Dalton Transactions View Article Online
PAPER
View Journal | View Issue
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
NO2 bond cleavage by MoL3 complexes† Cite this: Dalton Trans., 2014, 43, 1620
Miranda F. Shaw,a Narges Mahdizadeh Ghohe,b Alireza Ariafard,*a,b Nigel J. Brookes,a Robert Strangerc and Brian F. Yates*a The cleavage of one N–O bond in NO2 by two equivalents of Mo(NRAr)3 has been shown to occur to form molybdenum oxide and nitrosyl complexes. The mechanism and electronic rearrangement of this reaction was investigated using density functional theory, using both a model Mo(NH2)3 system and the full [N(tBu)(3,5-dimethylphenyl)] experimental ligand. For the model ligand, several possible modes of coordination for the resulting complex were observed, along with isomerisation and bond breaking pathways. The lowest barrier for direct bond cleavage was found to be via the singlet η2-N,O complex (7 kJ mol−1). Formation of a bimetallic species was also possible, giving an overall decrease in energy and a lower barrier for reaction (3 kJ mol−1). Results for the full ligand showed similar trends in energies for both isomerisation between the different isomers, and for the mononuclear bond cleavage. The lowest calculated barrier for cleavage was only 21 kJ mol−1 via the triplet η1-O isomer, with a strong thermodynamic driving force to the final products of the doublet metal oxide and a molecule of NO. Formation of the full ligand dinuclear complex was not accompanied by an equivalent decrease in energy seen with the model
Received 16th September 2013, Accepted 1st November 2013
ligand. Direct bond cleavage via an η1-O complex is thus the likely mechanism for the experimental reac-
DOI: 10.1039/c3dt52554f
tion that occurs at ambient temperature and pressure. Unlike the other known reactions between MoL3 complexes and small molecules, the second equivalent of the metal does not appear to be necessary, but
www.rsc.org/dalton
instead irreversibly binds to the released nitric oxide.
1.
Introduction
The reactivity of small, multiply bonded molecules has generated considerable interest in recent years, given the increasing relevance of the atmospheric concentration of the oxides of carbon, nitrogen and sulfur. Nitrogen dioxide is one of the most abundant pollutants in this context, primarily released during the combustion of fossil fuels. As well as its acute health effects, NO2 contributes to the formation of acid rain and toxic, tropospheric ozone.1 Several transition metal complexes have been successfully shown to coordinate and cleave these small molecules, with one of the more promising and active reagents being the tri-amide molybdenum complex MoL3.2–8 Following the discovery of its ability to bind and cleave dinitrogen, this complex has since been successfully used in analogous reactions with a range of other molecules
a School of Chemistry, University of Tasmania, Private Bag 75, Hobart, TAS 7001, Australia. E-mail: [email protected]; Fax: +61 3 6226 2858; Tel: +61 3 6226 2167 b School of Chemistry, Central Tehran Branch, Islamic Azad University, Gharb, Tehran, Iran. E-mail: [email protected] c Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia † Electronic supplementary information (ESI) available. See DOI: 10.1039/c3dt52554f
1620 | Dalton Trans., 2014, 43, 1620–1629
including SO2, CO, CS2, and N2O.9 The reaction with nitrogen dioxide was shown to be especially facile, with complete deoxygenation occurring rapidly under mild conditions.9 The structure and electronic state of the active complex and the mechanism of bond cleavage for several of these known observed reactions has been investigated by theoretical studies,10–27 however these do not provide a detailed understanding of the reaction between MoL3 and NO2. We therefore aimed to further examine and rationalize this reaction, which results in the molybdenum oxide and nitrosyl complexes shown in Fig. 1. The complexities within this reaction arise from the bent geometry of NO2, which allows coordination in a range of conformeric orientations, and the electronic arrangements of both
Fig. 1 Deoxygenation of NO2 by molybdenum complexes. Reaction occurs between ethereal MoL3 with 0.5 equivalents of NO2 at 25 degrees.9
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
Dalton Transactions
the starting materials and products. NO2 is paramagnetic, with the unpaired electron in a sigma orbital, and this must bind to the monomeric, trigonal planar, triply degenerate d3 quartet molybdenum metal centre of MoL3. Access to the binding site is also restricted by a protecting ligand environment. The end products of the reaction are the doublet O-MoL3 and singlet ON-MoL3. Clearly the electronic rearrangement necessary to form these products is non-trivial and requires further examination. We therefore aimed to establish the reaction mechanism that occurs when NO2 reacts with MoL3, with particular consideration to the range of conformers likely during initial coordination, the electronic rearrangement that is necessary for the homolytic cleavage of the N–O bond, and the potential for dinuclear intermediate formation. Experimentally, the high reactivity at the Mo centre is promoted by the bulky ligands forcing the metal to remain coordinatively unsaturated, by preventing L3Mo–MoL3 dimerisation.13 To examine the pure electronic interaction between MoL3 and NO2 a simplified model system in which the NRAr ligands were replaced with NH2 was employed to map a comprehensive potential energy surface. Based on these results, selected structures were then optimised with the full [N(tBu)(3,5-dimethylphenyl)] ligand to provide further information about the effects of steric crowding around the active site.
