Chalcogenidobis(ene–1,2–dithiolate)molybdenum(IV) complexes (chalcogenide E = S, Se): Probing Mo≡E and ene–1,2–dithiolate substituent effects on geometric and electronic structure

Abstract

New square–pyramidal bis(ene–1,2–dithiolate)MoSe complexes, [Mo^IVSe(L)₂]²⁻, have been synthesised along with their terminal sulfido analogues, [Mo^IVS(L)₂]²⁻, using alkyl (L^C4H8), phenyl (L^Ph) and methyl carboxylate (L^COOMe) substituted dithiolene ligands (L^R). These complexes now complete three sets of Mo^IVO, Mo^IVS and Mo^IVSe species that are coordinated with identical ene–1,2–dithiolate ligands. The alkyl substituted [Mo(S/Se)(L^C4H8)₂]^{2 −} complexes were reported in prior investigations (H. Sugimoto, T. Sakurai, H. Miyake, K. Tanaka and H. Tsukube, Inorg. Chem. 2005, 44, 6927, H. Tano, R. Tajima, H. Miyake, S. Itoh and H. Sugimoto, Inorg. Chem. 2008, 47, 7465). The new series of complexes enable a systematic investigation of terminal chalcogenido and supporting ene–1,2–dithiolate ligand effects on geometric structure, electronic structure, and spectroscopic properties. X–ray crystallographic analysis of these (Et₄N)₂[Mo(E)L₂] (E = terminal chalcogenide) complexes reveals an isostructural Mo centre that adopts a distorted square pyramidal geometry. The M≡E bond distances observed in the crystal structures and the ν(M≡E) vibrational frequencies indicate that these bonds are weakened with an increase in L→Mo electron donation (L^COOMe < L^Ph < L^C4H8), and this order is confirmed by an electrochemical study of the complexes. The ⁷⁷Se NMR resonances in MoSeL complexes appear at lower magnetic fields as the selenido ion became less basic from MoSeL^C4H8, MoSeL^Ph and MoSeL^COOMe. Electronic absorption and resonance Raman spectroscopies have been used to assign key ligand-field, MLCT, LMCT and intraligand CT bands in complexes that possess the L^COOMe ligand. The presence of low-energy intraligand CT transition in these MoEL^COOMe compounds directly probes the electron withdrawing nature of the −COOMe substituents, and this underscores the complex electronic structure of square pyramidal bis(ene–1,2–dithiolate)-Mo^IV complexes that possess extended dithiolene conjugation.

Introduction

Compounds with metal–ligand multiple bonds, M=E or Mo≡E (M = metal and E = element), function as key components of industrial and biological catalysts.¹ The recent structural characterisation of some mononuclear pyranopterin molybdenum enzymes has resulted in increased interest in the study of multiply bonded MoO, MoS, and MoSe units in the coordination chemistry of molybdenum.² A terminal oxo donor, possessing a formal triple bond, is found in many forms of pyranopterin molybdenum enzymes, which include the molybdenum hydroxylases of the xanthine oxidase enzyme family and the molybdenum oxotransferases that belong to the sulfite oxidase and DMSO reductase enzyme families.² In contrast to these terminal Mo-oxo bonds, a terminal sulfido donor with a reduced Mo-S_sulfido bond order is found oriented cis to the apical oxo ligand in the xanthine oxidase family.³ Also a terminal sulfido donor is included in the reduced formate dehydrogenase of molybdenum enzymes.⁴ Recently, Dobbek and coworkers showed that a terminal selenide donor is found in place of the terminal sulfido present in xanthine oxidase and nicotinic acid hydroxylase.⁵ Although a large number of monomeric molybdenum coordination compounds possess terminal oxo ligands, notably fewer that possess terminal sulfido ligands have been prepared and their reactivities and spectroscopic properties explored.^6,7 Even fewer studies have examined the geometric and electronic structure of molybdenum complexes that possess terminal selenido coordination due to the difficulty in synthesising MoSe complexes.^7,8,9 Additionally, there is a paucity of comparative studies on MoO, MoS and MoSe compounds that have been synthesised using the same or similar supporting ligands. The dearth of such compounds has limited our ability to make detailed geometric-electronic structure correlations as a function of the MoE unit, and this has provided the impetus for the synthesis of new monomeric compounds that possess terminal sulfido and selenido coordination.

Over a decade ago, Cotton and coworkers synthesised a series of six–coordinate distorted octahedral trans–Mo^IVE₂(P-P)₂ complexes (E = O, S, Se, and P-P = 1,2-bis(diphenylphosphino)ethylene or its ethane derivative (Fig. 1)).¹⁰ Cummins and coworkers also reported a series of four coordinate distorted tetrahedral Mo^VE complexes with an identical coligand in Mo^VE(N(R)Ar)₃ (E = O, S, Se; R = C(CD₃)₂CH₃, Ar = 3,5-C₆H₃Me₂) (Fig. 1).¹¹ However, no series of five–coordinate MoO, MoS, and MoSe complexes exist that possess identical supporting ligands. Furthermore, the P-P and N(R)Ar supporting ligands in the earlier studies did not possess electron withdrawing groups (EWGs) or electron donating groups (EDGs) that would allow for a systematic study of substituent effects on the electronic structure, spectroscopy and reactivities of the complexes. A classic series of Mo^IV/V/VI S complexes, [{HB(R₂pz)₃}MoS(L)_m]ⁿ⁻ (hydrotris(3,5-dialkylpyrazol-1-yl)borate, L = mono- or di-dentate ligands), has been prepared by Young and Enemark using sterically demanding tridentate ligands as models for Mo-sulfido centres in pyranopterin molybdenum enzymes.^12–17 This series of compounds has allowed for very meaningful systematic investigations of their geometric structures, spectroscopic properties, and reactivities by modifying both the alkyl groups, R, and the mono- and bidentate ligands, L. However, no corresponding MoSe derivatives have been reported to date.

Fig. 1.

Designation and abbreviation of MoE complex structures reported previously (E = O, S, Se). ^{ref. 10,11}

Recently, we reported the first series of multiply bonded square pyramidal MoE (E = O, S, Se) complexes possessing an identical supporting ligand, [Mo^IVE(L^C4H8)₂]²⁻ (L^C4H8 = cyclohexene-1,2-dithiolate).^18–20 These complexes make possible detailed investigations regarding the effects of the MoE unit on the structure, spectroscopy and reactivity of the complexes. Regarding the square pyramidal structure, the active sites of reduced (i.e. Mo(IV)) formate dehydrogenase and arsenite oxidase adopt similar square pyramidal bis(pyranopterin-dithiolene)-Mo^IVS and -Mo^IVO structures, respectively,^4,21 This has stimulated our interest in extending the [Mo^IV Se(ene-1,2-dithiolate)₂]²⁻ family to include the ene-1,2-dithiolates, 1,2-diphenyl-1,2-dithiolate (L^Ph) and 1,2-dicarbomethoxyethylene-1,2-dithiolate (L^COOMe), allowing us to probe MoE bonding as a function of dithiolene electronic structure.

In this manuscript, we describe the synthesis and X-ray structures of (Et₄N)₂[Mo^IVS(L^Ph)₂] (MoSL^Ph), (Et₄N)₂[Mo^IVS(L^COOMe)₂] (MoSL^COOMe), as well as their selenido derivatives, (Et₄N)₂[Mo^IVSe(L^Ph)₂] (MoSeL^Ph) and (Et₄N)₂[Mo^IVSe(L^COOMe)₂] (MoSeL^COOMe) (see Fig. 2). These new molecules now complete set of [MoEL₂]²⁻ (E = O, S, Se; L = L^C4H8, L^Ph, L^COOMe) compounds that allow the development of detailed electronic and geometric structure correlations as a function of the terminal E ligand, and this provides a means for evaluating the effects of dithiolene substituents on the structure and spectroscopic properties of these compounds. A long these lines, we have performed a detailed electronic absorption, resonance Raman and computational investigation of the MoEL^COOMe (E = O, S, Se) series in order to understand the nature of their low energy ligand-field and charge transfer excitations and to develop a basic electronic structure description of the complexes as a function of the apical E donor.

