New luminescent copper(I) complexes with extended π-conjugation

Polyhedron 140 (2018) 42–50

Contents lists available at ScienceDirect

Polyhedron journal homepage: www.elsevier.com/locate/poly

New luminescent copper(I) complexes with extended p-conjugation Kévin Soulis a, Christophe Gourlaouen b, Chantal Daniel b, Alessia Quatela c, Fabrice Odobel a, Errol Blart a, Yann Pellegrin a,⇑ a

Université UNAM, CNRS, Chimie Et Interdisciplinarité: Synthèse, Analyse, Modélisation (CEISAM), UMR 6230, 2, rue de la Houssinière – BP 92208, 44322 Nantes Cedex 3, France Laboratoire de Chimie Quantique Institut de Chimie UMR 7177 CNRS-Université de Strasbourg, 4, Rue Blaise Pascal CS 90032, F-67081 Strasbourg Cedex, France c HORIBA France SAS, Avenue de la Vauve – Passage Jobin Yvon CS 45002, 91120 Palaiseau, France b

a r t i c l e

i n f o

a b s t r a c t While copper(I)-bis(diimine) complexes [CuI(L)2]+ are considered as potent substitutes for [RuII(bpy)3]2+, they exhibit low molar extinction coefficients with respect to ruthenium parent analogues. One interesting possibility to improve the light collection ability of [CuI(L)2]+ consists in increasing the length of the Cu-L dipole. In order to achieve this goal, we propose in this contribution to fuse aromatic rings onto the 2,9-di-nbutyl-1,10-phenanthroline core and examine how the properties of the corresponding copper(I) complexes are impacted. Electrochemical, absorption and emission properties are assessed; rewardingly, the envisioned approach was successful since extinction coefficients above 10,000 M
1cm
1 were measured. All copper(I) complexes remain photoluminescent, with emission maxima greatly varying from 725 to 815 nm, strongly affected by the molecular structures. A rationale to explain the variations of the emission quantum yields is proposed. Ó 2017 Elsevier Ltd. All rights reserved.

Article history: Received 30 September 2017 Accepted 18 November 2017 Available online 24 November 2017 Keywords: Copper(I) complexes Phenanthroline Photoluminescence Extinction coefficients Extended conjugation

1. Introduction Copper(I) complexes [CuI(L)2]+ where L is a chelating diimine ligand such as 2,20 -bipyridine (bpy) or 1,10-phenanthroline (phen) have attracted considerable attention since McMillins’ breakthrough in 1980 [1], revealing that those molecular species can feature near IR (NIR) photoluminescence if the ligands L are sterically challenged in a of the chelating nitrogen atoms. There has been since an extensive work on the impact of bulky substituents around the copper(I) coordination sphere, leading to an overall understanding of the photo-induced processes within such copper complexes [2–17]. Briefly, [CuI(L)2]+ complexes undergo an internal charge transfer upon MLCT excitation following the equation below: þ

½CuI ðLÞ2  þ hm ! ½CuII ðLÞðL
Þ

þ

When the complex is excited, the copper(I) ion is transiently oxidized into copper(II) while an electron from the 3dCu orbitals is transferred onto a ligand centred p⁄ orbital. Given that copper (I) and copper(II) ions display different preferred geometries (tetrahedral and square planar respectively), the excitation of those complexes in their metal-to-ligand-charge-transfer (MLCT) states entails a dramatic flattening of the structure which is responsible ⇑ Corresponding author. E-mail address: [email protected] (Y. Pellegrin). https://doi.org/10.1016/j.poly.2017.11.026 0277-5387/Ó 2017 Elsevier Ltd. All rights reserved.

for quick, radiationless deactivation of the excited state [5,11,18]. This deactivation is due to the strong stabilization of the lowest excited states during flattening favouring non radiative decay by virtue of the gap law. Additionally, the distortion of the complex frees space above the copper(II) ion allowing the latter to be attacked by any kind of nucleophile to reach the stable five-coordinate square planar pyramidal geometry; solvent molecules play an important role in this process [6,19,20]. A collection of copper(I)-bis(diimine) luminescent complexes, both homoleptic ([Cu(L)2]+) and heteroleptic ([Cu(L)(L0 )]+) [21,22] exhibiting interesting properties in their excited states have thus been isolated and studied in various domains (dye sensitized solar cells [23–31], photo-induced water splitting [32], organic photochemistry [32–37], supramolecular photo-induced electron transfer) [38–46]. Those complexes are particularly well-suited for photo-reductive mechanisms since their photo-reductive power is higher than the champion photosensitizer [RuII(bpy)3]2+ [47]. However, bearing in mind that the ultimate aim is to replace the latter by copper(I)-bis(diimine) complexes, there remain a few issues to address. In particular, [CuI(L)2]+ complexes are weak emitters, kinetically labile and endowed with small extinction coefficients. Great progresses have been made with copper complexes [CuI(L)2]+ where L is tert-butyl [7,13], sec-butyl [4,9] or iso-propyl [10] 2,9-substituted phen. These complexes, however, remain weak absorbers with extinction coefficients ranging between 3100 and 7500 Lmol
1cm
1 [7]. Incidentally, grafting

