Synthesis, structure & diphenylacetylene reduction reactivity of a carbide-supported Fe₄Mo₂ carbonyl cluster: a higher Fe-valence hydride intermediate for enhanced selectivity

Abstract

The novel four-iron, two-molybdenum cluster [(µ₆-C)Fe₄Mo₂(CO)₁₈]²⁻ (2) containing an interstitial carbide has been structurally characterized and prepared from the corresponding Fe₄ dianion [(µ₄-C)Fe₄(CO)₁₂]²⁻ (1) supported by two crowned alkali units in [K(benzo-18-crown-6)]⁺. In the X-ray structure of 2, the two molybdenum atoms share an internal geometry of a cis orientation about the central carbide, which indicates the stability of the well-known ‘butterfly’ Fe₄C core found in the precursor 1, and this finding is consistent with other bis-heterometal variants of general formula [Fe₄M₂] (M = Ni, Cu, Rh, Au). Reactivity studies monitoring the reduction of diphenylacetylene (DPA) catalyzed by 2 showed selective reduction to cis-diphenylethylene. The effect of proton source pK_a and steric bulk on DPA reduction demonstrated that product selectivity is enhanced with increased steric bulk near the protonation site, and that conversion increases with more acidic proton sources. Overall, increased selectivity is observed with 2 compared with catalyst-free reactions, Fe-only clusters and the Fe₅Mo variant. We attribute the beneficial reactivity profile of cluster 2 to the electronic effect of the two Mo(0) centers, which lead to higher valent iron sites in the hydride intermediate cluster(s), thus decreasing non-specific reduction and increasing selectivity.

---

Introduction

Renewed interest in iron-carbide clusters in the last decade has spurred the development of heterometal congeners of such clusters, which may prove useful as a means of providing access to higher valent clusters while maintaining the interstitial carbide. Such clusters have found utility as catalysts in Fischer–Tropsch,¹ H₂ splitting and proton reduction^2–5 as well as CO and CO₂ reduction.^6,7 Additionally, structural relevance to the six-coordinate carbide of nitrogenase led DeBeer et al. to conduct X-ray emission studies that provided insight into the electronic structure of the delocalized orbitals of the iron-carbide core.^5,8,9 Research into converting bound carbon-based ligands (CO, CS₂, CN) or carbon-based electrophiles (CS, CI₄, CH₂I₂, CF₂Br₂ etc.) to a carbide unit has thus far proven unfruitful – our work included¹⁰ – and thus developing a fundamental understanding of carbide as a ligand as yet remains in the family of organometallic, low-valent metal carbonyl clusters – some with bio-relevant metals (from nitrogenase: namely Fe, Mo and V).

Since the report of the first tetra-iron-carbide cluster [Fe₄(μ₄-C(COOMe))(CO)₁₂]⁻,¹¹ the ‘open-face’ carbide cluster has demonstrated increased reactivity compared with its more ‘closed’ penta- and hexa-iron counterparts, thus serving as a useful synthon for the development of heterometal clusters.^3,12–14 A survey of the Cambridge Crystallographic Data Center (CCDC) reveals a number of published Fe₄M₂ clusters (M = d block metal) largely based upon heterometal addition to the tetra-iron-carbide cluster [Fe₄(μ₄-C)(CO)₁₂]²⁻ and the analogous penta-iron-carbide cluster, or alternatively heterometal substitution of hexa-iron-carbide clusters. Regarding 3d metals, reports employing copper as the heterometal demonstrate the utility of a variety of copper-ligand motifs. The limited number of structures containing any d block heterometal (only Ni, Cu, Rh, Au), however, indicate limited synthetic investigations and provide an opportunity for further investigation.^15–19 The Fe₄M₂ clusters bear a variety of alternative non-CO ligands, including halides, phosphines and nitriles. The oxidation state of such clusters ranges from 0 to −2 with the cluster [Fe₄(μ₆-C)(CO)₁₄(CuCl)₂]²⁻ exhibiting the highest average metal oxidation state of +1.¹⁷

Our group has examined ligand substitutions of the well-characterized Fe₆ clusters to understand the reactivity of hexairon-carbide clusters. Previous work has explored the substitution of electrophilic sulfur sources (S₂Cl₂ and S₈) and thiolate (PhS-Cl) on Fe₆²⁻ clusters, thus affording the first reported thiolato-iron-carbide complex.²⁰ We also demonstrated that by using in situ oxidation, reactive, polyhedral skeletal electron pair theory (PSEPT)–non-conforming clusters can be generated that are more substitutionally active, in contrast to the substitutionally inert Fe₆²⁻ cluster. For example, our recent report elucidated guidelines for cluster coordination of electron-starved clusters by multidentate phosphine ligands: single iron site coordination leads to cluster disproportionation, whereas multi-iron site chelation provides intact penta-iron clusters.¹⁴ Applying the same approach we have also investigated substitution with phosphides and most importantly thiolates, leading to a novel thiolate-bound intact hexa-iron carbide cluster [Fe₆(μ₆-C)(μ₂-Stol)(μ₂-CO)₂(CO)₁₄]³⁻.²¹ This previous work provides insight into successful routes for substitution reactions on six-iron clusters. In this work, we seek to establish a new platform for investigating downstream ligand/metal substitution reactions.

