Deep Level Investigation in Si-GaAs by Rapid Thermal Annealing

  • PDF / 293,566 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 85 Downloads / 210 Views

DOWNLOAD

REPORT


AScaVca [10]. Some recent reports proposed that there is a relation between EL2 and EL6 defect [9,12]. According to these reports, the atomic structure of EL2 and EL6 are assumed to be AsV,,V& and VGaVA, respectively. In spite of an enormous worldwide interest in these centers, a microscopic identification and explanation for their behaviors has not been provided so far. In this paper, the annealing behaviors of defects and variation of resistivity in the undoped LEC semi-insulating GaAs by rapid thermal annealing (RTA) are repoted. This work emphasizes that there are two defect gropus in SI-GaAs: EL2 and EL6 groups. EXPERIMENT Good-quality commercial liquid-encapsulated Chochraslki (LEC) SI-GaAs was used in this study. They were labeled as grown sample A and annealed sample B (at 950 'C for five hours). The wafer was cut into small pieces of approximately 5x5 mm2, on which ohmic contacts were made for PITS [12] and Hall measurements. The contacts consist of spots 0.55mm in diameter, which were sintered at 400 'C for 5 min under 112 atmosphere. Before fabricating Ohmic contacts, the surfaces layers of 0.5 um thickness were chemically etched in order to remove the thermaly damaged layers. The A samples were annealed at 500 °C-800 °C for 10 s under N2 atmosphere. To avoid arsenic-escaping, the sample was covered with a piece of LEC GaAs. The temperature was

407 Mat. Res. Soc. Symp. Proc. Vol. 470 01997 Materials Research Society

elevated at the rate of 100 'C/s and cooled to room temperature rapidly. PITS measurement was performed with a BIO-RAD DM4600 deep level transient spectroscopy system The wave length of pulse light is 850nm, the width 4 ms.6V voltage was applied between two Ohmic contacts. PITS temperature scan was from 77K to 400K Changing rate windows, we obtained a series of PITS curves. The essential feature of PITS is the ability to set an emission rate window so that the measurement apparatus responds only when it sees a transient with a rate within this window. Thus, if the emission rate of a trap is varied by varing the sample temperature, the instrument will show a response peak at the temperature where the trap emission rate is within the window. The peak intensity is proportional to trap concentration. If we take the ln(e,/T 2) as a function of temperature, where e, is electron emissivity. By plotting ln(e,/T 2) vs I/T, the activation energy Ea and the capture cross section of the defects can be determined. RESULTS AND DISCUSSION

Fig. 1 shows PITS signal for as-grown sample A and sample B annealed at 950 'C for five hours in our work. In Fig. 1, PITS signal is ploted on a logarithmic scale, and the rate window is 20 s-'. Eight PITS peaks are shown in Fig. 1. Their activation Ea is shown in table I. Because of the

EL9 EL8 EL6 EL5 E1A

EL12

I

I

EL2

(b)

xio0I



U

mj~

100

200

300

400

T(K)

Fig. I The PITS signal for SI-GaAs (a) sample A: as-grown (b) sample B: annealed at 950 °C, 5h

408

TABLE I. The activation energy ( Ea ) on LEC SI-GaAs studied in this work. Energy level Ec-Ea(e