10:00AM, C1
"Reflectance-Difference Spectroscopy of GaAs(100) Surfaces:" J.R. CREIGHTON, C.M. Truong, Department 1126, MS 0601, Sandia National Laboratory, Albuquerque, NM 87185-0601
One of the critical issues of III-V MOCVD is the nature of the surface during deposition conditions. Aspnes and others have shown the utility of reflectance-difference spectroscopy (RDS) for examining compound semiconductor surfaces in situ during MOCVD and MBE. For many conditions the surface observed during MOCVD appears to have the same reconstruction (or at least the same RD spectrum) as that observed during MBE. However, RDS was "benchmarked" in an MBE system only for the arsenic and gallium terminated surfaces, hence the effect of adsorbates such as methyl groups or hydrogen was not determined. We also note that there are MOCVD conditions where new RDS lineshapes occurthat were not seen in the original MBE "benchmarking" or "fingerprinting" experiments. We have added the RDS capability to our UHV/surface science system in order to measure the spectra for well-defined adsorbate-covered GaAs surfaces. The RD spectra for the (1X2)-CH3 and (1X4)-CH2 terminated Ga-rich surfaces were measured and found to be unique, i.e. they do not look like the RD spectra of the standard Ga-rich and As-rich reconstructions. There are similarities, however, as both CH3 and CH2 terminated surfaces exhibit a positive going RD peak at ~2.6-2.7 eV, which is also seen for the (2X4) As-rich surface. In fact, the lineshapes for the CH3 and CH2 terminated surfaces show many similarities to the RDS results Aspnes et al. obtained for ALE conditions during the TMGa cycle.
We have also found that the unusual RD lineshape seen by Aspnes (referred to as [[sigma]]) and Richter (referred to as "type II") during MOCVD is actually another type of "super" As-rich surface. Normally, the c(4X4) "super" As-rich surface is characterized by a negative RD peak at ~2.6 eV. Under many MOCVD conditions most of the RD lineshape between 1.5-5.5 eV is consistent with the c(4X4) lineshape measured during MBE. For other conditions, generally lower temperatures, the negative going peak appears at ~2.1-2.2 eV. Nothing like this was measured during the original MBE "benchmarking" work by Aspnes et al. We can generate a very similar RD lineshape by creating a "super" As-rich surface at 350-450deg.C with either arsine or As4. This surface often exhibits a reasonable quality c(4X4) LEED pattern, although some additional "V"-shaped streaks are usually observed. Higher temperature exposures to arsine or arsenic, followed by quenching, usually produces the traditional c(4X4) RD lineshape and the c(4X4) LEED pattern appears to be a little sharper with no "V"-shaped streaks. TPD of the two types of "super" As-rich surfaces have gross similarities, but the distribution of arsenic in the two binding sites characteristic of the c(4X4) surface is different. Further characterization of these two types of "super" As-rich surfaces is in progress.
*This work was supported by the US DOE under Contract DE-AC04-94AL8500
10:20AM, C2
"InP Grown by CBE Using Uncracked Tertiarybutylbisdimethylaminophosphine:" H.H. RYU, C.W. Hill, L.P. Sadwick, G.B. Stringfellow, College of Engineering, The University of Utah, Salt Lake City, UT 84112; T. Groshens, R.W. Gedridge, Jr., Naval Air Warfare Center Weapons Division, Chemical and Materials Branch, Code 474220D, China Lake, CA 93555; R.T. Lareau, Electronic & Power Source Director, Army Research Laboratories, Fort Monmouth, NJ 07703
We report the growth of indium phosphide (InP) by the chemical beam epitaxy (CBE) technique using a new phosphorus (P) precursor, tertiarybutylbisdimethylamino-phosphine (TBBDMAP). TBBDMAP was specifically designed for use in the CBE growth of P-containing compounds. A potentially important advantage of TBBDMAP compared to other P precursors is the ability to grow InP without thermally precracking the P precursor. To the best of our knowledge this is the first report of the CBE growth of single crystalline layers of InP without precracking the group P source. The indium source used in this work was ethyldimethylindium (EDMIn). The T50 of TBBDMAP is approximately 425deg.C. InP was successfully grown without P precursor precracking for substrate temperatures in the range of 450 to 520deg.C. The morphology was mirror-like at substrate temperatures above 450deg.C. Strong bound exciton (BE) photoluminescence (PL) was observed with only a relatively small donor-acceptorpair (DAP) impurity peak present. Values of PL full width at half maximum (FWHM) for InP grown with uncracked TBBDMAP were typically 10 meV at 16K. InP grown with cracked TBBDMAP had PL FWHM values as low as 6 meV at 16K but had a significant DAP impurity peak compared to InP grown with uncracked TBBDMAP.
The results of a systematic study of the effects of growth temperature and V/III ratio for InP grown using both uncracked and cracked TBBDMAP will be presented. Quantitative results to be presented include Hall, secondary ion mass spectroscopy (SIMS), and PL measurements on InP grown using both uncracked and cracked TBBDMAP.
A discussion of pyrolysis studies of TBBDMAP on quartz, InP, and GaP surfaces will be presented in the context of understanding the growth of InP, GaP, and GaInP using TBBDMAP.