Fig. 2.

Designation and abbreviation of complex structures.

Experimental section

General

All reagents and solvents were used as received unless otherwise noted. All reactions were carried out under a dinitrogen atmosphere using standard Schlenk techniques. Complexes of MoOL^C4H8,¹⁸ MoSL^C4H8,¹⁹ MoSeL^C4H8,²⁰ MoOL^Ph,²² MoOL^COOMe,²³ [Mo(CO)₂(L^Ph)]²⁴ and (Et₄N)[Mo^IV(OSitBuPh₂)(L^COOMe)₂] (Mo(OSitBuPh₂)L^COOMe)²⁵ were prepared by following established literature procedures.

Synthesis and characterization of complexes

(Et₄N)₂[Mo^IVS(L^Ph)₂] (MoSL^Ph)

To an acetonitrile suspension containing [Mo(CO)₂(L^Ph)₂] (46 mg, 0.072 mmol), an acetonitrile suspension containing Na₂S (52 mg, 0.67 mmol) was added. After stirring for 3 hours, Et₄NCl (24 mg, 0.15 mmol) was added to the suspension. Excess Na₂S was removed by filtration and the filtrate was concentrated to dryness. The obtained green solid was dissolved in minimum volume of DMSO and diffusion of ether to the solution gave green-brown crystals of MoSL^Ph. The crystals were collected by filtration and dried under reduced pressure. Yield 44 mg (70%). Anal. Calcd for C₄₄H₆₀MoN₂S₅ • DMSO (mol wt. 951.38): C, 58.07; H, 6.99; N, 2.94. Found: C, 57.82; H, 6.65; N; 2.75. ¹H NMR (CD₃CN): δ 7.24-6.92 (m, 20H), 3.09 (q, 16H), 2.53 (s, 6H), 1.08 (t, 24H). UV-vis spectrum (CH₃CN): λ_max = 258 (ε = 37000), 321 (29600), 385 (13000), 450 (sh, 4660), 569 (1370), 699 nm (1240 M⁻¹ cm⁻¹). ESI-MS (CH₃CN): m/z = 357 [M]²⁻. CV (0.1 V s⁻¹, CH₃CN): E_1/2 (rev.) = −0.43 V vs. SCE. IR (KBr): ν 486 (s), 691 (s), 723 (s), 759 (m), 836 (s), 870 (s), 997 (m), 1026 (s), 1077 (w), 1108 (vs), 1236 (vs), 1438 (vs), 1484 (s), 1586 (w), 1707 (s), 1716 cm⁻¹ (s).

(Et₄N)₂[Mo^IVSe(L^Ph)₂] (MoSeL^Ph)

This complex was synthesised using the same procedure as MoSL^Ph, except Na₂Se (27 mg, 0.22 mmol) was used instead of Na₂S. Recrystallization from DMSO/ether gave a microcrystalline powder, which was collected by filtration and dried in dinitrogen. Yield: 24 mg (34%). Anal. Calcd for C₄₄H₆₀MoN₂S₄Se • DMSO (mol wt. 998.25): C, 55.35; H, 6.66; N, 2.81. Found: C, 55.38; H, 6.65; N; 2.96. ¹H NMR (CD₃CN): δ 7.24-6.91 (m, 20H) 3.09 (q, 16H), 2.53 (s, 6H), 1.08 (t, 24H). UV-vis spectrum (CH₃CN): λ_max = 294 (ε = 33200), 328 (32300), 370 (sh, 14900), 414 (12000), 480 (sh, 3850), 588 (sh, 830), 705 nm (620 M¹⁻cm⁻¹). ESI–MS (CH₃CN): m/z = 792 {[M]²⁻ + Et₄N⁺}⁻. CV (0.1 V s⁻¹, CH₃CN): E_1/2(rev.) = −0.45 V vs. SCE. IR (KBr): ν 784 (m), 807 (m), 909 (s), 999 (m), 1017 (m), 1055 (m), 1152 (m), 1172 (m), 1234 (vs), 1325 (m), 1394 (s), 1431 (m), 1454 (m), 1483 (s), 1528 (s), 1696 (s), 1711 cm⁻¹ (s).

(Et₄N)₂[Mo^IVS(L^COOMe)₂] (MoSL^COOMe)

To 10 mL of an acetonitrile solution of Mo(OSitBuPh₂)L^COOMe (124 mg, 0.14 mmol), 1.5 mL of an acetonitrile solution of Et₄NSH (25 mg, 0.15 mmol) was added. The resultant brown solution was concentrated to ~1 mL. A brown powder was precipitated by addition of diethyl ether to the solution, and this was collected by filtration, washed with THF/acetone, and dried in vacuo. To 3 mL of an acetonitrile solution of the brown powder, ^tBuPh₂SiCl (14 mL, 0.065 mmol) was added, and the solution was stirred for 5 minutes and then concentrated to 1 mL. Addition of 6 mL of THF to the solution yielded a brown powder, which was collected by filtration and washed with THF/acetone. Yield: 50.5 mg (45%). Anal. Calcd for C₂₈H₅₂MoN₂O₈S₅ (mol wt. 801.00): C, 41.99; H, 6.54; N, 3.50. Found: C, 41.94; H, 6.55; N; 3.53. ¹H NMR (CD₃CN, 25 °C): δ 3.725 (s, 12H), 3.09 (q, 16H), 1.08 (t, 24H). UV-vis spectrum (CH₃CN): λ_max = 365 (ε = 8110), 429 (2020), 505 (sh, 370), 711 nm (62 M⁻¹ cm⁻¹). ESI-MS (CH₃CN): m/z = 672 {[M]²⁻ + Et₄N⁺}⁻. CV (CH₃CN): E^pa(irrev.) = −0.12 V vs. SCE. IR (KBr): ν 495 (s), 678 (w), 753 (w), 782 (m), 1014 (m), 1068 (w), 1172 (m), 1227 (vs), 1393 (m), 1433 (m), 1456 (w), 1478 (s), 1525 (s), 1684 (s), 1718 cm⁻¹ (s).

(Et₄N)₂[Mo^IVSe(L^COOMe)₂] (MoSeL^COOMe)

To an acetonitrile solution (4 mL) containing Mo(OSi^tBuPh₂)L^COOMe (153 mg, 0.17 mmol), an acetonitrile suspension (1 mL) containing Na₂Se (51 mg, 0.041 mmol) was added. After stirring this suspension for 20 hours, an acetonitrile solution (1 mL) of Et₄NCl (28 mg, 0.17 mmol) was added to the reaction mixture. Undissolved Na₂Se was removed by filtration and the brown filtrate was concentrated to ca. 1 mL. A brown powder was precipitated by addition of diethyl ether to the solution, which was collected by filtration, washed with THF/acetone, and dried in dinitrogen. To an acetonitrile solution (6 mL) of the brown powder, ^tBuPh₂SiCl (31 mL, 0.14 mmol) was added. After stirring the solution for 5 min, the solution was concentrated to 1 mL. Addition of 12 mL of THF to the solution gave a brown powder, which was collected by filtration, washed with THF/acetone, and dried in dinitrogen. Yield: 42 mg (29%). Anal. Calcd for C₂₈H₅₂MoN₂O₈S₄Se (mol wt. 847.89): C, 39.66; H, 6.18; N, 3.30. Found: C, 39.53; H, 6.13; N; 3.46. ¹H NMR (CD₃CN, 25°C): δ 3.725 (s, 12 H), 3.09 (q, 16H), 1.08 (t, 24H). UV-vis spectrum (CH₃CN): λ_max = 376 (ε = 8440), 458 (2020), 575 (sh, 280), 740 nm (42 M⁻¹ cm⁻¹). ESI–MS (CH₃CN): m/z = 719 {[M]²⁻+ Et₄N⁺}⁻. CV (CH₃CN): E^pa (irrev.) = −0.12 V vs. SCE. IR (KBr): ν 675 (w), 780 (m), 1013 (m), 1067 (w), 1171 (m), 1225 (vs), 1393 (m), 1432 (m), 1455 (w), 1478 (s), 1684 (s), 1718 cm⁻¹ (vs).

