Photo- and electrocatalytic H2 production by new first-row transition-metal complexes based on an aminopyridine pentadentate ligand

DOI: 10.1002/chem.201303317

Full Paper

&

Hydrogen Production

Photo- and Electrocatalytic H₂Production by New First-Row

Transition-Metal Complexes Based on an Aminopyridine

Pentadentate Ligand

Arnau Call, Zoel Codolꢀ, Ferran AcuÇa-Parꢁs, and Julio Lloret-Fillol*^[a]

Abstract: The synthesis and characterisation of the penta-

dentate ligand 1,4-di(picolyl)-7-(p-toluenesulfonyl)-1,4,7-tri-

azacyclononane (Py₂^Tstacn) and their metal complexes of

general formula [M(CF₃SO₃)(Py₂^Tstacn)][CF₃SO₃], (M=Fe (1_Fe),

Co (1_Co) and Ni (1_Ni)) are reported. Complex 1_Copresents ex-

cellent H₂photoproduction catalytic activity when using

[Ir(ppy)₂(bpy)]PF₆(PS_Ir) as photosensitiser (PS) and Et₃N as

electron donor, but 1_Niand 1_Feresult in a low activity and

a complete lack of it, respectively. On the other hand, all

three complexes have excellent electrocatalytic proton re-

duction activity in acetonitrile, when using trifluoroacetic

acid (TFA) as a proton source with moderate overpotentials

for 1_Co(0.59 V vs. SCE) and 1_Ni(0.56 V vs. SCE) and higher for

1_Fe(0.87 V vs. SCE). Under conditions of CH₃CN/H₂O/Et₃N

(3:7:0.2), 1_Co(5 mm), with PS_Ir(100 mm) and irradiating at

447 nm gives a turnover number (TON) of 690 (n_H/n₁) and

production. It should be noted that 1_Coretains 25% of the

catalytic activity for photoproduction of H₂in the presence

of O₂. The inexistence of a lag time for H₂evolution and the

absence of nanoparticles during the first 30 min of the reac-

tion suggest that the main catalytic activity observed is de-

rived from a molecular system. Kinetic studies show that the

reaction is ꢀ0.7 order in catalyst, and time-dependent dif-

fraction light scattering (DLS) experiments indicate formation

of metal aggregates and then nanoparticles, leading to cata-

lyst deactivation. By a combination of experimental and

computational studies we found that the lack of activity in

photochemical water reduction by 1_Fecan be attributed to

II/I

the 1_Feredox couple, which is significantly lower than the

PS_Ir^III/II, while for 1_Nithe pK_avalue (ꢀ0.4) is too small in com-

parison with the pH (11.9) imposed by the use of Et₃N as

electron donor.

2

Co

initial turnover frequency (TOF) (TONꢂt^ꢀ1) of 703 h^ꢀ1for H₂

Introduction

intense search has been dedicated to find earth-abundant

transition-metal complexes able to catalyse the reduction of

protons to hydrogen at low overpotentials or by using light as

a source of energy.^[5]Earth abundant molecular complexes

based on cobalt,^[5a]nickel,^[6]iron^[7]and molybdenum^[8]have

been shown to be able to reduce the protons of water to H₂

under electro- and photochemical conditions in an efficient

manner. However, a comparative study of the factors that

affect the reactivity in catalytic water reduction among a series

complexes based on different metals with the same ligand

framework to the best of our knowledge has not been ex-

plored in detail so far.^[9]

Hydrogen has gained special attention as the most promising

fuel for the future use in transport and energy storage in

terms of sustainability, carbon neutrality and energy conver-

sion. However, there are many challenges that still remain; one

of the most important is its efficient production. Water split-

ting promoted by sunlight is the most widely desired energetic

pathway for hydrogen production.^[1]In nature, iron and nickel

hydrogenase enzymes^[2]produce H₂, but they are inconven-

iently instable under aerobic conditions making their applica-

tion in H₂production devices difficult.^[3]Many efforts have

been devoted to develop biomimetic catalytic systems of such

enzymes, producing stimulating examples of H₂production en-

zymatic models, but usually such catalysts have a tendency to

be unstable against O₂and operate generally at relatively high

negative potentials.^[4]On the other hand, in the last years an

Very often water reduction requires intermediates in which

metals are in a low oxidation state, and which are frequently

stabilised by oxygen-sensitive or easily hydrolysable ligands. It

is then crucial to develop catalysts that are active in water as

the reaction media, are resistant to oxygen and which could

operate at a low overpotential and by light-driven. In this

regard, previous studies,^[10–12]relying on abundant metals and

robust polypyridyl ligands are encouraging for the electrocata-

lytic and photocatalytic reduction of water. More recently,

more versatile amino pyridine ligands have also shown impres-

sive catalytic performance. Up to now, systems that contain

from five chelating pyridines to three pyridines and two

amines have been studied (Figure 1).

[a] A. Call, Z. Codolꢀ, F. AcuÇa-Parꢁs, Dr. J. Lloret-Fillol

Institut de Quꢂmica Computacional i Catꢀlisi (IQCC) and

Departament de Quꢂmica

Universitat de Girona Campus Montilivi, 17071, Girona (Spain)

Fax: (+34)972418150

E-mail: julio.lloret@udg.edu

Supporting information for this article is available on the WWW under

http://dx.doi.org/10.1002/chem.201303317.

Chem. Eur. J. 2014, 20, 1 – 14

1

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

Results and Discussion

The synthesis of the ligand 1,4-di(picolyl)-7-(p-toluenesulfonyl)-

1,4,7-triazacyclononane (Py₂^Tstacn) has been achieved in excel-

lent yield (82%) from a direct dialkylation of the monotosylat-

ed 1,4,7-triazacyclononane with two equivalents of 2-(chloro-

methyl)pyridine hydrochloride. The reaction of the Py₂^Tstacn

ligand with one equivalent of M(OTf)₂·2CH₃CN (M=Co, Fe and

Ni) in THF at room temperature afforded the M^IIfive-coordinate

monocationic

complexes

of

general

formula

[M-

(CF₃SO₃)(Py₂^Tstacn)][CF₃SO₃], M=Co (1_Co), M=Fe (1_Fe) and M=

Ni (1_Ni) (Scheme 1).

Figure 1. Selected polypyridine and polypyridinamine cobalt complexes

studied in proton and water reduction.^[9a,10,12e]

We are interested in developing robust catalysts that could

retain their molecular nature under catalytic conditions with

the aim to study reaction mechanisms. In this regard, we previ-

ously reported a highly efficient water oxidation catalysts

based on iron complexes and highly chelating tetradentate

aminopyridine ligands.^[13]In particular, 1-pyridine-4,7-dimethyl-

triazacyclononane (Pytacn) has been particularly useful to in-

vestigate the reaction mechanism due to the conferred high

stability of its iron coordination complexes.

It is known that the triazacyclononane (tacn) moiety is

a highly electron-donating, redox-innocent, ligand moiety that

stabilises high oxidation states at metal centres. In principle,

the application of ligands derived from this unit to water re-

duction are counterintuitive considering the mechanisms pro-

posed for molecular complexes with polypyridine ligands that

usually require reduction of the metal centre to M^Ior even

M^o.^[14]In this regard, the introduction of electron-withdrawing

substituents could decrease the ligand stabilisation and favour

the low oxidation states, but preserve the high chelating

nature of the moiety. More interestingly is to explore catalysts

for both water oxidation (WO) and water reduction (WR) reac-

tions based on analogous ligand structures. To date, only a few

examples of catalysts have been reported that can carry out

both WO and WR catalysis.^[15]Water splitting devices could be

favoured by the use of both WO and WR catalysts build from

comparable structural motifs in order to achieve compatibili-

ties and synergies among homogeneous systems.

Scheme 1. Synthesis of ligand Py₂^Tstacn and complexes 1_Co, 1_Feand 1_Ni.

All three M^IIcoordination complexes 1_Co, 1_Feand 1_Nishowed

1

broad signals in their H NMR spectra that expand from ꢀ10 to

140 ppm, indicative of paramagnetic species (see Supporting

Information, Figures SI.3, SI.4 and SI.5). High-resolution electro-

spray ionisation mass spectrometry (HR-ESI-MS) shows the

monocharged molecular peak (673.1054 m/z for [1_CoꢀOTf]⁺,

670.1069 m/z for [1_FeꢀOTf]⁺and 672.1058 m/z for [1_NiꢀOTf]⁺)

and the double charged molecular peak of the complexes

without the triflates (262.0766 m/z for [1_Coꢀ2OTf]²⁺

,

260.5804 m/z for [1_Feꢀ2OTf]²⁺and 261.5797 m/z for

[1_Niꢀ2OTf]²⁺). In agreement, X-ray diffraction analysis of mono-

crystals from 1_Coand 1_Niconfirm the formation of distorted oc-

tahedral complexes formed by a pentacoordinate ligand

bonded to the metal by the three amines of the triazacyclono-

nane unit and two different pyridines; the last coordination

site occupied by a labile triflate anion (Figure 2). The selected

bond lengths and angles are summarised in Figure 2. MꢀN

bond lengths are in the range of the expected values for the

two pyridines (ca. 2.10 and 2.05 ꢄ for 1_Coand 1_Ni, respectively)

and the alkyl amines (ca. 2.14 and 2.07 ꢄ for 1_Coand 1_Ni, re-

spectively), but considerably larger for the tosylated amine

(2.320 and ca. 2.264 ꢄ for 1_Coand 1_Ni, respectively).