2. Computational details In accordance with previous studies, the majority of calculations, including all single point energy calculations and geometry optimisations for the model NH2 system, were performed using the def2-QZVP28 basis set for molybdenum, incorporating the SDD29 effective core potential, and 6-311+G(2d,p)30 for all other atoms. This basis set combination is referred to as BS2. A less computationally demanding basis set combination, BS1, was chosen for the full ligand geometry optimisations; SDD on the metal and 6-31G(d)31 on the remaining atoms. The M06 functional32–34 was used for all calculations, as this has been shown to generally perform well for organometallic and inorganometallic chemistry.35,36 The model system was also evaluated with the B3LYP functional37–40 (again with BS2) as an additional check for functional independence (see ESI†). All final geometries were optimised without constraint and with unrestricted wavefunctions in the case of radical systems. Frequency calculations were performed at the same level of theory as the geometry optimisations, with structures characterised as minima or transition states based on the observed number of imaginary frequencies. All transition structures contained exactly one imaginary frequency and were linked to reactant, products or intermediates using intrinsic reaction coordinate (IRC) calculations.41 Except where noted, all energies in this publication include Gibbs free energy corrections (reported as ΔG). All minimum energy crossing point (MECP) calculations were carried out using the code provided by Dr Jeremy Harvey,42,43 at the M06/BS2 level of theory. MECP calculations are reported
This journal is © The Royal Society of Chemistry 2014
Paper
uncorrected due to the absence of stationary points at crossing points. Solvation effects were evaluated using the IEFPCM model44,45 for all single point calculations, using diethylether as the solvent in accordance with related experiments. No significant differences were found between optimised geometries in the gas phase or with solvent (see ESI†); gas phase geometries were used for all calculations discussed. All calculations were carried out with the Gaussian0946 set of programs.
3. Results and discussion 3.1.
Reaction between NO2 and Mo(NH2)3
The bond cleavage reaction can proceed via two general pathways, involving either one or two of the starting MoL3 complexes. The potential energy profiles for these mononuclear and dinuclear pathways are shown in full in Fig. 2, 6 and 8. The structures shown are labelled in the format N_X or TSN_X in the case of transition structures, where X indicates the multiplicity of the complex (S = singlet, D = doublet, T = triplet and Q = quartet). For the full N(R)(Ar) ligand, the subscript F is added (i.e. N_XF). 3.2.
Mononuclear pathway
The initial step of the reaction was coordination of an NO2 molecule to the quartet complex 1_Q; the quartet has been previously demonstrated to be lower in energy than the doublet alternative.26 The single unpaired electron in free NO2 is shared between all three atoms, as shown by the calculated
Fig. 2 Relative Gibbs free energies (kJ mol−1) for the mononuclear, triplet surface.
Dalton Trans., 2014, 43, 1620–1629 | 1621
View Article Online
Paper
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
Fig. 3
Dalton Transactions
Isomers of NO2–Mo(NH2)3.