K. Soulis et al. / Polyhedron 140 (2018) 42–50

methyl groups in positions 3, 4, 7 and 8 on the already 2,9-disubstituted phen core was very beneficial from all points of view, since the emission lifetime was greatly improved and the extinction coefficient too (ca. 104 Lmol
1cm
1) [9,10]. The inductive effects of the methyl groups tend to increase the dipole length and this materializes into a more intense MLCT band [6,48]. A similar approach consists in using chromogenic diimine ligands [29,49] :by increasing the extent of p conjugation over the phen core and tethering electron donating groups, the dipole length is increased and additional ligand centred transitions can overlay with the MLCT band. Very high extinction coefficients have been obtained using this approach, but loss of the MLCT excited state properties were experienced in a few cases because the absorption features of the ligand override those of the copper-diimine chromophore. To sum up, increasing the extent of p delocalization is a good strategy to improve the extinction coefficients of the MLCT and we propose in this contribution to apply the latter by fusing aromatic rings to the phenanthroline ligand. Several examples can be found in the literature where concatenation of extra aromatic rings lead to improved e, the epitome of all being famous dipyrido [3,2-a:20 ,30 -c]phenazine ligand, best known as dppz [50,51]. The latter ligand results from the fusion of the phenazine electron acceptor with a bipyridine cavity. This ligand and complexes thereof have been abundantly studied owing to their intriguing electronic properties [52–55] and tested in many different fields such as cytotoxic agents design [56–58], photo-induced electron transfer platforms [59,60], catalysis [61–64] and of course DNA probes [65]. Although the phenazine spacer is known to be electronically decoupled from the dipyrido coordination sphere [52,66], the extinction coefficient of e.g. [RuII(bpy)2(dppz)]2+ is significantly higher than parent [RuII(bpy)3]2+ [60], probably thanks to the increased dipole length. Fusion of one additional benzene onto dppz yielded yet again improved extinction coefficients, to a lesser extent though [67]. Importantly, the ligand centred absorption features of [RuII(bpy)2(dppz)]2+ stand below the Ru ? L MLCT and do not interfere with the MLCT transition. Thus, dppz seems a good candidate to improve copper(I) complexes light harvesting capacity, since [CuI(L)2]+ complexes exhibit absorption profiles which are surprisingly similar to those of [RuII(bpy)3]2+. The latter was incidentally employed in the field of diphosphine-diimine [64] and diimineA-diimineB [68], heteroleptic copper(I) complexes and afforded indeed better absorbing complexes than their associated references. Besides, a little studied ligand, yet closely related to dppz and showing increased p conjugation too is ligand np (Fig. 1), consisting of an anthracene spacer fused to a bipyridine cavity [69–71]. Although np and dppz are very look-alike, their electronic properties are quite different. In particular, the anthracene moiety is a far less potent electron acceptor than phenazine, and this translates into very different electrochemical behaviors. Importantly, [RuII(bpy)2(np)]2+ features an improved MLCT extinction coefficient compared to [RuII(bpy)3]2+ [69]. Dppz and np ligands are thus our primary targets for this contribution. They however cannot be used as such and it is mandatory to tether bulky substituents in a of the chelating nitrogen atoms of dppz and np to design room temperature, solution phase luminescent copper(I) complexes. Interestingly, Guo et al. have published a theoretical article where heteroleptic copper(I) complexes bearing methyl-substituted dppz and np have been reported for future use in dye sensitized solar cells [70]. Critically, the complexes were praised for their enhanced MLCT oscillator strengths, justifying the interest in preparing these species. We thus embarked in the synthesis of ligands L1 and L3 (Fig. 1), consisting respectively of dppz and np molecules encumbered with two nbutyl chains on both sides of the diimine chelate. On our way, we isolated ligand L2, displaying a potentially high dipole

43

length too, and decided to explore the ground and excited states properties of the corresponding copper(I) complex. In this article, we evidence the fusing of aromatic rings onto phenanthroline as a good approach to increase the extinction coefficients of related copper(I) complexes without sacrificing to too large an extent the luminescence quantum yields. 2. Experimental section Chemicals were purchased from Sigma–Aldrich or Alfa Aesar and used as received. Thin-layer chromatography (TLC) was performed on aluminium sheets precoated with Merck 5735 Kieselgel 60F254. Column chromatography was carried out either with Merck 5735 Kieselgel 60F (0.040–0.063 mm mesh). 1H spectra were recorded on an AVANCE 300 UltraShield BRUKER. Chemical shifts for 1H NMR spectra are referenced relative to residual protium in the deuterated solvent (CDCl3 d = 7.26 ppm). NMR spectra were recorded at room temperature, chemical shifts are written in ppm and coupling constants in Hz. High-resolution mass (HRMS) spectra were obtained either by electrospray ionization coupled with high resolution ion trap orbitrap (LTQ-Orbitrap, ThermoFisher Scientific,) or by MALDI-TOF-TOF (Autoflex III, Bruker), working in ion-positive or ion-negative mode. Electrochemical measurements were made under an argon atmosphere in CH2Cl2 with 0.1 M Bu4NPF6. Cyclic voltammetry experiments were performed by using an Autolab PGSTAT 302 N potentiostat/galvanostat. A standard three-electrode electrochemical cell was used. Potentials were referenced to a saturated calomel electrode (SCE) as internal reference. All potentials are quoted relative to SCE. The working electrode was a glassy carbon disk and the auxiliary electrode was a Pt wire. In all the experiments the scan rate was 100 mVs
1. UV–Vis absorption spectra were recorded on a UV2401PC Shimadzu, using 1 cm path length cells. Emission spectra were recorded on a Fluoromax-4 Horiba Jobin Yvon spectrofluorimeter (1 cm quartz cells). Luminescence decays were recorded with a DELTAFLEX time correlated single photon counting system (HORIBA) on degassed dichloromethane solutions. All calculations were performed with ADF 2013 package [72] at the DFT level of theory with B3LYP functional [73]. The atoms were described by the all electrons slater-type TZP basis set [74]. Scalar relativistic corrections were included through the ZORA Hamiltonian [75]. Solvent corrections for dichloromethane were introduced using PCM model [76]. The structures were fully optimized and the absorption spectra were computed by means of TD-DFT [77] on these structures. For the spectra, only the nonequilibrium response of the solvent were included and Tamm-Dancoff [78] corrections introduced. In all the calculations, the butyl chains were replaced by methyl. 2,9-di-nbutyl-1,10-phenanthroline [2], 2,9-di-nbutyl-1,10phenanthroline-5,6-dione [79] (1) and complex C4 [80] were synthesized according to previously published literature procedures. 2.1. Ligand 3,6-di-nbutyl-dipyrido[3,2-a:20 ,30 -c]phenazine L1 0.12 mmol of 1 (40 mg) and 0.12 mmol of 1,2-diaminobenzene 2 (13.4 mg) were dissolved in ethanol (20 mL) and heated to reflux for 8 h. The solvent was then evaporated and dissolved in dichloromethane (10 mL). The latter solution was washed three times with aqueous 1 M HCl (100 mL  2), then pure water. The organic phase was subsequently dried on sodium sulphate, filtered and evaporated to dryness, affording a pale yellow powder. Yield: 38 mg (80%). 1H NMR (300 MHz, CDCl3, 25 °C): d = 9.56 (2H, d, J = 8.1 Hz), 8.34 (2H, dd, J = 3.3 Hz, J = 6.6 Hz), 7.90, (2H, dd, J = 3.3 Hz, J = 6.6 Hz), 7.68 (2H, d, J = 8.4 Hz), 3.28 (4H, m), 1.96 (4H, m), 1.55 (4H, m), 1.03 (6H, t, J = 7.5 Hz) ppm. HRMS

44

K. Soulis et al. / Polyhedron 140 (2018) 42–50

R

R

CN

N

N

N

N

N

N

R

N

R = H : np-CN R = nBu : L2

nBu nBu X

X

Cu X

N

N

R

R = H : np R = nBu : L3

+ N

N

N

CN

R

R = H : dppz R = nBu : L1

R

X

nBu nBu

PF6N

N Cu

N

N

C1: [CuI(L1)2]+, X = N C2: [CuI(L2)2]+, X = C-CN C3: [CuI(L3)2]+, X = CH

+ PF6

nBu nBu

z

nBu nBu

C4

Fig. 1. Structures of the three ligands L1, L2 and L3 and their associated homoleptic copper(I) complexes C1, C2 and C3. Structure of model complex C4.