A previous report from our group utilized molybdenum as a heteroatom for penta-iron-carbide clusters in an [Fe₅Mo(μ₆-C)(μ₂-CO)₃(CO)₁₄]²⁻ cluster^22,23 (selected for its nominal relevance to the active site of nitrogenase), and demonstrated the cluster's catalytic activity towards selective reductions of diphenyl acetylene. While previously reported,²⁴ [Fe₄Mo₂(μ₆-C)(μ₂-CO)₂(CO)₁₆]²⁻ (Fe₄Mo₂²⁻, 2) was not crystallographically characterized. In this work, we synthesize and utilize the novel di-molybdenum iron-carbide cluster [Fe₄Mo₂(μ₆-C)(μ₂-CO)₂(CO)₁₆]²⁻ (2) to examine its structural parameters and reactivity toward reductions of diphenylacetylene (DPA).

Results & discussion

Synthesis

Synthesis of the di-anionic [Fe₄(μ₄-C)(CO)₁₂]²⁻ was accomplished through a series known reactions starting from the well-characterized [Fe₆(μ₆-C)(μ₂-CO)₄(CO)₁₂]²⁻ cluster (Scheme 1), followed by selective removal of two iron sites using FeCl₃ (as an inner-sphere oxidant), thus generating the acetyl(ester)-capped carbide cluster [Fe₄(μ₄-C(COOMe))(CO)₁₂]⁻.²⁵ This four-iron cluster was protonated with triflic acid to extrude methanol, thus providing the carbide-authentic neutral cluster [Fe₄(μ₄-C)(CO)₁₃].²⁵ Subsequent reduction with KC₈ generates the key dianionic intermediate [Fe₄(μ₄-C)(CO)₁₂]²⁻ (1).²⁵

Scheme 1 Synthetic pathway to generate [K(benzo-18-crown-6)]₂[Fe₄Mo₂(μ₆-C)(CO)₁₆(μ₂-CO)₂] (2).

In our previous report of the synthesis of the penta-iron [Fe₅Mo(μ₆-C)(μ₂-CO)₃(CO)₁₄]²⁻ cluster, addition of a stoichiometric amount of Mo(cycloheptatriene)(CO)₃ to the five-iron cluster [Fe₅(μ₅-C)(μ₂-CO)₂(CO)₁₂]²⁻ resulted in the formation of the target mono-molybdenum cluster [Fe₅Mo(μ₆-C)(μ₂-CO)₃(CO)₁₄]²⁻. In contrast, stoichiometric addition of 1–2 equiv. of Mo(cycloheptatriene)(CO)₃ to 1 led only to a mixture of products (postulated from preliminary crystal structure data) that co-crystallized and proved inseparable. Thus, to isolate a pure crystalline sample of [K(benzo-18-c-6)]₂[Fe₄Mo₂(μ₆-C)(μ₂-CO)₂(CO)₁₆] (2), it proved necessary to perform the reaction with an excess of Mo(chpt)(CO)₃ (5 equiv.) and mild heating (80 °C) to drive the reaction to completion.

X-ray structures of 1 & 2

Similar to previous reports on the structure of tetra-iron-carbide clusters,²⁵ the crystal structure data for 1 (Fig. S3) revealed a four-coordinate carbide encompassed by four Fe atoms fixed in a butterfly geometry. The crystal structure of 2 (Fig. 1) reveals that the core Fe₄ ‘butterfly’ motif remains unchanged from 1, imposing cis-Mo coordination about the carbide. This is expected as there is no CCDC structure for a tetra-iron–heterometal carbide or nitride cluster that does not preserve the Fe₄ butterfly motif. Upon insertion of the two molybdenum centers, the average Fe–Fe bond contract and Fe–C_carbide bond lengths become elongated to 2.61(12) Å and 1.91(4) Å, respectively. A trend in these contacts becomes apparent upon comparing the Fe₆, Fe₅Mo and Fe₄Mo₂ cores (Table 1). As Fe atoms are discretely replaced with Mo atoms, the Fe–Fe bonds shorten, while the Fe–C_carbide bonds elongate. In contrast to the comparison of Fe₆ and Fe₄ (in which the Fe₄C motif becomes overall compressed), the trend suggests that the presence of the molybdenum atoms promote a displacement of the carbide away from the Fe₄ unit.

Fig. 1 ORTEP diagram (50% thermal ellipsoids) of 2. Hydrogens are omitted for clarity. A complete ORTEP diagram of the full asymmetric part of the unit cell is included in SI (Fig. S4).