With this aim, herein we present a robust platform assem-

bled by appending on the N atoms of the tacn ring with one

electron-withdrawing tosyl group and two methylpyridine

arms, giving a pentadentate ligand, as well as their corre-

sponding iron, cobalt and nickel complexes. The resultant Co

complex is active in both electro- and photochemical H₂pro-

duction, while the Fe and Ni complexes are only electrocatalyt-

ically active. The comparative study among complexes with

the same ligand framework allows us to establish basic princi-

ples for water reduction based on aminopyridine ligands.

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

2

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Paper

(TON_cath^ꢀ1) values of 260 and 174 h^ꢀ1(2.1 mL(H₂)h^ꢀ1), respec-

tively. In contrast, complex 1_Niproduces small amounts of hy-

drogen (6 TON_cat) and 1_Fedoes not show any photocatalytic ac-

tivity (see the Supporting Information, Figure SI.7). This is

a very interesting result, but the difference among the activi-

ties of the complexes is difficult to rationalise at this point and

will be discussed in more detail in the following sections. Con-

trol experiments show that the three components are needed

for the photocatalytic H₂evolution. Furthermore, when the re-

action was performed in dark conditions, no gas production

could be detected. Finally, only negligible amounts of H₂

where measured when using Co(OTf)₂·2CH₃CN, Ni-

(OTf)₂·2CH₃CN and Fe(OTf)₂·2CH₃CN under analogous photo-

chemical conditions. Therefore, these results provide strong

evidence that the H₂obtained in our experiments cannot be

attributed to the activity of free metal in solution, and there-

fore we conclude that the catalytic activity arises from the

metal complex.

Photosensitiser

H₂production by 1_Cowas drastically reduced when photosensi-

tisers such as [Ru(bpy)₃]²⁺(PS_Ru) or erythrosine B (PS_EB) were

used instead of PS_Ir(TON_cat=20 with an initial TOF_catof 26 h^ꢀ1

and negligible, respectively) (see Figure SI.8 in the Supporting

Information). This can be rationalised by the reduction poten-

tials of the photosensitiser (PS_Ir: ꢀ1.55,^[16]PS_Ru: ꢀ1.34,^[17]and

PS_EB: ꢀ1.05 V^[18]vs. SCE). The redox potential E8(1_Co^II/I), mea-

sured by cyclic voltammetry, was found ꢀ1.10 V versus SCE.

Therefore PS_Irhas a higher reduction potential, and it is appro-

priate to account for the catalytic activity.

Figure 2. ORTEP structures with ellipsoids set at the 50% probability level of

1_Co(top) and 1_Ni(bottom) from the X-ray diffraction analysis. Hydrogen

atoms have been omitted for clarity. Relevant distances [ꢄ] and bond angles

[8] for 1_Co: Co(1)ꢀO(1), 2.034(1); Co(1)ꢀN(1), 2.141(2); Co(1)ꢀN(2), 2.093(2);

Co(1)ꢀN(3), 2.320(2); Co(1)ꢀN(4), 2.130(2); Co(1)ꢀN(5), 2.133(2); O(1)-Co(1)-

N(1), 172.39(6); N(2)-Co(1)-N(4); 161.67(7); N(3)-Co(1)-N(5), 159.09(6); N(1)-

Co(1)-N(2), 80.05(6); N(1)-Co(1)-N(3), 81.61(6); N(1)-Co(1)-N(4), 82.20(6); N(1)-

Co(1)-N(5), 96.47(6); N(2)-Co(1)-N(3), 101.95(6); N(4)-Co(1)-N(5), 79.31(7) and

for 1_Ni: Ni(1)ꢀO(1), 2.0488(17); Ni(1)ꢀN(1), 2.071(2); Ni(1)ꢀN(2), 2.042(2);

Ni(1)ꢀN(3), 2.264(2); Ni(1)ꢀN(4), 2.076(2); Ni(1)ꢀN(5), 2.066(2); O(1)-Ni(1)-N(1),

173.15(7); N(2)-Ni(1)-N(4); 166.31(8); N(3)-Ni(1)-N(5), 161.82(7); N(1)-Ni(1)-

N(2), 82.18(8); N(1)-Ni(1)-N(3), 83.67(8); N(1)-Ni(1)-N(4), 84.51(8); N(1)-Ni(1)-

N(5), 96.52(8); N(2)-Ni(1)-N(3), 100.46(7); N(4)-Ni(1)-N(5), 80.75(8).

Effect of the sacrificial electron donor and the solvent mixture

Photocatalytic water reduction to H₂

The light-driven hydrogen production catalysed by 1_Cois also

dependent on the amount of sacrificial electron donor and the

solvent mixture. For instance, the yield of H₂production is

lower at higher Et₃N loadings than 2% in volume of the reac-

tion mixture (Figure 3). This reduction in activity can be ration-

alised by the increase in pH due to the increase on concentra-

tion of Et₃N (from pH 11.9 using 1% of Et₃N to pH 12.6 using

15% of Et₃N), which may hinder either the formation of the

CoꢀH intermediate or the reaction of CoꢀH with protons to

form H₂. Likewise, the ratio water/CH₃CN modifies the catalytic

activity due to the need to adjust the polarity of the medium

to have all the components well solubilised. The highest activi-

ty was found at 70% (v/v) of water in acetonitrile (TON: 260,

TOF: 174 h^ꢀ1) (Figure 3).

Effect of the metal

First, we tested the catalytic activity of 1_Fe, 1_Coand 1_Niunder

light-driven conditions by using a photosensitiser and a sacrifi-

cial electron-donor under irradiation (l=447 nm). Hydrogen

evolution was monitored online by the difference of pressure

between the reaction vessel containing the mixture of the re-

action and a blank sample under the same conditions, but

lacking the metal complex. Reactions were carried out by

using an acetonitrile/water mixture; acetonitrile was used to

ensure the complete solubilisation of the photosensitiser. The

amount of hydrogen obtained at the end of the reaction was

established in every case by analysing an aliquot of the head-

space using gas chromatography with a thermal conductivity

detector (GC-TCD).

Effect of the photosensitiser and catalyst concentration

The comparison among the activity of the three complexes

was carried out by irradiation of a mixture containing the com-

plex 1_m(50 mm), [Ir(ppy)₂(bpy)]PF₆(PS_Ir, 250 mm, ppy: phenyl

pyridine) as photosensitiser and Et₃N (200 mL) as a source of

electrons, in water/acetonitrile (7:3 mL) mixture. Under these

conditions complex 1_Cois highly active, producing 3.2 mL of H₂

The increase of the [PS_Ir] had a positive impact in the total

amount H₂obtained. The TON_catincreases linearly with the

concentration of the chromophore; however, the TON_PS(n_H/

2

n_PS) is almost independent of the photosensitiser concentra-

tion. For instance, TON_catvalues of 68 and 327 and TON_PS

values of 68 and 55 were found when using 50 and 300 mm of

after 1.5 h of reaction with a TON_cat(n_H/n_complex) and TOF_cat

2

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

3

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

Regeneration of the catalytic

activity

All efforts to regenerate the cat-

alytic activity at the end of the

reaction

were

unsuccessful.

Indeed, further addition of any

of the reaction components (PS_Ir,

1_Coor Et₃N) at the end of the re-

action did not restart the catalyt-

ic activity. Furthermore, the com-

bined addition of catalyst and

Et₃N or PS_Irand Et₃N results in

negligible H₂evolved. Only 25%

of the catalytic activity could be

restored upon addition of more

catalyst and PS and 40% when

all three components were

added at the same time (see the

Supporting Information, Fig-

ure SI.9). Altogether, these re-

sults suggest that the catalytic

activity degradation over time

may be due to a poisoning of

Figure 3. Study of the effect of solvent mixture, [Et₃N], [PS_Ir] and [1_Co] in the photocatalytic H₂formation by 1_Co

.

Top left: effect of Et₃N% v/v in the solvent mixture, CH₃CN/H₂O (2:8 mL), 1_Co(50 mm) and PS_Ir(250 mm). Top right:

effect of the CH₃CN/H₂O solvent mixture (total volume 10 mL), Et₃N (0.2 mL), 1_Co(50 mm), PS_Ir(250 mm). Bottom

left: TON_1Coand TON_PSas a function of [PS_Ir] in CH₃CN/H₂O/Et₃N (2:8:0.2 mL), 1_Co(50 mm). Bottom right: TON_1Co

and TON_PSas a function of [1_Co], CH₃CN/H₂O/Et₃N (2:8:0.2 mL), PS_Ir(100 mm). TON_cat=n_H/n_cat, TON_PS=n_H/n_PS.

2

Ir

PS_Ir, respectively (Figure 3). The loss of activity at PS_Irconcen-

trations higher than 300 mm can be attributed to the limited

PS_Irsolubility under these conditions. Initial reaction rates cal-

culated at 10% of conversion versus the [PS_Ir] revealed first-

order dependence (Figure 4), which is an indication that the re-

II

duction of the cobalt complex by the photoreduced PS_Ircon-

trols the H₂production rate.

