Ziel der vorliegenden Arbeit war die Herstellung, der Einsatz und die Bewertung von Durchfluss-Photoreaktoren für online Prüfungen zur photochemischen Stabilität im HPLC-System (hn- HPLC). Neben der Anfertigung eines für dieses Anforderungsprofil geeigneten Reaktors war es von Interesse, Strahlungsintensitäten zu bestimmen, die Photostabilität von UV-Filtern zu bewerten, sowie analytische Verfahren zu optimieren. Als wichtigste Ergebnisse dieser Arbeit sind: (1) erstmalig aktinometrische Messungen im Durchfluss durchgeführt worden, (2) die eingestrahlten Intensität im Durchfluss-Reaktor mit dem Standard-Sonnenspektrum (COLIPA) verglichen worden, (3) die Bestimmung von 4-Aminobenzoesäure (PABA) und 4-Acetamidobenzoesäure (PAABA) mittels HPLC-hn optimiert und (4) photochemische Umsätze an folgenden UV-Filtern festgestellt worden: 4-Methoxy-zimtsäure-2-ethyl-hexylester, 3-(4'-Methyl)benzyliden-bornan-2-on,1-(4-tert.-Butylphenyl)-3-(4-methoxyphenyl)propan-1.3-dion und 4-Aminobenzoesäure. / The aim of the present thesis was the construction and evaluation of an on-line photochemical reactor for photochemical stability tests in an HPLC-system (hn- HPLC). The work was focused on the production of a functional and effective photochemical reactor, the determination of radiation intensities and the optimisation of analytical methods. Furthermore the stability of UV-filters under exposure to UV-radiation was investigated. The main results of these thesis are presented in the following: (1) the determination of radiation intensities in online photochemical reactors, (2) the comparison of spectra with standard-sun spectra (COLIPA), (3) the optimisation of the determination of 4-aminobenzoic acid (PABA) and 4-acetamidobenzoic acid (PAABA) with an HPLC-hn method (4) the characterisation of photochemical changes of UV-filters: p-methoxy-cinnamic acid-2-ethyl-hexyl ester, 3-(4'-methyl)benzylidene-bornan-2-on,1-(4-tert.-butylphenyl)-3-(4-methoxyphenyl)propan-1,3-dion and p-aminobenzoic acid.

The Ph.D. project compiled in this thesis has focused on the role of the solvent in solid-liquid phase equilibria and in nucleation kinetics. Six organic substances have been selected as model compounds, viz. ortho-, meta- and para-hydroxybenzoic acid, salicylamide, meta- and para-aminobenzoic acid. The different types of crystal phases of these compounds have been explored, and their respective solid-state properties have been determined experimentally. The solubility of these crystal phases has been determined in various solvents between 10 and 50 oC. The kinetics of nucleation has been investigated for salicylamide by measuring the metastable zone width, in five different solvents under different experimental conditions. A total of 15 different crystal phases were identified among the six model compounds. Only one crystal form was found for the ortho-substituted compounds, whereas the meta-isomeric compounds crystallized as two unsolvated polymorphs. The para-substituted isomers crystallized as two unsolvated polymorphs and as several solvates in different solvents. It was discovered that the molar solubility of the different crystal phases was linked to the temperature dependence of solubility. In general, a greater molar solubility corresponds to a smaller temperature dependence of solubility. The generality of this relation for organic compounds was investigated using a test set of 41 organic solutes comprising a total of 115 solubility curves. A semi-empirical solubility model was developed based on how thermodynamic properties relate to concentration and temperature. The model was fitted to the 115 solubility curves and used to predict the temperature dependence of solubility. The model allows for entire solubility curves to be constructed in new solvents based on the melting properties of the solute and the solubility in that solvent at a single temperature. Based on the test set comprising the 115 solubility curves it was also found that the melting temperature of the solute can readily be predicted from solubility data in organic solvents. The activity of the solid phase (or ideal solubility) of four of the investigated crystal phases was determined within a rigorous thermodynamic framework, by combining experimental data at the melting temperature and solubility in different solvents and temperatures. The results show that the assumptions normally used in the literature to determine the activity of the solid phase may give rise to errors up to a factor of 12. An extensive variation in the metastable zone width of salicylamide was obtained during repeated experiments performed under identical experimental conditions. Only small or negligible effects on the onset of nucleation were observed by changing the saturation temperature or increasing the solution volume. The onset of nucleation was instead considerably influenced by different cooling rates and different solvents. A correlation was found between the supersaturation ratio at the average onset of nucleation and the viscosity of the solvent divided by the solubility of the solute. The trends suggest that an increased molecular mobility and a higher concentration of the solute reduce the metastable zone width of salicylamide. / QC 20100831