Physical measurements

FT–IR spectra were recorded with a Perkin–Elmer Spectrum One spectrometer. Resonance Raman spectra were taken on a Jasco NRD–1000 instrument using an Ar⁺ ion laser with excitation at 488 nm. Solid state resonance Raman (rR) spectra and associated rR excitation profiles were collected using a system comprised of an PI/Acton SpectraPro SP-2556 500 mm focal length imaging spectrograph with a triple grating turret and a PI/Acton Spec-10:100B back-illuminated 1340 × 100 pixel digital CCD spectroscopy system with a cryogenically cooled camera head. A Coherent Innova Ar⁺ ion laser was the excitation source. Samples were mixed with either NaCl or a NaCl/Na₂SO₄ mixture with Na₂SO₄ as an internal calibrant. ¹H NMR spectra were recorded with a JEOL Lambda 300, and TMS signal was adjusted to 0 ppm. ⁷⁷Se-NMR spectra were recorded on a JEOL FT-NMR Lambda 300WB spectrometer. (CH₃)₂Se was used as an internal standard. ESI–mass spectra were measured with a JEOL JMS-700S. Routine UV-vis spectra were recorded on a HP-8453 or Shimazu-UV 2550 spectrometer. Additional solution electronic absorption spectra were collected using a Hitachi U-3501 UV-Vis-NIR dual-beam spectrophotometer capable of scanning a wavelength region between 185 and 3200 nm. Spectral samples were dissolved in dry, degassed acetonitrile, and the electronic absorption spectra were measured in a 1 cm pathlength, 100 mL, black-masked, quartz cuvette (Starna Cells, Inc.) equipped with a Teflon stopper. All electronic absorption spectra were performed at room temperature and repeated at regular time intervals to ensure the structural stability and integrity of the complex in solution.

Electrochemistry

Cyclic voltammetric measurements were performed under dinitrogen with a Hokuto Denko HZ-3000 potentiostat. A set of a glassy-carbon working electrode (circular, 3 mm diameter), a SCE reference electrode, and a platinum counter electrode was employed in these experiments.

X-ray crystallography

Single crystals of MoSL^Ph•CH₃CN were obtained by diffusion of diethyl ether into an acetonitrile solution of MoSL^Ph. Single crystals of MoSeL^Ph•DMSO were obtained by diffusion of diethyl ether into a DMSO solution of MoSeL^Ph. Single crystals of MoSL^COOMe and MoSeL^COOMe were obtained by diffusion of diethyl ether into each acetonitrile solution. Single crystals of MoOL^Ph were obtained by recrystallization from DMSO/diethylether. Each single crystal obtained, except for MoOL^Ph, was mounted on a glass fiber, and all X-ray data were collected at −173 °C on a Rigaku CCD diffractometer with monochromatic Mo-Kα radiation. A single crystal of MoOL^Ph was mounted on a glass fiber, and all measurements were made on a Rigaku RAXIS RAPID imaging plate area detector with graphite monochromated Mo-Kα radiation. The structures were solved by direct methods²⁶ and expanded using DIRDIF 99.²⁷ The non–hydrogen atoms were refined anisotropically by full–matrix least squares on F². The hydrogen atoms were attached at idealized positions on carbon atoms and were not refined. All structures in the final stages of refinement showed no movement in the atom positions. The calculations were performed using Single–Crystal Structure Analysis Software, version 3.8.²⁸ Crystallographic parameters are summarised in Table 1.

Table 1.

Crystallographic Data for MoSL^Ph•CH₃CN, MoSeL^Ph•DMSO, MoSL^COOMe, MoSeL^COOMe and MoOL^Ph

	MoSLPh•CH3CN	MoSeLPh•DMSO	MoSLCOOMe	MoSeLCOOMe	MoOLPh
formula	C46H63MoN3S5	C46H66MoN2OS5Se	C28H52MoN2O8S5	C28H52MoN2O8S4Se	C44H60MoN2OS4
fw	914.26	998.24	800.97	847.87	857.15
cryst system	orthorhombic	monoclinic	orthorhombic	orthorhombic	monoclinic
space group	P212121	P21/c	Pbca	Pbca	P21
a, Å	19.519(4)	10.628(2)	12.8056(5)	12.7676(7)	14.733(9)
b, Å	20.011(4)	19.035(4)	18.2219(8)	18.3036(10)	8.778(5)
c, Å	11.887(2)	24.634(5)	31.5759(15)	31.9341(17)	17.409(11)
β, deg	90	107.256(2)	90	90	104.966(10)
V, Å3	4643.1(15)	4759.2(16)	7368.0(6)	7462.8(7)	2175.0(23)
Z	4	4	8	8	2
ρ, g cm−3	1.308	1.393	1.444	1.509	1.309
μ, mm−1	0.541	1.295	0.685	1.596	0.528
data	13116	37472	54114	56065	21205
unique data	11310	10659	7969	8435	9877
R1a	0.0510	0.0568	0.0712	0.0566	0.0443
wR2 (F2)b	0.1261	0.1686	0.1457	0.1689	0.1097
GOF	0.953	0.962	1.053	1.029	1.028

^aR1 = Σ(|F_o| − |F_c|)/Σ|F_o|.

^bwR2 = {Σ(w(Fo² − Fc²)²)/Σw(F_o²)²}^1/2.

Electronic structure calculations

Electronic structure and vibrational frequency calculations were performed at the density functional level of theory using the Gaussian 03W software package.²⁹ All calculations employed the B3LYP hybrid functional and used a LANL2DZ basis set with an effective core potential for Mo. A 6-31G* basis set was used for all light atoms. Input files were prepared using the molecule builder function in the Gaussview software package. The results of these calculations were further analyzed using the program AOMix. Electron density difference maps (EDDMs) were constructed using the GaussSum suite of programs.

Results and discussion

Preparation and characterisation of Mo^IVS/Mo^IVSe complexes with L^Ph and L^COOMe

The first series of the square pyramidal bis(ene-1,2-dithiolate)Mo^IVE (E = O, S, Se) complexes, [Mo^IVE(L^C4H8)₂]²⁻ (MoEL^C4H8), were synthesised from Mo(CO)₂(L^C4H8)₂ and 2 equivalent of hydroxide ion (Et₄NOH), 1 equivalent of sulfido ion (Na₂S) or 1 equivalent of selenido ion (Na₂Se) in dry CH₃CN.^18,19,20 In a similar manner, MoSL^Ph and MoSeL^Ph were synthesised. These two complexes, along with MoOL^Ph, complete the second series of Mo^IVE complexes including an identical ene-1,2-dithiolate ligand. The elemental analysis and negative-ion ESI-MS data for MoSL^Ph and MoSeL^Ph were consistent with the corresponding molecular formulas. For the synthesis of the Mo^IVS and Mo^IVSe species with L^dmed, Mo(OSi^tBuPh₂)L^COOMe was employed. When the compound was treated with one equivalent of EtN₄SH in acetonitrile, a mixture of two complexes was obtained. The negative ESI-MS of the acetonitrile solution containing Mo(OSi^tBuPh₂)L^COOMe and Et₄NSH in a 1: 1 ratio showed two peak clusters at m/z = 656 and 672 in ca. 1: 1 ratio in intensity (Fig. 3), of which one peak cluster at m/z = 656 corresponded to the mono-oxo molybdenum complex MoOL^COOMe ([M²⁻ + Et₄N⁺]⁻), and another one at m/z = 672 was attributed to the mono–sulfido molybdenum complex MoSL^COOMe ([M²⁻ + Et₄N⁺]⁻). The ¹H NMR spectrum of an acetonitrile–d₃ solution containing equimolar amount of Mo(OSi^tBuPh₂)L^COOMe and EtN₄SH exhibited two singlets at 3.685 and 3.725 ppm in a 1: 1 integrate ratio. The former chemical shift was identical to that of the ¹H NMR signal for −COOMe groups of the MoOL^COOMe reported in the literature.²³ The IR spectrum of a powder precipitated from the solution showed strong absorption bands at 911 and 495 cm⁻¹ assignable to ν(MoO) and ν(MoS) stretchings, respectively. Collectively, Mo(OSi^tBuPh₂)L^COOMe was revealed to change to the MoO and MoS compounds upon treatment with SH⁻. The MoS species was purified by washing the mixture with THF containing ^tBuPh₂SiCl since Mo(OSi^tBuPh₂)L^COOMe formed again here is soluble in THF. Similarly, MoSeL^COOMe was synthesised from Mo(OSi^tBuPh₂)L^COOMe and Na₂Se in CH₃CN. Monoselenido molybdenum(IV) complexes are extremely rare and [Mo^IVSe(Se₄)₂]²⁻, ³⁰ is the only example of such species besides the series reported in this paper.

