Impact of Organo‐Functionalization on the Light‐Driven Hydrogen Evolution Reactivity of Molecular Molybdenum Sulfides Authors: Moritz Jahn, Luis Kammerer, Luca Schleicher, Amelie Meyer, Mariet Sibi Puthanagady, Alexander K. Mengele, Sven Rau, Stephan Kupfer, Benjamin Dietzek‐Ivanšić, Carsten Streb Published: 2026 ABSTRACT ABSTRACTMolecular molybdenum sulfides, or thiomolybdates, are well‐established noble‐metal free molecular catalysts for the hydrogen evolution reaction (HER). To‐date, there is a knowledge‐gap regarding the impact of organo‐functionalization on reactivity, stability and heterogenization of this compound class. Here, we report the development of synthetic routes for controlled introduction of organic N‐donor ligands as a first step toward establishing structure‐property‐reactivity relationships in this new compound class. Photophysical studies combined with computational modelling are used to rationalize trends in the observed light‐driven homogeneous HER activity. This work lays the foundation to develop organo‐functionalized thiomolybdate HER catalysts suitable for covalent or supramolecular linkage to photosensitizers and for anchoring on heterogeneous supports, e.g. photocathodes. Introduction As global energy demand rises and the energy technologies shift toward carbon neutrality, the demand for clean energy storage systems is growing rapidly. One of the most sustainable pathways to this end is the use of sunlight to produce green hydrogen, so that solar energy is converted to chemical energy, enabling a carbon‐free alternative to fossil fuels [1, 2]. Traditionally, catalysts for the hydrogen evolution reaction (HER) rely on rare and expensive noble metals such as platinum, making this approach undesirable for large‐scale deployment in industry [3]. Consequently, earth‐abundant HER catalyst alternatives have received tremendous interest [4]. Here, amorphous molybdenum sulfide catalysts have emerged as promising noble‐metal free alternative catalysts which can show comparable HER activity to Pt catalysts [5, 6, 7]. However, mechanistic studies on the molecular level of amorphous materials are inherently difficult. Therefore, molecular molybdenum sulfides, so‐called thiomolybdates, for example, the prototype [Mo3S13]2− (hereafter {Mo3}) have emerged as suitable active site models for photochemical and electrochemical HER studies [7, 8, 9]. However, identifying the active sites of these molecular well‐defined catalysts is challenging as the activity is highly dependent on the catalytic conditions and multiple active species can be formed, the exact speciation and reactivity are not fully understood. Previous mechanistic studies on {Mo3} indicate that ligand exchange of the terminal disulfides has a major influence on the HER activity [10]. Also, substitution of the terminal disulfides with chloride or bromide ligands is reported and results in catalytic systems with lower HER activity compared to the native {Mo3} [10]. Mechanistic studies to‐date propose two plausible HER mechanisms which proceed either via a disulfide/sulfide mechanism [10, 11, 12] or a molybdenum hydride mechanism [6, 13]. Note that to‐date, the true nature of the active HER site in {Mo3} remains unknown, and there is experimental and theoretical evidence for both proposed mechanisms, so that in fact, both mechanisms might (co‐)exist, and the dominant mechanisms depends on the exact reaction conditions [10, 11, 12, 13]. It is noteworthy that despite the insight into the HER‐activity dependence on ligand substitution, to‐date, only few organo‐functionalized {Mo3} derivatives have been reported [6, 14, 15, 16, 17, 18] and HER studies on these systems are exceedingly rare [19, 20]. In 2019 Schmehl and colleagues [19] introduced dithiocarbamate ligands to the [Mo3S7]4+ core, providing insights into the HER performance of the catalyst in the absence of terminal disulfide ligands [19]. The group reported structural changes of the catalyst under irradiation, involving the conversion of bridging disulfides to sulfides [19]. Recently Lin and coworkers [20] reported 2‐mercaptobenzothiazole substituted {Mo3}‐based catalysts with high acid‐base stability, and reported initial photocatalytic HER data [20]. Here, we build on these pioneering studies and report synthetic access and mechanistic insights for a series of organo‐functionalized {Mo3} derivatives with the general formula TEA[Mo3S7Br4L]Br (TEA = tetraethylammonium, L = ligand). As ligand L, we used chelating N‐donor ligands. The more flexible 2,2′‐bipyridine type was employed to investigate the most optimal coordination geometry at the molybdenum center allowed for by the free rotation around the C2─C2’ bond. Specifically 2,2′‐bipyridine (bpy); 4,4'‐diphenyl‐2,2'‐bipyridine (dpbpy) were employed. The more rigid and therefore demanding 1,10‐phenanthroline type was investigated by using 