Photocatalytic Hydrogen Evolution Driven by a Heteroleptic

Article pubs.acs.org/IC

Cite This: Inorg. Chem. 2019, 58, 9127−9134

Photocatalytic Hydrogen Evolution Driven by a Heteroleptic Ruthenium(II) Bis(terpyridine) Complex Mira Rupp,†,‡ Thomas Auvray,‡ Elodie Rousset,‡,§ Gabriel M. Mercier,‡ Valeŕ ie Marvaud,§ Dirk G. Kurth,*,† and Garry S. Hanan*,‡ †

Downloaded via KEAN UNIV on July 24, 2019 at 05:38:09 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Lehrstuhl für Chemische Technologie der Materialsynthese, Julius-Maximilians-Universität Würzburg, Röntgenring 11, D-97070 Würzburg, Germany ‡ Département de Chimie, Université de Montréal, Montréal H3C 3J7, Québec Canada § Institut Parisien de Chimie Moléculaire, Sorbonne Université, CNRS UMR 8232, 4 place Jussieu, 75252 Paris Cedex 05, France S Supporting Information *

ABSTRACT: Since the initial report by Lehn et al. in 1979, ruthenium tris(bipyridine) ([Ru(bpy)3]2+) and its numerous derivatives were applied as photosensitizers (PSs) in a large panel of photocatalytic conditions while the bis(terpyridine) analogues were disregarded because of their low quantum yields and short excited-state lifetimes. In this study, we prepared a new terpyridine ligand, 4′-(4bromophenyl)-4,4‴:4″,4‴′-dipyridinyl- 2,2′:6′,2″-terpyridine (Bipytpy) and used it to prepare the heteroleptic complex [Ru(Tolyltpy)(Bipytpy)](PF6)2 (1; Tolyltpy = 4′tolyl-2,2′:6′,2′-terpyridine). Complex 1 exhibits enhanced photophysical properties with a higher quantum yield (7.4 × 10−4) and a longer excited-state lifetime (3.8 ns) compared to those of [Ru(Tolyltpy)2](PF6)2 (3 × 10−5 and 0.74 ns, respectively). These enhanced photophysical characteristics and the potential for PS−catalyst interaction through the peripheral pyridines led us to apply the complex for visible-light-driven hydrogen evolution. The photocatalytic system based on 1 as the PS, triethanolamine as a sacrificial donor, and cobaloxime as a catalyst exhibits sustained activity over more than 10 days under blue-light irradiation (light-emitting diode centered at 450 nm). A maximum turnover number of 764 was obtained after 12 days.

■

INTRODUCTION Ruthenium tris(bipyridine) complexes have become the archetypical model compound for transition-metal-ion-based photosensitizers (PSs) in photocatalytic reactions.1,2 For the ability to function as a PS, a long excited-state lifetime and a high luminescence quantum yield are necessary in order to efficiently transfer electrons to the catalyst.3 Ruthenium tris(bipyridine) complexes offer these properties, associated with a high degree of tunability, allowing one to adjust the redox potential of the complex and push its absorption to cover the entire visible spectrum.4−6 Numerous studies using ruthenium, as well as iridium or rhenium, PSs showed that the presence of interactions between the PS and catalyst leads to an enhancement of the overall photocatalytic activity.7−11 Recently, Mulfort and Tiede suggested that a PS capable of coordinating to the catalytic center could overcome diffusional constraints and enable efficient intermolecular electron transfer. Even though they did not observe activity in hydrogen production for their systems, their investigations and results confirm this to be a promising approach.12 Interestingly, they considered the use of a ruthenium bis(terpyridine) complex as the PS and showed that, through the suitable design of a donor−acceptor system, efficient electron transfer could be observed. This is surprising because ruthenium bis(terpyridine) complexes have been largely disregarded as PSs © 2019 American Chemical Society