Fig. 3.

ESI mass (a) and ¹H NMR (b) spectra obtained by a treatment of Mo(OSi^tBuPh₂)L^COOMe with 1 equiv of SH⁻ in acetonitrile.

Crystal structures of MoS and MoSe complexes of L^Ph and L^COOMe

The crystal structures of the anionic parts of MoSL^Ph and MoSL^COOMe are shown in Figs. 4 and 5, respectively, and the selected bond distances and angles are listed in Table 2 together with those of MoSL^C4H8. The Mo1 atom of MoSL^Ph is coordinated by a terminal sulfido, S5, and four sulfur atoms, S1–S4 of two L^Ph. The S1, S2, S3 and S4 atoms are almost coplanar and the four bond angles, S5-Mo1-S1 (110.20(3) °), -S2 (112.08(3) °), -S3 (106.16(3) °) and -S4 (110.19(3) °) are found to be very similar. Additionally, the S1-Mo1-S3 bond angle (143.64(3) °) is very similar to the S2-Mo1-S4 (137.71(3) °) angle. Taken together, these observations indicate that each Mo1 atom possesses a distorted square pyramidal geometry as commonly found in the Mo(IV) complexes, [MoS(didentate ligand)₂]²⁻, (bidentate ligand = S₄,²³ CS₄,³¹ L^C4H8,¹⁸ S₂C₂Me₂³²), [MoS(bdt)(S₄)]²⁻,³³ and [MoS(bdtCl₂)(S₄)]²⁻.³³ The Mo1 atom of MoSL^COOMe also adopts a similar distorted square pyramidal geometry as MoSL^Ph, and this is reflected by the S5-Mo1-S (L^COOMe) and S1-Mo1-S3 and S2-Mo1-S4 bond angles. With respect to the dimensions of L^Ph and L^COOMe, the four C-S bond lengths of L^Ph (1.766(3) – 1.789(3) Å) and the three C-S bonds (1.748(6) – 1.774(4) Å), except for the C1-S1 bond (1.735(5) Å) of L^COOMe, are in the typical range of C-S single bonds. The somewhat shorter L^COOMe S1-C1 bond length may be ascribed to the mesomeric effect of the −COOMe ester group attached on S2 atom of MoSL^COOMe, as suggested by the small 6.43(6) ° dihedral angle between the −COOMe and S1-C1-C2-S2 planes. As comparison, other three dihedral angles between −COOMe and S1-C1-C2-S2 or S3-C3-C4-S4 planes of L^COOMe are close to 90 °. The respective C1-C2 and C3-C4 bond lengths of MoSL^Ph (1.355(5) and 1.360(5) Å) and MoSL^COOMe (1.353(7) and 1.358(10) Å) suggest their double bond character. Collectively, these observations indicate that L^Ph and L^COOMe ligands of the MoS complexes have a structure of ene–1,2–dithiolate, as in the case of L^C4H8 of MoSL^C4H8. The Mo1–S5 bond length of the three complexes with L^Ph, L^COOMe and L^C4H8 increases in the order of MoSL^C4H8 (2.167(2) Å) > MoSL^Ph (2.1592(9) Å) > MoSL^COOMe (2.1495(16) Å), indicating that the Lewis acidity of the central molybdenum(IV) ion decreases with an increase in L→Mo electron donation yielding weaker Mo≡S bonds.

Fig. 4.

Crystal structure of the anionic part of MoSL^Ph • CH₃CN shown with 50% ellipsoids. The hydrogen atoms were omitted for clarity.

Fig. 5.

Crystal structure of the anionic part of MoSL^COOMe shown with 50% ellipsoids. The hydrogen atoms were omitted for clarity.

Table 2.

Selected Bond Distances (A) and Angles (deg) of MoSL^Ph•CH₃CN, MoSL^COOMe and MoSL^CC4H8

	MoSLPh•CH3CN	MoSLCOOMe	MoSLC4H8, a
Mo1–S5	2.1592(9)	2.1495(16)	2.167(2)
Mo1–S1	2.3577(8)	2.3740(13)	2.369(2)
Mo1–S2	2.3675(8)	2.3696(14)	2.364(2)
Mo1–S3	2.3455(9)	2.3623(13)	2.382(2)
Mo1–S4	2.3656(8)	2.3601(18)	2.369(2)
S1–C1	1.781(3)	1.735(5)	1.775(8)
S2–C2	1.789(3)	1.774(4)	1.75(1)
S3–C3	1.766(3)	1.748(6)	1.753(8)
S4–C4	1.779(3)	1.770(8)	1.772(9)
C1–C2	1.355(5)	1.353(7)	1.33(1)
C3–C4	1.360(5)	1.358(10)	1.35(1)
Mo1–4S plane	0.7981(5)	0.7968(9)	0.79(1)
S5–Mo1–S1	110.20(3)	108.17(5)	106.46(9)
S5–Mo1–S2	112.08(3)	109.36(6)	110.39(9)
S5–Mo1–S3	106.16(3)	111.53(5	109.12(9)
S5–Mo1–S4	110.19(3)	109.56(6)	111.82(9)
S1-Mo1-S2	81.86(2)	83.07(4)	82.20(8)
S1-Mo1-S3	143.64(3)	140.30(4)	144.42(8)
S1-Mo1-S4	84.95(2)	84.18(5)	84.92(7)
S2-Mo1-S3	85.33(3)	83.62(4)	85.70(8)
S2-Mo1-S4	137.71(3)	141.06(5)	137.78(8)
S3-Mo1-S4	81.99(2)	83.13(6)	81.90(8)

^aref. 19

The crystal structures of anionic parts of MoSeL^Ph and MoSeL^COOMe are shown in Figs. 6 and 7, respectively. The selected bond lengths and angles are indicated in Table 3 together with those of MoSeL^C4H8. The Mo1 atoms of MoSeL^Ph and MoSeL^COOMe are coordinated with one terminal selenido, Se1, and the four dithiolene sulfur atoms, S1–S4, derived from either L^Ph or L^COOMe. The dimensions about each Mo1 atom are similar to those of the corresponding sulfido analogues except for the Mo1–Se1 bond distances, yielding an isostructural distorted square-pyramidal geometry. In MoSeL^Ph, the C–S bond lengths range from 1.768(3) to 1.789(5) Å and the C1-C2 and C3-C4 bond distances are 1.345(5) and 1.380(7) Å. These data suggest that the two L^Ph ligands of MoSeL^Ph are also ene-1,2-dithiolates and not dithiones.³⁴ The two L^COOMe ligands in MoSeL^COOMe also possess an ene-1,2-dithiolate structure as reflected by the C-S bond distances (1.749(6) - 1.769(4) Å) and the C1-C2 and C3-C4 bond lengths (1.345(6) and 1.338(7) Å). The Mo1-Se1 bond (2.3069(5) Å) of MoSeL^C4H8 with the largest L→Mo charge donation is the longest in the Mo1-Se1 bonds of the three MoSeL complexes (see electrochemical section). All three of the Mo1-Se1 bond lengths are longer than the MoSe bond in square-pyramidal [Mo^IVSe(Se₄)₂]²⁻ (2.270(4) Å).³⁰

Fig. 6.

