Abstract
The first Lewis acid base adducts of MoF6 and an organic base have been synthesized, i. e., MoF6(NC5H5) and MoF6(NC5H5)2. These adducts are structurally characterized with X‐ray crystallography, showing that both adducts adopt capped trigonal prismatic structures. The MoF6(NC5H5) and MoF6(NC5H5)2 adducts are fluxional on the NMR time scale at room temperature. Two different fluorine environments could be resolved by 19F NMR spectroscopy at −80 °C for the 1 : 2 adduct, MoF6(NC5H5)2, whereas MoF6(NC5H5) remains fluxional at that temperature. Density functional theory (DFT) calculations aide the assignment of the infrared and Raman spectra. Natural Bond Order and Molecular Electrostatic Potential analyses elucidate the structures and properties of the MoF6 pyridine adducts. Regions of significantly higher molecular electrostatic potential, i. e., σ‐holes, in trigonal prismatic compared to octahedral MoF6 rationalize the capped trigonal prismatic geometry of the adducts. Whereas MoF6(NC5H5) is stable at room temperature under exclusion of moisture, MoF6(NC5H5)2 decomposes at 60 °C in pyridine solvent, and the solid slowly decomposes at room temperature after 24 h.
Keywords: Lewis acids, Adducts, Molybdenum, Fluorine, X-ray crystallography
Molybdenum hexafluoride was shown to form Lewis acid base adduct with the organic base pyridine, the first neutral transition metal adducts beyond those of WF6. The MoF6(NC5H5) and MoF6(NC5H5)2 adducts adopt capped trigonal prismatic geometries. The trigonal prismatic MoF6 moiety in the adducts can be explained by the high electrostatic potentials of the square faces, where pyridine can coordinate.
Introduction
The 16 confirmed hexafluorides range in properties from the inert SF6 to the extremely strong oxidizer PtF6. [1] A subgroup of hexafluorides are transition metal hexafluorides, which adopt octahedral geometries. Among the transition metal hexafluorides, MoF6, WF6, TcF6, and ReF6 exhibit Lewis‐acid properties as displayed by their adducts with fluoride ions. Among the transition metal hexafluorides, molybdenum hexafluoride and tungsten hexafluoride are the least oxidizing ones. The initial claim of the preparation of the other possible group‐6 hexafluoride, CrF6, has not been substantiated, and CrF6 has been predicted to be unstable under ambient pressure conditions, but has been predicted to be possibly accessible at high pressures.[ 1 , 2 , 3 ] Molybdenum hexafluoride is a somewhat weaker Lewis acid (fluoride‐ion affinities: MoF6 316 kJ/mol, WF6 325 kJ/mol), but a stronger oxidizer than WF6. [4] The latter property is exemplified by MoF6 being able to oxidize NO, whereas WF6 cannot.[ 5 , 6 ] The trigonal prismatic structures of MoF6 and WF6 are transition states for the interconversion between the octahedral ground‐state structures, called Bailar‐twist, having energy barriers of 28 and 46 kJ/mol, respectively.[ 7 , 8 ] On the other hand, Mo(CH3)6 and W(CH3)6 have trigonal prismatic ground states. Molecular orbital and valence bond models suggest that the octahedral geometry is favoured for six‐coordinate transition metal compounds over the trigonal prismatic geometry when the ligands can act as π‐donors, like the fluorido ligand. [9]
Derivatization of MoF6 via substitution of fluorido ligands has been studied. [10] Alkoxide‐fluoride substitution using (CH3)2Si(OCH3)2 and (CH3)Si(OCH3)3 yield [MoF(OCH3)5] and [MoF2(OCH3)4], respectively, although under certain conditions explosions were observed. [11] Full substitution of the fluorido ligands of MoF6 was accomplished with (CH3)3SiN3 producing the shock‐sensitive Mo(N3)6. [12] Chloride‐fluoride substitution is observed when MoF6 reacts with BCl3 forming MoCl6. [13] The MoVI oxide fluorides MoOF4 and MoO2F2 are Lewis acids and show extensive coordination chemistry with organic ligands.[ 10 , 14 ] For example, MoOF4 forms the hexacoordinate MoOF4(NC5H5) and heptacoordinate MoOF4(NC5H5)2 adducts. [15]
Among the transition metal hexafluorides, WF6 is the only one that has been reported to form neutral adducts, consistent with it having the lowest oxidizing strength. Neutral adducts of WF6 with nitrogen bases, phosphines, sulfides, and even arsines have been reported.[ 14 , 16 , 17 , 18 , 19 ]
The WF6 adducts with pyridine, i. e., WF6(NC5H5) and WF6(NC5H5)2, were synthesized by reacting WF6 with NC5H5 in a 1 : 1 and 1 : 2 molar ratios, respectively, in dichloromethane. [20] Single‐crystal X‐ray diffraction showed that both adducts have capped trigonal prismatic geometries.[ 20 , 21 ] Fluorine‐19 NMR spectroscopy at room temperature show a singlet resonance for each adduct even though two fluorine environments are expected for the rigid structures, suggesting rapid exchange of fluorine environments. The singlet splits into quintets and triplets at low temperatures. [20] For WF6(NC5H5) and WF6(NC5H5)2, these two multiplets are well resolved at 140 and 193 K, with approximate energy barriers for the exchange of 47 and 142 kJ mol−1, respectively. [20]
Other notable adducts of WF6 are those with P(CH3)3, P(CH3)2(C6H5), AsR3 (R=CH3, C2H5), and bidentate nitrogen bases.[ 17 , 18 , 22 , 23 ] The structure of WF6{P(CH3)3} is capped trigonal prismatic as is the case for most WF6 7‐coordinate adducts, while WF6{P(CH3)2(C6H5)} is the only reported WF6 adduct that adopts a different geometry, being capped octahedral. [17]
Whereas adducts of MoVF5 have been reported, i.e., MoF5(NCCH3) and MoF5(NC5H5), no definitive evidence for the Lewis acid behaviour of MoVIF6 towards organic bases has been obtained.[ 15 , 24 , 25 ] Herein, we report the synthesis and conclusive characterization of neutral adducts of MoF6 with pyridine, which are the first reported adducts of MoF6 with organic bases.
