L. Meda

Electrical and structural characteristics of thin buried oxides

Abstract In this paper a continuous thin BOX obtained by oxygen implantation into silicon at 200 keV is experimentally studied; the aim of this work is the correlation of BOX structural imperfections (inclusions and pinholes) as seen by transmission electron microscopy (TEM) with the electrical characteristics measured by capacitors with scaled areas. A method for obtaining information on the distribution of the BOX silicon inclusions from the breakdown voltage is proposed. The heterogeneity of these inclusions is found to affect the current-voltage characteristics.

Thin buried oxide in implanted silicon

Abstract Silicon-on-insulator (SOI) is a candidate substrate for future microelectronic devices. At present, the most reliable technology to obtain SOI is oxygen ion implantation (SIMOX); however, due to the implantation time, the material cost is high. A low dose SIMOX, with a thin((50–100 nm) buried layer, can decrease the present cost. The key to obtain low dose SIMOX is the control of oxide precipitation which takes place during ion implantation. A model is formulated to predict distribution and size of oxide precipitates formed during the implantation, and the feasibility of a very thin buried layer, at different depths, is demonstrated.

Displacement and recoil in ion implanted silicon

This work shows the existence of two kinds of implantation damage: the first one, easily produced by heavy ions, having a yield much higher than unity, being centered before the projected range (RP) and easily recoverable by thermal treatments; the second one, mostly produced by light ions, having roughly a 1:1 yield, being centered just on RP and not recoverable by the same treatments.

Silicon on thin insulator: prediction of the oxygen fluence required for the formation of a continuous buried oxide

Abstract A scheme is proposed to explain the fluence domain in which a continuous buried SiO 2 layer is formed to oxygen implantation and subsequent high temperature annealing. This scheme correctly describes the minimum fluence observed in single step implantation experiments and its temperatures and energy dependence; moreover, it also explains the seemingly extravagant structure of this fluence domain (showing a narrow window of allowed values at relatively low fluence) recently reported by Nakashima and Izumi ( Electron Lett., 26 (1990) 1647, J. Mater. Res., 8 (1993) 523) and the reasons making it possible the low fluence process reported by Meda et al. ( Nucl. Instrum. Methods B, 80/81 (1993) 813.

Current effects in BF2 implants in silicon

Boron implantion in (100) silicon with BF2 molecular ions was studied as a function of the implant current and fluence, using a medium current, commercial ion implanter equipped with a naturally cooled end station. Sample characterization was performed with Rutherford backscattering spectrometry (RBS), secondary ion mass spectrometry (SIMS), nuclear reaction analysis (NRA) and sheet resistance measurements. The thickness of the amorphous layer created by the impinging beam was found to increase, for fluences larger than the amorphization threshold, as the logarithm of the fluence for beam current densities lower than 50 μ A/CM2. For higher current densities the formerly created amorphous layer was found to recover by solid state epitaxy; the anneal is nearly complete for sample implanted at 500 μA/cm2 beam current density and at a fluence of 1 × 1016 cm−2. In high fluence samples boron and fluorine are snow-plowed by the travelling amorphous/crystalline interface both during the ion implantation step at high beam current and during furnace annealing at 500–600°C of low current implanted samples. For high current implanted samples fluorine is severely desorbed, and only a small fraction of the implanted dose is retained at the end of the implantation step. In addition, boron is redistributed up to the surface. During furnace annealing of high fluence, low current implanted samples, fluorine is not desorbed, and accumulates in two regions, at about half and twice the projected range of fluorine, where a high density of implantation originated defects are expected. In these samples boron does not reach the surface, but accumulates in the external fluorine-rich layer.

Bandgap widening in quantum sieves

Dense distributions of separate insulating regions in silicon may produce Anderson localization and hence a kind of confinement. This weak confinement is expected to be responsible for visible photoluminescence with different features from that observed in porous or nanocrystalline silicon. Dense distributions of separate insulating regions may be produced by ion implantation of a suitable species followed by an appropriate heat treatment. The visible photoluminescence observed in ion-implanted Si:H, Si:O and Si:He is explained in terms of exciton weak confinement.