Table 1 Isomers of NO2–Mo(NH2)3. Gibbs free energies (G) and electronic energies (E), relative to Mo(NH2)3 and unbound NO2 (kJ mol−1), calculated spin densities and selected bond lengths (Å)
NO2 + 1_Q 2_T 2_S 3_T 3_S 4_T 4_S 5_T 5_S 6_S
G
E
O1–N
O2–N
O1–Mo
N–Mo
O2–Mo
SMo
SNO2
0.0 −166.8 −196.3 −166.5 −210.7 −180.1 −167.2 −206.7 −211.6 −252.2
0.0 −229.8 −269.7 −224.0 −280.9 −236.8 −299.4 −265.4 −283.5 −325.1
1.194 1.235 1.275 1.341 1.362 1.351 1.409 1.291 1.298 1.380
1.194 1.215 1.204 1.184 1.193 1.183 1.167 1.258 1.253 1.380
— 2.712 2.321 2.046 2.024 2.020 1.924 2.075 2.052 1.951
— 2.110 1.945 2.117 1.946 2.959 2.922 2.599 2.597 1.959
— 3.047 3.012 3.294 3.114 4.015 3.941 2.187 2.155 1.951
2.968 1.843 — 1.579 — 1.861 — 1.701 — —
1.000 0.051 — 0.239 — 0.054 — 0.165 — —
spin densities: 0.492 on the nitrogen, and 0.254 on each of the oxygens, allowing all three atoms to be potential donors. The initial coordination occurs by pairing this single electron with one of those in molybdenum’s half-filled orbitals, leaving two remaining unpaired electrons on the metal. This can alternatively be interpreted as a single electron transfer from the metal to NO2, forming an intermediate nitrite anion which rapidly coordinates to the now cationic d2 molybdenum centre. In either case the end result is the formation of one of four possible linkage isomers, all with a triplet electronic arrangement (Fig. 3). A fifth η3-O,N,O complex 6_T was unable to be located, suggesting further donation from the NO2 requires an additional empty orbital on the metal. Relative energies and selected bond lengths are summarised in Table 1. Structures 2_T, 3_T, 4_T and 5_T are able to interchange with one another via the transition structures shown in Fig. 2, with barriers in the range of 10–50 kJ mol−1. In addition, each of these four complexes is able to undergo spin crossing to its singlet counterpart via a minimum energy crossing point (MECP). The energies of these MECPs, were 9, 6, 25 and 11 kJ mol−1 higher than the electronic energies of 2_T, 3_T, 4_T and 5_T respectively. The low barriers suggest that spin crossing is thus competitive with both isomerisation and bond breaking reactions. Pairing these electrons to form a singlet complex leaves an unoccupied orbital on the metal, allowing formation of the η3-O,N,O complex 6_S. Although 6_S was found to be the lowest energy isomer, the barrier to its formation via 5_S is relatively high at 55 kJ mol−1 (Fig. 6). The stabilising effect of this empty molybdenum orbital on the singlet surface can also be seen in the η2-N,O complex 3_S. The ability to donate the nitrogen
1622 | Dalton Trans., 2014, 43, 1620–1629
lone pair into the empty orbital results in a decrease in both the Mo–N bond length and the relative energy of 3_S compared to 3_T. Isomerisation to 3_S from both 2_S and 4_S was not found to have any discernable barrier. As rotation of the small NH2 ligands was unrestrained during optimisation, they were free to adopt the most favourable orientation in each complex. The three coordinate Mo(NH2)3 starting material has D3h symmetry, with each of the N–H bonds perpendicular to the MoL3 plane. Consistent with studies of related N2 and CS2 complexes, following coordination of the NO2 either one or two of the NH2 ligands rotates approximately 70°. This ligand rotation has been observed both computationally19,47 and in the crystal structures9 for several different NR2 ligands, suggesting that the driving force for this rotation is due to an electronic rather than steric interaction. NBO analyses have shown that in the trigonal MoL3 complex, the orientation of the amide ligands maximises the p ↔ d π interaction with the dxy and dx2−y2 orbitals, while minimising overlap with the partially filled dxz, dyz and dz2 orbitals. Once an additional ligand (such as NO2, N2, etc.) has coordinated, electron density is withdrawn from these partially filled orbitals. This allows one of the amide ligands to rotate and donate to the dxz, dyz and dz2 orbitals, thus promoting charge transfer to the fourth ligand, activating it towards bond cleavage (see Fig. 4). This rotation also brings the nitrogen 2p orbital of the rotated amide in close proximity to the NO2 donor atom and the Mo-donor atom σ* orbitals. Therefore whether one, two, or all three amides have rotated, as well as the exact NO2–Mo–N–H dihedral angles, is therefore different for each of the isomers and singlet–triplet pairs, as each isomer optimises the orbital overlap with the different donor atoms and Mo-donor atom σ* orbitals.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
Dalton Transactions
Fig. 4
Paper
Metal–nitrogen orbital interactions.
