(54) SEMICONDUCTOR MATERIAL AND (51) Int. Cl." ................................................ H01L 21/20 METHOD FOR ENHANCING SOLUBILITY (52) U.S. Cl. ........................ 438/542; 438/544; 438/795 OF A DOPANT THEREN (58) Field of Search ................................. 438/510,530, 438/544, 795. 478: 428/704. 469, 699 (75) Inventors: Babak Sadigh, Oakland, CA (US); /544, 795,478; 428/ 4723,450 Thomas J. Lenosky, Pleasanton, CA • as (US); Tomas Diaz de la Rubia, Danville, CA (US); Martin Giles, (56) References Cited Hillsborough, OR (US); Maria-Jose Caturla, Livermore, CA (US); Vidvuds U.S. PATENT DOCUMENTS Ozolins, Pleasanton, CA (US); Mark 4,181,538 A * 1/1980 Narayan et al. Asta, Evanston, IL (US); Silva Theiss, 5,110,373 A * 5/1992 Mauger St. Paul, MN (US); Majeed Foad, 5,212,101 A * 5/1993 Canham et al. 5,286.660 A * 2/1994 Chiou et al. sistics (S). Andrew Quong, 5,298.452 A * 3/1994 Meyerson

Optimization of the interface between high-k dielectrics and SiGe substrates is a challenging topic due to the complexity arising from the coexistence of Si and Ge interfacial oxides. Defective high-k/SiGe interfaces limit future applications of SiGe as a channel material for electronic devices. In this paper, we identify the surface layer structure of as-received SiGe and Al2O3/SiGe structures based on soft and hard X-ray photoelectron spectroscopy. As-received SiGe substrates have native SiOx/GeOx surface layers, where the GeOx-rich layer is beneath a SiOx-rich surface. Silicon oxide regrows on the SiGe surface during Al2O3 atomic layer deposition, and both SiOx and GeOx regrow during forming gas anneal in the presence of a Pt gate metal. The resulting mixed SiOx-GeOx interface layer causes large interface trap densities (Dit) due to distorted Ge-O bonds across the interface. In contrast, we observe that oxygen-scavenging Al top gates decompose the underlying SiOx/GeOx, in a selective fashion, leaving an ultrathin SiOx interfacial layer that exhibits dramatically reduced Dit.

The evolution of very high dose arsenic implants in silicon formed by plasma immersion ion implantation and deposition (PIIID) using a non-pulsed plasma source was studied over a time period of more than one year. The study focused on the effect of arsenolite micro-crystal formation, enhanced oxidation, and the significant As dose loss from as implanted samples at room temperature. The study was carried out combining analytical evidence from SIMS, XPS, INAA, SEM, and optical microscopy suggesting a two stage process of As dose loss, the first implying arsenolite crystal growth and the second an out-diffusion. In fact, comparison of samples fabricated using different PIIID parameters showed that crystal growth seems not to be the only process responsible for dose loss. (C) 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

The stability and the evolution of electrical properties of high concentration arsenic ultra-shallow junctions in silicon have been studied with regard to their effect on the evolution of point defects. The activation of 2 keV 1x10(15) cm(-2) As implants was performed using millisecond sub-melt laser annealing at two different temperatures, 1100 and 1300 degrees C. The electrical deactivation upon subsequent thermal treatment at 700 degrees C was indirectly monitored through the diffusion of five 10 nm-wide boron layers aimed to detect the injection of self-interstitials coming from dopant clustering. Thermal treatments were repeated on samples implanted with Ge at condition similar to the As ones. The comparison helped to discriminate between interstitials coming from lattice damage evolution and dopant clustering. The results show the relevance of the laser annealing temperature in order to ensure junction stability in terms of active carrier concentration and junction depth. (C) 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Samples produced by plasma immersion ion implantation of Arsenic in Silicon using a non-pulsed plasma source and subsequent laser annealing were investigated with respect to As depth distribution, oxide thickness, and As local order using SIMS, XPS, INAA and EXAFS analysis. A surface layer (similar to 10 nm), was identified as an As-rich Si oxide formed after implantation. The thickness of this layer was found to be larger for samples annealed using a low thermal budget up to a threshold where probably melting occurred. Dopant depth profile was re-distributed whereas the final oxide film of these samples showed thicknesses of a few nm. The retained As dose exhibited an apparent drastic increase. A hypothesis for the processes involved is presented based on experimental evidence.

Spontaneous growth of arsenolite micro-crystals at room temperature after high fluence, low energy arsenic trihydride implantation in silicon was observed on the wafer surface after exposure to air. The crystals have been identified unambiguously by x-ray absorption and fluorescence spectroscopy. Thermal treatment easily sublimates the crystals at temperatures as low as 200 °C without any relevant in-diffusion of As into the substrate. The deposition of a thin As-rich layer under high fluence implantation conditions is suggested as possible precursor for crystal formation. The same layer can explain the anomalous retained dose increase often observed after annealing.

Boron is the most important p-type dopant in Si and it is essential that, especially for low energy implantation, both as-implanted B distributions and those produced by annealing should be characterized in very great detail to obtain the required process control for advanced device applications. While secondary ion mass spectrometry (SIMS) is ordinarily employed for this purpose, in the present studies implant concentration profiles have been determined by direct B imaging with approximately nanometer depth and lateral resolution using energy-filtered imaging in the transmission electron microscopy. The as-implanted B impurity profile is correlated with theoretical expectations: differences with respect to the results of SIMS measurements are discussed. Changes in the B distribution and clustering that occur after annealing of the implanted layers are also described.