(MALDI) m/z: [M+H]+ calculated for C26H27N+4: 395.2236; found: 395.2227. D = 2.3 ppm. 2.2. Ligand 3,6-di-nbutyl-naphto[2,3-f][4,5]phenanthroline-9,14dicarbonitrile L2 1 (0.98 mmol, 315 mg) and 1,2-diacetonitrilebenzene (1.13 mmol, 176 mg) were dissolved in acetonitrile (25 mL) and heated to reflux. DBU (1,8-Diazabicyclo [5.4.0] undec-7-ene, 1 mmol, 156 mg) was then added to the solution which turned dark green at once. After 4 h of reflux, the medium was let to return to room temperature and water was added. The mixture was extracted twice with dichloromethane. The organic phase was then dried on sodium sulphate, filtered and evaporated to dryness. The crude was chromatographied on silica gel (eluent: dichloromethane/triethylamine 98/2 (v/v)). The first fraction contained the desired compound and 1,2-diacetonitrilebenzene (respectively 43% and 57% molar, estimated by NMR). The latter mixture was dissolved in dichloromethane, degassed by argon bubbling, and a degassed solution of Cu(CH3CN)4PF6 (0.21 mmol, 77 mg) was syringed herein. After 30 min, the deep red solution is evaporated to dryness and subjected to chromatography column on silica gel (eluent: gradient from pure dichloromethane to a mixture of dichloromethane and methanol (98/2 v/v)). The collected orange fraction was then dissolved in acetonitrile (5 mL) and a saturated aqueous solution of KCN (5 mL) was added. The solution changes from dark red to light yellow within a few seconds. The latter was extracted twice with dichloromethane, dried on sodium sulphate, filtered and evaporated to dryness. The resulting yellow powder was the desired compound. Yield: 189 mg (44% on the overall process). 1H NMR (300 MHz, CDCl3, 25 °C): d = 9.69 (2H, d, J = 8.7 Hz), 8.63 (2H, dd, J = 3.3 Hz, J = 6.6 Hz), 7.92, (2H, dd, J = 3.3 Hz, J = 6.6 Hz), 7.60 (2H, d, J = 8.7 Hz), 3.21 (4H, m), 1.94 (4H, m), 1.53 (4H, m), 1.03 (6H, t, 7.5 Hz) ppm. 2.3. Ligand 3,6-di-nbutyl-naphto [2,3-f][4,5]phenanthroline L3 L2 (0.34 mmol, 150 mg) was suspended in a mixture of ethanol (7 mL) and water (1 mL). KOH (54 mmol, 3 g) was added. The brown solution was set in an autoclave and heated under autogeneous pressure at 200 °C for two days. When back at room temper-

ature, the resulting slurry is poured into water (10 mL) and extracted with dichloromethane (20 mL, three times). The organic phase was washed with water, dried over sodium sulphate, filtered and evaporated under reduced pressure. The crude was then subjected to chromatography on silica gel (eluent: gradient of dichloromethane 100% to a mixture of dichloromethane and methanol 95/5 (v/v)) and a yellow powder was collected. Yield: 20 mg (15%). 1H NMR (300 MHz, CDCl3, 25 °C): d = 8.99 (2H, s), 8.94 (2H, d, J = 8.4 Hz), 8.09 (2H, dd, J = 3.3 Hz, J = 6.3 Hz), 7.59 (4H, m), 3.19 (4H, m), 1.93 (4H, m), 1.52 (4H, m), 1.02 (6H, t, J = 7.2 Hz). HRMS (MALDI) m/z: [M+H]+ calculated for C28H29N+2: 393.2331; found: 393.2346. D = 3.8 ppm. 2.4. Complex [Cu(L1)2]+(PF
6 ) C1 Ligand L1 (0.076 mmol, 30 mg) was dissolved in dichloromethane (5 mL) and thoroughly degasssed by argon bubbling in an ultrasonic bath. A degassed dichloromethane solution of Cu (CH3CN)4PF6 (0.038 mmol, 14.2 mg) was syringed into the former; the solution turned red and was left at room temperature for half an hour. The solvent was then removed by rotary evaporation under reduced pressure. The crude was filtered through a silica plug in dichloromethane and evaporation of the filtrate afforded C1. Yield: 34 mg (90%). 1H NMR (300 MHz, CDCl3, 25 °C): d = 9.87 (4H, d, J = 8.1 Hz), 8.48 (4H, dd, J = 3.3 Hz, J = 6.3 Hz), 8.05, (4H, dd, J = 3.6 Hz, J = 6.6 Hz), 8.02 (4H, d, J = 8.4 Hz), 2.86 (8H, m), 1.47 (8H, m), 0.83 (8H, m), 0.33 (12H, t, J = 7.2 Hz) ppm. HRMS (MALDI) m/z: [M]+ calculated for C52H52N8Cu+: 851.3611; found: 851.3618. D = 0.8 ppm. 2.5. Complex [Cu(L2)2]+(PF
6 ) C2 Complex C2 was isolated during the synthesis procedure of ligand L2. 1 H NMR (300 MHz, CDCl3, 25 °C): d = 10.18 (4H, d, J = 8.7 Hz), 8.80 (4H, dd, J = 3.0 Hz, J = 6.6 Hz), 8.11 (4H, dd, J = 3.3 Hz, J = 6.6 Hz), 7.98 (4H, d, J = 8.7 Hz), 2.85 (8H, m), 1.47 (8H, m), 0.84 (8H, m), 0.40 (12H, t, J = 7.2 Hz). HRMS (MALDI) m/z: [M]+ calculated for C60H52N8Cu+: 947.3611; found: 947.3592. D = 2.0 ppm.

45

K. Soulis et al. / Polyhedron 140 (2018) 42–50

2.6. Complex [Cu(L3)2]+(PF
6 ) C3 Ligand L3 (0,025 mmol, 10 mg) was dissolved in dichloromethane (5 mL) and thoroughly degassed by argon bubbling in an ultrasonic bath. A degassed dichloromethane solution of Cu (CH3CN)4PF6 (0.013 mmol, 4.7 mg) was syringed into the former; the solution turned red and was left at room temperature for half an hour. The solvent was then removed by rotary evaporation under reduced pressure. The crude was subjected to preparative thin layer chromatography, using as eluent the mixture acetonitrile/H2O/saturated aqueous KNO3 (100/8/2 in volumes). The bright orange stripe was collected and the silica was washed several times with a mixture of dichloromethane and methanol (95/5 v/ v) until it was colorless. The orange organic phase was evaporated under reduced pressure affording a pure batch of C3. Yield: 14.5 mg (58%). 1H NMR (300 MHz, CDCl3, 25 °C): d = 9.36 (4H, d, J = 8.7 Hz), 9.26 (4H, s), 8.23, (4H, m), 7.90 (4H, d, J = 8.4 Hz), 7.69 (4H, m), 2.77 (8H, m), 1.41 (8H, m), 0.75 (8H, m), 0.30 (12H, t, J = 7.2 Hz) ppm. HRMS (MALDI) m/z: [M]+ calculated for C56H56N4Cu+: 847.3801; found: 847.3785. D = 1.9 ppm. 3. Results and discussion 3.1. Synthesis The synthis of all ligands is depicted in Scheme 1, and rely on the same starting material, namely 2,9-di-nbutyl-1,10-phenanthroline-5,6-dione 1, obtained in good yields by oxidation of 2,9di-nbutyl-1,10-phenanthroline [79]. Following the well-established dppz synthesis protocol, condensation of 1 onto 1,2phenylenediamine 2 in refluxing ethanol afforded ligand L1. The presence of floppy butyl alkyl chains endows L1 with a higher solubility than dppz; hence, although the latter precipitated as a pure compound in the reaction mixture, a purification step by chromatography was necessary to get a batch of pure L1. The synthesis of L3 proved to be more challenging. The related np ligand (namely L2 with no alkyl chains, see Fig. 1) was synthesized twice before with quite low yields following two different strategies. Our own protocol is inspired by both, and is presented in Scheme 1. First, condensation of 1 with 1,2-diacetonitrilebenzene 3 in presence of 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU) yielded L2. Once again, because of the higher solubility provided by the nbutyl chains, L2 did not precipitate in the reaction medium. All our attempts to isolate L2 by precipitation invariably lead to mixtures of L2 and 3, a mixture which proved to be very difficult to purify by chromatography column. Taking advantage of the coordinating ability of L2 versus 3, we prepared complex C2 by adding the appropriate