Results and Discussion
Synthesis and Properties of MoF6(NC5H5) n (n=1, 2)
The reaction of MoF6 with pyridine in a 1 : 1 molar ratio in CH2Cl2 yielded an orange solid with clear orange supernatant and, when volatiles were removed, afforded bright orange solid MoF6(NC5H5) (Equation 1). The solid was stable over several months at room temperature under exclusion of moisture, as determined by Raman spectroscopy. The 1 : 1 adduct is soluble in CH3CN (orange solution) and pyridine (yellow solution), although when dissolved in pyridine MoF6(NC5H5) was converted to MoF6(NC5H5)2. Isolated MoF6(NC5H5) was sparingly soluble in CH2Cl2 or SO2 forming yellow solutions, although MoF6(NC5H5) exhibits higher solubility in CH2Cl2, when being prepared in this solvent, forming a bright orange solution.
![]() |
(1) |
The reaction of MoF6 with 2 molar equivalents of pyridine in CH2Cl2 afforded the MoF6(NC5H5)2 adduct (Equation 1) as a bright yellow‐orange powder upon removal of volatiles. During the synthesis, the reaction mixture should not be allowed to warm to room temperature except for a brief moment, because decomposition is observed in CH2Cl2. The 1 : 2 adduct slowly degrades in the solid state at ambient temperature with additional Raman bands appearing after three days. The adduct is soluble in SO2, pyridine, and sparingly soluble in CH2Cl2, all giving yellow solutions.
When heating a solution of MoF6(NC5H5)2 in pyridine to 60 °C the solution changes from bright yellow to a dull orange over the period of 1 h 20 min. The resulting solution did not show any 19F NMR resonances other than impurities, suggesting a paramagnetic decomposition product.
Molecular Geometries
The structures of MoF6(NC5H5) and MoF6(NC5H5)2 were studied by X‐ray crystallography and gas‐phase optimizations at the B3LYP and MN15/cc‐pVTZ(C, H)/aug‐cc‐pVTZ(N, F)/aug‐cc‐pVTZ‐PP (Mo) levels of theory (aVTZ). Selected crystallographic parameters are listed in Table 1, and selected experimental and calculated metric parameters for both adducts are given in Table 2. The complete lists of metric parameters are given in Tables S1 and S2. Figure 1 shows the thermal ellipsoid plots of MoF6(NC5H5) and MoF6(NC5H5)2 together with the respective optimized geometries. Figures S4 and S5 show the crystal packing of the two adducts with the π‐π stacking of the pyridine rings.
Table 1.
Crystallographic Data Collection and Refinement Parameters for MoF6(NC5H5) n (n=1, 2).
|
MoF6(NC5H5) |
MoF6(NC5H5)2 |
|
|---|---|---|
|
space group |
triclinic, P |
monoclinic, Cc |
|
Z |
2 |
4 |
|
a (Å) |
6.8039(3) |
12.2505(4) |
|
b (Å) |
7.9088(3) |
7.9251(2) |
|
c (Å) |
7.9640(3) |
12.5640(4) |
|
α (°) |
70.679(3) |
90 |
|
β (°) |
81.176(3) |
104.358(4) |
|
γ (°) |
78.215(3) |
90 |
|
R1 [I≥2σ(I)][a] |
0.0499 |
0.0329 |
|
wR2 [I≥2σ(I)][b] |
0.1450 |
0.0799 |
|
CCDC number |
2370068 |
2370069 |
[a] .
[b] .
Table 2.
Selected metric parameters for the experimental and optimized structures (B3LYP and MN15) for MoF6(NC5H5) and MoF6(NC5H5)2 with bond lengths in Å and angles in °.
|
Exptl. |
Calcd. |
||
|---|---|---|---|
|
B3LYP |
MN15 |
||
|
MoF6(NC5H5)[a] | |||
|
Mo−F(1) |
1.846(5) |
1.866 |
1.850 |
|
Mo−F(2) |
1.854(3) |
1.866 |
1.850 |
|
Mo−F(3) |
1.861(4) |
1.866 |
1.850 |
|
Mo−F(4) |
1.836(5) |
1.866 |
1.850 |
|
Mo−F(5) |
1.861(4) |
1.856 |
1.837 |
|
Mo−F(6) |
1.859(4) |
1.856 |
1.837 |
|
Mo−N |
2.260(5) |
2.346 |
2.278 |
|
MoF6(NC5H5)2 [a] | |||
|
Mo−F(1) |
1.875(2) |
1.882 |
1.875 |
|
Mo−F(2) |
1.879(3) |
1.882 |
1.875 |
|
Mo−F(3) |
1.895(3) |
1.881 |
1.867 |
|
Mo−F(4) |
1.897(2) |
1.881 |
1.867 |
|
Mo−F(5) |
1.898(2) |
1.881 |
1.867 |
|
Mo−F(6) |
1.897(3) |
1.881 |
1.867 |
|
Mo−N(1) |
2.331(4) |
2.478 |
2.364 |
|
Mo−N(2) |
2.326(3) |
2.478 |
2.364 |
|
N(1)−Mo−N(2) |
125.88(11) |
126.4 |
125.7 |
[a] Atom numbering as shown in Figure 1.
Figure 1.

Thermal ellipsoid plots (50 % probability level) of a) MoF6(NC5H5) and c) MoF6(NC5H5)2 and their B3LYP/aVTZP optimized structures b) and d), respectively. Optimized structures correspond to the local minima with the same conformations as the those of the experimental structures, i. e., conformations 1 (vide infra).
The 1 : 1 adduct, MoF6(NC5H5) is isomorphous to WF6(NC5H5), crystallizing in the triclinic P space group, adopting a monocapped trigonal prismatic molecular structure. The MoF6(NC5H5)2 adduct adopts a bicapped trigonal prismatic geometry similar to that of WF6(NC5H5)2. However, the two 1 : 2 adducts are not isomorphous. The MoF6(NC5H5)2 adduct crystallizes in the non‐centrosymmetric, polar Cc space group, which manifests in the MoF6(NC5H5)2 molecules being oriented with the pyridine groups pointing in the same direction of the c‐axis (Figure S6). In contrast, WF6(NC5H5)2 crystallizes in the centrosymmetric Pnma space group with the pyridine groups alternating in their orientations (Figure S7). The valence geometry index for octacoordinated complexes, τ8, [26] gives τ8 values of 1.10 and 0.13 for opposite faces of for MoF6(NC5H5)2, which is consistent with the structure being bicapped trigonal prismatic (Figure S8).
