Surface passivation effects of As2S3 glass on selfâ•’aligned AlGaAs/GaAs heterojunction bipolar transistors

A recently developed As2S3 chemical treatment was used to passivate the perimeters of self‐aligned heterojunction bipolar transistors (HBTs). The As2S3 chemical treatment significantly lowered the base current resulting in a two order of magnitude reduction in the collector current density at which dc current gain was observed (β=1). No degradation with time has been observed in the electrical characteristics of the chemically treated HBTs. This absence of degradation is attributed to the impermeability to oxygen of the As2S3 glass which coats the perimeter of the HBT after chemical treatment.

Surface passivation effects of AS 2 S 3 glass on self~aUgned AIGaAs/GaAs heterojunclion bipolar transistors H. l. Chuang A recently developed AS 2 S 3 chemical treatment was used to passivate the perimeters of self-aligned heterojunction bipolar transistors (HBTs). The As 2 S 3 chemical treatment significantly lowered the base current resulting in a two order of magnitude reduction in the collector current density at which de current gain was observed «(3 = 1). No degradation with time has been observed in the electrical characteristics of the chemically treated HBTs. This absence of degradation is attributed to the impermeability to oxygen of the AS 2 S 3 glass which coats the perimeter of the HBT after chemical treatment.
The AIGaAs/GaAs heterojunction bipolar transistor (HET) is a very promising device for high-speed applications. Both high-speed performance and low power consumption require the reduction of lateral dimensions of an HBT to minimize parasitic capacitances and resistances. However, it has been found that the current gain of an HBT degrades as the emitter size is scaled down. 1 This current gain degradation is caused by the increased importance of base current associated with high surface recombination of minority carriers around the unpassivated emitter-base junction perimeter and at the exposed base surface. I -3 This surface recombination can be minimized by the use of a graded base to establish a built-in electric field to steer the electrons away from the exposed base surface,3 or by the use of an AIGaAs layer to passivate the base surface. 4 . 5 In the last few years it has been shown that chemical treatments of GaAs surfaces with inorganic sulfides can produce dramatic modifications of the surface electrical properties.~!2 The gains of non-self-aligned HETs have been shown to increase by 60-fold with treatment of the emitter base junction with Na2S·9H20.6,7 Reductions in edge recombination currents have been observed with Na2S'9H20 and ammonium sulfide treatment of GaAs pn homojunction solar cells. 8 These reductions in perimeter currents can be attributed to the reduction in surface recombination of the GaAs surface after inorganic sulfide treatment. I 1 The chemical nature of one of these inorganic sulfide treatments, ammonium sulfide, has been investigated with x-ray photoelectron spectroscopy (XPS) . [13][14][15] The XPS studies found that the ammonium suifide treatment removes surface oxides and terminates the GaAs surface with about a monolayer of sulfur, Since the ammonium sulfide chemically modified GaAs surface is terminated with just a monolayer of sulfur, it is not surprising that the surface eventually degrades back to its initial electrical characteristics, due to reoxidation of the underlying GaAsY Recently, Yablonovitch et al. 16 have developed a sulfide chemical treatment which leaves the GaAs surface coated with an amorphous As 2 S 3 layer. This AS 2 S 3 chemical treatment resulted in a lOO-fold reduction in perimeter current for conventional doubleheterostructure lasers with no degradation after four months. This absence of degradation was attributed to the impermeability of the As 2 S} glass, which prevents oxygen from diffusing to the GaAs surface. The AS 2 S 3 chemical treatment has also been applied to GaAs pn homo junctions resulting in at least a tenfold reduction in the perimeter current. 17 In this letter we report our investigations of the passivation of self-aligned HBTs with the AS 2 S 3 chemical treatment.
