Cl 2 Reactive Ion Beam Etching of Heavy n-Type Si
- PDF / 1,075,221 Bytes
- 5 Pages / 417.6 x 639 pts Page_size
- 43 Downloads / 245 Views
C12 REACTIVE ION BEAM ETCHING OF HEAVY n-TYPE Si E. Eric Krueger and Arthur L. Ruoff Department of Materials Science, Cornell University, Ithaca, NY 14853
ABSTRACT 3
The etch yield vs n-type doping level up to 1.5 x 1021 As/cm was measured. Also, the etch yield as a function of ion energy (300 to 1500 eV) for the heaviest doped samples (1.5 x 1021 3 As/cm ) was measured. Only a small 10-15 percent enhancement in etch rate over low doped silicon was observed. Channeling experiments with Rutherford backscattering spectrometry showed 84 percent of the arsenic atoms were in substitutional positions in the silicon crystal. In addition, 2 silicon was implanted with arsenic to a dose of 1.25 x 1017 #/cm and then laser annealed. This produced samples with a 5000 angstrom layer of uniformly doped silicon at five atomic percent. Channeling experiments showed these samples to have 82 percent of the dopant in substitutional sites. Samples were then masked with chrome and etched at varied pressures, temperatures and ion energies. Within our experimental resolution no isotropic etching was observed. This small doping dependence of the etching behavior of n-type silicon in reactive ion beam etching suggests that the rate limiting step for C12 RIBE is different than for C12 RIE, rf-plasma and laser assisted etching. A discussion of this difference is given assuming a much lower concentration of atomic chlorine in RIBE.
INTRODUCTION Heavily doped n-type silicon is used in semiconductor technology for ohmic contacts, buried collector layers in bipolar transisitors and gate material in FETs. Several researchers have reported a doping dependence of the etching behavior of silicon in halogen based dry etching systems1- 7 . It has been observed with chlorine based gases in reactive ion etching (RIE), rf-plasma etching, and laser assisted etching, that heavily doped n-type silicon etches 3, 15, 100 times faster than lightly doped or p-type silicon, respectivelyl-e. Schwartz and Schaible 4 also observed isotropic etching of heavily doped n-type silicon in C12 RIE. Mogab and Levinstein 3 found that the dopants must be electrically active to observe the dopant level effects on the etching behavior. Several models have been suggested to describe this phenomenon, but the etching mechanism is still not understood. We have characterized the etching behavior of heavily doped n-type silicon during C12 reactive ion beam etching (RIBE). The etch rate of heavily doped n-type silicon was slightly enhanced but no evidence of isotropic etching was observed.We suggest that in C12 RIBE there is a lack of atomic chlorine that is rate limiting the etch, where as in C12 RIE, rf-plasma, and laser assisted etching the rate limiting step is the incorporation of chlorine into the silicon.
EXPERIMENTAL The experiments were performed in a 12 inch diffusion pumped chamber equipped with a cold trap. The C12 was admitted directly into a three grid 15 cm Kaufman9 type ion beam source. The current densities and chlorine pressures measured were corrected for resona
Data Loading...
