Rhodium(I) Pincer Complexes of Nitrous Oxide.

The synthesis of two well-defined rhodium(I) complexes of nitrous oxide (N2 O) is reported. These normally elusive adducts are stable in the solid state and persist in solution at ambient temperature, enabling comprehensive structural interrogation by 15 N NMR and IR spectroscopy, and single-crystal X-ray diffraction. These methods evidence coordination of N2 O through the terminal nitrogen atom in a linear fashion and are supplemented by a computational energy decomposition analysis, which provides further insights into the nature of the Rh-N2 O interaction.

The synthetic exploitation of nitrous oxide (N2O) is an enduring challenge that draws topical interest as a means to remediate the detrimental impact emission of this kinetically stable gas on the environment.1 Whilst the application of homogenous transition‐metal complexes is an attractive prospect, the underpinning inorganic chemistry is conspicuously under‐developed.2 Indeed, the number of discrete transition‐metal complexes of N2O is currently limited to a handful of examples (A–D), of which only two have been structurally characterised in the solid state using X‐ray diffraction (Figure 1).3, 4, 5, 6, 7 This paucity is attributed to the extremely poor ligand properties of N2O, conferred by a low dipole moment, weak σ‐donor and π‐acceptor characteristics, and the propensity of these adducts for subsequent N−N or N−O bond cleavage.2

Figure 1

Well‐defined transition‐metal complexes of nitrous oxide. Weakly coordinating [BArF 4]− anions have been omitted from the reaction scheme for clarity.

Inspired by work by Brookhart and Caulton,8, 9 and building upon that conducted in our laboratories,10 we chose cationic phosphine‐based pincer complexes of rhodium(I) as a platform for studying the coordination chemistry of N2O. Our approach utilises dimeric [{Rh(pincer)}2(μ‐η2:η2‐COD)][BArF 4]2 (1 a, pincer=2,6‐(tBu2PO)2C5H3N; 1 b, pincer=2,6‐(tBu2PCH2)2C5H3N; COD=cyclooctadiene, ArF=3,5‐(CF3)2C6H3) as synthons for reactive {Rh(pincer)}+ fragments in the weakly coordinating solvent 1,2‐F2C6H4 (DFB).11 Satisfactorily, reactions of 1 with N2O (1.5 bar) at room temperature afforded well‐defined adducts [Rh(pincer)(N2O)][BArF 4] 2 in quantitative yield by 31P NMR spectroscopy, as evidenced by resonances at δ 210.4 (2 a, 1 J RhP=134 Hz, t<3 h)/ δ 70.9 (2 b, 1 J RhP=127 Hz, t<5 min) that display diagnostic 103Rh coupling (Figure 1). These Rh–N2O complexes were subsequently isolated as analytically pure materials in good yield on precipitation with hexane at low temperature and extensively characterised (2 a, 65 %; 2 b 78 %). Both can be stored under argon in the solid state, but decompose slowly in DFB solution at room temperature (2 a, t 50 %dec≈4.0 h; 2 b, t 50 %dec≈2.5 h), with generation of the known dinitrogen complexes [Rh(pincer)(N2)][BArF 4] (3 a, δ 211.3, 1 J RhP=133 Hz; 3 b, δ 71.2, 1 J RhP=126 Hz).10, 12 By drawing parallels with the reaction of a neutral rhodium PNP analogue with N2O, where formation of a discrete adduct is inferred but not experimentally corroborated, we suggest 2 decomposes by a bimetallic oxygen atom transfer mechanism that is initiated by dissociation of N2O and proceeds via {(pincer)RhII–N=N–O–RhII(pincer)}2+.9, 13 Consistent with this assertion, enhanced solution stability was observed under a N2O atmosphere (2 %/30 % decomposition of 2 a/b after 24 h).

