Through-Space Electrostatic Interactions Surpass Classical Through-Bond Electronic Effects in Enhancing CO2 Reduction Performance of Iron Porphyrins

Article abstract of DOI:10.1002/cssc.202002718

In his pioneering work to unravel the catalytic power of enzymes, Warshel has pertinently validated that electrostatic interactions play a major role in the activation of substrates. Implementing such chemical artifice in molecular catalysts may help impr

Full text of DOI:10.1002/cssc.202002718

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
Very Important Paper
1
2
Through-Space Electrostatic Interactions Surpass Classical
Through-Bond Electronic Effects in Enhancing CO2
Reduction Performance of Iron Porphyrins
3
4
5
6
7
8
9
Asma Khadhraoui,[a] Philipp Gotico,[b] Winfried Leibl,[b] Zakaria Halime,*[a] and
Ally Aukauloo*[a, b]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
In his pioneering work to unravel the catalytic power of
enzymes, Warshel has pertinently validated that electrostatic
interactions play a major role in the activation of substrates.
Implementing such chemical artifice in molecular catalysts may
help improve their catalytic properties. In this study, a series of
tetra-, di-, and mono-substituted iron porphyrins with cationic
imidazolium groups were designed. Their presence in the
second coordination sphere helped stabilize the [FeÀ CO2]
intermediate through electrostatic interactions. It was found
herein that the electrocatalytic overpotential is a function of the
number of embarked imidazolium. Importantly, a gain of six
orders of magnitude in turnover frequencies was observed
going from a tetra- to a mono-substituted catalyst. Further-
more, the comparative study showed that catalytic perform-
ances trend of through-space electrostatic interaction, a first
topological effect reported for iron porphyrins, outperforms the
classical through-structure electronic effect.
Introduction
hemic models have long been studied as models for the
Cytochrome c Oxidase to perform the four-electron reduction
of O2 or the reductive activation of O2 by the oxygenases family
of enzymes.[3c,f,4] Savéant and co-workers provided a first
thorough electrochemical investigation of the electrocatalytic
activity of the iron (III) tetra-phenylporphyrin, FeTPP.[4b,5] The
synthetic development of functionalized tetra-arylporphyrins
has led to the sophistication of these molecular catalysts to
optimize the electrocatalytic reduction of CO2, with catalytic
overpotential and turnover frequency (TOF) as the main
characteristics for performance (structural modifications sum-
marized in Figure S1).
A temporary solution to deal with the mitigation of carbon
dioxide (CO2) levels in our atmosphere resides in the grand-
scale capture of CO2 to be buried into rocks.[1] Although several
incentives on this strategy have been launched worldwide,
many concerns still persist on the long-run stability and cost
efficiency of such programs. Such solutions are being consid-
ered primarily because of the lack of knowledge on how to
massively and sustainably transform CO2 to energy-rich forms of
carbon. We are currently witnessing a substantial effort from
chemists to address the issue of capturing and converting CO2
into a fuel or stitching it in a reduced form as a C1 unit in
organic synthesis.[2] This promising research field is a priority for
our society with the target to set a virtuous cycle through
valorization of exhausted CO2.
In molecular chemistry, a great effort is dedicated to design
new molecular catalysts inspired by active sites of families of
enzymes that have been efficiently developed by nature to
achieve these chemical transformations. Many metal complexes
with different sets of ligands have already shown interesting
activities.[3] Among these, metalloporphyrins have captured a
particular attention for the CO2 reduction inasmuch as the
At first, electron-withdrawing groups were introduced on
the porphyrin platform to lower the overpotential of the CO2
reduction.[4c,6] However, the beneficial lowering of the over-
potential was counterbalanced by lower reaction rate constants
and TOFs.[7] The positioning of proton or hydrogen bonding
donors in the second coordination sphere has led to a
remarkable gain in both TOFs and overpotentials.[8] We and
others further introduced cationic functions on the phenyl
groups on the meso positions of the porphyrins with the
targeted objective to stabilize the FeÀ CO2.[6,9] This intermediate
was later spectroscopically characterized by Dey and co-workers
2À [10]
as FeIIÀ CO2
.
The benefit of the presence of these cationic
functions was demonstrated by an important lowering of the
overpotential together with enhanced electrocatalytic activities.
Most of these achievements have been realized with catalysts
bearing second coordination sphere groups on the ortho
position of the porphyrin meso aryl rings in fully substituted
systems with a C4 symmetry. The ortho position was selected to
place these groups above the metallic center where interactions
with a FeÀ CO2 intermediate are expected to take place. Of note,
the symmetrical introduction of these groups on the four meso
[a] A. Khadhraoui, Dr. Z. Halime, Prof. A. Aukauloo
Université Paris-Saclay, CNRS
Institut de Chimie Moléculaire et des Matériaux d’Orsay (ICMMO)
91405, Orsay, France
E-mail: zakaria.halime@universite-paris-saclay.fr
ally.aukauloo@universite-paris-saclay.fr
[b] Dr. P. Gotico, Dr. W. Leibl, Prof. A. Aukauloo
Université Paris-Saclay, CEA
Institute for Integrative Biology of the Cell (I2BC)
91198, Gif-sur-Yvette, France
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/cssc.202002718
ChemSusChem 2021, 14, 1308–1315
1308
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
aryls of the porphyrin platform is mainly motivated by synthetic
practicality.
the mono-substituted imidazolium derivative 1À ImÀ Fe was
compared to its analogue bearing six fluorine atoms as
inductive electron-withdrawing groups (abbreviated as
1À ImÀ F6À Fe) to investigate the cumulative action of these
groups.