Table 1 Average metal-carbide and metal–metal bond distances (Å) for [K(benzo-18-crown-6)]₂[Fe₄(μ₄-C)(CO)₁₂] (1), (NEt₄)₂[Fe₆(μ₆-C)(μ₂-CO)₄(CO)₁₂],²⁶ [K(benzo-18-crown-6)]₂[Fe₅Mo(μ₆-C)(μ₂-CO)₃(CO)₁₄]²³ and [K(benzo-18-c-6)]₂[Fe₄Mo₂(μ₆-C)(μ₂-CO)₂(CO)₁₆] (2)

Bond	Fe4 (1)	Fe6	Fe5Mo	Fe4Mo2 (2)
Fe–Ccarbide	1.87 ± 0.11	1.88 ± 0.01	1.90 ± 0.05	1.91 ± 0.04
Fe–Fe	2.62 ± 0.04	2.66 ± 0.06	2.65 ± 0.01	2.61 ± 0.12
Mo–Ccarbide	—	—	2.11 ± 0.06	2.11 ± 0.02
Mo–Fe	—	—	2.91 ± 0.03	2.89 ± 0.05

---

Diphenylacetylene reduction

The reactivity of the Fe₄Mo₂ cluster was examined and compared to the iron-only Fe₆ cluster in its activity for the reduction of diphenylacetylene (DPA). A similar reduction of ethyne is reported with the native nitrogenase enzyme and follows literature precedent for the activation of alkynes by iron carbonyls by bonding of the alkyne to the iron centers, enforcing certain conformations leading to enhanced selectivity.^27,28 The reduction of DPA was pursued to provide insight into selective reductions of triply-bonded functional groups. Results are included below in Table 2, for the following general (unbalanced) reaction (Scheme 2).

Scheme 2 Reduction of DPA and possible product outcomes.

Table 2 Selected results from reductions of DPA to cis/trans-diphenylethylene and diphenylethane

Catalyst	Reductant	Proton source	pKa	Temp (°C)	Convn (%)	C=C/C–C ratio	cis/trans ratio
None	Na2(per)	[Me3AnH]OTf	4.37	rt	0.09	C=C only	0.89
Fe6	Na2(per)	[Me3AnH]OTf	4.37	rt	17.70	1.18	0.05
Fe4Mo2	Na2(per)	[Me3AnH]OTf	4.37	rt	16.98	1.09	0.03
Fe4Mo2	Na2(per)	[Me3AnH]OTf	4.37	−20	0.43	0.22	0.50
Fe4Mo2	Na2(per)	[Me3AnH]OTf	4.37	60	1.08	0.19	0.95
None	Na2(per)	[^tBu3AnH]OTf	3.30	rt	15.88	12.58	0.01
Fe6	Na2(per)	[^tBu3AnH]OTf	3.30	rt	2.68	7.62	0.22
Fe4Mo2	Na2(per)	[^tBu3AnH]OTf	3.30	rt	1.51	9.95	0.96
Fe4Mo2	Na2(per)	[^tBu3AnH]OTf	3.30	−20	0.27	9.82	0.61
Fe4Mo2	Na2(per)	[^tBu3AnH]OTf	3.30	60	1.20	9.37	0.28
Fe6	Na2(per)	[Me2AnH]OTf	3.95	rt	60.23	1.32	∼0
Fe4Mo2	Na2(per)	[Me2AnH]OTf	3.95	rt	11.84	3.31	0.01

---

Initial investigations consisted of control reactions without cluster for each proton source. Sodium perylinide [Na₂(per)] was selected over a stronger reductant (e.g. KC₈), as our previous report demonstrated that while stronger reductants lead to increased conversion, they also result in lower selectivity for the cis-alkene product; thus, we opted for the reductant that provides greatest selectivity. For the catalyst-free reaction of Na₂(per) with 2,4,6-trimethylanilinium triflate ([Me₃AnH]OTf, Scheme 3) a near-zero conversion of 0.09% was obtained, but with surprising selectivity towards the intermediately reduced diphenylethylene products (no observed diphenylethane), and relatively high selectivity towards the desired cis product (0.89 cis/trans). Upon addition of the Fe₆ cluster under the same conditions, conversion was substantially increased (17.7%), but selectivity was decreased for both alkene/alkane selectivity (1.18) and for cis/trans selectivity (0.05) compared with the catalyst-free reaction. The novel Fe₄Mo₂ exhibited nearly identical conversion and selectivity as Fe₆, despite the differences in structure. The addition of either cluster as a catalytic unit to these reaction conditions increased conversion while decreasing selectivity, in contrast to previous work with the analogous Fe₅Mo cluster, which resulted in increased conversion and selectivity for the alkene and cis products.

Scheme 3 Anilinium triflate salts used as proton sources, including the pK_a values for the relevant acidic protons.

Notably, comparison of the same reductions with a more sterically encumbered proton source (2,4,6-tri-tert-butylanilinium triflate, [^tBu₃AnH]OTf) resulted in a different trend: addition of catalytic Fe₆ resulted in a 5-fold decrease in conversion (15.9% to 2.7%) and a moderate decrease in alkene/alkane selectivity (12.6 to 7.6) yet provided greater selectivity of the cis versus trans product (0.01 to 0.22). Fe₄Mo₂ provided a lower conversion (1.5%) than either the iron-only or catalyst-free reactions with a marginally lower selectivity (9.9) for alkene/alkane compared to the catalyst-free reaction (12.6), but with the highest selectivity for the cis product observed (0.96 cis/trans) overall in this report. This trend towards lower conversion but higher selectivity for the cis-alkene product is the inverse of the trend observed with the less bulky [Me₃AnH]OTf; however, it does parallel the trend found previously for Fe₅Mo, which also exhibited increased selectivity despite increased overall conversion. Comparison of [Me₃AnH]OTf and [^tBu₃AnH]OTf reactions demonstrated the same trend found previously, with the more sterically encumbered proton source resulting in decreased conversion but increased selectivity for the cis-alkene product.