Crystal structure of the anionic part of MoSeL^Ph • DMSO shown with 50% ellipsoids. The hydrogen atoms were omitted for clarity.

Fig. 7.

Crystal structure of the anionic part of MoSeL^COOMe shown with 50% ellipsoids. The hydrogen atoms were omitted for clarity.

Table 3.

Selected Bond Distances (A) and Angles (deg) of MoSeL^Ph • DMSO, MoSeL^COOMe and MoSeL^C4H8

	MoSeLPh•DMSO	MoSeLCOOMe	MoSeLC4H8, a
Mo1–Se1	2.2915(5)	2.2900(5)	2.3069(5)
Mo1–S1	2.3552(11)	2.3637(11)	2.3517(12)
Mo1–S2	2.3400(10)	2.3645(10)	2.3627(11)
Mo1–S3	2.3520(13)	2.3530(13)	2.3639(11)
Mo1–S4	2.3452(8)	2.3636(13)	2.3618(11)
S1–C1	1.768(3)	1.769(4)	1.786(4)
S2–C2	1.789(5)	1.746(4)	1.770(4)
S3–C3	1.785(3)	1.749(6)	1.772(4)
S4–C4	1.784(4)	1.761(5)	1.765(4)
C1–C2	1.345(5)	1.345(6)	1.331(7)
C3–C4	1.380(7)	1.338(7)	1.338(7)
Mo1–4S plane	0.7976(7)	0.7933(7)	0.79(1)
Se1–Mo1–S1	109.45(3)	109.78(3)	110.86(2)
Se1–Mo1–S2	109.66(3)	109.78(3)	107.59(3)
Se1–Mo1–S3	109.35(3)	108.80(3)	113.20(3)
Se1–Mo1–S4	111.00(3)	110.96(3)	106.50(3)
S1-Mo1-S2	82.09(3)	82.91(3)	82.50(4)
S1-Mo1-S3	141.20(4)	141.41(4)	135.94(4)
S1-Mo1-S4	84.64(3)	84.33(4)	85.77(4)
S2-Mo1-S3	84.56(4)	83.74(4)	84.70(3)
S2-Mo1-S4	139.34(4)	139.99(4)	145.90(3)
S3-Mo1-S4	82.20(3)	83.06(4)	81.78(4)

^aref. 20

We report the crystal structure here and find the square-pyramidal geometry and the Mo≡O bond length to be 1.709(2) Å (Fig. 8 and Table 4) although MoOL^Ph had been synthesised previously.²² In summary, the crystal structures presented here show that the bis(L)Mo^IVE series provides a set of isostructural complexes that can be used to evaluate the strength of the Mo≡E bonds as a function of L→Mo charge donation.

Fig. 8.

Crystal structure of the anionic part of MoOL^Ph shown with 50% ellipsoids. The hydrogen atoms were omitted for clarity.

Table 4.

Selected Bond Distances (A) and Angles (deg) of MoOL^Ph, MoOL^COOMe and MoOL^C4H8

	MoOLPh	MoOLCOOMe, a	MoOLC4H8, b
Mo1–O1	1.709(2)	1.686(6)	1.745(6)
Mo1–S1	2.3874(7)	2.380 c	2.418(3)
Mo1–S2	2.3761(7)		2.384(7)
Mo1–S3	2.3824(8)		2.418(5)
Mo1–S4	2.3778(7)		2.39(1)
S1–C1	1.792(2)	1.758 d	1.76(1)
S2–C2	1.782(3)		1.777(8)
S3–C3	1.785(2)		1.81(1)
S4–C4	1.786(3)		1.783(8)
C1–C2	1.345(4)	1.33(1)	1.34(1)
C3–C4	1.369(5)	1.31(1)	1.32(1)
Mo1–4S–plane	0.7871(5)	e	0.76(1)
O1–Mo1–S1	110.45(8)	108.9 f	106.6(4)
O1–Mo1–S2	108.37(7)		106.2(6)
O1–Mo1–S3	110.64(8)		106.5(3)
O1–Mo1–S4	107.04(7)		108.1(4)
S1-Mo1-S2	82.20(2)	83.1(1)	82.4(2)
S1-Mo1-S3	138.89(2)	140.7(1)	146.8(5)
S1-Mo1-S4	86.70(2)	84.3(1)	88.5(4)
S2-Mo1-S3	83.92(2)	85.3(1)	86.7(2)
S2-Mo1-S4	144.57(3)	143.5(1)	145.7(5)
S3-Mo1-S4	82.67(2)	83.0(1)	83.0(4)

^aref. 23,

^bref. 18,

^caverage distance of the four Mo-S bonds,

^daverage distance of the four C-S bonds,

^enot reported,

^faverage angle of the four O1-Mo1-S angles.

Electrochemical properties

Although the MoSL and MoSeL complexes exhibit a redox wave assignable to the Mo(V)/Mo(IV) couple, the reversibility of the redox wave is influenced by the nature of the dithiolene, L. The Mo(V)/Mo(IV) couple for MoSL^C4H8 (E_1/2 = −0.75 V vs. SCE) is reversible on the cyclic voltammetry (CV) time scale using scan rates greater than 20 mV s⁻¹. The Mo(V)/Mo(IV) couple for MoSeL^C4H8 was observed at E_1/2 = −0.74 V and was also observed to be reversible on the CV time scale under the same experimental conditions as MoSL^C4H8. However, both MoSL^Ph and MoSeL^Ph exhibited a noticeable scan rate dependent reversibility in the CV. Using scan rates greater than 300 mV s⁻¹, the Mo(V)/Mo(IV) redox process of these complexes exhibited a reversible wave at −0.43 V for MoSL^Ph and −0.45 V for MoSeL^Ph. The redox wave became irreversible at scan rates less than 300 mV s⁻¹. The dependence of scan rate on the CV of MoSeL^Ph is shown in Fig. 9. These observations indicate that the structure of the monomeric Mo(IV) species is unstable relative to one-electron oxidation processes. This instability likely results from dimerization, as both MoSL^C4H8 and MoSeL^C4H8 have been reported to yield dinuclear {[Mo^V(L^C4H8)₂]₂(μ −S/Se)₂}²⁻ structures upon one-electron oxidation,^19,20 and the UV–vis spectra of solutions obtained by chemical oxidation of MoSL^Ph and MoSeL^Ph with ferrocenium hexafluorophosphate are identical to those of dimeric {[Mo^V(L^Ph)₂]₂(μ − S/Se)₂}²⁻.³⁵ The MoSL^COOMe and MoSeL^COOMe complexes both exhibit an irreversible Mo(V)/Mo(IV) redox couple at E^pa = −0.12 and −0.12 V, respectively, even with scan rates greater than 500 mV s⁻¹, indicating rapid dimerization of a putative square pyramidal Mo(V) species to generate {[Mo^V(L^COOMe)₂]₂(μ − S/Se)₂}²⁻. The marked scan rate dependency on the reversibility of the Mo(V)/Mo(IV) redox process for MoSL and MoSeL supports an argument that an increase in dithiolene L→Mo charge donation can stabilise the square pyramidal Mo(V) species. In the case of MoOL, the one-electron oxidized mononuclear Mo(V) form is stable and does not yield dinuclear {[Mo^V(L)₂]₂(μ-O)₂}²⁻ species. Therefore, the Mo(V)/Mo(IV) redox couple in the CV is reversible. The E_1/2 values for the Mo(S/Se)L series and the corresponding literature values for the MoOL series are summarised in Table 5. Here it is observed that the Mo(V)/Mo(IV) redox potential increases in the order L^COOMe > L^Ph > L^C4H8, indicating that L→Mo charge donation increases in the order of L^C4H8 > L^Ph > L^COOMe. This is the expected trend based on the electronic donating/withdrawing nature of the substituents on the dithiolene, L. In square pyramidal bis(dithiolene)Mo^{IV, V, VI} complexes that possess an apical multiply bonded ligand E, the Mo redox orbital is Mo(d_x2−y2) (see Table 8). This redox orbital is oriented orthogonal to the Mo≡E bond with the principal lobes bisecting all four Mo-S bonds. In the C_4v high-symmetry limit, the Mo(d_x2−y2) orbital is non-bonding with respect to the apical E ligand. As a result, the nature of the Mo≡E bond should not have a dominant influence on the redox potential of a given MoEL complex (E = O, S, Se) at parity of the equatorial ligands. We observe that the redox potentials for the L^Ph and L^C4H8 compounds as a function of E are essentially uneffected by the nature of the E donor, with potentials that vary by only 30–50 mV across a given series. This should be compared with the much larger (i.e. ~300 mV) potential shift as a function of the dithiolene for a given terminal E donor. For the L^COOMe compounds, MoOL^COOMe possesses a redox potential that is 90 mV more positive than for MoOL^COOMe and MoOL^COOMe. In summary, the redox potentials of the MoEL series are dominated by the nature of the dithiolene, but are much less affected by the nature of the terminal E donor despite the fact that the pK_b values of O²⁻ (≪0), S²⁻ (−0.85), and Se²⁻ (~3) are significantly different.³⁶

