Anomalous Impurity Diffusion in Epitaxial Silicon near the Substrate

Kahng, D.; Thomas, Candice; Manz, R. C.

doi:10.1149/1.2425248

Cited by 25 publications

(12 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This also means that any total removal of oxide in order to regrow a fresh oxide will result in an N-skin layer before the oxide can be regrown. This is in agreement with previous observations and models (1)(2)(3)(4). A protective oxide remaining over the p-regions surrounding the N + island is sufficient to prevent the phosphorus from penetrating to the p-region.…”

supporting

confidence: 93%

See 1 more Smart Citation

Evidence of Phosphorus N-skin on Silicon from Vapor Transport

Edwards

1969

J. Electrochem. Soc.

View full text Add to dashboard Cite

These stresses too are intrinsic and must depend on the deposition mechanism. Indeed they are dependent on the rf power density, but not noticeably on the substrate temperature. ConclusionsThe measurement of the thermal expansion coefficient of silicon nitride over a large temperature range shows it to be compatible with silicon. Efforts to relate residual stress in deposited films to thermal expansion coefficients must take into account intrinsic stresses produced during deposition. These stresses themselves may have temperature dependence that could make them difficult to separate from thermal expansion stresses.

show abstract

supporting

confidence: 93%

“…Also using a differential capacitance-voltage method for determining doping profiles in depth in epitaxial semiconductor films, Kahng, Thomas, and Manz (3,4) presented experimental results for a proposed model of transfer of dopant from the silicon backside (and the silicon pedestal used in the epitaxiaI process) into the working gas phase.…”

mentioning

confidence: 99%

Evidence of Phosphorus N-skin on Silicon from Vapor Transport

Edwards

1969

J. Electrochem. Soc.

View full text Add to dashboard Cite

show abstract

“…The distance xj is referred to in this paper as the junction lag. For a wide range of conditions, it may be much larger than that predicted from diffusion considerations alone, even allowing for anomalously high diffusion rates near the interface (10). For those situations where the process junction lag is greater than that predicted by diffusion, the value of xj is determined almost entirely by the growth process itself.…”

Section: Inix = [Nd--naix = ]N*o E-<~x--a(1--e-~x)[x [4]mentioning

confidence: 95%

“…The first term on the right is the substrate contribution to the film doping, and the second term on the right is the total gas phase contribution to the film doping. Equation [2] defines the film doping distribution, exclusive of a thin diffusion dominated region near the film substrate interface (10), in terms of three measurable system parameters, N*o, r and A. Previously, we have discussed the measurement of these three parameters and experimental determination of the doping profiles for n-type films on n-type substrates.…”

mentioning

confidence: 99%

Epitaxial Silicon Junctions

Kahng¹,

Thomas²,

Manz³

1963

J. Electrochem. Soc.

Self Cite

View full text Add to dashboard Cite

In a preceding paper, the effect of the epitaxial silicon growth process on impurity distribution in the epitaxial film was treated in terms of a countercurrent model, and experimental results were presented for n-type films on n-type substrates. In the present paper, additional experimental data are presented for p-on-p, n-on-p, and p-on-n growth conditions.For n-on-p, or p-on-n growth, the electrical junctions are shifted by the process a distance of xj microns (the junction lag) from the substrate-film interface. The process-dominated junction lag may be much larger than that predicted from diffusion considerations. It is dependent on the film growth temerature and the doping levels in the substrate and in the gas phase. As predicted by the countercurrent distribution expression, it is greater for lower growth temperatures.Junction lags from immeasurably small values up to greater than 30~ have been generated. There is quantitative agreement between the observed magnitude of the junction lag and the value predicted by use of the countercurrent distribution expression. Also, for those samples measured electrically (with junction lags of 1.8-4.8~) the agreement between the measured mean impurity concentration level (by four-point probe measurements) in the epitaxial layer and the mean impurity level predicted on the basis of the distribution expression is quantitative.

show abstract

“…Kahng et al [122] discovered that the boundary rcgion of epitaxial silicon films grown on heavily arsenic-doped Si substrates (e = 5 x Qcm) contains a high density of macrodefects. The impurity diffusion is characterized here by an anomalously high rate while in the remaining part of the film it is close to the values typical for the bulk material.…”

Section: Diffusion-induced Redistribution Of Impurities and Structuramentioning

confidence: 99%