Homoleptic low-valent polyazides of group 14 elements

Abstract

First examples of coordinatively unsaturated, homoleptic azido complexes of low-valent group 14 elements are reported. A simple strategy uses low-valent precursors, ionic azide transfer reagents and bulky cations to obtain salt-like compounds containing E(N₃)₃⁻ of Ge(II)/Sn(II) which are fully characterised, including XRD. Remarkably, these compounds are kinetically stable at r.t. and isolable in sub-gram quantities.

---

Binary azides are known for all elements in group 14 and exist as covalent E(N₃)₄ compounds (E = C, Si),² as hyper-coordinate E(N₃)₆²⁻ complexes (Si–Pb)¹ and as E(N₃)₂ compounds (Sn, Pb).^5c However, no low-valent, homoleptic group 14 complex has yet been reported. All known binary p-block azides are highly endothermic primary explosives most of which possess exceedingly high electrostatic and friction sensitivities and a propensity to release N₂. As covalent, N-rich compounds, their isolation is generally challenging and experimental characterisation is limited.⁵ In contrast, stability-inducing effects of hyper-coordination and of bulky, weakly coordinating counter ions (WCC)⁶ allow many salt-like homoleptic polyazides to be synthesised in bulk and characterised fully, including via X-ray crystallography and IR spectroscopy,^5a owing to azide groups (N₃) giving rise to intense bands in the mid-IR region. It has been shown that azide anions (N₃⁻) are able to coordinate to low-valent centres in compounds such as E(L)(N₃) and E(L′)(N₃)₂, E = Ge, Sn.^3,4 On the other hand, the stability of low valent molecules, e.g. carbenes, silylenes, germylenes, stannylenes,⁷ increases by saturating the electron deficient centre with sterically demanding, electron donating groups, such as N-based C(NⁱPr)₂(NⁱPr₂) and HC{(CMe)(2,6-ⁱPr₂C₆H₃N)}₂ ligands.^7a,d This insight has led to tri- and tetracoordinate complexes bearing uni-, bi- and terdentate ligands, e.g. E(NHC)X₂, Ge(NHC)₂Cl⁺ and Ge{HB(Me₂pz)₃}Cl, E = Si, Ge; X = Cl-I, N₃; NHC = N-heterocyclic carbene.^4a,8 Exploitation of these concepts has permitted the synthesis and characterisation of the first low-valent and homoleptic Ge and Sn azides described in this paper.

Compounds already containing WCC ions and low-valent germanium, AsPh₄GeCl₃, PPh₄GeCl₃,⁹ PPNGeCl₃ (ref. 10) (1a–c), were prepared in high yield from the GeCl₂(diox) adduct¹⁸ and WCC chlorides¹⁹ (Scheme 1, route A).‡¹¹ These colourless, moderately air sensitive crystalline trichlorogermanates react readily with THF suspensions of NaN₃. In situ IR spectra of the reaction (2b) show bands due to asymmetric NNN stretches, ν_as(N₃), typical for coordinated N₃ groups at [small nu, Greek, macron]_max/(cm⁻¹) = 2092 and 2058, which have grown fully after a reaction time of 1 h. Exposure of the reaction solution to fresh NaN₃ results in no further spectral change. From the solution, a highly air sensitive, colourless solid (3b) was precipitated, the IR spectrum of which exhibits the finger print of PPh₄⁺ and the ν_as(N₃) bands. The ν_as(N₃) frequencies lie within the range of those reported previously for semi-covalent germanium(II) azides (2027–2077 cm⁻¹, Fig. 1), below those of Ge(IV) azides (Ge(N₃)₄, PPN₂Ge(N₃)₆, 5c)^1b and above that of the N₃⁻ ion. While solution ¹H, ¹³C and ³¹P NMR spectra of 3b show signals of the WCC cations only, two peaks are observed in the ¹⁴N NMR spectra at −263 and −207 ppm next to the solvent (−136 ppm) with FWHM line widths of 552, 147 and 24 Hz, respectively. These characteristics are typical for the N_α and N_γ nuclei of coordinated N₃ groups while the signal for N_β is obscured by solvent.^5a

Scheme 1 Synthesis of azido(chloro) germanates (1−) and stannates (1−), E = Ge, n = 1 (1, 3); E = Sn, n = 0 (2, 4); WCC = AsPh₄ (a), PPh₄ (b), N(PPh₃)₂ (PPN, c).

