Mechanisms of layer-transfer related to silicon-on-insulator structures

The objective of this dissertation was to study the mechanisms that affect an efficient hydrogenation process in silicon and to validate a hypothesis concerning the hydrogenation mechanism of pre-implanted silicon wafers under hydrogen-plasma processing. These studies are related to a general process. A trapping layer for hydrogen was introduced by ion beam implantation into the silicon wafer. Next, this wafer was hydrogenated using a hydrogen plasma. It was hypothesized that the trapping layer acts as a getterer for the hydrogen diffusing into the silicon wafer. It was found that a large amount of hydrogen could be absorbed in the trapping layer by hydrogen plasma processing. The depth of layer transfer and surface blistering could be controlled by the trapping layer after annealing. This research focused on the mechanism of hydrogen plasma reacting with the buried, heavily disordered silicon layer, deep level defects in the wafer, and nano-/microcrack growth enhanced by the effect of inertial gas bubbles and hydrogen plasma processing. This dissertation also studied suitable wafer bonding methods for a novel method to produce nanoscale silicon-on-insulator materials. Silicon wafers were implanted by different elements (He, N, Ne, Ar) at appropriate energy. A hydrogen trapping layer formed in the depth of ~100 nm based on calculations. Two different types of trapping layers were formed, one composed of a vacancy cluster, and the other was a gas bubble formation. In the trap layer, there were many vacancies, interstitials and micro-voids. Due to the presence in the plasma of molecular and atomic hydrogen and extremely strong acids H_3^+ and H_2^+ the surface of Si wafer could be modified by H^+, H_2^+, H_3^+ in H-plasma. The surface strained Si-Si bonds were damaged. H_2 readily dissociated, bound to the surface and diffused into the Si bulk at low temperatures (150 - 200 C). At higher temperatures (300- 350 C), the trapped in silane-like species was detrapped, and hydrogen atoms diffused deeper into the Si bulk. When these atoms met the buried disorder layer and bubbles, they formed another Si-H structure and molecular hydrogen accumulated in the interstitial voids. In plasma-ion-immersion-implantation processing, compared with non-implantation samples, less hydrogen was trapped in the disordered structure and many hydrogen was trapped in the internal surface of the voids. Many small defects could be generated in normal H Pill processing. Inertial gas pre-implantation may help remove these small defects and produce good quality transferred layer in the later layer exfoliation. It is found that many voids could be generated in bonding wafers by using only RCA clean activation, annealed at temperature above 200 C. These voids could not disappear until annealed to above 1050 C. These voids were induced from dissociated H_2 from water and evaporated CH_X. Plasma activation and hot nitric acid activation for direct wafer bonding may be a good choice for SOT fabrication.

Name thesis:degree_name

Doctor of Philosophy in Materials Science and Engineering - (Ph.D.)

Discipline thesis:degree_discipline

Committee for the Interdisciplinary Program in Materials Science and Engineering

Year

2004

Author dc:creator

• Chen, Bo

Contributors dc:contributor

• William N. Carr
• Alexander Y. Usenko
• Yves Jean Chabal

• Plasma processing
• Silicon-on-insulator
• Wafer bonding
• Hydrogenation
• Layer exfoliation
• Materials Science and Engineering

Repository record dc:identifier

https://digitalcommons.njit.edu/dissertations/622

OAI identifier oai:identifier

oai:digitalcommons.njit.edu:dissertations-1677