because of their low quantum yield and short excited-state lifetime resulting from the low-lying nonradiative metalcentered excited state, which is easily populated at room temperature.13−15 However, the linear D2d symmetry of bistridentate complexes, which comes without stereoisomers compared to the 3-fold D3 symmetry of tris-bidentate complexes, offers a suitable connectivity for directional electron transfer. In addition, the coordination geometry of the bis(terpyridine) complex allows the design of linear arrays or polymers,16−18 whose properties can be tuned by using different spacers as well as substituents.19−21 Despite intensive efforts dedicated to improving their photophysical properties, ruthenium bis-tridentate complexes have, to the best of our knowledge, not been shown to photocatalyze hydrogen production.22−25 Herein, we report a newly designed heteroleptic complex capable of coordinating to the catalyst via peripheral pyridine substituents and discuss its photophysical and electrochemical properties as well as its potential as a PS in hydrogen production experiments. Received: March 12, 2019 Published: June 25, 2019 9127

DOI: 10.1021/acs.inorgchem.9b00698 Inorg. Chem. 2019, 58, 9127−9134

Article

Inorganic Chemistry Scheme 1. Synthetic Pathway toward Complex 1

■

C31H20BrN5 ([M + H]+): m/z 542.0975. Found: m/z 542.0887. Difference: 16.2 ppm. [Ru(Tolyltpy)Cl3]. A solution of ruthenium(III) chloride trihydrate (261 mg, 1.00 mmol, 1 equiv) and Tolyltpy (323 mg, 1.00 mmol, 1 equiv) in ethanol (125 mL) was heated to reflux for 3 h. Subsequently, the reaction mixture was filtered and the residue washed with ethanol to yield a red-brown solid (470 mg, 885 μmol, 89%). The product was used without further purification. [ R u ( T o l y l t p y ) (B ip y t py ) ] ( P F 6 ) 2 ( 1) . A s u s p e n s i o n o f [RuCl3(Tolyltpy)] (75.0 mg, 141 μmol, 1 equiv), Bipytpy (76.6 mg, 141 μmol, 1 equiv), and 4 drops of triethylamine in 15.0 mL of ethylene glycol was heated to 170 °C for 25 min using microwave irradiation. After cooling to room temperature, water and an aqueous KPF6 solution were added to the solution, and the precipitate was filtered off over Celite. After washing with water, it was dissolved in acetonitrile and dried over magnesium sulfate, and the solvent was removed under reduced pressure. Traces of ethylene glycol were removed by washing the crude product with water. It was further purified by column chromatography [silica gel, 7:1 MeCN:KNO3(aq)] to yield a dark-red solid (138 mg, 110 μmol, 78%). 1H NMR (400 MHz, CD3CN): δ 9.22 (s, 2H), 9.06 (s, 2H), 8.99 (d, 4J = 1.6 Hz, 2H), 8.79 (dd, 3J = 4.5 Hz, 4J = 1.6 Hz, 4H), 8.70 (d, 3J = 8.1 Hz, 2H), 8.25 (d, 3J = 8.6 Hz, 2H), 8.16 (d, 3J = 8.2 Hz, 2H), 7.99 (m, 4H), 7.76 (dd, 3J = 4.5 Hz, 4J = 1.7 Hz, 4H), 7.63 (d, 3J = 7.9 Hz, 2H), 7.60 (d, 3J = 5.9 Hz, 2H), 7.54 (dd, 3J = 5.9 Hz, 4J = 2.0 Hz, 2H), 7.48 (d, 3J = 4.9 Hz, 2H), 7.23 (ddd, 3J = 7.1 Hz, 3J = 5.6 Hz, 4J = 1.2 Hz, 2H), 2.55 (s, 3H). 13C{1H} NMR (151 MHz, CD3CN): δ 159.9, 159.2, 156.6, 156.3, 155.6, 153.8, 153.5, 151.8, 149.8, 148.2, 148.0, 143.7, 142.1, 139.2, 136.9, 133.8, 131.3, 130.6, 128.7, 128.5, 125.9, 125.5, 123.4, 123.3, 122.9, 122.5, 122.5, 21.4. ESI-MS. Calcd for C53H37BrF6N8PRu ([M − PF6]+): m/z 1111.10044. Found: m/z 1111.09792. Difference: 3.1 ppm. Anal. Calcd for C53H37BrF12N8P2Ru·2H2O: C, 49.24; H, 3.20; N, 8.67. Found: C, 48.84; H, 3.06; N, 8.36.

