Report

Synthesis and characterization of P2N3 : An aromatic ion composed of phosphorus and nitrogen

See allHide authors and affiliations

Science  29 May 2015:
Vol. 348, Issue 6238, pp. 1001-1004
DOI: 10.1126/science.aab0204

An aromatic phosphorus and nitrogen ring

In chemistry, the term “aromatic” denotes the energy stabilization associated with electrons being shared among atoms in a ring. Benzene is the best-known aromatic compound, although numerous related hydrocarbons also manifest the property. Velian and Cummins now report a rare instance of an inorganic aromatic compound: a negatively charged pentagonal ring composed of three nitrogen and two phosphorus atoms.

Science, this issue p. 1001

Abstract

Aromaticity is predominantly associated with carbon-rich compounds but can also occur in all-inorganic ones. We report the synthesis of the diphosphatriazolate anion, a rare example of a planar aromatic inorganic species. Treatment of azide (N3) in tetrahydrofuran solution with P2A2 (A = C14H10), a source of P2, produced P2N3, which we isolated as its [Na-kryptofix-221]+ salt in 22% yield and characterized by single-crystal x-ray diffraction. Salts [Na-kryptofix-221] [P2N3] and [Na-kryptofix-221] [P215NN2] were analyzed by infrared and Raman spectroscopy, 15N and 31P nuclear magnetic resonance spectroscopy, and mass spectrometry. The formation of the P2N3 anion was investigated using density functional theory, and its aromatic character was confirmed by NICS (nucleus-independent chemical shift) and QTAIM (quantum theory of atoms in molecules) methods.

In a chemical context, the term aromaticity refers to the special stability and electronic delocalization exhibited by planar hydrocarbons having 4n + 2 π electrons, the most famous example of which is benzene (C6H6, six π electrons, n = 1) (1, 2). The prevalence of aromatic hydrocarbons, many of which are naturally occurring in fossil fuel deposits, makes aromaticity largely the province of organic chemistry. Put forward as “inorganic benzene,” borazine (B3N3H6) is a colorless liquid and possesses high stability but is not deemed aromatic by the criterion of electronic delocalization (35). Aromaticity does not only occur in the case of six-membered planar rings; another instance is the pentagonal cyclopentadienide ion (C5H5, six π electrons). Theory suggests that an all-nitrogen pentagonal ion N5, isoelectronic to cyclopentadienide, would be aromatic; however, pentazolate has so far been observed only as a gas-phase ion by tandem mass spectrometry experiments. There, the pentazolate anion was generated from parahydroxyphenylpentazole by cleavage of the organic residue from the N5 unit using a high collision voltage (6). SN2P2, a sulfur-pnictogen ring with six π electrons, was recently identified in the gas phase by infrared (IR) spectroscopy as a product of the flash pyrolysis of SP(N3)3 (7).

By contrast, the all-phosphorus aromatic five-membered anionic ring P5 was identified as a persistent species in solution (8); it could even be used in ensuing synthetic studies giving rise to metallocene analogs in which a P5 ring replaces a cyclopentadienide ligand, as exemplified by the ferrocene-analogous complex Cp*Fe(P5) (Cp* = C5Me5; Me = methyl) (9). The Cp*Fe(P5) complex has proven to be a valuable building block in supramolecular chemistry, assembling into nanometer-scale balls upon interaction with copper(I) halides (10).

Recently, we prepared a versatile, anthracene-based source of P2, P2A2 (A = anthracene or C14H10), shown to release anthracene and diphosphorus upon mild heating (11). In the context of accessing all-pnictogen heterocycles, we wondered whether the diphosphatriazolate anion P2N3 (Fig. 1) could be prepared in a “click” reaction between diphosphorus (provided by P2A2) and the azide anion N3, similar to triazole formation in the [3+2] cyclization of azides with alkynes or phosphaalkynes (1215).

Fig. 1 Solid-state molecular structure of [Na-kryptofix-221] [P2N3].

Hydrogen atoms are omitted for clarity; the ellipsoids are plotted at the 50% probability level. The P2N3 unit is disordered over two positions with occupancies of 92% (major component) and 8% (minor component), respectively (inset A). Selected interatomic distances and angles: (major component) N2-N1, 1.326 ± 0.002 Å; N1-P1, 1.677 ± 0.002 Å; P1-P2, 2.069 ± 0.001 Å; P2-N3, 1.680 ± 0.002 Å; N3-N2, 1.310 ± 0.002 Å; P1-P2-N3, 93.17° ± 0.08°; P2-N3-N2, 117.7° ± 0.2°; N1-N2-N3, 118.5° ± 0.2°; (minor component) N3B-N2B, 1.32 ± 0.03 Å; N2B-N1B, 1.32 ± 0.02 Å; N1B-P1B, 1.64 ± 0.02 Å; P1B-P2B, 2.08 ± 0.02 Å; P2B-N3B, 1.69 ± 0.02 Å. A model of P2N3 anion was optimized using DFT methods (27, 32); selected interatomic distances and angles: P-P, 2.07 Å; P-N, 1.69 Å; N-N, 1.31 Å; P-P-N, 93°; P-N-N, 117°; N-N-N, 120°.

