In situ-Messungen von thermodynamischen Zustandsgrößen im ammonothermalen Prozess

The ammonothermal methode is an interesting process for the crystallisation of vari-ous products, especially for Nitrides, i.g. Gallium Nitride. In this work, three topics for the measurement of thermodynamic variables in the ammonothermal process were dealt with. Firstly, solubility measurements under process conditions using an adapted autoclave with the first-time use of a heated high-pressure high-temperature valve for ammonothermal use. Secondly, theoretical considerations on the investigation of the viscosity of ammonothermal media, which led to the first experimental determination of the viscosity of a NaN3/NH3-mixture. Finally, spectro-scopic in situ measurements were used to gain an interesting insight into the influence of filling methods, the dissociation of NH3 under ammonothermal conditions and the intermediates of ammonobasic mixtures. The most important findings of these three topics are briefly summarized here. The GaN/NH4F/NH3 mixture was investigated in detail ex situ as an experimental medium, i.e. a sample quantity was measured before and after the experiment and conclusions were drawn about the conditions during the experiment. According to the literature, time dependence, pressure dependence and temperature dependence were identified as critical influencing factors. In general, the solubility of the GaN/NH4F/NH3 mixture is in the range of the GaN/NH4Cl/NH3 mixture, a pressure dependence was only observed up to 53 MPa and the time for equilibrium was observed after 34 hours. Of particular interest is the identification of the retrograde sol-ubility behaviour from 530 °C, which is known from ammonobasic solutions at low temperatures and was postulated by Jiang et al [56]. Due to the as yet unresolved corrosion problem, the experiments were carried out in an autoclave that was not coated nor had some kind of liner. Corresponding corrosion tests have shown, how-ever, that the corrosion in the GaN/NH4F/NH3 mixture is significantly lower than that of other ammonoacid mixtures and that the solubility tests are therefore meaningful. Due to the disadvantages of the ex situ measurements, a spectroscopic experiment was carried out to narrow down the pressure and temperature ranges when using NH4F as a mineralizer in the ammonothermal process. This showed a change in the chemical composition from 538 °C, so that the conclusion was drawn that crystalliza-tion experiments only make sense above this temperature. This experiment is partic-ularly important with regard to further investigations of new nitrides in the ammono-thermal process. With this method, the temperature and pressure ranges can be nar-rowed down with short screening tests prior to intensive series of measurements. For comparison, an investigation with this method takes about two days, the ex situ measurement series requires up to 16 tests with test times between 34 h and 72 h per test. A central topic of this work was the extraction of samples from ammonothermal mix-tures under high temperature and pressure conditions. This type of sampling auto-clave has not yet been realized. One step in this direction is the initial testing of a high-pressure, high-temperature valve. The first successful tests for sampling an ammonothermal mixture under synthesis conditions have shown that sampling is possible with this setup. At the same time, however, there are limits to this set-up. As a conclusion, a combination of an autoclave and a viewing cell was presented. The analysis of viscosity in the ammonothermal process is crucial for a successful simulation of crystallization due to its influence on the flow behaviour in the reactor and the diffusion coefficient in the environment of the crystal. The ultrasonic-pulse-echo viscometer and rolling-ball viscometer were identified as possible variants and adapted to the conditions of ammonothermal synthesis. The ultrasonic-pulse-echo viscometer remains at the status of a study based largely on literature research. The general basic functions were used in situ for further experiments together with Baser et al. and are therefore available. For the rolling ball viscometer, four critical aspects for the high-pressure, high-temperature application are analysed and overcome step by step. As a result, the viscosity of ammonia in the range above 400 to 600 °C at a maximum pressure of 252 MPa is presented. In addition, some measurements of ammonia-ammonium fluoride mixtures are compared with pure ammonia, where the viscosity with ammonium fluoride is about 1.4 times higher than without. However, it has not yet been possible to establish a generally valid formula as described for salts in aqueous solution. Optical investigations into ammonothermal synthesis have already been successfully carried out by Alt [24], the knowledge acquired was therefore further expanded. Firstly, various methods of filling the autoclaves with nitrogen and ammonia are in-vestigated. The order of filling should a priori have no influence on the subsequent experiment, which was disproved by Raman measurements. Raman measurements were also used to investigate the influence of increasing temperatures on a known mixture. This allowed determination of the composition of unknown mixtures of ammonia and nitrogen at temperatures between 400 and 550 °C. This knowledge was used to analyse the dissociation of ammonia using Raman spectroscopy. From the recorded spectra, it can be concluded that the addition of NaN3 and GaN has no effect on the decomposition of NH3. At the same time, the experiments show that the decomposition of the ammonia stopped long before the equilibrium concentration was reached. Since the decomposition requires a catalyst, which is abundantly available due to the nickel in the autoclave material, this is ex-plained by an inhibition of the catalyst by the nitrification. Finally, intermediates are analysed in situ. The experiments are carried out in an ammonobasic environment, with caesium amide as mineralizer. The detection limit of Cs[Ga(NH2)4] is not yet exceeded at a concentration that is similar to the crystallisation condition. So for fur-ther experiments with caesium, already synthesized Cs[Ga(NH2)4] is used. Video-optical images and spectral analysis show for the first time that there is no longer a two-phase mixture above 450 °C. The intermediate, on the other hand, is still detected in the single-phase region above 450 °C. Many of the measurement methods presented here are still in their infancy, but many new insights into ammonothermal synthesis have already been gained. Especially with regard to the synthesis of new nitrides and the time-consuming search for suitable crystallization conditions, new helpful tools could be provided in this work.