EXPERIMENTAL SECTION

Materials. The ligand 4′-tolyl-2,2′:6′,2′-terpyridine (Tolyltpy) was prepared according to a published procedure.26 The synthesis of the precursor 2-acetyl-4,4′-bipyridine was optimized in our laboratory from a previous work.27,28 Complex 1 was prepared via a two-step procedure adapted from the literature.29 All reagents and solvents were obtained from commercial sources (VWR, Fisher Scientific, Acros, or Sigma-Aldrich) and used as received, unless stated otherwise. RuCl3·3H2O was obtained from Pressure Chemicals. Synthesis. 2-Acetyl-4,4′-bipyridine. A solution of 4,4′-bipyridine (3.00 g, 19.2 mmol, 1 equiv), iron(II) sulfate heptahydrate (267 mg, 960 μmol, 0.05 equiv), acetaldehyde (5.40 mL, 4.23 g, 96.0 mmol, 5 equiv), tert-butyl hydroperoxide (70% by wt in H2O, 26.6 mL, 17.3 g, 192 mmol, 10 equiv), and trifluoroacetic acid (7.40 mL, 11.0 g, 96.0 mmol, 5 equiv) in acetonitrile (60 mL) was heated to reflux for 90 min. Subsequently, acetonitrile was removed under reduced pressure, and the pH of the residue was raised to approximately 9 by adding 1 M aqueous sodium carbonate. After extraction with dichloromethane, the organic phase was dried over magnesium sulfate and the solvent removed under reduced pressure. The crude product was purified by column chromatography (silica gel, ethyl acetate) to yield the pure product as an off-white solid (1.60 g, 8.07 mmol, 42%). 1H NMR (400 MHz, CDCl3): δ 8.81 (d, 1H, 3J = 5.1 Hz, PyH6), 8.77 (dd, 2H, 3 J = 4.4 Hz, 4J = 1.6 Hz, Py′H2,6), 8.31 (d, 1H, 4J = 1.2 Hz, PyH3), 7.72 (dd, 1H, 3J = 5.1 Hz, 4J = 1.9 Hz, PyH5), 7.59 (dd, 2H, 3J = 4.5 Hz, 4J = 1.7 Hz, Py′H3,5), 2.78 (s, 3H, CO(CH3)). 13C{1H} NMR (101 MHz, CDCl3): δ 26.0, 119.4, 121.5, 124.6, 144.9, 146.7, 150.1, 150.9, 154.5, 199.9. Elem anal. Calcd for C12H10N2O: C, 72.71; H, 5.08; N, 14.13. Found: C, 72.43; H, 5.11; N, 13.67. HRMS (ESITOF). Calcd for C12H10N2NaO ([M + Na]+): m/z 221.0685. Found: m/z 221.0685. 4′-(4-Bromophenyl)-4,4‴:4″,4‴′-dipyridinyl-2,2′:6′,2″-terpyridine (Bipytpy). To a solution of 2-acetyl-4,4′-bipyridine (500 mg, 2.52 mmol, 2.5 equiv) in ethanol (25 mL) were added 4bromobenzaldehyde (187 mg, 1.01 mmol, 1 equiv), potassium hydroxide (283 mg, 5.04 mmol, 5 equiv), and aqueous ammonia (28% by wt, 7.02 mL, 6.32 g, 50.4 mmol, 50 equiv). The reaction mixture was heated to reflux for 2 weeks. Subsequently, water was added to precipitate the product, which was filtered over Celite, washed with water and methanol, and dissolved in dichloromethane. The solution was dried over magnesium sulfate, and the solvent was removed under reduced pressure. The crude product was suspended in methanol and the slurry filtered to give the product as a white solid (251 mg, 463 μmol, 46%). 1H NMR (300 MHz, CDCl3): δ 8.89 (d, 2H, 4J = 1.8 Hz, PyH3/Py″H3), 8.86 (dd, 2H, 3J = 5.1 Hz, 4J = 0.7 Hz, PyH6/Py″H6), 8.80 (dd, 4H, 3J = 4.5 Hz, 4J = 1.7 Hz, Py‴H2,6/ Py‴′H2,6), 8.79 (s, 2H, Py′H3,5), 7.81 (dd, 3J = 6.6 Hz, 4J = 2.0 Hz, 2H, PhH), 7.68 (m, 6H, PhH/Py‴H3,5/Py‴′H3,5), 7.62 (dd, 3J = 5.1 Hz, 4J = 1.8 Hz, 2H, PyH5/Py″H5). 13C{1H} NMR (101 MHz, CDCl3): δ 157.1, 156.0, 150.9, 150.3, 149.6, 146.8, 146.1, 137.3, 132.4, 129.0, 123.9, 121.8, 121.7, 119.5, 119.3. ESI-MS. Calcd for

