Local Proton Source in Electrocatalytic CO2 Reduction with [Mn(bpy–R)(CO)3Br] Complexes

A Journal of

Accepted Article

Title: Local Proton Source in the Electrocatalytic CO2 Reduction by

Mn(bpy-R)(CO)3Br Complexes

Authors: Carlo Nervi, Federico Franco, Claudio Cometto, Luca Nencini,

Claudia Barolo, Fabrizio Sordello, Claudio Minero, Jan

Fiedler, Marc Robert, and Roberto Gobetto

This manuscript has been accepted after peer review and appears as an

Accepted Article online prior to editing, proofing, and formal publication

of the final Version of Record (VoR). This work is currently citable by

using the Digital Object Identifier (DOI) given below. The VoR will be

published online in Early View as soon as possible and may be different

to this Accepted Article as a result of editing. Readers should obtain

the VoR from the journal website shown below when it is published

to ensure accuracy of information. The authors are responsible for the

content of this Accepted Article.

To be cited as: Chem. Eur. J. 10.1002/chem.201605546

Link to VoR: http://dx.doi.org/10.1002/chem.201605546

Supported by

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

Local Proton Source in the Electrocatalytic CO₂Reduction by

Mn(bpy-R)(CO)₃Br Complexes

Federico Franco^[a]$, Claudio Cometto^[a,c]$, Luca Nencini^[a]$, Claudia Barolo^[a], Fabrizio Sordello^[a],

Claudio Minero^[a], Jan Fiedler^[b], Marc Robert^[c], Roberto Gobetto*^[a]and Carlo Nervi*^[a]

Abstract: The electrochemical behavior of fac-[Mn(pdbpy)(CO)₃Br]

(pdbpy 4-phenyl-6-(phenyl-2,6-diol)-2,2'-bipyridine), 1, in

fac-[M(bpy-R)(CO)₃X] (M = Mn^I, Re^I; bpy-R = 2,2’-bipyridine-

based ligands; X = Cl^–or Br^–) complexes have been widely

studied as precursors to efficient electrocatalysts for the

reduction of CO₂to CO. In particular, Mn complexes bearing

4,4’-disubstituted bipyridines with H, CH₃and t-Bu (since all

complexes herein mentioned are fac- we will omit this label) are

valid low-cost alternatives to the Re^Icounterparts in terms of

stability, selectivity and efficiency.^[8]The combination of IR/UV-

Vis spectroelectrochemistry (SEC)^[9]with non-conventional

spectroscopic techniques^[10-12]and computational methods,^[13]

provided valuable complementary tools to electrochemistry for

elucidating the electrocatalytic mechanism of CO₂reduction by

Group VII metal based catalysts. These studies suggest that

doubly reduced pentacoordinated species [M(bpy-R)(CO)₃]^–(M

= Mn, Re) are responsible for the catalytic conversion of CO₂

into CO. Nevertheless, there are several differences in the

electrocatalytic behavior of the two classes of complexes. The

Mn catalysts are formed upon the reductive cleavage of the

[Mn(bpy-R)(CO)₃]₂dimer, which is generated after the 1e

reduction of the starting species and subsequent rapid

dissociation of the halogen ion. Contrariwise, the structurally

similar [Re(bpy-R)(CO)₃]₂dimer is known to be a side-product in

the electrocatalytic cycle of Re complexes.^[14]In an effort to

reduce the overpotential required for the formation of the

catalytically active species, Kubiak and coworkers have recently

exploited the bulky nature of the mesbpy ligand (6,6′-dimesityl-

2,2′-bipyridine) to prevent dimerization in [Mn(mesbpy)-

(CO)₃Br].^[15]However, the [Mn(mesbpy)(CO)₃(COOH)] adduct

requires a third “extra-electron” and catalysis occurs more

negatively even though the [Mn(mesbpy)(CO)₃]^–anion is formed

via a 2e transfer (ECE mechanism) at less negative potentials

than in the case of bpy. Moreover, unlike the polypyridyl Re

catalysts, Mn-based catalytic CO₂conversion is commonly

observed only in the presence of an added proton source (water,

methanol or TFE).^[8,15]Nevertheless, we have recently reported

the first case of a bromotricarbonyl Mn^Icatalyst, namely [Mn-

(pdbpy)(CO)₃Br] (1) that is capable of reducing CO₂even in

anhydrous acetonitrile,^[16]without the need for deliberate addition

of Brønsted acids.^[17]This unique behavior has been ascribed to

the structure of the pdbpy ligand, in which the pendant phenolic

groups near to the metal center may act as intramolecular

proton sources (Figure 1). We found that the presence of local

protons not only provided a dramatic enhancement in the

electrocatalytic activity towards CO₂reduction, but also an

unexpected change in selectivity, producing a non-negligible

amount of formate in addition to CO. Two distinct pathways have

therefore been proposed along with the formation of an

electroinduced Mn hydride species, which was assumed to be

responsible for HCOO^–production, as reported in recent

photocatalytic studies on [Mn(bpy)(CO)₃Br].^[18,19]

=

acetonitrile under Ar and its catalytic performances for CO₂reduction

with added water, 2,2’,2’’-trifluoroethanol (TFE) and phenol are

discussed in detail. Preparative-scale electrolysis experiments,

carried out at –1.5 V vs. SCE in CO₂-saturated acetonitrile solutions,

reveal that the process selectivity is extremely sensitive to the acid

strength, providing CO and formate in different faradaic yields. A

detailed spectroelectrochemical (IR and UV-Vis) study under Ar and

CO₂atmospheres shows that 1 undergoes fast solvolysis; however

dimer formation in acetonitrile is suppressed, providing an atypical

reduction mechanism in comparison with other reported Mn^I

catalysts. Spectroscopic evidence of Mn hydride formation supports

the existence of different electrocatalytic CO₂reduction pathways.

Furthermore, a comparative investigation performed on the new fac-

[Mn(ptbpy)(CO)₃Br] (ptbpy

=

4-phenyl-6-(phenyl-3,4,5-triol)-2,2'-

bipyridine) catalyst, 2, bearing a bipyridyl derivative with OH groups

in different positions to those in 1, provides complementary

information about the role that the local proton source plays during

the electrochemical reduction of CO₂.

Introduction

The electrochemical conversion of carbon dioxide into value-

added chemicals mediated by transition metal complexes has

attracted growing interest in recent years.^[1-2]The rational design

of molecular catalysts that possess low overpotentials for the

selective conversion of CO₂is a crucial target, while knowledge

of the main factors affecting this process will provide the basis

for future improvements and advances in this field. Many

organometallic catalysts have been found to selectively reduce

CO₂into such two-electron products as CO and HCOOH in non-

aqueous systems. The former is widely used in mixtures with H₂

(syngas) in large-scale industrial processes.^[3]Liquid solutions of

formic acid can instead be directly used in green technologies

(fuel cells),^[4]or as hydrogen storage materials.^[5]However, most

of the reported homogeneous catalysts for CO₂electrochemical

reduction are selective towards CO formation,^[2,6]whereas only a

few of them give formate in high yields.^[7]

[a]

Dr. F. Franco, C. Cometto, L. Nencini, Prof. Dr. C.Barolo, Dr. F.

Sordello, Prof. Dr. C. Minero, Prof. Dr. C. Nervi and Prof. Dr. R.

Gobetto.

University of Turin, Department of Chemistry and NIS, Via P. Giuria

7, 10125 Turin, Italy, E-mail: carlo.nervi@unito.it

These authors equally contributed.

$

[b]

Ing. CSc. J.Fiedler,

Heyrovský Institute of Physical Chemistry of ASCR, v.v.i.,

Dolejškova 3, 18223 Prague, Czech Republic

Prof. Dr. M. Robert., C. Cometto

[c]

Encouraged by the remarkable electrocatalytic activity of 1

under CO₂, we firstly aimed to in-depth investigate its unusual

electrochemical behavior under inert atmosphere by using cyclic

voltammetry (CV) and IR/UV-Vis SEC.

Univ. Paris Diderot, Sorbonne Paris Cité, UMR CNRS 7591,

Laboratoire Electrochimie Moléculaire, F-75205 Paris 13, France

Supporting information for this article is given via a link at the end of

the document.

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

Ph

experiments under Ar performed just after R1 (–1.25 V vs. SCE)

consumed one electron per molecule of 1. More information

about the chemical processes involved in the course of R1 will

be given below.

Ph

Br

HO

N

I

N

I

N

Mn

OC

CO

OC

CO

OH

CO

OH

Mn^I

Br

Mn^I

Br

(ptbpy)(CO)₃

(pdbpy)(CO)₃

2

( )

1

( )

Figure 1. Chemical sketches of the complex 1 and 2.