1,10‐phenanthroline (phen); bathophenanthroline (bphen); dipyrido[3,2‐a:2',3'‐c]phenazine (dppz); 7,8‐dimethyldipyridophenazine (dppzMe2). By utilizing the 4,5‐diazafluoren‐9‐one (fluo) similar rigidity of the ligand as observed with phenanthroline can be anticipated, however the relative position of the N‐donor atoms is significantly altered in comparison to the former. All ligands are shown in Figure 1b. We compared the light‐driven HER performance of these compounds and evaluated their activity with electrochemical, photophysical and theoretical studies, linking their HER activity to their reduction potential, catalyst‐photosensitizer interactions and frontier orbital energies. (a) Synthesis scheme of the organo‐functionalization of thiomolybdate clusters and (b) illustration of the N‐donor ligands employed, that is, 2,2′‐bipyridine (bpy); 1,10‐phenanthroline (phen); 4,4'‐diphenyl‐2,2'‐bipyridine (dpbpy); bathophenanthroline (bphen); dipyrido[3,2‐a:2',3'‐c]phenazine (dppz); 7,8‐dimethyldipyridophenazine (dppzMe2); 4,5‐diazafluoren‐9‐one (fluo). (c) Crystal structure of {Mo3Br}bpy, C (grey), H (white), Br (green), Mo (dark blue), N (blue), S (yellow), probability ellipsoids at 50%. Solvent molecules and TEABr were removed for clarity. (d) Crystal structure of {Mo3Br}bphen, C (grey), H (white), Br (green), Mo (dark blue), N (blue), S (yellow), probability ellipsoids at 50%. Solvent molecules and TEABr were removed for clarity. Results and Discussion Synthesis and Characterization {Mo3} was synthesized based on a literature‐reported procedure [10]. Next, {Mo3} was brominated using hydrobromic acid as reported by Fedin and colleagues [21] (resulting in [Mo3S7Br6]2−, {Mo3Br}), followed by substitution of the bromide ligands with the respective organic ligand (Figure 1a). The reaction of {Mo3Br} with the respective ligands was performed similar to a literature‐reported method [17] and adapted for each ligand. This gave the compounds TEA[Mo3S7Br4L]Br (hereafter {Mo3Br}L) (L = N‐donor ligand, for details see Figure 1a). All clusters were characterized by 1H‐NMR spectroscopy, IR spectroscopy, UV‐Vis spectroscopy, elemental analysis (EA) and where possible by single crystal x‐ray diffraction (SC‐XRD). Single crystals suitable for SC‐XRD were obtained for {Mo3Br}bpy and {Mo3Br}bphen (Figure 1c,d; see Supporting Information for crystallographic details) [22]. For both {Mo3Br}bpy and {Mo3Br}bphen one ligand is coordinated to a molybdenum center, resulting in a charge‐neutral cluster. Additionally, a bromide ion is bound near the cluster via the bridging disulfides, which is charge‐balanced by a TEA cation in the lattice. This unusual phenomenon has been observed previously [23] and assigned to weak disulfide‐bromide interactions. The crystal structure of {Mo3Br}bpy was previously reported by Nandi and colleagues [16]. As shown in Figure 1c,d, the ligands show virtually no structural distortion. 1H‐NMR spectroscopy of all compounds was performed in DMSO‐d6. The functionalized clusters were compared to the free ligand, all showing characteristic downfield shifts as well as splitting of the signals due to reduced symmetry of resulting species (for complete data, see Supporting Information). As expected, the strongest signal splitting is observed for the protons in proximity of the thiomolybdate cluster, indicating the spatial proximity of ligand and cluster also in solution. Light‐Driven HER Studies Light‐driven HER studies were performed for all organo‐functionalized catalysts and compared to {Mo3Br} as a reference. Measurements were performed using samples containing the respective catalyst (0.3 µM), [Ru(bpy)3](PF6)2 (20 µM) as photosensitizer and ascorbic acid (0.25 M) as sacrificial electron donor (SED) in MeOH containing 5 vol‐% N,N‐dimethyl formamide (DMF). Ascorbic acid was established previously as suitable SED for comparable catalytic systems using {Mo3} as catalyst [24]. Additionally ascorbic acid is known for tolerating acidic conditions, whereas basic conditions are usually necessary for other commonly used amine‐based SEDs to prevent protonation, which severely limits their electron‐donating capabilities [25]. The samples were irradiated at λ = 467 nm (LED) for 6 h, for details see Supporting Information. Hydrogen production was quantified via head‐space gas chromatography (GC) and all data shown were recorded multiple times to obtain meaningful averages and standard deviations (for details see, Table S2). The resulting turnover numbers (TONs) are shown in Figure 2. The native species {Mo3Br} shows a TON of ca. 3790 and for all ligand substitutions studied, we observe only modest drops in TON, highlighting that ligand substitution is possible while maintaining the overall light‐driven HER activity of the species. Notably, for {Mo3Br}fluo, we observe slightly higher TONs (ca. 3920) compared to the native {Mo3Br} species. Reference