717−3797.
17. Pegis ML, Wise CF, Martin DJ and Mayer JM.
Chem. Rev. 2018; 118: 2340−2391.
32. Spek AL. J. Appl. Cryst. 2003; 36: 7–13.
18. Zhao YM, Yu GQ, Wang FF, Wei PJ and Liu JG.
Chem. — Eur. J. 2019; 25: 3726–3739.
Copyright © 2019 World Scientific Publishing Company
J. Porphyrins Phthalocyanines 2019; 23: 10–10