Geometry optimizations at the DFT level of theory using the experimental structures as starting points gave minimum‐energy structures that reproduce the experimental structures well. The conformers obtained this way, however, are only local minima on the potential energy surface (vide infra). The τ8 values from the optimized structure of MoF6(NC5H5)2 are 1.25 and 0 (B3LYP) for the opposing faces, which are in excellent agreement with those from the experimental structure. While the calculated M−F bond lengths agree closely with the experimental bond lengths for both adducts, the B3LYP Mo−N bond lengths are overestimated, as has been previously reported for other transition metal adducts. [21] The MN15 calculations better reproduce the dative Mo−N bonds (Table 2). DFT calculations for the MoF6(NC5H5) adduct predict the Mo−F bond lengths of the capped square face to be 0.01 Å longer than the Mo−F bond lengths to the two distal fluorine atoms. The ranges of the experimental bond lengths overlap (distal fluorines: Mo−F(5,6) 1.859(4), 1.861(4) Å; square face: Mo−F(1,2,3,4) 1.836(5) to 1.861(4) Å), with the shortest Mo−F bond length being in the capped square face, suggesting that the solid‐state bond lengths are affected by packing in the crystal structure. In MoF6(NC5H5)2, the experimental Mo−F(1,2) bonds that are shared between the two capped square faces were found to be shorter than the other four Mo−F bonds, while DFT calculations predict the reverse trend (MN15) or essentially the same lengths (B3LYP). The Mo−N experimental bond lengths in the 1 : 2 adduct (2.331(4), 2.326(3) Å) are significantly larger than the one in the 1 : 1 adduct (2.260(5) Å), which is consistent with increased steric congestion for the octacoordinated adduct and is best reproduced by the MN15 computational results (MN15: 2.364 and 2.278 Å, respectively). The Mo−F bonds also lengthen with increasing coordination number; the bonds in free MoF6 (exptl: 1.8147(6)–1.8201(8) Å) [27] increase upon addition of one and two pyridine moieties (1.836(4)–1.861(4) Å and 1.875(2)–1.898(2) Å, respectively). Interestingly, the M−N (M=W or Mo) bond length (2.251(7) Å) in WF6(NC5H5) is indistinguishable of that in MoF6(NC5H5).
In addition to the conformations observed experimentally, several minima were found by DFT (B3LYP and MN15) calculations for the two MoF6 adducts (Figures 2 and 3). The lowest calculated energy structures differ from those observed in the crystal structures by rotation of the pyridine ligands along the Mo−N bonds. In the crystal structure of MoF6(NC5H5), pyridine adopts a conformation where the pyridine is eclipsed with the two opposing Mo−F bonds, as was observed for WF6(NC5H5) (conformer 1 in Figure 2). However, in the DFT‐predicted lowest‐energy structure, the pyridine is staggered with respect to the distal Mo−F bonds (conformer 2 in Figure 2), which is more stable by 4.94 (B3LYP) or 5.85 kJ/mol (MN15) in free energy. This parallels conformer 2 of WF6(NC5H5) being 4.23 kJ/mol lower (B3LYP) in energy than its conformation 1. [21] The crystal structure of MoF6(NC5H5)2 shows the pyridines both staggered towards the opposing MoF2 groups, resulting in a close to coplanar arrangement of the two pyridine rings. As expected from the steric interference of the two pyridine rings in the experimentally observed conformation, the two pyridine groups are perpendicular to each other in the calculated global energy minimum structure (conformer 3 in Figure 3), which is only 2.80 (B3LYP) or 1.19 kJ/mol (MN15) lower in free energy than conformer 1. The discrepancy between the experimental and predicted lowest‐energy conformer is likely due to crystal packing effects and the stacking of the pyridine rings, which can easily overwrite the small calculated energy difference. When using MN15, conformation 2 is predicted to be essentially isoenergetic with conformation 1 (0.28 kJ/mol higher in free energy), while conformation 2 is predicted to be slightly more stable than 1 when using B3LYP (1.46 kJ/mol lower in free energy). Conformer 4 is the only local‐minimum structure that is not capped trigonal prismatic while still having exclusively real frequencies, i. e., a trigonal dodecahedral structure, being 50.59 (B3LYP) or 54.27 kJ/mol (MN15) higher in free energy than conformer 1. These calculations indicate that the MoF6 in neutral adducts exhibit preferences for adopting capped trigonal prismatic structures over other 7‐ or 8‐coordinated geometries.
Figure 2.

DFT predicted gas‐phase geometries (local minima) of MoF6(NC5H5) at the B3LYP and MN15/aVTZ level of theory.
Figure 3.

DFT predicted gas‐phase geometries (local minima) of MoF6(NC5H5)2 at the B3LYP and MN15/aVTZ level of theory.
Vibrational Spectroscopy
Raman and infrared spectra were obtained for MoF6(NC5H5) and MoF6(NC5H5)2 in the solid state at room temperature (Figure 4). The experimental vibrational frequencies are correlated to the calculated vibrational frequencies for the optimized geometry corresponding to the experimentally observed structure, i. e., conformation 1, at the B3LYP or MN15/aVTZ level of theory. Selected experimental and calculated vibrational frequencies at the B3LYP level of theory are listed in Table 3 whereas complete lists are given in Tables S3 and S4. Both adducts have molecular C 2v point symmetry with only the A2 modes being infrared inactive and all modes being Raman active. Calculated frequencies tend to overestimate the experimental values as was also the case with the WF6(NC5H5) adduct and its derivatives. [21] The MN15 functional results in a more substantial overestimation of vibrational frequencies, particularly those of the MoF6 moiety. Therefore, the B3LYP‐calculated frequencies are used in the discussion.
Figure 4.

Raman spectra (lower trace) and IR spectra (upper trace) of MoF6(NC5H5) (top) and MoF6(NC5H5)2 (bottom).
Table 3.