Fig. 1 IR spectra of Ge(N₃)₃⁻ (black), Sn(N₃)₃⁻ (red) in THF; ν_as(N₃) frequency ranges of related azides, Ge(N₃)₄,¹² HN₃,¹⁰ Ge(N₃)₆²⁻,^1b L_nGe(II) azides L₁ = {Me₂(^tBuO)Si}₂N;^4c L₂ = ⁿPr₂ATI, Mes₂DAP,^4f,g (NHC)₂;³ L₃ = HB(R₂pz)₃, (C₅R₅)Co{P(O)(OEt)₂}₃,^4a,e,10 (2,6-ⁱPr₂C₆H₃)₂C₂H₂N₂CGe(N₃)₂ (ref. 8d) and N₃⁻,^4e are indicated in the top bar; see Table S1 (ESI†) for exact values.

Alternative routes to 3 and 4 use WCC azides as azide transfer reagents. GeCl₂(diox) reacts directly also with (PPN)N₃ in MeCN solution (Scheme 1, route B). Intriguingly, equimolar reactant mixtures produce only one ν_as(N₃) band (2078 cm⁻¹); increasing the stoichiometric ratio (1 : 2) results in two additional bands (2088, 2066 cm⁻¹), while at ratios of 1 : 9 and above only bands at 2095, 2064 cm⁻¹ and that of N₃⁻ were detected. These observations are interpreted tentatively in terms of the formation of mono, di- and triazido complexes.

Tin dichloride was subjected to a similar treatment as GeCl₂(diox) using WCC(N₃) and NaN₃; however, complete Cl/N₃ exchange requires a larger excess of azide transfer reagent. Similar observations as with 3b were made, including the intermediate rise and decay of a ν_as(N₃) band (2064 cm⁻¹) and the ultimate rise of bands of the final product 4b (2081, 2050 cm⁻¹) in the expected region between Sn(N₃)₆²⁻ and charge-neutral Sn(II) monoazides (Table S1, ESI†), and ¹⁴N resonances at −218.5 ppm (FWHM ≈ 32 Hz) and −260.0 ppm (166 Hz). The ¹⁴N NMR signals of 3b and 4b, in particular those assigned to N_α, are deshielded in comparison to those of E(N₃)₆²⁻ dianions.^1b,5a,173b and 4b are soluble in MeCN, THF and CH₂Cl₂.

The synthetic strategy was extended to AsPh₄⁺ and PPN⁺ counter ions affording compounds 3a,c and 4c (Scheme 1A) which all have spectroscopic properties analogous to those of 3b and 4b described already. The combined analytical evidence, including the absence of chlorine in 3b and the ¹¹⁹Sn NMR signal of 4b (δ = −220 ppm, see ESI†) point to the formation of anionic complexes in compounds of the type (WCC)E(N₃)₃ as the final products of Cl/N₃ exchange.