Fig. 9.

Scan rate dependent cyclic voltammogram of MoSeL^Ph in CH₃CN, 500 (black), 400 (brown), 300 (green), 200 (purple), blue (100), and 50 mV s⁻¹ (red).

Table 5.

Redox potentials (V vs. SCE) of Mo(V)/Mo(IV) process for MoEL

Complexes	LCOOMe	LPh	LC4H8
MoOL	−0.03	−0.46	−0.70
MoSL	−0.12	−0.43	−0.75
MoSeL	−0.12	−0.45	−0.74

Table 8.

Percentage Fragment Orbital Character to the Frontier Molecular Orbitals of MoEL^COOMe

MoOLCOOMe
Molecular Fragment	% Mo	% Sdithiolene	% Dithiolene (total)	% Oxo
HOMO-1	0.8	68.3	95.0	4.1
HOMO	81.2	10.1	18.5	0.3
LUMO	5.9	9.9	90.3	2.8
LUMO+1	3.3	13.5	95.6	0.0
LUMO+2	36.4	7.1	53.9	8.6
LUMO+3	43.3	9.2	45.7	10.2

MoSLCOOMe
Molecular Fragment	% Mo	% Sdithiolene	% Dithiolene (total)	% Sulfido
HOMO-1	0.4	61.0	83.3	15.7
HOMO	80.3	9.8	18.8	0.8
LUMO	21.7	8.7	65.9	11.8
LUMO+1	3.9	13.7	94.4	0.6
LUMO+2	57.7	17.5	22.7	19.5
LUMO+3	39.9	14.4	47.5	12.1

MoSeLCOOMe
Molecular Fragment	% Mo	% Sdithiolene	% Dithiolene (total)	% Selenido
HOMO-1	0.5	56.2	76.6	22.8
HOMO	80.0	9.9	19.1	1.0
LUMO	25.8	8.7	60.0	13.6
LUMO+1	6.2	13.8	91.2	1.7
LUMO+2	55.8	18.2	26.7	17.5
LUMO+3	37.7	15.0	51.4	10.4

⁷⁷Se NMR spectra of MoSeL complexes

Given the effects of E on the redox potentials of the MoEL series at parity of L are minimal, we now turn our investigation toward the effects of L on the Mo≡E bond at parity of E. For the MoSeL complexes, this can conveniently be accomplished using ⁷⁷Se NMR spectroscopy (Fig. 10). To the best of our knowledge, MoSe(NR)([N]Ar)₂ (R = C(CD₃)₂CH₃, Ar = 3,5–C₆H₃Me₂) and trans–Mo(PMe₃)₄(Se)₂ are the only examples of terminal selenido complexes that have been probed by ⁷⁷Se NMR spectroscopy.^10,37 Our experiments reveal the ⁷⁷Se NMR resonances to shift to lower magnetic field in the order MoSeL^C4H8 (1827.8) < MoSeL^Ph (1895.4) < MoSeL^COOMe (2175.4 ppm). This clearly indicates that electron withdrawing substituents on L result in an increase in Se²⁻→ Mo(IV) charge donation, which in turn results in a less basic terminal selenide. These results are important, since they suggest that the nature of the dithiolene ligand can fine-tune the acid/base and the atom transfer reactivity of terminal MoE sites in enzymes and related systems.

Fig. 10.

⁷⁷Se NMR spectra of MoSeL^C4H8 (above), MoSeL^Ph (middle), and MoSeL^COOMe (bottom) in (CD₃)₂SO.

Spectroscopic and electronic structure studies

Solution electronic absorption spectra for MoSL and MoSeL were collected in CH₃CN. The electronic absorption spectra of the MoOL series have been reported previously.^18,22,23 Inspection of the absorption spectra for MoSL and MoSeL complexes reveals that they can be divided into three sections: the first region occurs at wavelengths lower than ca. 350 nm where the complexes possesses intense absorption bands with molar extinction coefficients (ε) over ~10,000 M⁻¹ cm⁻¹, the second region occurs between 350 – 600 nm and displays a well-defined band with an extinction coefficient of ca. 2000 M⁻¹ cm⁻¹, and finally a third low-energy region that consists of a series of weak absorption bands. The well-defined band that appears in the second spectral region was previously tentatively assigned as an E→Mo(IV) charge transfer band, with respect to the prior spectral assignments for trans–Mo(E)₂(PMe₃)₄ (E= O, S, or Se) and trans–Mo(E)₂(P–P)₂ (E = O, S, and Se).^37,10 The solution electronic absorption spectra of MoEL^COOMe (E = O, S, Se) and MoSeL are shown in Fig. 11. Interestingly, the prominent band in the second spectral region is shifted progressively to lower energy in the MoEL^COOMe series as a function of the terminal chalcogen donor atoms (MoOL^COOMe, 380 nm; MoSL^COOMe, 429 nm; MoSeL^COOMe, 458 nm), and this reflects a significant degree of E wavefunction mixing with the dithiolene frontier orbitals, vide infra. Finally, this CT band shifts to higher energy with a decrease in L → Mo charge donation (E = S; 462 nm for L^C4H8, 450 nm for L^Ph, 429 nm for L^COOMe: E = Se; 489 nm for L^C4H8, 480 nm L^Ph, 458 nm for L^COOMe).

Fig. 11.

Absorption spectra in visible region in CH₃CN: (a) MoEL^COOMe (O (red), S (blue), and Se (black)), (b) MoSeL (L^C4H8 (red), L^Ph (blue), and L^COOMe (black)).

Understanding the nature of key M-L vibrations in these MoEL (E = O, S, Se) complexes is important in developing a greater understanding of important bonding interactions in both synthetic model systems and the enzymes, particularly those of the DMSOR enzyme family that possesses bis-dithiolene coordination. Solid-state rR spectra for the MoEL^COOMe series are presented in Fig. S1 (Supplementary Information) and summarised in Table 6. Additionally, solution resonance Raman (rR) spectra of MoOL, MoSL and MoSeL^{Ph, COOMe} in CH₃CN were obtained using 488 nm excitation. The Raman data are presented in Figs. S2 and S3 (Supplementary Information) and are summarized in Table 7 for comparative purposes. We were not able to obtain a Raman spectrum for MoSeL^C4H8 due to decomposition as a result of the incident laser radiation. Vibrational assignments are consistent with calculated virbrations and earlier vibrational assignments made by Spiro et al for MoOL^COOMe.³⁸

Table 6.

Solid State Resonance Raman Data (cm⁻¹) for MoEL^COOMe

Complexes	ν(Mo≡E)	ν(Mo-S)dithiolene	ν(C=C)sym	ν(C=O)sym
MoOLCOOMe	903	364	1530	1691
MoSLCOOMe	493	364	1527	1702
MoSeLCOOMe	370a	344 a	1525	1700

^avibrational frequency calculations indicate is significant degree of mixing between ν(Mo≡Se) and ν(Mo-S)_ditholene resulting from the small energy difference between these two totally symmetric vibrations.