The structures of 2 were definitively established in DFB solution using 15N NMR spectroscopy, aided by samples prepared using isotopically labelled 15N2O (98 % 15N, Figure 2 and the Supporting Information). Intact coordination of N2O through the terminal nitrogen atom is evident by an upfield shift of Δδ 43.3/37.0 for the corresponding 15N resonances, comparable to that of B,4 which exhibit 1 J RhN coupling of 28/27 Hz14 and 2 J PN coupling of 4/5 Hz for 2 a/b, respectively. The internal 15N resonances are conversely shifted downfield by Δδ 15.3/14.3, but retain coupling to both 103Rh (2 J RhN=8 Hz) and 31P (3 J PN=1 Hz), albeit with commensurate reductions in magnitude. The 1 J NN coupling constant of free N2O (9 Hz) is small and appreciably reduced on complexation to rhodium; to the extent that it is only just resolved (1 Hz). Analysis of natural abundance and isotopically labelled (powdered) solid samples of 2 using ATR IR spectroscopy enabled unambiguous assignment of the principal ν(N‐N) and ν(N‐O) bands of 2 a (2279, 1252 cm−1, respectively) and 2 b (2267, 1228 cm−1, respectively). The former are significantly blue shifted, relative to the free ligand, whilst the latter are red shifted (2224, 1285 cm−1, respectively).15 Spectra of 2 can also be acquired in DFB and give similar values (full details are provided in the Supporting Information). Higher energy ν(N‐N) bands are also observed for C and D,5, 6 with this phenomenon running contrary to normal expectations for (meaningful) metal‐to‐ligand π‐back bonding.

Figure 2

Solid‐state structures (150 K) and 15N NMR spectra (15N2O atmosphere, DFB, 61 MHz, 298 K) of 2. Thermal ellipsoids drawn at 50 % (2 a) and 30 % (2 b) probability; minor disordered components (pincer ligand in 2 b) and anions omitted for clarity. Selected bond lengths and angles: 2 a Rh1–P2 2.2677(5) Å, Rh1–P3 2.2688(5) Å, Rh1–N4 1.981(2) Å, N4–N5 1.108(3) Å, N5–O6 1.194(3) Å, Rh1–N20 2.007(2); N20–Rh1–N4 178.05(8)°, Rh1–N4–N5 173.4(2)°, N4–N5–O6 178.5(3)°; 2 b Rh1–P2 2.282(2) Å, Rh1–P3 2.288(2) Å, Rh1–N4 1.962(7) Å, N4–N5 1.111(11) Å, N5–O6 1.192(11) Å, Rh1–N20/N20A 2.071(7)/2.052(8) Å; N20/N20A–Rh1–N4 178.1(5)/175.4(5)°, Rh1–N4–N5 176.8(11)°, N4–N5–O6 178.7(14)°.22

The solid‐state structures of 2 a and 2 b have been determined using single‐crystal X‐ray diffraction (150 K) and verify that they are both discrete N2O complexes, with the ligand binding through the terminal nitrogen atom in a linear fashion (Rh–N–N>173° and N–N–O>178°; Figure 2). The former is well‐ordered, but the latter features an extensively disordered pincer ligand symptomatic of dynamic isomerism in the lattice (C 2 ⇌ C s conformations).16 Whilst this disorder was modelled satisfactorily, the metal–ligand metrics in 2 b are inevitably determined with reduced precision in comparison to 2 a. Nevertheless, the pertinent data associated with coordination of N2O in 2 point to very similar bonding characteristics. When normalising by the sum of the covalent radii,17 the extent of the M–N interactions in 2 a (1.981(2) Å) and 2 b (1.962(7) Å) are in close agreement with those previously determined in C (2.1389(10) Å) and D (1.890(8) Å): r(M‐N)/[r(M)+r(N)]=0.93, 2 a; 0.92, 2 b; 0.95, C; 0.93, D.5, 6 There is a trend for the N−N bonds (1.108(3)/1.111(11) compared to 1.128 Å) to be shortened and the N−O bonds (1.194(3)/1.192(11) compared to 1.184 Å) to be elongated in 2 a/b relative to free N2O,15 but these changes are marginal.