Afterwards, we studied the dependence of the H₂produc-

tion versus the concentration of 1_Co. Experiments were carried

out varying the concentration of catalyst (5–100 mm) and fixing

the concentration of PS_Ir(100 mm) with CH₃CN/H₂O/Et₃N

(2:8:0.2) as solvent mixture (Figures 3 and 4). The activity clear-

ly increases with the decrease of the 1_Coconcentration. At low

catalyst concentrations (1_Co=5 mm, PS_Ir=100 mm) a TON₁of

Co

690 and a TOF₁of 703 h^ꢀ1were obtained. At lower catalysts

Co

concentrations, higher TON are clearly produced; however, the

inaccuracy of the measurements under these conditions avoids

further analysis. On the other hand, at higher catalyst concen-

trations (100 mm) the hydrogen production rate is still remark-

able (1.46 mLh^ꢀ1of H₂). Initial reaction rates shows that the re-

action becomes faster at lower 1_Coconcentration. The double

ln plot of the k_iversus 1_Coconcentration shows an

inverse catalytic-activity/metal-complex concentration correla-

tion (Figure 4, top, inset). This might suggest the formation of

inactive dimers or oligomeric species, but also cobalt nanopar-

ticles, although the mononuclear species may have the higher

activity and increase in number when the concentration is de-

creased.

Figure 4. Plots of on-line pressure monitored versus time for reactions using

CH₃CN/H₂O/Et₃N (2:8:0.2) as solvent mixture under irradiation at l=447 nm.

Top : PS_Ir(100 mm) and [1_Co] from 5 mm to 100 mm. Inset: double-ln plot of in-

itial H₂evolution rates versus [1_Co]. Bottom: 1_Co(50 mm) and [PS_Ir] from

25 mm to 125 mm. Inset: double-ln plot of initial H₂evolution rates versus

[PS_Ir].

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

4

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Paper

II

the catalytic system by decomposition products formed during

the H₂evolution.

changes. Therefore, this implies the reduction of 1_Coby PS_Iris

a one-electron process. Continuous irradiation of the solution

II

produced the recovery of PS_Ir. In the case of the addition of

1_Ni(1 equiv), again the UV/Vis features of PS_Irvanished with

a concomitant rising of a new broad band with a l_maxat

523 nm, which was attributed to the reduction of the 1_Nicom-

plex by one electron (Figure SI.13 and SI.16 in the Supporting

Information). However, in the case of the addition 1_Fe(1 equiv)

II

Effect of the O₂presence in the catalytic activity

One of the main problems of the water reduction catalysts is

that for most catalysts the presence of O₂inhibits completely

the H₂formation.^[10b]Interestingly, the reactions of 1_Counder

air retain a significant amount of photoinduced H₂-evolution

activity. For instance, a reaction carried out with 1_Co(50 mm)

and PS_Ir(250 mm) under air in a solvent mixture of CH₃CN/H₂O/

Et₃N (3:7:0.2) gives a TON_catvalue of 63, and a TOF value of

44 h^ꢀ1. Nevertheless, the same reaction carried out under strict

inert conditions produce a total TON_catvalue of 259 and a TOF

value of 147 h^ꢀ1. Despite the fact that O₂clearly reduces the

evolution of H₂, up to 25% of the activity is maintained (see

the Supporting Information, Figure SI.10). Recently, Reisner

et al. reported the first first-row transition-metal catalyst that

can produce hydrogen in presence of high levels of O₂.^[19]This

was achieved by using a homogeneous catalytic system based

on a cobaloxime type of catalyst and eosin Y as photosensitis-

er.^[20]Nevertheless, our system operates under distinct condi-

tions.

II

only partial loss and then only partial recovery of PS_Iroccurs

under continuous irradiation (Figure SI.13 and SI.14 in the Sup-

II

porting Information). This clearly indicates that PS_Irdo not

have enough potential to reduce 1_Fe. The extent of the reactiv-

II

ity with the PS_Irfollows the order 1_Coꢁ1_Ni@1_Fe, which is in

agreement with the redox potentials measured in CH₃CN

(E(1_Co^II/I)=ꢀ1.10, E(1_Ni^II/I)=ꢀ1.07 and E(1_Fe^II/I)=ꢀ1.38 V vs.

SCE). This behaviour infers that under catalytic conditions the

formation of reducing species in case of 1_Coand 1_Niis fav-

oured, while is unlikely for 1_Fe. Consequently, one of the rea-

sons of photochemical inactivity in H₂production of 1_Feis the

II

inability of PS_Irto generate the necessary reduced species of

1_Fe.

Mechanistic studies

To get insight into the first steps

of the mechanism we studied

the reaction of the reduced pho-

tosensitiser

(PS_Ir^II=[Ir^II(ppy)₂-

(bpy)][PF₆]; ppy=2-phenylpyri-

dine) with 1_m(M=Co, Ni and Fe)

II

by UV/Vis spectroscopy. PS_Ir

was generated in situ by irradia-

tion of

a

solution containing

(0.1 mm) and Et₃N

in anhydrous

III

PS_Ir

(1000 equiv)

CH₃CN under N₂atmosphere. Re-

actions were performed at

ꢀ208C to minimise the decom-

II

position of PS_Ir, which occurs

fast at 258C. Upon irradiation,

three bands located at 385, 495

and 527 nm belonging to the

II

PS_Irraised (Figure 5 and Fig-

ure SI.12 in the Supporting Infor-

mation). After quantitative for-

II

mation of PS_Ir(ca. 30 s) one

equivalent of 1_mwas added. At

this point, the addition of 1_Co

(1 equiv) produced the vanishing

of the spectral features corre-

II

sponding to PS_Irin approxi-

mately 1.5 s (Figures SI.13 and

SI.15 in the Supporting Informa-

tion) with a concomitant recov-

III

II

Figure 5. UV/Vis monitoring of PS_Ir(0.1 mm) reductive quenching with Et₃N (1000 equiv) to form PS_Irat

l=447 nm, and subsequent addition of one equivalent of 1_Fe, 1_Coor 1_Niin anhydrous CH₃CN at ꢀ208C, under N₂

ery of PS_Ir^III, as judged by UV/Vis atmosphere.

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

5

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

Additional UV/Vis spectroscopic experiments were per-

formed monitoring the amount of H₂produced under catalytic

conditions (CH₃CN/H₂O (3:7) at 258C). The irradiation of the

the reaction and no induction time is observed under the

same conditions, we conclude that the H₂production activity

is molecular. Moreover, we found that the concentration of

nanoparticles in solution was very low, usually at the limit of

detection for this instrument (the lowest nanoparticle concen-

tration that can be detected for a reliable measurement

>0.1 ppm), and it was not always possible to have a reliable

size distribution measurement. The same conclusions were

reached by studying the same experimental conditions by

using nanoparticle tracking analysis (NTA). NTA showed values

of nanoparticlesmL^ꢀ1in the 1ꢂ10⁸–1ꢂ10⁹range. We note that

nanoparticle concentration did not significantly vary before

and after irradiation (ca. 1.5- to 2-fold increase), with oscilla-

tions under the same order of magnitude. Moreover, the same

experiments conducted in the presence of Hg⁰reduce the

nanoparticle concentration by a similar proportion. Further-

more, if we consider the obtained concentration of 1ꢂ10⁸–1ꢂ

10⁹particlesml^ꢀ1by NTA as being responsible for the catalytic

water reduction, a TOF value of 10⁶–10⁷s^ꢀ1is obtained. This

value is evidently unrealistic and hence the low concentration

of nanoparticles formed can be excluded as being responsible

for the observed catalytic activity.

II

sample (l=447 nm) forms PS_Ir. Further stoichiometric addi-

tion of 1_Coproduced a clear consumption of the reduced pho-

II

tosensitiser (PS_Ir) with a concomitant formation of H_2,up to

170 mL H₂were formed after 1.5 h of irradiation. Most impor-

tantly no induction time was observed (Figure 6, bottom/

inset). This is very relevant, since it is an indication that 1_Cois

the competent catalytic species. In contrast, when 1_Niwas

added, 20 mL H₂were formed, while in case of 1_Feno H₂was

detected (Figure 5), which qualitatively correlates with the

II

[PS_Ir] decay in each case, as judged by the UV/Vis.

DLS analysis on catalytic reactions of 1_Counder high concen-

tration conditions (500 mm) and typical concentrations of PS_Ir

(200 mm) and Et₃N (2% v/v), reveal more clearly the formation

of nanoparticles, which again increase in size during the reac-

tion (Figure 7). NTA under these conditions reveal nanoparticle

Figure 6. Top: UV/Vis monitoring of the reductive quenching of PS_Ir^IIspecies

generated by irradiating at l=447 nm, upon addition of one equivalent of

1_Fe, 1_Coand 1_Ni. Conditions: [PS_Ir]=0.1 mm, Et₃N=1000 equiv, reaction volu-

me=2 mL, CH₃CN/H₂O (0.6:1.4) at 258C. Bottom: Time-dependent photoca-

talytic H₂evolution activities measured in the UV/Vis cell upon addition of

~

&

^

( ), 1_Co( ) and 1_Ni( ). Inset: Magnification in the case

one equivalent of 1_Fe

of 1_Co.

Figure 7. Number of nanoparticles distribution after different irradiation

times: before irradiation, 5 min, 30 min, 1 h and 2 h. Conditions: 1_Co

(500 mm), PS_Ir(200 mm), CH₃CN/H₂O/Et₃N (5:5:0.2) as solvent mixture irradiat-

ed at 447 nm.

