Hexagonal boron nitride (hBN) has many unique properties such as a strong interaction with thermal neutrons, a wide energy bandgap (>6.0 eV), and surface phonon polaritons. These properties make it a good candidate for compact, efficient, and low-cost neutron detectors, deep UV light emitters and detectors, and infrared nanophotonic devices. Thus, high quality hBN single crystals are needed for national defense (detection and monitoring of nuclear weapons by neutron detectors), health (sterilization of surfaces by radiation from deep UV LEDs), and microscopy (deep sub-wavelength optical imaging).
Molten metal solutions in nitrogen at atmospheric pressure have proven effective in growing large, monoisotopic, high purity, and high quality hBN single crystals. Single crystals have been grown that are up to 3mm across, hundreds of microns thick, and with etch pit densities of 107 to 108 cm-2. Monoisotopic 10B and 11B enriched hBN crystals were also grown simply by using isotopically pure boron as the source material. While epitaxial growth techniques (CVD and MBE) have also proven effective at growing hBN, these methods require more complex apparatus and processes than solution growth and require a substrate which introduces lattice constant and thermal expansion coefficient mismatches. Furthermore, they cannot produce crystals as thick as solution growth, nor readily produce monoisotopic hBN, as they require specialized chemical sources, which are rarely available and quite expensive. However, to become commercially viable, solution growth must be further refined to produce much larger crystals with lower defect densities, and lower residual impurity concentrations.
One of the primary obstacles to optimizing this process is the lack of a cohesive understanding of how the melt composition impacts hBN crystal growth. For example, several different combinations of metals (Ni, Ni/Cr, Fe, Fe/Cr), two different boron sources (pure boron powder and hot-pressed boron nitride), and other additives such as silicon have been used in varying concentrations within the melt. In each case, the crystal nucleation density, grain size area, thickness, and other characteristics have been affected by the particular composition used. However, how the melt composition affects these characteristics and what specific properties are responsible for these effects has not been thoroughly analyzed. Thus, it is difficult to explain why certain compositions of boron, metals, and nitrogen yield excellent quality crystals while others do not.
In this work, processes employing various combinations of Ni, Cr, Fe, and other additives with a boron source were systematically analyzed according to the crystal size (obtained from micrographs), defect density (determined via defect-selective etching), and quality (determined via a variety of spectrographic tools including XRD and Raman spectroscopy). Specific features of the melt were analyzed with respect to these criteria to pinpoint exactly which properties were most important in the process. For example, pure Ni and Fe both have high boron solubilities, but nitrogen solubility is negligible in pure Ni. While alloying with Cr greatly increases the nitrogen solubility, it also increases the nucleation density. Thus, relatively large hBN single crystals can be produced in pure Fe, but not pure Ni. Thick hBN layers can be produced in Ni/Cr solutions, but with small crystal grain sizes. In addition, residual impurities present in the source materials, intentionally added impurities, and the interaction of the melt with the furnace and crucible materials significantly affects the process and the crystal properties. This project gives new insights into the interplay between melt composition and hBN crystal growth and reveal new, direct pathways toward optimizing the process.