Further insight into the nature of intermediates and products of the exchange reactions was obtained from quantum chemical calculations²⁰ on the ECl_(3−n)(N₃)_n⁻ species, which were performed at the B3LYP/cc-pVTZ level²¹ with effective core potentials²² for Ge and Sn. Solvent (THF) was described using PCM.²³ The calculations found conformational isomerism resulting in several minima for n = 1, 2, 3, that were close in energy. These conformers are related by rotation of ligands. Since rotational barriers of sterically unhindered N₃ groups are small (cf. GeH₃N₃, ∼1 kJ mol⁻¹),¹³ fast interconversion involves all significantly thermally populated rotamers above the minimum energy conformation (E_rel < 5.8 kJ mol⁻¹). This process is likely to result in averaged absorption bands weighted by the rotamer population (rotamers may have more than one degenerate, absolute spatial configuration, and inter-rotamer vibrational energy transfer is unaccounted for). Taking account of the theoretical equilibrium mole fractions, absorption intensities and scaled vibrational frequencies,²⁴ approximate average frequencies of the in-phase and out-of-phase ν_as(N₃) stretches and the qualitative intensity ratios could be determined (Table S1, ESI†), which match those observed (e.g. Ge(N₃)₃⁻, 2059, 2091 vs. 2060, 2093; Sn(N₃)₃⁻, 2051, 2080 vs. 2050, 2078 cm⁻¹). This approach leads to the assignment of the observed bands of intermediates to GeCl₂(N₃)⁻, GeCl(N₃)₂⁻, SnCl₂(N₃)⁻ and SnCl(N₃)₂⁻. Calculations using the Gauge-Independent Atomic Orbital method²⁵ verify the assignment of ¹⁴N NMR data (see ESI†).

Crystals of azido germanates were grown from THF–Et₂O (1 : 10) solutions at −18 °C (3b, needles) or by diffusion of Et₂O into concentrated THF solutions (3c).§¶ According to single crystal X-ray diffraction studies, 3b consists of PPh₄⁺ and Ge(N₃)₃⁻ ions (Fig. 2). The shortest interionic Ge⋯N and N⋯N distances were found to be 4.13 and 5.07 Å, respectively (Fig. 4, left), hence, covalent {[Ge(N₃)₃]⁻}⋯{[Ge(N₃)₃]⁻} interactions are absent (Fig. 3). Germanium is coordinated by three, essentially linear N₃ ligands and occupies the apical position in a trigonal-pyramidal Ge[N]₃ framework. The ligands are bound in the fashion typical of covalent azides and adopt Ge–N_α–N_β angles between 116° and 121°. All inter-ligand angles are close to 90° which indicates stereochemical inactivity of the lone electron pair at germanium.¹⁴

Fig. 2 Thermal ellipsoid plot (50%) of Ge(N₃)₃⁻ in the crystal of PPh₄Ge(N₃)₃ (3b). Bond lengths [Å] Ge1–N7 1.984(2), Ge1–N1 1.988(3), Ge1–N4 2.011(3), N1–N2 1.213(3), N2–N3 1.148(3), N4–N5 1.209(3), N5–N6 1.142(3), N7–N8 1.206(3), N8–N9 1.140(3), angles [°] N7–Ge1–N1 93.59(10), N7–Ge1–N4 94.29(10), N1–Ge1–N4 91.05(11), N2–N1–Ge1 116.4(2), N5–N4–Ge1 118.7(2), N8–N7–Ge1 121.3(2).

Fig. 3 Thermal ellipsoid plot (50%) of {Sn₂(N₃)₆}²⁻ in crystals of PPh₄Sn(N₃)₃ (4b). Bond lengths [Å] Sn–N1 2.262(3), Sn–N4 2.193(3), Sn–N7 2.207(3), 2.674(3), N1–N2 1.203(4), N2–N3 1.148(5), N4–N5 1.200(5), N5–N6 1.143(5), N7–N8 1.185 (5), 1.193(5), N8–N9 1.157(6), 1.168(6) Snⁱ–N6 3.567(4), angles [°] N1–Sn–N4 88.35(12), N4–Sn–N7 89.14(13), N1–Sn–N7 88.59(12), N2–N1–Sn 123.6(3), N5–N4–Sn 119.4(2), N8–N7–Sn 128.2(9), 120.2(8), N7–Sn–N7 68.32(16), Sn–N7–Sn 111.68(18).

Fig. 4 Packing diagrams of PPh₄Ge(N₃)₃ (3b, left) and PPh₄Sn(N₃)₃ (4b, right); H (bright grey), C (dark grey), N (light blue), P (orange), Ge and Sn (turquoise).