■

RESULTS AND DISCUSSION Synthesis and Structural Characterizations. Synthesis of the ligand Bipytpy was carried out using 2-acetyl-4,4′bipyridine analogously to a one-pot reaction described previously.26 Synthesis of this monoacetylated precursor was previously mentioned as traces in the acetylation of 4,4′bipyridine.27 The careful investigation leading to the new protocol used here allowed us to reduce the amount of polyacetylated products and isolate the desired compound in good yield and purity. The heteroleptic complex is prepared via a two-step synthesis, adapted from Fallahpour et al.29 First, a ruthenium(III) complex is formed by the reaction of hydrated RuCl3 with Tolyltpy. This complex is then combined with Bipytpy under microwave conditions to yield the heteroleptic ruthenium(II) complex 1 in 78% yield (Scheme 1). Both the 9128


Article

Inorganic Chemistry ligand and complex are characterized by 1H and 13C NMR spectroscopy (Figures S1, S2/S3, and S4/S5 for 2-acetyl 4,4′bipyridine, Bipytpy, and complex 1, respectively) as well as high-resolution mass spectrometry (HRMS). X-ray analysis was performed on a single crystal of complex 1 obtained by the slow diffusion of diethyl ether into the concentrated acetonitrile solution of 1 [see section 3 of the Supporting Information (SI) and Figure 1]. The central

with torsion angles around 23 and 45°, as observed in other polypyridyl-based systems.9 The geometry of complex 1 was compared to the optimized geometry at the density functional theory (DFT) level using the PBE0 hybrid functional31 and a LanL2DZ basis set.32−35 The solvent (acetonitrile) was modeled as a conductor-like polarized continuum. A bond length and angle comparison between the experimental and calculated values showed good agreement, validating the reliability of the theoretical model (Table S2). Electronic Properties. The nonsymmetric design of complex 1 involves two terpyridine ligands with different electron densities in order to facilitate the transfer of electron density toward the more electron-accepting Bipytpy ligand in the excited state, thus favoring directional electron transfer. The success of our design is confirmed by the electrochemical experiments conducted in acetonitrile, whose results are summarized in Table 1. The cyclic voltammogram of complex 1 (Figure S8) shows one reversible oxidation and three reversible reduction processes. The oxidation process at +1.21 V versus saturated calomel electrode (SCE) usually corresponds to the RuII-toRuIII oxidation. Three reduction waves were assigned by comparison to those of the homoleptic complex [Ru(Tolyltpy)2]2+ measured in the same conditions. The first reduction of 1 is at a slightly less negative potential than that for the homoleptic complex and is expected to correspond to reduction of the Bipytpy ligand. Indeed, Bipytpy is easier to reduce because of its electron-withdrawing substituents and the extended conjugated π system. The second reduction process happens at a potential similar to that of the homoleptic complex. It thus likely corresponds to reduction at the Tolyltpy ligand. The third cathodic wave is absent in the homoleptic complex and would correspond to a second reduction of the Bipytpy ligand. These assignments are corroborated by the theoretical DFT calculations performed on complexes 1 and [Ru(Tolyltpy)2]2+ (see the details in section 5 of the SI). Selected molecular orbitals of complex 1 are presented in Figure 2 (additional orbitals along with their respective molecular contributions are presented in Table S3 for 1 and Table S4 for [Ru(Tolyltpy)2]2+). While in [Ru(Tolyltpy)2]2+ the lowest unoccupied molecular orbital (LUMO) and LUMO +1 are delocalized on both ligands, in 1 the LUMO and LUMO+1 are centered on Bipytpy and Tolyltpy, respectively, as shown in Figure 2. This clear separation between the Bipytpy and Tolyltpy centered orbitals confirms the success of our design for directional electron transfer because the first injected electron will be localized on Bipytpy, which bears the peripheral coordination site introduced for catalyst−PS interaction. The impact of our design on the photophysical properties of the complex was then analyzed using UV−vis absorption spectroscopy. The spectrum is shown in Figure 3, and the

