INFLUENCE OF ARSENIC VAPOR SPECIES ON ELECTRICAL AND OPTICAL PROPERTIES OF MBE GROWN GaAs

The in f luence o f d imer ic (As2) vs te t ramer ic (As4) arsenic vapour spec i e s on the e l e c t r i c a l and o p t i c a l p roper t ies o f undoped and of Ge-doped GaAs grown by molecular beam ep i taxy (MBE) has been studied by using H a l l e f f e c t , photoluminescence (PL) and DLTS measurements. The arsenic molecular beam was obtained from separate As2 and AS+ sources, resp., and from a s i n g l e source p rov id ing adjustable As4/As2 f lux r a t i o s . The occurrence o f the defect r e l a t e d bound exc i ton l i n e s i n the 2K PL spectra was found t o be d i r e c t l y co r re la ted w i t h t h e existence o f th ree deep s ta tes (MI, M3, M4) which a re e h a r a c t e r i s t i c o f MBE grown GaAs. The i n t e n s i t y o f these e x t r a PL l i n e s and simultaneously the e l e c t t o n t r a p concentrat ion could be reduced s u b s t a n t i a l l y by decreasing the Asq vs As2 f l u x r a t i o . I n add i t i on , a cons ide rab ly lower autocompensation r a t i o i n Ge-doped n-GaAs was achieved w i t h As2 molecu?ar beam species which prov ide a such h igher steady-state arsenic sur face popul a t i o n . These r e s u l t s demonstrate t h a t the incorpora t ion o f defect r e l a t e d centers as we l l as o f amphoteric dopants s t rong ly depends on the sur face chemistry involved i n the f i l m growth process. 1. Int roduct ion. I n the growth o f GaAs f i l m s by molecular beam ep i taxy (MBE) the arsenic molecular beam cons is ts o f e i t h e r te t ramer ic (As4) [l] o r d imer ic (As2) spec i e s C21. General ly, AS+ mol ecular beams produced by evaporat ing po lyc rys ta l 1 i n e arsenic source mate r ia l are employed. For the more labor ious generat ion o f As2 molecu les dur ing MBE, the re a r e naw th ree methods avai lab le: (a) t h e incongruent evapora t ion o f GaAs source mate r ia l [31, (b) the thermal decomposition o f gaseous AsH3 i n a specia l l eak -e f fus ion source [41 o r crack ing furnace [51, and (c) the crack ing o f As4 i n t o As2 species i n a two-zone e f f u s i o n c e l l a r rayement [6,7]. Previous s tudies ind ica ted t h a t the q u a l i t y o f MBE GaAs f i l m s i s probably Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1982517 C5-136 JOURNAL DE PHYSIQUE superior when grown from dimeric arsenic species [7, 81. Low-temperature photoluminescence (PL) measurements showed the existence of additional sharp PL l ines in the range 1.504 1.511 eV which a re due to radiative recombination of bound excitons induced by point defects 181. The occurrence of these defects was correlated with the complex second-order interaction process of the impinging As4 molecules on the (100) GaAs ~rowth surface [I!. Except for the l ine a t 1.5109 eV, the extra PL features disappeared to t a l ly when As2 instead of As4 beam species were used, thus indicating a substantial decrease of growth induced defects. A comparable reduction of the concentration of deep electron traps in As2-grown MBE GaAs was derived from DLTS measurements [7] The apparent advantage of As2 over As4 species in tlBE of GaAs was attr ibuted to the elemectary first-order reaction process of dissociative chemisorption of As2 involving only a single Ga surface atom [2]. This reduces the probability of incorporation of point7GTEts related to Ga vacancies a.nd thei r descendants C81. In addition, i t has been speculated that a higher steady-state arsenic surface population during growth exis ts with As2 beam species. If so, t h i s should also have a strong impact on the incorporation of amphoteric dopants in MBE GaAs. In t h i s paper we present the resul ts of combined Hall e f fec t , low-temperature PL and DLTS measurements on not intentionally doped and on Ge-doped GaAs layers grown from separate As4 and AS:! sources, resp., or from a single source providing adjustable As4/As2 flux rat ios. In addition to the incorporation of defect related centers, we examine the autocompensation r a t io and thereby the s i t e occupancy of the amphoteric dopant Ge. The resul ts are then correlated with the different surface chemistry involved with As2 and As4 molecular beams. 