1
2
3
4
5
6
7
8
9
Using our previously reported catalyst bearing four imidazo-
lium arms in the ortho position of the meso aryl (4À ImÀ Fe,
Scheme 1) as a reference catalyst,[9] in this work, our first
interrogation concerns the question whether the exaltation of
the observed electrocatalytic properties is the collective con-
tribution of all four groups on the porphyrin rim. To address
this important question, we have designed two iron porphyrins
holding two and one imidazolium motifs hereby abbreviated as
2À ImÀ Fe and 1À ImÀ Fe, respectively (Scheme 1).
We found that the overpotential of 4À ImÀ Fe, 2À ImÀ Fe and
1À ImÀ Fe catalysts increases with the number of embarked
imidazolium units ranging from 230 to 430 to 620 mV,
respectively, compared to 680 mV for the parent tetra-phenyl
iron porphyrin, FeTPP. Importantly, we evidenced a gain of six
orders of magnitude for the TOFs going from the tetra- to the
mono-substituted catalyst. Accordingly, in line with Warshel’s
electrostatic theory in explaining the unmatched reactivity of
enzymes,[11] we found that the electrocatalytic performance
trend of through-space electrostatic interaction models outper-
forms the classical electronic effects strategy. This is further
exemplified for the 1À ImÀ F6À Fe complex whose TOF is three
orders of magnitude higher than the FeTPPF8 derivative at
similar overpotential.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
To further examine whether combining effects of different
chemical functions may provide new reactivity patterns, we
have also been interested in this study to combine the through-
structure electronic effects with an electrostatic one. As such,
Results and Discussion
Synthesis of iron-porphyrins holding imidazolium moities
Catalysts 2À ImÀ Fe, 1À ImÀ Fe and 1À ImÀ F6À Fe were prepared
starting from three different aminophenylporphyrins having
two (3) or one (7 and 7’) amine functions (Scheme 2). Porphyrin
3 was obtained by the reduction of the nitro groups of synthon
2 that was prepared according to the 2+2 Lindsey condensa-
tion
of
benzaldehyde
and
the
ortho-nitrophenyl-
dipyrromethane,[12] which in turn resulted from an acid-
catalyzed condensation of pyrrole and ortho-nitrobenzaldehyde.
To keep a similar general topology, only the αα atropisomer,
having two amino groups on the same side with respect to the
porphyrin plane was included in this study. The condensation
of 4 equivalents of pyrrole and 1 equivalent of ortho-nitro-
benzaldehyde with 3 equivalents benzaldehyde or di-ortho-
Scheme 1. Schematic representation of modified iron-porphyrin catalysts
bearing four (4À ImÀ Fe), two (2À ImÀ Fe), and one (1À ImÀ Fe and 1À ImÀ F6À Fe)
imidazolium arms.
Scheme 2. Synthesis of iron porphyrin catalysts 2À ImÀ Fe, 1À ImÀ Fe, and 1À ImÀ F6À Fe.
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1309
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
fluorobenzaldehyde results in the mono-(ortho-nitrophenyl)
porphyrin 6 and 6’ respectively, which were subsequently
reduced to the amino analogues 7 and 7’ respectively. Starting
from these different aminophenylporphyrins, we used a similar
synthetic tactic that we previously reported for catalyst
4À ImÀ Fe.[9] In the first step, the benzylic linker is introduced in a
reaction between the porphyrin’s amine groups and the 3-
(chloromethyl)benzoyl chloride. When the amino groups are in
the ortho position of the meso aryls of the porphyrin (4, 8, and
8’), the obtained semi-rigid aryl-amide-aryl motif was shown to
be optimal for positioning functional groups in pre-organized
fashion close to the coordination sphere of the metal ion lying
in the center of the porphyrin ring.[13] In the next step of this
synthesis, a ferrous chloride salt was used in presence of a base
to insert the iron ion in the N4 coordinating cavity of the
porphyrins. The obtained iron complexes 5, 9 and 9’ were then
treated with stoichiometric amounts of N-methylimidazole to
substitute the benzylic chlorides with methyimidazolium groups
and yield the targeted catalysts 2À ImÀ Fe, 1À ImÀ Fe, and
1À ImÀ F6À Fe as chloride salts. The above-mentioned porphyrin
described in the Experimental Section and Supporting Informa-
tion.
1
2
3
4
5
6
7
8
9
Influence of the number of methylimidazolium groups
Under argon atmosphere, in presence of 0.1 m of tetra-N-butyl-
ammonium hexafluorophosphate (TBAPF6) as supporting elec-
trolyte, and dimethylformamide (DMF) as solvent, all the
functionalized iron porphyrins in Scheme 1 display three
reversible waves corresponding to the formal redox couples
FeIII/II, FeII/I, and FeI/0 (Figure 1 and Table 1). Recent reports
suggest that the second and the third redox waves can be
mainly centered on the porphyrin because of the redox non-
innocent character of the tetra-pyrrolic ligand and therefore can
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
*
À
be alternatively formulated as FeIIporph/FeIIporph
and
*
**
FeIIporph /FeIIporph
couples (porph=porphyrin).[14] Cata-
À
2À
lysts 2À ImÀ Fe and 1À ImÀ Fe were designed in a way to divide
the number of methylimidazolium groups in the reference
catalyst 4À ImÀ Fe by two and by four, respectively, while
maintaining a similar general topology. Compared to this
reference catalyst, the first direct impact of decreasing the
number of methylimidazolium groups can be observed under
argon atmosphere with a 200 and 390 mV cathodic shift of the
third reduction wave (FeI/0) for catalysts 2À ImÀ Fe and 1À ImÀ Fe,
1
ligands were characterized by H and 13C NMR spectroscopy,
UV/Vis spectroscopy, and high-resolution electron spray ioniza-
tion mass spectrometry (HRÀ ESI MS), and the metallated
porphyrins were characterized by UV/Vis spectroscopy and HR-
ESI MS. Detailed synthetic procedures and characterizations are
°
Figure 1. CVs of 1 mm of modified iron porphyrin complexes 4À ImÀ Fe (green), 2À ImÀ Fe (blue), and 1À ImÀ Fe (red) in DMF containing 0.1 m TBAPF6 at 25 C (a)
under argon and (b) under 1 atm of CO2 with 5.5 m H2O. (c) Comparison of the TOF calculated from FOW analysis as a function of catalytic overpotentials.