amount of Cu(CH3CN)4PF6 to the latter mixture. C2 was then easily separated from 3 by chromatography. C2 was finally treated in a mixture of CH3CN, water and potassium cyanide; the red solution quickly turned opaque as a bright yellow precipitate of pure L2 formed. A batch of pure C2 was retrieved for further analysis. The latter was recovered by extraction with dichloromethane, with an overall yield of 44% over the whole process. The low yield could partly be explained by the fact that the starting dione 1 is not stable in basic medium, all the more if there are traces of water (formation of fluorenone) [68]. The synthesis of L3 consists in a one-pot hydrolysis of both nitrile groups on L2 followed by decarboxylation in a strongly alkaline medium at high temperature. Only small amounts of ligand L3 were obtained, despite all our efforts to increase the yield of the reaction. Increasing the duration of the reaction did not improve the yield. On the other hand, the synthesis of complexes C1 and C3 was straightforward. Cu(CH3CN)4PF6 (1 eqv.) was added to dichloromethane solutions of L1 and L3 (2 eqv.) under argon at room temperature. The solutions immediately turned orange-red, unmistakable sign that coordination of copper(I) by two diimine ligands has taken place. As mentioned above, C2 was incidentally obtained during the purification process of L2. High resolution mass and NMR spectra of all ligands and complexes are all in agreement with the proposed structures. 3.2. Electrochemistry Cyclic voltammetry was performed on dichloromethane solutions of each complex in order to probe their electronic properties in the ground state. Voltammograms are given in Fig. 2 and relevant potentials are reported in Table 1. In all cases, a pseudo-reversible oxidation wave is observed around 1 V versus SCE, which can be unambiguously assigned to the copper(II)/copper(I) redox couple. The difference between the anodic peak potential Eap and the cathodic peak potential Ecp is a good indicator of the reversibility of the electrochemical process and should be close to 60 mV for a one-electron exchange process; it is much larger in this case and for copper(I)-diimine complexes in general owing to the flattening of the coordination sphere occurring when tetrahedral copper(I) is oxidized to square planar copper(II) [15,81]. The butyl chains prevent this distortion to some extent, thus justifying the Cu(II)/Cu(I) half wave potential is higher for sterically challenged copper(I) complexes than for ‘‘naked” archetypal [CuI(phen)2]+ [17]. Accordingly, the steric bulk imposed by bulky substituents in a of the chelating nitrogen atoms has a great influence on the Cu(II)/Cu(I) half wave potential [4,12]; however, since C1, C2, C3 and model C4 feature the same coordination cage, we can safely

nBu

nBu N

O

H 2N

EtOH, reflux

+ N

O

nBu

nBu

H 2N

nBu

CN N

CH3CN, DBU

O

N

N

N

nBu

2

1

N

L1

nBu

CN

EtOH, KOH

N

N

+ N

RT

O CN

nBu

1

3

200°C, 3 days

N nBu

CN

L2 Scheme 1. Synthetic routes for ligands L1, L2 and L3.

N nBu

L3

46

K. Soulis et al. / Polyhedron 140 (2018) 42–50

5 μA

E (V) / SCE -1.6

-1.1

-0.6

-0.1

0.4

0.9

1.4

Fig. 2. Cyclic voltammograms of dichloromethane solutions of C1, C2 and C3 (from top to bottom) in presence of tetra-nbutylammonium hexafluorophosphate (0.1 M) at room temperature. Potentials are referenced vs. SCE, platinum disc used as working electrode. Sweep rate: 100 mVs
1.

Table 1 Relevant data extracted from electrochemistry and spectroscopic measurements.

a

Complex

E/V vs. SCE (DEp/mV)

C1 C2 C3 C4

1.11 1.16 0.99 0.92

(140) (110) (120) (150)


1.19 (180)
0.88a n.d. n.d.

kabs/nm (e/Lmol
1cm
1)

kem/nm (Uem  10
4/%)

s/ns

456 456 470 457

725 815 730 725

53 n.d. 67 1504

(10900) (10460) (8600) (7000)4

(3.6) (n.d.) (5.8) (9)4

Irreversible wave.

assume that the relative oxidation and reduction potentials of all complexes are mainly governed by electronic effects. The latter likely accounts for the more positive half wave potentials measured for all complexes compared to reference C4. The electron withdrawing effects brought by the phenazine (for C1), the dicarbonitrileanthracene (for C2) and the anthracene (for C3) spacers deplete the electron density on the metal centre, thus shifting the oxidation potential to more anodic values. C1 and C2 feature more positive oxidation potentials than C3 because the anthracene unit is less p-accepting than phenazine or dicarbonitrileanthracene. This is in contrast with observations made on [RuII(bpy)2(dppz)]2+, [RuII(bpy)2(np-CN)]2+ and [RuII(bpy)2(np)]2+ where the presence of phenazine, dicarbonitrileanthracene or anthracene moieties did not influence the oxidation potentials of the complexes [69]. In the case of C1 and C2, a rather prominent wave at ca. 0.8 V versus SCE, absent when the potential sweeping is stopped before the CuII/CuI oxidation wave, can be attributed to the partial destruction of the complexes upon oxidation, likely the loss of one ligand upon oxidation. The strong electron withdrawing character of the phenazine and dicarbonitrileanthracene fragments may deplete the electron density on the nitrogen atoms, therefore affecting the r-donation ability of the latter. This phenomenon is not observed in the case of C3 (nor C4) [4], suggesting anthracene has a milder effect on the electronic properties of the complex. Copper(I)-diimine complexes usually display very negative reduction potentials, well below the limit of electroactivity window of dichloromethane (ca.
1.5 V versus SCE in our conditions), and corresponding to one-electron ligand centred reductions. Yet, complexes C1 and C2 feature well defined cathodic waves and by comparison with previously published work on related ruthenium complexes [60,69,82], those reduction waves were assigned to the addition of an extra electron on the phenazine and dicarbonitrileanthracene spacers, respectively. Logically, the more easily reduced complex C2 is the more difficult to oxidize too. Conversely,