Temperature dependence was also investigated, initially with [^tBu₃AnH]OTf. At both decreased (−20 °C) and elevated (60 °C) temperature, the selectivity for the cis alkene product decreased while selectivity for the alkene product over the alkane product remained similar. At lower temperatures conversion also decreased, however conversion remained unchanged at elevated temperature—in both cases contradictory to our previous work with Fe₅Mo. Interestingly, a different trend emerged with [Me₃AnH]OTf, which exhibited greatly increased selectivity for the cis product in the case of both lowered and elevated temperature—despite decreased conversion and selectivity for the alkene product (in both cases).

We also investigated the effect of pK_a by using 2,6-dimethylanilinium triflate ([Me₂AnH]OTf), a proton source with identical steric bulk near the acidic site, but with a lower pK_a due to the absence of the para methyl group. The more acidic proton source with Fe₆ provided a remarkable 60.2% conversion—the highest conversion observed in our work with these iron clusters—but providing very poor cis and alkene selectivity. Fe₄Mo₂ resulted in decreased conversion, but notably a three-fold increase in alkene selectivity. The seemingly outsized effect of such a small increase in acidity indicates the presence of a barrier to cluster protonation—which once overcome, greatly increases reaction rate. This suggests that the iron-only cluster Fe₆ provides a more accessible protonated state than the Fe₄Mo₂ cluster.

Mechanistic insight from electrochemistry, X-ray photoelectron spectroscopy & ¹H NMR

In our previous report we provided evidence that the catalytic method proceeds first via two-electron reduction of the cluster to a tetra-anionic cluster followed by subsequent protonation to form a dihydride di-anionic cluster. To ensure this cluster remained intact, IR spectra were collected before and after reduction, showing a shift of an intense feature from 1928 cm⁻¹ to 1894 cm⁻¹, consistent with a two electron reduction (Fig. S5). To further implicate a protonated H₂Fe₄Mo₂²⁻ cluster was formed, the IR spectrum was monitored and a blue-shift from 1894 cm⁻¹ to 2018 cm⁻¹ was observed, consistent with a two-electron oxidation back to a di-anionic species. Evidence of the protonated cluster was also observed in the ¹H NMR spectrum after protonation (Fig. 2), exhibiting resonances at δ = −22 and −26 ppm that correspond to bridging hydride and dihydride species respectively—as observed in the analogous (previously reported by Zacchini) reaction of Fe₆ (δ = −21 ppm) or Fe₅Mo (δ = −26 ppm), as well as the recent work with Fe₆⁴⁻ (δ = −21, −27 ppm for mono- and di-hydride species respectively).^23,29

Fig. 2 ¹H NMR spectrum (400 MHz, CD₃CN) in the hydride region after reduction and protonation of 2, indicating the presence of hydride species as the putative catalytic intermediate in the hydrogenation of DPA.

In our previous report, we provided evidence that the catalytic mechanism proceeds first via two-electron cluster reduction (eqn (1)) followed by protonation to generate a hydride-supported di-anionic cluster (eqn (2)):

[M₆]²⁻ + 2e⁻ → [M₆]⁴⁻ (1)

[M₆]⁴⁻ + 2H⁺ → [H₂M₆]²⁻ (2)

[H₂M₆]²⁻ + 2e⁻ → [H₂M₆]⁴⁻ (3)

The hydride supported cluster can, itself, directly act upon DPA, or undergo further two-electron reduction to then react with substrate (eqn (3)). Binding of the DPA is theorized to occur analogously to a previous report utilizing iron-carbonyl species with a single iron site binding the DPA π system followed by hydrogenation utilizing cluster bound hydrides.³⁰ As a control to investigate DPA binding the reduced 4-clusters ¹³C NMR was collected after reacting Fe₆ with KC₈ and upon addition of DPA the only carbide stretches observed corresponded to the Fe₆²⁻ and Fe₆⁴⁻ species (δ = 483.71 ppm and 488.89 ppm respectively), consistent with previous work and precluding a DPA bound cluster forming.³¹

Our previous report³² demonstrated that the Fe₅Mo cluster is an ‘electronic hybrid’ between Fe₆²⁻ and Fe₅²⁻ clusters, nominally serving as a slightly reduced Fe₅²⁻ cluster capped with a slightly oxidized Mo rather than as an Fe₅Mo²⁻ cluster with even charge distribution. As such, the iron sites in Fe₅Mo²⁻ exhibit an average oxidation state of approximately Fe^+0.4, with a higher average oxidation state than in the case of the iron-only Fe₆²⁻ (Fe^+0.33). We thus hypothesize there is a similar, but even more pronounced effect with Fe₄Mo₂: capping the Fe₄²⁻ fragment with two neutral Mo(0) centers promotes access to the most oxidized iron sites in the Fe_xMo_y series, namely (or nominally, at least) as Fe^+0.5. To test this hypothesis, X-ray photoelectron spectra (XPS) were collected for Fe₆, Fe₅Mo and Fe₄Mo₂ and analyzed in the Fe 2p region (j=3/2 and 1/2) to determine the extent of oxidation at the iron sites (Fig. 3a).³³

Fig. 3 High-resolution X-ray photoelectron spectrum (XPS) of (a) the iron 2p region and (b) the molybdenum 3d region: pink, Fe₆; green, Fe₅Mo; blue, Fe₄Mo₂.