Catalytic performance of nonfunctionalized FeTPP (black) and fluorinated derivates FeTPPÀ F8 (dark gray) and FeTPPÀ F20 (gray) are included for comparison
under the same experimental conditions in (b).
Table 1. Summary of electrochemical properties of the modified iron-porphyrin catalysts.
Catalysts
E(FeI/0) [V]
η
[a] [mV]
TOFmax[b] [sÀ 1
]
log TOFmax
FE[c] [%]
KCO2[d] [mÀ 1
]
KIE[e]
4À ImÀ Fe
2À ImÀ Fe
1À ImÀ Fe
1À ImÀ F6À Fe
FeTPP
FeTPPF8
FeTPPF20
À 0.969
À 1.168
À 1.357
À 1.266
À 1.420
À 1.296
À 1.106
229
428
617
526
680
556
366
3.10×102
8.57×105
2.05×108
9.80×105
1.07×104
1.57×102
1.88×101
2.49
5.93
8.31
5.99
4.03
2.20
1.27
91
93
89
14.3 (�0.3)
18 (�5)
0.43 (�0.08)
1.12 (�0.13)
1.02 (�0.12)
2.71 (�0.30)
1.81 (�0.06)
1.27 (�0.05)
1.16 (�0.05)
2.1 (�0.2)
2.1 (�0.2)
0.9 (�0.1)
0.9 (�0.1)
0.9 (�0.1)
92
94–100[18]
–
96[20]
[a] Overpotential calculated as the difference between the E(FeI/0) and E (CO2/CO)=0.74 V vs. NHE. [b] TOFmax estimated from FOW analysis. [c] FE measured
from bulk electrolysis experiments (see Supporting Information). [d] CO2 binding constant estimated using CVs in dry DMF purged with CO2 measured at
1 VsÀ 1 (calculations in the Supporting Information). [e] KIE (kcat,H2O/kcat,D2O) calculated as average using different concentrations of H2O/D2O (1.1 to 5.5 m).
°
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1310
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
respectively (Figure 1a and Table 1). This significant potential
shift indicates that, similar to the through-structure inductive
effect, the through-space charge stabilization of the catalyst’s
anionic forms, that is, formal FeI and Fe0 states, by the
imidazolium arms is also a cumulative effect.
co-workers have shown, in the case of O2 reduction by iron
porphyrins, that the modification of different parameters
involved in the rate-determining step such as the intrinsic
catalyst parameter of EFe(III/II) (through ligand modifications of
the catalyst) or the catalytic conditions such as acid pKa and
concentration can result in different values of correlation
coefficients.[16] However, Costentin and Savéant reported that
such correlations only exist within families of catalyst in which
the through-structure electronic effects influence in a similar
manner the catalyst standard potential on one hand, and the
thermodynamics and kinetics of the rate-determining catalytic
sequence on the other hand.[17] Remarkably, we report here for
the first time that in this family of imidazolium-functionalized
iron porphyrins, the overall driving force brought by through-
space substituent effects influence also the driving force of the
rate-determining sequence in CO2-to-CO catalytic reduction
which translates in the correlation observed in the log(TOF) vs.
overpotential plot (Figure 1c).
1
2
3
4
5
6
7
8
9
The cyclic voltammograms (CVs) of all three catalysts in a
CO2-saturated solution and in presence of H2O as proton source
show an important catalytic current on the third reduction
wave (Figure 1b). GC analysis of the electrochemical reaction
headspace during bulk electrolysis experiments confirmed that
for all three catalysts, the observed catalytic current corre-
sponds exclusively to the 2eÀ reduction of CO2 to CO with a
faradaic efficiency (FE) ranging from 89 to 93% (Figures S29–
32). The catalytic waves of the three catalysts were also
analyzed using the Foot-of-the-Wave (FOW) method, intro-
duced by Savéant and co-workers,[6,7b,15] to estimate the TOF
free from secondary phenomena (substrate and co-substrate
depletion, product inhibition, catalyst degradation, etc.), which
would otherwise require increasing scan rates during CV
measurements to reach the ideal canonical S-shaped catalytic
wave. A plot of the kinetic parameter log(TOF) against the
thermodynamic parameter of overpotential, η= jE(FeI/0)–
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
To better understand the effect of the number of
imidazolium arms on the overall reaction rate, we investigated
on a possible change of the nature of the rate-determining step
of the electrocatalytic reduction. CO2 binding and proton
transfer are in general the two main processes with a significant
contribution to the rate-determining step of the reaction
because electron transfer processes are considered to be fast at
the potential where the electrocatalytic reaction proceeds. The
CO2 binding constant of the active form, that is, the formal Fe0
state of the iron porphyrin derivatives, was estimated by
recording the CV responses in CO2-saturated dry DMF and in
the absence of a proton source, which prevents any catalytic
activity (Figures S9–S12). As shown in Figure 2, there is no direct
correlation between the CO2 binding constant and the number
of methylimidazolium groups. However, 4À ImÀ Fe and 2À ImÀ Fe
°
E (CO2/CO)j, with a reported value of 0.740 V vs. normal
hydrogen electrode (NHE) for the latter in DMF/H2O,[7a]
illustrates important effects of the porphyrin substituents on
the electrocatalytic behavior (Figure 1c).