This structural feature has been found in the valence isoelectronic complexes of Ge₂(μ-pz*)₃⁺ GeCl₃⁻ (ref. 11f) and pilocarpine-trichlorogermanate hemihydrate.^11g The Ge–N_α bonds of 3b are shorter than those of tetracoordinate Ge(II) azides 2.088(6)–2.094(7) Å (Fig. 1), longer than those of the homoleptic Ge(IV) azide PPN₂Ge(N₃)₆ (5c, Table S1, ESI†) and rather within the range of previously investigated, tricoordinate Ge(II) azides (1.969–2.047(2) Å, Table 1). All other bond lengths and angles are close to those of 5c (Table 1). The crystallographic structure of Ge(N₃)₃⁻ is consistent with one of the geometries predicted by DFT (vide supra).

Table 1 Bond lengths D/Å in the salt-like homoleptic azides of the type (WCC)⁺_n[E(N₃)_3n]ⁿ⁻, E = Ge, Sn; n = 1, WCC = PPh₄⁺; n = 2, WCC = N(PPh₃)₂⁺

E, n	E–Nα	Nα–Nβ	Nβ–Nγ	ΔNNav^a
a , in parentheses. b s ≪ σ, error estimated by . c This work. d Ref. 1b. e Ref. 15.
Ge, 1	1.984(3)–2.011(2)	1.206(4)–1.213(4)	1.140(4)–1.148(4)	6.6(5)^b
Ge, 2	1.969(2)–1.980(3)	1.210(4)–1.214(3)	1.143(3)–1.151(3)	6.5(4)
Sn, 1	2.193(3)–2.262(3)	1.189(6)–1.203(4)	1.143(5)–1.163(6)	4.6(17)
Sn, 2	2.117(3)–2.134(2)	1.182(3)–1.213(3)	1.111(4)–1.148(3)	5.7(10)

---

Single crystalline needles of 4b were obtained and investigated with the methods used for 3b.¶ The asymmetric unit of 4b also contains a E(N₃)₃⁻ moiety; however, the packing is at variance with 3b, which allows Sn(N₃)₃⁻ to interact via two long E⋯N_α bonds and form {Sn(N₃)₃}₂²⁻ dimers (Fig. 3). The interaction involves asymmetric μ_1,1-N₃ bridges with short (2.207(3) Å) and long (2.674(3) Å) Sn–N_α bonds, the latter being considerably shorter than the sum of the van der Waals radii (3.72 Å).¹⁶ Weak intermolecular interactions have been observed previously between neutral Sn(ⁿPr₂ATI)N₃ complexes (vide supra),^4b where a slightly longer Sn⋯N_α bond (2.87 Å) was found. The sum of bond angles involving the bridging N_α indicates planarity and effective sp² hybridisation. As in the crystal of 3b, the primary E(II)–N_α bonds are significantly longer (2.193(3)–2.262(3) Å) than those found in the homoleptic E(IV) azide 6 (2.125 Å).¹⁵ The potential for dimerisation was studied by DFT using the geometry of {E(N₃)₂(μ_1,1-N₃)}₂²⁻ in crystalline 4b as a starting point. Optimisation results in separate anions devoid of covalent interionic interactions in the case of Ge(N₃)₃⁻, whereas a dimeric structure was found for Sn(N₃)₃⁻ that resembles the molecular structure in the crystal. Estimates of the basis set superposition error for the solution phase were obtained from BSSE calculations²⁶ in the gas phase. After BSSE correction, {Sn(N₃)₂(μ_1,1-N₃)}₂²⁻ was found to be at least 4 kcal mol⁻¹ less stable than two monomers, rendering the existence of a dimer in solution highly unlikely.