Study of nanoparticle formation during catalysis

Dynamic light scattering (DLS) experiments and tracking analy-

sis of nanoparticles (NTA) were performed during the reaction

to ascertain whether catalyst degradation could result in nano-

particle formation during the irradiation time.^[21]DLS analysis

on catalytic reactions of 1_Co(50 and 200 mm), in which PS_Ir

(200 mm) is used as a chromophore and Et₃N (2% v/v) as a sacri-

ficial electron-donor reagent, reveal the formation of nanopar-

ticles that increase in size during the reaction (from 1 nm at

30 min irradiation to 100 nm after 3 h of irradiation) (Fig-

ure SI.17 in the Supporting Information). Since no clear evi-

dence of nanoparticles were found during the first 30 min of

concentrations in the 10⁹order of magnitude, which again

imply an unrealistic TOF. DLS and NTA experiments suggest

that the loss of the catalytic activity overtime may be related

with the decomposition of the molecular system to form nano-

particles, which is in agreement with the ꢀ0.7 reaction order

obtained for the 1_Co.

Finally, following a reviewer suggestion, to investigate the

effect on the catalytic activity of possible nanoparticles formed

during the reaction was evaluated by poisoning experiments

with a large excess of Hg⁰(1000 equiv or even 1.5 mL which

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

6

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Paper

corresponds to the 15% of the total reaction volume) added

5 min before starting the irradiation. Experiments in triplicate

were carried out under standard conditions (1_Co(50 mm), PS_Ir

(250 mm), CH₃CN/H₂O/Et₃N (3:7:0.2 mL)) and with and without

Hg⁰for comparison reasons (see Figure SI.17 in the Supporting

Information). The results evidenced the same catalytic behav-

iour of H₂production independently of the amount of Hg⁰

used, within the experimental error. Furthermore, we also ex-

amined the effect of addition of Hg⁰during the reaction (2ꢂ

0.5 mL, 10% v/v), which did not produce any significant effect

on the time-profile nor on the hydrogen production TOF

(see Figure SI.17 in the Supporting Information). Arguably,

nanoparticles are unlikely to be responsible for the observed

hydrogen evolution activity; however, these results could also

indicate that the formation of nanoparticles may be related

with the catalyst decomposition and should be considered.

Table 1. Redox potentials for the studied complexes 1_Fe, 1_Coand 1_Niin

anhydrous CH₃CN, and CH₃CN/H₂O mixtures. Glassy carbon and SCE were

the working and reference electrodes, scan rate=100 mVs^ꢀ1. Potentials

are quoted versus SCE.

CH₃CN^[a]

CH₃CN/H₂O

(3:7)^[c]

E [V]

M^II/I

CH₃CN/H₂O/Et₃N

(3:7:0.2)^[d]

E [V]

M^II/I

E [V]

M^I/0

E [V]

M^III/II

h_TFA

[V]^[b]

M^II/I

1_Fe

1_Co

1_Ni

ꢀ1.68

ꢀ1.49

–

ꢀ1.38

ꢀ1.10

ꢀ1.07

1.14

0.72

1.75

0.87

0.59

0.56

ꢀ1.58

ꢀ1.23

ꢀ1.20

ꢀ1.63

ꢀ1.34

ꢀ1.21

[a] CV measured in CH₃CN containing Bu₄NPF₆(0.1m) as supporting elec-

trolyte (SE). [b] H₂reduction overpotentials (h) were calculated consider-

ing the protons reduction of TFA as proton source, which is E^o

=

(TFA/H₂)

ꢀ0.51 V (see section SI.3.1 in the Supporting Information). [c] CVs mea-

sured in CH₃CN/H₂O (3:7) containing KNO₃(0.1m) as SE. [d] CVs measured

in CH₃CN/H₂O/Et₃N (3:7:0.2) containing KNO₃(0.1m) as SE.

Electrochemical proton reduction to H₂

plexes is the expected and can be understood from the gener-

al increase in ionisation energies from left to right across

a period in the d block.

To get further insight into the capacity of metal complexes 1_Fe,

1_Coand 1_Niin the production of H₂, we explored the reduction

of protons under electrochemical conditions. First, redox po-

tentials were determined for 1_Fe, 1_Coand 1_Ni(depicted in

Figure 8) by using a glassy carbon as working and a saturated

The electrocatalytic proton reduction capacity of the three

complexes was studied in acetonitrile after addition of increas-

ing amounts of trifluoroacetic acid (TFA, pK_a=12.7 in acetoni-

trile)^[22]as the proton source. The CV of all three 1_mcomplexes

presents the appearance of new irreversible cathodic waves

upon the addition of increasing amounts of TFA triggers

(Figure 9). The grown of the reduction peak versus increase of

acid concentration is a clear indication of electrocatalytic be-

haviour of proton reduction to H₂. At low acid concentrations

the shape of the catalytic wave was similar to the wave of the

respective complex in the absence of acid; however, at higher

acid concentrations the electrocatalytic wave becomes broader

and more intense.

In all cases the onset of the catalytic wave matches the po-

tential of the M^II/Iredox couple, and therefore this is consistent

with M^Ispecies as the active ones in electrocatalysis through

the formation of M^IIIꢀH by protonation.^[11,23]The proton reduc-

tion overpotential (h) for complexes 1_Co(590 mV) and 1_Ni

(560 mV) turned to be considerable lower than for 1_Fe

(870 mV) (Table 1).

Figure 8. CVs of 1_Fe, 1_Coand 1_Ni. Experimental conditions: Bu₄NPF₆(0.1m) as

supporting electrolyte in acetonitrile, using a glassy carbon as working and

SCE as reference electrodes. Scan rate of 100 mVs^ꢀ1

.

calomel (SCE) as reference electrode in CH₃CN containing

Bu₄NPF₆(0.1m) as supporting electrolyte. A summary of the

M^I/0, M^II/Iand M^III/IIreduction potentials is presented in Table 1.

Complex 1_Fepresents the lowest reduction potential M^II/I

from the series (ꢀ1.38 V), and therefore it is the less favoured

to be reduced to oxidation state (I). In both cases 1_Coand 1_Ni

showed similar reduction potentials for the M^II/Icouple (ꢀ1.10

and ꢀ1.07 V, respectively). The CV of the same complexes in

the solvent mixture used in the photocatalysed H₂formation

(CH₃CN/H₂O 3:7) experiments exhibit a reduction in the redox

potentials for the M^II/Icouple in the range of 100–200 mV. This

is slightly accentuated by the basification of the media after

the addition of Et₃N. We note that the ligand alone does not

show any reduction wave above ꢀ1.8 V. The trend found in

the relative reduction potential M^II/Iamong the three com-

The dependence of the peak intensity versus the concentra-

tion of acid gives valuable information about the H₂formation

mechanism. In this regard, we studied the i_cat/i_pratio as a func-

tion of the TFA concentration in a 1 mm solution of 1_Co, 1_Fe

and 1_Niin acetonitrile, (in which i_catstands for the peak current

in the presence of acid and i_pfor the peak current in the ab-

sence of acid). As shown in Figure 9, the i_cat/i_pratio of currents

of 1_Coincreases with the acid concentration until it becomes

independent versus the concentration of protons. Below a TFA

concentration of 100 mm the i_cat/i_pversus [TFA] is linear and

consistent with a second-order process on [H⁺] (Figure 9). At

a TFA concentration higher than 100 mm, i_cat/i_pit is independ-

ent of the [TFA] (Figure 10), suggesting that under these condi-

tions the rate-limiting step is an intramolecular proton transfer

or an elimination of H₂.

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

7

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

Figure 10. i_cat/i_pratio as a function of the [TFA] in a 1 mm concentration so-

lution of 1_Fe, 1_Coor 1_Ni. Scan rate=100 mVs^ꢀ1. Glassy carbon working elec-

trode and SCE as reference. i_cat/i_pvalues were measured at the M^II/Iredox po-

tential.

face of the working electrode, a new CV was recorded form

a freshly prepared TFA (100 mm) solution, which did not con-

tained 1_Co(Figure SI.20 in the Supporting Information). In this

case, the electrocatalytic wave completely disappeared. Conse-

quently, the i_cat/i_pvalues obtained cannot be attributed to elec-

trodeposition of cobalt species.

The good electrochemical behaviour of 1_Cofacilitates the cal-

culation of the electrocatalytic rate constant for hydrogen pro-

s

duction, k_e(95m^{ꢀ2 ꢀ1}) and the corresponding turnover fre-

quency (TOF=k[H⁺]²=3420 mol(H₂)mol(1_Co)^ꢀ1h^ꢀ1) by plotting

the slopes obtained of i_cat/i_pversus [TFA] at different scan rates

(n) as a function of n^ꢀ1/2(Figures 10 and 11 and Figures SI.21–

SI.30 in the Supporting Information).^[11]

Figure 9. Successive CVs of 1_Fe(top), 1_Co(middle) and 1_Ni(bottom) (1 mm) in

acetonitrile containing Bu₄NPF₆(0.1m) in the presence of TFA (0–60 mm).

Scan rate=100 mVs^ꢀ1, glassy carbon as working electrode. Insets: i_cat/i_p

versus [TFA].

On the other hand, complexes 1_Feand 1_Nihave similar be-

haviours at low acid concentration (up to 40 mm for 1_Fe, and

60 mm for 1_Ni), while at higher acid concentration the i_cat/i_pdra-

matically decreases, which may be due to catalyst degradation

provoked by the high concentration of protons in the reaction

medium.

