VOLUME 77, NUMBER 16 PHYSICAL REVIEW LETTERS 14 OCTOBER 1996 Identification of the Native Vacancy Defects in Both Sublattices of ZnS x Se 12x by Positron Annihilation K. Saarinen, T. Laine, K. Skog, J. Mäkinen, and P. Hautojärvi Laboratory of Physics, Helsinki University of Technology, 02150 Espoo, Finland K. Rakennus, P. Uusimaa, A. Salokatve, and M. Pessa Department of Physics, Tampere University of Technology, P.O. Box 692, 33101 Tampere, Finland (Received 15 February 1996) We show how positron annihilation can distinguish vacancies in the different sublattices of a binary compound by performing experiments in ZnS x Se 12x layers. We identify the Se vacancies V Se  in N-doped and the Zn vacancies V Zn  in Cl-doped material by the shape of the core electron momentum distribution. The charge of the defect involving V Se is neutral or negative in p-type ZnS x Se 12x , suggesting that V Se is complexed with an acceptor. The concentration of the V Se complexes is high $10 18 cm 23 , indicating that their role is important in the electrical compensation of p-type ZnS x Se 12x . [S0031-9007(96)01393-2] PACS numbers: 71.55.Gs, 73.61.Ga, 78.70.Bj ZnSe has attracted strong interest as a potential mate- rial for blue-green diode lasers, since its band gap is wide (2.67 eV) and its lattice constant matches closely to that of GaAs substrates. However, fundamental problems ex- ist in the doping: ZnSe can be easily doped n type but the fabrication of p-type material is still difficult. With nitro- gen doping the active acceptor concentration saturates at about 10 18 cm 23 , although more than 10 19 cm 23 N atoms have been incorporated during the growth. In fact, simi- lar doping problems have been detected in many wide- band-gap sulfides and selenides, which suggests that these phenomena may have a common microscopic nature. Several theoretical models have been proposed for the origin of the doping problems. The simplest explanation is that native defects compensate the doping atoms. Theoretical calculations have suggested, however, that the concentration of spontaneously formed point defects is too low to cause substantial compensation [1]. On the other hand, the recent calculations of Garcia and Northrup [2] show that a high concentration of defect complexes involving the N dopant and donor defects can be formed in the lattice. Another kind of explanation for the doping problems is the existence of a limit for the solubility of impurities [3]. It has also been proposed that large atomic displacements near the dopant atoms can play an important role in passivation [4,5]. These models include the formation of defects analogical to the DX center in AlGaAs [6] and the recently proposed lattice instability leading to two broken bonds in the proximity of the acceptor atom [7]. However, very little experimental data exist on the microscopic origin of the compensation in ZnSe. In this work we apply positron annihilation spec- troscopy to study the vacancy defects in ZnS x Se 12x and their role in the compensation. We show how the sub- lattices of the vacancies can be directly identified in the experiments by the shape of the core electron momentum distribution. We detect Se vacancies in N-doped and Zn vacancies in Cl-doped ZnS x Se 12x . The charge state of the Se vacancy is neutral or negative in p-type ZnS x Se 12x , suggesting that it is complexed with an acceptor, possibly with the nitrogen dopant. The vacancy concentration in heavily N-doped ZnS x Se 12x is at least 10 18 cm 23 , which indicates that the Se vacancy can play an important role in the electrical compensation of p-type ZnS x Se 12x . The samples in this work were 2 mm ZnS 0.06 Se 0.94 over- layers grown by molecular beam epitaxy (MBE) as lat- tice matched on a GaAs substrate. The alloy composition x  0.06 varied less than 1% from one sample to another according to the x-ray diffraction experiments. A ZnCl 2 effusion cell was used for Cl doping and a N 2 plasma source for the N doping. The impurity concentrations were determined by secondary ion mass spectrometry (SIMS). The N concentrations varied in the range of N  1 30 3 10 18 cm 23 in the six studied samples, depend- ing on the power of the N 2 plasma source. The elec- trochemical capacitance-voltage profiling (ECV) showed that the hole concentration is at maximum 2 3 10 17 cm 23 , thus indicating that most of the nitrogen acceptors are electrically inactive. In the two Cl-doped samples both the n-type carrier and the Cl concentrations were roughly 10 18 cm 23 according to ECV and SIMS experiments. The positron experiments were performed using a positron beam at the energy of 15 keV corresponding to a mean positron stopping depth of 0.5 mm. This energy was chosen so that all positrons annihilate in the ZnS x Se 12x overlayer. The Doppler broadened shape of the 511 keV annihilation radiation was measured with a Ge detector and described with the conventional valence and core annihilation parameters S and W [8]. The S parameter represents the fraction of positrons annihilating mainly with the valence electrons with a longitudinal momentum component of p L # 3.7 3 10 23 m 0 c, where m 0 is the electron mass and c the speed of light. The 0031-9007 96 77(16) 3407(4)$10.00 © 1996 The American Physical Society 3407