Supplementing the experimental work, the geometries and thermodynamics of 2 were interrogated in silico at the DLPNO‐CCSD(T)/def2‐TZVPP//ωB97X‐D3/def2‐TZVP(‐f) level of theory.18 Whilst the trend for a longer Rh–N contact in 2 a (2.006 Å) relative to 2 b (1.989 Å) established by X‐ray diffraction is reproduced, it is for the former that binding of N2O is predicted to be most exergonic (ΔG 298K=−68.5 kJ mol−1, 2 a; −67.6 kJ mol−1 2 b). The magnitude of these values is consistent with slow exchange on the 15N NMR timeframe (61 MHz, 298 K; Figure 2), with the difference congruent with the relative rate of decomposition observed in solution. Only very small perturbations to the N–N and N−O bond lengths are computed on coordination (less than 0.005 Å), but the associated vibrations corroborate the experimental pattern and are discernibly blue and red shifted, respectively (see the Supporting Information). To gain deeper insight into the nature of the Rh–N2O interaction, a local energy decomposition (LED) analysis was carried out using ORCA 4.1.2. (Table 1; Supporting Information).18, 19, 20 The results reiterate marginally stronger N2O binding in 2 a (D e=+122.3 kJ mol−1) compared to 2 b (D e=+120.1 kJ mol−1) and highlight the important role of dispersion, which accounts for approximately 12 % of the total stabilising interactions.21 The interfragment orbital energies are small and reflect the presence of weak σ‐donation and π‐back bonding; with the former predominating (ca. 56 % vs. 40 %). When the two complexes are compared, the combined stabilising interactions are most pronounced for 2 b, but counteracted by even more extensive Pauli repulsion (i.e. sterics) than in 2 a. The latter difference is reconciled by the more obtuse bite angle of the PNP (P–Rh–P=169.24(8)°, expt) vs. PONOP (P–Rh–P=162.77(2)°, expt) pincer ligand, which causes greater buttressing between the tBu substituents and the coordinated N2O ligand.

Table 1

LED results for 2 → {Rh(pincer)}+ + N2O (kJ mol−1).18

	2 a	2 b
ΔE interaction	−124.8	−119.8
E dispersion	−70.0	−76.5
ΔE Pauli	+476.8	+519.1
E electrostatic	−359.0	−385.5
ΔE orbital	−140.9	−148.0
ΔE orbital (σ donation)[a]	−79.8 (57 %)	−83.3 (56 %)
ΔE orbital (π backbonding)[a]	−56.2 (40 %)	−59.2 (40 %)
ΔE residual	−31.7	−33.1
ΔE preparation	+2.5	+3.8
ΔE binding (=−D e)	−122.3	−120.1

[a] Determined by application of the extended transition state method for energy decomposition analysis combined with the natural orbitals for chemical valence theory (ETS‐NOCV). The character of the interaction is deduced from visual inspection of the natural orbitals. Percentage of total orbital interaction ΔE orbital in parenthesis.

In summary, the synthesis and comprehensive characterisation of two rhodium(I) pincer complexes of N2O are reported. Through an unprecedented combination of 15N NMR and IR spectroscopy, and single crystal X‐ray diffraction the discrete nature of these complexes and the coordination of N2O to the metal through the terminal nitrogen atom in a linear fashion is unequivocally established. Subtle differences in the characteristics of the Rh–N2O interaction associated with the ancillary pincer ligand employed have been reconciled using a computational energy‐decomposition analysis, which highlights the weakly interacting nature of N2O, the important stabilising role of dispersion interactions, and the effect of steric buttressing with the pincer substituents.

Conflict of interest

The authors declare no conflict of interest.

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.

Supplementary

Click here for additional data file.

Supplementary

Click here for additional data file.