In general the chemical composition of multicomponent alloy surfaces exhibits significant deviations from the bulk composition due to thermally activated segregation processes. Upon cosegregation epitaxially stabilized two-dimensional surface compounds are formed on substrate surfaces of suitable orientation, e.g. CrN on Fe-15%Cr-N(100) and CrC on Fe-15%Cr-C(100) . The important role of the epitaxial stabilization manifests itself by the fact that on Fe-15%Cr-C(100) crystals the formation of cubic CrC surface precipitates is possible upon nonequilibrium cosegregation of the constituent components. Such a chromium carbide with rocksalt structure is not stable in the bulk of bcc Fe-Cr-C alloys . Angle-resolved photoemission of Cr 2 p 3/2, N 1s and C 1s core level photoelectrons is used to determine the structure of the CrN surface compound and of the CrC surface precipitate on a Fe-15%Cr-C,N(100) single crystal. Both surface phases exhibit a sharp 1×1 LEED pattern. Polar angle intensity distributions of photoelectrons (x-ray photoelectron diffraction) are recorded in the [001] and [011] azimuths, respectively. For CrN the angular intensity distributions exhibit forward scattering peaks for Cr 2 p 3/2 but not for N 1s. Most likely, the CrN phase consists of a CrN compound layer with significant N outward relaxation of about 0.6 Å and a second completed Cr layer. Single scattering cluster (SSC) calculations indicate that the Cr-Cr interlayer distance is expanded by about 26% with respect to the bcc lattice. For CrC , on the other hand, Cr 2 p 3/2 and C 1s show both strong forward scattering features in all distributions. The CrC phase consists of at least three complete compound layers with NaCl structure.