To ensure that the electrocatalytic behaviour of 1_Cocorre-

sponds to a molecular catalytic activity and thus discarding an

electrodeposition process over the surface of the glassy

carbon electrode, we first measured the CV of 1_Co(1 mm) in

the presence of TFA (100 mm), and then after rinsing the sur-

Figure 11. Top: plots of i_cat/i_pversus [TFA] for a 1 mm solution of 1_Coat differ-

ent scan rates. Bottom: Plot of the i_cat/i_pslopes at different scan rates versus

[TFA] as function of the n^ꢀ1/2

.

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

8

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Paper

Mechanistic considerations

Complexes 1_m(M=Co, Ni and Fe) are all electrocatalysts for H₂

production, while only 1_Cois highly active under photochemi-

cal conditions. Under the latter conditions, 1_Nipresents much

reduced activity and 1_Feis completely inactive. The inactivity of

1_Feunder photochemical conditions could be attributed to the

low redox potential observed that it is needed for the forma-

tion of the Fe^Ispecies. This was shown by the reaction of the

II

in situ generated PS_Irwith 1_mcomplexes. Complexes 1_Niand

II

1_Coinstantaneously react with PS_Ir, while 1_Feapparently is

II

mainly inert to PS_Ir. This was also consistent with the M^II/Ire-

Figure 12. Successive CVs at increasing [1_Co]. Conditions: [TFA]=100 mm,

Supporting electrolyte: Bu₄NPF₆(0.1m), scan rate=100 mVs^ꢀ1, GC working

electrode. Inset: Electrocatalytic current versus [1_Co] (0.1–0.8 mm) showing

a first order kinetics.

duction potential of 1_Fein acetonitrile and the observed elec-

trocatalytic H₂production, which only occurs at very low po-

tentials. However, 1_Nipresents similar M^II/Ireduction potential

and overpotential in electrocatalytic H₂formation as 1_Coand

II

In addition, the dependence of the catalytic current on the

concentration of catalyst at sufficiently high TFA concentration

(100 mm) was found to be first-order (Figure 12). In summary,

these studies lead us to formulate an overall rate law for the

production of hydrogen of v=k_e[H⁺]²[1_Co].

Finally, complex 1_Cowas soluble and stable in pure aqueous

solution. The cyclic voltammograms in water showed a redox

couple Co^II/Iat ꢀ1.29 V versus SCE. This is slightly shifted to

lower potentials in comparison with the one obtained in

CH₃CN (ꢀ1.10 V vs. SCE). The addition of TFA to a 1 mm solu-

tion of 1_Cocauses an increase of the electrocatalytic current

near the Co^II/Icouple (Figure 13) comparable to the previously

moreover reacts instantaneously with in situ generated PS_Ir.

Therefore, other limitations should be considered for the low

H₂production under photochemical conditions exhibited by

1_Ni. One must consider that electro- and photochemical condi-

tions are rather different. For instance, the electrochemical

studies were performed in organic solvents or water, but eval-

uating the electrocatalytic current based on the [H⁺], usually

provided by the addition of an strong acid (TFA in our case).

This acid supply sets up the medium to a low pH. In contrast,

in photochemical reactions the electrons are provided by the

incorporation of a sacrificial agent (Et₃N in our case). This

makes the reaction medium highly basic (pHꢁ12). The large

pH differences between electro- and photochemical condi-

tions, makes pH as one of the factor which diverged the H₂

formation activity among complexes and conditions (1_Niand

1_Co). Electro- and photochemical experiments provided evi-

dence that under these conditions both 1_Ni,Coare reduced by

one electron to form M^Ispecies. Then, subsequent protonation

of those species will generate M^IIIꢀH intermediates and further

protonation will produce H₂. Alternatively, under certain condi-

tions M^IIIꢀH intermediates may not be enough reactive to pro-

duce H₂(for instance under basic conditions) and will first re-

quire an additional reduction by one electron to form M^IIꢀH,

which then by protonation produce H₂.^[12c]

To provide some light on the reactivity differences among

the 1_mseries of complexes, structures of M^I, M^II, M^IꢀL, M^IIꢀL

(L=H₂O or CH₃CN) and M^IIIꢀH species were DFT modelled with

the Gaussian 09 program^[24]using the B3LYP hybrid exchange-

correlation functional and taking into account the effect of

water solvation and London dispersion effects (see the Sup-

porting Information section SI.4 for complete details of the

employed methodology). Theoretical efforts focused on the

evaluation of the solvent coordination, the M^II/Ireduction po-

tentials and the pK_avalues of hydride species to explore the

energetic accessibility of the proposed intermediates as well as

their relative stability. The results are summarised in Tables 2

and 3.

Figure 13. Successive CVs of 1_Co(1 mm) in water containing KNO₃(0.1m)

and TFA acid: from bottom to top 0, 1, 3, 10, 20, 40, and 60 mm. Scan

rate=100 mVs^ꢀ1, glassy carbon as working electrode.

observed in CH₃CN. Remarkably, the i_cat/i_pin water (65 at

60 mm) was found to be significantly higher than in acetoni-

trile (48 at 60 mm). The i_pwas measured in a non-pH-buffered

aqueous solution of KNO₃(0.1m, pH6) and therefore a part of

the i_pvalue could be partly associated with H₂generation ac-

tivity. Consequently, 1_Cois not only stable in water, but also is

more electrocatalytically active for proton reduction to H₂than

in pure acetonitrile.

Under photochemical conditions (CH₃CN/H₂O 3:7 solvent

mixture) the metal centre may interact with both solvent mole-

cules and the redox activity of the resulting aqua or acetoni-

trile complexes could be substantially different. Therefore, the

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

9

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

in the range of 11–12. Therefore, these results suggest that the

protonation is disfavoured, thus blocking the H₂production. In

contrast, the reduction potential of ꢀ1.42 V for the Fe^II/Icouple

agree with the electrochemically observed of ꢀ1.58 V, reveal-

ing that the Fe^Ispecies may not be photochemically generat-

ed. This situation bans the possible formation of Fe^IIIꢀH species

(pK_a1.8, also disfavoured), and then the capacity to photopro-

duce H₂when using PS_Iras photosensitiser (Scheme 2).

Table 2. Computed free energy (DG) for coordination of CH₃CN and H₂O

to 1_mand for ligand exchange reactions.^[a]

DG [kcalmol^ꢀ1

]

Fe^I

Fe^II

Co^I

Co^II

Ni^I

Ni^II

[M^II/I(Py₂^Tstacn)]+L![M^II/I(Py₂^Tstacn)(L)]

L=CH₃CN

L=H₂O

2.9

2.2

3.4

1.8

ꢀ0.2

ꢀ1.1

5.5

9.6

ꢀ5.1

ꢀ1.9

3.9

4.1

CH₃CN+[M^II/I(Py₂^Tstacn)(H₂O)]$[M^II/I(Py₂^Tstacn)(CH₃CN)]+H₂O

ꢀ0.6

1.6

ꢀ3.8

ꢀ5.1

ꢀ4.1

ꢀ3.2

[a] All values have been obtained using the ground spin state for each

structure.

Table 3. DFT computed reduction potentials [V] and pK_avalues of the

proposed cobalt, nickel and iron intermediates.^[a]

M=Fe

M=Co

M=Ni

[b]

E^o(M^II/I

)

ꢀ1.42

ꢀ0.86

ꢀ0.97

ꢀ0.4

pK_a(M^IIIꢀH)^[c]

1.8

7.0

[a] All values were obtained using the ground spin state for each struc-

ture. E^ovalues correspond to the standard reduction potentials relative to

the SCE electrode. [b] Complexes without solvent coordination were em-

ployed for the calculation. [c] pK_avalues were calculated at 298 K.

identification of the most stable aqua, acetonitrile or solvent-

free intermediate for each metal oxidation state is necessary to

ensure a correct description of the initial redox step. In gener-

al, the energies gained by coordination of water or acetonitrile

to the metal complexes in oxidation state II were only small or

even disfavoured for 1_Fe. Binding of water to 1_Cowas also ener-

getically disfavoured. This trend is clearer for the oxidation

state I. With some exceptions, the energies fall in the limits of

the average error of the used methodology (Table 2). We can

therefore conclude that the absence of solvent ligand coordi-

nated is generally favoured for oxidation state I in both water

and acetonitrile and also favoured for oxidation state II in

water.

The reduction potentials M^II/Ifor 1_Co, 1_Niand 1_Fewithout sol-

vent bonded were calculated to be ꢀ0.86, ꢀ0.97 and ꢀ1.42 V

(SCE), respectively, in qualitative agreement with the experi-

mental reduction potentials of ꢀ1.23, ꢀ1.20 and ꢀ1.58 V

(SCE), respectively. Although these values are 150–250 mV

lower than the experimental results, they still fall in the usual

range of the method accuracy.

Scheme 2. Differences in reactivity among 1_Fe, 1_Coand 1_Niin single-electron

reduction process to M^Ispecies and the subsequent protonation to yield

the M^III-H species under photochemical conditions. pK_avalues have been ob-

tained by DFT calculations.

Conclusion

We synthesised and characterised a new ligand 1,4-di(picolyl)-

7-(p-toluenesulfonyl)-1,4,7-triazacyclononane (Py₂^Tstacn) and

the corresponding metal complexes 1_Fe, 1_Coand 1_Ni. All

showed to be good electrochemical catalysts for proton reduc-

tion to H₂. However, only 1_Cois an efficient catalyst for water

reduction when using [Ir(ppy)₂(bpy)]PF₆as photosensitiser and

Et₃N as electron donor. In contrast, under the same conditions

1_Nipresents low photochemical water reduction activity and

1_Feis inactive. More interestingly, to some extent the catalyst

1_Cocan operate as photocatalyst for H₂production in presence

of O₂.