Selected Raman active vibrational frequencies (in cm−1) of the MoF6 pyridine adducts.
|
Mode Description[a] |
MoF6(NC5H5) |
||
|---|---|---|---|
|
Raman |
infrared |
B3LYP |
|
|
vs(NC5), v(A1)[b] |
1023(40) |
1021[m] |
1040(32)[6] |
|
vas(MoF6), v(B2) |
682sh |
696(<1)[249] |
|
|
vs(MoF2), v(A1) |
674(100) |
669[s] |
691(46)[176] |
|
vas(MoF4), v(B1) |
641[vs] |
670(<0.1)[192] |
|
|
vs(MoF4)−vs(MoF2), v(A1) |
651(11) |
648[vs] |
665(15)[101] |
|
vas(MoF4), v(A2) |
566(3) |
570sh |
567(0)[0] |
|
vas(MoF2)+vas(MoF4), v(B2) |
530(4) |
517[m] |
537(3)[2] |
|
MoF6(NC5H5)2 |
|||
|
Raman |
infrared |
B3LYP |
|
|
vs(NC5), v(A1) |
1021(20) |
1034(21)[<1] |
|
|
vs(NC5), v(B2) |
1018[m] |
1034(1)[8] |
|
|
vas(MoF4), v(B1) |
635sh |
658(<1)[190] |
|
|
vs(MoF6), v(A1) |
640(100) |
653(107)[42] |
|
|
vas(MoF6), v(A1) |
606(35) |
601[vs] |
617(37)[269] |
|
vas(MoF4), v(B2) |
567[vs] |
617(<1)[51] |
|
|
vas(MoF4)−vas(MoF2), v(B1) |
536(3) |
535sh |
542(2)[<1] |
|
vas(MoF4), v(A2) |
487(1) |
499sh |
506(2)[0] |
The Raman and infrared spectra of MoF6(NC5H5) and MoF6(NC5H5)2 contain bands associated with the pyridine ligands, which are shifted with respect to those of free pyridine. The most diagnostic pyridine band is that of the ring breathing mode, which shifts to higher frequencies upon adduct formation (Raman: free pyridine 991, MoF6(NC5H5) 1023, MoF6(NC5H5)2 1021 cm−1). Any expected vibrational coupling for the pyridine modes in MoF6(NC5H5)2 is predicted to be very small, below the resolution of experimental Raman (2 cm−1) and infrared (4 cm−1) spectra.
The most intense Raman band for MoF6(NC5H5) appears at 674 cm−1 (corresponding infrared band: 669 cm−1), which mainly corresponds to the vs (MoF2) of the distal difluoro moiety (calcd. 691 cm−1). This symmetric Mo−F stretching frequency is red‐shifted with respect to vs (MoF6) of free MoF6 (742 cm−1), [30] reflecting the longer, more ionic Mo−F bonding upon adduct formation. Coordination of a second pyridine ligand results in a further low‐frequency shift to 640 cm−1 for vs (MoF6) of MoF6(NC5H5)2 (calcd. 653 cm−1). Intense infrared bands at 648 and 641 cm−1 for MoF6(NC5H5) (calcd. 670, 665 cm−1) and at 601 and 567 cm−1 for MoF6(NC5H5)2 (calcd. 617, 617 cm−1) are assigned to asymmetric stretching modes.
NMR Spectroscopy
The 19F NMR spectrum of MoF6(NC5H5) in CH2Cl2 shows a broad singlet resonance at 294.7 ppm (Δν1/2=160 Hz) at 22 °C that broadens to Δν1/2=240 Hz and shifts to 259.9 ppm at −80 °C, indicating rapid exchange of the two fluorine environments that are expected based on the capped trigonal prismatic structure observed by X‐ray crystallography. The 1 : 2 adduct, MoF6(NC5H5)2, also shows a singlet resonance in the 19F NMR spectrum at room temperature in CH2Cl2 at 251.7 ppm (Δν1/2=400 Hz), reflecting the exchange between the two expected fluorine environments. At −80 °C, the exchange can be slowed down to allow for the observation of a triplet at 228.9 ppm and a quintet at 158.0 ppm (2 J F–F=121 Hz) in a 2 : 1 ratio (Figure 5), representing the two different fluorine environments observed in the crystal structure. Variable‐temperature NMR spectroscopy shows that the triplet and quintet start to lose their multiplicity at −70 °C and coalesce between −40 °C and −30 °C (Figure S12a). Upon raising the temperature the relative integration of the two resonances deviates from the expected 2 : 1 ratio observed at −80 and −70 °C to ratios of 4 : 1.54 (−60 °C) and 4 : 1.3 (−50 °C), suggesting a secondary exchange process. While the determination of accurate activation enthalpy and entropy values are thwarted by the secondary exchange process, analysis of the low‐temperature signals yielded an approximate activation energy of 46 kJ mol−1 for the exchange of the two fluorine environments based on an Arrhenius plot (see Supporting Information). The significant high‐frequency shift of the room‐temperature resonance for MoF6(NC5H5)2 suggests a dissociative equilibrium between MoF6(NC5H5)2 and MoF6(NC5H5).
Figure 5.

19F NMR spectrum of MoF6(NC5H5)2 at −80 °C in CH2Cl2 showing a A2X4 spin system.
In the 1H and 13C NMR spectra of both adducts, single sets of resonances are observed for ortho, meta and para nuclei of the adducted pyridine. The fact that only a single set of three resonances is observed for the 1 : 2 adduct at room temperature, as well as low temperature, indicates rapid rotation of the pyridine rings along the Mo−N bonds, consistent with the small calculated energy difference between the conformers.
NBO Analysis
The natural population analysis (NPA) charges and Wiberg Bond Indices (WBIs) were calculated for MoF6(NC5H5) and MoF6(NC5H5)2, as well as MoF6 in the Oh ground state minimum and D 3h transition state (Table 4). The NPA charge on Mo is larger for MoF6 in the trigonal prismatic structure (+2.48) than in the octahedral ground‐state geometry (+2.40). As pyridine ligands coordinate to trigonal prismatic MoF6, the charge becomes less positive with values of +2.36 (1 : 1 adduct) and +2.28 (1 : 2 adduct), as a consequence of negative charge donation from the ligand. The charges on the F atoms become more negative upon coordination of pyridine, consistent with increasing ionic contributions to Mo−F bonding. The total charge transferred from pyridine to the molybdenum fluoride in MoF6(NC5H5) is 0.29 and in MoF6(NC5H5)2 it is 0.25 per pyridine. The WBI for Mo−N is almost half of that for Mo−F, indicating a strong dative bond with the weakest dative bond being found for MoF6(NC5H5)2 (0.33). For WF6(NC5H5), the WBI for the W−F bond is 0.72–0.76 and for the W−N bond is 0.36, which are both slightly lower than those in MoF6(NC5H5). [21]
Table 4.