R1

R2

R3

Furthermore, the effect due to the addition of different proton

sources (H₂O, TFE and phenol) on either the electrocatalytic

activity and selectivity of 1 under CO₂was also evaluated.

Finally,

a comparative study on the structurally similar

[Mn(ptbpy)(CO)₃Br] (2), containing three local OH groups in

meta and para positions of the phenolic ring in 6, i.e. a little

farther from the Mn center than in 1 (Figure 1), provided further

useful information to better understand the role of an

intramolecular proton source on the electrochemical and

electrocatalytic properties of this class of catalysts.

Figure 2. CVs of a MeCN solution of 1 (0.5 mM) at 100 mV s^-1under Ar (red

curve) and CO₂(blue curve), at a glassy carbon electrode. The black line is

the CO₂-saturated blank.

Conversely, CV of 2 in MeCN under Ar significantly differs from

Results and Discussion

that of

1 (Figure 3), resembling that of the well-known

[Mn(bpy)(CO)₃Br].^[8]It exhibits a first irreversible reduction, R1’,

at E_p= -1.30 V, followed by fast Br^–dissociation and the

subsequent formation of the Mn-Mn dimer, which is then

reoxidized at -0.49 V (O1’, Figure 3). As evidenced by the IR-

SEC data (see below) the replacement of Br^–by MeCN at open

circuit is much slower in 2.

Synthesis and characterization. The pdbpy and ptbpy ligands

were synthesized using the Kröhnke reaction and by coupling

the corresponding pyridinium iodides and chalcones according

to the reported procedure.^[17]A demethylation reaction, based on

refluxing the precursor in a CH₃COOH solution of HBr (33%),

was used to synthesize the final ptbpy ligand. Full details can be

found in the Experimental Section.

Cyclic voltammetry (CV) under Ar. CV of complex 1 in MeCN

under inert atmosphere^[17]exhibits three consecutive reduction

waves (Figures 2 and S1). The first two reduction processes, R1

and R2, are chemically irreversible (E_p= –1.21 V, –1.50 V vs.

SCE), whereas the third reduction R3 appears to be quasi-

reversible (E_1/2= –1.66 V). R2 and R3 current peaks decrease

with respect to R1 (Figure S2) upon increasing the scan rate.

O1’

This is consistent with

processes following R1 generate the species undergoing the

reductions R2 and R3.

a mechanism in which chemical

R2’

R1’

As we will discuss later, the SEC experiments performed under

Ar highlighted two key aspects that strongly influence the

electrochemical behavior of 1 in MeCN. First of all, rapid

solvolysis of the Mn-Br bond even occurs at open circuit in the

dark at room temperature. This behavior has been reported for

similar complexes.^[20-21]Solutions of 1 in MeCN are thus

mixtures of 1 and the positively charged [Mn(pdbpy)(CO)₃-

Figure 3. CVs of a MeCN solution of 2 (0.5 mM) at 100 mV s^-1under Ar (red

curve) and CO₂(blue curve), at a glassy carbon electrode. The black line is

the CO₂-saturated blank.

(MeCN)]⁺(1a). Moreover, the dimerization of

1 was not

The second irreversible reduction of 2, R2’, occurs more

negatively, at E_p= -1.64 V. The relationship between the peak

current of R2’ and the CV scan rate is similar to findings for

waves R2 and R3 in 1 (Figure S3). In a 1 mM solution the

apparent single R2’ peak splits into two almost overlapped

spectroscopically detected by IR and UV-Vis SEC upon R1 (see

below). This is in agreement with CV, where no anodic

reoxidation peak of the dimer, commonly found at around –0.3

V,^[8]is observed (Figure S1). Thus, the first reduction is

chemically irreversible, but no dimer is formed. Exhaustive

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

peaks that are ca. 70 mV separated from each other (E_p= -1.61

and -1.68 V, respectively) (Figure S4). The Mn catalyst

containing the 6-(2-hydroxyphenyl)-2,2′-bipyridyl ligand showed

an analogous CV response in MeCN + 5% H₂O, but the nature

of the two reductions (at about -1.30 V in that case) was not

explained.^[22]They might be related to the presence of partially

deprotonated intermediate species in solution which may be

formed in the course of the reductive scan on the CV timescale.

Further discussion of this point is given below.

and the MeCN-containing cation 1a (2044, 1961 cm^–1), in a time-

dependent ratio due to the proceeding of the solvolysis

equilibrium (Figure S5a). The two low-energy ν_CObands of 1

overlap with those of 1a. In a typical run, the solvent complex,

1a, was found to be the major species at the beginning of the IR-

SEC experiments (Figure S5b). Weaker bands at 1613 and

1623 cm^–1are assigned to the plane stretching of the aromatic

rings (bpy and Ph modes are mixed) of 1 and 1a, respectively.

When a potential slightly more negative than R1 is applied, both

the series of bands corresponding to the starting forms 1 (2026,

1935, 1925(sh), 1613 cm^-1) and 1a (2044, 1961, 1623 cm^-1)

progressively disappear, leading to growth of a new species as

major reduction product with absorptions at 2021, 1920(br) (CO

stretches) and 1607 cm^-1(ligand-based stretch mostly

overlapped with those belonging to 1) (Figure 4).

Infrared Spectroelectrochemistry (IR-SEC) under Ar and

DFT calculations of ν_COstretches.

IR-SEC of 1 in MeCN/TBAPF₆. IR-SEC experiments were

performed in an OTTLE cell in order to get further insights about

the reduction paths of 1 and 2. This technique has already been

used to characterize key intermediates during the electroche-

mical reduction of carbonyl organometallic electrocatalysts by

monitoring the spectral changes of the CO stretching bands,

[9]

ν_CO

.

DFT calculations were employed as an additional tool to

support the assignment of the intermediate species observed

during the IR-SEC experiments.

Table 1. Selected experimental and calculated ν_COin MeCN.

Complex^]

Experimental

DFT^]

[Mn(pdbpy)(CO)₃Br] (1)

2026, 1935, 1925(sh) 2013, 1939, 1928

[Mn(pdbpy)(CO)₃(MeCN)]⁺(1a) 2044, 1961

[Mn(pdbpy)(CO)₃Br]^–

2033, (1962, 1951)

1984, 1911, 1900

2014, 1938, 1926

2010, (1925, 1916)

[Mn(pdbpy)(CO)₃(MeCN)]

[Mn(pdbpy−H⁺)(CO)₃] (1b)

[HMn(pdbpy−H⁺)(CO)₃]^–(1c)

[HMn(pdbpy)(CO)₃] (1c’)

[Mn(pdbpy−H⁺)(CO)₃]²⁻(1d)

[Mn(pdbpy−2H⁺)(CO)₃]^–(1e)

[Mn(pdbpy−2H⁺)(CO)₃]³⁻(1f)

[Mn(ptbpy)(CO)₃Br] (2)

2021, 1920(br)

1987, 1899, 1878(sh) 1972, 1891, 1879

1988, 1908, 1897

Figure 4. IR-SEC of a 1 mM MeCN solution of 1 under Ar: reduction at the first

peak (R1). The inset shows a concomitant decrease of the νOH

1910, 1818(br)

1892, (1832, 1808)

2012, 1916, 1900(sh) 1998, 1912, 1899

1910, 1805^a

1874, (1814, 1785)

2023, 1936, 1914

The latter species can be reasonably assigned to a singly-

deprotonated neutral species, [Mn(dhbpy−H⁺)(CO)₃] (1b),

formed by a reductive deprotonation of the starting 1a upon R1.

Accordingly to this interpretation, the newly formed bands fit well

with the spectroscopic data of Re and Mn complexes that

contain the 4dhbpy ligand (4,4′-dihydroxy-2,2′-bipyridyl) and in

which a reductive deprotonation mechanism was reported.^[21,23]

In particular, a shift in the high-energy ν_COfrom 2019 to 2012

cm^–1(7 cm^–1red shift) was observed in [Re(4dhbpy)(CO)₃Cl] as

the starting species transformed into the singly-deprotonated

complex [Re(4dhbpy−H⁺)(CO)₃Cl]^–(Table S1).^[23]Similarly, a

corresponding 5 cm⁻¹red shift is experimentally observed

passing from 1 (2026 cm^–1) to [Mn(pdbpy−H⁺)(CO)₃] 1b (2021

cm^–1). This interpretation is also supported by the decrease of

the broad OH stretching band at around 3359 cm^–1(calculated

at 3363 cm^–1) during the first reduction (Figure 4, inset). The R1

process is also chemically irreversible in IR-SEC, since

reoxidation does not restore the initial spectrum (Figure S6).