We found that P2N3 formed in quantitative spectroscopic yield (along with anthracene) when a solution of P2A2 (Fig. 2) in tetrahydrofuran (THF) was heated to 70°C for 3 hours in the presence of excess [TBA] [N3] (6.2 equiv; TBA = tetrabutylammonium). The P2N3 anion displays a phosphorus chemical shift at +334 ppm and is readily detected by electrospray ionization (ESI) spectrometry (mass-charge ratio m/z = 104.9). Similar solubility properties impeded the separation of the desired [TBA] [P2N3] salt from excess [TBA] [N3] and anthracene, but when the reaction was repeated with a substoichiometric amount of [Na-kryptofix-221] [N3], pure [Na-kryptofix-221] [P2N3] could be isolated by crystallization in 22% yield as colorless crystals sensitive to air and moisture.

Fig. 2 Characterization of the [Na-kryptofix-221] [P215NN2] salt.

(A) Preparation by thermolysis of P2A2 in the presence of [Na-kryptofix-221] [15NN2]. (B and C) Experimental (up) and simulated (down) 15N NMR and 31P NMR spectra, respectively, of the [P215NN2] ion. (D and E) IR and Raman vibrational spectra, respectively, of solid [Na-kryptofix-221] [P2N3] and [Na-kryptofix-221] [P215NN2].

The identity of the [Na-kryptofix-221] [P2N3] salt was confirmed by x-ray diffraction analysis (Fig. 1) (16). Of remarkable simplicity, the P2N3 anion reveals itself as an essentially planar ring approximating C2v point group symmetry, with metrical parameters that compare well to those predicted using density functional theory (DFT). The P-P and P-N interatomic distances [2.068 ± 0.001 Å and 1.679 (avg.) Å, respectively] fall in between typical single- and double-bond values (17, 18); this multiple bonding character is also reflected by the delocalization indices calculated for the P-P and P-N bonds (1.48 and 1.19, respectively) (1921). The delocalized bonding present in the P2N3 anion is best visualized by plotting the Laplacian of the calculated electronic charge density in the plane of the ring (Fig. 3) (22); here, the areas of local charge concentration are distributed over all atomic basins.

Fig. 3 Theoretical bonding analysis.

Natural resonance theory (NRT) (33) resonance weights (top) calculated for the diphosphatriazolate anion, P2N3 (34, 35). The topology of the Laplacian distribution of charge density (bottom) in the plane of the P2N3 anion is shown with regions of charge depletion in solid curves and those of charge concentration in dashed curves. Bond paths, bond and ring critical points, and atomic basin paths are also depicted (21).

To probe the expected aromaticity of the diphosphatriazolate anion relative to other five-membered rings with six π electrons (N5, P5, C5H5, and N2S32+) that are deemed aromatic (23, 24), we calculated its nucleus-independent chemical shift along the z axis (NICSzz) (24). Similar to the aforementioned rings, the NICSzz profile for the P2N3 anion was found to be class 1, with the highest absolute value for NICS at ~0.6 Å above the ring critical point, diagnostic for a π aromatic system (Fig. 4) (25). Moreover, magnetic susceptibility analysis by QTAIM (quantum theory of atoms in molecules) showed that π orbitals made the largest contribution (57%) to the total magnetic aromaticity of the P2N3 anion (26, 27).

Fig. 4 Probing aromaticity.

Plotted is NICS(z) versus r (Å) (27) for Cp, P2N3, P5, N5, and N2S32+.

The downfield 31P nuclear magnetic resonance (NMR) shifts observed for the P2N3 and P5 anions (+334 and +467 ppm, respectively) are characteristic for these aromatic, all-pnictogen anionic rings. The C2v symmetry of P2N3 observed in polar solvents such as THF and acetonitrile, where the anion appears to tumble freely, is broken in nonpolar solvents such as benzene, where the phosphorus atoms become inequivalent and two peaks in a 1:1 ratio are detected by 31P NMR spectroscopy. This desymmetrization is also observed in the solid-state structure, where N3 and N3b are preferentially oriented toward C23 of the cryptand (N3-C23 distance = 3.329 ± 0.002 Å); this may be the effect of a weak donor-acceptor interaction (C23-H···N3) (28). Consistent with this hypothesis is the natural bond orbital (NBO) analysis of the diphosphatriazolate anion, which reveals N1 and N3 as the most basic sites of the P2N3 ring, each bearing a partial −0.64 charge (Fig. 3).

