Effect of complex formation on diffusion of arsenic in silicon

Self Cite

In a previous paper, the diffusion of ion-implanted As in <100> Si was discussed as well as the electrical quality of implanted-diffused layers. It is the purpose of this paper to derive equations that describe the important characteristic profile parameters, and to support these equations with experimental data. A discussion of total As surface concentration, junction depth, and profile gradient will be presented, which will be followed by an analysis of the sheet resistance of implanted-diffused As layers and the surface concentration of the electrically active As. AnalysisTotal As surface concentration, CTO, and junction depth, xj.--It was shown previously (1) that the implanted-diffused total As profile shape is approximately described by the equationt is the diffusion time (sec), Di is the intrinsic As diffusivity (cm2/sec), n1 is the intrinsic electron concentration, and CTO is the As surface concentration (atom/cm 8) assumed to be greater than ,~1 X 1019 cm-~ Equation [1] is a Chebyshev polynomial approximation to the solution of the diffusion equation with a linear concentration-dependent diffusion coefficient and constant surface concentration, CT0. It is probable that the polynomial coefficients are time-dependent for the case of nonconstant CTO, such as the case of a redistributing impurity distribution. For the case of implanted As in St, CT0 decreases as t -~/~, and the con-

Section: Discussionmentioning

confidence: 98%

Profile Parameters of Implanted‐Diffused Arsenic Layers in Silicon

Fair¹,

Tsai²

1976

Self Cite

“…They were generally deduced applying the Boltzmann-Matano method to the total arsenic profile and show (see Fig. 4) an apparent reduction of the diffusivity for concentrations exceeding the above-mentioned value (2,4,6,15). This trend is due to the fact that the fraction of the dopant which is present in the form of precipitates (or clusters, according to the authors who proposed this model) is not mobile.…”

Section: Fig 1 Carrier Concentration Profiles Of Ion-implanted and mentioning

confidence: 99%

“…Well-known models hypothesize the formation of complex point defects, or "clusters," in thermal equilibrium (1)(2)(3)(4)(5)(6).…”

mentioning

confidence: 99%

Precipitation and Diffusivity of Arsenic in Silicon

Angelucci¹,

Celotti²,

Nobili³

et al. 1985

Small angle x-ray scattering measurements were performed on arsenic-implanted laser-annealed silicon slices to investigate the features of precipitation at 900~ for different dopant supersaturations. With decreasing As concentration from 3.2 to 1.1 • 10 ~l cm -3, the size distribution of the precipitates shifts toward larger values and their shape evolves from that of thin platelets, becoming three-dimensional and approaching at last that of a sphere. The high concentration diffusivity of As at 900~ deduced from the size of the largest precipitates in specimens doped 1 • 102' cm -3, turns out to be 2 • 10 -'4 cm2-s -~. Determinations of the shift of the carrier concentration profiles and of the junction depth, after heating at 1000~ show that the diffusivity of As is not affected by the precipitation and, hence, is controlled by the concentration of the dopant.The physical nature of the electrically inactive As, which is present in Si at high doping levels, is receiving increasing attention, primarily due to the demands of VLSI. Well-known models hypothesize the formation of complex point defects, or "clusters," in thermal equilibrium (1-6).Previous work (7) performed on ion-implanted and laser-annealed samples, doped with As in a wide composition range, put into evidence that the electrically active concentration depends only on temperature, a finding leading to the conclusion that a two-phase equilibrium takes place.The formation of a second phase was also indicated by the isochronal heating behavior, typical of a precipitation process. Precipitates were in fact detected in very heavily doped specimens by small angle x-ray scattering (SAXS), a technique which is suitable because of the noticeably different atomic numbers of Si and As. These determinations provided information on the size distribution and shape of these precipitates, which were found to be in the form of thin platelets. These particles should be coherent with the silicon matrix, as suggested by the absence of extra lines or spots in the diffraction patterns.As an additional result, the displacement of the carrier concentration profiles due to the heat-treatments showed that the diffusivity of the dopant strongly increased with the implanted dose. This effect, which was particularly evident at high temperature, could be attributed either to the increase of As concentration or to the increased amount of precipitates per unit surface, i.e., to lattice defects in excess generated by the precipitation process.The present work extends SAXS examinations to less heavily doped compositions (a difficult task owing to the small thickness of the implanted layer) with the aim of controlling the occurrence of precipitation at lower dopant concentrations and of getting further information on this process. In addition, it was intended to verify whether the precipitation process can influence the high temperature diffusivity of As. ExperimentalFor these experiments, we used (100)-oriented, CZpulled, p-type, boron-doped silicon slices of 1 t2-cm resistivity.The s...

“…Most of these authors suggest the formation of complex point defects or clusters in thermal equilibrium with the As in solution (1)(2)(3)(4)(5)(6). The existence of arsenic clusters, however, is presently not more than a hypothesis.…”

mentioning

confidence: 99%

Electron Microscopy of As Supersaturated Silicon

Armigliato¹,

Nobili²,

Solmi³

et al. 1986

Silicon specimens have been implanted with 1 • 1016 and 5 • 1016 As + ions/cm 2 and annealed with a pulsed ruby laser, thus giving rise to a supersaturated solution. Subsequent heat-treatments carried out in an inert ambient at either 450 ~ or 900~ led to a strong deactivation of the dopant. The physical nature of the inactive arsenic has been investigated by transmission electron microscopy (TEM), both in the weak beam and the high resolution imaging modes. For both doses and annealing temperatures, the presence of precipitates, having a diameter ranging from 1.5 to 3 nm, has been observed. Other defects, such as dislocations, perfect and faulted loops, and rod-like defects have also been detected in our TEM investigations. Despite the high density of the detected precipitates, the corresponding amount of arsenic atoms, as deduced by assuming an SiAs composition, can account for only a fraction of the inactive dopant. A possible explanation is that particles in the subnanometer range are invisible even in the high resolution images. This hypothesis is supported by the absence of contrast found in corresponding simulated high resolution images.Arsenic, due to the high solubility and low diffusivity, is the silicon n-type dopant more commonly used in VLSI circuits, both for the source and drain of MOS transistors and the emitter of bipolar transistors.Several works performed in recent years have investigated the physical nature of electrically inactive As, which is present in silicon at high doping levels. Most of these authors suggest the formation of complex point defects or clusters in thermal equilibrium with the As in solution (1-6). The existence of arsenic clusters, however, is presently not more than a hypothesis. Chu and Masters (7), from-angle-scanned channeling studies, reported results that are compatible with cluster formation, but there is not, up to now, direct evidence of the arsenic clusters.Recent work (8, 9) performed on ion-implanted and laser-annealed silicon, doped with As in a wide range of compositions, provided evidence that the electrically active concentration depends only on temperature, a finding leading to the conclusion that a two-phase equilibrium takes place. The formation of a second phase was also indicated by the isochronal heating behavior, showing a reverse annealing typical of a precipitation process (8). Precipitates were, in fact, detected by small angle x-ray scattering (SAXS), a technique that is suitable because of the noticeably different atomic numbers of Si and As. These determinations provided information on the size distribution and shape of these precipitates that was found to be in agreement with the classical theory of nucleation, which predicts a decrease in particle density and an increase in particle size as the supersaturation decreases (9). Moreover, most of these particles should be coherent with the silicon matrix, as expected from the RBS experiments performed in channeling conditions (7, 10), showing that the off-axis concentration is markedly lower than...