From a structural point of view, DFT calculations suggest that

the phenolate group in 1b may occupy the vacant coordination

position (after the release of MeCN from 1a, Mn-O = 2.052 Å).

[Mn(ptbpy)-(CO)₃(MeCN)]⁺(2a) 2044, 1958

[Mn(ptbpy)(CO)₃]₂(2b)

1932,

1847

1880,

1868,

[HMn(ptbpy–H⁺)(CO)₃]^–(2c)

[Mn(ptbpy)(CO)₃]^–(2d)

[HMn(bpy)(CO)₃]

1987

1912, 1817, 1804^a

1989, 1892 ^[a]

[HMn(bpy)(CO)₃]

[Mn(CO)₃(bpy)]^–

1991, 1892, 1888sh

1916,1814.5 ^[b]

1911, 1811

1978, 1893, 1884

-

1890, (1816, 1812)

[a] THF as solvent, data from ref. ²⁴; [b] THF as solvent, data from ref. ^9a

Tables 1 and S1 summarize the experimental and DFT calcula-

ted infrared data for all the transient species proposed here,

together with those reported for similar Mn complexes in the

literature. In dry acetonitrile and before applying any potentials,

1 shows two different sets of three ν_CO, which are both typical of

facially coordinated [Mn(bpy-R)(CO)₃X] complexes (Figure S5a).

This is a clear indication^[20-21]that the axial Br^–in 1 is replaced by

the solvent even in the dark and at room temperature. Thus, the

IR spectra of MeCN solutions of 1 exhibit mixtures of the original

Br^–containing complex (ν_CObands at 2026, 1935, 1925sh cm^–1)

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

The interaction between the Mn^Icenter and the negatively

charged phenolate unit in the axial position is expected to

stabilize the pseudo-octahedral (distorted) structure of 1b, thus

accounting for the lack of dimerization upon R1. Furthermore,

Mn retains its original +1 oxidation state in 1b, maintaining a

similar electron density on the Mn(CO)₃moiety to that of 1 (as

reflected by a very small shift in ν_CO). DFT ν_COpredictions of 1,

1a and 1b well reproduce the experimental data (Table 1).

The two sets of calculated transitions at 1962/1951 cm^–1and

1925/1916 cm^–1are in good agreement with the broad

experimental absorptions at 1961 and 1920 cm^–1for 1a and 1b,

respectively. Finally, it is worth mentioning that the hypothetical

assignment of 1b to the neutral six-coordinate [Mn(dhbpy−H⁺)-

(CO)₃(CH₃CN)] can be ruled out by the absence of any

observable band in the region between the ν_COstretches of 1

and 1a, respectively, after R1.^[21]

by simultaneous growth of two different sets of ν_COstretches,

indicating the formation of two new species (Figure 5). The ν_CO

stretches at 1916 and 1818 cm^-1, which are in excellent

agreement with several reported doubly-reduced five-coordinate

Mn anions,^[8b,9a,15]can be assigned to the partially deprotonated

[Mn(dhbpy−H⁺)(CO)₃]²⁻anion (1d), which still possesses one

local OH group attached to the ligand moiety. Furthermore,

concomitantly to 1d formation, a 9 cm^–1shift of the IR signal at

2021 cm^–1to lower energy indicates a partial conversion of 1b

into another species, characterized by a growing ν_COstretch at

2012 cm^–1and a poorly resolved shoulder at 1900 cm^–1(Figure

5). The ligand mode at 1607 cm^–1of 1b is also slightly shifted to

1605 cm^–1. These spectroscopic features are consistent with the

doubly-deprotonated intermediate [Mn(dhbpy−2H⁺)(CO)₃]^–(1e),

derived from further reductive deprotonation of 1b and

structurally similar to the latter. This interpretation is in

accordance with further decrease of the ν_OHstretch at 3359 cm^–1,

which is experimentally seen during R2 in the recorded spectra

(Figure 5, inset). Moreover, an analogous red-shift (∼10 cm⁻¹) of

the high energy ν_COstretch was reported for the reductive

conversion of the related complex [Re(4dhbpy−H⁺)(CO)₃Cl]⁻to

the doubly-deprotonated [Re(4dhbpy−2H⁺)(CO)₃]⁻(Table S1).^[23]

The DFT calculated ν_COstretches match well with the

experimental ones, for either 1d (1892, 1832, 1808 cm^–1) and 1e

(1998, 1912, 1899 cm^–1).

Figure 5. IR-SEC of a 1 mM MeCN solution of 1 under Ar: reduction at the

second peak (R2). The inset shows further decrease of ν_OH

.

Actually, less intense ν_COat 1987 and 1878(sh) cm^–1appear

after reduction at R1 in addition to the bands of 1b (Figure 4).

We assign these weak ν_COto the Mn hydride [HMn^I(pdbpy−H⁺)-

(CO)₃]^–1c, formed by intramolecular proton transfer from pdbpy

to the metal (with a negative charge formally localized on the

phenolate). Although rare cases of carbonyl Mn hydrides have

been spectroscopically characterized, similar ν_COwere reported

by Riera et al. for the analogous [HMn^I(bpy)(CO)₃] in THF (1989

and 1892 cm^–1).^[24]We were able to reproduce the literature data

by performing the IR-SEC of [Mn(bpy)(CO)₃Br] in MeCN with 0.1

M of phenol (Figure S7). The reduction of the dimer that was

initially formed (ν_COat 1976, 1963vw, 1933, 1879, 1857 cm^–1),

gave the [Mn(bpy)(CO)₃]^–anion (ν_COat 1911 and 1811 cm^–1),

which is quickly transformed into the hydride species in the

presence of 0.1 M phenol (ν_COat 1991, 1892, 1888sh cm^–1). In

more acidic conditions the production of the dimer is strongly

reduced and the hydride is more directly formed (Figure S7b).

These data are in good agreement with our experimental and

theoretical data on 1c (Table 1).

Figure 4. IR-SEC of a 1 mM MeCN solution of 1 under Ar: reduction at the

third peak (R3).

The ν_COat 2012 and 1900 cm^–1disappear at slightly more

negative potentials, while ν_COcentered at 1910, 1817 and 1805

cm^–1appear (Figure 6). The net result is that the IR spectrum at

the end of the reduction shows a very intense sharp ν_COat 1910

cm^–1and a broader one, split into two maxima at 1817 and 1805

cm^–1, respectively. These spectral changes likely indicate that

the 1d anion coexists with its deprotonated counterpart

[Mn(pdbpy−2H⁺)(CO)₃]³⁻(1f) in solution, resulting from the

reduction of 1e. 1d and 1f share the same five-coordinate

geometry as the [Mn(bpy-R)(CO)₃]^–species, as suggested by

the similarity in their IR features (see Table 1).^[8b,9a,15]It should

When the cell potential is stepped more negative to R2 (~–1.5 V),

decay of both species 1b and 1c produced upon R1 is balanced

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

be also mentioned that direct interconversion between 1d and 1f

via acid-base equilibria in solution is non-negligible due to the

pH-dependent properties of the pdbpy ligand. For this reason, it

is hard to clearly distinguish the differently protonated anionic

forms in the experimental IR spectra.^[21]These chemical

complications, as well as the highly reducing potentials, probably

contribute to the appearance of some minor Mn species in the

final spectrum, characterized by very weak ν_CObands (e.g. at

1982 and 1869 cm^–1, Figure 6).

consistent with a reductive cleavage of the dimer to generate the

five-coordinate [Mn(ptbpy)(CO)₃]^–(2d) anion. The similarity

between the IR spectrum shown in Figure 8 and the one

obtained after the IR-SEC experiment on 1 (Figure 6), prompted

us to assign the increasing bands at 1912, 1817 and 1804 cm^–1

to a mixture of differently protonated anionic species (mostly

derived from acid-base processes).

IR-SEC of 2 in CH₃CN/TBAPF₆. As preliminarily suggested by

CV data, an IR-SEC study of 2 confirmed the remarkable

differences in the reduction mechanism with respect to 1,

highlighting the crucial role played by the local proton source.

The solvolysis equilibrium in 2 is less marked than in 1, so that

the IR spectrum of a MeCN solution of 2 shows three main ν_CO

at 2023, 1936 and 1914 cm^–1, corresponding to the Br-

containing initial form. However, bands at 2044 and 1958 cm^–1

(Figure S8) rapidly grow just before the first reduction,

suggesting that the cationic [Mn(ptbpy)(CO)₃(CH₃CN)]⁺(2a)

complex is rapidly generated by a catalytic ETC (electron

transfer chain) reaction.^[25]

Upon the first reduction R1’, decrease in the ν_COof 2 (2023,

1936, 1914 cm^–1) and 2a (2044, 1958 cm^–1) is balanced by the

growth of ν_COat 1932, 1880, 1868 and 1847 cm^–1, likely

corresponding to the dimer [Mn(ptbpy)(CO)₃]₂(2b) (Figure 7).