Table 7.

Solution Resonance Raman Data (cm⁻¹) for MoEL

Complexes	ν(Mo ≡E) stretch	ν(C=C) stretch
MoOL	897 (LC4H8), 903 (LPh), 911 (LCOOMe)a	1596 (LC4H8), c, 1533 (LCOOMe)
MoSL	480 (LC4H8), 487 (LPh), 500 (LCOOMe)	1593 (LC4H8), c, 1531 (LCOOMe)
MoSeL	b c 347 (LCOOMe)d	b, c, 1531 (LCOOMe)

^aref. 38.

^bMoSeL^C4H8 was decomposed upon irradiation.

^cnot assigned unambiguously.

^dvibrational frequency calculations indicate both Mo≡Se and Mo-S_dithiolene stretches contribute to the 347 cm⁻¹ mode.

With respect to the MoSeL^C4H8 and MoEL^Ph (E = O, S, Se) solution rR spectra, neither ν(Mo≡Se) for MoSeL^C4H8 or ν(C=C) for MoEL^Ph could be assigned due to spectral overlap of the intrinsic compound vibrations with modes associated with the solvent. For the MoEL^COOMe (E = O, S, Se) series, there is very good agreement between the solution and solid state spectra and this indicates the integrity and geometry of the complexes are maintained in solution. The Raman data for the MoEL^COOMe series reveal that the ν(Mo≡E) stretching frequency is shifted to lower energy according to O>S>Se, and reflect reduced mass differences within the Mo≡E unit. This is in marked contrast to the ν(C=C) and ν(M-S_dithiolene) stretching frequencies across this series which remain essentially constant. These observations indicates that E→Mo charge donation does not dominate Mo-S_dithiolene bonding interactions or bonding interactions within the dithiolenes. However, the ν(Mo≡O/S) stretch in the Mo(O/S)L series does shift to lower frequency as a function of increased L→Mo charge donation, and this effect is greatest for the MoSL series. Specifically, this reflects the relative degree of L→Mo charge donation into the Mo(d_xz, d_yz) orbitals, which are π-antibonding with respect to the Mo≡E bonding scheme, and the amount of E character admixed into the Mo(d_xz, d_yz) orbitals. These results are also consistent with the change in Mo≡S bond length for the MoSL complexes (2.167(2) Å for L^C4H8 > 2.1592(9) Å for L^Ph > 2.1495(16) Å for L^COOMe). Finally, our vibrational data indicate that electron withdrawing substituents on the dithiolene noticeably reduce the ν(C=C) stretching frequency (1593 cm⁻¹ for L^C4H8 > 1531 cm⁻¹ for L^COOMe), and this is due to electron withdrawal from occupied frontier dithiolene orbitals that possess C=C π-bonding character.

Electronic structure calculations have been performed on the MoEL^COOMe series in order to obtain insight into the nature of the frontier molecular orbital bonding scheme and spectroscopy of these complexes. The relevant frontier molecular orbitals for MoOL^COOMe and MoSeL^COOMe are shown in Fig. 12, and the fragment orbital contributions to these MOs are listed in Table 8. The frontier molecular orbitals for MoSL^COOMe are provided in Fig. S4 since their appearance is essentially identical to those of MoSeL^COOMe. Assuming idealized C_2v symmetry, the HOMO-1 for these MoEL^COOMe complexes is principally a dithiolene S based MO of b₁ symmetry with a noticeable degree of apical E(p_x) character (Table 8). The E character in the HOMO-1 orbital is observed to decrease from E=Se to E=O, and this reflects the progressively higher valence ionization energy and electronegativity of the terminal E donor that is commensurate with a decrease in principle quantum number. As mentioned previously, the HOMO is the Mo(d_x2−y2) redox orbital of a₁ symmetry (idealized C_2v) and is essentially invariant across the series with ~80% Mo character and <1% terminal E character (Table 8), consistent with our vibrational and electrochemical results, vide supra. The LUMO and LUMO+1 for the MoEL^COOMe complexes are dithiolene based orbitals with C=C π* character. However, the LUMO also possesses some Mo(d_yz)-E π* character which decreases appreciably (MoOL^COOMe: 2.8% O; MoEL^COOMe: 11.8% S; MoEL^COOMe: 13.6% Se; Table 8) with increasing apical ligand electronegativity. These results are interesting in that there exists a pair of low energy acceptor orbitals with significant ligand character that lie between the filled Mo(d_x2−y2) HOMO and the rest of the Mo(IV) d-orbital manifold. Importantly, this provides a mechnanism for low-energy MLCT transitions in MoEL^COOMe compounds. The LUMO+2 and LUMO+3 orbitals for MoOL^COOMe are the Mo(d_yz) and Mo(d_xz) orbitals, respectively, and these also possess π antibonding interactions with the terminal oxo p_y and p_x orbitals. This orbital ordering is reversed for Mo(S/Se)L^COOMe with the Mo(d_yz) observed at higher energy than the Mo(d_xz). This results from stronger orbital mixing with the dithiolene-based LUMO, which also possesses some Mo(d_yz)-E π* character. The LUMO+4 for Mo(S/Se)L^COOMe and the LUMO+6 for Mo(O)L^COOMe are higher energy unoccupied Mo(d_z2) orbitals which are σ* with the apical E donor. Due to the stronger ligand field of the terminal oxo donor, the Mo(d_z2) orbital is at higher energy for Mo(O)L^COOMe.^39,40 The LUMO+10 for Mo(S/Se)L^COOMe and the LUMO+9 for Mo(O)L^COOMe are the unoccupied Mo(d_xy) orbitals which are σ* with the equatorial dithiolene S donors. The high energy of the Mo(d_z2) and Mo(d_xy) orbitals restricts their ability to function as acceptor orbitals in low-energy LMCT transitions.

Fig. 12.

Frontier Kohn-Sham orbitals for MoSeL^COOMe. The corresponding orbitals for Mo(O/S)L^COOMe are very similar with important differences detailed in the text.

The electronic absorption spectrum for MoOL^COOMe is presented in Fig. 13a along with resonance Raman excitation profiles for the ν(Mo-S_dithiolene), ν(Mo≡O), ν(C-O + C-C), ν(C=C) and ν(C=O) vibrations. Gaussian resolution of the electronic absorption spectrum reveals the presence of four bands below ~27,000 cm⁻¹. These bands have been assigned (Table 9) based on their <Fig. 13, Table 7> relative intensities and energies, the nature of the rR excitation profiles, and via a computationally assisted approach using the results of our bonding (Fig. 13) and time dependent density functional theory calculations. From these data, we can assign bands 1 and 2 as the HOMO → LUMO and HOMO → LUMO+1 MLCT transitions. These assignments are strongly supported by the nature of the rR profile, which shows resonance enhancement of ligand ν(C-O + C-C), ν(C=C) and ν(C=O), but no resonance enhancement of ν(Mo-S_dithiolene) or ν(Mo≡O) under band 1. The fact that the ν(Mo≡O) mode is not resonantly enhanced under band 1 confirms the results of our bonding calculations which show the expected result that the Mo(d_x2−y2) HOMO wavefunction possesses no oxo character. Furthermore, the lack of a ν(Mo≡O) stretch indicates a minimal admixture of Mo(d_xz, d_yz)-O π* character in the ligand-based LUMO and LUMO+1 orbitals. Bands 3 and 4 are assigned as HOMO-1→LUMO and HOMO-1→LUMO+1 transitions, respectively. Since the HOMO-1 is non-bonding with the Mo centre, we anticipate no enhancement of the ν(Mo-S_dithiolene) stretch and enhancement of ν(C-O + C-C), ν(C=C) and ν(C=O) vibrations and this is observed experimentally, confirming these band assignments. The nature of the HOMO-1→LUMO and HOMO-1→LUMO+1 transitions are unusual for this type of complex as they are intraligand transitions described by a charge transfer from the dithiolene S donors to the carbon backbone of the dithiolene ligand. A similar intraligand charge transfer transition has very recently been observed in Tp*MoO(S₂BMOQO) (BMOQO = 2-(3-butynyl-2-methyl-2-ol)quinoxaline), which possesses a donor-acceptor type dithiolene ligand.⁴¹ The nature of the intraligand charge transfer is apparent in the electron density difference map (EDDM) constructed for the HOMO-1→LUMO+1 transition (Fig. 13, band 4). The observation of such intraligand charge transfer bands in MoOL^COOMe underscores the electron withdrawing nature of the −COOMe groups in L^COOMe dithiolenes.