complex C3 does not exhibit any electrochemical event when sweeping towards negative potentials. The concatenation of two phenyl rings onto phenanthroline does not increase to a sufficient extent the p accepting power of ligand L3 to promote the reduction of the latter above
1.5 V versus SCE. This is in line with the observations from Albano et al. made on corresponding ruthenium complexes [69]. 3.3. Electronic absorption spectroscopy The absorption spectra of complexes C1, C2 and C3 were recorded in dichloromethane (Fig. 3), the associated data are gathered in Table 1. All three new complexes absorb in the visible domain, owing to the famous broad MLCT spanning from 400 to 650 nm. Additional bands take place between 350 and 420 nm; these are monitored on the spectra of the free ligands (see supporting information, Fig. S1) and were thus assigned to ligand centred p–p⁄ transitions. Other intense ligand centred transitions take place in the UV; the latter are red-shifted for the complexes with respect to the free ligands, a common feature confirming the coordination of the ligands to the copper(I) ion. The MLCT of complexes C1 and C2 are look-alike and exhibit a maximum absorbance wavelength very similar to model C40 s. However, the transitions are substantially broader (see supporting information, Fig. S2) and the absorption onsets appear thus red-shifted with respect to C4, showing that the phenazine and anthracenedicarbonitrile moieties exert a stabilizing effect on the MLCT. The MLCT absorption maximum of complex C3 is obviously red-shifted compared to C4 on the one hand, and C1 and C2 on the other. The transition appears thinner than C1 and C20 s and the absorption onset follows the order: C2 > C1  C3 where ‘‘>” means ‘‘exhibits a longer absorption onset wavelength than”. Thus, it seems difficult to relate the electronic effects monitored by electrochemistry and those estimated by UV–Vis spectroscopy. The difference between ‘‘optical” and ‘‘redox” orbitals is here again exemplified; this is all the more true

47

K. Soulis et al. / Polyhedron 140 (2018) 42–50

160000

exncon coefficient (L.mol-1.cm-1)

140000 120000 100000 80000 60000 40000 20000 0 250

350

450

550

650

750

850

wavelength / nm Fig. 3. UV–Vis absorption spectra of complexes C1 (solid black line), C2 (solid grey line) and C3 (dotted black line) in dichloromethane.

since the MLCT for [CuI(L)2]+ complexes is not a S0 to S1 transition, the latter being symmetry forbidden [83]. Rewardingly, the extinction coefficients of all three complexes are higher than those of common copper(I) complexes such as C4. This is likely a result of the increased dipole moment along the z-axis of the molecule. Accordingly, L1 and L2 probably exhibit higher dipolar moments than L3 because of the strong electronic effects imposed by the intracyclic nitrogen atoms (for L1) and the nitrile groups (for L2). This translates into a larger extinction coefficient for C1 and C2 compared to C3. 3.4. Calculations The structures of the three complexes were fully optimized. The general structures of the three complexes are very similar with one major difference. The complex C1 and C3 are perfectly symmetric with a D2d symmetry. The complex C2 has a lower symmetry: the anthracenedicarbonitrile part is no longer planar. The presence of the cyano groups on the anthracene part induces an out of plane twist of the ligand (see supporting information, Fig. S3). However, the symmetric structure is very close in energy, the D2d equivalent of C2 is only 5 kcalmol
1 less stable than distorded structure. Such a small difference suggests that the ligands are oscillating between their two possible distorted structures and probably the complex is dynamically planar; this is confirmed by the NMR data showing only one species on the spectra. All the structures were then optimized without symmetry and the absorption spectra were computed on these structures for the three complexes (see supporting information, Fig. S4 and Table S1). The theoretical results are in good agreement with the experimental ones. A first set of transitions is present between 450 and 500 nm in the three complexes. This is followed by a second set of transitions in the region of 325–400 nm. Then the most intense band appears. We retrieve the qualitative experimental order. The lowest energy peak is for C2 at 311 nm (294 and 318 nm exp.) followed by that of C3 at 293 nm (exp: 294 nm). The most energetic transition is that of C1 at 282 nm (exp: 280 nm). However, we do not retrieve the splitting of the peak of C2, which may arise from the dynamical behavior of the ligand. The spectrum of C3 is the most simple and will be described first. The first absorbing band is located at 484 nm and corresponds to a MLCT from the copper(I) ion to the phenantroline part of the ligand (Fig. 4c). Two lower energy states are present (534 and

516 nm) but have negligible intensities and have the same nature. A massif is present between 325 to 400 nm. The peak at 375 nm is the combination of four electronic transitions, two of them being MLCT from a deeper d orbital of copper again towards the phenantroline part of the ligand. The other two are charge transfer inside the ligands from the external part towards the central one (Fig. 4c). The third peak at 346 nm is generated by a single transition and is a p–p⁄ transition delocalized on the whole ligand. Finally, the very intense peak at 293 nm is generated by a single transition corresponding to a p–p⁄ excitation delocalized on the whole ligand. In C1, the replacement of one C–H group by a N atom in the ligand leads to a qualitatively similar spectra with small differences. The first intense peak is slightly blue-shifted at 470 nm but has exactly the same nature as the peak at 484 nm in C3. However, a weak transition appears in C1 at 539 nm due to the presence of the nitrogen. This new peak corresponds to a MLCT from copper to the phenazine part of the ligand (Fig. 4a). A similar transition was already present in C3 but was strongly blue shifted and had no absorption intensity. The second intense band in C1 appears at 339 nm and is a p–p⁄ transition delocalized on the whole ligand similar to the weak peaks at 346 nm of C3 but much more intense. The presence of the nitrogen in the ligand C1 instead of C–H in C3 modifies the orbital distribution. The LUMO and LUMO+1, which were localized on the phenantroline part in C3 (as in most [CuI(L)2]+ model complexes), are now localized on the phenazine part of the ligand in C1, the LUMO+2 and LUMO+3 corresponding to the phenantroline part. This is corroborated by the electrochemical measurements, and explains the presence of new peaks at low energy in C1. The presence of the CN group in C2 again change the orbital distribution, making the anthracenedicarbonitrile part of the ligand much more electrophilic. Indeed, the LUMO and LUMO+1 are localized on the anthracenedicarbonitrile part of the ligand in C2 as in C1 but they are also delocalized on these CN groups. This is the reason of the existence of low energy transitions of MLCT type (Fig. 4b) at 624 nm. Then we retrieve the intense MLCT band at 473 nm present in C1 and in C3. New bands appear around 405 nm which correspond to p–p⁄ excitations localized on the anthracenedicarbonitrile. The changes of the ligand in the complexes affect the nature of the lowest unoccupied orbital and therefore the nature and energy of the lowest transitions. Compared to the reference C3, there are additional LUMOs at lower energy, sources of new MLCT in C1 and C2, which are not present in C3. In all case the main transitions

48

K. Soulis et al. / Polyhedron 140 (2018) 42–50

complex

Transition / nm (computed)