Figure 8. IR-SEC of a 1 mM MeCN solution of 2 under Ar: reduction at the

second peak (R2’).

UV-Vis Spectroelectrochemistry (UV-Vis SEC) under Ar. In

the OTTLE cell, at open circuit, the UV-Vis spectrum of 1 in

MeCN shows a series of absorptions at 382, 283(sh), 273 and

250 nm (FigureS9a), which do not undergo any appreciable

changes during the electrochemical reduction R1. This is

consistent with the formation of 1b upon R1, as the theoretical

UV-Vis spectra of 1a and 1b (Figure S10) resulted to be very

similar to each other. Importantly, the absence of any intense

absorption at around 800 nm confirmed the lack of dimerization

after R1, as already evidenced by the electrochemical and IR-

SEC results.^[8a,9a]When the applied potential is close to R2 (ca.

–1.5 V), a broad absorption at 627 nm and other minor bands at

375, 243 nm start to grow to the detriment of the maxima at 283,

273, 250 nm (Figure S9b), indicating the formation of a five-

coordinate anion, analogous to [Mn(bpy)(CO)₃]^–produced upon

2e reduction of [Mn(bpy)(CO)₃Br] in MeCN.^[8a]At slightly more

negative potentials, the band at 243 nm decreases, with an

isosbestic point at 290 nm, while a new, intense band appears at

506 nm and the maximum at 627 nm is slightly blue-shifted to

622 nm (Figure S9c). These spectral changes are in agreement

with the formation of a mixture of analogous pentacoordinated

Mn species after R3.^[8a]

Figure 5. IR-SEC of a 1 mM MeCN solution of 2 under Ar: reduction at the first

peak (R1’).

This interpretation matches not only with the CV data of 2, but

also with the experimental IR-SEC data reported for analogous

systems.^[9a,12]A small peak at 1987 cm^–1also grows during R1’,

indicating the partial formation of a Mn hydride species ([HMn-

(ptbpy−H⁺)(CO)₃]^–, 2c (or protonated counterparts). Consistently

with a less spatial interaction between the intramolecular

phenolic groups and the metal atom in 2, it is apparent that the

hydride 2c is generated in-situ in a considerably smaller amount

than the analogue 1c.

Proposed reduction mechanisms under Ar. The set of

experimental data in MeCN led us to propose two different

reduction routes for 1 (Scheme 1) and 2 (Scheme 2). A rational

understanding of the reduction pathway for 1 should take into

account the solvolysis process which provides an initial mixture

of 1 and 1a. Since it seems reasonable to assume that the

reduction of 1a occurs more positively than 1 and IR-SEC clearly

At increasingly negative potential values, the ν_COof the dimer

decrease, while the very intense ν_COat 1912, 1817 and 1804

cm^–1start to increase (Figure 8). These spectral changes are

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

showed that 1a is thermodynamically favored over 1, we will

consider R1 simply describing the 1e reduction of 1a. As

evidenced by IR-SEC, chemical complications interfere with the

redox process occurring at R1, yielding the neutral 1b species

as the major product via a reductive deprotonation of 1a (after

release of a solvent molecule).

accounting for the overall consumption of one electron per

molecule of 1b produced.

At the onset potential of R2, 1b is reduced into a transient

radical species (not experimentally observed), which undergoes

two competing chemical processes. Firstly, as an effect of

increasing negative charge over the diimine moiety, the axial

phenolate coordination to Mn becomes more labile in the

reduced form of 1b, promoting dissociation of the Mn-O bond.

Upon R2, the resulting species further reduces to produce the

doubly-reduced, singly-deprotonated anion 1d. In parallel, the

aforementioned short-lived radical species may undergo another

reductive deprotonation, providing the doubly-deprotonated

intermediate 1e. The net result is that, as observed by IR-SEC,

1d and 1e coexist at the potential of R2 (Figure 5). The

electrochemical conversion of 1e into 1f is analogue to the

conversion of 1b into 1d. The proposed mechanism is also in

agreement with the voltammetric behavior of 1 at different scan

rates (Figure S2).

+

Ph

OH

Ph

OH

Br

N

+MeCN -Br^-

N

I

Mn

OC

OH

OC

OH

CO

1a

1

e-

-

1

e-

R1

2

-

H

MeCN 1/2

2

-

Ph

The electrochemical behavior of 2 under Ar is straightforward

(Scheme 2), and similar to other [Mn(diimine)(CO)₃X] complexes,

where the formation of the dimer 2b is the only relevant process

experimentally observed upon the first 1e reduction (R1’). The

hydride 2c, albeit detected in IR-SEC, is produced only in very

small amount, if compared with 1c. Unlike 1, CV and IR-SEC

gave no indication of deprotonation during the first reduction of 2.

In the subsequent reduction, R2’, the reductive cleavage of 2b

produces the pentacoordinated anion 2d (or a mixture of

differently protonated anions).

H

O

N

O

OC

N

I

Mn

I

Mn

HO

OC

OH

CO

1c

CO

major

CO

1b

minor

hydride (

)

e-

2

e^-

R2

1

H

-1/2

2

-

Ph

O

Ph

Br

2-

HO

N

I

N

O

OC

N

OC

N

I

Mn

0

Mn

O

OC

CO

-Br^-

CO

OH

2

CO

1e

1d

1e^-

2e^-

R1'

e^-

-Br^-

+H⁺

2

-

-H⁺

R3

Ph

H

HO

OH

CO

O

N

I

3-

OC

0

Mn

CO

N

Ph

Mn

OC

CO

HO

N

1/2

O

CO

N

OH

0

Mn

Ph

2c

minor

hydride (

)

OC

CO

N

0

Mn

O

1f

HO

CO

OC

CO

Scheme 1. Electrochemical mechanism for 1 under Ar.

OH

2b

major

1e^-

R2'

-

As discussed above, reductive deprotonation is not uncommon

in Group VII carbonyl complexes containing pyridinol

moieties.^[21,23]No experimental evidence of dimerization after R1

was obtained. However, a small amount of the hydride 1c was

detected by IR-SEC at this reduction potential. Formally, it

should be produced via an overall two electron reduction, but

electrolysis (at longer timescale than SEC) apparently

consumed only one electron per molecule of 1. By the way, as

suggested by DFT calculations, the reactions of 1c with the

starting complexes 1 or 1a leading to 1b are thermodynamically

spontaneous (ΔG = –103.7 kJ/mol for 1c + 1  2 1b + H₂+ Br^–

and ΔG = –110.7 kJ/mol for 1c + 1a  2 1b + H₂+ MeCN), thus

Ph

HO

N

0

Mn

OC

CO

OH

2d

Scheme 2. Electrochemical mechanism for 2 under Ar.

Catalytic behavior under CO₂. To investigate the catalytic

process, we performed IR-SEC on a MeCN/TBAPF₆solution of

1 under CO₂-saturated conditions. During the first reduction,

where no catalytic current in CV was observed, the recorded

spectra are similar to those obtained under Ar, as the

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

consumption of the starting species 1 and 1a gives the

catalytically inactive neutral intermediate 1b, with typical ν_COat

2021, 1918 cm^–1and the ligand-based absorption at 1607 cm^–1

(Figure S12). Interestingly, the characteristic ν_COof 1c are not

observed at R1 and are replaced by growth of new bands in the

1700-1500 cm^–1region of the FTIR spectra. This seems to

suggest that either the Mn hydride 1c or its radical precursor

may interact with the substrate. Unfortunately this new set of ν_CO

cannot be univocally assigned to CO₂reduction products (e.g.

free HCOO^–) or to a Mn-CO₂adduct due to partial overlap with

the bands of 1b,^[15,26]so that we cannot rule out the formation of

a new fac-tricarbonyl Mn^I-CO₂complex, as found for an

analogous Mn^Iand Re^Icatalysts.^{[15,23,26,29]}Besides this, the

asymmetric ν_COOstretches of such adducts (between 1700-1600

to a decrease of the CO₂stretching mode at 2342 cm^–1(Figure

S16). Nevertheless, the detection of 2d even under CO₂-

saturated conditions confirms the reduced catalytic activity of 2

in comparison with 1. Furthermore, growth of the IR signal at

1607 cm^–1during catalytic CO₂reduction is less evident than in

the case of 1. This may suggest a slower formation of the

hydride species, believed to be responsible for catalytic CO₂

conversion into formate, and is consistent with a more difficult

intramolecular H⁺transfer for 2.

Electrolysis under CO₂in the presence of Brønsted acids.