Experimental and computational results determine that both

1_Coand 1_Nicomplexes form species with a M^Ioxidation state

upon reacting with the reduced PS_Ir. Consequently, the origin

of the photochemical inactivity of nickel species for H₂produc-

tion may be due to other factors such as the inaccessibility of

the [Ni^IIIꢀH] species under basic pH, since under electrocatalyt-

ic acidic pH conditions the H₂evolution reaction does take

place. In this regard, the DFT calculated pK_avalues for [Co^IIIꢀH]

and [Ni^IIIꢀH] were 7.0 and ꢀ0.4, respectively. This clearly under-

lines the more acidic character of the [Ni^IIIꢀH]. We should note

that under photocatalytic conditions the pH of the mixture is

The active species during the photocatalytic water reduction

are molecular. This was clear by the lack of lag time in the H₂

evolution, and the absence of nanoparticles during the first

minutes of the reaction, when the system was most active.

Moreover, the formation of nanoparticles is ascribed to the re-

duction of catalytic activity.

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

10

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Paper

cle-size distribution from 1 to 2000 mm). The analyses were carried

out in the Institut de Ciꢆncia dels Materials de Barcelona.

Mechanistic investigations, electrochemical studies and DFT

calculations allowed us to elucidate the factors that control the

activity among the three complexes with the same aminopyri-

dine ligand. The Py₂^Tstacn ligand not only controls the relative

stabilisation of the metal oxidation state, but also the basicity.

Both factors are important and should be considered for the

boundary reaction conditions imposed by the photosensitiser,

electron donor and pH. Indeed, the use of [Ir(ppy)₂(bpy)]PF₆as

X-ray crystallography

Single crystals of 1_Coand 1_Niwere mounted on a nylon loop for X-

ray structure determination. The measurements were carried out

on a BRUKER SMART APEX CCD diffractometer using graphite-

monochromated Mo_Karadiation (l=0.71073 ꢄ). Programs used:

data collection, Smart version 5.631 (Bruker AXS 1997-02); data re-

duction, Saint+ version 6.36A (Bruker AXS 2001); absorption cor-

rection, SADABS version 2.10 (Bruker AXS 2001). Structure solution

and refinement was done using SHELXTL Version 6.14 (Bruker AXS

2000–2003). The structure was solved by direct methods and re-

fined by full-matrix least-squares methods on F². The non-hydro-

gen atoms were refined anisotropically. The hydrogen atoms were

placed in geometrically optimised position and forced to ride on

the atom to which they are attached.

II

photosensitiser allows the one-electron reduction of 1_Coand

1_Ni^II, but not in the case of 1_Fe^II. On the other hand, the use of

Et₃N as electron donor basifies the media making the protona-

tion of the M^Imetal centre for 1_Nidifficult and, therefore, the

H₂evolution. Altogether, this study gives light on important

factors to be considered for the development of more efficient

first-row transition-metal catalysts for water reduction.

CCDC-957199 (1_Co) and CCDC-957198 (1_Ni) contain the supplemen-

tary crystallographic data for this paper. These data can be ob-

tained free of charge from The Cambridge Crystallographic Data

Centre via www.ccdc.cam.ac.uk/data_request/cif.

Experimental Section

General methods

All procedures were carried out under N₂using standard vacuum

line, Schlenk, and inert atmosphere glovebox techniques. Reagents

and solvents were purchased from commercial sources as used as

received unless otherwise stated. Triethylamine (Et₃N) ꢂ99% purity

and ascorbic acid (ꢂ99%) were purchased from Sigma–Aldrichꢅ

and used without further purification. 1-(p-Toluensulfonyl)-1,4,7-tri-

Gas-evolution monitoring studies

All catalytic reactions were performed in a 15 mL vial caped with

a septum. Each experiment was conducted in a volume-calibrated-

vial equipped with stir-bars and containing the solvent mixture

with the reagents, and then was connected to one port of a differ-

ential pressure transducer sensor (Honeywell-ASCX15DN, ꢃ15 psi).

Each reaction had its own reference reaction, which was conduct-

ed at the other port of the differential pressure transducer sensor.

The reaction and reference vials were kept under the same experi-

mental conditions to minimise the system noise due to tempera-

ture–pressure fluctuations. In this sense, each vial was submitted

and located just over a LED radiation source. Furthermore, in order

to ensure a constant and stable irradiation, the LED sources were

equipped with a water refrigeration system.

azacyclononane (^Tstacn),^[25]and [Ir(bpy)(ppy)₂]PF₆were synthes-

[26]

ised according to the literature procedures. Anhydrous acetonitrile

was purchased from Scharlab. Water (18.2MWcm) was purified

with a Milli-Q Millipore Gradient AIS system. All the solvents were

strictly degassed and stored in anaerobic conditions. All water re-

duction catalytic reactions were performed under N₂otherwise no-

tified.

Physical methods

NMR spectra were recorded on Bruker spectrometers operating at

300 or 400 MHz as noted at 300 K. All H chemical shifts are report-

Just before starting the reaction, the corresponding catalyst was

added by syringe through the septum from a stock solution pre-

pared in acetonitrile. The reaction began when the LEDs were

turned on. At this point, the hydrogen evolved from the reactions

was monitored by recording the increase in pressure of the head-

space (1 s interval). The pressure increment was the result of the

difference in pressure between the reaction and reference vials.

After the hydrogen evolution reached a plateau the amount of the

formed gas was captured and measured, equilibrating the pressure

between reaction and reference vials. The hydrogen contained in

each of the reactions was measured by analysing an aliquot of gas

of the headspace (0.2 mL) by gas chromatography. GC measure-

ments of H₂corroborated the values obtained by pressure incre-

ments.

1

ed in ppm and were internally calibrated to the monoprotio impur-

ity of the deuterated solvent. The coupling constants are measured

in Hz. The ¹³C chemical shifts were internally calibrated to the

carbon atoms of the deuterated solvent. All experimental proce-

dures were conducted at ambient temperature under nitrogen at-

mosphere and using degassed acetonitrile. A standard three-elec-

trode configuration was employed in conjunction with CH Instru-

ments potentiostat interfaced to a computer with CH Instruments

600D software. Using a one-compartment cell, all voltammetric

scans were recorded using glassy carbon working electrode, which

was treated between experiments by means of a sequence of pol-

ishing with MicroPolish Powder (0.05 micron) before washing and

sonification. A saturated calomel electrode (SCE) and Pt wire were

used as reference and counter electrodes, respectively. All materials

for electrochemical experiments were dried at room temperature

before use. UV/Vis spectra were recorded on an Agilent 8453

diode array spectrophotometer (190–1100 nm range) in 1 cm

quartz cells. A cryostat from Unisoku Scientific Instruments was

used for the temperature control. Electrospray ionisation mass

spectrometry (ESI-MS) experiments were performed on Bruker Dal-

tonics Esquire 3000 Spectrometer, by introducing samples directly

into the ESI-source using a syringe. DLS and NTA experiments were

performed on a Zetasizer Nano ZS, Malvern Instruments (particle-

size distribution from 0.6 to 6.000 nm) and Nanosight LM20 (parti-

Gas-evolution studies performed in the presence of Hg⁰

Catalytic reactions performed in the presence of Hg⁰were carried

out in a 15 mL vial capped with a septum. Each experiment was

conducted in a volume-calibrated-vial equipped with stir-bars and

containing the CH₃CN/H₂O (3:7) as the solvent mixture with all the

reagents under N₂. The experiments were conducted in one port

of a differential pressure transducer sensor, with a reference reac-

tion at the other port. The reaction and reference vials were kept

under the same experimental conditions to minimise the system

noise due to temperature-pressure fluctuations. Before starting the

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

11

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

reaction Hg⁰was added in large excess (1000 equiv or 1.5 mL; 15%

of the total reaction volume) by syringe though the septum. The

solution was stirred for 5 min under and then the solution was

submitted to LED irradiation at 447 nm. Hg⁰was left in contact

with the solution during the whole catalytic assay in order to

poison the heterogeneous catalyst formed during the irradiation

time. The hydrogen evolved from the reaction was monitored in

real time by recording the increase in pressure of the headspace

(1 s interval). The pressure increment was the result of the differ-

ence in pressure between the reaction and reference vials. After

the hydrogen evolution reached a plateau the amount of the

formed gas was captured and measured, equilibrating the pressure

between reaction and reference vials. The hydrogen contained in

each of the reactions was measured by analysing an aliquot of gas

of the headspace (0.2 mL) by gas chromatography. GC measure-

ments of H₂corroborated the values obtained by pressure incre-

ments.

Catalyst concentration dependence study

TFA was added to a solution of 0.1m Bu₄NPF₆in CH₃CN in an elec-

trochemistry cell to provide an acid concentration of 100 mm. To

this solution, aliquots of 20 mm stock solution of 1_Cowere added

and CVs were collected. The order of the proton reduction with re-

spect to [catalyst] was determined by plotting the current at the

potential M^II/IIversus the concentration of catalyst.