Along with isomerisation and spin-crossing reactions, the majority of the complexes described above are also subject to the non-reversible, N–O bond cleavage. Considering first the triplet surface, bond breaking is most active from the relatively unstable η2-N,O complex 3_T, occurring with a barrier of only 8 kJ mol−1. The N–O bond is already weakened in 3_T, as indicated by the increase in N–O bond length from 1.194 to 1.384 Å. The transition structure (TS6_T) is also stabilised by the weak Mo–N interaction. Elongation (to 1.351 Å) and resulting activation of the N–O bond is also seen in 4_T, but without the Mo–N interaction, the barrier for bond cleavage is increased to 35 kJ mol−1. This is higher again (at 51 kJ mol−1) for the more symmetrical η2-O,O complex 5_T, in which the two N–O bonds remain relatively similar (1.291 and 1.258 Å). Each of these pathways results in the molybdenum oxide 7_D and a molecule of nitric oxide. NO is itself a potential ligand, and can coordinate to another molecule of 1_Q, pairing their respective electrons to give the triplet 10_T (Fig. 5). This complex then passes through a very low minimum energy crossing point (
PAPER
View Journal | View Issue
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
NO2 bond cleavage by MoL3 complexes† Cite this: Dalton Trans., 2014, 43, 1620
Miranda F. Shaw,a Narges Mahdizadeh Ghohe,b Alireza Ariafard,*a,b Nigel J. Brookes,a Robert Strangerc and Brian F. Yates*a The cleavage of one N–O bond in NO2 by two equivalents of Mo(NRAr)3 has been shown to occur to form molybdenum oxide and nitrosyl complexes. The mechanism and electronic rearrangement of this reaction was investigated using density functional theory, using both a model Mo(NH2)3 system and the full [N(tBu)(3,5-dimethylphenyl)] experimental ligand. For the model ligand, several possible modes of coordination for the resulting complex were observed, along with isomerisation and bond breaking pathways. The lowest barrier for direct bond cleavage was found to be via the singlet η2-N,O complex (7 kJ mol−1). Formation of a bimetallic species was also possible, giving an overall decrease in energy and a lower barrier for reaction (3 kJ mol−1). Results for the full ligand showed similar trends in energies for both isomerisation between the different isomers, and for the mononuclear bond cleavage. The lowest calculated barrier for cleavage was only 21 kJ mol−1 via the triplet η1-O isomer, with a strong thermodynamic driving force to the final products of the doublet metal oxide and a molecule of NO. Formation of the full ligand dinuclear complex was not accompanied by an equivalent decrease in energy seen with the model
Received 16th September 2013, Accepted 1st November 2013
ligand. Direct bond cleavage via an η1-O complex is thus the likely mechanism for the experimental reac-
DOI: 10.1039/c3dt52554f
tion that occurs at ambient temperature and pressure. Unlike the other known reactions between MoL3 complexes and small molecules, the second equivalent of the metal does not appear to be necessary, but
www.rsc.org/dalton
instead irreversibly binds to the released nitric oxide.
1.