experiments performed without [Ru(bpy)3]2+ show strongly reduced catalytic activity, with TONs <50 (see Table S3), emphasizing the need of a photosensitizer for effective visible light HER. When compared to other thiomolybdate HER systems we note that the systems reported show comparable reactivity with other organo‐functionalized species (e.g., TON ca. 300 for dithiocarbamate functionalized systems [19]), however, we also note that these systems do not achieve the activity reported for the native [Mo3S13]2− (TONs > 23,000) [10]. TONs of all catalysts studied. Functionalized catalysts are abbreviated with the respective ligand. Conditions: catalyst (0.3 µM), photosensitizer [Ru(bpy)3](PF6)2 (20 µM), SED: ascorbic acid (0.25 M) solvent: MeOH containing 5 vol‐% DMF. LED irradiation, λ = 467 nm, irradiation time 6 h. To gain quantitative insights into the energy necessary to produce hydrogen in our system, we determined the photonic efficiency (PE, details see Supporting Information). The PE reaches from 2.87 × 10−3 for {Mo3Br}dppzMe2 to 3.78 × 10−3 for {Mo3Br}fluo. Thus, the PE is in the same order of magnitude as for previously reported data for {Mo3} [24]. In order to assess the stability of the catalysts, we investigated structural changes of the catalyst in solution by 1H‐NMR before and after irradiation (see Figures S24–S30). Since characterization of the catalyst under catalytic conditions is very challenging, we used a simplified system, enabling suitable concentrations for NMR spectroscopy and avoiding interfering peaks from other compounds. Therefore, 10 mM solutions of the respective catalysts were prepared in DMSO‐d6 and irradiated for 6 h using a 467 nm LED. The data show that the major NMR signals associated with the intact species are present after irradiation, while in some cases, some minor additional signals are observed. While the exact nature of these signals is currently unknown, the data demonstrate the overall stability and integrity of the systems, and no ligand exchange (i.e., no signals for “free” ligands) is observed in our studies. Electrochemistry and Computational Modelling To explore possible differences in electron transfer behavior, we explored the electrochemistry of the catalysts. In order to access the onset potential for cluster reduction, linear sweep voltammetry (LSV) was performed (Figure 3). All compounds except {Mo3Br} and {Mo3Br}fluo show a clear first reduction event between −0.92 V and −1.00 V vs the Fc+ /Fc° couple. The reduction event for {Mo3Br} and {Mo3Br}fluo are both cathodically shifted to more negative potentials (−1.35 V and −1.44 V, respectively) and show a more complex reduction pattern, which could indicate possible structural changes or more complex electron‐transfer processes upon reduction. The compounds with significantly more negative reduction potential also show higher average HER activity, but no clear trend between reduction onset potential and HER is observable, indicating other competing effects. These could include supramolecular interaction of photosensitizer and catalyst, aggregate formation, or photoinduced catalyst / photosensitizer degradation. (a) Linear sweep voltammograms of all catalysts (1 mmol/L) in de‐aerated DMF containing 0.1 mM nBu4NPF6; scan rate 100 mV/s; internal reference: Fc+/0. (b) Relationship between the reduction onset potential Ered. onset and the observed turnover number TON for the catalysts studied. (c) DFT‐calculated frontier orbital energy levels of {Mo3Br} and ligand‐substituted derivatives. Energies of ligand‐centered π(L)/π*(L) and metal‐centered d(CAT)/n(CAT) orbitals are shown relative to the d(CAT) orbital of {Mo3Br} (set to 0 eV). Vertical dashed lines indicate ligand groupings; orbitals are visualized for {Mo3Br}bpy. Furthermore, density functional theory (DFT) calculations (see Supporting Information for computational setup) were performed to investigate the energetics as well as the electronic characters of the frontier molecular orbitals along the series of thiomolybdate clusters. In particular, these quantum chemical simulations aimed to assess the energy levels of the lowest virtual orbitals localized on the {Mo3Br} cluster and the introduced aromatic ligand, and thus to elucidate the molecular moieties which are associated with the first reduction event (Figure 3c); the calculated orbital energies are shown in the Supporting Information (Table S8). The lowest‐unoccupied molecular orbitals (LUMOs) range from 1.68 eV for {Mo3Br}dppz to 3.34 eV for {Mo3Br} relative to the d(CAT) orbital of {Mo3Br} (set to 0 eV). Upon introduction of the organic ligand all thiomolybdate clusters localize the LUMO on the organic ligands. Notably, such ligand modification also affects the energy levels of the {Mo3Br}‐centered highest‐occupied molecular orbital (HOMO)—which is mainly given by the lone‐pairs of the bromide ligands—as