Formation and stability of continuous buried SiO2 layers in SIMOX

Abstract Separation by implanted oxygen (SIMOX) is possible only when a continuous buried oxide (BOX) is synthesized. The formation of the buried oxide is controlled by SiO2 precipitation, which may occur in three ways: homogeneous, heterogeneous, or “etherogeneous” precipitations. These mechanisms take place in competition and the resulting BOX will be influenced by the prevalence of one or the other in as-implanted condition, and by their ability to drive the subsequent evolution during final high-temperature (⋍ 1350°C) annealing. The prevailing mechanism (which is controlled by machine parameters such as the implantation energy, target temperature and fluence rate) affects the evolution with fluence of the Si-SiO2 interface depth and the presence of silicon inclusions in the BOX. A tentative scheme for prevision of the SIMOX spectrum (i.e., the set of fluence values allowing the formation of a continuous BOX) as a function of the operation parameters is proposed. The scheme is validated in a few cases by studying the evolution with fluence of buried Si-SiO2 interfaces.

A paradigm for damage recovery in ion-implanted silicon

Consider a process of ion implantation into a phase X of silicon at or below room temperature. For instance, this phase may be a single crystal as well as a polycrystal with given orientation and grain-size distribution. Provided that: 1) the implant is carried out at a fluence low enough to generate vacancy-self-interstitial (v-i) pairs in negligible concentration, 2) the silicon phase X is stable up to the local temperature reached by the layer during the implant, and 3) the anneal temperature is below this temperature, but high enough to activate recrystallization phenomena, then the amorphous implanted layer reverts, after anneal, to the same phase X pre-existing to the implantation. The validity of this paradigm was experimentally demonstrated in the following cases: 1) X = diamond cubic silicon (trivial case); 2) X = (220) preferentially-oriented polysilicon stabilized by annealing at 1000°C; 3) X = strongly-damaged, self-annealed silicon obtained by high fluence, high current ion implantation. It was also demonstrated that if one or other of the hypotheses 1) to 3) is ruled out, then the reconstructed phase is not the saint as the one pre-existing the implantation. Therefore, the amorphous phase obtained by low fluence, low temperature ion implantation retains memory of the previous structure.

C2H4 INSERTION INTO ION-BOMBARDED POROUS SILICA

Abstract A study via infrared and X-ray-photoelectron spectroscopies of ethylene reactions with porous silica during high-energy (300 keV) argon bombardment in a C 2 H 4 atmosphere, followed by exposure to air at room temperature or heating at 500°C in a vacuum, is reported. This study has revealed in the as-prepared sample the presence of CC, CO, CH 2 and CH 3 groups. Aging at room temperature in a nitrogen atmosphere as well as in air produced an increase of CH 3 and CO amount, while a heat treatment at 500°C produced the following: the disappearance of CC and CO groups, an increase in the CH 2 amount, and a decrease in the CH 3 amount. An interpretation for this behaviour is proposed. A previous conclusion concerning CO 2 addition to ion-bombarded porous silica, that approximately 10 2 oxygen-bridge vacancy per impinging ion are stabilized through the addition of gas molecules, is confirmed by the present work.

Etherogeneous precipitation in oxygen-implanted silicon

Abstract SiO 2 precipitation in high-fluence oxygen-implanted silicon takes place in three regions: the region centred on the depth x dam of maximum damage production, the region centred on the most probable penetration depth x max , and a region at greater depth x ‡ . At fluences Φ for which compositional changes and sputtering are of lower importance, x max and x dam do not depend on Φ (they depend only on the implantation energy E ); in contrast, x ‡ varies significantly with Φ. The mechanism of SiO 2 precipitation at x dam is heterogeneous precipitation; the mechanism active at x max is homogeneous precipitation; the mechanism active at x ‡ (Φ), referred to as etherogeneous precipitation, is a new mechanism characteristics of these mechanisms or for the prevalence of one or the other of them after high temperature annealin is given for Φ in the interval 5 × 10 15 - 2.5 × 10 17 cm −2 and for E between 100 keV and 200 keV.