Transformation of CO₂ and isocyanates mediated by N-borane-substituted cyclic phosphine imides (BCPIs) via λ⁵-oxazaphosphetanes

Abstract

We herein report reliable evidence that λ⁵-oxazaphosphetane species are a key intermediate in the transformation of CO₂ and isocyanates through their reaction with N-borane-substituted cyclic phosphine imides (BCPIs). We have isolated and fully characterized several λ⁵-oxazaphosphetane species prepared via formal [2 + 2] cycloaddition reactions between BCPIs and CO₂ or isocyanates. The transformation of these λ⁵-oxazaphosphetanes via retro-ring opening reaction afforded an isocyanate and a carbodiimide from the CO₂- and isocyanate-derived λ⁵-oxazaphosphetanes.

---

Introduction

Phosphine imides (R′-N=PR₃), a nitrogen analogue of phosphine oxides, are widely used in the aza-Wittig reaction¹ and the Staudinger reaction/ligation.^2,3 The aza-Wittig reaction has been proposed to be initiated from a formal [2 + 2] cycloaddition between the phosphine imides and CO₂ (or other heterocumulenes with an O=C=X moiety; X = O or NR′′) to yield phosphine oxides and R′-N=C=X via λ⁵-oxazaphosphetane species as key intermediates (Fig. 1A).^1b,4 However, the direct observation and isolation of λ⁵-oxazaphosphetane species from the reaction of a phosphine imide and CO₂ have not yet been reported,⁵ even though related studies on their synthesis, characterization, and reactivity have been published by e.g., Röschenthaler et al.,⁶ Schmidpeter and von Criegern et al.,⁷ as well as Kawashima et al.⁸ Moreover, a plausible mechanism for the aza-Wittig reaction has been examined theoretically using DFT calculations, which supported the formation of λ⁵-oxazaphosphetane intermediates.⁹ Thus, direct evidence of these species in the transformations of heterocumulenes including CO₂ would provide valuable fundamental insight into the reaction mechanism. Here, we report the isolation and characterization of λ⁵-oxazaphosphetane species via the direct reaction of CO₂ or isocyanates with N-borane-substituted cyclic phosphine imides (BCPIs), which have recently been developed in our group (Fig. 1B).¹⁰ We also discuss the retro-ring opening reactions from the λ⁵-oxazaphosphetanes, which unambiguously confirm their intermediacy in aza-Wittig reactions.

Fig. 1 (A) Simplified scheme of the aza-Wittig reaction between phosphine imides and heterocumulenes such as CO₂ (X = O) and isocynates (X = NR′′). (B) This work (X = O and NR′′).

Results and discussion

We first explored the reaction between BCPI 1 and CO₂ (Fig. 2A). Treatment of 1 with CO₂ (5 atm) at room temperature (rt) in α,α,α-trifluorotoluene (TFT) resulted in the quantitative formation of λ⁵-oxazaphosphetane 2 (isolated yield: >99%) within 20 minutes.¹¹ Compound 2 was characterized using multinuclear-NMR-spectroscopy techniques as well as single-crystal X-ray diffraction (SC-XRD) analysis. The ³¹P NMR spectrum of 2 shows a characteristic chemical shift (δ_P −18.5) that is consistent with the formation of a penta-coordinated phosphorus species. The molecular structure of 2 in the solid state, which was determined by SC-XRD analysis (Fig. 2B), confirmed the formation of the P–N3–C4–O1 ring. The phosphorus atom adopts a distorted trigonal bipyramidal geometry with the N2 and O1 atoms at the axial positions (N2–P–O1 = 158.3(1)°), while the N3 atom and two C atoms reside in the trigonal plane. The N3–P (1.673(3) Å) and N2–P (1.899(2) Å) bonds in 2 are notably elongated compared to those in 1, which suggests a decrease in their bond order. In contrast, the length of the B1–N3 bond remains almost unchanged (∼1.56 Å) through the transformation from 1 to 2. At rt, 2 gradually decomposed to yield N-phosphinoyl isocyanate 3 with concomitant formation of a dimer of an imidazole-substituted 9-borafluorene (4)₂. These results constitute the direct observation of a λ⁵-oxazaphosphetane species as the intermediate in the aza-Wittig reaction between phosphine imides and CO₂. Given that the thermolysis of 2 can be expected to be promoted by the formation of the thermodynamically favorable P=O bond, we subsequently explored the cleavage of the P–O1 bond in 2. The reaction of 2 and HB(C₆F₅)₂ afforded 5 in 97% yield through the coordination of the carboxylate moiety to the borane (O2–B2 = 1.518(4) Å; Fig. 2A and C).^12,13 Thus, reconfiguration of the geometry around the phosphorus center to form a distorted tetrahedron occurs, together with a contraction of the N2–P bond (1.720(3) Å) and an elongation of the B1–N3 bond (1.608(5) Å). The length of the N3–P bond remained virtually unchanged (1.642(3) Å) compared to that of 2 and is thus significantly longer than that in 1. NMR analyses provided insight into the electronic state of the P atom in 5. The resonance at δ_P 59.9 indicates that the electron density on the P atom in 5 is nearly identical to that in 1, albeit that it should be slightly higher.¹⁴ Based on these experimental data, we conclude that the electron delocalization between the N3 atom and the adjacent carbonyl group partially weakens the negative hyperconjugation between the N3 and P atoms, while the increased π-donation from the imidazolium unit to the P atom compensates for the loss of electron density on the P atom.