Figure 1. Ellipsoid representation of 1 at 50% probability. Hydrogen atoms, one PF6− counterion, and a cocrystallized acetonitrile molecule are omitted for clarity.

ruthenium ion is embedded in a distorted octahedral coordination sphere, imposed by the restricted bite angle of the two meridionally coordinated terpyridine ligands. The Tolyltpy ligand is slightly bent because of intramolecular interactions: weak stacking between the terminal tolyl group and one of the lateral pyridines of Tolyltpy of a neighboring complex and hydrogen bonding involving the pendant pyridines (Figures S6 and S7). A similar bending induced by interactions in the crystal packing was recently reported by Beves et al. in their work of crystal engineering using heteroleptic ruthenium bis(terpyridine) complexes as building blocks.30 Finally, the structure shows that the pendant pyridines are not fully coplanar with the terpyridine unit,

Table 1. Electrochemical Half-Wave Redox Potentials of Complex 1 Compared to Those of [Ru(Tolyltpy)2]2+ in Acetonitrilea [Ru(Tolyltpy)(Bipytpy)] [Ru(Tolyltpy)2]2+

2+

EOx 1/2

ERed1 1/2

ERed2 1/2

ERed3 1/2

+1.21 (85) +1.22 (69)

−1.17 (75) −1.24 (42)

−1.46 (71) −1.48 (67)

−1.66 (86)

a Potentials are given in volts versus SCE; measurements in N2-purged acetonitrile solutions containing 0.1 M tetrabutylammonium hexafluorophosphate at a sweep rate of 100 mV s−1; the differences between the cathodic and anodic peak potentials are given in parentheses in millivolts.

9129


Article

Inorganic Chemistry

Figure 3. UV−vis absorption (black) and normalized emission spectra (gray) of complex 1 (solid line) and the reference complex [Ru(Tolyltpy)2]2+ (dashed line) measured in acetonitrile at 20 °C.

