The applicability of quenched phosphorescence as a detection mode in capillary electrophoresis (CE) was explored for a number of analyte classes and buffer systems. The detection method is based on the quenching of biacetyl phosphorescence (biacetyl is a constituent of the CE buffer) by the analytes via various mechanisms (energy transfer, electron transfer and, possibly, hydrogen donation) and gives rise to negative peaks in the electropherograms. A number of buffers in the pH range 4.7-11.5, frequently used in CE, were tested for their compatibility with this detection mode. Berate, succinate, malonate, acetate, and phosphate buffers (pH 4.7-8.5) could be used without any problems. With a pH of ca. 8.5 or higher the baseline declined with time, while at a pH higher than 9.5 no signal at all was obtained. Obviously, the noise on the phosphorescence signal (i.e., the baseline) determines the ultimate analyte detection limits (LODs), The baseline signal-to-noise ratio, usually denoted as the dynamic reserve (DR), was enhanced ca. 25-fold compared to direct biacetyl excitation by sensitization of the biacetyl phosphorescence by 1,5-naphthalenedisulfonic acid, and by application of a total emission mirror (TEM). A concentration of 1 x 10(-3) M 1,5-naphthalenedisulfonic acid was found to be optimal. For the buffer systems considered, the DR was typically ca. 300-600 under optimized conditions (noise defined as 1 x sigma). Investigated analytes include naphthalenesulfonic acids (NS), nitrophenols, hydroxybenzoic acids, amino acids, and dithiocarbamates (DTCs.). For most of these, the LODs were in the 10(-7)-10(-8) M range, which is significantly lower than with direct or indirect absorption detection.