Introduction
The reactivity of small, multiply bonded molecules has generated considerable interest in recent years, given the increasing relevance of the atmospheric concentration of the oxides of carbon, nitrogen and sulfur. Nitrogen dioxide is one of the most abundant pollutants in this context, primarily released during the combustion of fossil fuels. As well as its acute health effects, NO2 contributes to the formation of acid rain and toxic, tropospheric ozone.1 Several transition metal complexes have been successfully shown to coordinate and cleave these small molecules, with one of the more promising and active reagents being the tri-amide molybdenum complex MoL3.2–8 Following the discovery of its ability to bind and cleave dinitrogen, this complex has since been successfully used in analogous reactions with a range of other molecules
a School of Chemistry, University of Tasmania, Private Bag 75, Hobart, TAS 7001, Australia. E-mail: [email protected]; Fax: +61 3 6226 2858; Tel: +61 3 6226 2167 b School of Chemistry, Central Tehran Branch, Islamic Azad University, Gharb, Tehran, Iran. E-mail: [email protected] c Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia † Electronic supplementary information (ESI) available. See DOI: 10.1039/c3dt52554f
1620 | Dalton Trans., 2014, 43, 1620–1629
including SO2, CO, CS2, and N2O.9 The reaction with nitrogen dioxide was shown to be especially facile, with complete deoxygenation occurring rapidly under mild conditions.9 The structure and electronic state of the active complex and the mechanism of bond cleavage for several of these known observed reactions has been investigated by theoretical studies,10–27 however these do not provide a detailed understanding of the reaction between MoL3 and NO2. We therefore aimed to further examine and rationalize this reaction, which results in the molybdenum oxide and nitrosyl complexes shown in Fig. 1. The complexities within this reaction arise from the bent geometry of NO2, which allows coordination in a range of conformeric orientations, and the electronic arrangements of both
Fig. 1 Deoxygenation of NO2 by molybdenum complexes. Reaction occurs between ethereal MoL3 with 0.5 equivalents of NO2 at 25 degrees.9
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
Dalton Transactions
the starting materials and products. NO2 is paramagnetic, with the unpaired electron in a sigma orbital, and this must bind to the monomeric, trigonal planar, triply degenerate d3 quartet molybdenum metal centre of MoL3. Access to the binding site is also restricted by a protecting ligand environment. The end products of the reaction are the doublet O-MoL3 and singlet ON-MoL3. Clearly the electronic rearrangement necessary to form these products is non-trivial and requires further examination. We therefore aimed to establish the reaction mechanism that occurs when NO2 reacts with MoL3, with particular consideration to the range of conformers likely during initial coordination, the electronic rearrangement that is necessary for the homolytic cleavage of the N–O bond, and the potential for dinuclear intermediate formation. Experimentally, the high reactivity at the Mo centre is promoted by the bulky ligands forcing the metal to remain coordinatively unsaturated, by preventing L3Mo–MoL3 dimerisation.13 To examine the pure electronic interaction between MoL3 and NO2 a simplified model system in which the NRAr ligands were replaced with NH2 was employed to map a comprehensive potential energy surface. Based on these results, selected structures were then optimised with the full [N(tBu)(3,5-dimethylphenyl)] ligand to provide further information about the effects of steric crowding around the active site.
2. Computational details In accordance with previous studies, the majority of calculations, including all single point energy calculations and geometry optimisations for the model NH2 system, were performed using the def2-QZVP28 basis set for molybdenum, incorporating the SDD29 effective core potential, and 6-311+G(2d,p)30 for all other atoms. This basis set combination is referred to as BS2. A less computationally demanding basis set combination, BS1, was chosen for the full ligand geometry optimisations; SDD on the metal and 6-31G(d)31 on the remaining atoms. The M06 functional32–34 was used for all calculations, as this has been shown to generally perform well for organometallic and inorganometallic chemistry.35,36 The model system was also evaluated with the B3LYP functional37–40 (again with BS2) as an additional check for functional independence (see ESI†). All final geometries were optimised without constraint and with unrestricted wavefunctions in the case of radical systems. Frequency calculations were performed at the same level of theory as the geometry optimisations, with structures characterised as minima or transition states based on the observed number of imaginary frequencies. All transition structures contained exactly one imaginary frequency and were linked to reactant, products or intermediates using intrinsic reaction coordinate (IRC) calculations.41 Except where noted, all energies in this publication include Gibbs free energy corrections (reported as ΔG). All minimum energy crossing point (MECP) calculations were carried out using the code provided by Dr Jeremy Harvey,42,43 at the M06/BS2 level of theory. MECP calculations are reported
This journal is © The Royal Society of Chemistry 2014
Paper
uncorrected due to the absence of stationary points at crossing points. Solvation effects were evaluated using the IEFPCM model44,45 for all single point calculations, using diethylether as the solvent in accordance with related experiments. No significant differences were found between optimised geometries in the gas phase or with solvent (see ESI†); gas phase geometries were used for all calculations discussed. All calculations were carried out with the Gaussian0946 set of programs.
