Identification and Sources of Impurities in InGaAs Grown by Liquid Phase Epitaxy
- PDF / 1,637,998 Bytes
- 8 Pages / 417.6 x 639 pts Page_size
- 18 Downloads / 228 Views
IDENTIFICATION AND SOURCES OF IMPURITIES IN InGaAs GROWN BY LIQUID PHASE EPITAXY
D.G. KNIGHT, C.J. MINER AND A. MAJEED Bell-Northern Research Ltd., P.O. Box 3511, Canada, KIY 4H7
Station C, Ottawa,
Ontario,
ABSTRACT High purity In. 5 3 Ga. 4 7 As and InP with carrier concentrations 15 [ND-NA] ' 5x10 cm3 has been grown by the LPE technique on both n-type and semi-insulating substrates to detect and identify trace donor and low Acceptor impurities have been detected in acceptor impurities. temperature photoluminescence spectra where LPE melt baking and growth programs indicate a melt origin for two of these species, one of which is zinc. Data from semiconductor profiles provides evidence for sulfur and tin donor impurities, which comes from the rinse melt used to etch back Silicon and sulfur substrates doped with the respective contaminants. contaminants have been detected by SIMS measurements; and may arise not only from the indium and III-V materials, but also the graphite boat used Volatile sulfur-containing compounds have been to grow the epilayers. detected during high temperature bake-out of high purity graphite boats.
INTRODUCTION For the production of avalanche photodiodes (APD's) and PIN photodiodes for X = 1.O-1.60m fiber optic communication systems, it is 5 3 necessary that In. 5 3 Ga. 4 7 As with IND - NA] < 5xl0 cmbe grown on InP as a light absorbing material. However, the stringent purity requirements makes it necessary to identify the contaminants and their probable sources, so that production of wafers for PIN diodes can be achieved repeatably. The wafers grown by liquid phase epitaxy (LPE) in this work consist of an n-In. 5 3 Ga.4 7 As layer lattice matched to an n-InP buffer layer, which is in' turn grown on an n+InP substrate. Unintentionally doped layers grown without precautions are always n-type, and are typically 1-5xl0 6cm-3 for InGaAs and - 101cm-3 for InP. The two donor impurities most often implicated for this unintentional doping are sulfur and silicon, which have distribution coefficients of .4-10(l and 30[2] respectively. There is a difference of opinion in the literature as to which impurity is more dominant and exactly what precautions should be taken during crystal growth to produce high purity epilayers[3-11]. However, it is reasonable to assume that the high distribution coefficient for these elements makes them both suspect as contaminants if it can be shown that they are present in trace amounts in either the crystal growth materials or the LPE system. The most commonly reported acceptor in LPE grown InGaAs is zinc [12, 13]. Its most probable sources are the InAs and GaAs used in melt preparation. When present, acceptors degrade the mobility of the InGaAs layers and thereby reduce device speed. All authors agree that baking of the indium and III-V materials is necessary before crystal growth. Oliver and Eastman[3] were the first to propose the neutralization of silicon by H20 as the purification mechanism. This is reasonable since silicon is present at ~.lppma in the indium u
Data Loading...