Consistent with our hypothesis, the all-iron Fe₆ cluster demonstrated the lowest average iron oxidation state with an binding energy of 708.05 eV.³⁴ Upon substitution of iron for an additional molybdenum, an increase in binding energy is observed with Fe₅Mo (708.31 eV) → Fe₄Mo₂ (708.39 eV) indicating a slightly higher average iron oxidation state in each case. Unexpectedly, the difference between Fe₆ and Fe₅Mo is 0.26 eV (attributable to a difference in oxidation state of ∼0.07), whereas the nominally more oxidized iron sites in Fe₄Mo₂ differed from Fe₅Mo by only 0.08 eV—despite an expected oxidation state difference of ∼0.1.

The corresponding Mo 3d XPS spectra provide insight regarding charge distribution across the clusters (Fig. 3b). Previous work showed that the Mo site in Fe₅Mo is slightly oxidized, and this work further supports that conclusion with an Fe₅Mo Mo 3d binding energy of 228.18 eV compared to literature Mo(0) of 228.0 eV.^32,35 Interestingly, the Fe₄Mo₂ cluster exhibits further oxidized Mo sites (228.29 eV) and provides insight regarding the small shift in Fe 2p features in the Fe₄Mo₂ XPS spectrum (Fig. 3a): namely, that the iron sites do not exhibit proportionally higher binding energy features (i.e. average oxidation state does not increase as much as expected) as the two Mo(0) sites experience some loss of electron density in Fe₄Mo₂, compared with an invariant Mo(0)-‘like’ oxidation state in the Fe₅Mo cluster. Despite this, Fe₄Mo₂ still exhibits the highest energy Fe 2p binding energies (proxy for most oxidized iron sites), and it would thus be expected for the cluster to exhibit reactivity similar to the ‘free’ (uncapped) Fe₄ unit.

This is notable as the protonated Fe₄²⁻ was crystallographically characterized³⁶; however, protonation results in one Fe–H–Fe bridging hydride and one carbide-based C–H ‘protonation’ in HFe₄(η²-CH)(CO)₁₂; this cluster is readily deprotonated and has limited stability in solution. However, further and/or excess acid treatment results in carbide protonation (eventually released as CH₄), thus precluding the ‘bare’ Fe₄ cluster's efficacy as a reduction catalyst.³⁷ In contrast, the Mo₂-capped Fe₄ unit in 2 in this work serves as an ‘inorganic protecting group’, preventing carbide protonation and driving the formation of an exclusively metal-based (no carbide protonation) dihydride intermediate, overall producing a catalytically competent species.

Overall, the XPS data indicate that Mo substituted clusters with fewer iron sites have higher average oxidation states, thus resulting in more tightly bound hydrides (compared with Fe₆ or Fe₅Mo) and rendering reactivity more selective. The catalysis results above (Table 2) also indicate that the barrier to protonation for Fe₄Mo₂ is important, with less sterically hindered proton sources evidently overcoming the barrier and ‘over-protonating’ the cluster, thus resulting in more reactive but less selective hydride intermediates. Alternatively, more sterically hindered proton sources evidently ‘under-protonate’ the cluster leading to less reactive but more selective substrate conversions.

Cyclic voltammetry data was also collected was collected for Fe₆, Fe₅Mo and Fe₄Mo₂ to provide further insight into the reduction and oxidation potentials of the clusters—and possible accessibility of the two-electron reduced cluster intermediates (Fig. 4 and SI Fig. S10–20). Consistent with previous work by Zacchini et al., the Fe₆ cluster showed multiple and poorly resolved redox features corresponding to the formation of Fe₆⁴⁻ and subsequent re-oxidation back to Fe₆²⁻; similar features were observed for Fe₅Mo, albeit at slightly higher potentials. However in the case of Fe₄Mo₂, the highest oxidation potential in the series is obtained—consistent with the oxidation of more electropositive iron centers in Fe₄Mo₂ relative to Fe₆ and Fe₅Mo. This observation further supports assignment of higher average oxidation state iron sites in the Fe₄Mo₂ cluster.

Fig. 4 Comparison of cyclic voltammograms (vs. Fc/Fc⁺) of 5 mM of cluster in MeCN containing 0.2 M TBAPF₆. Experiment conditions: WE, glassy carbon; CE, Pt wire; RE, Ag wire quasi-reference; scan rate: 10 mV s⁻¹.