According to differential scanning calorimetry measurements (Fig. S11 and S12, ESI†), compound 3b decomposes in two exothermic processes with (ΔH = −270 and −467 J g⁻¹).|| Remarkably, step 1 occurs at temperatures (T^ex1_on = 99 °C) that are drastically below the decomposition onset of the related, hypercoordinate azide 5c (T^ex1_on = 256 °C),^1b whereas step 2 sets in at T^ex2_p = 310 °C, which is nearly identical with the temperature found in 5c (312 °C).** Furthermore, the molar enthalpies of step 2 (−251 vs. −482 kJ mol⁻¹) scale approximately with the complex charge; however, step 1 releases much less energy than expected (145 kJ mol⁻¹, 3bvs. 705 kJ mol⁻¹, 5c). This phenomenon is still under investigation. Further experiments show that heating of 3b at 150 °C produces PPh₄N₃, which suggests that the release of N₃⁻ initiates the decomposition of Ge(N₃)₃⁻. The tin homologue 4b melts at T_on = 115 °C and decays at 215 °C and 308 °C and thus behaves as expected relative to 6c.¹⁵ No sensitivity was noted during preparation and analysis of compounds 3b and 4b on the stated reaction scale. The material remains unchanged when struck by a hammer. Upon lighting up, it burns rapidly with an orange flame leaving black residues.

In solution, Ge(N₃)₃⁻ and Sn(N₃)₃⁻ react with hydrazoic acid leading to a precipitate with the IR ν_as(N₃) absorptions characteristic for Ge(N₃)₆²⁻ and Sn(N₃)₆²⁻ complexes, respectively, which can be verified by comparison with spectra of the fully characterized salts 5c, 6c (Scheme 2, Fig. S13 and S14, ESI†).

Scheme 2 Oxidation of the tri(azido) complexes 3b and 5b.

The first low-valent homoleptic azido complexes of group 14 have been synthesized and fully characterised. The preparative approach to salt-like compounds containing these complexes has been demonstrated on a 0.2–0.7 g scale for a range of weakly coordinating cations and involves chloro(azido) species, ECl_x(N₃)_y⁻. Unlike their hypercoordinate E(N₃)₆²⁻ analogues and despite the presence of innocent cations, the new class of compounds is highly reactive, exhibiting low thermal stability and a propensity to oxidation. Crystallographic analysis revealed that in the solid state, E(N₃)₃⁻ complexes of group 14 may dimerise via azido ligand bridges. Neither of the low-valent coordination centres exhibits a stereochemically active lone electron pair. DFT calculations correctly predict the dimerisation and suggest furthermore that the dimers are unstable in polar solvents.

The authors thank the EPSRC (EP/E054978/1), the University of Sheffield and Humboldt-Universität zu Berlin for support and Prof. A. C. Filippou for advice.

