In 1997 Hubert et al. reported the solid-state synthesis¹ of a new boron subnitride B₆N by the reaction of boron with hexagonal graphite-like boron nitride (hBN) at 7.5 GPa and 1700 °C in a Walker-type multianvil apparatus [3]. According to the quantitative PEELS analysis, the oxygen-free grains of the product have a mean composition of B₆N_0.92, and its X-ray powder diffraction pattern seems to be similar to that of B₆O and B₄C. On this basis, the structure similar to that of other icosahedral boron-rich phases, more specifically to B₆O, has been suggested. However, the X-ray powder diffraction data reported in Ref. [3] does not seem to be unambiguous. The lines at low angles, which are suppose to be intensive in the case of B₆O- and B₄C-like structures, are of very low intensity and only two lines at higher angles may be attributed to the B₆N phase. At the same time, these two lines coincide with the 240 and 042 reflections of β-rhombohedral boron (β-rh B).

In Ref. [3], the samples have been powdered before taking the diffraction patterns; so, the absolute intensities of the lines should be close to the theoretical ones (how we see it for the patterns of B₆O and B₄C in Ref. [3]). Since preferred orientation is not expected for the B₆N phase, such weak 101, 003 and 012 lines of ‘B₆N’ observed in the pattern reported in Ref. [3] cannot be explained in the framework of the B₆O- and B₄C-like structures. Our crystallographic simulations using PowderCell program [4] have shown that neither nitrogen occupancy nor nitrogen/icosahedron positioning lead to such a strong decrease of intensities of all three lines in relation to the 104 and 021 reflections (Fig. 1). Therefore, the powder diffraction patterns reported in Ref. [3] can hardly belong to a B₆N phase with B₆O- or B₄C-like structure.

Very recently we found that at high pressures and temperatures β-rh boron forms the bulks with strong texture [5]. Even after powdering between two WC discs (the same has been done in Ref. [3]), the X-ray diffraction patterns of the boron samples show unpredictable intensities of some reflections. Most probably, it is caused by the tendency of boron to form single crystals. Only a fine powdering in a hard alloy mortar has allowed us to observe the powder diffraction pattern of β-rh boron with reasonable line intensities. So, the possible explanation of the ‘B₆N’ pattern reported in Ref. [3] is a strong preferred orientation of β-rh boron (or another boron phase that is stabilized by nitrogen impurity).

In order to clarify the situation, we have performed the experiments similar to that of Hubert et al. [3], but with an advantage of the in situ observation. The chemical interaction between boron and hBN at 7.4 GPa and 1700 °C have been studied using X-ray diffraction with synchrotron radiation at the European Synchrotron Radiation Facility. The experiments have been carried out using a new compact multianvil device combining a T-cup module with a Paris–Edinburgh press [6] at beamline ID27, ESRF. The mixture of β-rh boron² and hBN (99.9%, Alfa) of the B₆N stoichiometry has been subsequently loaded into a capsule of hBN and introduced into an octahedral cell assembly, which has been described in Ref. [6]. hBN has been used as an internal pressure calibrant [7]. Angle-dispersive diffraction patterns have been collected on an on-line MAR 345 image plate detector, which allows to record the high-quality data within less than a minute. High-brilliance synchrotron radiation from the ID27 two-phased undulator has been set to a wavelength of 0.3738 Å using an in-vacuum channel-cut Si (111) monochromator and collimated down to 100 × 30 μm² (horizontal by vertical) by two sets of tungsten carbide slits. An oscillating multichannel radial collimator [8] has been used to avoid diffraction from the materials surrounding the sample. Exposure times have been between 60 and 120 s depending on the electron current.

Sample temperature has been measured with the W3%Re–W25%Re thermocouple without correction for the pressure effect on the thermocouple emf. If the thermocouple failed, the temperature was estimated by the power-temperature calibration curve. Also, pressures and temperatures have been independently determined by analytically cross-calibrating the equations of state of hBN and Re with the program PTX-cal [9] using the unit-cell volumes obtained by the least-squares fits of the diffraction patterns obtained in situ.

The precise calibration of the sample-to-detector distance (369.649 mm) has been made by measuring a LaB₆ (99.999%) powder. Correction of the two-dimensional diffraction images for spatial distortions and integration of the Debye–Scherrer rings have been performed using the FIT2D software [10]. Lattice parameters of the crystal phases have been obtained by LeBail full profile refinement of integrated patterns using the GSAS [11] program package.

Fig. 2 shows the set of diffraction patterns obtained in situ during the heating of the boron–hBN mixture at 7.4 GPa. No interaction occurs in the mixture; only the random change in relative intensities of various boron reflections is observed. During the 30-min heating at 1700 °C, the boron lines do not disappear and the expected B₆N reflections do not arise. In the powder diffraction pattern of the quenched sample (G3000 TEXT, INEL) we do not see the lines consistent with the pattern of B₆N, although some reflections of boron have the same position as the reflections of hypothetical B₆O-like boron subnitride.

Thus, the results of our in situ studies of the chemical interaction between boron and hBN at 7.4 GPa and 1700 °C has allowed us to conclude that the evidence for the solid-state synthesis of boron subnitride B₆N with B₆O-like structure reported in Ref. [3] is inconclusive.