3. Results and discussion 3.1.
Reaction between NO2 and Mo(NH2)3
The bond cleavage reaction can proceed via two general pathways, involving either one or two of the starting MoL3 complexes. The potential energy profiles for these mononuclear and dinuclear pathways are shown in full in Fig. 2, 6 and 8. The structures shown are labelled in the format N_X or TSN_X in the case of transition structures, where X indicates the multiplicity of the complex (S = singlet, D = doublet, T = triplet and Q = quartet). For the full N(R)(Ar) ligand, the subscript F is added (i.e. N_XF). 3.2.
Mononuclear pathway
The initial step of the reaction was coordination of an NO2 molecule to the quartet complex 1_Q; the quartet has been previously demonstrated to be lower in energy than the doublet alternative.26 The single unpaired electron in free NO2 is shared between all three atoms, as shown by the calculated
Fig. 2 Relative Gibbs free energies (kJ mol−1) for the mononuclear, triplet surface.
Dalton Trans., 2014, 43, 1620–1629 | 1621
View Article Online
Paper
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
Fig. 3
Dalton Transactions
Isomers of NO2–Mo(NH2)3.
Table 1 Isomers of NO2–Mo(NH2)3. Gibbs free energies (G) and electronic energies (E), relative to Mo(NH2)3 and unbound NO2 (kJ mol−1), calculated spin densities and selected bond lengths (Å)
NO2 + 1_Q 2_T 2_S 3_T 3_S 4_T 4_S 5_T 5_S 6_S
G
E
O1–N
O2–N
O1–Mo
N–Mo
O2–Mo
SMo
SNO2
0.0 −166.8 −196.3 −166.5 −210.7 −180.1 −167.2 −206.7 −211.6 −252.2
0.0 −229.8 −269.7 −224.0 −280.9 −236.8 −299.4 −265.4 −283.5 −325.1
1.194 1.235 1.275 1.341 1.362 1.351 1.409 1.291 1.298 1.380
1.194 1.215 1.204 1.184 1.193 1.183 1.167 1.258 1.253 1.380
— 2.712 2.321 2.046 2.024 2.020 1.924 2.075 2.052 1.951
— 2.110 1.945 2.117 1.946 2.959 2.922 2.599 2.597 1.959
— 3.047 3.012 3.294 3.114 4.015 3.941 2.187 2.155 1.951
2.968 1.843 — 1.579 — 1.861 — 1.701 — —
1.000 0.051 — 0.239 — 0.054 — 0.165 — —
spin densities: 0.492 on the nitrogen, and 0.254 on each of the oxygens, allowing all three atoms to be potential donors. The initial coordination occurs by pairing this single electron with one of those in molybdenum’s half-filled orbitals, leaving two remaining unpaired electrons on the metal. This can alternatively be interpreted as a single electron transfer from the metal to NO2, forming an intermediate nitrite anion which rapidly coordinates to the now cationic d2 molybdenum centre. In either case the end result is the formation of one of four possible linkage isomers, all with a triplet electronic arrangement (Fig. 3). A fifth η3-O,N,O complex 6_T was unable to be located, suggesting further donation from the NO2 requires an additional empty orbital on the metal. Relative energies and selected bond lengths are summarised in Table 1. Structures 2_T, 3_T, 4_T and 5_T are able to interchange with one another via the transition structures shown in Fig. 2, with barriers in the range of 10–50 kJ mol−1. In addition, each of these four complexes is able to undergo spin crossing to its singlet counterpart via a minimum energy crossing point (MECP). The energies of these MECPs, were 9, 6, 25 and 11 kJ mol−1 higher than the electronic energies of 2_T, 3_T, 4_T and 5_T respectively. The low barriers suggest that spin crossing is thus competitive with both isomerisation and bond breaking reactions. Pairing these electrons to form a singlet complex leaves an unoccupied orbital on the metal, allowing formation of the η3-O,N,O complex 6_S. Although 6_S was found to be the lowest energy isomer, the barrier to its formation via 5_S is relatively high at 55 kJ mol−1 (Fig. 6). The stabilising effect of this empty molybdenum orbital on the singlet surface can also be seen in the η2-N,O complex 3_S. The ability to donate the nitrogen
