Practical aspects of fast reversed-phase high-performance liquid chromatography using 3 microm particle packed columns and monolithic columns in pharmaceutical development and production working under current good manufacturing practice

The potential and limitations of fast reversed-phase high-performance liquid chromatographic separations for assay and purity of drug substances and drug products were investigated in the pharmaceutical industry working under current good manufacturing practice using particle packed columns and monolithic columns. On particle packed columns, the pressure limitation of commercially available HPLC systems was found to be the limiting factor for fast separations. On 3 microm particle packed columns, HPLC run times (run to run) for assay and purity of pharmaceutical products of 20 min could be achieved. As an interesting alternative, monolithic columns were investigated. Monolithic columns can be operated at much higher flow rates, thus allowing for much shorter run times compared to particle packed columns. Compared to particle packed columns, the analysis time could be reduced by a factor up to 6. However, some compounds investigated showed a dramatic loss of efficiency at higher flow rates. This phenomenon was observed for some larger molecules supporting the theory that mass transfer is critical for applications on monolithic columns. At flow rates above 3 ml/min some HPLC instruments showed a dramatic increase in noise, making quantifications at low levels impossible. For very fast separations on monolithic columns, the maximum data acquisition rate of the detector is the limiting factor.     

A vision architecture for unconstrained and incremental learning of multiple categories

We present an integrated vision architecture capable of incrementally learning several visual categories based on natural hand-held objects. Additionally we focus on interactive learning, which requires real-time image processing methods and a fast learning algorithm. The overall system is composed of a figure-ground segregation part, several feature extraction methods and a life-long learning approach combining incremental learning with category-specific feature selection. In contrast to most visual categorization approaches, where typically each view is assigned to a single category, we allow labeling with an arbitrary number of shape and color categories. We also impose no restrictions on the viewing angle of presented objects, relaxing the common constraint on canonical views.     

Facile ring opening of siloxy cyclopropanes by photoinduced electron transfer. A new way to β-keto radicals

Siloxy cyclopropanes undergo ring opening and fragmentation of formal silyl cations under formation of β-keto radicals. These reactive intermediates can be used in inter- and intramolecular addition reactions leading to complex ring systems if more than one unsaturated side chain is present in the starting material. Beside some synthetic examples mainly the mechanism will be discussed focusing on the structure of the primarily formed radical cations, the regioselectivity of cyclopropane cleavage (endo vs exo ring opening), leaving of the silyl group, and termination by H-transfer.     

Structural phase transition in the 2D spin dimer compound SrCu 2 ( BO 3 ) 2

A displacive, 2nd order structural phase transition at T _s = 395 K from space group I 2 m below T _s to I 4/m c m above T _s has been discovered in the two-dimensional spin dimer compound SrCu₂(BO₃)₂. The temperature evolution of the structure in both phases has been studied by X-ray diffraction and Raman scattering, supplemented by differential scanning calorimetry and SQUID magnetometry. The implications of this transition and of the observed phonon anomalies in Raman scattering for spin-phonon and interlayer coupling in this quantum spin system will be discussed. Received 24 July 2000 and Received in final form 2 November 2000     

Light-matter interactions via the exact factorization approach

The exact factorization approach, originally developed for electron-nuclear dynamics, is extended to light-matter interactions within the dipole approximation. This allows for a Schrödinger equation for the photonic wavefunction, in which the potential contains exactly the effects on the photon field of its coupling to matter. We illustrate the formalism and potential for a two-level system representing the matter, coupled to an infinite number of photon modes in the Wigner-Weisskopf approximation, as well as to a single mode with various coupling strengths. Significant differences are found with the potential used in conventional approaches, especially for strong couplings. We discuss how our exact factorization approach for light-matter interactions can be used as a guideline to develop semiclassical trajectory methods for efficient simulations of light-matter dynamics.