Conclusion

In summary, we report a synthetic route to the structurally characterized iron-dimolybdenum carbide cluster [K(benzo-18-crown-6)]₂[(μ₆-C)Fe₄Mo₂(μ₂-CO)₂(CO)₁₆] (2) by addition of Mo(0) centers to a reduced four-iron carbide cluster (1), wherein the Fe₄ skeleton retains its ‘butterfly’ motif and enforces cis-Mo₂ coordination about the carbide. The carbide is structurally displaced relative to the starting Fe₄ framework and is contracted towards the Mo sites. Reduction/protonation of 2 affords a cluster-di/hydride species that provides enhanced selectivity for regioselective partial reduction of DPA to cis-diphenylethylene product due to increased stability of higher valent iron-hydride species, thus decreasing conversion but enhancing selectivity. Future work will investigate substitution of the strong field carbonyl ligand environment for Fe_mMo_n clusters with for more biologically relevant sulfur-based supports such as thiolates and sulfides (or phosphine/phosphide) through our recently developed in situ, outer-sphere oxidation method that renders clusters substitutionally active due to their non-PSEPT (Wade–Mingos rules) electron count.^14,21 We expect that the capping effect of the molybdenum sites will provide an increased extent of sulfide/thiolate (or phosphine/phosphide) substitution compared with iron-only clusters by promoting external L- and X-type ligand binding to the more electropositive (higher average oxidation state) Fe₄ iron centers.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: additional NMR, IR, XPS, cyclic voltammetry and X-ray structures. See DOI: https://doi.org/10.1039/d6dt00524a.

CCDC 2440297 (2) and 2440298 (1) contain the supplementary crystallographic data for this paper.^38a,b

Acknowledgements

We acknowledge the National Science Foundation (CHE-2109175) and the Robert A. Welch Foundation (F-1822) for support of this work. We acknowledge use of a Bruker AVIII HD 500 with Prodigy liquid N₂ Cryoprobe supported by NIH Grant S10 OD021508. EJD acknowledges a University of Texas at Austin, College of Natural Sciences Dean's Strategic Fellowship. We also thank Dr Ian Riddington for his assistance with the collection of GC-MS data for product characterization and Dr Hugo Celio and Jeremy Brinker for their assistance with the collection of XPS data.