NPA charges and WBIs of MoF6 and its pyridine adducts calculated at the B3LYP/aVTZ level of theory.
|
MoF6 (Oh) |
MoF6 (D3h) |
MoF6(NC5H5) |
MoF6(NC5H5)2 |
|
|---|---|---|---|---|
|
charges |
||||
|
Mo |
+2.40 |
+2.48 |
+2.36 |
+2.28 |
|
F |
−0.40 |
−0.41 |
−0.42(F5,6) −0.45(F1,2,3,4) |
−0.46(F3,4,5,6) −0.48(F1,2) |
|
N |
– |
– |
−0.43 |
−0.40 |
|
NC5H5 |
– |
– |
+0.29 |
+0.25 |
|
WBI |
||||
|
Mo−F |
0.84 |
0.83 |
0.76–0.80 |
0.71–0.74 |
|
Mo−N |
– |
– |
0.38 |
0.33 |
Molecular Electrostatic Potentials (MEPs)
The MEP isosurfaces of MoF6 in the Oh and D 3h structures, MoF6(NC5H5), and MoF6(NC5H5)2 are shown in Figure 6. For octahedral MoF6, the eight regions of high electrostatic potential (EP), i. e., σ‐holes, have values of 94 kJ mol−1. In comparison, trigonal prismatic MoF6 (D 3h ) has three regions of substantially higher EP (σ‐holes) with values of 256 kJ mol−1, localized in each of the square faces, which is consistent with the MoF6(NC5H5) n adducts preferring the capped trigonal prismatic geometry as the pyridine can coordinate more readily to the higher EP region. Once a pyridine coordinates to MoF6, the two remaining σ‐holes in the square faces of the monocapped trigonal prismatic structure are lowered to 95 kJ mol−1 in MoF6(NC5H5), which is similar to that of the σ‐holes on free octahedral MoF6 and affords the coordination of a second pyridine. In MoF6(NC5H5)2 the open coordination site has an electrostatic potential value of −13 kJ mol−1, which prevents a third pyridine from coordinating and rationalizes why there is no evidence for a 1 : 3 adduct even with excess pyridine. The size of the σ‐holes is paralleled by the NPA charges on Mo, with the Mo in MoF6 (D 3h ) having the most positive charge and σ‐hole and MoF6(NC5H5)2 having the least positive charge on Mo and the lowest EP on the σ‐hole.
Figure 6.

Molecular electrostatic potential surfaces (MEPs) calculated at the 0.001 e bohr−3 isosurfaces of MoF6 (Oh and D 3h ), MoF6(NC5H5), and MoF6(NC5H5)2. The arrows indicated areas of relatively high electrostatic potential values. Calculations were done at the B3LYP/aTVZ level of theory.
Molecular Orbitals and Electronic Transition
The molecular orbital representing the Mo−N bond in MoF6(NC5H5) is represented by the HOMO‐2, in which the lone pair on N interacts with the dz 2 orbital on Mo (Figure S18). For the MoF6(NC5H5)2 adduct, the Mo−N bonding orbitals are represented by the HOMO‐2 and HOMO‐5 orbitals (Figure S19) which transform as B2 and A1 in the C 2v point group.
Frontier orbital energies and corresponding absorption energies, calculated from the HOMO‐LUMO gap using regular DFT and the time‐dependent (TD) DFT methods, are listed in Table 5. Figure 7 shows the frontier molecular orbitals for MoF6 Oh , MoF6(NC5H5), MoF6(NC5H5)2. For MoF6(NC5H5), MoF6(NC5H5)2, the HOMOs are localized on the pyridines, whereas the LUMOs are mainly localized on the MoF6 moiety.
Table 5.
Calculated HOMO and LUMO energies, HOMO‐LUMO energy gaps (ΔEHOMO/LUMO), the HOMO‐LUMO transition using time‐dependent (TD) DFT, and corresponding wavelengths (λHOMO/LUMO and λTDDFT).
|
MoF6 (Oh) |
MoF6(NC5H5) |
MoF6(NC5H5)2 |
||
|---|---|---|---|---|
|
MO Energy[a] |
HOMO |
−12.46 |
−8.54 |
−7.90 |
|
(eV) |
LUMO |
−5.98 |
−4.63 |
−3.85 |
|
ΔEHOMO/LUMO |
DFT[a] |
6.48 |
3.90 |
4.05 |
|
(eV) |
TDDFT[b] (f)[c] |
– |
4.08 (0) |
4.22 (0) |
[a] Calculated at the B3LYP/aVTZ level of theory. [b] Calculated at the camB3LYP/def2‐TZVP. [c] Calculated oscillator strength, f, is given in parentheses.
Figure 7.

Frontier molecular orbitals (DFT/B3LYP) of MoF6 Oh (left), MoF6(NC5H5) (centre), MoF6(NC5H5)2 (right). Isosurface values are drawn at 0.02 e Å−3.
Going from MoF6(NC5H5) to MoF6(NC5H5)2, the HOMO and LUMO energies increase, resulting in similar HOMO‐LUMO gaps for the two adducts (3.90 and 4.05 eV), which are significantly smaller than that in MoF6 (6.48 eV). TDDFT calculations (using the B3LYP/aVTZ conformation 1 geometry) at the camB3LYP/def2‐TZVP level of theory (Tables S7 and S8) show that the HOMO‐LUMO (MO #55 to 56 for 1 : 1, 76 to 77 for 1 : 2) transitions are 4.08 (1 : 1 adduct) and 4.22 eV (1 : 2 adduct). Although these represent ligand‐to‐metal charge transfer (LMCT) transitions for both adducts, they are symmetry‐forbidden A2‐symmetric, with zero oscillator strength and, hence, cannot be causing the intense colour of the adducts. For MoF6(NC5H5), the two transitions with a significant oscillator strength (f=0.05) are the HOMO to LUMO+3 (B2, 5.56 eV, 223 nm), which is a pyridine‐based π‐π transition, and HOMO‐2 to LUMO+2 (A1, f=0.1, 6.33 eV, 196 nm), which is an excitation from a fully delocalized MO to a MoF6‐based MO (Figure S20). For MoF6(NC5H5)2, the two transitions with the largest oscillator strengths are HOMO‐2 to LUMO+1 (B2, 4.65 eV, 266 nm, f=0.23) and HOMO‐5 to LUMO+2 (B2, 5.72 eV, 217 nm, f=0.2), which are both excitations from fully delocalized MOs to MoF6‐based MOs (Figure S21). Calculations of spin‐orbit‐coupling corrected excited state energies are given in Tables S9 and S10 and show that such corrections need to be considered; an excited state of 297 nm with oscillator strength of 0.015 is most likely expected to be observable (vacuum‐UV spectroscopy) for MoF6(NC5H5), but overall the oscillator strengths remain low for these compounds. The largest oscillator strength (0.14) is predicted for a 197 nm transition. For MoF6(NC5H5)2, a notable excitation is at 334 nm with an oscillator strength of 0.015. The transition with the highest oscillator strength (0.22) occurs at 269 nm. The calculated electronic transitions close to the visible region of light for the adducts are in agreements with the yellow and orange color of the compounds, whereas MoF6 is colourless.
