Internal gettering heat treatments and oxygen precipitation in epitaxial silicon wafers
- PDF / 391,910 Bytes
- 5 Pages / 594 x 810 pts Page_size
- 7 Downloads / 217 Views
I. INTRODUCTION In silicon technology internal gettering is a wellaccepted method for removing heavy metal contamination from the front surface of the wafers during device fabrication. Well-controlled internal gettering processes can reduce the surface defects caused by heavy metal precipitation and also improve the minority carrier lifetime.1'2 In polished silicon wafers the internal gettering process involves three major mechanisms: (i) the outdiffusion of oxygen at the surface region and formation of a defect-free (denuded) zone, (ii) nucleating of oxygen precipitates, and (iii) generation of precipitateinduced bulk stacking faults to act as gettering sites. The kinetics of the diffusion and precipitation of oxygen can be affected by several variables, such as the initial oxygen concentration in the silicon wafers, temperature, time, and heat treatment ambients.3 In the device fabrication process, heat treatments are normally performed in various ambients. Therefore the effect of ambients on oxygen precipitation becomes important. The actual effects of ambients are unclear. Hu 4 reported on the retardation of the oxygen precipitation process by silicon self-interstitials generating during heat treatment in an oxidizing ambient. On the other hand, deKock etal.5 suggested that self-interstitials play an important role only when the oxygen supersaturation ratio is less than 5. The primary purpose of this paper is to report the results of our investigation on the effect of ambients on the postepitaxial deposition internal gettering processes.
ONE-STEP
II. EXPERIMENTAL PROCEDURES
1050°C/X HRS.
The samples used were prepared from 100 mm P / P + (100) epitaxial silicon wafers between 500 and 525 fi thick. The epitaxial silicon layer was approximately 13.5 fi thick and had a resistivity between 30 and 50 fl cm. The substrate wafer was heavily doped with boron and had a resistivity ranging between 0.01 and 0.03 ft cm. The initial oxygen concentration in the epitaxial material can not be measured by the Fourier transform J. Mater. Res. 1 (5), S e p / O c t 1986
http://journals.cambridge.org
infrared (FTIR) method due to free carrier absorption of the dopant. However, the substrate wafers were from crystal ingots that had been grown using a procedure known to produce an oxygen concentration in the range 29-32 ppma. We will assume that this target has been achieved. The substrate wafers were mechanically backside damaged and encapsulated with a thin layer of chemical vapor deposition (CVD) silicon oxide on the backside. The CVD oxide layer was deposited to prevent the out-diffusion of boron from the substrate wafer. Each wafer was first cleaved into four quarter sections and then cleaned. Two quarter sections from a wafer were heat treated in a split run, using a one-step cycle, and the rest were heat treated using a two-step cycle. The samples were normally heat treated at high temperature at the same time. Split runs, in different heat treatment ambients, were also performed. The one- and two-step heat treatment cyc
Data Loading...