C1

539

C2

624

Nature of the transition

484

C3

375

293

Fig. 4. Nature of the transitions determined by electron density differences between the excited state and the ground state in C1, C2 and C3. In red, area electronically depleted upon transitions and in green area electronically enriched upon transition.

are similar and weakly affected by the modification introduced on the ligand, though it may be less true for C2. These extra MLCT present in C1 and C2 are at the origin of the broadening of the MLCT band. 3.5. Luminescence properties The steady state emission spectra of dichloromethane solutions of complexes C1, C2 and C3 with matching absorbance at the excitation wavelength are given in Fig. 5 and related data are reported in Table 1. All complexes are luminescent but show different behaviors. The emission spectra of C1, C3 and reference compound C4 are very comparable as regards their shape and emission maxima, despite their very different molecular structures. A similar

observation has been made on parent ruthenium complexes [69]. A very slight red-shift (ca. 5 nm) of the emission maximum wavelength for C3 compared to C4 could be grounded in the extended p electronic system of ligand L3 compared to plain phenanthroline. As regards C1, the dppz ligand is reputed for its peculiar electronic structure, as mentioned above; it has been indeed well demonstrated that the phenazine spacer and the chelating pyridyl rings were not electronically coupled. Thus, the emission maxima wavelengths of solutions of the famous complexes [RuII(bpy)2(dppz)]2+ and [RuII(bpy)3]2+ are virtually identical. A similar behavior can be observed here in the case of C1. On the other hand, the maximum emission wavelength of complex C2 is substantially red-shifted, which likely reflects the strong p-accepting nature of the anthracenedicarbonitrile moiety.

49

K. Soulis et al. / Polyhedron 140 (2018) 42–50

450000

luminescence intensity (a.u.)

400000 350000 300000 250000 200000 150000 100000 50000 0 500

550

600

650

700

750

800

850

wavelength / nm Fig. 5. Emission spectra of complexes C1 (black solid line), C2 (grey solid line) and C3 (dotted black line) in degassed dichloromethane. Solutions are absorption matched at the wavelength of excitation (456 nm for C1 and C2, 470 nm for C3).

Intriguingly, the luminescence behavior of C2 is markedly different from C1 and C3 whereas their absorption spectra are quite similar. Nevertheless, in the case of C2 the spectrum shape is broader and the onset of absorption occurs at longer wavelength than C1 or C3 (see Fig. S2 and TD-DFT calculations reported in Table S1). Moreover, one must bear in mind that the observed MLCT band is not a S0 to S1 transition, the latter being symmetry-forbidden for tetrahedral copper(I) complexes [5,83]. As such, we only have few experimental data on the energy of the S1 state from the absorption spectra of C1, C2 and C3 and drawing conclusions from the latter could be irrelevant. Relationships between the absorption and emission energies are besides difficult to establish in the particular case of copper(I)-bis(diimine) complexes [17]. Additionally, one cannot exclude a contribution from a ligand centred emission as proposed by Albano et al. for the related ruthenium complex [69]. The strong variations of the emission quantum yields among the three complexes are remarkable features on Fig. 5. The position of C20 s emission band is too far in the red part of the visible to allow measuring a trustworthy quantum yield with our spectrometer. Nevertheless, one can deduct from the absorption matched emission spectra of the three complexes the following trend: U (C3) > U(C1) > U(C2) where U is the emission quantum yield. A lower luminescence quantum yield for C2 is expectable given the strongly red-shifted emission band by virtue of the gap law. On the other hand, the latter cannot account for the fact that both C1 and C3 have lower emission quantum yields than reference C4. Partial quenching of the luminescence by photo-induced intramolecular electron transfer can be ruled out although copper(I)-bis(diimine) complexes are justly considered as very potent photo-reductants. In our case, a direct photo-induced electron transfer from copper(I) to L1, L2 or L3 upon excitation in the MLCT is endergonic by several hundreds of meV (see supplemental information, Table S2 and comments). Interestingly though, the emission quantum yields and the extinction coefficients for C1, C3 and C4 are progressing in opposite directions: the higher e, the lower the quantum yield. It is well established now that copper(I) complexes such as C4 exhibit a thermally activated delayed fluorescence mechanism (TADF) [10,84–86]. As such, the emission band of C4 is mainly composed of a S1 to S0 transition. Yet in our approach we aimed at increasing the extinction coefficients of copper complexes by enlarging the moment of the copper-ligand dipole. S0 to Sn transitions are indeed

favoured but in return S1 to S0 de-excitation transition too, leading to an overall decrease of the emission quantum yield [6,12,17]. Time resolved emission decays, measured by time-correlated single photon counting after excitation at 440 nm show that both complexes C1 and C3 have indeed much shorter emission lifetimes than complex C4 (ca. 50 ns versus 150 ns respectively) [4]. However, the emission lifetimes of C1 and C3 are very similar (unfortunately we could not manage to record a suitable fluorescence decay for complex C2 because of the appearance of an intense parasitic signal during acquisition, likely grounded in partial dissociation of the complex). The decays are given in Fig. S5, and are in agreement with the usual behavior of parent copper(I) complexes such as C4: a first short component can be assigned to the various deactivation pathways following excited state formation (flattening, internal conversion, intersystem crossing, prompt fluorescence). Longer lived phases correspond to the radiative decay and were estimated at 53 and 47 ns for complexes C1 and C3 respectively. Another phenomenon should therefore take place explaining the weaker emission quantum yield of C1 compared to C3. Typically, since the radiative constants are roughly similar, then the non-radiative constants are to be incriminated, since U = kr/(kr + knr), where U, kr and knr are the emission quantum yield, the radiative and the non-radiative constants, respectively. Knowing that lower lying MLCT excited states have been highlighted in the case of C1 (and C2), their participation to the non-radiative deexcitation pathway is plausible [82].

4. Conclusion In order to increase the notably low extinction coefficients of the MLCT states of [CuI(L)2]+ complexes (where L is a diimine ligand), we designed three ligands derived from the phenanthroline, aiming at increasing the length of the molecular dipole. The strategy was to fuse aromatic rings onto the phenanthroline core. The consequences on the electronic structure of the complexes were rationalized. In particular, all complexes are luminescent in the near infrared domain, with variable emission quantum yields. It seems that enlarging the length of the dipole has two main effects: on one hand, it indeed increases the extinction coefficient (above 10000 M
1cm
1) but on the other it increases the nonradiative deactivation pathways. In both cases, this is due to the