well as of the respective cluster‐based LUMOs. Thus, introduction of these ligands allows us to tailor the electrochemical properties and in consequence the catalytic performance (see Table S7). Since the active H‐binding site is expected to be on the [Mo3S7]4+ core of the catalyst [6], a further stabilization of the cluster's lowest virtual orbital toward or even below the π(L)* orbital might be beneficial for HER activity. The LUMOs of {Mo3Br}fluo, {Mo3Br}dppz and {Mo3Br}dppzMe2 are mainly ligand‐centered with only little electron density on the cluster part, while the LUMOs for all other catalysts are distributed over the whole catalyst. It is noteworthy that in the case of {Mo3Br}fluo the π(L)* orbital is the closest in energy to the d(CAT) orbital. This observation may account for the enhanced catalytic performance of {Mo3Br}fluo. As the π‐system of the ligand extends, reduction at the ligand becomes progressively more favorable, thereby competing with and disfavoring reduction at the cluster. In the case of {Mo3Br}fluo, cluster‐centered reduction is therefore least impeded. Hence, electron‐withdrawing ligands featuring comparatively small π‐systems, like fluo, may have a positive effect on catalysis, whereas electron‐rich ligands with extended π‐systems, such as dppz or dppzMe2, may lead to reduced catalytic activity. Notably, this initial computational modelling merely allows us to assess the respective electrochemical properties along the series of molecular molybdenum sulfide clusters, while future joint synthetic‐spectroscopic‐theoretical investigations will focus on an in‐depth evaluation of the subsequent reaction pathways associated with HER activity. Stern–Volmer Analysis Next, we were interested to study solution interactions between the photosensitizer and the respective catalysts, which is relevant to understand catalytic performance, particularly since supramolecular π‐stacking interactions can control intermolecular / supramolecular charge‐transfer in related systems [26]. Here, we used emission quenching studies and Stern–Volmer analyses to gain insights into the solution behavior and interactions between the photosensitizer, [Ru(bpy)3]2+ and the three model catalysts {Mo3Br} (reference), {Mo3Br}fluo (highest TONs) and {Mo3Br}dppzMe2 (lowest TONs). Figure 4 summarises the results from emission quenching experiments where decay kinetics of the photosensitizer at 610 nm in the presence of three catalysts (a, c, e) shows a clear quenching trend in emission intensity. The decrease in the excited state lifetime of the photosensitizer (τ) with increase of quencher concentration [Q] suggests a dynamic quenching mechanism due to collisional interactions between excited state photosensitiser and catalyst and implies an oxidative quenching pathway under the given conditions. For all three systems, the bimolecular quenching rate constants are on the order of 108 M−1 s−1 (b, d, f), which confirms a moderate level of quenching with {Mo3Br}fluo showing a slightly higher bimolecular quenching rate (K q = 8.1 × 108 M−1 s−1) in comparison to {Mo3Br}dppzMe2 which shows the lowest (K q = 3.9 × 108 M−1 s−1) while {Mo3Br} (reference) lies in between these two extremes (K q = 5.5 × 108 M−1 s−1). This is in line with the observed light‐induced catalytic HER performance of the three systems, thereby, suggesting that organo‐functionalization of thiomolybdates can be used to control solution‐interactions between a photosensitizer and the respective catalyst. Decay kinetics (610 nm) of photosensitiser in presence of different concentration of the quencher at λ ex = 470 nm (a) {Mo3Br} (c) {Mo3Br}fluo (e) {Mo3Br}dppzMe2 . Stern–Volmer plots showing quenching efficiency of the quencher (b) {Mo3Br}, (d) {Mo3Br}fluo, and (f) {Mo3Br}dppzMe2 . Conclusion In summary, we report the first systematic synthetic access to thiomolybdate HER catalysts organo‐functionalized with aromatic N‐donor ligands. Catalytic studies show that the catalysts retain their HER activity after ligand exchange, thereby opening new paths for the development of organic‐inorganic hybrid HER catalysts. Initial mechanistic studies using photophysical analyses and theoretical computations provide first insights into the factors which affect HER‐activity and point to areas which require further study to rationalize the tuning scope for this new compound class. In future, we expect that these species can be used for the design of covalently or non‐covalently linked photosensitizer‐catalyst dyads, as well as for the cluster anchoring on heterogeneous substrates, for example, photoelectrodes. Conflicts of Interest The authors declare no conflicts of interest. Supporting information The authors have cited additional references within the Supporting Information [27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37]. Supporting file: chem71178‐sup‐0002‐DataFile.cif.