For example, the classical through-structure strategy of
decreasing the overpotential by decorating the porphyrin ring
with inductive electron-withdrawing fluorine groups results in
linear decrease of the log(TOF) (comparison within FeTPP,
FeTPPÀ F8, and FeTPPÀ F20).[7b,9] The beneficial lowering of the
overpotential because of the thermodynamic ease of reducing
the iron porphyrin comes at a detrimental cost of lower
catalytic rate constant.
have
a
higher CO2 binding constant (14.3 and 18 mÀ 1,
Interestingly, our findings revealed an obvious structure/
reactivity trend illustrated by an increase of the TOF and the
reaction overpotential with decreasing numbers of cationic
methylimidazolium groups when going from 4À ImÀ Fe to
2À FeÀ Im and to 1À ImÀ Fe. From these findings, a linear
correlation like that observed for the through-structure induc-
tive effect can also be drawn for the through-space electrostatic
effect, but the latter is characterized by significantly higher
reaction rates. We previously attributed the catalytic
enhancement observed with the imidazolium-modified catalyst
4À ImÀ Fe, in comparison with FeTPP and its fluorinated
derivatives, to the additional role played by the positively
charged methylimidazolium groups in stabilizing the negatively
charged FeÀ CO2 catalytic intermediate by through-space elec-
trostatic interactions.[9] In addition to confirming this through-
space effect, this study shows that the amplitude of this effect
can be controlled by the number of methylimidazolium groups.
Importantly, a noticeable difference in the slope of the through-
space and through-structure linear correlations can be clearly
observed. The correlation coefficient (inverse slope), extracted
from the plot of log(TOF) vs. overpotential (Figure 1c), varies
from 118 mVdecÀ 1 in TOF for the fluorinated derivates to
66 mVdecÀ 1 in TOF for the imidazolium derivatives. Mayer and
respectively) compared to 1À ImÀ Fe (2.1 mÀ 1) containing only
one of the methylimidazolium groups, but the latter has still a
higher value compared to the non-functionalized FeTPP
Figure 2. Comparison of the CO2 binding constant and KIE for the modified
iron porphyrin complexes as a function of the number of imidazolium
groups.
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1311
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
(0.9 mÀ 1). The close values of the CO2 binding constant between
catalysts with two and four imidazolium arms suggest that a
minimum of two methylimidazolium groups are involved in
“efficient” CO2 capture. The CO2 binding ability of these
modified iron porphyrins can be thought of as a synergistic
effect between the cationic imidazolium groups stabilizing the
anionic CO2 adduct and, to lesser extent, the hydrogen-bonding
aptitude of the amide groups as such functionalities have been
reported to enhance CO2 binding.[8c,e]
2À ImÀ Fe and to 1À ImÀ Fe. However, to probe such differences
in-depth quantum chemical calculations and physico-mathe-
matical analysis of the reaction steps in the catalytic cycle with
relevant experimental validation through intermediate charac-
terization are necessary, which are beyond the scope of this
study.
1
2
3
4
5
6
7
8
9
Combining through-structure electronic effects and
through-space electrostatic interactions
The protonation steps following CO2 binding are often
reported to be the rate-limiting step in the electrocatalytic CO2
reduction by iron porphyrins.[8b,18] Kinetic isotope effect (KIE=
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
As discussed above, both through-structure inductive effects
and through-space electrostatic interactions can be employed
to adjust the redox potential of the FeI/0 redox couple and thus
shift the catalytic potential of CO2 reduction along the
established correlations. Additionally, the through-space elec-
trostatic interactions fulfill an extra role of enhancing the
reaction rate due to the through-space stabilization of the
FeÀ CO2 intermediate. To investigate what catalytic modifica-
tions the association of these two different effects may lead to,
we developed a new analogue catalyst 1À ImÀ F6À Fe combining
six fluorine groups as through-structure electron-withdrawing
groups and one imidazolium arm to induce the through-space
electrostatic interactions.
FOW analysis of the catalytic current observed with
1À ImÀ F6À Fe catalyst under CO2 atmosphere confirms a dual action
of the six fluorine groups and the methylimidazolium group
(Figure 3). Indeed, 1À ImÀ F6À Fe displays a catalytic potential
(90 mV) more positive than that of its nonfluorinated 1À ImÀ Fe
analogue and thus 1À ImÀ F6À Fe operates with a catalytic potential
closer to that of FeTPPÀ F8 [E0cat(1À ImÀ F6À Fe)=À 1.266 V and
E0cat(FeTPPÀ F8)=À 1.296 V]. However, thanks to the through-space
stabilization of the FeÀ CO2 intermediate by the methylimidazolium
group, 1À ImÀ F6À Fe displays a much higher TOF value (Figure 3c).
