Combinatorial synthesis and screening of very large numbers of organic compounds has been widely applied in the pharmaceutical industry for drug discovery(1). Recently, combinatorial arrays of inorganic materials with known or potential superconductivity(2) and giant magnetoresistance(3) have been synthesized and screened. The combinatorial approach is particularly well suited to ternary and higher-order inorganic materials, for which efforts to predict basic properties have been unsuccessful(4). Here we describe an automated combinatorial method for synthesizing and characterizing thin-film libraries of up to 25,000 different materials, on a three-inch-diameter substrate, as candidates for new phosphors. The discovery and development of new compounds for ultraviolet-excited phosphors is of great importance for the development of flat-panel displays(5) and lighting(6). As there are no reliable theories to predict the relation between composition and phosphor colour and efficiency, the less than 100 useful commercial phosphor materials have been discovered through one-by-one serial synthesis and testing(7,8). Our approach, in contrast, offers rapid screening of many compositions, and it has enabled us to identify a new red phosphor, Y0.845Al0.070La0.060Eu0.025VO4, which has a quantum efficiency comparable or superior to those of existing commercial red phosphors.