Hexacoordinated nitrogen(V) stabilized by high pressure

In all of its known connections nitrogen retains a valence shell electron count of eight therefore satisfying the golden rule of chemistry - the octet rule. Despite the diversity of nitrogen chemistry (with oxidation states ranging from + 5 to −3), and despite numerous efforts, compounds containing nitrogen with a higher electron count (hypervalent nitrogen) remain elusive and are yet to be synthesized. One possible route leading to nitrogen’s hypervalency is the formation of a chemical moiety containing pentavalent nitrogen atoms coordinated by more than four substituents. Here, we present theoretical evidence that a salt containing hexacoordinated nitrogen(V), in the form of an NF6− anion, could be synthesized at a modest pressure of 40 GPa (=400 kbar) via spontaneous oxidation of NF3 by F2. Our results indicate that the synthesis of a new class of compounds containing hypervalent nitrogen is within reach of current high-pressure experimental techniques.


Hexacoordinated nitrogen(V) stabilized by high pressure
Dominik Kurzydłowski, Patryk Zaleski-Ejgierd Table of Table S1 gives the ZPE corrected atomization energies (in eV) of NF n m+ molecules obtained with the use of the PBE and HSE06 functionals, compared to previously reported CCSD(T) calculations. 1 Our calculations were made using Gaussian09 2 with the cc-pVQZ basis set. For both PBE and HSE06 full geometry optimization was conducted. The main difference between HSE06 and PBE results ( transition pressure shifting more than 160 GPa from the HSE06 value of 243 GPa to 52 GPa at the PBE level. Similarly, the P-1 to I-4 transition pressure shifts from 78 GPa to 52 GPa. The same difference can be seen when comparing pressures at which the R3m phase, of (NF 4 + )(F -) composition, transforms to I4/m, of (NF 4 + ) 2 (NF 6 -)(F -) composition: the HSE06 value for this phase transition is 48 GPa compared to a much smaller value of 19 GPa at the PBE level of theory.

I. Benchmark calculations
We note that in the case of transitions between polymorphs containing non-hypervalent nitrogen the pressure shifts are much smaller, as can be seen in Table S3. Finally, it must be noted that at the PBE level of theory the enthalpy change associated with the reaction NF 3 + F 2 → NF 5 (I4/m polymorph) becomes negative already at 23 GPa compared to a value of 40 GPa for HSE06. Therefore, the pressure at which NF 5 might be synthesized is underestimated by a factor of about 2 when comparing PBE and HSE06 results.  In order to assess the stability of NF 5 against decomposition into NF 4 we have calculated the convex hull of the N/F system at 0, 100, 200, and 300 GPa. To find the best structures of NF 4 at a given pressure we have employed the evolutionary searches as described in the main text of the article, and re-optimized the low-enthalpy structures with the HSE06 functional.
As can be seen from

IV. Phonon stability of I4/m and P4/n
Calculations of the phonon density of state and dispersion curves were conducted with the use of the CASTEP code 4 implemented in the Materials Studio package. These calculations utilized the PBE functional and ultrasoft pseudopotentials, and were carried out on structure optimized with the same functional (PBE). The k-point spacing was set at 2π x 0.01 Å -1 , the cut-off energy of the plane waves was equal to 330 eV with a self-consistent-field convergence of 10 -6 eV per atom. Phonon calculations were conducted with the finite-displacement method for pre-optimized structures. Supercells of 2x2x2 size were used; they contained 120 and 160 atoms, respectively for I4/m at 40 GPa and P4/n at 150 GPa.
Below we show the phonon DOS and dispersion curves for I4/m optimized at 40 GPa, and P4/n optimized at 150 GPa. As can be seen no imaginary modes are observed at these pressures. We also do not find any imaginary modes at higher pressures (P > 40 for I4/m and P > 150 for P4/n).