10:40AM, C3
"Flow Modulation Growth and Tellurium Doping of Al0.48In0.52As/InP for InP Based HEMTs:" D.T. EMERSON, V.A. Williams, J.R. Shealy, J.B. Shealty*, OMVPE Facility, School of Electrical Engineering, 20 Thornwood Dr., Cornell University, Ithaca, NY 14850; *Hughes Research Laboratories, 3011 Malibu Canyon Rd., Malibu, CA 90265-4799
11:00AM, C4+
"Interface Control During MOCVD Growth of InGaAsP/InGaP Heterostructures:" J.G. NEFF, R.V. Chelakara, M.R. Islam, R.D. Dupuis, Microelectronics Research Center, The University of Texas at Austin, MER 1.606D-R9900, Austin, TX 78712-1100; R. Hull, Department of Materials Science, University of Virginia, Charlottesville, VA 22903
We have implemented Flow Modulation Epitaxy, for the first time, to deposit AlInAs lattice matched to InP for use in High Electron Mobility Transistors (HEMTs). In addition, this is the first report of atomic planar doping in AlInAs/InP heterostructures using diethyltelluride to supply the donor impurities. A variety of growth parameters are varied to optimize material quality. Optical, structural, and electrical quality of the AlInAs are assessed using photoluminescence, double crystal x-ray diffraction, atomic force microscopy, Hall measurements, and capacitance-voltage measurements. Impurity distribution in device structures is studied with secondary ion mass spectroscopy. Finally, the advantages and disadvantages associated with the use of tellurium in HEMTs will be addressed using device results for illustration.
Because the transport properties of the two-dimensional electron gas at the AlInAs/InP heterointerface are very sensitive to the growth conditions, the sheet density and electron mobility are used to optimize material purity and flow modulation growth parameters in undoped AlInAs with respect to growth temperature and V/III ratio. The electron mobility is early constant [3(104)cm2/vsec] for growth temperatures below 650deg.C and drops sharply for temperatures above 650deg.C, while the residual carrier concentration is minimized at growth temperatures around 650deg.C. The V/III ratio is optimized near values of 20 for growth at 650deg.C. The variation in electrical properties of the undoped films will be discussed in relation to unintentional impurity incorporation and interface formation. Hall transport properties are also correlated with capacitance-voltage measurements, film morphology, and optical quality.
In our doped AlInAs studies, diethyltelluride diluted in hydrogen was used to supply the donor impurities. Both uniform and atomic planar doping were assessed. The tellurium was found to incorporate much more effectively as the V/III ratio was reduced, and we have achieved electron concentrations exceeding 7.5(1018) cm-3 in lattice matched AlInAs at a V/III ratio of 5. Furthermore, we have recently pioneered planar doping with tellurium. We consistently observe better mobilities at higher sheet densities using atomic planar tellurium doping in HEMTs than those obtained with uniform doping. The control of sheet charge density and uniformity with arsine flow, impurity supply time, and growth temperature will be addressed, as will the assessment and control of the memory effect associated with diethyltelluride.
11:20AM, C5
"Characterization of Interfacial Dopant Layer for High Purity InP Grown by MOCVD:" D.G. KNIGHT, G. Kelley, Bell Northern Research, PO Box 3511 Station C, Ottawa, Ontario, Canada K1Y 4H7; S.P. Watkins, Department of Physics, Simon Fraser University, Burnaby, BC, Canada V5A 1S6
Control of residual impurities is a key issue in the fabrication of compound semiconductor devices based on GaAs and InP. In recent years, the metal organic chemical vapor deposition (MOCVD) technique has achieved great success in reducing bulk impurity incorporation to very low levels. However, the presence of a dopant layer at the interface between the substrate and the epitaxial layer can severely degrade the performance of devices such as high electron mobility transistors (HEMTS), field effect transistors (FET's) and metal-semiconductor-metal (MSM) photodetectors. We will show that just such an interfacial dopant layer is formed by the presence of silicon at the interface for the growth of high purity InP, and that this interfacial dopant layer gives rise to anomalous transport properties as determined by the Hall technique. The two layer conduction caused by the presence of the interfacial dopant layer has a temperature dependent carrier concentration which mimics the freezeout behavior expected for a deep donor, and anomalously high carrier concentration and low mobility data are obtained at a single value of the magnetic field. The application of a two layer conduction model gives an excellent prediction ofthe change in Hall data as a function of the magnetic field at 77 K.
The presence of silicon at the growth interface is confirmed by secondary ionmass spectroscopy (SIMS), where high sheet concentrations of Si correlate with two layer conduction behavior as determined by Hall data. Analysis of samples with very low interfacial contamination, or samples where a two layer conduction model is used, indicates that the bulk properties of InP are excellent. The bulk carrier concentration is typically 2 x 1014 cm-3 and has a mobility of 1.5 x 105 cm2/Vs. The consistently pure nature of the trimethyl indium and phosphine precursors can be shown by the linear dependence of the room temperature Hall carrier concentration on the interfacial Si dosage, where the 2 x 1014 cm-3 value is obtained for the limiting case of a contamination free interface. The silicon is only partially active, as determined by comparison of total Si dosage obtained by SIMS and the fit value of the sheet carrier concentration as obtained from the two layer conduction model. This result is confirmed by comparison of total Si dosage and the active dosage obtained by a semiconductor profile plotter. The presence of inactive SiOx species is believed to be responsible for the partial inactivity of the silicon, since the interfacial oxygen dosage as determined by SIMS is linearly dependent on the SIMS Si dosage when several wafers are analyzed.
11:40AM, C6
"Interface Dependence on AsH3 Concentration in OMVPE Growth of InGaAs/InP
Superlattices:" A.R. CLAWSON, University of California San Diego, ECE
Dept.-0407, 9500 Gilman Dr., La Jolla, CA 92093-0407; D.P. Mullin
| Search | TMS Specialty Meetings Page | TMS Meetings Page | About TMS | TMS OnLine |
|---|