The 1À ImÀ F6À Fe catalytic performance falls in between the
through-structure and through-space linear correlations, demon-
strating that extra level of tunability of the catalytic activity of iron
kcat,H2O/kcat,D2O) was measured for each of the imidazolium
derivatives using H2O or D2O as a proton or deuterium source.
Both 2À ImÀ Fe and 1À ImÀ Fe showed a normal primary KIE of
1.12 and 1.02, respectively, but interestingly, the 4À ImÀ Fe
showed a strong inverse primary KIE of 0.43 (further evidence
pointed by Figures S20–22). The latter effect possibly indicates
that for the 4À ImÀ Fe derivative, a pre-equilibrium (initial
substrate stabilization), preceding the rate-determining proto-
nation step, has a significant contribution to the overall kinetics
of the reaction. Compared to 2À ImÀ Fe and 1À ImÀ Fe derivatives,
the 4À ImÀ Fe catalyst has a unique configuration where the four
imidazolium pillars form a positively charged channel above the
metal center. This configuration can induce a H2O/D2O molec-
ular nanoconfinement, which could possibly explain the
peculiar behavior of the 4À ImÀ Fe catalyst. Recently, such H2O
molecular nanoconfinement are being actively investigated in
various ionic liquids including those based on imidazolium
groups.[19] The nonconforming trend in CO2 binding constant
and KIE with respect to the number of imidazolium groups in
the second coordination sphere of the iron porphyrins indicates
that behind the observed remarkable correlation between the
overall driving force and the catalytic rate may also lie
significant changes in the rate-determining step. Initial exper-
imental data presented here dictates that there is a switch of
rate-determining step when changing from 4À ImÀ Fe to
Figure 3. CVs of 1 mm of modified iron porphyrin complexes 1À ImÀ Fe (red), 1À ImÀ F6À Fe (orange), in comparison with FeTPPÀ F8 (dark gray) in DMF containing
°
0.1 m [Bu4N]PF6 at 25 C (a) under argon and (b) under CO2 with 5.5 m H2O. (c) Comparison of the TOF calculated from FOW analysis as a function of catalytic
overpotentials (E0CO2/CO =À 0.74 V vs. NHE). Catalytic performance of nonfunctionalized FeTPP (black) and FeTPPÀ F8 (dark gray) are included for comparison. (d)
Comparison of the CO2 binding constant and KIE between 1À ImÀ Fe and 1À ImÀ F6À Fe.
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1312
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
porphyrins can be achieved by combining through-structure
inductive effects and through-space electrostatic interactions.
We previously reported that despite the difference in their
E(FeI/0) values (see Table 1), complexes FeTPP, FeTPPÀ F8, and
FeTPPÀ F20 display similar CO2 binding constants (KCO2 =0.9 mÀ 1)
indicating that the nucleophilic strength of Fe0 species in these
three complexes has no significant effect on CO2 binding. As
mentioned earlier, the presence of an imidazolium arm in
1À ImÀ Fe increases the CO2 binding constant (KCO2 =2.1 mÀ 1) but
here too, the introduction of six fluorine atoms as electron-
withdrawing groups (1À ImÀ F6À Fe), although it shifts E(FeI/0)
anodically by 90 mV, does not affect the CO2 binding constant
(KCO2 =2.1 mÀ 1) (Figure 3d). In addition, the fluorinated-deriva-
tive 1À ImÀ F6À Fe has a higher normal primary KIE value of 2.71
compared to either 1À ImÀ Fe (1.02) or FeTPPÀ F8 (1.27) indicating
substituted catalyst. Of particular interest, we observed that the
trend in the electrocatalytic performance of through-space
electrostatic interaction models outperforms the classic elec-
tronic effects strategy, as summarized by the catalytic Tafel
plots in Figure 4. Furthermore, when both electrostatic and
electronic features of one imidazolium unit and six fluorine
atoms, respectively, are embarked in a model, here too we
noticed a combined effect resulting in a three order magnitude
gain in the turnover frequency with a consequential fall in the
overpotential. Further work is in progress in our labs to provide
both theoretical insights and the best combination of chemical
functions in the quest for cost-efficient catalytic systems for CO2
reduction.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
that the combination of electron-withdrawing fluorine groups Experimental Section
and cationic methylimidazolium groups induces a more signifi-
Syntheses of precursor ligands (compounds 1–9) and further details
on experimental techniques and characterization are reported in
the Supporting Information.
cant dependence of the overall rate of the reaction on the
protonation step. The notable increase in the TOF of the
1À ImÀ F6À Fe, despite the dragging effects of fluorine groups on
the reaction rate, but without reaching the newly established
through-space linear correlation with imidazolium groups
points out that a different contribution to the reaction rate
from the CO2 binding and the protonation steps is at play here
and not just the result of summation of each of the
independent effects. This will be further probed in additional
experimental and theoretical investigations currently being
performed in our group.
Synthesis of 2À ImÀ Fe
Compound 5 (43 mg, 43 μmol) was dissolved in 5 mL of dry DMF
under argon atmosphere, then 1-methylimidazol was added (7 μL,
86 μmol) to the reaction mixture at room temperature. After 24 h
°
stirring at 100 C, the reaction mixture was brought to room
temperature, then approximately 100 mL of diethylether was
added. The obtained purple precipitate was filtered and washed
with 2×25 mL of diethylether to give 2À ImÀ Fe (47 mg, yield 91%).
UV/Vis (CH2Cl2): λmax (ɛ)=693 (3400), 588 (4000), 518 (11900), 421
(89500 molÀ 1 dm3 cmÀ 1); HRMS (ESI): m/z calcd for C68H52FeN10O2:
365.4537 [M]3+; found: 365.4534.