1622 | Dalton Trans., 2014, 43, 1620–1629
lone pair into the empty orbital results in a decrease in both the Mo–N bond length and the relative energy of 3_S compared to 3_T. Isomerisation to 3_S from both 2_S and 4_S was not found to have any discernable barrier. As rotation of the small NH2 ligands was unrestrained during optimisation, they were free to adopt the most favourable orientation in each complex. The three coordinate Mo(NH2)3 starting material has D3h symmetry, with each of the N–H bonds perpendicular to the MoL3 plane. Consistent with studies of related N2 and CS2 complexes, following coordination of the NO2 either one or two of the NH2 ligands rotates approximately 70°. This ligand rotation has been observed both computationally19,47 and in the crystal structures9 for several different NR2 ligands, suggesting that the driving force for this rotation is due to an electronic rather than steric interaction. NBO analyses have shown that in the trigonal MoL3 complex, the orientation of the amide ligands maximises the p ↔ d π interaction with the dxy and dx2−y2 orbitals, while minimising overlap with the partially filled dxz, dyz and dz2 orbitals. Once an additional ligand (such as NO2, N2, etc.) has coordinated, electron density is withdrawn from these partially filled orbitals. This allows one of the amide ligands to rotate and donate to the dxz, dyz and dz2 orbitals, thus promoting charge transfer to the fourth ligand, activating it towards bond cleavage (see Fig. 4). This rotation also brings the nitrogen 2p orbital of the rotated amide in close proximity to the NO2 donor atom and the Mo-donor atom σ* orbitals. Therefore whether one, two, or all three amides have rotated, as well as the exact NO2–Mo–N–H dihedral angles, is therefore different for each of the isomers and singlet–triplet pairs, as each isomer optimises the orbital overlap with the different donor atoms and Mo-donor atom σ* orbitals.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 01 November 2013. Downloaded by Monash University on 27/10/2014 17:16:26.
Dalton Transactions
Fig. 4
Paper
Metal–nitrogen orbital interactions.
Along with isomerisation and spin-crossing reactions, the majority of the complexes described above are also subject to the non-reversible, N–O bond cleavage. Considering first the triplet surface, bond breaking is most active from the relatively unstable η2-N,O complex 3_T, occurring with a barrier of only 8 kJ mol−1. The N–O bond is already weakened in 3_T, as indicated by the increase in N–O bond length from 1.194 to 1.384 Å. The transition structure (TS6_T) is also stabilised by the weak Mo–N interaction. Elongation (to 1.351 Å) and resulting activation of the N–O bond is also seen in 4_T, but without the Mo–N interaction, the barrier for bond cleavage is increased to 35 kJ mol−1. This is higher again (at 51 kJ mol−1) for the more symmetrical η2-O,O complex 5_T, in which the two N–O bonds remain relatively similar (1.291 and 1.258 Å). Each of these pathways results in the molybdenum oxide 7_D and a molecule of nitric oxide. NO is itself a potential ligand, and can coordinate to another molecule of 1_Q, pairing their respective electrons to give the triplet 10_T (Fig. 5). This complex then passes through a very low minimum energy crossing point (