The catalytic performances of 1 and 2 were also investigated in

the presence of different Brønsted acids. In particular, water,

TFE or phenol (2.7 M) were added to MeCN solutions (0.5 mM)

of 1 and 2. Bulk electrolysis was performed at –1.50 V (for 2h)

and –1.70 V (for 3h with water and phenol and for 50 min with

TFE), for 1 and 2, respectively. Table 2 summarizes the

quantitative results obtained during the exhaustive experiments.

cm^–1) would also completely overlap with those of HCO₃/CO₃

–

2–

.

^[9b,27]In agreement with the CV data, the catalytic CO₂reduction

process starts at more negative potentials than R1, as

suggested by an evident decay of the CO₂signal at 2342 cm^–1

(asymmetric stretching mode). In the ν_COregion the high-energy

band of 1b is slightly shifted to 2019 cm^–1, whereas the broad

absorption at 1918 cm^–1is slowly replaced by two different ν_COat

1921 and 1907 cm^–1(Figure S13). The recorded IR spectra do

not change significantly once these three bands are fully grown

up, as expected for the proceeding of a typical catalytic run. No

bands around 1800 cm^–1related to the anions 1d or 1f,

considered catalytically active for CO production, were observed

(Figure S13). At the same time, the rapid increase of the band at

1607 cm^–1, which is partially overlapped with the bpy-based

mode of 1b, gives a clear indication of the electrocatalytic

formation of free formate.^[9b,27]This is in agreement with the

results of the bulk electrolysis performed on 1 under CO₂in dry

MeCN.^[17]Moreover, two other bands, increasing at 1684 and

1646 cm^–1, can be assigned to free bicarbonate,^[9b,9d,27]and are

often related to catalytic CO production. Finally, as previously

observed,^[28]evolution of gas on the working electrode surface

and a small rising peak, caused by CO adsorbed on Pt, was

seen at 2138 cm^–1. The band at 1646 cm^–1may result from the

overlap with that of the carbonate ion (as an ion pair), since it

appears to grow independently of the signal at 1684 cm^–1.^[9b,29]At

lower CO₂concentrations, the three ν_COat 2019, 1921 and 1907

cm^–1are isosbestically converted into a mixture of 1d and 1e

only after the consumption of approximately 50% of the starting

CO₂(Figure S14).

Table 2. TON and faradic efficiencies (η) from bulk electrolysis (applied

potentials E in V vs. SCE) of a 0.1 M TBAPF₆MeCN solutions of 1 and 2 (0.5

mM) in the presence of different Brønsted acids (2.7 M).

TON

(η %)

E

Time

Acid

Catalyst

1

CO HCOO^–

H₂CO HCOO^–H₂

(V) (min) (2.7 M)

–1.50 120

–1.70 180

–1.70 50

H₂O

28

11

4

1.4

9

0.7 90

0.8 48

5.5 15

0.2 74

0.04 74

0.9 56

4

2

3

TFE

36

39

4

phenol

H₂O

12

0.5

0.3

0.8

21

2

7

2

TFE

2

10

15

1

–1.70 180

phenol

2

17

A CO₂flow of 50 mL min^–1was kept constant during the

experiments, whereas gaseous and liquid CO₂reduction

products were determined by gas (GC) and ion (IC)

chromatography, respectively. CVs carried out on solutions of 1

and 2 in the presence of H₂O show a significant catalytic current

enhancement under CO₂(Figure 9a), which is in agreement with

previously reported Mn catalysts. Notably, the catalytic currents

observed for 1 at the peak potentials of R2 and R3, in aqueous

MeCN under CO₂, are considerably higher than the peak current

achieved under an inert atmosphere at the same potentials,

indicating that 1 is able to efficiently and selectively reduce CO₂,

even in the presence of a considerable amount of water (Figure

Analogously to 1, CV of the complex 2 in a CO₂saturated MeCN

solution exhibits an increased peak current upon the second

reduction R2’ (at –1.50 V, Figure 3), even though the observed

electrocatalytic current enhancement is much lower in

9a). In this case the wave seems to have reached a plateau, so

[30]

that we can roughly estimate the catalytic rate constant, k_cat

,

equal to 81 s^–1. An analogous selective catalytic mechanism for

CO₂reduction is expected also for 2, being characterized

however by a considerably lower catalytic current under CO₂

(Figure 9b).

comparison to

1 under the same conditions (Figure 2).

Accordingly with the electrochemical data, no catalytic process

is observed after the first reduction in IR-SEC in a CO₂-saturated

MeCN solution of

2 (Figure S15), but the dimer 2b is

Bulk electrolysis experiments on CO₂-saturated aqueous MeCN

solutions well reproduce the CV data, revealing selective CO

production. In particular, faradaic efficiencies η_COof 90% and

74% with TON_COof 28 and 7 were found for 1 and 2,

respectively (Figure S17a-b). Conversely, the competing

catalytic processes that yield HCOO^–and H₂are almost

quantitatively produced from the starting 2 and 2a. However, as

previously discussed for 1, the small ν_COat 1987 cm^–1, assigned

to the Mn hydride 2c, is no longer detected during reduction

under CO₂. As in 1, a catalytic behavior is observed at the foot of

the second reduction, showing a growth of bands at 1684, 1646

(free HCO₃/CO₃^2–) and 1607 (free formate) cm^–1concomitantly

–

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

suppressed, providing η_HCOO= 4% (TON_HCOOof 1.4 for 1 and 0.5

for 2) and η_H2= 2% (TON_H2of 0.7 for 1 and 0.2 for 2) for both

catalysts.

minor amounts, with η_HCOO= 10% (TON_HCOO= 0.3 with TFE) and

15% (TON_HCOO= 0.8 with phenol). Finally, bulk electrolysis in

MeCN + 2.7 M TFE also produced traces of molecular hydrogen

by using 2 as a catalyst (η_H2= 1%, TON_H2= 0.04), and in higher

quantities when phenol is employed as external proton source

(η_H2= 17%, TON_H2= 0.9).

Conclusions

In order to shed light on the influence of a ligand-centered

proton relay on the CO₂electroreduction catalyzed by tricarbonyl

Mn diimine complexes, the electrochemical and spectroscopic

features of the complexes

1 and 2 were systematically

investigated under inert atmosphere as well as under CO₂. In

particular, the results here presented highlight that how little

modification in the position of the bpy-localized phenolic groups

(in ortho for 1 and in meta/para for 2) is able to produce

substantial differences in the reductive mechanism under Ar,

even inducing a different reactivity towards CO₂. Moreover, the

selectivity dependence for the electrocatalytic CO₂reduction

process upon the strength of the external proton source was

investigated in the presence of three different added acids (H₂O,

TFE, phenol).

a)

Taken together, CV and UV-Vis/IR-SEC data gave evidence of

different electrochemical mechanisms for 1 and 2 under Ar. A

fast Br^–dissociation of 1 at zero applied potential produces the

acetonitrile complex 1a as the main starting species in solution,

whose 1e reduction does not lead to dimer formation but rather

occurs via a reductive deprotonation process. As a result, the

neutral intermediate 1b is proposed as the main species formed

after R1, in agreement with the experimental and DFT results.

Furthermore, spectroscopic evidence of small amounts of a Mn

hydride (1c) is found after the first reduction, being related to the

OH proximity to the Mn center. No traces of such hydride

species are detected under CO₂atmosphere. IR-SEC under

CO₂in dry MeCN qualitatively well reproduces the quantitative

loss of selectivity highlighted during bulk electrolysis, providing a

mixture of CO and HCOO^–.^[17]Notably, the well-known proton-

independent mechanism of CO₂disproportionation to CO and

b)

Figure 9. CVs at 100 mV s^-1of 0.5 mM MeCN solutions of 1 and 2 under Ar

(blue), under Ar and with H₂O 2.7 M (red) and under CO₂and with H₂O 2.7 M

(black).

The presence of stronger Brønsted acids like TFE and phenol

highlighted important differences in the product distribution of 1

and 2. Unlike catalyst 2, which remains basically selective

towards CO production regardless of the acid employed, 1 exhi-

bits a much more pronounced acid-dependent change in selecti-

vity. In particular, in the presence of TFE, CO is still found as the

major product (η_CO= 48%, TON_CO= 11) (Figure S17c), whereas

the catalytic production of HCOO^–increases significantly (η_HCOO

= 36%, TON_HCOO= 9), giving an almost 1:1 CO/HCOO^–ratio. In

these conditions, H₂is still produced in negligible quantities (η_H2

= 3%, TON_H2= 0.8). This trend is amplified upon using an even

stronger Brønsted acid like phenol (2.7 M), leading to HCOO^–as

the major product (η_HCOO= 39%, TON_HCOO= 12) and to CO in

minor quantities (η_CO= 15%, TON_CO= 4). At the same time, H⁺

reduction becomes significant (η_H2= 21% TON_H2= 5.5, Figure

S18). Under the same experimental conditions, the catalytic

selectivity of 2 is found to be less sensitive to the strength of the

added acid than 1. In particular, the electrocatalytic reduction of

2–

CO₃cannot be ruled out under these conditions, as carbonate

and bicarbonate anions are spectroscopically observed.