Conclusions
Despite the significant oxidizing strength of MoF6, the first neutral Lewis acid base adducts of MoF6 have been synthesized and structurally characterized with pyridine as the organic ligand, i. e., MoF6(NC5H5) and MoF6(NC5H5)2. These adducts are only the second type of neutral adducts of a transition metal hexafluoride other than those of WF6. Single‐crystal X‐ray diffraction revealed that both adducts adopt capped trigonal prismatic structures, mirroring the geometries of the tungsten analogues. The molecular electrostatic potential surfaces (DFT/B3LYP level of theory) provide a rationale for the preference of the capped trigonal prismatic structure, as the σ‐holes on a trigonal prismatic MoF6 are significantly larger than those of free octahedral MoF6. Addition of pyridine reduces the size of the remaining σ‐holes, with a negative potential on the uncapped square face of MoF6(NC5H5)2, preventing the coordination of a third pyridine. Raman spectroscopy showed red shifts of the Mo−F stretching frequencies upon addition of pyridine. At room temperature, MoF6(NC5H5) and MoF6(NC5H5)2 are fluxional at the NMR time scale with singlet resonances in the 19F NMR spectra. For the 1 : 2 adduct, triplet and quintet resonances could be observed at −80 °C for the two 19F environments of MoF6(NC5H5)2 allowing for the estimate of the energy barrier of exchange (46 kJ mol−1), whereas exchange was still rapid for MoF6(NC5H5) at that temperature. The isolation of these pyridine adducts suggests that further Lewis acid base chemistry of MoF6 is possible.
Experimental Section
Caution! Elemental F2 and MoF6 are highly toxic and corrosive, rapidly evolving HF upon exposure to moisture. Appropriate safety measures should be taken during their handling.
Materials and Apparatus
Reactors were made of heat‐sealed and flared 1/4 in. o.d. tetrafluoroethene‐hexafluoropropene copolymer (FEP) reactors attached to a stainless‐steel valve. The reactors were evacuated for at least 7 h and passivated with 100 % F2(g) for at least 8 h. The solvents were volatile enough to be distilled through a Pyrex vacuum line equipped with PTFE Chemglass stopcocks. The solvents dichloromethane, SO2, pyridine and acetonitrile were vacuum distilled on the glass vacuum line. In contrast, MoF6 was handled on a prepassivated nickel/316 stainless‐steel vacuum line equipped with 316 stainless‐steel valves (Autoclave Engineers). A drybox (Omni Lab, Vacuum Atmospheres) with dry nitrogen atmosphere was used for handling solids.
All solvents used (acetonitrile, dichloromethane, pyridine) were dried over 3 Å sieves except for SO2 which was dried over CaH2. All solvents were kept in glass bulbs with either J‐Young or Chemglass stopcocks.
Synthesis of MoF6(NC5H5) n (n=1, 2)
Excess MoF6 (0.130 g, 0.619 mmol) was distilled onto pyridine (0.038 g, 0.48 mmol) in CH2Cl2 (0.567 g) at −196 °C. The reaction was started at −85 °C and warmed up to room temperature to afford MoF6(NC5H5) as a bright orange powder with a 97 % yield (0.135 g, 0.467 mmol). 19F NMR (in CH2Cl2 at 22 °C) δ 294.7 ppm and (at −80 °C) 259.9 ppm {br, s, MoF6(NC5H5)}. Impurities of [MoF7]− (267.3 ppm, 1.9 mol %) and MoOF4(NC5H5) (142.3 ppm, 0.3 mol %) at 22 °C; and of MoF6 (280.5 ppm, 1 J MoF=49 Hz, 2.1 mol %), [MoF7]− (266.5 ppm, 0.2 mol %), unknown impurity (262.6 ppm, 4 % integration with respect to MoF6(NC5H5)) and MoOF4(NC5H5) (141.1 ppm, 1.7 mol %) at −80 °C, small amount of white precipitate at −80 °C. 1H NMR (in CH2Cl2 at 22 °C) 8.56 ppm {H1, d, 3 J HH=5.0 Hz}, 7.59 ppm {H3, t, 3 J HH=7.5 Hz}, 7.15 ppm {H2, t, 3 J HH=6.8 Hz}. 13C NMR (in CH2Cl2 at −30 °C) 144.3 ppm {C1, d, 1 J=188.1 Hz}, 142.4 ppm {C3, dt, 1 J CH=166.8 Hz, 2 J CH=6.2 Hz}, 126.2 ppm {C2, dt, 1 J CH=170.2 Hz, 2 J CH=5.4 Hz}.
The synthesis for MoF6(NC5H5)2 used pyridine (0.076 g, 0.96 mmol) to MoF6 (0.090 g, 0.43 mmol) in CH2Cl2 (0.289 g) afforded MoF6(NC5H5)2 (0.164 g, 0.44 mmol, 100 % yield) as a yellow‐orange powder which slowly decomposed over a few days at room temperature within the drybox. 19F NMR (in CH2Cl2 at 22 °C) δ 251.7 ppm {br, s, MoF6(NC5H5)2} and (at −80 °C) 228.9 ppm {t, 2 J FF=121 Hz, MoF6(NC5H5)2} and 158.0 ppm {q, 2 J FF=121 Hz, MoF6(NC5H5)2}. Impurities of MoOF4(NC5H5) (142.5 ppm, 0.5 mol %) and [MoOF5]− (Feq, 130.4 ppm, d, 2 J FF=48 Hz, 1.6 mol %) at 22 °C; and of [MoF7]− (269.0 ppm, 1.5 mol %) and [MoOF5]− (Feq, 129.2 ppm, d, 2 J FF=51 Hz, 0.8 mol %) at −80 °C, small amount of white precipitate at −80 °C. 1H NMR (in CH2Cl2 at −20 °C) 7.98 ppm {H1, d, 3 J HH=4.9 Hz} 6.97 ppm {H2, t, 3 J HH=7.05 Hz} 6.55 ppm {H3, t, 3 J HH=5.7 Hz}. 13C NMR (in CH2Cl2 at −30 °C) 143.76 ppm {C1, ddd, 1 J CH=184.6 Hz, 2 J CH=12.3 Hz} 139.7 ppm {C3, dt, 1 J CH=166.5 Hz, 2 J CH=5.7 Hz} 125.0 ppm {C2, dt, 1 J CH=168.3 Hz, 2 J CH=5.9 Hz}.
