F. F. Morehead

Diffusion of zinc in gallium arsenide: A new model

Experimental results on the diffusion of Zn in GaAs which could not be satisfactorily explained in terms of a Frank–Turnbull mechanism involving vacancies can be understood with a ’’kick‐out model’’ in which the equilibrium between interstitial and substitutional Zn is established via gallium interstitials.

Shallow junctions by high‐dose As implants in Si: experiments and modeling

Shallow (<0.2 μm) n+ layers in Si with high conductivity (<40 Ω/⧠) have been formed by high‐dose (2×1016 cm−2) As implants. Experimental observations of As distributions and carrier concentrations are successfully simulated by a computer program which accounts for both the concentration dependent diffusion and As clustering effects. Reduction of electrical carriers in high‐dose As implanted Si during moderate temperature (∼800 ° C) heat treatments is readily explained by the kinetics of As clustering. Physical limitations on the conductivity which can be achieved by thermally annealed As implants in Si are also discussed.

Enhanced tail diffusion of phosphorus and boron in silicon: self-interstitial phenomena

It is well known that high surface concentration phosphorus diffusion leads to deeply penetrating ‘‘tails’’ in its concentration profile. At 700 °C the tail diffusivity exceeds that of low concentration phosphorus by a factor of 1000. Less spectacular, but very significant tailing also affects boron, making the conventional models contained in commonly available process simulation programs quite inaccurate for high concentrations of boron. We show that the observed tailing can be accounted for by a model whose central assumption is the local equality of dopant and oppositely directed defect fluxes.

FORMATION OF AMORPHOUS SILICON BY ION BOMBARDMENT AS A FUNCTION OF ION, TEMPERATURE, AND DOSE.

The dose (fluence) of 200‐keV boron, phosphorous, and antimony ions required to produce a continuous amorphous layer in silicon is determined as a function of target temperature. EPR measurements are used to monitor the process which is also then related to annealing effectiveness. The continuous amorphous layer recrystallizes at 550°C, after which only the implanted ions within that layer are completely electrically active. Carrier concentration profiles indicate the position of the amorphous layer and allow an approximate determination of the distribution with depth of damage. At the low dose rates used, reasonable agreement with a simple model for the formation of amorphous silicon as a function of ion, temperature, and dose is obtained.

Diffusion of gold in silicon: A new model

Experimental results from the diffusion of Au in Si, which could not be explained before, in terms of the Frank‐Turnbull mechanism involving vacancies, can be quantitatively understood with the assumption that the equilibrium between interstitial and substitutional Au is established via Si self‐interstitials. The results suggest that these defects dominate diffusion in Si.

ANNEALING CHARACTERISTICS OF n-TYPE DOPANTS IN ION-IMPLANTED SILICON.

Investigations of the conditions under which the donors, P, As, and Sb, are incorporated into Si by by ion implantation (260–300 keV) in an electrically active form are reported. Above a critical dose, room‐temperature implantations followed by a 600°C post anneal are substantially more effective than implantations at 600°C.

Self‐interstitial and vacancy contributions to silicon self‐diffusion determined from the diffusion of gold in silicon

We present an analysis of gold diffusion profiles in silicon taking into account that both self‐interstitials and vacancies are present at thermal equilibrium. We find that at 1000 °C the contribution of self‐interstitials to silicon self‐diffusion is about equal to that of vacancies.

Boron diffusion in silicon at high concentrations

Depth profiles measured by secondary ion mass spectrometry have been used to compare boron diffusion from three different sources for temperatures ranging from 850 to 1050 °C. The sources included boron in situ doped and ion‐implanted polycrystalline silicon as well as vapor using an evacuated capsule with highly doped powder. The junction depths and surface concentrations demonstrated little source dependence. Boltzmann–Matano analysis has been used to show that the concentration dependence of the diffusivity on source was minimal. We have clearly shown that conventional models of boron diffusion cannot fit the experimental data or the Boltzmann–Matano results, regardless of source. A new model has been used to describe the boron diffusion profiles more accurately.

Exact description and data fitting of ion‐implanted dopant profile evolution during annealing

We solve analytically the problem of dopant redistribution from an arbitrary initial implanted profile. The diffusivity, assumed concentration independent, can be an otherwise arbitrary function of temperature. Next, we derive a closed‐form expression for the annealed concentration distribution in the special case of an initial truncated Gaussian. Our solution is valid over a much wider range of experimental conditions than is Seidel and MacRae’s approximation [Trans. AIME 245, 491 (1969)]. We offer a simple, yet precise, criterion for the validity of this approximation, and we guard against its indiscriminate use. Last, we fit short‐time annealed P profiles in implanted Si to get an average diffusivity D≂3×10−12 cm2/s. We describe simple and accurate one‐parameter data‐fitting procedures.

The steady‐state model for coupled defect‐impurity diffusion in silicon

We extend and generalize our earlier model which was proposed to explain tails in the diffusion profiles of high‐concentration boron and phosphorus in silicon. Our quasi‐steady‐state approach is generalized here to include both vacancies (V) and interstitials (I) at equivalent levels. I‐V recombination is regarded as near local equilibrium, occurring through reactions of the defects with defect‐impurity pairs. This approach leads to important details such as the well‐known plateau, kink, and tail in high‐surface‐concentration P diffusions in Si, and to the less well‐recognized tails in B as well, with no additional ad hoc assumptions. Our extended model, in its simplest form, allows a more complete and less restrictive treatment than the usual models of Au in‐diffusion in Si. An important advantage is the direct inclusion of these defect‐impurity interactions and the resulting gradients in the defect concentrations, which are ignored in even the latest versions of popular process simulation programs, such...