On the other hand, the electrochemical reduction of 2 under Ar

follows the general scheme commonly reported for other

[Mn(bpy-R₂)(CO)₃Br] catalysts, indicating that dimerization

occurs upon R1’. Only traces of another Mn hydride species (2c)

are detected after R1’ in the absence of CO₂, but not under CO₂-

saturated conditions. Moreover, IR-SEC of

confirmed a non-selective CO₂reduction process in anhydrous

MeCN, as well as reduced electrocatalytic activity in

2 under CO₂

a

comparison with 1. Since pdbpy and ptbpy display similar

bulkiness, the deviation between the electrochemical behavior of

1 and 2 are consistent with the differences in the proximity of the

local proton source from the metal site and in the acidity of the

ligand moiety.

CO₂to CO is still the main process, giving η_CO= 74% (TON_CO

2, Figure S17d) and 56% (TON_CO= 2, Figure S19) at –1.70 V

with 2.7 M TFE and phenol, respectively. HCOO^–is formed in

=

Finally, controlled-potential electrolysis with different acids

confirms the existence of competing electrocatalytic pathways,

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

leading to CO₂conversion to CO and HCOO^–, in a ratio which is

highly dependent on the acidity of the solution. Whereas the

five-coordinate anions (e.g. 1d, 1f, 2d) are likely responsible for

the catalytic CO production, the HCOO^–formation is supposed

to be catalyzed by the transient Mn hydride species (1c, 2c).

Notably, addition of water (2.7 M) leads to selective CO₂

reduction to CO by both 1 and 2, but side HCOO^–and H₂

productions become more prominent as the strength of the

Brønsted acid increases. Accordingly to these experimental data,

acids stronger than water are able to give a fast reprotonation of

the ligand-centered phenolate groups, favoring the in-situ

formation of transient hydride species. Going from added water

to phenol, the CO:HCOO^–ratio can be tuned from ca. 23:1 to 1:3

with 1 as catalyst. In the case of 2, such trend is much less

evident than for 1 and this may be explained by a slower

intramolecular proton transfer to form the catalytically active

hydride. We believe that these findings contribute to a better

understanding of the factors that control the selectivity of a

promising class of earth-abundant transition metal catalysts.

(s, 1H), 8.26 (d, J= 7.3 Hz, 2H), 7.98 (t, J = 7.9 Hz, 1H), 7.86 (s, 1H),

7.53 (t, J = 7.7 Hz, 2H), 7.47-7.41 (m, 2H), 6.41 (s, 2H), 3.90 (s, 3H),

3.78 (s, 6H).

Synthesis

of

4-phenyl-6-(3,4,5-trihydroxyphenyl)-2,2'-bipyridine

(ptbpy): 0.9 mmol of (c) was added to a solution of HBr (33%) in acetic

acid and heated to reflux for 72 hours. The mixture was then cooled to

room temperature and neutralized using an aqueous solution of NaOH.

The green precipitate was filtered and washed with water and chloroform

(yield 93%). ¹H-NMR (400 MHz, d₆-DMSO): δ/ppm = 9.14 (s, 2H), 8.73 (d,

1

2

J = 4.7 Hz, 1H), 8.58-8.56 (m, 2H), 8.48 (s, 1H), 8.03 (td, J = 7.9 Hz, J

= 1.4 Hz, 1H), 7.98 (s, 1H), 7.92 (d, 7.3 Hz, 2H), 7.58 (t, J = 7.6 Hz, 2H),

7.53-7.49 (m, 2H), 7.31 (s, 2H). ¹³C-NMR (400 MHz, d₆-DMSO): δ/ppm =

156.97, 155.59, 155.44, 149.50, 149.39, 146.39, 138.05, 137.54, 135.15,

129.47, 128.99, 127.16, 124.55, 120.84, 117.04, 115.63, 106.20. MS

(ESI⁺): m/z = 357.28 [M-H]⁺.

Synthesis of [Mn(ptbpy)(CO)₃Br] (2): Mn(CO)₅Br (1 eq) and ptbpy

(1.01 eq) were refluxed for 4 hours in diethyl ether under stirring. The

reaction mixture was then cooled to room temperature and the orange

1

product was filtered and washed once with diethyl ether (yield 82%). H-

NMR (400 MHz, (CD₃)₂CO): δ/ppm = 9.31 (d, J = 4.7 Hz, 1H), 8.83-8.79

(m, 2H), 8.40-8.14 (m, 3H), 8.08 (d, J = 6.2 Hz, 2H), 7.90 (s, 1H), 7.76-

7.71 (m, 2H), 7.58 (m, 3H), 6.75 (s, 2H). ¹³C-NMR (400 MHz, (CD₃)₂CO):

δ/ppm = 166.81, 158.66, 157.56, 153.96, 150.73, 146.69, 139.45, 136.91,

135.65, 135.00, 131.14, 130.20, 128.43, 126.73, 125.37, 124.71, 119.56,

109.87, 109.78. IR (ATR): 3499, 3306, 2032, 1963, 1912 cm^–1. MS

(ESI⁺): m/z = 573.08, 575.02 [M-H]⁺.

Experimental Section

General considerations

NMR spectra were recorded on a JEOL EX 400 spectrometer (¹H

operating frequency 400 MHz) at 298 K. ¹H and ¹³C chemical shifts are

reported relative to TMS (δ = 0) and referenced against solvent residual

peaks. UV-Vis spectra were recorded in the 190-1100 nm range on an

Agilent 8453 spectrophotometer. Infrared spectra were recorded in the

4000-1000 cm^–1range (with a resolution 2 cm^–1) on a Nicolet iS50 FT-IR

spectrometer equipped with a KBr beam splitter and an MCT/A detector.

The microanalysis samples were dried under vacuum to constant weight

(20°C, ca. 0.1 Torr). Elemental analyses (C, H, N) were performed on a

Fisons Instruments 1108 CHNS-O Elemental Analyzer. Tetrabutyl-

ammonium hexafluorophosphate (TBAPF₆, Sigma-Aldrich, 98%) was

recrystallized twice from ethanol and dried before use. All reagents were

purchased from Sigma-Aldrich and used as received. Solvents were

freshly distilled and purged with argon before use. All Mn complexes

under study were carefully protected from light during use. Syntheses of

the precursor (E)-picolinoyl-phenyl-ethene (a) and 1 have previously

been reported.^[17]

Electrochemistry

All the solvents used for electrochemical experiments were freshly

distilled. Cyclic voltammetry experiments were performed using

a

Metrohm Autolab 302N potentiostat. 0.5-1.0 mM solutions of the

compounds were used in MeCN, with tetrabutylammonium

hexafluorophosphate (TBAPF₆) as the supporting electrolyte (0.1 M). A

single-compartment cell, with a glassy carbon (GC) working electrode (Ø

= 1 mm) was employed, while a Pt counter electrode and a SCE (KCl

3M) reference electrode were also used. In our experimental conditions,

the reference ferrocene/ferrocinium (Fc/Fc⁺) redox is at E_1/2= 0.39 V

(ΔE_p= 65 mV). The Ar- and CO₂-saturated conditions were achieved by

purging gases for 5 minutes before each potential sweep.

Spectroelectrochemistry

IR-SEC experiments were performed using the optically transparent thin-

layer electrode (OTTLE) cell, equipped with a Pt minigrid working and

auxiliary electrodes, an Ag microwire pseudo reference electrode and a

CaF₂window.^[33]IR spectra were recorded on a Nicolet iS50 FT-IR

spectrometer. UV-Vis spectroelectrochemistry was performed using the

OTTLE cell equipped with quartz optical windows and spectra were

measured with diode-array spectrometer Agilent 8453. The course of

spectroelectrochemical experiments was controlled and monitored using

a PA4 potentiostat and recorder (Laboratory Devices, Prague, Czech

Republic). TBAPF₆was used as the supporting electrolyte during the

experiments.