References

1. M. A. Beno, J. M. Williams, M. Tachikawa and E. L. Muetterties, Fischer-Tropsch Chemistry: Structure of a Seminal η²-CH Cluster Derivative, HFe₄(η²-CH)(CO)₁₂, J. Am. Chem. Soc., 1980, 102(13), 4542–4544,  DOI:10.1021/ja00533a052.
2. M. A. Beno, J. M. Williams, M. Tachikawa and E. L. Muetterties, A Closed Three-Center Carbon-Hydrogen-Metal Interaction. A Neutron Diffraction Study of HFe₄(η²-CH)(CO)₁₂, J. Am. Chem. Soc., 1981, 103(6), 1485–1492,  DOI:10.1021/ja00396a032.
3. J. H. Davis, M. A. Beno, J. M. Williams, J. Zimmie, M. Tachikawa and E. L. Muetterties, Structure and Chemistry of a Metal Cluster with a Four-Coordinate Carbide Carbon Atom, Proc. Natl. Acad. Sci. U. S. A., 1981, 78(2), 668–671.
4. S. Ghosh, K. B. Holt, S. E. Kabir, M. G. Richmond and G. Hogarth, Electrocatalytic Proton Reduction Catalysed by the Low-Valent Tetrairon-Oxo Cluster [Fe₄(CO)₁₀(κ²-dppn)( μ₄-O)]²⁻ [dppn = 1,1′-bis(diphenylphosphino)naphthalene], Dalton Trans., 2015, 44(11), 5160–5169,  10.1039/C4DT03323J.
5. I. Čorić and P. L. Holland, Insight into the Iron–Molybdenum Cofactor of Nitrogenase from Synthetic Iron Complexes with Sulfur, Carbon, and Hydride Ligands, J. Am. Chem. Soc., 2016, 138(23), 7200–7211,  DOI:10.1021/jacs.6b00747.
6. E. M. Holt, K. H. Whitmire and D. F. Shriver, The Role of Metal Cluster Interactions in the Proton-Induced Reduction of CO. the Crystal Structures of [PPN]{HFe₄(CO)₁₂} and HFe₄(CO)₁₂(η-COCH₃), J. Organomet. Chem., 1981, 213(1), 125–137,  DOI:10.1016/S0022-328X(00)93954-8.
7. A. Taheri and L. A. Berben, Tailoring Electrocatalysts for Selective CO₂ or H⁺ Reduction: Iron Carbonyl Clusters as a Case Study, Inorg. Chem., 2016, 55(2), 378–385,  DOI:10.1021/acs.inorgchem.5b02293.
8. K. M. Lancaster, M. Roemelt, P. Ettenhuber, Y. Hu, M. W. Ribbe, F. Neese, U. Bergmann and S. DeBeer, X-Ray Emission Spectroscopy Evidences a Central Carbon in the Nitrogenase Iron-Molybdenum Cofactor, Science, 2011, 334(6058), 974–977,  DOI:10.1126/science.1206445.
9. M. U. Delgado-Jaime, B. R. Dible, K. P. Chiang, W. W. Brennessel, U. Bergmann, P. L. Holland and S. DeBeer, Identification of a Single Light Atom within a Multinuclear Metal Cluster Using Valence-to-Core X-Ray Emission Spectroscopy, Inorg. Chem., 2011, 50(21), 10709–10717,  DOI:10.1021/ic201173j.
10. C. Joseph, J. P. Shupp, C. R. Cobb and M. J. Rose, Construction of Synthetic Models for Nitrogenase-Relevant NifB Biogenesis Intermediates and Iron-Carbide-Sulfide Clusters, Catalysts, 2020, 10(11), 1317,  DOI:10.3390/catal10111317.
11. J. S. Bradley, G. B. Ansell and E. W. Hill, Homogeneous Carbon Monoxide Hydrogenation on Multiple Sites: A Dissociative Pathway to Oxygenates, J. Am. Chem. Soc., 1979, 101(24), 7417–7419,  DOI:10.1021/ja00518a055.
12. S. Deabate, P. J. King and E. Sappa, Activation of Carbon Monoxide, Water, and Alcohols on Metal Carbonyl Clusters. Homogeneous and Surface–Mediated Reactions, in Metal Clusters in Chemistry, ed. P. Braunstein, L. A. Oro and P. R. Raithby, Wiley, 1999, pp 796–843.  DOI:10.1002/9783527618316.ch2h.
13. J. W. Kolis, F. Basolo and D. F. Shriver, Reactivity of Metal Carbide Clusters: Alkylation and Protonation of [Fe5(CO)14C]2-, J. Am. Chem. Soc., 1982, 104(21), 5626–5630,  DOI:10.1021/ja00385a011.
14. C. R. Cobb, R. K. Ngo, E. J. Dick, V. M. Lynch and M. J. Rose, Multi-Phosphine-Chelated Iron-Carbide Clusters via Redox-Promoted Ligand Exchange on an Inert Hexa-Iron-Carbide Carbonyl Cluster, [Fe₆(μ₆-C)(μ₂-CO)₄(CO)₁₂]²⁻, Chem. Sci., 2024, 15(29), 11455–11471,  10.1039/D4SC01370K.
15. S. Saha, L. Zhu and B. Captain, Synthesis and Structural Characterization of Bimetallic Iron−Nickel Carbido Cluster Complexes, Inorg. Chem., 2010, 49(7), 3465–3472,  DOI:10.1021/ic100057x.
16. C. Femoni, R. D. Pergola, M. C. Iapalucci, F. Kaswalder, M. Riccò and S. Zacchini, Copolymerization of Fe₄Cu₂C(CO)₁₂ Moieties with Bidentate N-Ligands: Synthesis and Crystal Structure of the [Fe₄Cu₂(μ₆-C)(CO)₁₂(μ-Bipy)]₄·8THF Square Tetramer and the Infinite [Fe₄Cu₂(μ₆-C)(CO)₁₂(μ-L–L)]_∞ Zigzag Chains, Dalton Trans., 2009, )(No. 9), 1509–1511,  10.1039/B821639H.
17. R. Della Pergola, A. Sironi, L. Garlaschelli, D. Strumolo, C. Manassero, M. Manassero, S. Fedi, P. Zanello, F. Kaswalder and S. Zacchini, Fe–Cu Octahedral Carbide Clusters, and the Replacement of Their Labile Halide Ligands: Synthesis, Solid State Structure, Substitution and Electrochemical Reactivity of [Fe₅C(CO)₁₄(CuBr)]²⁻, [Fe₄C(CO)₁₂(CuCl)₂]²⁻, [{Fe₄Cu₂C(CO)₁₂(μ-Cl)}₂]²⁻, [Fe₅C(CO)₁₄(CuOC₄H₈)]⁻ and [Fe₄C(CO)₁₂(CuNCMe)₂], Inorg. Chim. Acta, 2010, 363(3), 586–594,  DOI:10.1016/j.ica.2009.01.025.
18. L. A. Polyakova, S. P. Gubin, A. Churakov and L. G. Kuz’mina, [Fe₃(μ₃-O)(CO)₉]^2-, Dianion as a Suitable Material for the Preparation of Homo- and Heterometallic Carbonyl Clusters, Russ. J. Coord. Chem., 1999, 25, 711–720.