X‐ray Crystallography. Crystal Growth and Mounting
0.011 g of MoF6(NC5H5) was dissolved in pyridine to afford a yellow solution consisting of dissolved MoF6(NC5H5)2 that, when kept at −35 °C over two days, afforded a cluster of orange crystals. The solvents were removed under dynamic vacuum.
Bright orange crystals of MoF6(NC5H5) were grown by reacting 0.038 g (0.48 mmol) of pyridine with 0.128 g (0.610 mmol) of MoF6 in 0.372 g of dichloromethane to afford a bright orange solution which was then cooled down from −3 to −35 °C. Note that MoF6(NC5H5) dissolves in dichloromethane when synthesized but redissolving the solid in dichloromethane only shows partial solubility.
An aluminium trough was cooled between −50 and −80 °C with N2(g) that was generated from a Dewar filled with N2(l) by passing a flow of nitrogen gas through the liquid. The crystals were picked with a nylon cryo‐loop previously dipped in perfluorinated polyether oil (Fomblin Z‐15).
Data Collection and Reduction
The Rigaku SuperNova diffractometer equipped with a Dectris Pilatus 3R 200K−A hybrid‐pixel‐array detector, a four‐circle κ goniometer, an Oxford Cryostream 800, and sealed MoKα and CuKα X‐ray sources where the MoKα source (λ=0.71073 Å) at −173 °C was used. CrysAlisPro was used to process the data. [31]
The crystals were centered on a Rigaku SuperNova diffractometer equipped with a Dectris Pilatus 3R 200 K−A hybrid‐pixel‐array detector, a four‐circle κ goniometer, an Oxford Cryostream 800, and sealed Mo Kα and Cu Kα X‐ray sources. Data were collected using the Mo Kα source (λ=0.71073 Å) at −173 °C. Crystals were screened for quality before a preexperiment was run to determine the unit cell, and a data collection strategy was calculated based on the determined unit cell and intensity of the preliminary data. The data were processed using CrysAlisPro, which applied the necessary Lorentz and polarization corrections to the integrated data and scaled the data. A numerical (Gaussian‐grid) absorption correction was generated based on the indexed faces of the crystal.
Structure Solution and Refinement
The intrinsic phasing method (ShelXT) and least‐squares refinement (ShelXL) were used for the refinement. Structure and solution were done with Olex2 (version 1.5).[ 32 , 33 , 34 ] Non‐H atoms were refined anisotropically, and recommended weights for the atoms were determined before H atoms were introduced using a riding model (HFIX). The maximum and minimum electron densities in the Fourier difference maps were located close to the Mo atoms.
Deposition Numbers 2370068 and 2370069 contain the supplementary crystallographic data for this paper. These data are provided free of charge by the Cambridge Crystallographic Data Centre.
Raman Spectroscopy
Raman spectra were recorded in the solid state at 21 °C in flame‐sealed melting point capillaries with a Bruker RFS‐100 Raman spectrometer outfitted with a quartz beam‐splitter and liquid‐N2‐cooled germanium detector. The 1064 nm line of a Nd:YAG laser was used for excitation of the sample, and backscattered (180°) radiation was sampled. The usable Stokes range of the collected data was 85–3500 cm−1. The spectral resolution is 2 cm−1 and a laser power of 100 mW was used.
Infrared Spectroscopy
Infrared spectra were recorded on solid samples at 21 °C inside a dry box using a Bruker Alpha‐P FT‐IR spectrometer outfitted with a KBr beam splitter, a RT‐DLATGS detector, and a single‐bounce ATR module. A spectral resolution of 4 cm−1 was used.
NMR Spectroscopy
The NMR samples were in a heat sealed (under dynamic vacuum) 4 mm o.d. FEP tube that is then inserted into a 5 mm o.d. glass NMR tube and run on a Bruker Avance II 300 MHz spectrometer with a 5 mm broadband probe. All spectra were run unlocked in dichloromethane with an external reference (CFCl3 for 19F and Si(CH3)4 for 1H and 13C at 21 °C).
Computational Methods
Geometry optimizations and the corresponding vibrational frequencies were performed in the gas phase at the DFT/B3LYP (or MN15) level of theory using Gaussian 16. The basis set used is the aug‐cc‐pVTZ on N, O, F, and Mo with a Stuttgart pseudopotential also included on Mo, and the cc‐pVTZ basis set on H and C (aVTZ). [35] Subsequent calculations were done with the optimized geometry that matched the experimental conformations. Natural bond order (NBO) analysis was done using NBO (version 6.0). [36] Gaussview and Avogadro were used to visualize vibrational modes and molecular orbitals.[ 37 , 38 ] Time‐dependent DFT (TDDFT) was done at the camB3LYP/def2‐TZVP level of theory using the Gaussian 16 program package. [39] The TDDFT calculations with spin‐orbit coupling were done at the camB3LYP/ZORA‐def2‐TZVP level of theory using ORCA 5.0.[ 40 , 41 ]
Supporting Information Summary
The data that support the findings of this study are available in the supplementary material of this article.
Conflict of Interests
The authors declare no conflict of interest.
1.
Supporting information
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Supporting Information
Acknowledgments
We thank the Natural Sciences and Engineering Research Council of Canada for awarding Discovery grants to P.H., S.D.W. and M.G. In addition, we would like to thank the University of Lethbridge for awarding a Graduate Research Award to M.D.v.H. The computational studies were performed using equipment funded through the Canada Foundation for Innovation, as well as resources made available through the Digital Research Alliance of Canada (the Alliance).
van Hoeve M. D., Bell R., O'Donnell F., Hazendonk P., Wetmore S. D., Gerken M., Chem. Eur. J. 2024, 30, e202402749. 10.1002/chem.202402749
Data Availability Statement
The data that support the findings of this study are available in the supplementary material of this article.