50

K. Soulis et al. / Polyhedron 140 (2018) 42–50

increased probability of the transitions between singlet states. The emission quantum yields are still modest, but great improvements could be made by increasing the steric bulk around the copper(I) ion first at the level of the substituents in a of nitrogen atoms (e.g. with isopropyl groups) and second by decorating the phenanthroline moiety with methyl groups. This would improve not only the luminescence intensity of the corresponding complexes, but the extinction coefficient too as a consequence of enlarged dipole length [6,9,10,48]. Acknowledgements The authors wish to thank the HORIBA-Jobin Yvon Facility for giving access to a Deltaflex TC-SPC device, Pr. Chantal Daniel for fruitful discussions, the ANR (PERCO program n°ANR-16-CE070012-01), the HPC of Strasbourg for computing facilities and the CNRS for funding. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.poly.2017.11.026. References [1] M.W. Blaskie, D.R. McMillin, Inorg. Chem. 19 (1980) 3519. [2] R.M. Everly, R. Ziessel, J. Suffert, D.R. McMillin, Inorg. Chem. 30 (1991) 559. [3] A.K. Ichinaga, J.R. Kirchhoff, D.R. McMillin, C.O. Dietrich-Buchecker, P.A. Marnot, J.P. Sauvage, Inorg. Chem. 26 (1987) 4290. [4] M.K. Eggleston, D.R. McMillin, K.S. Koenig, A. Pallenberg, J. Inorg. Chem. 36 (1997) 172. [5] M. Iwamura, S. Takeuchi, T. Tahara, Acc. Chem. Res. 48 (2015) 782. [6] G. Capano, U. Rothlisberger, I. Tavernelli, T.J. Penfold, J. Phys. Chem. A 119 (2015) 7026. [7] O. Green, B.A. Gandhi, J.N. Burstyn, Inorg. Chem. 48 (2009) 5704. [8] J. Huang, M.W. Mara, A.B. Stickrath, O. Kokhan, M.R. Harpham, K. Haldrup, M.L. Shelby, X. Zhang, R. Ruppert, J.P. Sauvage, L.X. Chen, Dalton Trans. 43 (2014) 17615. [9] C.E. McCusker, F.N. Castellano, Inorg. Chem. 52 (2013) 8114. [10] S. Garakyaraghi, P.D. Crapps, C.E. McCusker, F.N. Castellano, Inorg. Chem. 55 (2016) 10628. [11] G.B. Shaw, C.D. Grant, H. Shirota, E.W. Castner, G.J. Meyer, L.X. Chen, J. Am. Chem. Soc. 129 (2007) 2147. [12] M.T. Miller, P.K. Gantzel, T.B. Karpishin, Inorg. chem. 38 (1999) 3414. [13] N.A. Gothard, M.W. Mara, J. Huang, J.M. Szarko, B. Rolczynski, J.V. Lockard, L.X. Chen, J. Phys. Chem. A 116 (2012) 1984. [14] M.W. Mara, N.E. Jackson, J. Huang, A.B. Stickrath, X. Zhang, N.A. Gothard, M.A. Ratner, L.X. Chen, J. Phys. Chem. B 117 (2013) 1921. [15] G. Accorsi, N. Armaroli, C. Duhayon, A. Saquet, B. Delavaux-Nicot, R. Welter, O. Moudam, M. Holler, J.-F. Nierengarten, Eur. J. Inorg. Chem. 2010 (2010) 164. [16] A. Lavie-Cambot, M. Cantuel, Y. Leydet, G. Jonusauskas, D.M. Bassani, N.D. McClenaghan, Coord. Chem. Rev. 252 (2008) 2572. [17] D.V. Scaltrito, D.W. Thompson, J.A. O’Callaghan, G.J. Meyer, Coord. Chem. Rev. 208 (2000) 243. [18] M. Iwamura, H. Watanabe, K. Ishii, S. Takeuchi, T. Tahara, J. Am. Chem. Soc. 133 (2011) 7728. [19] A. Agena, S. Iuchi, M. Higashi, Chem. Phys. Lett. 679 (2017) 60. [20] G. Smolentsev, A.V. Soldatov, L.X. Chen, J. Phys. Chem. A 112 (2008) 5363. [21] M. Schmittel, A. Ganz, Chem. Commun. (1997) 999. [22] M.T. Miller, P.K. Gantzel, T.B. Karpishin, J. Am. Chem. Soc. 121 (1999) 4292. [23] N. Alonso-Vante, J.-F. Nierengarten, J.-P. Sauvage, J. Chem. Soc., Dalton Trans. (1994) 1649. [24] T.E. Hewat, L.J. Yellowlees, N. Robertson, Dalton Trans. 43 (2014) 4127. [25] T. Bessho, E.C. Constable, M. Graetzel, A. Hernandez Redondo, C.E. Housecroft, W. Kylberg, M.K. Nazeeruddin, M. Neuburger, S. Schaffner, Chem. Commun. (2008) 3717. [26] S. Sakaki, T. Kuroki, T. Hamada, J. Chem. Soc., Dalton Trans. (2002) 840. [27] C.E. Housecroft, E.C. Constable, Chem. Soc. Rev. 44 (2015) 8386. [28] M. Sandroni, L. Favereau, A. Planchat, H. Akdas-Kilig, N. Szuwarski, Y. Pellegrin, E. Blart, H. Le Bozec, M. Boujtita, F. Odobel, J. Mater. Chem. A 2 (2014) 9944. [29] K.A. Wills, H.J. Mandujano-Ramirez, G. Merino, G. Oskam, P. Cowper, M.D. Jones, P.J. Cameron, S.E. Lewis, Dyes Pigm. 134 (2016) 419. [30] M. Magni, P. Biagini, A. Colombo, C. Dragonetti, D. Roberto, A. Valore, Coord. Chem. Rev. 322 (2016) 69. [31] F.J. Malzner, M. Willgert, E.C. Constable, C.E.J. Housecroft, Mater. Chem. A 5 (2017) 13717.