Notes and references

1. (a) A. C. Filippou, P. Portius and G. Schnakenburg, J. Am. Chem. Soc., 2002, 124, 12396; (b) A. C. Filippou, P. Portius, D. U. Neumann and K.-D. Wehrstedt, Angew. Chem., Int. Ed., 2000, 39, 4333; (c) D. Fenske, H. D. Dörner and K. Dehnicke, Z. Naturforsch., B: J. Chem. Sci., 1983, 38, 1301; (d) K. Polborn, E. Leidl and W. Beck, Z. Naturforsch., B: J. Chem. Sci., 1988, 43, 1206.
2. (a) K. Banert, Y.-H. Joo, T. Rüffer, B. Walfort and H. Lang, Angew. Chem., Int. Ed., 2007, 46, 1168; (b) P. Portius, A. C. Filippou, G. Schnakenburg, M. Davis and K.-D. Wehrstedt, Angew. Chem., Int. Ed., 2010, 49, 8013.
3. Y. Xiong, S. Yao and M. Driess, Chem. Commun., 2014, 50, 418.
4. (a) A. C. Filippou, P. Portius and G. Kociok-Köhn, Chem. Commun., 1998, 2327; (b) H. V. R. Dias and A. E. Ayers, Polyhedron, 2002, 21, 611; (c) M. Veith and A. Rammo, Z. Anorg. Allg. Chem., 2001, 627, 662; (d) V. N. Khrustalev, I. A. Portnyagin, N. N. Zemlyansky, I. V. Borisova, Y. A. Ustynyuk and M. Y. Antipin, J. Organomet. Chem., 2005, 690, 1056; (e) A. C. Filippou, P. Portius, G. Kociok-Köhn and V. Albrecht, Dalton Trans., 2000, 1759; (f) A. E. Ayers, D. S. Marynick and H. V. R. Dias, Inorg. Chem., 2000, 39, 4147; (g) A. E. Ayers, T. M. Klapötke and H. V. R. Dias, Inorg. Chem., 2001, 40, 1000.
5. (a) P. Portius and M. Davis, Coord. Chem. Rev., 2013, 257, 1011; (b) W. P. Fehlhammer and W. Beck, Z. Anorg. Allg. Chem., 2013, 639, 1053 , ref. cited; (c) T. Müller, F. Karau, W. Schnick and F. Kraus, Angew. Chem., Int. Ed., 2014, 53, 13695.
6. (a) C. J. Price, H.-Y. Chen, L. M. Launer and S. A. Miller, Angew. Chem., Int. Ed., 2009, 48, 956; (b) I. Krossing and A. Reisinger, Coord. Chem. Rev., 2006, 250, 2721.
7. (a) M. Asay, C. Jones and M. Driess, Chem. Rev., 2011, 111, 354; (b) S. Nagendran and H. W. Roesky, Organometallics, 2008, 27, 457; (c) Y. Mizuhata, T. Sasamori and N. Tokitoh, Chem. Rev., 2009, 109, 3479; (d) B. Blom, M. Stoelzel and M. Driess, Chem. – Eur. J., 2013, 19, 40; (e) M. Mück, K. Junold, J. A. Baus, C. Burschka and R. Tacke, Eur. J. Inorg. Chem., 2013, 5821.
8. (a) R. S. Ghadwal, H. W. Roesky, S. Merkel, J. Henn and D. Stalke, Angew. Chem., Int. Ed., 2009, 48, 5683; (b) A. C. Filippou, O. Chernov and G. Schnakenburg, Angew. Chem., Int. Ed., 2009, 48, 5687; (c) Y. Xiong, S. Yao and M. Driess, Chem. Commun., 2014, 50, 418; (d) B. Lyhs, D. Bläser, C. Wölper, S. Schulz, R. Haack and G. Jansen, Inorg. Chem., 2013, 52, 7236.
9. U. M. Tripathi, G. L. Wegner, A. Schier, A. Jockisch and H. Schmidbaur, Z. Naturforsch., B: J. Chem. Sci., 1998, 53, 939.
10. P. Portius, PhD thesis, Humboldt-Universität zu Berlin Weißensee Verlag, Berlin, 2002, ISBN 3-934479-63-4.
11. (a) X. Tian, T. Pape and N. W. Mitzel, Z. Naturforsch., B: J. Chem. Sci., 2004, 59, 1524; (b) S. Nogai, A. Schriewer and H. Schmidbaur, Dalton Trans., 2003, 3165; (c) G. Kociok-Köhn, J. G. Winter and A. C. Filippou, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1999, 55, 351; (d) G. L. Wegner, A. Jockisch and H. Schmidbaur, Z. Naturforsch., B: J. Chem. Sci., 1998, 53, 430; (e) M. Karnop, W. W. du Mont, P. G. Jones and J. Jeske, Chem. Ber., 1997, 130, 1611; (f) A. Steiner and D. Stalke, Inorg. Chem., 1995, 34, 4846; (g) S. Fregerslev and S. E. Rasmussen, Acta Chem. Scand., 1968, 22, 2541.
12. J. E. Drake and R. T. Hemmings, Can. J. Chem., 1973, 51, 302.
13. D. T. Durig, M. S. Durig and J. R. Durig, Spectrochim. Acta, Part A, 2005, 61, 1287.
14. (a) D.-K. Seo, N. Gupta, M.-H. Whangbo, H. Hillebrecht and G. Thiele, Inorg. Chem., 1998, 37, 407; (b) U. Schwarz, H. Hillebrecht, M. Kaupp, K. Syassen and H.-G. v. Schnering, J. Solid State Chem., 1995, 118, 20.
15. R. Campbell, B. Peerless and P. Portius, unpublished results.
16. A. Bondi, J. Phys. Chem., 1964, 68, 441.
17. (a) W. Beck, W. Becker, K. F. Chew, W. Derbyshire, N. Logan, D. M. Revitt and D. B. Sowerby, Dalton Trans., 1972, 245; (b) W. Beck, W. P. Fehlhammer, P. Pöllmann, E. Schuierer and K. Feldl, Chem. Ber., 1967, 100, 2335.
18. (a) J. Kouvetakis, A. Haaland, D. J. Shorokhov, H. V. Volden, G. V. Girichev, V. I. Sokolov and P. Matsunaga, J. Am. Chem. Soc., 1998, 120, 6738; (b) S. P. Kolesnikov, I. S. Rogozhin and O. M. Nefedov, Izv. Akad. Nauk SSSR, Ser. Khim., 1974, 23, 2379.
19. (a) V. Y. Kukushkin and A. I. Moiseev, Inorg. Chim. Acta, 1990, 176, 79; (b) A. Martinsen and J. Songstad, Acta Chem. Scand., Ser. A, 1977, 31, 645.
20. M. J. Frisch, et al., Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford, CT, 2013.
21. (a) A. D. Becke, J. Chem. Phys., 1993, 98, 5648; (b) C. Lee, W. Yang and R. Parr, Phys. Rev. B, 1988, 37, 785; (c) T. H. Dunning Jr., J. Chem. Phys., 1989, 90, 1007.
22. K. A. Peterson, J. Chem. Phys., 2003, 119, 11099.
23. (a) B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 106, 5151; (b) M. Cossi, V. Barone, B. Mennucci and J. Tomasi, Chem. Phys. Lett., 1998, 286, 253 and references therein.
24. K. K. Irikura, R. D. Johnson and R. N. Kacker, J. Phys. Chem. A, 2005, 109, 8430.
25. R. Ditchfield, Mol. Phys., 1974, 27, 789.
26. S. F. Boys and F. Bernardi, Mol. Phys., 1970, 19, 553.