Synthesis of 1-(2-oxo-2-(3,4,5-trimethoxyphenyl)ethyl)pyridinium

iodide (b): 20 mmol of 3′,4′,5′-trimethoxyacetophenone and 24 mmol of

iodine were refluxed for 3 hours in 50 ml of pyridine. The reaction mixture

was cooled to 0° C and the yellow product was precipitated, filtered and

washed with cold pyridine (yield 80%).^{[31] 1}H-NMR [400 MHz, (CD₃)₂CO]:

δ/ppm = 9.36 (d, J = 5.8 Hz, 2H), 8.89 (t, J = 7.9 Hz,1H), 8.40 (t, J = 7.0

Hz, 2H), 7.50 (s, 2H), 7.07 (s, 2H), 3.96 (s, 6H), 3.85 (s, 3H).

Synthesis of 4-phenyl-6-(3,4,5-trimethoxyphenyl)-2,2'-bipyridine (c):

2.8 mmol of (a) and an equivalent amount of (b) were added to a flask

with an excess of ammonium acetate (28 mmol) and 20 ml of methanol.

The mixture was purged with Ar for ten minutes and was then heated to

reflux for 6 hours. The solution was dried on vacuum and the solid was

dissolved in 30 ml of ethyl acetate. The organic phase was washed three

times using an aqueous solution of NaHCO₃(10%) and then dried using

anhydrous MgSO₄. The solid collected after removal of the solvent was

purified on a silica gel column (petroleum ether/ethyl acetate 7:1; yield:

40%).^{[32] 1}H-NMR (400 MHz, (CD₃)₂CO): δ/ppm = 8.73-8.67 (m, 2H), 8.42

Controlled-potential electrolysis

A double compartment H-type cell was used with a Pt wire as the counter

electrode in a bridge separated from the cathodic compartment by a

glass frit. A glassy carbon rod was used as the working electrode along

with an aqueous SCE reference electrode. CPE experiments were

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

performed in acetonitrile with 0.1 M TBAPF₆as the supporting electrolyte

and 0.5 mM of the catalysts 1 and 2 at –1.5 V and –1.7 V vs. SCE,

respectively. The acids H₂O, TFE and phenol were added to the solution

in equimolar quantity (2.7 M). A controlled flow of CO₂(50 mL min^–1),

measured just before arrival into the cell, was maintained during the CPE

measurements by means of a Smart Trak 100 (Sierra) flow controller. All

the electrochemical cells were airtight and equipped with a bubbler that

maintained the inner atmosphere but avoided gas overpressure.

Sanpaolo) is gratefully acknowledged. J.F. thanks the Grant

Agency of the Czech Republic (grant 14-05180S) for support.

Keywords: carbon dioxide • manganese • electrochemistry •

DFT calculations • electrocatalysis

[1]

a) J. Qiao, Y. Liu, F. Hong, J. Zhang, Chem. Soc. Rev. 2014, 43, 631-

675; b) J. Ronge, T. Bosserez, D. Martel, C. Nervi, L. Boarino, F.

Taulelle, G. Decher, S. Bordiga, J. A. Martens, Chem. Soc. Rev. 2014,

43, 7963-7981.

Quantitative analysis of CO₂reduction products

[2]

a) C. Costentin, S. Drouet, M. Robert, J. M. Saveant, Science 2012,

338, 90-94; b) C. Costentin, G. Passard, M. Robert, J. M. Saveant,

Proc. Natl. Acad. Sci 2014, 111, 14990-14994; c) C. Costentin, M.

Robert, J.-M. Savéant, A. Tatin, Proc. Natl. Acad. Sci. 2015, 112,

6882–6886.

CO and H₂were detected and quantified using an Agilent 490 Micro GC

gas chromatograph equipped with a Molsieve 5 Å column, which was

kept at 90° C and a pressure of 200 kPa, and a thermal conductivity

detector. The carrier gas was Ar. The gas inside the measurement cell

was sampled for 90s every five minutes and fill the Micro GC 10 μL

sample loop, and eventually 500 nL were injected into the column for the

analyses. It was possible to determine the amounts of CO and H₂

produced during the experiment from the concentrations of CO and H₂in

the gas samples and knowledge of the gas flow and the cell volume. We

used Ar, He, N₂, O₂, CO₂and CO pure gases (>99.9995%), from Sapio,

for the operation and calibration of the flow controller and GC apparatus,

while H₂was produced using a Claind HyGen hydrogen generator. The

detection limits for CO and H₂were 20 ppmv and 0.5 ppmv respectively.

Formic acid production was evaluated using ion chromatography;

specifically, a Dionex DX 500 instrument equipped with a Dionex

IonPac® AS9-HC column (200 mm × 4 mm i.d.), GP40 pump (Dionex),

an electrochemical detector ED40 (Dionex) and an Anion Self-

Regenerating Suppressor-Ultra (ARSR®-ULTRA, 4-mm, Dionex). Elution

was performed in isocratic conditions with an aqueous solution of K₂CO₃

9.0 10^–3M at 0.70 mL min^–1flow rate. We observed a retention time of

6.23 min for the formate anion in these conditions. The TBAPF₆matrix

had to be removed for the analysis. Therefore, the samples were filtered

and brought to 25.0 mL in volumetric flasks after basification with

ammonia. We then injected 2.00 mL of each sample into a disposable

SPE column (containing BAKERBOND spe™ Quaternary Amine (N+)

and Carlo Erba Amberlite® CG-120, previously conditioned with HCl 0.01

M) and eluted with 8.00 mL of HCl 0.010 M. The resulting solutions were

brought to pH = 12 using NaOH 2 M and analyzed by means of ion

chromatography.

[3]

[4]

[5]

[6]

a) I. Wender, Fuel Process. Technol. 1996, 48, 189–297; b) M. Dry, E.

Catal. Today 2002, 71, 227-241.

C. Rice, S. Ha, R. I. Masel, P. Waszczuk, A. Wieckowski, T. Barnard, J.

Power Sources 2002, 111, 83–89.

A. S. Agarwal, Y. Zhai, D. Hill, N. Sridhar, ChemSusChem 2011, 4,

1301–1310.

a) J. Hawecker, J.-M. Lehn, R. Ziessel, J. Chem. Soc. Chem. Commun.

1984, 984, 328; b) J. W. Raebiger, J. W. Turner, B. C. Noll, C. J. Curtis,

A. Miedaner, B. Cox, D. L. Dubois, N. Renewable, C. B. V, N. Dame, et

al., Organometallics 2006, 25, 3345–3351; c) Z. Chen, C. Chen, D. R.

Weinberg, P. Kang, J. J. Concepcion, D. P. Harrison, M. S. Brookhart,

T. J. Meyer, Chem. Commun. 2011, 47, 12607; d) J. D. Froehlich, C. P.

Kubiak, J. Am. Chem. Soc. 2015, 137, 3565–3573.

[7]

a) C. M. Bolinger, B. P. Sullivan, D. Conrad, J. A. Gilbert, N. Story, T. J.

Meyer, J. Chem. Soc. Commun. 1985, 796–797; b) H. Ishida, H.

Tanaka, K. Tanaka, T. Tanaka, J. Chem. Soc. Chem. Commun. 1987,

131; c) C. Arana, S. Yan, M. Keshavarz-K., K. T. Potts, H. D. Abruna,

Inorg. Chem. 1992, 31, 3680–3682; d) A. Taheri, E. J. Thompson, J. C.

Fettinger, L. A. Berben, ACS Catal. 2015, 5, 7140–7151; e) L. Chen, Z.

Guo, X. G. Wei, C. Gallenkamp, J. Bonin, E. Anxolabéhère-Mallart, K.

C. Lau, T. C. Lau, M. Robert, J. Am. Chem. Soc. 2015, 137, 10918–

10921; f) P. Kang, C. Cheng, Z. Chen, C. K. Schauer, T. J. Meyer, M.

Brookhart, J. Am. Chem. Soc. 2012, 134, 5500–5503; g) P. Kang, S.

Zhang, T. J. Meyer, M. Brookhart, Angew. Chem. Int. Ed. 2014, 53,

8709–8713.

[8]

a) M. Bourrez, F. Molton, S. Chardon-Noblat, A. Deronzier, Angew.

Chem. Int. Ed. 2011, 50, 9903–9906; b) J. M. Smieja, M. D. Sampson,

K. A. Grice, E. E. Benson, J. D. Froehlich, C. P. Kubiak, Inorg. Chem.

2013, 52, 2484–2491; c) F. Franco, C. Cometto, C. Garino, C. Minero,

F. Sordello, C. Nervi, R. Gobetto, Eur. J. Inorg. Chem. 2015, 2015,

296–304; d) F. Franco, C. Cometto, F. Sordello, C. Minero, L. Nencini,

J. Fiedler, R. Gobetto, C. Nervi, ChemElectroChem 2015, 2, 1372–

1379; e) M. D. Sampson, C. P. Kubiak, J. Am. Chem. Soc. 2016, 138,

1386–1393.