2. Experimental. The GaAs films were grown in a UHV system of the vert ical evapor a t m i p p e d with a sample exchange load-lock device [91. For the present study we operated four effusion ce l l s to evaporate elemental arsenic, gallium and germanium a s well as polycrystalline GaAs. The epitaxial GaAs layers with a thickness of 2 ym were deposited without any buffer layer simultaneously onto (100) criented Cr-doped SI and heavily Si-doped N -GaAs substrates. During deposition, the substrate temperature was typically maintained a t 550 C , i f not quoted otherwise. The temperatures of the effusion c e l l s containing Ga and elemental arsenic (lower part of the ce l l only!) or GaAs, resp., were kept constant throughout the growth runs resulting in a constant growth ra te of 1.0 pm/h and a constant arsenic-to-gallium f lux r a t io of about two, which produces the As-stabilized (2x4) reconstruction on the (100) growth surface. During th i s study we adjusted the temperature of the Ge effusion ce l l to yield a constant d p ~ a n t 5lux f o r a gree:electron concentration in the grown layers of between 1x10 cmand 4x10~ cmdepending on the autocompensation rat io. 3 The elemental arsenic was evaporated in a two-zone 20 cm capacity effusion ce l l made from high-purity quartz glass uhich has been described in Ref. 10. The two zones could be heated independently and thei r temperatures measured by W-Re thfrmocoupl es were recorded separately. The temperature of the 1 ower (evaporation) zone determines the total f lux, while the temperature of the upper (cracking) zone determines the emerging As4/As2 ra t io . This arrangement made feasible a very e f f i cient cracking of the evaporating tetrameric species into dimeric species, the amount depending only on the temperature of the removable upper zone. The generation of adjustable As4 /As2 flux ra t ios from a single source avoids the introduction of any additional impurities from separate A m As4 sources. Operation of the cracking zone a t temperatures below m i e l d s -a pure As4 flux, whereas operation of tha t zone a t 850 C generates a beam consisting of a t l eas t 90% As2. The undoped GaAs films grown from the single elemental arsenic source exhibilgd coypensated or s l ight l y n t y p ~ behaviour with electron concentrations below 1x10 cm . Even a t 850 OC cracker temperature there was no measurable influence on the layer quality. After evaluation through 300 K and 77 K Hall ef fec t measurements, the samples were studied by detailed photoluminescence measuremen$s a t temperatures below 2 K. Excitatioq was performed by the 6471 A l ine of a Kr laser using a power density of 1 Wcm . The excitation power was adjusted by neutral density f i l t e r s . The luminescence l igh t was analysed by a grating monochromator with 1 A resolution and detected by a photon counting system attached to a cooled GaAs photolathode photomultiplier. The DLTS measurements of layers grown on n -substrates were performed on Au Schottky barrier diodes usinq a fu l ly automated measuring system, which has been described elsewhere t l l l . During measurement the samples were biased a t -2.0 V. The whole capacitance transient was recorded a f t e r applying a short voltage puls of 2.0 V to zero-bias the sample. The emission rates of the electron traps were then calculated from the time constant of the capacitance transient. The temperature dependence of the thermal emission rates yielded the activation energy of the traps and allowed a comparison with data from the l i t e ra tu re L121. The t rap concentrations were determined directly from the peak height of the DLTS signal, the total capacitance, and the free-electron concentration a s determined by Hall measurements. 3. Results and Discussion. -Investigations. Two typical 1 ow-temperature PL spectra of FlBE GaAs in the near band gap energy region are shown in Fig. 1. The upper spectrum was obtained from a nominally undoped GaAs layer grown w i t h the cracking zone of MBE the arsenic efgusion ce l l kept a t a tempePh,=1wcm-2 reture of 300 C. The well-known emission lines of the f ree exciton, FE, the d ~ n o r bound-exciton, (DO, X ) , and the acceptor bound-exci ton, (A', X ) , are clearly resol ved [131. In addition, the emission structures a t energies below 1.51 eV appear, which have been attr ibuted to defect induced bound excitons [81. The ent i re PL spectrum shows thus the same peculiaif t ies as i f the sample was grown from a standard onezone Asb source, as shown in Fig. 2. Fig. 1 : Comparison of 1.8 K photoluminescence spec t ra obtained from nomin a l l y undoped MBE GaAs l ayers grown from predeminantly As+ ( top) and As2 (bottom) molecular beam species of t h e two-zone effusion