[32] R.S. Khnayzer, C.E. McCusker, B.S. Olaiya, F.N. Castellano, J. Am. Chem. Soc. 135 (2013) 14068. [33] S. Paria, O. Reiser, ChemCatChem 6 (2014) 2477. [34] O. Reiser, Acc. Chem. Res. 49 (2016) 1990. [35] J.-M. Kern, J.-P. Sauvage, J. Chem. Soc., Chem. Commun. (1987) 546. [36] M. Pirtsch, S. Paria, T. Matsuno, H. Isobe, O. Reiser, Chemistry 18 (2012) 7336. [37] A. Baralle, L. Fensterbank, J.-P. Goddard, C. Ollivier, Chemistry 19 (2013) 10809. [38] L. Flamigni, V. Heitz, J.-P. Sauvage, Struct. Bonding 121 (2006) 217. [39] M.W. Mara, D.N. Bowman, O. Buyukcakir, M.L. Shelby, K. Haldrup, J. Huang, M. R. Harpham, A.B. Stickrath, X. Zhang, J.F. Stoddart, A. Coskun, E. Jakubikova, L.X. Chen, J. Am. Chem. Soc. 137 (2015) 9670. [40] J. Huang, O. Buyukcakir, M.W. Mara, A. Coskun, N.M. Dimitrijevic, G. Barin, O. Kokhan, A.B. Stickrath, R. Ruppert, D.M. Tiede, J.F. Stoddart, J.-P. Sauvage, L.X. Chen, Angew. Chem. Int. Ed. 51 (2012) 12711. [41] M. Sandroni, A. Maufroy, M. Rebarz, Y. Pellegrin, E. Blart, C. Ruckebusch, O. Poizat, M. Sliwa, F. Odobel, J. Phys. Chem. C 118 (2014) 28388. [42] J.E. Barnsley, S.O. Scottwell, A.B.S. Elliott, K.C. Gordon, J.D. Crowley, Inorg. Chem. 55 (2016) 8184. [43] N. Armaroli, G. Accorsi, J.-P. Gisselbrecht, M. Gross, J.-F. Eckert, J.-F. Nierengarten, New J. Chem. 27 (2003) 1470. [44] A. Listorti, G. Accorsi, Y. Rio, N. Armaroli, O. Moudam, A. Gégout, B. DelavauxNicot, M. Holler, J.-F. Nierengarten, Inorg. Chem. 47 (2008) 6254. [45] J.D. Megiatto, K. Li, D.I. Schuster, A. Palkar, M.Á. Herranz, L. Echegoyen, S. Abwandner, G. de Miguel, D.M. Guldi, J. Phys. Chem. B 114 (2010) 14408. [46] M.S. Lazorski, R.H. Gest, C.M. Elliott, J. Am. Chem. Soc. 134 (2012) 17466. [47] N. Armaroli, Chem. Soc. Rev 30 (2001) 113. [48] C.C. Phifer, D.R. McMillin, Inorg. Chem. 25 (1986) 1329. [49] M. Sandroni, M. Kayanuma, M. Rebarz, H. Akdas-Kilig, Y. Pellegrin, E. Blart, B.H. Le, C. Daniel, F. Odobel, Dalton Trans. 42 (2013) 14628. [50] M.N. Ackermann, L.V. Interrante, Inorg. Chem. 23 (1984) 3904. [51] J.C. Chambron, J.P. Sauvage, E. Amouyal, P. Koffi, Nouv. J. Chim. 9 (1985) 527. [52] H. van der Salm, A.B.S. Elliott, K.C. Gordon, Coord. Chem. Rev. 282 (2015) 33. [53] R. Horvath, K.C. Gordon, Inorg. Chim. Acta 374 (2011) 10. [54] S. Berger, J. Fiedler, R. Reinhardt, W. Kaim, Inorg. Chem. 43 (2004) 1530. [55] J. Fees, M. Ketterle, A. Klein, J. Fiedler, W. Kaim, J. Chem. Soc., Dalton Trans. (1999) 2595. [56] P.K. Sasmal, S. Saha, R. Majumdar, S. De, R.R. Dighe, A.R. Chakravarty, Dalton Trans. 39 (2010) 2147. [57] T. Ma, J. Xu, Y. Wang, H. Yu, Y. Yang, Y. Liu, W. Ding, W. Zhu, R. Chen, Z. Ge, Y. Tan, L. Jia, T. Zhu, J. Inorg. Biochem. 144 (2015) 38. [58] G. Scalese, I. Correia, J. Benítez, S. Rostán, F. Marques, F. Mendes, A.P. Matos, J. Costa Pessoa, D. Gambino, J. Inorg. Biochem. 166 (2017) 162. [59] K.L. Wouters, N.R. Tacconi, R. Konduri, R.O. Lezna, F.M. MacDonnell, Photosynth. Res. 87 (2006) 41. [60] E. Amouyal, A. Homsi, J.C. Chambron, J.P. Sauvage, J. Chem. Soc., Dalton Trans. (1990) 1841. [61] L. Troian-Gautier, C. Moucheron, Molecules 19 (2014) 5028. [62] T.A. White, S.E. Witt, Z. Li, K.R. Dunbar, C. Turro, Inorg. Chem. 54 (2015) 10042. [63] T.A. White, S. Maji, S. Ott, Dalton Trans. 43 (2014) 15028. [64] M. Heberle, S. Tschierlei, N. Rockstroh, M. Ringenberg, W. Frey, H. Junge, M. Beller, S. Lochbrunner, M. Karnahl, Chem. Eur. J. 23 (2017) 312. [65] A.E. Friedman, J.C. Chambron, J.P. Sauvage, N.J. Turro, J.K. Barton, J. Am. Chem. Soc. 112 (1990) 4960. [66] M.K. Brennaman, T.J. Meyer, J.M. Papanikolas, J. Phys. Chem. A 108 (2004) 9938. [67] R.M. Hartshorn, J.K. Barton, J. Am. Chem. Soc. 114 (1992) 5919. [68] Y. Pellegrin, M. Sandroni, E. Blart, A. Planchat, M. Evain, N.C. Bera, M. Kayanuma, M. Sliwa, M. Rebarz, O. Poizat, C. Daniel, F. Odobel, Inorg. Chem. 50 (2011) 11309. [69] G. Albano, P. Belser, L. De Cola, M.T. Gandolfi, Chem. Commun. (1999) 1171. [70] S. Wei, Y. Shao, X. Shi, X. Lu, K. Li, Z. Zhao, C. Guo, H. Zhu, W. Guo, Org. Electron. 29 (2016) 142. [71] W. Lu, D.A. Vicic, J.K. Barton, Inorg. Chem. 44 (2005) 7970. [72] G. te Velde, F.M. Bickelhaupt, E.J. Baerends, C. Fonseca Guerra, S.J.A. van Gisbergen, J.G. Snijders, T. Ziegler, J. Comput. Chem. 22 (2001) 931. [73] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, The Journal of Physical Chemistry 98 (1994) 11623. [74] E. Van Lenthe, E.J. Baerends, J. Comput. Chem. 24 (2003) 1142. [75] E.V. Lenthe, A. Ehlers, E.-J. Baerends, J. Chem. Phys. 110 (1999) 8943. [76] C.C. Pye, T. Ziegler, Theor. Chem. Acc. 101 (1999) 396. [77] S.J.A. van Gisbergen, J.G. Snijders, E.J. Baerends, Comput. Phys. Commun. 118 (1999) 119. [78] S. Hirata, M. Head-Gordon, Chem. Phys. Lett. 314 (1999) 291. [79] S. Ott, R. Faust, Synthesis 2005 (2005) 3135. [80] A.J. Pallenberg, K.S. Koenig, D.M. Barnhart, Inorg. Chem. 34 (1995) 2833. [81] P. Federlin, J.M. Kern, A. Rastegar, C. Dietrich-Buchecker, P.A. Marnot, J.P. Sauvage, New J. Chem. 14 (1990) 9. [82] T. Véry, D. Ambrosek, M. Otsuka, C. Gourlaouen, X. Assfeld, A. Monari, C. Daniel, Chemistry 20 (2014) 12901. [83] W.L. Parker, G.A. Crosby, The Journal of Physical Chemistry 93 (1989) 5692. [84] J.R. Kirchhoff, R.E. Gamache, M.W. Blaskie, A.A. Del Paggio, R.K. Lengel, D.R. McMillin, Inorg. Chem. 22 (1983) 2380. [85] R.M. Everly, D.R. McMillin, The Journal of Physical Chemistry 95 (1991) 9071. [86] Z.A. Siddique, Y. Yamamoto, T. Ohno, K. Nozaki, Inorg. Chem. 42 (2003) 6366.