Conclusion
The through-structure electronic effects of donor and with-
drawing groups have commonly been plugged in molecular
based complexes to modulate the electronic and catalytic
properties of metal complexes. In the field of CO2 reduction, the
presence of electron-withdrawing groups led to a bargain
between a consequent drop in the overpotential and the
diminished electrocatalytic rate. With the mindset to improve
the electrocatalytic properties of molecular catalysts, chemists
have successfully implemented different functionalities such as
hydrogen bonding groups, proton relays, and electrostatic
interactions in the second coordination sphere that have
contributed to modulate both electrochemical parameters. As a
state of the art, two iron porphyrin derivatives holding cationic
tetra-methylammonium and imidazolium functions stand out in
the targeted window for catalysts with low overpotential and
high turnover frequencies. Such an approach was inspired from
the electrostatic interactions that prevail in the functioning of
enzymes. We have reported herein a series of substituted iron
porphyrin systems holding different number of imidazolium
units, the tetra-, di-, and mono-substituted derivatives with the
target to bring insights in the structure/reactivity pattern. We
found that the drop in overpotential of these catalysts is a
function of the number of embarked imidazolium units.
Importantly, we evidenced a gain of six orders of magnitude for
the turnover frequencies from the tetra- to the mono-
Figure 4. Catalytic Tafel plots of the imidazolium-decorated iron porphyrin
catalysts (4À ImÀ Fe in green, 2À ImÀ Fe in blue, 1À ImÀ Fe in red, 1À ImÀ F6À Fe
in orange) showing the electrocatalytic improvements (enhancement of TOF
and decrease of overpotential) in comparison with the classical through-
structure strategy (FeTPP in black, FeTPPÀ F8 in dark gray, FeTPPÀ F20 in
gray). Condition: CO2-saturated DMF with 5.5 m H2O. The arrow represents
the upper left shift (higher TON and lower overpotential) of the Tafel plots
when going from through-structure to through-space strategy.
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1313
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
molecular sieve porous layer open tubular (PLOT) column having
an internal diameter of 0.53 mm, helium carrier gas, and a thermal
conductivity detector (TCD). Peak separation between H2, CO, and
Synthesis of 1À ImÀ Fe
Compound 9 (212 mg, 0.23 mmol) was dissolved in 5 mL of dry
DMF under argon atmosphere, then 1-methylimidazol (19 μL,
0.23 mmol) was added to the reaction mixture at room temper-
1
2
3
4
5
6
7
8
9
CO2 were achieved by programming the oven temperature from
À 1
°
°
an initial 40 C for 2 min, ramped by 10 C min until a final
°
ature. After 24 h stirring at 100 C, the reaction mixture was
°
temperature of 100 C held for 2 min. H2 is detected at 2.16 min,
brought to room temperature, then approximately 200 mL of
diethylether was added. The obtained purple precipitate was
filtered and washed with 2×50 mL of diethylether to give
1À ImÀ Fe (200 mg, yield 82%). UV/Vis (CH2Cl2): λmax (ɛ)=586
(4600), 516 (14500), 418 (115300 molÀ 1 dm3 cmÀ 1); HRMS (ESI): m/z
calcd for C56H40FeN7O: 441.1317 [M]2+; found: 441.1294.
CO is detected at 2.88 min, and CO2 is detected at 6.10 min. A
splitless injector line was utilized to maximize the injection
volume and become sensitive to possible trace amounts. A
calibration curve relating peak area and concentration of CO (or
H2) was established by injecting known amounts of pure CO (or
H2) into the experimental set-up (electrolysis cell) and after 10 min
equilibration time, 50 μL gas aliquots were drawn from the
headspace and injected into the GC. This calibration method
accounts for the headspace volume of the set-up and any CO
dissolved in the solution and as such, it can be used to determine
actual amounts of CO (or H2) produced.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Synthesis of 1À ImÀ F6À Fe
Compound 9’ (50 mg, 511 μmol) was dissolved in 5 mL of dry DMF
under argon atmosphere, then 1-methylimidazol (4 μL, 511 μmol)
was added to the reaction mixture at room temperature. After
°
24 h stirring at 100 C, the reaction mixture was brought to room
temperature, then approximately 100 mL of diethylether was
added. The obtained purple precipitate was filtered and washed
with 2×25 mL of diethylether to give 1À ImÀ F6À Fe (48 mg, yield
89%). UV/Vis (CH2Cl2): λmax (ɛ)=645 (7800), 582 (8700), 510
(21500), 416 (166400 molÀ 1 dm3 cmÀ 1); HRMS (ESI): m/z calcd for
C56H34F6FeN7O: 495.1034 [M]2+; found: 495.1022.
Acknowledgements
This work has been supported by the French National Research
Agency (ANR-19-CE05-0020-02, LOCO). We thank CNRS, CEA
Saclay, LABEX CHARMMMAT, ICMMO and University Paris-Saclay
for the financial support and Ecole Doctorale 2MIB for PhD
fellowship (for A. Khadhraoui). We also thank the synthesis
support facility (Pôle Synthèse) at ICMMO for their help with the
synthesis of the starting porphyrin platform.