Fig. 13.

Gaussian resolved solution electronic absorption spectra for MoOL^COOMe (A), MoSL^COOMe (B), and MoSeL^COOMe (C). Resonance Raman excitation profiles for MoOL^COOMe and MoSeL^COOMe are overlayed on their respective absorption spectra. An electron density difference map that details the nature of the intraligand transition (HOMO-1→LUMO+1) in MoOL^COOMe (red: electron density loss in transition, green: electron density gain in transition) is presented in A (inset).

Table 9.

Band Assignments for Low-Energy Transitions in MoEL^COOMe Compounds

	MoOLCOOMe		MoSLCOOMe		MoSeLCOOMe
Band	Energy	Assignment	Energy	Assignment	Energy	Assignment
1	13,320	HOMO→LUMO	12,750	HOMO→LUMO	12,850	HOMO→LUMO
2	17,720	HOMO→LUMO+1	14,530	HOMO→LUMO+2	14,525	HOMO→LUMO+2
3	21,980	HOMO-1→LUMO	17,242	HOMO-1→LUMO	17,140	HOMO-1→LUMO
4	25,980	HOMO-1→LUMO+1	20,290	HOMO→LUMO+1 HOMO-1→LUMO+2	19,880	HOMO→LUMO+1 HOMO-1→LUMO+2
5		a	23,430	HOMO-1→LUMO+1 HOMO-1→LUMO+3	21,180	HOMO-1→LUMO+1 HOMO-1→LUMO+3

^anot observed

The electronic absorption spectra of MoSL^COOMe and MoSeL^COOMe are presented in Figs. 13b and 13c, respectively, and their transition energies and band assignments are summarised in Table 9. Resonance Raman excitation profiles have been constructed for the ν(Mo-S_dithiolene), ν(Mo≡Se), ν(C-O + C-C), ν(C=C) and ν(C=O) vibrations of MoSeL^COOMe and these are overlayed on the absorption spectrum of MoSeL^COOMe in Fig. 13c. Gaussian resolution of the electronic absorption spectra for MoSL^COOMe and MoSeL^COOMe yield five bands at energies below ~27,000 cm⁻¹. Due to the similar spectral features observed for MoSL^COOMe and MoSeL^COOMe we will discuss their band assignments together. Band 1 is assigned as the HOMO → LUMO ligand field transition (Mo(d_x2−y2) → Mo(d_yz)) that possesses some MLCT character (Fig. 12) while band 2 is assigned as the HOMO → LUMO+2 ligand field transition (Mo(d_x2−y2) → Mo(d_xz)). The small energy splitting of these bands, coupled with their similar transition energies, indicates nearly equivalent ligand field strengths for the terminal sulfido and selenido donors in MoSL^COOMe and MoSeL^COOMe. Band 3 is the first LMCT band and is primarily described as a HOMO-1 → LUMO one-electron promotion with some intraligand CT character. Both the HOMO → LUMO+1 and HOMO-1 → LUMO+2 contribute to the spectral region defined by band 4. This is supported by resonance enhancement of both the ν(Mo-S_dithiolene) and the ν(C=C) modes in MoSeL^COOMe. Interestingly, the ν(Mo≡Se) mode is only weakly resonance enhanced with excitation into band 4 and this likely results from a smaller contribution of the HOMO-1 → LUMO+2 one electron promotion to this band compared to the HOMO → LUMO+1. Large resonance enhancement of ν(C=C), ν(C=O), ν(Mo-S_dithiolene) and ν(Mo≡E) vibrational modes supports an assignment of both the HOMO-1 → LUMO+1 and HOMO-1 → LUMO+3 one-electron promotions as dominant contributors to band 5. The EDDMs for the two transitions that principally contribute to this band in MoSeL^COOMe are given in Fig. 14. Collectively, the assignments for band 5 in MoSL^COOMe and MoSeL^COOMe, and for the corresponding band 4 in MoOL^COOMe indicate that the one-electron promotions that contribute to these transitions originate from the HOMO-1 orbital, and this orbital possesses increased E character in the order O < S < Se. Therefore, the E character in the HOMO-1 orbitals contribute to the progressive shift of this band to higher energies, and this is consistent with the electronegativity and valence ionization energy of the terminal E donor.⁴⁰

Fig. 14.

Electron density difference maps for the two transitions that comprise band 5 that detail the nature of the intraligand (HOMO-1→LUMO+1) and LMCT (HOMO-1→LUMO+3) one-electron promotions in MoSeL^COOMe (red: electron density loss in transition, green: electron density gain in transition) is presented in A (inset). Notice the increased in the Se→Mo charge transfer contribution to this transition relative to MoOL^COOMe (Figure 13A inset).

Conclusions

We have synthesised and structurally characterised new square pyramidal bis(ene-1,2-dithiolate)Mo^IV(S/Se) complexes with L^C4H8, L^Ph and L^COOMe ligands that, when coupled with the known Mo^IVO complexes of L^Ph and L^COOMe, have enabled a systematic study of the effects of terminal chalcogenido and ene-1,2-dithiolate ligands on geometric and electronic structure. The structural data indicate that the Mo≡E bond lengths decrease according to L^C4H8 > L^Ph > L^COOMe, reflecting the relative electron donating and withdrawing nature of the substituents on the dithiolene ligands. Dithiolene ligand effects on the nature of the Mo≡E bonds have been probed in our electrochemical investigations, where it was shown that the redox potential for the Mo(V)/Mo(IV) couple was shifted more positive as the dithiolene ligand L changed from L^C4H8 to L^Ph to L^COOMe. Electrochemical, spectroscopic, and bonding calculations are consistent with a redox orbital that is dominantly Mo(d_x2−y2) in character and possesses no contribution from the terminal E donor. ⁷⁷Se NMR spectroscopy on MoSeL complexes showed that the ⁷⁷Se resonances shifted to lower magnetic field as the terminal selenido donor became less basic. Optical spectroscopic studies on the MoEL^COOMe (E = O, S, Se) series have allowed detailed band assignments to be made for all three compounds, and this has resulted in a greater understanding of their electronic structures. Specifically, it was found that certain low-energy transitions possessed considerable intraligand charge transfer character described by a substantial dithiolene S charge donation to the carbon backbone of the dithiolene ligand, underscoring the electron withdrawing nature of the −COOMe substituents in the L^COOMe dithiolenes. The Mo(d_x2−y2) → Mo(d_xz, d_yz) ligand field transitions were assigned for the MoEL^COOMe (E = S, Se) complexes yielding a spectroscopic t_2g splitting of ~14,500 cm⁻¹. This may be compared with an ~12,000 cm⁻¹ Mo(d_x2−y2) → Mo(d_xz, d_yz) ligand field for des-oxo Mo^IV(ER)(dithiolene) (E = O, S, Se) complexes studied by Holm, Kirk and coworkers.⁴⁰ Finally, the origin of spectral shifts in the MoEL^COOMe (E = O, S, Se) series can be explained by 1) the much stronger ligand field exerted by the terminal oxo ligand that results in markedly higher energy ligand field bands for MoOL^COOMe, and 2) the nature of the dithiolene-based LUMO wavefunction, which possesses increased E character and leads to lower energy HOMO→LUMO and HOMO-1→LUMO CT transitions as the apical chalcogenido electronegativity decreases.