optically dilute measurement method37 with [Ru(Tolyltpy)2]2+ as the reference.38 The data are summarized in Table 2. The maximum of emission is slightly red-shifted in comparison to the homoleptic Tolyltpy complex, with values of 658 and 636 nm for 1 and [Ru(Tolyltpy)2]2+, respectively. As expected, the quantum yield Φc and excited-state lifetime are significantly lower than those for ruthenium(II) tris-bidentate complexes, with the quantum yield and lifetime of [Ru(bpy)3]2+, for instance, reported at 9.5% and 855 ns, respectively (tabulated in Table 2).38 However, when complex 1 is compared with [Ru(Tolyltpy)2]2+, the quantum yield is increased by an order of magnitude, from 0.003% to 0.074%, and the excited-state lifetime is increased from 0.74 to 3.8 ns (a luminescence decay analysis is presented in Figure S5).38 This excited-state lifetime can allow reductive or oxidative quenching to take place more efficiently in the context of a photocatalytic experiment. Interestingly, Sauvage et al. reported a bis(terpyridine) complex, namely, [Ru(tptpy)2]2+ (where tptpy is 4,4′,4″triphenyl-2,2′:6′,2″-terpyridine), with similar photophysical properties. This complex was shown to be capable of reducing methyl viologen under light irradiation in the presence of triethanolamine (TEOA) as a sacrificial electron donor, while [Ru(Tolyltpy)2]2+ gave no radical formation in similar conditions.39 To further characterize the complex, spectroelectrochemical measurements monitoring changes in absorption during oxidation and reduction were conducted, and the results are shown in parts A and B of Figure 4, respectively. In oxidation, the most significant change is the decrease of the MLCT absorption band around 500 nm. Concomitantly, an increasing absorption band around 400 nm is observed and can be assigned to a ligand-to-metal charge-transfer (LMCT) transition, possibly because of oxidation of the metal-centered HOMO. On the other end, in reduction, the intensity of the MLCT band at 500 nm hardly changes, but it becomes broader, indicating that this transition is not as affected by reduction as by oxidation of the complex. The electrochemically reversible reduction processes observed in cyclic voltammetry already led to the assumption that the reduction processes are LC, resulting in a stable d6 complex. The spectroelectrochemical experiments confirm this conclusion. The new peak around 420 nm could be assigned to the LMCT

Figure 2. Representation of selected molecular orbitals of complex 1 with their calculated energy (isocontour value 0.03 au).

major bands are summarized in Table 2, along with the value for the homoleptic Tolyltpy complex. The lower-energy transition is slightly red-shifted compared to the homoleptic complex (497 against 491 nm for 1 and [Ru(Tolyltpy)2]2+ , respectively); in the literature, this transition is attributed to metal-to-ligand charge transfer (MLCT).14 A comparison between the experimental absorption spectrum and time-dependent DFT (TD-DFT)-calculated transitions is available in Figure S9 (the values are reported in Table S5 and the natural transition orbitals36 of the lowerenergy transitions in Tables S6 and S7 for both complexes). Calculations validate the attribution of the low-energy transition, identifying this band as a superimposition of transitions between highest occupied molecular orbital (HOMO), HOMO−1, and HOMO−2, mainly centered on the metal, and the LUMO, LUMO+1, LUMO+2, and LUMO +3, centered on the ligands. The higher-energy bands visible in the absorption spectrum are attributed to ligand-centered (LC) transitions. Emission spectra were recorded in argon-purged acetonitrile solutions at 20 °C (temperature controlled with a thermostat system). The quantum yields were determined using the 9130


Article

Inorganic Chemistry

Table 2. UV−Vis Absorption and Emission Data of Complex 1 Compared to Those of [Ru(Tolyltpy)2]2+ and [Ru(bpy)3]2+ (Spectra Measured in Acetonitrile at 20° C) λmax abs/nm (ε/×103 L mol−1 cm−1) LC 2+

[Ru(Tolyltpy)(Bipytpy)] [Ru(Tolyltpy)2]2+ [Ru(bpy)3]2+

234 (76) 285 (58)

301 (70) 310 (64)

331 (62)

MLCT

λmax em/nm

Φc × 10−5

τc/ns

497 (30) 491 (23) 450 (16)a

658 636 619a

74 ± 33 3 ± 1a 9500a

3.8 ± 0.1 0.74 ± 0.02a 855a

a

From ref 38.

Figure 4. UV−vis absorption spectra of 1 in argon-purged acetonitrile: (A) at oxidation potentials from 0.0 to 1.5 V versus Ag/Ag+; (B) at reduction potentials from 0.0 to −2.0 V versus Ag/Ag+. The reduction potentials are indicated by different colors.

