Implant isolation of both n-type InP and InGaAs by iron irradiation: Effect of post-implant annealing tem - Electron Devices for Microwave and Optoelectronic Applications, 2003. EDMO 2003. The 11th IEEE Int

1 MeV Fe' was implanted into n-type InP and InGaAs layers at different substrate temperatures, -196'C, room temperature (RT), 100°C and 2OO0C to obtain highresistivity regions. The sheet resistivity of the InP and InGaAs epilayers grown on semiinsulating (SI) InP substrates was measured as a function of substrate temperature and post-implantation annealing temperature (100 SOO'C). For InP, a maximum sheet resistivity of -lx107 W O was achieved for samples implanted at -196OC, RT and 100°C after annealing at 400'C. For InGaAs samples, a maximum sheet resistivity of M O 7 and 2 . 3 ~ 1 0 ~ WO is obtained for -196'C and RT implants respectively after annealing at 650'C for 60s. In both InP and InGaAs, the isolated regions exhibit good stability to heat treatment up to 5OO' C for all cases irrespective of the irradiation temperature. The iron depth profile obtained by secondary ion mass spectrometry (SIMS) shows that iron does not diffuse up to an annealing temperature of 550'C in both InP and InGaAs for all implantation temperatures. These results are novel since high sheet resistivity (-5x106 N O ) is obtained in both InP and InGaAs samples implanted at -196' and RT, and annealed at 400'C. These data demonstrate the potential usefulness of iron implantation for isolation of InPDnGaAs based devices. Introduction In hulk InP and InGaAs materials, highly resistive behaviour is usually obtained by doping the crystal with Fe during the growth. Fe atoms occupy In sites and act as deep acceptor centres for free electrons [l]. In epitaxial materials, the achievement of SI behaviour in selected areas is of crucial importance for lateral device isolation in integrated circuits and for the realisation of laser devices with current blocking regions beside the active region [2]. Thus ion implantation is an obvious candidate to selectively obtain SI regions in InP and InGaAs. With the increasing use of InGaAs and InP layers in heterojunction devices, there is a critical need to understand the behaviour of these materials upon ion bombardment for isolation purposes. The resistive behaviour can be due to the damage related levels introduced with the implant and/or to chemical compensation due to the implanted species (typically Fe or other transition metals) [3]. In nt InP and especially n+ InGaAs, very low sheet resistivity (-IO4 WO) was reported by several authors after implantation of light ion species such as proton, helium, boron and nitrogen [4,5,6]. Isolation was mainly caused by a damage-induced compensation mechanism. In an effolt to obtain higher sheet resistivities in n-type materials, we have *Fax : +44 1483689404, E-mail: [email protected] 0-7803-7904-7/03/$17.00 02003 IEEE. 18 Authorized licensed use limited to: University of Surrey. Downloaded on March 23,2010 at 06:44:09 EDT from IEEE Xplore. Restrictions apply. investigated the implantation of Fe, an impurity which is known to result in highresistivity InP when used as a dopant during the growth of bulk crystals [7]. Many authors in the past reported on anomalous behaviour of iron during annealing [8,9,10]. It seems to diffuse rapidly within the implanted layer gettering at the surface and at the damaged regions. These authors claim that high resistivity behaviour cannot be achieved by Fe implantation on initially n' doped layers. The formation of good and thermally stable electrical isolation in Si-doped InP and InGaAs layers using 1 MeV iron at different substrate temperatures is studied in this work. Experimental Procedure Semi-insulating Fe-doped InP wafers of (100) orientation were used as substrates for the growth of both n-type InGaAs and InP epilayen, with the (100) axis 2' off normal orientation, using a Solid Source Molecular Beam Epitaxy (SSMBE) reactor. The n-type epilayers were doped with silicon with a concentration and thickness of 1 ~ 1 0 ' ~ cm" and lpm respectively. The wafers were cleaved to obtain several samples of approximately 1 cmz for the prepamtion of the resistors. All samples were cleaned in organic solvents and the clover-leaf pattem was printed on them using optical lithography. The area of InP and InGaAs samples not covered with photoresist was etched to a depth of approximately 3pm using standard etching solutions. The photoresist was then removed in acetone leaving the cloverleaf Hall pattern on the samples. The samples were divided into four different groups with implant isolation at temperatures of -196OC, RT, 100°C;and 200°C using a 2MV High Voltage Engineering Europa (HVEE) implanter. For -196'C, 100nC and 2OO0C implants, ;the.samples were mounted on a special temperature control stage. The centre of the Hall pattem for all the samples was irradiated with Fe' using a dose and ener y of ~ x I O ' ~ cm-2 and lMeV respectively, with a beam current density < 0.33 pVcm , The post-implant annealing was performed in ' the range 100°C 800°C ( W C ) for a time of 60s in a nitrogen atmosphere, following a ramp up time of 60s to temperature. Ohmic contacts to the samples were fabricated by applying indium and sintering at approximately 2OO0C for 2min. The sheet resistivity, was measured using a Bio-Rad HL.5500 Hall effect system employing Van der Pauw geometry at 300K under a magnetic field strength of 0.32 T. All measurements were done at RT for samples implanted at 77K, 100nC and 2OO0C. The Fe depth profiles of the implant isolated samples were obtained by SIMS using a CAMECA IMS6f instrument. It was used to follow the diffusion behaviour of iron at different annealing temperatures. A 12.5keVO; primary beam was used to monitor Fe in both InP ,and InGaAs, analysing the 50pn diameter central part of a 150pn xl50pmcrater. Results and Discussion The projected range of .lMeV iron in InP and InGaAs is about 0.6pm and 0.4pm respectively. The energy of lMeV for the iron beam is chosen to place most of the iron atoms well inside the doped layer. In this way, the chemical compensation will be more effective for the electrical isolation of the InP and InGaAs epilayers. The effect of postimplant annealing' temperature on n-type InP samples for the four different substrate temperatures is shown in figure 1. The initial sheet resistivity of the n-type InP layer for all the samples is -15 WD. After iron implantation, an as-implanted sheet resistivity of -5x106 WO is obtained for substrate temperatures of -196'C, RT, and 100°C and that of 4