Electrochemical analysis
CV measurements were performed in an electrochemical cell
composed of a glassy carbon (3 mm diameter) working electrode,
a platinum wire as a counter electrode and an aqueous saturated
calomel electrode (SCE) as the reference electrode. All experi-
ments were carried out under argon or carbon dioxide atmos-
Conflict of Interest
°
phere at 25 C. DMF was used as solvent and solutions of catalysts
were prepared at a concentration of 1 or 0.5 mm. Tertbutylammo-
nium hexafluorophosphate ([Bu4N]PF6) was used as supporting
electrolyte and its concentration was maintained at hundred-fold
excess compared to the sample. The solutions were all purged
with inert argon gas. AUTOLAB PGSTAT320 was utilized to control
the applied voltages and to measure resulting current and the
scan rate was chosen at 100 mVsÀ 1 (unless stated otherwise).
Ohmic drop was compensated using the positive feedback
compensation implemented in the instrument. Estimations of the
kinetic parameters derived from the CVs are detailed in the
Supporting Information.
The authors declare no conflict of interest.
Keywords: carbon dioxide
· electrocatalysis · electrostatic
interactions · iron porphyrin · second coordination sphere
[1] D. A. N. Ussiri, R. Lal, in Carbon Sequestration for Climate Change
Mitigation and Adaptation, Springer, Cham, 2017, pp. 497–545.
[2] a) T. P. Senftle, E. A. Carter, Acc. Chem. Res. 2017, 50, 472–475; b) M.
Aresta, A. Dibenedetto, A. Angelini, Chem. Rev. 2014, 114, 1709–1742;
c) Q. Liu, L. Wu, R. Jackstell, M. Beller, Nat. Commun. 2015, 6, 5933;
d) S. C. Roy, O. K. Varghese, M. Paulose, C. A. Grimes, ACS Nano 2010, 4,
1259–1278; e) G. A. Olah, Angew. Chem. Int. Ed. 2005, 44, 2636–2639;
Angew. Chem. 2005, 117, 2692–2696; f) Q.-L. Zhou, Angew. Chem. Int. Ed.
2016, 55, 5352–5353; Angew. Chem. 2016, 128, 5438–5439; g) O. S.
Bushuyev, P. De Luna, C. T. Dinh, L. Tao, G. Saur, J. van de Lagemaat,
S. O. Kelley, E. H. Sargent, Joule 2018, 2, 825–832.
[3] a) R. Francke, B. Schille, M. Roemelt, Chem. Rev. 2018, 118, 4631–4701;
b) F. Wang, ChemSusChem 2017, 10, 4393–4402; c) H. Takeda, C.
Cometto, O. Ishitani, M. Robert, ACS Catal. 2017, 7, 70–88; d) M. Cokoja,
C. Bruckmeier, B. Rieger, W. A. Herrmann, F. E. Kühn, Angew. Chem. Int.
Ed. 2011, 50, 8510–8537; Angew. Chem. 2011, 123, 8662–8690; e) N.
Elgrishi, M. B. Chambers, X. Wang, M. Fontecave, Chem. Soc. Rev. 2017,
46, 761–796; f) K. E. Dalle, J. Warnan, J. J. Leung, B. Reuillard, I. S. Karmel,
E. Reisner, Chem. Rev. 2019, 119, 2752–2875.
[4] a) T. Katsuhiro, H. Kazuya, S. Hideo, T. Shinobu, Chem. Lett. 1979, 8, 305–
308; b) M. Hammouche, D. Lexa, M. Momenteau, J. M. Saveant, J. Am.
Chem. Soc. 1991, 113, 8455–8466; c) C. Costentin, J.-M. Savéant, Nat.
Chem. Rev. 2017, 1, 0087; d) A. W. Nichols, C. W. Machan, Front. Chem.
2019, 7; e) F. Franco, S. Fernández, J. Lloret-Fillol, Curr. Opin. Electro-
chem. 2019, 15, 109–117; f) P. Gotico, Z. Halime, A. Aukauloo, Dalton
Trans. 2020, 49, 2381–2396; g) Y. Matsubara, ACS Energy Lett. 2019,
Bulk electrolysis and product characterization
Bulk electrolysis was performed in CO2-saturated DMF containing
5.55 and 0.1 m [Bu4N]PF6 as supporting electrolyte. A gas-tight
two-compartment cell was used for this experiment, each cell
having a volume of 43 mL. The first compartment was filled with
20 mL electrolyte solution (DMF/H2O) containing 0.5 mm catalyst
and the working electrode (glassy carbon rod, effective surface
area 1.41 cm2) and reference electrode (Ag/AgNO3), while the
second compartment was filled with 20 mL electrolyte solution
(DMF) and contained the counter electrode (10 cm×3 cm titanium
grid with nominal spacing of 0.19 mm and wire diameter of
0.23 mm). Ferrocene was used as
comparison in all experiments and then potentials were converted
to NHE.
a reference for standard
Products analysis during the chronoamperometric experiment was
performed by withdrawing 50 μL gas aliquots from the headspace
of the first compartment cell with a gas-tight syringe and injected
into a GC (TraceGC Ultra, ThermoScientific) equipped with a 30 m
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1314
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cssc.202002718
ChemSusChem
1999–2004; h) S. Bhunia, A. Rana, P. Roy, D. J. Martin, M. L. Pegis, B. Roy,
A. Dey, J. Am. Chem. Soc. 2018, 140, 9444–9457.
[5] I. Bhugun, D. Lexa, J.-M. Savéant, J. Phys. Chem. 1996, 100, 19981–
Angew. Chem. 2007, 119, 5212–5216; d) I. Hijazi, T. Roisnel, P. Even-
Hernandez, E. Furet, J.-F. Halet, O. Cador, B. Boitrel, J. Am. Chem. Soc.