transition, while the peak around 370 nm would correspond to a red-shifted LC transition. Finally, considering the use of 1 in photocatalytic applications, the redox potentials for the excited complex 1* were estimated using the following equations: E*ox = Eox − Eem and E*red = Ered + Eem, where Eem (eV) ≈ 1240/ λem (nm), giving potentials of −0.66 and 0.71 V versus SCE for oxidation and reduction, respectively. Interestingly, these values are comparable to those of [Ru(bpy)3]2+ estimated at −0.72 and 0.68 V versus SCE for oxidation and reduction of the excited complex, respectively. Photocatalytic Experiments. On the basis of its photophysical properties, the potential of complex 1 to act as a PS in hydrogen evolution was investigated. We used standard photocatalytic conditions, with DMF as a solvent, TEOA as a sacrificial electron donor, and tetrafluoroboric acid (HBF4) as a proton source. The activity was followed by gas chromatography (experimental details are given in the SI, section 6). The turnover frequency (TOF) variation (in mmol of H2 per mol of PS per minute) over 24 h for complex 1 as a PS with different catalysts is presented in Figure 5, while the photocatalyzed hydrogen evolution results are summarized in Table 3. No hydrogen was observed in these photocatalytic conditions without light or catalyst or when using [Ru(Tolyltpy)2]2+ (entries 5−7 in Table 3). Two types of catalysts were used: either K2[PtCl4] as the precursor for platinum colloids as heterogeneous active species or cobaloxime complexes as homogeneous catalysts. The cobaloxime catalysts are used with excess free dimethylglyoxime ligand (dmgH2). We used [Co(dmgH)(dmgH2)Cl2], which is a cobalt(III) catalyst, or [Co(dmgH)2(OH2)2](BF4)2 prepared in situ with [Co(H2O)6](BF4)2 as a cobalt(II) source. These catalysts are known from previous investigations to be efficient in hydrogen production experiments.40−42 The maximum TOF for complex 1, i.e., at peak activity, appears to be similar to those of all catalysts within experimental error (±10%), around 50 mmolH2 molPS min−1.

Figure 5. Comparison of the hydrogen evolution activity of complex 1 with different catalysts. Inset: First 4 h showing the induction period for the cobalt(III) catalyst. Conditions: 0.1 mM complex 1 in DMF with 1 M TEOA (sacrificial electron donor) and 0.1 M HBF4 (acid source), irradiated from below with a blue LED (centered at 450 nm, output 62 mW).

This leads us to believe that reductive quenching by TEOA is the limiting step, which would be coherent with the modest photophysical properties of our PS. When using Pt as a catalyst (black dotted curve in Figure 5: entry 1 in Table 3), no induction period is observed, while hydrogen production with [CoIII(dmgH)(dmgH2)Cl2] (blue curve: entry 2 in Table 3) only starts after an induction period of several hours. The absence of induction with platinum is expected because no intermediate species needs to accumulate. However, if one considers the mechanism proposed for cobalt catalysts in the literature and briefly presented in Scheme 2,43,44 it appears that the initial cobalt(III) complex has to be reduced to its cobalt(I) state, which can form an hydride complex. Thus, the cobalt(II) species needs to accumulate first; otherwise, any 9131


Article

Inorganic Chemistry Table 3. Hydrogen Production Results under Blue-Light Irradiation (450 nm and 62 mW)a entry 1 2 3 4 5 6 7 8 9 10 11

PS

(pre)catalyst

1 1 1 1 1 1 [Ru(Tolyltpy)2]2+ 1 1 [Ru(bpy)3]2+ 1

K2[PtCl4] (50 μM) [Co(dmgH)(dmgH2)Cl2]d [Co(dmgH)(dmgH2)Cl2]e [Co(H2O)6](BF4)2 (1 mM)d none [Co(dmgH)(dmgH2)Cl2]d,f [Co(dmgH)(dmgH2)Cl2]d [Co(dmgH)(dmgH2)Cl2]d,g [Co(H2O)6](BF4)2d,g [Co(H2O)6](BF4)2d [Co(dmgH)(dmgH2)Cl2]d

TOFmax, mmolH2 molPS min−1

c

56 53 54 57

49 46 6850 51

nH2,b μmol

TON

induction h

60 63 62 69 N.D. N.D. N.D. 31h 29h 400i 395j

116 122 119 134

Photocatalytic Hydrogen Evolution Driven by a Heteroleptic

Recommend Documents