Ammonothermal Synthesis and Crystal Growth of Nitrides

This chapter is intended to introduce the ammonothermal synthesis as an alternative technique to other methods for nitride materials production. Properties of liquid and supercritical ammonia with focus on use as a solvent for nitride synthesis and crystal growth are discussed and compared to those of water. Finally, inherent drawbacks of the use of fluidic ammonia arising from its chemical properties are considered.

The following chapter provides an overview of the most important application fields of nitride and oxonitride materials and outlines briefly their structural features as well as their materials properties. (Oxo)nitrides are employed in a variety of important technological areas such as structural ceramics, heat conductors, semiconductor technology, solid-state lighting, water splitting and solid-state battery materials. Their significance for daily life as well as possible advancement of (oxo)nitride materials with respect to these applications is examined.

Due to the thick-walled autoclaves, the ammonothermal system has long been regarded as a black box and the crystal quality was only improved through systematic, time-consuming experiments [1]. One of the main goals therefore was to bring “light into the dark” with new window design for the extreme process parameters of larger than 650 °C and simultaneously 300 MPa.

The ammonothermal method involves an upper and lower temperature difference that is provided in an autoclave by a baffle plate, and GaN, which is dissolved in supercritical ammonia in the raw material dissolution region and deposited on a seed crystal in the crystal growth region. Because dissolution and deposition can be continuously performed in an autoclave, this is a suitable method for producing large crystals. The growth of bulk GaN single crystals by the ammonothermal method is currently under development and there remain many indeterminate factors. This chapter describes the technological developments to achieve a high-speed growth and high quality of grown crystals.

Although the crystal growth of GaN under ammonothermal conditions is being performed for quite some years now, the physical processes going on in the autoclave are still debated. Insight in the autoclave by in situ techniques is difficult and numerical simulations are based on physical models where the experimental prove is eventually vague. This chapter reports a period of experimental work leading to a new vision of the basic ammonothermal process. We shortly summarize the 3D thermal- and transport model and a first chemical model. Subsequently, we propose an alternative picture for the ammonobasic crystal growth of GaN, which not only leads to consequences with regard to required pressures and temperatures but also to a potential new growth process. The proposed hypothesis and empirical model involves the presence of a liquid phase in the autoclave in form of an amidogallate complex.

Ammonothermal syntheses require high pressures and temperatures to realize the supercritical conditions. For this a special autoclave technology is required, which places special demands on both the material and the geometrical design. Characteristics are thick walls, high weight, opaque, narrow interior. Therefore the exact determination of initial weights and in particular filling degrees of liquids, which are liquid only under pressure, becomes difficult. In this section, we are going to compare three methods (i) pumping time measurement, (ii) determination by weighing and (iii) ultrasound for adjusting and/or determining the ammonia filling degree of ammonia in an autoclave are presented and the advantages and disadvantages of the individual methods are compared. In addition it is important to have the Ammonia amount very precise to describe the all the concentrations in the autoclave, even under the ammothermal conditions.

The following chapter is mainly aimed at simulators and crystal growers, as viscosity has an influence on the flow behaviour in the reactor and the diffusion coefficient in the crystal’s vicinity. So, the chapter gives an overview of influencing factors to viscosity in ammonothermal media, which are the pressure and temperature as well as the concentration of used mineralizers. Therefore, different possible viscometers are described and discussed in detail for its potential use in ammonothermal media. Hereby two promising options are presented in detail: a modified rolling ball viscometer as well as an adaptation of the ultrasonic pulse-echo method for viscosity measurement for ammonothermal systems. While the last is mostly based on literature research and only some general prove of principle are carried out, the first one is fully described and analysed during operation. This means for the adaptation of this principle four critical aspects have to be overcome. As a result, the viscosity of ammonia in the range above 400 °C up to 600 °C at maximum pressure of 252 MPa is shown. Additionally, some measurements of ammonia-ammonium-fluoride-mixtures are compared with pure ammonia, whereas the viscosity is about 1.4 times lager with ammonium fluoride then without.