References
- 1. Seppelt K., Chem. Rev. 2015, 115, 1296–1306. [DOI] [PubMed] [Google Scholar]
- 2.M. J. Molski, K. Seppelt, J. Chem. Soc., Dalton Trans. 2009, 3379–3383. [DOI] [PubMed]
- 3. Lin J., Yang Q., Li X., Zhang X., Li F., Yang G., Phys. Chem. Chem. Phys. 2022, 24, 1736–1742. [DOI] [PubMed] [Google Scholar]
- 4. Christe K. O., Dixon D. A., McLemore D., Wilson W. W., Sheehy J. A., Boatz J. A., J. Fluor. Chem. 2000, 101, 151–153. [Google Scholar]
- 5. George P. M., Beauchamp J. L., Chem. Phys. 1979, 36, 345–351. [Google Scholar]
- 6. Bartlett N., Angew. Chem. Int. Ed. 1968, 7, 433–439. [Google Scholar]
- 7. Quiñones G. S., Hägele G., Seppelt K., Chem. Eur. J. 2004, 10, 4755–4762. [DOI] [PubMed] [Google Scholar]
- 8. Bailar J. C., J. Inorg. Nucl. Chem. 1958, 8, 165–175. [Google Scholar]
- 9. Seppelt K., Acc. Chem. Res. 2003, 36, 147–153. [DOI] [PubMed] [Google Scholar]
- 10. Bresciani G., Marchetti F., Pampaloni G., Coord. Chem. Rev 2023, 496, 215399. [Google Scholar]
- 11. Walker D. W., Winfield J. M., J. Fluorine Chem. 1972, 1, 376–378. [Google Scholar]
- 12. Haiges R., Boatz J. A., Bau R., Schneider S., Schroer T., Yousufuddin M., Christe K. O., Angew. Chem. Int. Ed. 2005, 44, 1860–1865. [DOI] [PubMed] [Google Scholar]
- 13. Tamadon F., Seppelt K., Angew. Chem. Int. Ed. 2013, 52, 767–769. [DOI] [PubMed] [Google Scholar]
- 14. Levason W., Monzittu F. M., Reid G., Coord. Chem. Rev. 2019, 391, 90–130. [Google Scholar]
- 15. Bykowski J., Turnbull D., Hahn N., Boeré R. T., Wetmore S. D., Gerken M., Inorg. Chem. 2021, 60, 15695–15711. [DOI] [PubMed] [Google Scholar]
- 16. Tebbe F. N., Muetterties E. L., Inorg. Chem. 1968, 7, 172–174. [Google Scholar]
- 17. El-Kurdi S., Al-Terkawi A.-A., Schmidt B., Dimitrov A., Seppelt K., Chem. Eur. J. 2010, 16, 595–599. [DOI] [PubMed] [Google Scholar]
- 18. Levason W., Monzittu F. M., Reid G., Zhang W., Chem. Commun. 2018, 54, 11681–11684. [DOI] [PubMed] [Google Scholar]
- 19. Benjamin S. L., Levason W., Reid G., Chem. Soc. Rev. 2013, 42, 1460–1499. [DOI] [PubMed] [Google Scholar]
- 20. Arnaudet L., Bougon R., Buu B., Lance M., Nierlich M., Thuéry P., Vigner J., J. Fluorine Chem. 1995, 71, 123–129. [Google Scholar]
- 21. Turnbull D., Kostiuk N., Wetmore S. D., Gerken M., J. Fluorine Chem. 2018, 215, 1–9. [Google Scholar]
- 22. Turnbull D., Wetmore S. D., Gerken M., Angew. Chem. 2019, 131, 13169–13172. [DOI] [PubMed] [Google Scholar]
- 23. Arnaudet L., Bougon R., Ban B., Lance M., Navaza A., Nierlich M., Vigner J., J. Fluorine Chem. 1994, 67, 17–25. [Google Scholar]
- 24.M. Mercer, T. J. Ouellette, C. T. Ratcliffe, D. W. A. Sharp, J. Chem. Soc. A Inorg. Phys. Theor. Chem. 1969, 2532–2534.
- 25. Fuggle J. C., Sharp D. W. A., Winfield J. M., J Fluorine Chem. 1972, 1, 427–431. [Google Scholar]
- 26. Turnbull D., Wetmore S. D., Gerken M., Chem. Eur. J. 2021, 27, 11335–11343. [DOI] [PubMed] [Google Scholar]
- 27. Drews T., Supeł J., Hagenbach A., Seppelt K., Inorg. Chem. 2006, 45, 3782–3788. [DOI] [PubMed] [Google Scholar]
- 28. Nieboer J., Yu X., Chaudhary P., Mercier H. P. A., Gerken M., Z Anorg. Allg. Chem. 2012, 638, 520–525. [Google Scholar]
- 29. Kumar M., Srivastava M., Yadav R. A., Spectrochim. Acta A Mol. Biomol. Spectrosc. 2013, 111, 242–251. [DOI] [PubMed] [Google Scholar]
- 30. Claassen H. H., Goodman G. L., Holloway J. H., Selig H., J. Chem. Phys. 1970, 53, 341–348. [Google Scholar]
- 31.CrysAlisPro, Version 1.171.41.122a. Rigaku Oxford Diffraction, Rigaku Corporation, Oxford, UK 2021.
- 32. Sheldrick G. M., Acta Crystallogr. A Found. Adv. 2015, 71, 3–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Sheldrick G. M., Acta Crystallogr. C Struc. Chem. 2015, 71, 3–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Dolomanov O. V., Bourhis L. J., Gildea R. J., Howard J. A. K., Puschmann H., J. Appl. Crystallogr. 2009, 42, 339–341. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Peterson K. A., Figgen D., Dolg M., Stoll H., J. Chem. Phys. 2007, 126, 124111. [DOI] [PubMed] [Google Scholar]
- 36. Glendening E. D., Landis C. R., Weinhold F., J. Comput. Chem. 2013, 34, 1429–1437. [DOI] [PubMed] [Google Scholar]
- 37.GaussView, version 6.0, R. Dennington, T. A. Keith, J. M. Millam, Semi-chem Inc., Shawnee Mission, KS, USA 2019.
- 38. Hanwell M. D., Curtis D. E., Lonie D. C., Vandermeersch T., Zurek E., Hutchison G. R., J. Cheminform. 2012, 4, 1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman, D. J. Fox, Gaussian 16, Revision C.01, Gaussian, Inc., Wallingford, CT, USA 2016.
- 40. Neese F., Wennmohs F., Becker U., Riplinger C., J. Chem. Phys. 2020, 152, 224108. [DOI] [PubMed] [Google Scholar]
- 41. Pantazis D. A., Chen X.-Y., Landis C. R., Neese F., J. Chem. Theory Comput. 2008, 4, 908–919. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Supporting Information
Data Availability Statement
The data that support the findings of this study are available in the supplementary material of this article.