The AIGaAs/GaAs HET structure used in this experiment is shown in Fig. !. The epitaxial layers were grown in a Varian GEN II molecular beam epitaxy system on a 2in.-diam semi-insulating (lOO)GaAs substrate. The epitaxiallayers were grown at a substrate temperature of 600 °C. 8i was used as the n-type dopant and Be as the p-type dopant. The doping densities of the epitaxial layers were determined from the calibrated temperatures of the dopant ovens. The top and bottom of the AIGaAs emitter were parabolically graded to an Al mole fraction of 0.3 over 300 A. All the HETs were oriented at 45° to the (110) planes (the natural cleavage planes) so that wet-etched vertical sidewalls were obtained. Alloyed AuGe/Ni/Ti/ Au was used for the 'Hype contacts to the emitter and collector, and nonalloyed Til Au for contact to the p-type base. The emitter contact was used as a mask for both the base etch and the base metallization. A photoresist mask covering the emitter and base contacts was used to define the base and also used as a lift-off mask for the collector metallization. Another photoresist mask was then used to define the collector contact and device isolation. Wet etching of the AIGaAs and GaAs layers was performed using 20:7:1000 NH 4 0H:H 2 0 2 :H 2 0, which provided well defined vertical sidewalls and an undercut ratio of 1: 1, The base and collector contacts were on opposite sides of the emitter as shown in Fig. 1. After fabrication, the HBTs were electrically characterized. Contacts were made to the HETs by probing on a micromanipulator model 6000 probe station and electrical data were taken with a Hewlett-Packard model 4145 semiconductor parameter analyzer. Following the electrical characterization, the HBTs underwent AS)S3 chemical 16 -treatment. Briefly, the AS 2 S 3 chemical treatment consisted of an oxide removal etch in 1: 10 NH 4 0H:H 2 0 followed by a 5 min soak in ammonium sulfide. The sample was then rinsed in de-ionized water and placed in a solution of 0.16 M AS 2 S 3 dissolved in a 1: 1 solution of NH 4 0H:CH 3 0H. The sample was then spun dry under flowing nitrogen. The final step was an anneal at 280°C on a graphite strip heater in flowing nitrogen for 2 min. Following AS 2 S 3 chemical treatment, the HBTs were again electrically characterized. Shown in Fig. 2 is the collector (Id and the base (IB) currents as a function of the base-emitter (V BE ) voltage with the base collector junction shorted for an HBT with an emitter size of 14 f-Lm X 14 f-Lm before and after the AS 2 S 3 chemical treatment. (HBTs with emitter sizes of 14 f-i,mX 14 f-i,m were the smallest HBTs we were able to test after the AS 2 S 3 treatment without damaging the devices when probing through the AS 2 S 3 film.) No change was observed in the collector current except for a difference in roU-off at high current levels due to a dif- ference in series resistance. (This difference in series resistance is probably a consequence of the probe contact.) In contrast, there is over an order of magnitude reduction in the base current at biases of V SE = 0.6-0.7 V following the As 2 S 3 chemical treatment. Also seen in Fig. 2 is that the leakage current initially observed in the base current has been eliminated by the AS 2 S 3 chemical treatment. Another significant change due to the As 2 S 3 chemical treatment was a reduction in the ideality factor of the base current from n = 1.81 to 1.54 for biases of VEE = 0.7-1.1 V. (The ideality factor ofthe collector current for biases of VUE = 0.9-1.2 V was n = 1.01 before and after the chemical treatment). Similar behavior was observed with all the treated HETs.
To illustrate the effects of chemical treatment on dc current gain of the HBT, we have plotted the ratios of the collector-to-base currents before and after treatment as a function of the collector current density in Fig. 3. Observed in Fig. 3 is that the collector current density at which gain is first observed (f3 = 1) has been lowered by about two orders of magnitude following the As 2 S, chemica! treatmcnt. Also, the dc current gain shows an-increase of more than a factor of 10 at the lower current densities. However, the dc current gain after the AS 2 S 3 chemical treatment still exhibits a strong dependence on current density, which is caused by a continued presence of a base current component due to recombination in the emitter-base depletion region. There is either 8till some perimeter recombination or the recombination is dominated by the bulk of the junction in the emitter-base depletion regicn after the AS 2 S 3 chemical treatment.
In summary, we have investigated the effects of treating the exposed perimeters of self-aligned HBTs with a new AS 2 S 3 chemical passivation technique. The AS 1 S 3 chemical treatment significantly lowered the base current, resulting in a two order of magnitUde reduction in the collector current density at which dc current gain is observed (f3 = 1). Furthermore, no degradation has been observed in the treated HBT characteristics after two months exposure to room temperature and atmospheric conditions. The work at Purdue University was supported by the