Materials Science and Engineering B50 (1997) 148 – 152 III – V – N compounds for infrared applications J. Salzman a, *, H. Temkin b a Department of Electrical Engineering, Solid State Institute and Microelectronics Center, Technion, The Israel Institute of Technology, Haifa 32000, Israel b Electrical Engineering Department, Texas Tech Uniersity, Lubbock, TX 79409, USA Abstract III – V alloys containing nitrogen and As or P are potential candidate materials for infrared applications. The most studied material in this system is GaAs 1 -x N x . After reviewing the early experiments, and some theoretical predictions, we describe growth experiments by metalorganic molecular beam epitaxy, in which highly crystalline, single phase material was obtained for x 0.18. Room temperature photoluminescence was measured for layers with x =0.73%. The GaAsN alloys seem to exhibit a composition dependent bowing parameter. © 1997 Elsevier Science S.A. Keywords: Infrared applications; Metalorganic molecular beam epitaxy; Bowing parameter 1. Introduction Most of the research on III–N alloys has been, thus, far directed towards the wide bandgap materials GaN, InN, AlN, and their alloys [1]. These III–N compounds can also be alloyed with the arsenide and phosphide III–V materials, to form zinc blende III–V–N ternary alloys such as GaAsN, InAsN, AlAsN, GaPN, InPN, AlPN, and quaternaries, such as GaInAsN, GaAlAsN, etc. [2]. Since the lattice constant of the wide bandgap III–N compounds in their cubic phase is 5A ˚ and the lattice constant of the conventional III – V compounds is 5.45 A ˚ , III–V–N alloys can, in principle, be synthesized lattice matched to Si (a =5.4301 A ˚ ). The possibility to produce a direct bandgap material epitax- ially grown on Si, the dominant electronic material, with lattice match between them, attracts considerable interest as a potential platform for photonic devices on Si substrates (J. Salzman, J. Ballantyne; unpublished). Additional motivations for the study of III–V–N’s, are the potential realization of ohmic contacts to p-type GaN [3], and the synthesis of In-containing long wave- length devices lattice matched to GaAs [4]. In the early experiments of Weyers et al. [5], and of Kondow et al. [6], a considerable bandgap reduction was observed for GaAs 1 -x N x with small nitrogen frac- tion (x 0.02). The large difference in lattice constant between GaAs and GaN ( 20%), and the anomalous bandgap reduction in the dilute alloy motivated theo- retical interest in the consequences of lattice mismatch on alloy properties such as bandgap bowing, phase stability, and ordering [7 – 10]. Previously, the synthesis of GaAsN on GaAs and on GaP was reported [5,6,11]. Also InAsN lattice matched to GaAs [12], and GaPN [13] were attempted, but the experimental study of III–V–N alloys is still in its early stages, with most of the composition range (As/N ratio, and P/N ratio) largely unexplored. Here, we review some of the main issues in GaAsN research, and we describe systematic experiments on the growth of high quality GaAs 1 -x N x by metalorganic molecular beam epitaxy (MOMBE). We find that the nitrogen-to-ar- senic flux ratio must be carefully controlled to assure growth of single phase GaAs 1 -x N x alloys with signifi- cant nitrogen content. The alloy bandgap exhibits a composition-dependent bowing parameter, with an esti- mated (extrapolated) bandgap for GaAs 0.8 N 0.2 (lattice matched to Si) of 0.9–1.0 eV. 2. The GaAs–GaN system The bandgap of a ternary alloy AB 1 -x C x can be in most cases written in terms of that of the binary constituents E g (AB) and E g (AC) as: * Corresponding author. 0921-5107/97/$17.00 © 1997 Elsevier Science S.A. All rights reserved. PII S09 21- 5 1 07(97)00 1 53 - 0