Fig. 2 (A) Synthesis and transformations of 2; the yield of the isolated products is given. ^a Based on the molar loading of 2. (B) Molecular structure of 2 with thermal ellipsoids at 30% probability; H atoms are omitted for clarity. Selected bond lengths [Å] and angles [°]: C1–B1 1.626(4), N3–B1 1.556(4), P–O1 1.831(2), O1–C4 1.334(5), C4–O2 1.212(4), C4–N3 1.371(3), N2–P–N3 84.6(1), N3–P–O1 73.7(1), P–O1–C4 89.2(2), O1–C4–N3 102.2(2), C4–N3–P 94.8(2). (C) Molecular structure of 5 with thermal ellipsoids at 30% probability; H atoms (except B2–H) are omitted for clarity. Selected bond lengths [Å] and angles [°]: C1–B1 1.624(5), N2–P–N3 92.7(2), P–N3–B1 119.8(2), N3–B1–C1 96.2(3), B1–N3–C4 122.8(3).

Next, we investigated reactions between 1 and isocyanates. When p-toluenesulfonyl isocyanate (Ts–N=C=O) was employed, the expected λ⁵-oxazaphosphetane (6a) was isolated in >99% yield within 15 minutes at rt (Fig. 3). The molecular structure of 6a in the solid state was determined by SC-XRD analysis and also supports the formation of the P–N3–C2–O ring, in which the phosphorus atom adopts a distorted trigonal bipyramidal geometry, wherein the N1 and O atoms at located at the axial positions, whereas the N3 atom and two C atoms reside in the trigonal plane. In stark contrast to 2, 6a exhibits significant thermal stability, i.e., no transformation occurred even after 12 h at 100 °C. At 135 °C, 6a gradually decomposed to yield 1-(N-phosphinoyl)-3-tosylcarbodiimide 7a and (4)₂. Again, these results clearly demonstrate the key intermediacy of λ⁵-oxazaphosphetane species in the aza-Wittig reaction between phosphine imides and isocyanates.

Fig. 3 Synthesis and transformations of 6a derived from TsNCO; the yield of the isolated products is given. ^a Based on the molar loading of 6a. Molecular structure of 6a with thermal ellipsoids at 30% probability; H atoms and solvated CH₂Cl₂ are omitted for clarity. Selected bond lengths [Å] and angles [°]: C1–B 1.618(2), N1–P 1.855(1), P–N3 1.687(1), N3–B 1.563(2), P–O 1.895(1), O–C2 1.320(2), C2–N3 1.356(2); N1–P–N3 85.48(5), P–N3–B 126.63(9), N3–B–C1 94.5(1), N3–P–O 72.15(5), P–O–C2 87.73(8), O–C2–N3 104.5(1), C2–N3–P 95.65(8).

We also used phenyl isocyanate in the reaction with 1 and confirmed the formation of the corresponding λ⁵-oxazaphosphetane (6b), which was isolated in 90% yield (Fig. 4A). Thus, we expected to obtain carbodiimide 7b from 6b, in analogy to the case of 6a (Fig. 3); however, the thermolysis of 6b at 40 °C resulted in the quantitative formation of guanidine 8b, which was isolated in 85% yield (Fig. 4A). Compounds 6b and 8b were unambiguously characterized using multinuclear-NMR-spectroscopy techniques and SC-XRD analyses; the latter results are shown in Fig. 4B and C, respectively. The molecular structure of 8b includes an N-phosphinoylguanidine unit comprising the N1, N3, N4, and C2 atoms.