---

Footnotes

† Electronic supplementary information (ESI) available: Spectra, thermograms, full crystallographic data and computational details. CCDC 1030031 and 1030032. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c5cc00259a

‡ GeCl₃⁻ (ref. 9 and 11) and SnCl₃⁻ salts (ref. 11c and e–g) with various organic counter ions have been reported previously.

§ All attempts to crystallize compound 3a have been futile.

¶ Crystallographic data: 3b, CCDC 1030032, C₂₄H₂₀GeN₉P, 538.07 g mol⁻¹, P1̄, a = 7.7712(2) Å, b = 11.4711(4) Å, c = 14.2003(4) Å, α = 93.278 (2)°, β = 99.357(2)°, γ = 100.865(2)°, Z = 2, V = 1221.59(6) ų, D_c = 1.463 g cm⁻³, T = 120(2) K, F(000) = 548, R₁ = 0.0376 (316 param.), wR₂ = 0.0807, GOOF = 1.090. 3c, P2₁, a = 10.7640(11) Å, b = 12.732(2) Å, c = 25.713(3) Å, α = γ = 90°, β = 100.682(12)°, D_c = 1.414 g cm⁻³, T = 180(2) K, extensive disorder of Ge(N₃)₃⁻ part (see ESI† and ref. 10). 4b, CCDC 1030031, C₂₄H₂₀N₉PSn, M = 584.15 g mol⁻¹, P1̄, a = 10.7560(6) Å, b = 11.0605(6) Å, c = 12.3540(7) Å, α = 91.668(4)°, β = 108.414(4)°, γ = 116.545(3)°, Z = 2, V = 1222.11(12) ų, D_c = 1.587 g cm⁻³, T = 100(2) K, F(000) = 584, R₁ = 0.0521 (334 param.), wR₂ = 0.1020, GOOF = 1.053.

|| Estimated error approx. ±10%.

** This step is assigned tentatively to the thermolysis of PPN(N₃) and PPh₄ (N₃) since enthalpies and onset temperatures are comparable with those of genuine samples of these salts: PPh₄N₃, mp = 250 °C, T^ex_on = 291 °C, ref. 15.

---