To access more information about the properties and formation of the P2N3 ion, we used terminally labeled [Na-kryptofix-221] [15NN2] as a reaction partner together with P2A2, giving rise to the formation of [Na-kryptofix-221] [P215NN2] (15N between P and 14N; Fig. 2) (27). The IR and Raman vibrational spectra for the 15N-labeled diphosphatriazolate anion are predictably red-shifted with respect to the unlabeled P2N3 ion and are well reproduced computationally (table S2). Involving primarily the vibration of the P-P bond, the A1 Raman mode is largely unaffected by the incorporation of the 15N label and is detected at 519 cm−1 for both the labeled and the unlabeled anion. For comparison, the breathing mode of the P5 ion was observed at 463 cm−1 in solution (8), and the stretching frequency of the P-P bond in diphosphene ArP=PAr (Ar = 2,4,6-tri-tert-butylphenyl) was observed at 610 cm−1 (29), which suggests that the P-P bond in P2N3 is stronger than in its all-phosphorus analog but weaker than in the diphosphene. The B1 Raman vibration of the diphosphatriazolate anion, involving primarily the stretching and contracting of the P-N bonds, red-shifts from 714 cm−1 for P2N3 to 709 cm−1 for P215NN2.

With three spin ½ nuclei, the P215NN2 anion is an ABX spin system and displays second-order coupling (30, 31). The 15N NMR spectrum of P215NN2 consists of an apparent doublet of doublets centered at +38.0 ppm, while its 31P NMR spectrum reveals an apparent doublet (Jvirtual = 52 Hz) flanked by two sets of outer wings (±460 Hz; Fig. 2). With the aid of the Spin Simulation package of the MNova NMR program, we estimated 1JP1−P2, 1JP2−N3, 2JP1−N3, and ΔνP1P2 to be −460, +102, +2, and 17 Hz, respectively (fig. S6 and table S1). The large P1-P2 coupling constant between the almost equivalent phosphorus atoms effects a stronger coupling of the P1 atom to the 15N nucleus in position 3 than otherwise (31). The presence of the outer wings in the 31P NMR spectrum, centered at ±460 Hz from the center of the apparent doublet (Jvirtual = 52 Hz), is characteristic for the ABX spin system (31). 15N NMR spectroscopy data rule out the presence of the 1,3 isomer of the P215NN2 ion in the reaction mixture, as the 1,3 isomer can incorporate the 15N atom in either position 2 or 4 of the heterocycle. This would give rise to two distinct signals in the 15N NMR spectrum, contrary to the experimental findings.

Using DFT methods [M06-2X, def2-TZVPP+, COSMO(THF)] (27), we investigated two routes through which P2N3 could form: (i) an associative route in which the reactive intermediate P2N3A forms from P2A and N3 and subsequently loses A, and (ii) a dissociative route in which P2, formed in the fragmentation of P2A, reacts directly with N3. Similar to the P2 transfer from P2A2 to 1,3-cyclohexadiene (11), we found the two routes to have comparable energetic profiles (fig. S13), likely both contributing to the formation of the P2N3 anion.

This work shows that it is possible to synthesize, using an original P2 transfer reaction, a planar aromatic heterocycle composed exclusively of nitrogen and phosphorus. Until now, to render isolable a molecular entity having substantial P=P double-bond character, the approach has been to use an architecture with bulky adjacent substituents to block the site, thereby preventing access to the reactive moiety and attendant degradation (18). Here, the small title anion P2N3 is entirely free of blocking groups and the observed stability must be attributed to electronic stabilization. Considering the data in aggregate, this stabilization is best construed as aromaticity—an effect traditionally reserved for the domain of organic chemistry.

Supplementary Materials

www.sciencemag.org/content/348/6238/1001/suppl/DC1

Materials and Methods

Figs. S1 to S13

Tables S1 to S15

References (3656)

References and Notes

  1. Metrical parameters for the structure of [Na-kryptofix-221] [P2N3] are available free of charge from the Cambridge Crystallographic Data Centre under reference CCDC-1051576.
  2. See supplementary materials on Science Online.
  3. Acknowledgments: This material is based on work supported by NSF grant CHE-1362118.
View Abstract

Navigate This Article