Fig. 4 (A) Synthesis and transformations of 6b derived from PhNCO; the yield of the isolated products is given. (B) Molecular structure of 6b, depicted with thermal ellipsoids at 30% probability; H atoms and CH₂Cl₂ are omitted for clarity. Selected bond lengths [Å] and angles [°]: C1–B 1.617(3), N1–P 1.908(2), P–N3 1.673(2), N3–B 1.548(2), P–O 1.817(1), O–C2 1.360(2), C2–N3 1.378(2); N1–P–N3 84.19(8), P–N3–B 128.8(1), N3–B–C1 95.1(1), N3–P–O 74.49(7), P–O–C2 89.4(1), O–C2–N3 101.1(2), C2–N3–P 95.0(1). (C) Molecular structure of 8b with thermal ellipsoids at 30% probability; H atoms and solvated CH₂Cl₂ are omitted for clarity. Selected bond lengths [Å] and angles [°]: C1–B 1.618(3), C2–N4 1.340(2), N4–B 1.581(2), C2–N3 1.275(2); N1–C2–N4 105.5(1), C2–N4–B 115.9(2), N4–B–C1 94.4(1), N4–C2–N3 127.7(2), C2–N3–P 138.3(1).

A plausible reaction mechanism affording 7a or 8b from 6 is shown in Fig. 5. First, the thermal decomposition of 6 occurs via the cleavage of the B–N, N–P, and C–O bonds, which results in the in situ formation of imidazole-substituted 9-borafluorene 4 and carbodiimides 7a or 7b. Subsequently, 4 dimerizes to form (4)₂ when the Lewis basicity of the nitrogen atom in 7a is insufficient to form B–N adduct due to the electron-withdrawing nature of the N–Ts group. On the other hand, the N atom in 7b exhibits sufficient Lewis basicity to coordinate to the boron center in 4 before its dimerization, which eventually yields guanidine 8b.

Fig. 5 A plausible mechanism for the conversion of 6.

Conclusions

In conclusion, we have gathered direct evidence and thus demonstrated for the first time that λ⁵-oxazaphosphetane species serve as key intermediates in aza-Wittig reactions involving CO₂ and isocyanates. We used N-borane-substituted cyclic phosphine imides (BCPIs) for the isolation and characterization of several λ⁵-oxazaphosphetanes derived from CO₂ and isocyanates via formal [2 + 2] cycloaddition reactions. Retro-ring opening reactions from the isolated λ⁵-oxazaphosphetanes afforded the expected isocyanates (from CO₂) or carbodiimides (from TsNCO) with concomitant formation of a dimer of an imidazole-substituted 9-borafluorene. In their entirety, these results successfully provide evidence for a missing link in contemporary organic chemistry. Further research into the applications of BCPIs in the fields of organic and organometallic chemistry is currently in progress in our group.

Data availability

The data supporting this article have been included as part of the ESI.†

Metrical data for the solid-state structures are available from Cambridge Crystallographic Data Centre: CCDC 2240798 (2), 2240797 ((4)₂), 2240796 (5), 2378677 (6a), 2378678 (6b), 2378679 (8b).†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This project was supported by Grants-in-Aid for Transformative Research Area (A) Digitalization-driven Transformative Organic Synthesis (22H05363 to Y. H.), the JST FOREST Program (JPMJFR2222 to Y. H.), and a JSPS Research Fellowship (JSPS KAKENHI grant JP23KJ1443 to S. N.). The authors would like to thank Mr Takaya Hinogami (Department of Applied Chemistry, Faculty of Engineering, Osaka University) for his contribution to exploring the reactivity of BCPIs toward CO₂.