In this study, the solubility of GaN in the ammonobasic MNH₂–GaN–NH₃ (M = Na, K) systems in the temperature range of 350–550 °C was studied. The target value for the ammonia filling degree was 45% and 55% for mineralizers NaNH₂ and KNH₂, respectively. The mineralizer concentration was varied at all temperatures from 1 M to 3 M, in respect to the liquid ammonia. It has been found that the solubility of GaN with NaNH₂ as a mineralizer increases slightly with increasing temperature in the range 400–600 °C, but is generally very low. A dependence of solubility on mineralizer concentration was not observed. An increase in the ammonia filling degree from 45% to 55% only slightly increased the solubility of GaN. The use of KNH₂ as a mineralizer brought an enormous increase in the solubility of GaN. Here, 7 times higher solubilities than those for NaNH₂ were observed. A retrograde solubility behavior of GaN in the temperature range 400–550 °C was determined for KNH₂ as a mineralizer. In the range of 350–400 °C the solubility showed a positive temperature dependence. The maximum solubility was reached at 400 °C. The solubility of GaN at 400 °C was found three times higher than at 550 °C.

X-ray based in situ monitoring techniques for ammonothermal processes are reviewed. Technological aspects are discussed, including general aspects of in situ X-ray visualization technology as well as the corrosion resistance of prospective materials for X-ray transparent windows under ammonothermal conditions (for more comprehensive information on corrosion resistance see Chap. 11). In situ X-ray visualization methods have proven to be extremely useful for gaining insights that are inaccessible through other methods to date. Results obtained by in situ X-ray imaging comprise information on solubility and dissolution kinetics as well as insights into the transport of solutes and local concentration changes of solutes. Moreover, phase changes of the fluid can be monitored and the technique has been adapted for solubility studies of novel materials that are unavailable as bulk materials to date. Findings of particular interest include a revision of solubility data for GaN, insights into face-selective and mineralizer-selective dissolution kinetics of GaN, and the visualization of Ga transport within the fluid upon dissolution of GaN. The observed extremely slow motion of Ga-containing species suggests a pronounced influence of solutes on the viscosity of the fluid, which has largely been treated as negligible due to the lack of data for the respective mixtures so far.

In addition to the high pressure and temperature, process equipment for ammonothermal syntheses has to withstand ammonothermal reaction media: an aggressive mixture of supercritical ammonia with basic or acidic additives. Severe corrosive attack of process equipment can occur under ammonothermal conditions but needs to be minimized for two reasons. Firstly, corrosive attack can lead to safety issues that must not be neglected. Secondly, corrosion products represent a major source of impurities, which represents a critical issue especially for the synthesis of semiconductors. Consequently, measures for minimizing corrosion are of utmost importance for the adequate design of ammonothermal equipment. In this chapter, materials suitable for ammonothermal applications, for pressure bearing parts as well as for internal setups and corrosion protection, and recent investigations on their corrosion behavior will be reviewed with suggestions for suitable applications and protection systems.

This chapter provides a brief overview of the synthesis of nitrides and oxonitrides by the ammonothermal method. Numerous binary, ternary and multinary nitrides as well as oxonitrides are discussed. The synthesis conditions with regard to the temperatures, pressures, precursors and mineralizers are mentioned. In addition, the crystal structure of the respective compounds will be briefly described. Since most of these compounds possess interesting electronic and optical properties, the bandgaps of the compounds are discussed in more detail and are summarized at the end.

Mineralizers possess a central relevance in ammonothermal synthesis and formation of soluble species for material transport and crystal growth in particular, governing the solubility, transport direction and deposition processes. In this chapter we review the knowledge on solubilities and chemical behavior of common mineralizers for ammonothermal synthesis. Additionally, we present the current knowledge on intermediates during ammonothermal gallium nitride crystal growth, depending on the nature of the applied mineralizer, as well as during a conceivable ammonothermal synthesis of zinc nitride. Additionally, crystal growth of indium nitride is discussed with focus on chemical processes within the ammonia medium.