2010, 132, 10652–10653.
1
2
3
4
5
6
7
8
9
19985.
[14] a) C. Römelt, J. Song, M. Tarrago, J. A. Rees, M. van Gastel, T.
Weyhermüller, S. DeBeer, E. Bill, F. Neese, S. Ye, Inorg. Chem. 2017, 56,
4745–4750; b) C. Römelt, S. Ye, E. Bill, T. Weyhermüller, M. van Gastel, F.
Neese, Inorg. Chem. 2018, 57, 2141–2148.
[15] a) C. Costentin, M. Robert, J.-M. Savéant, Acc. Chem. Res. 2015, 48, 2996–
3006; b) V. Artero, J.-M. Saveant, Energy Environ. Sci. 2014, 7, 3808–3814;
c) C. Costentin, J.-M. Savéant, J. Am. Chem. Soc. 2017, 139, 8245–8250.
[16] a) M. L. Pegis, B. A. McKeown, N. Kumar, K. Lang, D. J. Wasylenko, X. P.
Zhang, S. Raugei, J. M. Mayer, ACS Cent. Sci. 2016, 2, 850–856; b) M. L.
Pegis, C. F. Wise, B. Koronkiewicz, J. M. Mayer, J. Am. Chem. Soc. 2017,
139, 11000–11003; c) M. L. Pegis, C. F. Wise, D. J. Martin, J. M. Mayer,
Chem. Rev. 2018, 118, 2340–2391.
[17] C. Costentin, J.-M. Savéant, J. Am. Chem. Soc. 2018, 140, 16669–16675.
[18] C. Costentin, S. Drouet, G. Passard, M. Robert, J.-M. Savéant, J. Am.
Chem. Soc. 2013, 135, 9023–9031.
[19] a) K. Saihara, Y. Yoshimura, S. Ohta, A. Shimizu, Sci. Rep. 2015, 5, 10619;
b) A. Verma, J. P. Stoppelman, J. G. McDaniel, Int. J. Mol. Sci. 2020, 21,
403; c) C. Ma, A. Laaksonen, C. Liu, X. Lu, X. Ji, Chem. Soc. Rev. 2018, 47,
8685–8720.
[6] I. Azcarate, C. Costentin, M. Robert, J.-M. Savéant, J. Am. Chem. Soc.
2016, 138, 16639–16644.
[7] a) I. Azcarate, C. Costentin, M. Robert, J.-M. Savéant, J. Phys. Chem. C
2016, 120, 28951–28960; b) C. Costentin, J.-M. Savéant, Nat. Chem. Rev.
2017, 1, 0087.
[8] a) C. Costentin, S. Drouet, M. Robert, J.-M. Savéant, Science 2012, 338,
90–94; b) C. Costentin, G. Passard, M. Robert, J.-M. Savéant, J. Am. Chem.
Soc. 2014, 136, 11821–11829; c) E. Nichols, J. S. Derrick, S. K. Nistanaki,
P. T. Smith, C. J. Chang, Chem. Sci. 2018, 9, 2952–2960; d) P. Sen, B.
Mondal, D. Saha, A. Rana, A. Dey, Dalton Trans. 2019, 48, 5965–5977;
e) P. Gotico, B. Boitrel, R. Guillot, M. Sircoglou, A. Quaranta, Z. Halime,
W. Leibl, A. Aukauloo, Angew. Chem. Int. Ed. 2019, 58, 4504–4509;
Angew. Chem. 2019, 131, 4552–4557.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[9] A. Khadhraoui, P. Gotico, B. Boitrel, W. Leibl, Z. Halime, A. Aukauloo,
Chem. Commun. 2018, 54, 11630–11633.
[10] B. Mondal, A. Rana, P. Sen, A. Dey, J. Am. Chem. Soc. 2015, 137, 11214–
11217.
[11] a) A. Warshel, Acc. Chem. Res. 1981, 14, 284–290; b) P. K. Sharma, Z. T.
Chu, M. H. M. Olsson, A. Warshel, Proc. Nat. Acad. Sci. 2007, 104, 9661–
9666; c) A. Warshel, Proc. Nat. Acad. Sci. 1978, 75, 5250–5254.
[12] a) C.-H. Lee, J. S. Lindsey, Tetrahedron 1994, 50, 11427–11440; b) J. S.
Lindsey, I. C. Schreiman, H. C. Hsu, P. C. Kearney, A. M. Marguerettaz, J.
Org. Chem. 1987, 52, 827–836.
[13] a) J. P. Collman, R. R. Gagne, C. Reed, T. R. Halbert, G. Lang, W. T.
Robinson, J. Am. Chem. Soc. 1975, 97, 1427–1439; b) A. Didier, L.
Michaudet, D. Ricard, V. Baveux-Chambenoît, P. Richard, B. Boitrel, Eur. J.
Org. Chem. 2001, 2001, 1927–1926; c) Z. Halime, M. Lachkar, T. Roisnel,
E. Furet, J. F. Halet, B. Boitrel, Angew. Chem. Int. Ed. 2007, 46, 5120–5124;
[20] B. Zhao, H. Lei, N. Wang, G. Xu, W. Zhang, R. Cao, Chem. Eur. J. 2020, 26,
4007–4012.
Manuscript received: November 24, 2020
Revised manuscript received: December 24, 2020
Accepted manuscript online: January 2, 2021
Version of record online: January 21, 2021
ChemSusChem 2021, 14, 1308–1315
www.chemsuschem.org
1315
© 2021 Wiley-VCH GmbH