References

1. (a) N. I. Gusar, Russ. Chem. Rev., 1991, 60, 146; (b) S. Eguchi, Y. Matsushita and K. Yamashita, Org. Prep. Proced. Int., 1992, 24, 209; (c) P. Molina and M. J. Vilaplana, Synthesis, 1994, 1197; (d) H. Wamhoff, G. Richardt and S. Stölben, Adv. Heterocycl. Chem., 1995, 64, 159; (e) P. M. Fresneda and P. Molina, Synlett, 2004, 1; (f) F. Palacios, C. Alonso, D. Aparicio, G. Rubiales and J. M. de los Santos, Tetrahedron, 2007, 63, 523.
2. (a) H. Staudinger and J. Meyer, Helv. Chim. Acta, 1919, 2, 635; (b) H. Staudinger and E. Hauser, Helv. Chim. Acta, 1921, 4, 861; (c) Y. G. Gololobov, I. N. Zhmurova and L. F. Kasukhin, Tetrahedron, 1981, 37, 437; (d) E. W. Abel and S. A. Mucklejohn, Phosphorus, Sulfur Silicon Relat. Elem., 1981, 9, 235; (e) Y. G. Gololobov and L. F. Kasukhin, Tetrahedron, 1992, 48, 1353.
3. (a) M. Köhn and R. Breinbauer, Angew. Chem., Int. Ed., 2004, 43, 3106; (b) E. Saxon and C. R. Bertozzi, Science, 2000, 287, 2007; (c) S. S. van Berkel, M. B. van Eldijk and J. C. M. van Hest, Angew. Chem., Int. Ed., 2011, 50, 8806; (d) Z.-P. A. Wang, C.-L. Tian and J.-S. Zheng, RSC Adv., 2015, 5, 107192; (e) C. Bednarek, I. Wehl, N. Jung, U. Schepers and S. Bräse, Chem. Rev., 2020, 120, 4301.
4. (a) L. Horner and H. Oediger, Chem. Ber., 1958, 91, 437; (b) G. Aksnes and P. Frøyen, Acta Chem. Scand., 1969, 23, 2697; (c) J. Barluenga and F. Palacios, Org. Prep. Proced. Int., 1991, 23, 1; (d) L. Sarah, Doctoral thesis, Durham University, 1998.
5. (a) T. Sasaki, S. Eguchi and T. Okano, J. Am. Chem. Soc., 1983, 105, 5912; (b) F. A. Merckling and P. Rüedi, Tetrahedron Lett., 1996, 37, 2217; (c) U. Seneviratne, L. C. Godoy, J. S. Wishnok, G. N. Wogan and S. R. Tannenbaum, J. Am. Chem. Soc., 2013, 135, 7693; (d) H. B. Abed, O. Mammoliti, O. Bande, G. V. Lommen and P. Herdewijn, Org. Biomol. Chem., 2014, 12, 7159.
6. (a) W. Storzer, G.-V. Röschenthaler, R. Schmutzler and W. S. Sheldrick, Chem. Ber., 1981, 114, 3609; (b) R. Francke, D. Dakternieks, R. W. Gable, B. F. Hoskins and G.-V. Röschenthaler, Chem. Ber., 1985, 118, 922; (c) D. V. Sevenard, E. Lork, K. I. Pashkevich and G.-V. Röschenthaler, Heteroat. Chem., 2002, 13, 97.
7. (a) A. Schmidpeter and T. von Criegern, Angew. Chem., Int. Ed. Engl., 1978, 17, 55; (b) W. S. Sheldrick, D. Schomburg, A. Schmidpeter and T. von Criegern, Chem. Ber., 1980, 113, 55.
8. (a) N. Kano, X. J. Hua, S. Kawa and T. Kawashima, Tetrahedron Lett., 2000, 41, 5237; (b) N. Kano, J.-H. Xing, S. Kawa and T. Kawashima, Polyhedron, 2002, 21, 657; (c) N. Kano, J.-H. Xing, A. Kikuchi, S. Kawa and T. Kawashima, Phosphorus, Sulfur Silicon Relat. Elem., 2002, 177, 1685.
9. (a) J. Koketsu, Y. Ninomiya, Y. Suzuki and N. Koga, Inorg. Chem., 1997, 36, 694; (b) W. C. Lu, C. C. Sun, Q. J. Zang and C. B. Liu, Chem. Phys. Lett., 1999, 311, 491; (c) W. C. Lu, C. B. Liu and C. C. Sun, J. Phys. Chem. A, 1999, 103, 1078; (d) Y. Xue, D. Xie and G. Yan, J. Phys. Chem. A, 2002, 106, 9053; (e) Y. Xue and C. K. Kim, J. Phys. Chem. A, 2003, 107, 7945; (f) F. P. Cossío, C. Alonso, B. Lecea, M. Ayerbe, G. Rubiales and F. Palacios, J. Org. Chem., 2006, 71, 2839; (g) M. Fianchini and F. Maseras, Tetrahedron, 2019, 75, 1852; (h) A. Adda and H. Sediki, Chem. Proc., 2021, 3, 47.
10. S. Nagai, T. Hinogami, S. Ogoshi and Y. Hoshimoto, Bull. Chem. Soc. Jpn., 2023, 96, 1346.
11. For a related example using amidophosphoranes in CO₂ fixation, see: L. J. Hounjet, C. B. Caputo and D. W. Stephan, Angew. Chem., Int. Ed., 2012, 51, 4714.
12. λ⁵-Oxazaphosphetanes could also be a key intermediate in the phosphine-imide-catalyzed conversion of CO₂; for details, see: M.-A. Courtemanche, M.-A. Légaré, É. Rochette and F.-G. Fontaine, Chem. Commun., 2015, 51, 6858.
13. We also carried out a reaction between 2 and HSiEt₃, which unfortunately resulted merely in the thermal decomposition of 2 to 3 and (4)₂.
14. O. Kühl, Phosphorus-31 NMR Spectroscopy: A Concise Introduction for the Synthetic Organic and Organometallic Chemist, Springer-Verlag, Berlin, Heidelberg, 2008.

---

Footnote

† Electronic supplementary information (ESI) available: Synthetic, spectroscopic, and X-ray crystallographic details. CCDC 2240796–2240798 and 2378677–2378679. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4ob01565g

---