Thermodynamic modeling of the ammonia decomposition under ammonothermal conditions (T > 600 K, P > 100 MPa) is presented given recent advances in developing a new, simple equation of state (EOS) describing ammonia, hydrogen, nitrogen and their mixtures under these conditions. The simplified EOS is based on the traditional form of the Beattie-Bridgeman (BB) EOS and expands relevant parameters to second order in density. As a consequence, a seven-parameter EOS is presented and compared to experimental data for ammonia up to 250 MPa and 810 K. Experimental data for ammonia, hydrogen and nitrogen are simulated using the reference multiparameter EOS for each individual gas. Gas mixtures are formed by applying mixing rules with separated contributions for polar and non-polar interactions. The accuracy of the expanded BB EOS is suggested to be 1–2% in pressure for temperatures greater than 700 K for ammonia, hydrogen, nitrogen and NH₃–N₂–H₂ mixtures. Additionally, a general thermodynamic expression for the equilibrium constant is presented and applied to the ammonia decomposition reaction by using non-ideal mixing contributions from the second virial coefficient using the expanded BB EOS. Comparison with experimental data (P <210 MPa, T <810 K) suggests an accuracy of ~2% in pressure.

Molecular Simulations are increasingly entering the realm of materials syntheses. While pioneering studies were bound to simple models which could only address selected aspects of ‘real chemistry’ in the lab, recent advances in simulation methodology and computing hardware indeed paved the way to also modelling complex systems. Yet, we are hardly more than at the beginning of establishing molecular simulations as a routine tool for guiding syntheses. In the present contribution, we discuss the progress that has been made to understand ammonothermal syntheses of nitrides. This encompasses molecular dynamics simulations based on non-reactive force-fields—such as studies of liquid ammonia as a solvent, and its supercritical nature at high temperature and pressure. Moreover, we report on recent work on quantum and hybrid quantum/classical approaches for modelling the auto-protolysis of ammonia and ammonia protolyses in the course of metal ion solvation. This forms a basis for rationalizing the association of ion aggregates, size-induced proton transfer and the self-organization of amides, imides and nitrides from molecular simulations.

This chapter summarizes the present status of bulk GaN crystals grown by ammonothermal basic and acidic methods, and reviews their intrinsic physical properties. Crystals with low dislocation densities, typically well below 10⁵ cm⁻², high crystal lattice flatness, and sharp X-ray rocking curves (typically below 20 arcsec) are reproducibly grown by both methods. High quality strain-free homoepitaxial films have been successfully deposited on both polar and non-polar ammonothermal substrates. Characteristics of testing devices produced using these epitaxial templates confirm the potential of ammonothermal substrates for fabrication of high performance and high yield devices.

As the ammonothermal method often requires technically challenging conditions such as high temperature, high pressure and reaction media that are rather corrosive towards most metals (see Chap. 11), the further development of ammonothermal process equipment is of vital importance for tapping the full potential of the ammonothermal method. Suitable high-pressure/high-temperature autoclaves are therefore the basic requirement for a successful process. Ammonothermal syntheses are conventionally carried out in autoclaves, which do not allow monitoring of the physical and chemical processes during the experiments. Therefore, in situ measurement techniques at process conditions are crucial for obtaining a holistic understanding of ammonothermal processes. The following chapter presents optical cells for in situ monitoring enabling UV/Vis and Raman spectroscopy, X-ray imaging and X-ray diffraction as well as ultrasonic measurements. Additionally, special equipment like a rotatable feedthrough and a ceramic-crucible-based systems for minimization of corrosion and contamination are described.

Even more than fifty years after the first ammonothermal syntheses, the synthetic potential of this technique is still far from established. Even in the already commercialized crystal growth of GaN substrates, various technical obstacles remain, partly because of lack of materials for high pressure equipment sufficiently resistant against the aggressive medium and at the same time persistent at the process conditions, but clearly because of insufficient understanding of the chemical and physical processes in supercritical ammonia. Still, many novel nitrides already emerge from ammonothermal synthesis. This technique has already proven to hold great prospects in crystal growth of the other group III nitrides, AlN and InN, relevant from a technical point of view, and substitution variant among those nitrides with further trivalent ions. Huge potential was also demonstrated for synthesis of further hard to produce nitrides as nitridosilicates, nitridophosphates and similar compounds, or oxide nitride perovskites. With increasing understanding of the physiochemical background and concomitant extension of the accessible process parameters we definitely will see great advances in the ammonothermal synthesis of novel materials for future applications.