Computational Details

DFT calculations were performed using Gaussian 09 Rev. D.01,^[34]

including the solvent effect by the conductor-like polarizable continuum

model (CPCM)^[35]with acetonitrile as solvent. Geometry optimizations

were carried out without any constraints using the B3LYP functional,^[36]

the optimized def2-TZVP basis set for Mn and Br and the def2-SVP basis

set^[37]for all other atoms. The D3 version of Grimme’s dispersion method

was applied adopting the Becke-Johnson damping scheme.^[38]Gibbs

Free Energies were determined by thermal corrections for entropy and

enthalpy at 298 K to the electronic energies. In these calculations, the

computed harmonic frequencies were scaled by 0.952 to account for

anharmonicity. For radical anions, unrestricted Kohn–Sham formalism

(UKS) was adopted. The nature of all stationary points were confirmed by

normal-mode analysis (no imaginary frequencies were found).

[9]

a) F. Hartl, B. D. Rossenaar, G. J. Stor, D. J. Stufkens, Recl. des Trav.

Chim. des Pays-Bas 2010, 114, 565–570; b) F. P. A. Johnson, M. W.

George, F. Hartl, J. J. Turner, Organometallics 1996, 15, 3374–3387; c)

C. W. Machan, M. D. Sampson, S. A. Chabolla, T. Dang, C. P. Kubiak,

Organometallics 2014, 33, 4550–4559; d) Q. Zeng, J. Tory, F. Hartl,

Organometallics 2014, 33, 5002–5008.

[10] E. E. Benson, M. D. Sampson, K. A. Grice, J. M. Smieja, J. D. Froehlich,

D. Friebel, J. A. Keith, E. A. Carter, A. Nilsson, C. P. Kubiak, Angew.

Chem. Int. Ed. 2013, 52, 4841–4844.

[11] M. Bourrez, M. Orio, F. Molton, H. Vezin, C. Duboc, A. Deronzier, S.

Chardon-Noblat, Angew. Chem. Int. Ed. 2014, 53, 240–243.

Acknowledgements

[12] D. C. Grills, J. Farrington, B. H. Layne, S. V. Lymar, B. A. Mello, J. M.

Preses, J. F. Wishart, J. Am. Chem. Soc. 2014, 136, 5563–5566.

[13] a) J. a. Keith, K. a. Grice, C. P. Kubiak, E. a. Carter, J. Am. Chem. Soc.

Financial support from project PHOTORECARB (Progetti di

Ateneo/CSP 2012, Call 03, Università di Torino & Compagnia

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

2013, 135, 15823–15829; b) C. Riplinger, M. D. Sampson, A. M.

Ritzmann, C. P. Kubiak, E. A. Carter, J. Am. Chem. Soc. 2014, 136,

16285–16298; c) C. Riplinger, E. A. Carter, ACS Catal. 2015, 5, 900–

908.

[26] a) D. H. Gibson, X. L. Yin, H. Y. He, M. S. Mashuta, Organometallics

2003, 22, 337–346; b) M. D. Sampson, J. D. Froehlich, J. M. Smieja, E.

E. Benson, I. D. Sharp, C. P. Kubiak, Energy Environ. Sci. 2013, 6,

3748–3755; c) T. Morimoto, T. Nakajima, J. Am. Chem. Soc. 2013, 135,

16825–16828.

[14] E. E. Benson, C. P. Kubiak, Chem. Commun. 2012, 48, 7374–6.

[15] M. D. Sampson, A. D. Nguyen, K. A. Grice, C. E. Moore, A. L.

Rheingold, C. P. Kubiak, J. Am. Chem. Soc. 2015, 137, 3718.

[16] We assume that protons were taken from solvent to reprotonate the

catalyst.¹⁷The terms anhydrous and dry will be used to indicate freshly

distilled organic solvents (thoroughly used for all electrochemical

experiments), without added Brønsted acids.

[27] S. C. Cheng, C. A. Blaine, M. G. Hill, K. R. Mann, Inorg. Chem. 1996, 2,

7704–7708.

[28] J. Tory, B. Setterfield-Price, R. A. W. Dryfe, F. Hartl, ChemElectroChem

2015, 2, 213–217..

[29]

P. Christensen, A. Hamnett, A. V. G. Muir, J. A. Timney, J. Chem. Soc.

Dalt. Trans 1992, 1455–1463.

[17] F. Franco, C. Cometto, F. Ferrero Vallana, F. Sordello, E. Priola, C.

Minero, C. Nervi, R. Gobetto, Chem. Commun. 2014, 50, 14670–3.

[18] H. Takeda, H. Koizumi, K. Okamoto, O. Ishitani, Chem. Commun. 2014,

50, 1491–3.

[30] I. Azcarate, C. Costentin, M. Robert, J.-M. Savéant , J. Phys. Chem. C

2016, 120, 28951-28960.

[31] A. Basnet, P. Thapa, R. Karki, Y. Na, Y. Jahng, B.-S. Jeong, T. C.

Jeong, C.-S. Lee, E.-S. Lee, Bioorg. Med. Chem. 2007, 15, 4351–4359.

[32] F. Neve, M. Ghedini, O. Francescangeli, S. Campagna, Liq. Cryst.

1998, 24, 673–680.

[19] H. Fei, M. D. Sampson, Y. Lee, C. P. Kubiak, S. M. Cohen, Inorg.

Chem. 2015, 54, 6821–6828.

[20] J.-D. Compain, M. Bourrez, M. Haukka, A. Deronzier, S. Chardon-

Noblat, C. Costentin, M. Robert, J. M. Saveant, B. Kumar, M. Llorente,

et al., Chem. Commun. 2014, 50, 2539.

[33] M. Krejčik, M. Daněk, F. Hartl, J. Electroanal. Chem. Interfacial

Electrochem. 1991, 317, 179–187.

[34] Frisch, M. J. et al Gaussian 09, Rev. D.01 2009.

[35] a) S. Miertus, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117–129;

b) M. Cossi, G. Scalmani, N. Rega, V. Barone, J. Chem. Phys. 2002,

117, 43–54.

[21]

J. J. Walsh, C. L. Smith, G. Neri, G. F. S. Whitehead, C. M. Robertson,

A. J. Cowan, Faraday Discuss. 2015, 183, 147–160.

[22] J. Agarwal, T. W. Shaw, H. F. Schaefer, A. B. Bocarsly, Inorg. Chem.

2015, 54, 5285–5294.

[36] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648; b) C. Lee, W. Yang, R.

G. Parr, Phys. Rev. B 1988, 37, 785–789.

[23] G. F. Manbeck, J. T. Muckerman, D. J. Szalda, Y. Himeda, E. Fujita, J.

Phys. Chem. B 2015, 119, 7457–7466.

[37]

a) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7,

3297–305; b) F. Weigend, Phys. Chem. Chem. Phys. 2006, 8, 1057–

1065.

[24] F. J. Garcia Alonso, A. Llamazares, V. Riera, M. Vivanco, S. Garcia

Granda, M. R. Diaz, Organometallics 1992, 11, 2826–2832.

[25] a) K. A. Grice, N. X. Gu, M. D. Sampson, C. P. Kubiak, Dalt. Trans.

2013, 42, 8498; b) D. Osella, C. Nervi, M. Ravera, J. Fiedler, V. V.

Strelets, Organometallics 1995, 14, 2501–2505.

[38] S. Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32, 1456–

1465.

10.1002/chem.201605546

Chemistry - A European Journal

FULL PAPER

Entry for the Table of Contents

FULL PAPER

Federico Franco, Claudio Cometto, Luca

Nencini, Claudia Barolo, Fabrizio

Sordello, Claudio Minero, Jan Fiedler,

Marc Robert, Roberto Gobetto* and

Carlo Nervi*

Page No. – Page No.

The electrochemical behavior of fac-[Mn(pdbpy)(CO)₃Br] (pdbpy = 4-phenyl-6-

(phenyl-2,6-diol)-2,2'-bipyridine) in MeCN under Ar and its catalytic performances

for CO₂reduction with added water, TFE and phenol are discussed. Bulk electro-

lysis reveals that the process selectivity is sensitive to the acid strength, providing

CO and formate in different faradaic yields. The key role of the local proton source

is demonstrated by a detailed spectroelectrochemical study under Ar and CO₂.

Local Proton Source in the

Electrocatalytic CO₂Reduction by

Mn(bpy-R)(CO)₃Br Complexes

Article Doi

Local Proton Source in Electrocatalytic CO2 Reduction with [Mn(bpy–R)(CO)3Br] Complexes

DOI: 10.1002/chem.201605546

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201605546

Full text of DOI:10.1002/chem.201605546

Products guided by the article

R&D Labs maybe for 1448527-66-9

Relevant to this article

Hot Product