Syntheses

Synthesis of 1,4-di(picolyl)-7-(p-toluenesulfonyl)-1,4,7-triazacy-

clononane (Py₂^Tstacn): 2-Pycolyl chloride hydrochloride (1.16 g,

7.04 mmol), ^Tstacn (1 g, 3.53 mmol) and anhydrous acetonitrile

(40 mL) were mixed in a 100 mL flask. Na₂CO₃(2.05 g) and tetrabu-

tylammonium bromide (80 mg) were added directly as solids and

the resulting mixture was heated at reflux under N₂for 22 h. After

cooling to room temperature, the resulting orange mixture was fil-

tered and the filter cake was washed with CH₂Cl₂. The combined

filtrates were evaporated under reduce pressure. NaOH (2m,

15 mL) was added to the resulting residue, and the mixture was

extracted with CH₂Cl₂(4ꢂ40 mL). The combined organic layers

were dried over MgSO₄and the solvent was removed under re-

duced pressure. The resulting residue was treated with n-hexane

(100 mL) and stirred for 12 h. A fine white solid appeared which

was filtered off and dried under vacuum to yield 1.34 g of the de-

Parallel pressure transducer hardware

The parallel pressure transducer is composed by eight differential

pressure transducers (Honeywell-ASCX15DN, ꢃ15 psi) connected

to a hardware data-acquisition system (base on Atmega microcon-

troller) controlled by a home-developed software program. The dif-

ferential pressure transducer Honeywell-ASCX15DN was set with

a

100 ms response, signal-conditioned (high-level span, 4.5 V)

1

output, and a calibrated and temperature compensated (0–708C)

sensor. The differential sensor had two sensing ports that could be

used for differential pressure measurements. The pressure-calibrat-

ed devices, to within ꢃ0.5 matm, was offset and span calibrated

by using software for a high-precision pressure transducer (PX409-

030GUSB, 0.08% accuracy). Each of the eight differential pressure

transducers (Honeywell-ASCX15DN, ꢃ15 psi) produced voltage

outputs that could be directly transformed to a pressure difference

between the two measuring ports. The voltage outputs were digi-

talised with a resolution of 0.25 matm from 0 to 175 matm and

1 matm from 176 to 1000 matm by using an Atmega microcontrol-

ler with an independent voltage auto-calibration. Firmware

Atmega microcontroller and control software were home-devel-

oped. The sensitivity of H₂analytics allowed the quantification of

the gas formed when low H₂volumes were generated. Therefore,

it could not be discarded that small amounts of H₂were produced

by inactive complexes.

sired product (2.88 mmols, 82%). H NMR (CDCl₃, 300 MHz, 300 K):

d=8.50 (d, J=4.8 Hz, 2H; H₂of py), 7.66–7.60 (m, 4H; H_Tsand H₄

of py), 7.47 (d, J=7.8 Hz, 2H; H_Ts), 7.26 (d, 2H; J=8.4 Hz, H₅of py),

7.15–7.11 (m, 2H; H₃of py), 3.86 (s, 4H; CH₂-py), 3.22 (m, 4H; N-

CH₂-CH₂), 3.12 (m, 4H; N-CH₂-CH₂), 2.79 (s, 4H; N-CH₂-CH₂),

2.40 ppm (s, 3H; CH₃). ¹³C NMR (CDCl₃, 300 MHz, 300 K): d=159.75,

149.00, 142.99, 136.40, 135.94, 129.60, 127.12, 123.23, 121.94,

63.84, 55.85, 50.87, 21.47 ppm; ESI-MS: m/z: 466.2 [M+H]⁺; ele-

mental analysis calcd (%) for C₂₅H₃₁N₅O₂S·0.5H₂O: C 63.27, H 6.80,

N 14.76; found: C 63.47, H 6.68, N 13.91.

Synthesis of [Fe(OTf)(Py₂^Tstacn)](OTf) (1_Fe): In a glovebox, a solu-

tion of Fe(OTf)₂·2CH₃CN (235 mg, 0.539 mmol) in anhydrous THF

(1 mL) was added dropwise to a vigorously stirred solution of

Py₂^Tstacn (251.2 mg, 0.539 mmol) in THF (1 mL). After few minutes,

the solution become cloudy and a pale yellow precipitate ap-

peared. After stirring for an additional 5 h the solution was filtered

off and the resulting solid dried under vacuum. This solid was dis-

solved with CH₂Cl₂, and the slow diffusion of diethyl ether over the

resultant solution afforded, in few days, 340 mg of a pale yellow

solid (0.414 mmol, 76.9%).¹H NMR (CD₃CN, 400 MHz, 300 K): d=

56.62, 11.18, 8.42, 7.88, 3.45, 2.17, 1.12 ppm; HR-ESI-MS: m/z:

Gas chromatography identification and quantification of

gases

670.1069 [MꢀOTf]⁺, 260.5804 [Mꢀ2OTf]²⁺

; elemental analysis

calcd (%) for C₂₇H₃₁F₆FeN₅O₈S₃·0.5CH₂Cl₂: C 38.21, N 8.21, H 3.74;

Hydrogen at the headspace was analysed with an Agilent 7820 A

GC System equipped with columns Washed molecular sieves 5A,

2 mꢂ1/8’’ OD, Mesh 60/80 SS and Porapak Q, 4 mꢂ1/8’’ OD, SS.

Mesh: 80/100 SS and a thermal conductivity detector. The H₂

amount obtained was calculated through the interpolation of the

previous calibration using different H₂/N₂mixtures.

found: C 38.54, N 8.27, H 3.73.

Synthesis of [Co(OTf)(Py₂^Tstacn)](OTf) (1_Co): In a glovebox, a solu-

tion of Co(OTf)₂·2CH₃CN (236 mg, 0.537 mmol) in anhydrous THF

(1 mL) was added dropwise to a vigorously stirred solution of

Py₂^Tstacn (250 mg, 0.537 mmol) in THF (1 mL). After few minutes,

the solution become cloudy and a pale pink precipitate appeared.

After stirring for an additional 2 h the solution was filtered off and

the resulting solid dried under vacuum. This solid was dissolved in

CH₂Cl₂and the slow diffusion of diethyl ether into this solution

produced 355 mg of pink crystals (0.432 mmol, 83%). ¹H NMR

(CD₃CN, 300 MHz, 300 K): d=118.28, 55.72, 3.92, 3.76, 2.22,

9.42 ppm; HR-ESI-MS: m/z: 673.1054 [MꢀOTf]⁺, 262.0766

[Mꢀ2OTf]²⁺; elemental analysis calcd (%) for C₂₇H₃₁F₆CoN₅O₈S₃: C

39.42, N 8.51, H 3.80; found: C 39.23, N 8.56, H 3.68.

Acid concentration dependence study

A 0.8m stock solution of TFA was prepared in a CH₃CN solution of

0.1m Bu₄NPF₆. Aliquots of this solution were added to a degassed

solution of 1 mm catalyst in CH₃CN. After the addition the solution

was purged with N₂before carrying out the cyclic voltammetry.

The order with respect to [acid] was determined by plotting the

current at the potential M^II/Iversus the concentration of acid.

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

12

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Paper

Synthesis of [Ni(OTf)(Py₂^Tstacn)](OTf) (1_Ni): In a glovebox, a solution

of Ni(OTf)₂·2CH₃CN (205.9 mg, 0.429 mmol) in anhydrous THF

(1 mL) was added dropwise to a vigorously stirred solution of

Py₂^Tstacn (200 mg, 0.429 mmol) in THF (1 mL). After few minutes,

the solution become cloudy and a purple precipitate appeared.

After stirring for an additional 4 h the solution was filtered off and

the resulting solid dried under vacuum. This solid was dissolved in

CH₂Cl₂and the slow diffusion of diethyl ether into this solution

[9] a) C. Bachmann, M. Guttentag, B. Spingler, R. Alberto, Inorg. Chem.

2013 , 52 , 605 5 –6 061; b) M. Nippe, R. S. Khnayzer, J. A. Panetier, D. Z.

Zee, B. S. Olaiya, M. Head-Gordon, C. J. Chang, F. N. Castellano, J. R.

Long, Chem. Sci. 2013 , 4 , 393 4 –3 945.

[10] a) Y. Sun, J. P. Bigi, N. A. Piro, M. L. Tang, J. R. Long, C. J. Chang, J. Am.

Chem. Soc. 2011 , 133 , 921 2 –9 21 5; b) W. M. Singh, T. Baine, S. Kudo, S.

Tian, X. A. N. Ma, H. Zhou, N. J. DeYonker, T. C. Pham, J. C. Bollinger, D. L.

Baker, B. Yan, C. E. Webster, X. Zhao, Angew. Chem. 2012 , 124 , 604 3 –

6046; Angew. Chem. Int. Ed. 2012 , 51 , 594 1 –5 944.

1

produced 285 mg of purple crystals (0.346 mmol, 80.65%). H NMR

[11] J. P. Bigi, T. E. Hanna, W. H. Harman, A. Chang, C. J. Chang, Chem.

Commun. 2010 , 46 , 95 8 –9 60.

(CD₃CN, 400 MHz, 300 K): d=56.41, 52.13, 45.61, 15.46, 14.89, 9.21,

6.46, 5.70, 2.17 ppm; HR-ESI-MS: m/z: 672.1058 [MꢀOTf]⁺,

[12] a) M. Guttentag, A. Rodenberg, C. Bachmann, A. Senn, P. Hamm, R. Al-

berto, Dalton Trans. 2013 , 42 , 33 4 –3 37; b) A. E. King, Y. Surendranath,

N. A. Piro, J. P. Bigi, J. R. Long, C. J. Chang, Chem. Sci. 2013 , 4 , 157 8 –

1587; c) Y. Sun, J. Sun, J. R. Long, P. Yang, C. J. Chang, Chem. Sci. 2013 ,

4 , 11 8 –1 24; d) P. Zhang, M. Wang, F. Gloaguen, L. Chen, F. Quentel, L.