19. B. F. G. Johnson, D. A. Kaner, J. Lewis, P. R. Raithby and M. J. Rosales, Syntheses and Structural Characterisations of Some Novel Mixed-Metal Iron-Gold Carbido Clusters; X-Ray Crystal Structures of Fe4AuC(η-H)(CO)12(PPh3) and Fe4Au2C(CO)12(PEt3)2, J. Organomet. Chem., 1982, 231(3), C59–C64,  DOI:10.1016/S0022-328X(00)92900-0.
20. C. Joseph, C. R. Cobb and M. J. Rose, Single–Step Sulfur Insertions into Iron Carbide Carbonyl Clusters: Unlocking the Synthetic Door to FeMoco Analogues, Angew. Chem., 2021, 133(7), 3475–3479,  DOI:10.1002/ange.202011517.
21. C. R. Cobb, R. K. Ngo, S. Vasylevskyi, Y. Hur, E. J. Dick, J. A. Balderas and M. J. Rose, Single-Step Access to Thiolate-Supported, Six-Iron Clusters with an Authentic Carbide: Synthesis of [Fe₆(μ₆-C)(μ₂-Stol)(μ₂-CO)₂(CO)₁₂]^3- via Non-PSEPT, Redox Intermediates, Inorg. Chem., 2025, 64(33), 16734–16746 , https://doi.org/10.1021/acs.inorgchem.4c05362.
22. M. Tachikawa, A. C. Sievert, E. L. Muetterties, M. R. Thompson, C. S. Day and V. W. Day, Metal Clusters., 24. Synthesis and Structure of Heteronuclear Metal Carbide Clusters, J. Am. Chem. Soc., 1980, 102(5), 1725–1727,  DOI:10.1021/ja00525a047.
23. C. Joseph, S. Kuppuswamy, V. M. Lynch and M. J. Rose, Fe 5 Mo Cluster with Iron-Carbide and Molybdenum-Carbide Bonding Motifs: Structure and Selective Alkyne Reductions, Inorg. Chem., 2018, 57(1), 20–23,  DOI:10.1021/acs.inorgchem.7b02615.
24. M. Tachikawa, R. L. Geerts and E. L. Muetterties, Metal Carbide Clusters Synthesis Systematics for Heteronuclear Species, J. Organomet. Chem., 1981, 213(1), 11–24,  DOI:10.1016/S0022-328X(00)93947-0.
25. J. S. Bradley and E. W. Randall, Carbidocarbonyl Clusters of Iron [and Discussion], Philos. Trans. R. Soc., A, 1982, 308(1501), 103–113.
26. S. Kuppuswamy, J. D. Wofford, C. Joseph, Z. Xie, A. K. Ali, V. M. Lynch, P. A. Lindahl and M. J. Rose, Structures, Interconversions, and Spectroscopy of Iron Carbonyl Clusters with an Interstitial Carbide: Localized Metal Center Reduction by Overall Cluster Oxidation, Inorg. Chem., 2017, 56(10), 5998–6012,  DOI:10.1021/acs.inorgchem.7b00741.
27. J. M. Rivera-Ortiz and R. H. Burris, Interactions among Substrates and Inhibitors of Nitrogenase, J. Bacteriol., 1975, 123(2), 537–545.
28. A. C. Filippou and T. Rosenauer, A Reaction Pathway of [Fe(CO)₅] with Alkynes via Ferrabicyclobutenones, Angew. Chem., Int. Ed., 2002, 41(13), 2393–2396,  DOI:10.1002/1521-3773(20020703)41:13%253C2393::AID-ANIE2393%253E3.0.CO;2-3.
29. M. Bortoluzzi, I. Ciabatti, C. Cesari, C. Femoni, M. C. Iapalucci and S. Zacchini, Synthesis of the Highly Reduced [Fe₆C(CO)₁₅]⁴⁻ Carbonyl Carbide Cluster and Its Reactions with H⁺ and [Au(PPh₃)]⁺, Eur. J. Inorg. Chem., 2017, 2017(25), 3135–3143,  DOI:10.1002/ejic.201700169.
30. M. Haberberger and S. Enthaler, Straightforward Iron-Catalyzed Synthesis of Vinylboronates by the Hydroboration of Alkynes, Chem. – Asian J., 2013, 8(1), 50–54,  DOI:10.1002/asia.201200931.
31. L. Liu, T. B. Rauchfuss and T. J. Woods, Iron Carbide-Sulfide Carbonyl Clusters, Inorg. Chem., 2019, 58(13), 8271–8274.
32. J. McGale, G. E. Cutsail, C. Joseph, M. J. Rose and S. DeBeer, Spectroscopic X-Ray and Mössbauer Characterization of M₆ and M₅ Iron(Molybdenum)-Carbonyl Carbide Clusters: High Carbide-Iron Covalency Enhances Local Iron Site Electron Density Despite Cluster Oxidation, Inorg. Chem., 2019, 58(19), 12918–12932,  DOI:10.1021/acs.inorgchem.9b01870.
33. G. Beamson, B. T. Pickup, W. Li and S.-M. Mai, XPS Studies of Chain Conformation in PEG, PTrMO, and PTMG Linear Polyethers, J. Phys. Chem. B, 2000, 104(12), 2656–2672,  DOI:10.1021/jp9905629.
34. B. A. Sosinsky, N. Norem and J. Shelly, Spectroscopic Study of a Series of Iron Carbido Clusters, Inorg. Chem., 1982, 21(1), 348–356,  DOI:10.1021/ic00131a063.
35. C. L. Bianchi, M. G. Cattania and P. Villa, XPS Characterization of Ni and Mo Oxides before and after “in Situ” Treatments, Appl. Surf. Sci., 1993, 70–71, 211–216,  DOI:10.1016/0169-4332(93)90429-F.
36. M. Tachikawa and E. L. Muetterties, Metal Clusters. 25. A Uniquely Bonded C-H Group and Reactivity of a Low-Coordinate Carbidic Carbon Atom, J. Am. Chem. Soc., 1980, 102(13), 4541–4542 , https://doi.org/10.1021/ja00533a051.
37. S. Deabate, P. J. King and E. Sappa, Activation of Carbon Monoxide, Water, and Alcohols on Metal Carbonyl Clusters. Homogeneous and Surface–Mediated Reactions, in Metal Clusters in Chemistry, ed. P. Braunstein, L. A. Oro and P. R. Raithby, Wiley, 1999, pp. 796–843.  DOI:10.1002/9783527618316.ch2h.
38. (a) CCDC 2440297: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2mxb85; (b) CCDC 2440298: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2mxb96.

---

Footnote

† Equal author contribution.

---