Sun, Chem. Commun. 2013 , 49 , 945 5 –9 457; e) M. Singh, M. Mirmo-

hades, R. Jane, T. White, L. Hammarstrom, A. Thapper, R. Lomoth, S. Ott,

Chem. Commun. 2013 , 49 , 863 8 –8 640.

261.5797 [Mꢀ2OTf]²⁺

;

elemental analysis calcd (%) for

C₂₇H₃₁F₆NiN₅O₈S₃·0.5CH₂Cl₂: C 38.19, N 8.10, H 3.73; found: C 38.17,

N 8.15, H 3.51.

Acknowledgements

[13] J. Lloret-Fillol, Z. Codola, I. Garcia-Bosch, L. Gꢉmez, J. J. Pla, M. Costas,

Nat. Chem. 2011 , 3 , 80 7 –8 13.

We would like to thank Dr. N. Rodriguez, Dr. X. Ribas, Dr.

Josep M. Luis and Dr. Miquel Costas for their comments, the

ERC (grant no. FP7-PEOPLE-2010-ERG-268445) (J.Ll.-F.), MICINN

(grant no. CTQ2012–37420-C02-01/BQU), for the RyC contract

(J.Ll.-F.) and INNPLANTA project (INP-2011-0059-PCT-420000-

ACT1) to Dr. X. Ribas. The authors thank A. Bernabꢁ (ICMAB-

CSIC/CIBER-BBN) for his help in particle size measurements

with Zetasizer Nano and Nanosight LM20.

[14] a) J. L. Dempsey, B. S. Brunschwig, J. R. Winkler, H. B. Gray, Acc. Chem.

Res. 2009 , 42 , 199 5 –2 00 4; b) B. H. Solis, S. Hammes-Schiffer, J. Am.

Chem. Soc. 2011 , 133 , 1903 6 –1 903 9; c) S. C. Marinescu, J. R. Winkler,

H. B. Gray, Proc. Natl. Acad. Sci. USA 2012 , 109 , 1512 7 –1 5131.

[15] C.-F. Leung, S.-M. Ng, C.-C. Ko, W.-L. Man, J.-S. Wu, L.-J. Chen, T.-C. Lau,

Energy Environ. Sci. 2012 , 5 , 790 3 –7 907.

[16] Y. Ohsawa, S. Sprouse, K. A. King, M. K. DeArmond, K. W. Hanck, R. J.

Watts, J. Phys. Chem. 1987 , 91 , 104 7 –1 054.

[17] C. R. Bock, J. A. Connor, A. R. Gutierrez, T. J. Meyer, D. G. Whitten, B. P.

Sullivan, J. K. Nagle, J. Am. Chem. Soc. 1979 , 101 , 481 5 –4 824.

[18] W. Zhang, J. Hong, J. Zheng, Z. Huang, J. S. Zhou, R. Xu, J. Am. Chem.

Soc. 2011 , 133 , 2068 0 –2 0683.

Keywords: cobalt · electrochemistry · homogeneous catalysis ·

hydrogen · photochemistry · water reduction

[19] F. Lakadamyali, M. Kato, N. M. Muresan, E. Reisner, Angew. Chem. 2012 ,

124 , 951 5 –9 518; Angew. Chem. Int. Ed. 2012 , 51 , 938 1 –9 384.

[20] M. C. DeRosa, R. J. Crutchley, Coord. Chem. Rev. 2002 , 23 3 –2 34 , 35 1 –

371.

[1] a) J. Barber, Chem. Soc. Rev. 2009 , 38 , 18 5 –1 96; b) J. Barber, P. D. Tran, J.

R. Soc. Interface 2013, 10, 20120984.

[21] V. Artero, M. Fontecave, Chem. Soc. Rev. 2013 , 42 , 233 8 –2 356.

[22] A. J. Bard, L. R. Faulkner, Electrochemical Methods: Fundamentals and Ap-

plications, 2 ed., Wiley, New York, 2001.

[23] O. Pantani, E. Anxolabꢁhꢆre-Mallart, A. Aukauloo, P. Millet, Electrochem.

Commun. 2007 , 9 , 5 4 –5 8.

[2] a) J. W. Peters, W. N. Lanzilotta, B. J. Lemon, L. C. Seefeldt, Science 1998 ,

282 , 185 3 –1 858; b) A. Volbeda, M.-H. Charon, C. Piras, E. C. Hatchikian,

M. Frey, J. C. Fontecilla-Camps, Nature 1995 , 373 , 58 0 –5 87; c) Y. Nicolet,

C. Piras, P. Legrand, C. E. Hatchikian, J. C. Fontecilla-Camps, Structure

1999 , 7 , 13.

[24] Gaussian 09, Revision A.1, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.;

Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.;

Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratch-

ian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada,

M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Per-

alta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.;

Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell,

A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam,

J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo,

J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;

Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski,

V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels,

A. D.; Farkas, ꢊ.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.

Gaussian, Inc., Wallingford CT, 2009.

[3] J. C. Fontecilla-Camps, A. Volbeda, C. Cavazza, Y. Nicolet, Chem. Rev.

2007 , 107 , 427 3 –4 303.

[4] C. Tard, C. J. Pickett, Chem. Rev. 2009 , 109 , 224 5 –2 274.

[5] a) V. Artero, M. Chavarot-Kerlidou, M. Fontecave, Angew. Chem. 2011 ,

123 , 737 6 –7 405; Angew. Chem. Int. Ed. 2011 , 50 , 723 8 –7 266; b) W. T.

Eckenhoff, R. Eisenberg, Dalton Trans. 2012 , 41 , 1300 4 –1 3021.

[6] a) M. L. Helm, M. P. Stewart, R. M. Bullock, M. R. DuBois, D. L. DuBois, Sci-

ence 2011 , 333 , 86 3 –8 66; b) Z. Han, F. Qiu, R. Eisenberg, P. L. Holland,

T. D. Krauss, Science 2012 , 338 , 132 1 –1 324; c) Z. Han, W. R. McNamara,

M.-S. Eum, P. L. Holland, R. Eisenberg, Angew. Chem. 2012 , 124 , 169 9 –

1702; Angew. Chem. Int. Ed. 2012 , 51 , 166 7 –1 670.

[7] a) F. Gꢇrtner, A. Boddien, E. Barsch, K. Fumino, S. Losse, H. Junge, D.

Hollmann, A. Brꢈckner, R. Ludwig, M. Beller, Chem. Eur. J. 2011 , 17 ,

642 5 –6 436; b) T. Matsumoto, H.-C. Chang, M. Wakizaka, S. Ueno, A. Ko-

bayashi, A. Nakayama, T. Taketsugu, M. Kato, J. Am. Chem. Soc. 2013 ,

135 , 864 6 –8 654; c) R. Lomoth, S. Ott, Dalton Trans. 2009 , 995 2 –9 959;

d) A. M. Kluwer, R. Kapre, F. Hartl, M. Lutz, A. L. Spek, A. M. Brouwer,

P. W. N. M. van Leeuwen, J. N. H. Reek, Proc. Natl. Acad. Sci. USA 2009 ,

106 , 1046 0 –1 046 5; e) Y. Na, M. Wang, J. Pan, P. Zhang, B. ꢄkermark, L.

Sun, Inorg. Chem. 2008 , 47 , 280 5 –2 810.

[25] C. Flassbeck, K. Z. Wieghardt, Z. Anorg. Allg. Chem. 1992 , 608 , 6 0 –6 8.

[26] E. D. Cline, S. E. Adamson, S. Bernhard, Inorg. Chem. 2008 , 47 , 1037 8 –

10388.

Received: August 23, 2013

Revised: January 16, 2014

[8] a) H. I. Karunadasa, C. J. Chang, J. R. Long, Nature 2010 , 464 , 132 9 –

1333; b) H. I. Karunadasa, E. Montalvo, Y. Sun, M. Majda, J. R. Long, C. J.

Chang, Science 2012 , 335 , 69 8 –7 02.

Published online on && &&, 0000

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

13

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Paper

FULL PAPER

&

Hydrogen Production

A. Call, Z. Codolꢀ, F. AcuÇa-Parꢁs,

J. Lloret-Fillol*

&& – &&

Photo- and Electrocatalytic H₂

Production by New First-Row

Transition-Metal Complexes Based on

an Aminopyridine Pentadentate

Ligand

Cobalt is king! The Fe, Co, and Ni com-

plexes based on the same ligand have

shown excellent electrocatalytic H₂ac-

tivity (see figure). While only the Co

complex has excellent H₂photoproduc-

tion catalytic activity, when using

sensitiser and Et₃N as electron donor.

Analytic and kinetic studies show that

the catalytic activity is molecular. The

lack of photochemical activity of the Fe

and Ni complexes is ascribed to their

redox chemistry and the pK_a,

[Ir(ppy)₂(bpy)]PF₆(PS_Ir) as photo-

respectively.

&

Chem. Eur. J. 2014, 20, 1 – 14

www.chemeurj.org

14

ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Article Doi

Photo- and electrocatalytic H2 production by new first-row transition-metal complexes based on an aminopyridine pentadentate ligand

DOI: 10.1002/chem.201303317

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201303317

Full text of DOI:10.1002/chem.201303317

Products guided by the article

R&D Labs maybe for 1022983-26-1

Relevant to this article

Hot Product